GE regulations

4000-01-U

DEPARTMENT OF EDUCATION

34 CFR Parts 600 and 668

RIN 1840-AD15

Docket ID ED-2014-OPE-0039

Program Integrity: Gainful Employment

AGENCY: Office of Postsecondary Education, Department of Education.

ACTION: Final regulations.

SUMMARY: The Secretary amends regulations on institutional eligibility under the Higher Education Act of 1965, as amended (HEA), and the Student Assistance General Provisions to establish measures for determining whether certain postsecondary educational programs prepare students for gainful employment in a recognized occupation, and the conditions under which these educational programs remain eligible under the Federal Student Aid programs authorized under title IV of the HEA (title IV, HEA programs).

DATES: These regulations are effective July 1, 2015.

FOR FURTHER INFORMATION CONTACT: John Kolotos, U.S. Department of Education, 1990 K Street NW., room 8018, Washington, DC 20006-8502. Telephone: (202) 502-7762 or by email at:  [email protected].

If you use a telecommunications device for the deaf (TDD) or a text telephone (TTY), call the Federal Relay Service (FRS), toll free, at 1-800-877-8339.

SUPPLEMENTARY INFORMATION:

Executive Summary:

Purpose of This Regulatory Action: The regulations are intended to address growing concerns about educational programs that, as a condition of eligibility for title IV, HEA program funds, are required by statute to provide training that prepares students for gainful employment in a recognized occupation (GE programs), but instead are leaving students with unaffordable levels of loan debt in relation to their earnings, or leading to default. GE programs include nearly all educational programs at for-profit institutions of higher education, as well as non-degree programs at public and private non-profit institutions such as community colleges.

Specifically, the Department is concerned that a number of GE programs: (1) do not train students in the skills they need to obtain and maintain jobs in the occupation for which the program purports to provide training, (2) provide training for an occupation for which low wages do not justify program costs, and (3) are experiencing a high number of withdrawals or “churn” because relatively large numbers of students enroll but few, or none, complete the program, which can often lead to default. We are also concerned about the growing evidence, from Federal and State investigations and qui tam lawsuits, that many GE programs are engaging in aggressive and deceptive marketing and recruiting practices. As a result of these practices, prospective students and their families are potentially being pressured and misled into critical decisions regarding their educational investments that are against their interests.

For these reasons, through this regulatory action, the Department establishes: (1) an accountability framework for GE programs that defines what it means to prepare students for gainful employment in a recognized occupation by establishing measures by which the Department will evaluate whether a GE program remains eligible for title IV, HEA program funds, and (2) a transparency framework that will increase the quality and availability of information about the outcomes of students enrolled in GE programs. Better outcomes information will benefit: students, prospective students, and their families, as they make critical decisions about their educational investments; the public, taxpayers, and the Government, by providing information that will enable better protection of the Federal investment in these programs; and institutions, by providing them with meaningful information that they can use to help improve student outcomes in their programs.

The accountability framework defines what it means to prepare students for gainful employment by establishing measures that assess whether programs provide quality education and training to their students that lead to earnings that will allow students to pay back their student loan debts. For programs that perform poorly under the measures, institutions will need to make improvements during the transition period we establish in the regulations.

The transparency framework will establish reporting and disclosure requirements that increase the transparency of student outcomes of GE programs so that students, prospective students, and their families have accurate and comparable information to help them make informed decisions about where to invest their time and money in pursuit of a postsecondary degree or credential. Further, this information will provide the public, taxpayers, and the Government with relevant information to better safeguard the Federal investment in these programs. Finally, the transparency framework will provide institutions with meaningful information that they can use to improve student outcomes in these programs.

Authority for This Regulatory Action: To accomplish these two primary goals of accountability and transparency, the Secretary amends parts 600 and 668 of title 34 of the Code of Federal Regulations (CFR). The Department’s authority for this regulatory action is derived primarily from three sources, which are discussed in more detail in “§668.401 Scope and Purpose” and in the notice of proposed rulemaking (NPRM) published on March 25, 2014 (79 FR 16426). First, sections 101 and 102 of the HEA define an eligible institution, as pertinent here, as one that provides an “eligible program of training to prepare students for gainful employment in a recognized occupation.” 20 U.S.C. 1001(b)(1), 1002(b)(1)(A)(i), (c)(1)(A). Section 481(b) of the HEA defines “eligible program” to include a program that “provides a program of training to prepare students for gainful employment in a recognized profession.” 20 U.S.C. 1088(b). Briefly, this authority establishes the requirement that certain educational programs must provide training that prepare students for gainful employment in a recognized occupation in order for those programs to be eligible for title IV, HEA program funds -- the requirement that the Department defines through these regulations.

Second, section 410 of the General Education Provisions Act provides the Secretary with authority to make, promulgate, issue, rescind, and amend rules and regulations governing the manner of operations of, and governing the applicable programs administered by, the Department. 20 U.S.C. 1221e-3. Furthermore, under section 414 of the Department of Education Organization Act, the Secretary is authorized to prescribe such rules and regulations as the Secretary determines necessary or appropriate to administer and manage the functions of the Secretary or the Department. 20 U.S.C. 3474. These authorities, together with the provisions in the HEA, thus include promulgating regulations that, in this case: set measures to determine the eligibility of GE programs for title IV, HEA program funds; require institutions to report information about the program to the Secretary; require the institution to disclose information about the program to students, prospective students, and their families, the public, taxpayers, and the Government, and institutions; and establish certification requirements regarding an institution’s GE programs.

As also explained in more detail in “§668.401 Scope and Purpose” and the NPRM, the Department’s authority for the transparency framework is further supported by section 431 of the Department of Education Organization Act, which provides authority to the Secretary, in relevant part, to inform the public regarding federally supported education programs; and collect data and information on applicable programs for the purpose of obtaining objective measurements of the effectiveness of such programs in achieving the intended purposes of such programs. 20 U.S.C. 1231a.

The Department’s authority for the regulations is also informed by the legislative history of the provisions of the HEA, as discussed in the NPRM, as well as the rulings of the U.S. District Court for the District of Columbia in Association of Private Sector Colleges and Universities v. Duncan, 870 F.Supp.2d 133 (D.D.C. 2012), and 930 F.Supp.2d 210 (D.D.C. 2013) (referred to in this document as “APSCU v. Duncan). Notably, the court specifically considered the Department’s authority to define what it means to prepare students for gainful employment and to require institutions to report and disclose relevant information about their GE programs.

Summary of the Major Provisions of This Regulatory Action: As discussed under “Purpose of This Regulatory Action,” the regulations establish an accountability framework and a transparency framework.

The accountability framework, among other things, creates a certification process by which an institution establishes a GE program’s eligibility for title IV, HEA program funds, as well as a process by which the Department determines whether a program remains eligible. First, an institution establishes the eligibility of a GE program by certifying, among other things, that the program is included in the institution’s accreditation and satisfies any applicable State or Federal program-level accrediting requirements and State licensing and certification requirements for the occupations for which the program purports to prepare students to enter. This requirement will serve as a baseline protection against the harm that students could experience by enrolling in programs that do not meet all State or Federal accrediting standards and licensing or certification requirements necessary to secure the jobs associated with the training.

Under the accountability framework, we also establish the debt-to-earnings (D/E) rates measure¹ that will be used to determine whether a GE program remains eligible for title IV, HEA program funds. The D/E rates measure evaluates the amount of debt (tuition and fees and books, equipment, and supplies) students who completed a GE program incurred to attend that program in comparison to those same students’ discretionary and annual earnings after completing the program. The regulations establish the standards by which the program will be assessed to determine, for each year rates are calculated, whether it passes or fails the D/E rates measure or is “in the zone.”

Under the regulations, to pass the D/E rates measure, the GE program must have a discretionary income rate² less than or equal to 20 percent or an annual earnings rate³ less than or equal to 8 percent. The regulations also establish a zone for GE programs that have a discretionary income rate greater than 20 percent and less than or equal to 30 percent or an annual earnings rate greater than 8 percent and less than or equal to 12 percent. GE programs with a discretionary income rate over 30 percent and an annual earnings rate over 12 percent will fail the D/E rates measure. Under the regulations, a GE program becomes ineligible for title IV, HEA program funds, if it fails the D/E rates measure for two out of three consecutive years, or has a combination of D/E rates that are in the zone or failing for four consecutive years. We establish the D/E rates measure and the thresholds, as explained in more detail in “§668.403 Gainful Employment Framework,” to assess whether a GE program has indeed prepared students to earn enough to repay their loans, or was sufficiently low cost, such that students are not unduly burdened with debt, and to safeguard the Federal investment in the program.

The regulations also establish procedures for the calculation of the D/E rates and for challenging the information used to calculate the D/E rates and appealing the determination. The regulations also establish a transition period for the first seven years after the regulations take effect to allow institutions to pass the D/E rates measure by reducing the loan debt of currently enrolled students.

For a GE program that could become ineligible based on its D/E rates for the next award year, the regulations require the institution to warn students and prospective students of the potential loss of eligibility for title IV, HEA program funds and the implications of such loss of eligibility. Specifically, institutions would be required to provide warnings to enrolled students that describe, among other things, the options available to continue their education at the institution if the program loses its eligibility and whether the students will be able to receive a refund of tuition and fees. The regulations also provide that, for a GE program that loses eligibility or for any failing or zone program that is discontinued by the institution, the loss of eligibility is for three calendar years.

These provisions will: ensure that institutions have a meaningful opportunity and reasonable time to improve their programs for a period of time after the regulations take effect, and ensure that those improvements are reflected in the D/E rates; protect students and prospective students and ensure that they are informed about programs that are failing or could potentially lose eligibility; and provide institutions and other interested parties with clarity as to how the calculations are made, how institutions can ensure the accuracy of information used in the calculations, and the consequences of failing the D/E rates measure and losing eligibility.

In addition, the regulations establish a transparency framework. First, the regulations establish reporting requirements, under which institutions will report information related to their GE programs to the Secretary. The reporting requirements will facilitate the Department’s evaluation of the GE programs under the accountability framework, as well as support the goals of the transparency framework. Second, the regulations require institutions to disclose relevant information and data about the GE programs through a disclosure template developed by the Secretary. The disclosure requirements will help ensure students, prospective students, and their families, the public, taxpayers, and the Government, and institutions have access to meaningful and comparable information about student outcomes and the overall performance of GE programs.

Costs and Benefits:  There are two primary benefits of the regulations. Because the regulations establish an accountability framework that assesses program performance we expect students, prospective students, taxpayers, and the Federal Government to receive a better return on the title IV, HEA program funds. The regulations also establish a transparency framework which will improve market information that will assist students, prospective students, and their families in making critical decisions about their educational investment and in understanding potential outcomes of that investment. The public, taxpayers, the Government, and institutions will also gain relevant and useful information about GE programs, allowing them to evaluate their investment in these programs. Institutions will largely bear the costs of the regulations: paperwork costs of complying with the regulations, costs that could be incurred by institutions if they attempt to improve their GE programs, and costs due to changing student enrollment. See “Discussion of Costs, Benefits, and Transfers” in the regulatory impact analysis in Appendix A to this document for a more complete discussion of the costs and benefits of the regulations.

On March 25, 2014, the Secretary published the NPRM for these regulations in the Federal Register (79 FR 16426). In the preamble of the NPRM, we discussed on pages 16428-16433, the background of the regulations, the relevant data available, and the major changes proposed in that document. Terms used but not defined in this document, for example, 2011 Prior Rule and 2011 Final Rules, have the meanings set forth in the NPRM. The final regulations contain a number of changes from the NPRM. We fully explain the changes in the Analysis of Comments and Changes section of the preamble that follows.

Public Comment: In response to our invitation in the NPRM, we received approximately 95,000 comments on the proposed regulations. We discuss substantive issues under the sections of the proposed regulations to which they pertain. Generally, we do not address technical or other minor changes.

Analysis of Comments and Changes: An analysis of the comments and of any changes in the regulations since publication of the NPRM follows.

§668.401 Scope and Purpose

Comments: A number of commenters stated that, in promulgating the regulations, the Department exceeds its delegated authority to administer programs under the HEA. Some commenters asserted that the legislative history of the gainful employment provisions in the HEA does not support the Department’s regulatory action to define gainful employment and that the Department gave undue weight to testimony presented to Congress at the time the gainful employment provisions were enacted. Some commenters stated that Congress did not intend for the Department to measure whether a program leads to gainful employment based on debt or earnings.

Several commenters argued that, even if the Department has the legal authority, the issues addressed by the regulations should be addressed instead as a part of HEA reauthorization or by other legislative action. One commenter contended that members of Congress have asked the Department to refrain from regulating on gainful employment programs pending reauthorization of the HEA and that the proposed regulations constitute a usurping of legislative authority.

Other commenters asserted that identifying educational programs in the career training sector that do not prepare students for gainful employment and terminating their eligibility for title IV, HEA program funds is mandated by the HEA.

Discussion: The Department’s statutory authority for this regulatory action is derived primarily from three sources. First, sections 101 and 102 of the HEA define “eligible institution” to include an institution that provides an “eligible program of training to prepare students for gainful employment in a recognized occupation.” 20 U.S.C. 1001(b)(1), 1002(b)(1)(A)(i), (c)(1)(A). Section 481(b) of the HEA defines “eligible program” to include a program that “provides a program of training to prepare students for gainful employment in a recognized profession.” 20 U.S.C. 1088(b). These statutory provisions establish the requirement that certain educational programs must provide training that prepares students for gainful employment in a recognized occupation in order for those programs to be eligible for title IV, HEA program funds--the requirement that the Department seeks to define through the regulations.

Second, section 410 of the General Education Provisions Act provides the Secretary with authority to make, promulgate, issue, rescind, and amend rules and regulations governing the manner of operations of, and governing the applicable programs administered by, the Department. 20 U.S.C. 1221e-3. Furthermore, under section 414 of the Department of Education Organization Act, the Secretary is authorized to prescribe such rules and regulations as the Secretary determines necessary or appropriate to administer and manage the functions of the Secretary or the Department. 20 U.S.C. 3474. These provisions, together with the provisions in the HEA regarding GE programs, authorize the Department to promulgate regulations that: set measures to determine the eligibility of GE programs for title IV, HEA program funds; require institutions to report information about GE programs to the Secretary; require institutions to disclose information about GE programs to students, prospective students, and their families, the public, taxpayers, and the Government, and institutions; and establish certification requirements regarding an institution’s GE programs.

Third, the Department’s authority for establishing the transparency framework is further supported by section 431 of the Department of Education Organization Act, which provides authority to the Secretary, in relevant part, to inform the public about federally supported education programs and collect data and information on applicable programs for the purpose of obtaining objective measurements of the effectiveness of such programs in achieving the intended purposes of such programs. 20 U.S.C. 1231a.

The U.S. District Court for the District of Columbia confirmed the Department’s authority to regulate gainful employment programs in Association of Private Sector Colleges and Universities (APSCU) v. Duncan, 870 F.Supp.2d 133 (D.D.C. 2012), and 930 F.Supp.2d 210 (D.D.C. 2013). These rulings arose out of a lawsuit brought by APSCU challenging the Department’s 2010 and 2011 gainful employment regulations. In that case, the court reached several conclusions about the Department’s rulemaking authority to define eligibility requirements for gainful employment programs that have informed and framed the Department’s exercise of that authority through this rulemaking. Notably, the court agreed with the Department that the Secretary has broad authority to make, promulgate, issue, rescind, and amend the rules and regulations governing applicable programs administered by the Department, such as the title IV, HEA programs, and that the Secretary is “authorized to prescribe such rules and regulations as the Secretary determines necessary or appropriate to administer and manage the functions of the Secretary or the Department.” APSCU v. Duncan, 870 F.Supp.2d at 141; see 20 U.S.C. 3474. Furthermore, in answering the question of whether the Department’s regulatory effort to define the gainful employment requirement falls within its statutory authority, the court found that the Department’s actions were within its statutory authority to define the gainful employment requirement. Specifically, the court concluded that the phrase “gainful employment in a recognized occupation” is ambiguous; in enacting a requirement that used that phrase, Congress delegated interpretive authority to the Department; and the Department’s regulations were a reasonable interpretation of an ambiguous statutory command. APSCU v. Duncan, 870 F.Supp.2d at 146-49. The court also upheld the disclosure requirements set forth by the Department in the 2011 Final Rule, which are still in effect, rejecting APSCU’s challenge and finding that these requirements “fall comfortably within [the Secretary’s] regulatory power,” and are “not arbitrary or capricious.” Id. at 156.

Contrary to the claims of some commenters, the Department’s authority to promulgate regulations defining the gainful employment requirement and using a debt and earnings measure for that purpose is also supported by the legislative history of the statutory provisions regarding gainful employment programs. The legislative history of the statute preceding the HEA that first permitted students to obtain federally financed loans to enroll in programs that prepared them for gainful employment in recognized occupations demonstrates the conviction that the training offered by these programs should equip students to earn enough to repay their loans. APSCU v. Duncan, 870 F.Supp.2d at 139. Allowing these students to borrow was expected to neither unduly burden the students nor pose “a poor financial risk” to taxpayers. Specifically, the Senate Report accompanying the initial legislation (the National Vocational Student Loan Insurance Act (NVSLIA), Pub. L. 89-287) quotes extensively from testimony provided by University of Iowa professor Dr. Kenneth B. Hoyt, who testified on behalf of the American Personnel and Guidance Association. On this point, the Senate Report sets out Dr. Hoyt’s questions and conclusions:

Would these students be in a position to repay loans following their training? . . .

If loans were made to these kinds of students, is it likely that they could repay them following training? Would loan funds pay dividends in terms of benefits accruing from the training students received? It would seem that any discussion concerning this bill must address itself to these questions. . . . .

We are currently completing a second-year followup of these students and expect these reported earnings to be even higher this year. It seems evident that, in terms of this sample of students, sufficient numbers were working for sufficient wages so as to make the concept of student loans to be [repaid] following graduation a reasonable approach to take. . . . I have found no reason to believe that such funds are not needed, that their availability would be unjustified in terms of benefits accruing to both these students and to society in general, nor that they would represent a poor financial risk.

Sen. Rep. No. 758 (1965) at 3745, 3748-49 (emphasis added).

Notably, both debt burden to the borrower and financial risk to taxpayers and the Government were clearly considered in authorizing federally backed student lending. Under the loan insurance program enacted in the NVSLIA, the specific potential loss to taxpayers of concern was the need to pay default claims to banks and other lenders if the borrowers defaulted on the loans. After its passage, the NVSLIA was merged into the HEA, which in title IV, part B, has both a direct Federal loan insurance component and a Federal reinsurance component that require the Federal Government to reimburse State and private non-profit loan guaranty agencies upon their payment of default claims. 20 U.S.C. 1071(a)(1). Under either HEA component, taxpayers and the Government assume the direct financial risk of default. 20 U.S.C. 1078(c) (Federal reinsurance for default claim payments), 20 U.S.C. 1080 (Federal insurance for default claims). We therefore disagree that the legislative history does not support the Department’s action here nor do we see any basis, and commenters have provided none, for us to question that history or the information Congress relied upon in enacting the statutory provisions.

We appreciate that Congress may have a strong interest in addressing the issues addressed by these regulations in the reauthorization of the HEA or other legislation and we look forward to working with Congress on its legislative proposals. However, we do not agree that the Department should not take, or should defer, regulatory action on this basis until Congress reauthorizes the HEA or takes other action. In light of the numerous concerns about the poor outcomes of students attending many GE programs, and the risk that poses to the Federal interest, the Department must proceed now in accordance with its statutory authority, as delegated by Congress, to protect students and taxpayers.

Changes: None.

Comments: Some commenters suggested that the phrase “to prepare students for gainful employment” is unambiguous and therefore not subject to further interpretation. Commenters stated that the Department's interpretation of the phrase is incorrect because it is contrary to the ordinary meaning of the phrase “gainful employment,” to congressional intent, and to the rules of statutory construction. These commenters asserted that the dictionary definition of the phrase does not comport with the Department’s proposed definition or the definition of the term “gainful employment” in other provisions of the HEA. Commenters also stated that Congress has not made any changes to the HEA triggering a requirement by the Secretary to define the term “gainful employment” and claimed that the term cannot now be defined since Congress left it undisturbed during its periodic reauthorizations of the HEA.

Some commenters expressed the view that the framework of detailed program requirements under title IV of the HEA, including institutional cohort default rates, institutional disclosure requirements, restrictions on student loan borrowing, and other financial aid requirements, prevents the Department from adopting debt measures to determine whether a gainful employment program is eligible to receive title IV, HEA program funds.

One commenter claimed that the Department has previously defined the phrase “gainful employment in a recognized occupation” in the context of conducting administrative hearings and argued that the Department did not adequately explain in the NPRM why it was departing from its prior use of the term.

Discussion: As the court found in APSCU v. Duncan, Congress has not spoken through legislative action to the precise question at issue here: whether the statutory requirement that programs providing vocational training “prepare students for gainful employment in a recognized occupation” may be measured by reference to students’ ability to repay their loans. Congress did not provide a definition for the phrase “gainful employment” or “gainful employment in a recognized occupation” in either the statute or its legislative history. Thus, the phrase is ambiguous and Congress left further definition of the phrase to the Department.

There also is no common meaning of the phrase, contrary to the assertion of the commenters. The commenters’ argument that “gainful employment” has one meaning in all circumstances--“a job that pays”--is belied by other dictionaries that define “gainful” as “profitable.” See, e.g., Webster’s New Collegiate Dictionary 469 (1975). “Profitable” means the excess of returns over expenditures, or having something left over after one’s expenses are paid. Id. at 919. This definition supports the idea embodied in the regulations that “gainful employment in a recognized occupation” is not just any job that pays a nominal amount but a job that pays enough to cover one’s major expenses, including student loans.

Nor is there a common definition of the phrase in the HEA. Although Congress used the words “gainful employment” in other provisions of the HEA, the operative phrase for the purpose of these regulations is “gainful employment in a recognized occupation.” The modifying words “in a recognized occupation” qualify the type of job for which students must be prepared. “A recognized occupation” suggests an established occupation, not just any job that pays. In addition, the phrase “gainful employment” means different things based on its context in the statute. For example, the requirement that a recipient of a graduate fellowship not be “engaged in gainful employment, other than part-time employment related to teaching, research, or a similar activity” (20 U.S.C. 1036(e)(1)(B)(ii)) has a different meaning than the requirement that vocationally oriented programs “prepare students for gainful employment in a recognized occupation,” just as both requirements necessarily have a different meaning than a statutory requirement that a program for students with disabilities focus on skills that lead to “gainful employment” (20 U.S.C. 1140g(d)(3)(D)).

As the court stated in APSCU v. Duncan, “[t]he power of an administrative agency to administer a congressionally created ... program necessarily requires the formulation of policy and the making of rules to fill any gap left, implicitly or explicitly, by Congress. The means of determining whether a program ‘prepare[s] students for gainful employment in a recognized occupation’ is a considerable gap, which the Department has promulgated rules to fill.”  APSCU v. Duncan, 870 F. Supp. 2d 133, 146 (D.D.C. 2012) (internal quotations and citations omitted).

The commenters are incorrect in their assertion that the HEA’s provisions on loan default rates, student borrowing, and other financial aid matters prevent the Department from regulating on what it means for a program to provide training that prepares students for gainful employment in a recognized occupation. The Department’s regulations are not an attempt to second guess Congress or depart from a congressional plan but rather will fill a gap that Congress left in the statute--defining what it means to prepare a student for gainful employment in a recognized occupation--in a manner consistent with congressional intent. The regulations supplement and complement the statutory scheme. And, although there are differences between the regulations and other provisions, such as those regarding institutional cohort default rates (CDR), the regulations do not fundamentally alter the statutory scheme.

Rather than conflicting, as asserted by commenters, the CDR and GE regulations complement each other. Congress enacted the CDR provision as “one” mechanism--not the sole, exclusive mechanism--for dealing with abuses in Federal student aid programs. See H.R. Rep. No. 110-500 at 261 (2007) (“Over the years, a number of provisions have been enacted under the Higher Education Act to protect the integrity of the federal student aid programs. One effective mechanism was to restrict federal loan eligibility for students at schools with very high cohort loan default rates” (emphasis added).) Congress did not, in enacting the CDR provision or at any other time, limit the Department’s authority to promulgate regulations to define what it means to “prepare students for gainful employment in a recognized occupation.” Compare 20 U.S.C. 1015b(i), concerning student access to affordable course materials (“No regulatory authority. The Secretary shall not promulgate regulations with respect to this section.”). Nor did it alter this existing statutory language when it passed the CDR provision. Indeed, the court in APCSU v. Duncan specifically addressed the issue of whether the CDR provisions would preclude the Department from effectuating the gainful employment requirement by relying on other debt measures at the programmatic level and concluded that the “statutory cohort default rule . . . does not prevent the Department from adopting the debt measures.” APSCU v. Duncan, 870 F. Supp. 2d at 147 (citing to Career Coll. Ass’n v. Riley, 74 F.3d 1265, 1272-75 (D.C. Cir. 1996), where the D.C. Circuit held that the Department’s authority to establish “‘reasonable standards of financial responsibility and appropriate institutional capability’ empowers it to promulgate a rule that measures an institution’s administrative capability by reference to its cohort default rate--even though the administrative test differs significantly from the statutory cohort default rate test.”)

The GE regulations are also consistent with other provisions of the HEA aimed at curbing abuses in the title IV, HEA programs. Prompted by a concern that its enormous commitment of Federal resources would be used to provide financial aid to students who were unable to find jobs that would allow them to repay their loans, Congress enacted several statutory provisions to ensure against abuse. Congress specified that participating schools cannot “provide any commission, bonus, or other incentive payment based directly or indirectly on success in securing enrollments or financial aid to any persons or entities engaged in any student recruiting or admission activities or in making decisions regarding the award of student financial assistance.” 20 U.S.C. 1094(a)(20). “The concern is that recruiters paid by the head are tempted to sign up poorly qualified students who will derive little benefit from the subsidy and may be unable or unwilling to repay federally guaranteed loans.” United States ex rel. Main v. Oakland City Univ., 426 F.3d 914, 916 (7th Cir. 2005). To prevent schools from improperly inducing people to enroll, Congress prohibited participating schools from engaging in a “substantial misrepresentation of the nature of its educational program, its financial charges, or the employability of its graduates.” 20 U.S.C. 1094(c)(3)(A). Congress also required a minimum level of State oversight of eligible schools.

In sum, the GE regulations simply build upon the Department’s regulation of institutions participating in the title IV, HEA programs and the myriad ways in which the Department, as authorized by Congress, protects students and taxpayers from abuse of the Federal student aid program.

We further disagree that the Department has previously defined what “gainful employment in a recognized occupation” means for the purpose of establishing accountability and transparency with respect to GE programs and their outcomes. In support of this argument, the commenters rely on a 1994 decision of an administrative law judge regarding whether a program in Jewish culture prepared students enrolled in the program for gainful employment in a recognized occupation. As the district court noted, the administrative law judge did not fully decide what it means to prepare a student for gainful employment in a recognized occupation but merely stated that any preparation must be for a specific area of employment. APSCU v. Duncan, 870 F. Supp. 2d 133, 150 (D.D.C. 2012). Further, the Department did not depart from the administrative law judge’s interpretation in the 2011 Final Rules, as the court in APSCU v. Duncan agreed. See id. Nor is the Department departing from that interpretation with these regulations.

Changes: None.

Comments: Some commenters claimed that the proposed regulations violate the HEA because they would require an institution to ensure a student is gainfully employed in a recognized occupation. The commenters stated that the HEA requires only that vocational schools “prepare” students for gainful employment in a recognized occupation and not that they ensure they obtain such employment. Commenters also stated that the HEA does not hold institutions responsible for a student’s post-graduation employment choices but the proposed regulations would. The commenters stated that under the proposed regulations, an institution would be penalized if a student chose not to seek gainful employment after graduation or chose to seek employment in another field that did not result in sufficient earnings to repay their debt.

Discussion: The commenters ignore the legislative history demonstrating that, in enacting the gainful employment statutory provisions, Congress intended that students who borrowed Federal funds to obtain such training would be able to repay the debt incurred because they would have been prepared for gainful employment in a recognized occupation. Contrary to commenters’ claims, the D/E rates measure the Department adopts here neither requires a school to ensure that an individual student obtains employment nor holds schools responsible for a student’s career decisions. Rather, the measure evaluates whether a particular cohort of students completing a program has received training that prepares those students for gainful employment such that they are able to repay their student loans, not whether each student who completed the program obtains a job that enables that student to pay back his or her loans.

Changes: None.

Comments: One commenter asked how the Department defines “recognized occupation.” According to the commenter, this question is of particular concern for schools offering cosmetology programs. The commenter said that there are many individuals who use their cosmetology degrees to obtain employment in a field that is indirectly related, such as beauty school administration. The commenter stated that some companies frequently hire beauty school graduates to work in their financial and student advisor offices; these students do not possess degrees in finance, career counseling, or administration, but their background and education in cosmetology has been found to be sufficient to properly fulfill the job requirements. The commenter asked whether these indirectly related jobs would be considered a recognized occupation.

Discussion: The proposed and final regulations in §600.2 define recognized occupation as an occupation that is either (a) identified by a Standard Occupational Classification (SOC) code established by OMB or an Occupational Information Network O*Net-SOC established by the Department of Labor or (b) determined by the Secretary in consultation with the Secretary of Labor to be a recognized occupation. Institutions are expected to identify a CIP code for their programs that represents the occupations for which the institution has designed its program. The Bureau of Labor Statistics (BLS) has developed a crosswalk that identifies the occupations (SOCs) associated with the education and training provided by a program (www.onetonline.org/crosswalk), and these would be “recognized occupations” for the purposes of these regulations. However, regardless of whether an occupation is associated with a particular program so long as the occupation is identified by a SOC code, it is a recognized occupation.

Changes: None.

Comments: Some commenters claimed that the proposed regulations would require institutions to lower their tuition in order to meet the D/E rates measure. Referencing a House of Representatives committee report from 2005, the commenter stated that this was contrary to Congress’ decision not to regulate institutions’ tuition. One commenter stated that the proposed regulations attempt to address the costs of deferments and other repayment options, but that Congress has already created mechanisms to address the issue of increasing student debt load and rising tuition costs. The commenter claimed that the proposed regulations would require institutions to reduce tuition and therefore are contrary to congressional action in this area.

Discussion: The regulations do not require institutions to lower their tuition. Reducing tuition and fees may be one way for an institution to meet the D/E rates measure but it is not the only way. Institutions can also meet the D/E rates measure by having high-quality program curricula and engaging in robust efforts to place students.

The regulations also are not contrary to Congress’ findings in H.R. Rep. 109-231. That report states “[i]t is the Committee’s position that . . . the Federal Government does not have the ability to set tuition and fee rates for colleges and universities.” H.R. Rep. 109-231, at 159 (emphasis added). Given that these regulations do not “set tuition and fee rates for colleges and universities,” there is no conflict with the congressional findings in this report.

Changes: None.

Comments: Several commenters contended that the Department failed to satisfy its obligations under the Administrative Procedure Act in conducting negotiated rulemaking. Specifically, the commenters asserted that representatives of for-profit institutions and business and industry, as well as representatives from law, medical, and other professional schools, were not adequately represented on the negotiating committee. They further argued that the Department did not listen to the views of negotiators during the negotiated rulemaking sessions. Some commenters stated that the Department did not conduct the negotiations in good faith because the negotiation sessions were held for seven days when other negotiated rulemaking sessions have taken longer.

Discussion: The negotiated rulemaking process ensures that a broad range of interests is considered in the development of regulations. Specifically, negotiated rulemaking seeks to enhance the rulemaking process through the involvement of all parties who will be significantly affected by the topics for which the regulations will be developed. Accordingly, section 492(b)(1) of the HEA, 20 U.S.C. 1098a(b)(1), requires the Department to choose negotiators from groups representing many different constituencies. The Department selects individuals with demonstrated expertise or experience in the relevant subjects under negotiation, reflecting the diversity of higher education interests and stakeholder groups, large and small, national, State, and local. In addition, the Department selects negotiators with the goal of providing adequate representation for the affected parties while keeping the size of the committee manageable. The statute does not require the Department to select specific entities or individuals to be on the committee. As there was a committee member representing each of for-profit institutions and business and industry interests, we do not agree that these groups were not adequately represented on the committee. We also do not agree that specific areas of training, such as law and medicine, required specific representation, as institutions with such programs were represented at the sector level.

While it is to be expected that some committee members will have interests that differ from other members and that consensus is not always reached, as in the case of these regulations, the negotiated rulemaking process is intended to provide stakeholders an opportunity to present alternative ideas, to identify areas where compromises can be reached, and to help inform the agency’s views. In the negotiated rulemaking sessions for these regulations, there was robust discussion of the draft regulations, negotiators including those representing the commenters submitted a number of proposals for the committee to consider, and, as we described in detail in the NPRM, the views and suggestions of negotiators informed the proposed and these final regulations.

With respect to the length of the negotiations, the HEA does not require negotiated rulemaking sessions to be held for a minimum number of days. Seven days was a sufficient amount of time to conduct these negotiations.

Changes: None.

Comments: A number of commenters stated that the proposed regulations were arbitrary and capricious and therefore violate the Administrative Procedure Act. Commenters raised this concern both generally and with respect to specific elements of the proposed regulations. For example, several commenters argued that the thresholds for the D/E rates measure lack a reasoned basis. As another example, some commenters claimed that the Department was arbitrary and capricious in proposing regulations that were different from those promulgated in the 2011 Final Rules.

Discussion: We address commenters’ arguments with respect to specific provisions of the regulations in the sections of this preamble specific to those provisions. However, as a general matter, in taking this regulatory action, we have considered relevant data and factors, considered and responded to comments, and articulated a reasoned basis for our actions. Marsh v. Oregon Natural Res. Council, 490 U.S. 360, 378 (1989); Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983); see also Pub. Citizen, Inc. v. Fed. Aviation Admin., 988 F.2d 186, 197 (D.C.Cir.1993); PPL Wallingford Energy LLC v. FERC, 419 F.3d 1194, 1198 (D.C.Cir.2005). Further, for those provisions of the regulations that differ from those established in the 2011 Final Rules, we have provided a reasoned basis for our departure from prior policy. Motor Vehicle, 463 U.S. at 57; see also Williams Gas Processing–Gulf Coast Co., L.P. v. FERC, 475 F.3d 319, 326 (D.C.Cir.2006); Rust v. Sullivan, 500 U.S. 173, 187 (1991); F.C.C. v. Fox Television Stations, Inc., 556 U.S. 502, 514-516 (2009); Investment Co. Inst. v. Commodity Futures Trading Comm’n, 720 F.3d 370, 376 (D.C. Cir. 2013). 

Changes: None.

Comments: Various commenters argued that the regulations are impermissibly retroactive. These commenters contended that the accountability metrics reflect historical performance and not current program performance and, at least initially, would apply standards to measure a program’s performance at a time when the standards were not in effect. Commenters suggested that this approach deprives institutions of any ability to make improvements that would be reflected in those programs’ initial D/E rates.  Some commenters noted that this issue is more significant for programs that are of longer duration, as there will be a longer period after implementation of the regulations during which the D/E rates are based on student outcomes that predate the regulations.  Some commenters also noted that the manner in which program performance is measured could result in programs being required to provide warnings to students that would depress enrollment at times when the program had already been improved. 

Commenters proposed that the Department lengthen the transition period to avoid any sanctions against low-performing programs based upon periods when the new regulations were not in effect.  Other commenters urged that some mechanism be used to take more recent program performance into consideration.

Discussion:  Eligibility determinations based on past program performance, even performance that predates the effective date of the regulations, does not present a legal impediment to these regulations. A law is “not retroactive merely because the facts upon which its subsequent action depends are drawn from a time antecedent to the enactment.” Reynolds v. United States, 292 U.S. 443, 449 (1934).  This principle applies even when, as is the case with these regulations, the statutes or regulations at issue were not in effect during the period being measured. Career College Ass’n v. Riley, No. 94-1214, 1994 WL 396294 (D.D.C. July 19, 1994). This principle has been confirmed in the context of the Department’s use of institutional cohort default rates. Ass’n of Accredited Cosmetology Schools v. Alexander, 979 F.2d 859, 860-62 (D.C. Cir. 1992); Pro Schools Inc. v. Riley, 824 F.Supp. 1314 (E.D. Wis. 1993).  The courts in these matters found that measuring the past default rates of institutions was appropriate because the results would not be used to undo past eligibility, but rather, to determine future eligibility. See, e.g., Ass’n of Accredited Cosmetology Schools, 979 F.2d at 865. As with the institutional cohort default rate requirements, as long as it is a program’s future eligibility that is being determined using the D/E rates measure, the assessment can be based on prior periods of time. Indeed, the court in APSCU v. Duncan rejected this retroactivity argument with respect to the 2011 Prior Rule. 870 F. Supp. 2d at 151-52.

We discuss the comments relating to the transition period under “§668.404 Calculating D/E Rates.”

Changes: None.

Comments: We received many comments in support of the proposed regulations, including both general expressions of support and support with respect to specific aspects of the proposed regulations. Commenters stated that the proposed regulations would help ensure that more students have the opportunity to enter programs that prepare them for gainful employment and that students would be better positioned to repay their educational loans. Several commenters also believed that the regulations will help curtail the abusive recruiting tactics that were revealed by the Senate Permanent Subcommittee on Investigations and the Senate Committee on Health, Education, Labor and Pensions (HELP) in 2012. One commenter expressed support on the basis that, by preventing students from enrolling in low-performing programs, the regulations would curb predatory recruiting practices that target veterans in particular.

Discussion: We appreciate the support of these commenters.

Changes: None.

Comments: We received a number of comments suggesting that the regulations were not sufficiently strong to ensure programs prepare students for gainful employment and to protect students. One commenter argued that the regulations set a low bar for compliance and would do little to stem the flow of Federal dollars to poorly performing institutions. This commenter argued that Federal investment in a program carries an implied endorsement that the program has been “approved” and that the Department has determined it worthwhile. Similarly, several commenters advocated for stronger regulations that close loopholes by which programs could “game” the accountability metrics.

Discussion: We disagree that the regulations set too low a bar for compliance. We believe that the accountability framework strikes a reasonable balance between holding institutions accountable for poor student outcomes and providing institutions the opportunity to improve programs that, if improved, may offer substantial benefits to students and the public.

The Department acknowledges the concern among several commenters about potential loopholes in the proposed accountability metrics and notes that many of these concerns related to program cohort default rates, which in the final regulations will not be used as an accountability metric but, rather, will be used only as a potential disclosure item. We address the commenters’ other specific concerns in the sections of the preamble to which they pertain. As a general matter, however, although we cannot anticipate every situation in which an institution could potentially evade the intent of the regulations, we believe the regulations will effectively hold institutions accountable for a program’s student outcomes and make those outcomes transparent to students, prospective students, the public, taxpayers, and the Government.

Changes: None.

Comments: Several commenters argued that the regulations create overly burdensome reporting and compliance requirements that will be an enormous drain on programs and result in higher tuition costs. One commenter asserted that the regulations add 1.65 million additional hours of workload for institutions. Commenters contended that the regulations would harm community colleges by creating heavy regulatory and financial burdens and stifle innovation and employment solutions for both students and businesses. One commenter argued that, to avoid the administrative burden created by the regulations, foreign institutions with a small number of American students would likely cease to participate in the title IV, HEA programs.

Discussion: We appreciate the commenters’ concerns. Throughout the regulations, we have balanced our interest in minimizing burden on institutions with our interest in achieving our dual objectives of accountability and transparency. The reporting and disclosure requirements are integral to achieving those goals. We discuss concerns about burden throughout this preamble, including in “§668.411 Reporting Requirements for GE Programs,” “§668.412 Disclosure Requirements for GE Programs,” and Paperwork Reduction Act of 1995.

Changes: None.

Comments: Commenters expressed several concerns about specific elements of the definition of “gainful employment (GE) program.” Commenters recommended that graduate programs be excluded from the definition and, specifically, from evaluation under the accountability metrics. One commenter suggested that the HEA framework relating to gainful employment programs was established at a time when most qualifying programs were short term and job focused. The commenter asserted that it is unfair to apply this framework to graduate-level programs where the same program, for example, a Masters of Business Administration program, may be offered by a for-profit institution--and qualify as a GE program--and by a public institution--but not qualify as a GE program. Another commenter argued that a stated purpose of the regulations is to focus on the employability of students enrolled in entry-level postsecondary programs, and that evaluating graduate programs, where there are not the same employment challenges and return-on-investment considerations, would be inconsistent with this purpose. One commenter asserted that based on its analysis, graduate programs would be minimally affected by the proposed metrics and therefore should be exempt from them. Commenters also argued that graduate students are mature students and often experienced workers familiar with the debt and earnings potential of various educational and career paths who do not require the protections offered by the regulations. Commenters argued that the D/E rates measure and program Cohort Default Rate (pCDR)⁴ measure are not reliable metrics for many graduate programs because, according to the commenters, there tends to be a longer lag in time between when students enter these programs and when they experience increased earnings gains.

One commenter recommended that the Department exempt all law programs accredited by the American Bar Association because, according to the commenter, students who complete accredited law programs rarely have difficulty in avoiding default on loans. We received similar comments with respect to graduate medical programs. One commenter recommended that the Department conduct a study on the impact of the D/E rates measure on medical programs and release that with the final regulations.

Some commenters argued generally that it is unfair for the Department to set requirements for some programs and not others. One commenter, focusing on degree programs, questioned treating for-profit institutions and public institutions differently based on whether the degree programs are subject to the gainful employment requirements.

Some commenters suggested that “GE programs” should be defined more narrowly. These commenters suggested that, instead of grouping programs by classification of instructional program (CIP) code and credential level, GE programs should be evaluated by campus location, or at the individual program level, because program performance may vary by campus location or program format due to differences in, for example, student demographics, local market conditions, and instructional methods.

One commenter noted that community colleges may offer programs where certificates and associate degrees are conferred concurrently upon completion, and recommended excluding these types of programs from the definition of “GE program” as they are primarily degree programs offered by a public institution, which would not otherwise constitute GE programs.

Discussion: To the extent a program constitutes an “eligible program” that “provides a program of training to prepare students for gainful employment in a recognized profession” under the HEA, the program by statute constitutes a “GE program,” and we do not have the authority to exclude it from the regulations. We note, for example, that Congress amended the HEA in 2008 to exempt from the gainful employment provisions programs leading to a baccalaureate degree in liberal arts that had been offered by a regionally accredited proprietary institution since January 1, 2009. We view this relatively recent and very specific amendment as an indication that the Department lacks discretion to exempt other types of programs. This applies to graduate programs, including ABA-accredited law schools or medical schools, regardless of the results of such programs under the D/E rates measure. The Department is not providing a separate study analyzing the impact of the D/E rates measure on medical programs with these regulations. As the regulations are implemented, we will monitor the impact of the D/E rates measure on all GE programs, including graduate medical programs.

We also do not agree that the purposes of the regulations are served by excluding graduate programs. Specifically, the issues of accountability for student outcomes, including excessive student debt, and transparency are as relevant to graduate programs and students as they are to undergraduate programs and students. Whether or not it is the case that many graduate programs prepare students for occupations where earnings gains are delayed, we do not believe that this justifies an exemption from the regulations. As discussed in the NPRM, earnings must be adequate to manage debt both in the early years after entering repayment and in later years. Future earnings gains are of course a desirable outcome, but borrowers could default on their loans soon after entering repayment, or experience extreme hardship that leads to negative consequences, well before these earnings gains are realized. Further, as discussed in the NPRM, borrowers may still be facing extreme hardship in repaying their loans even though they have not defaulted, and so, a low default rate by itself is not necessarily an indication that a program is leading to manageable student debt.

In response to commenters’ concerns that similar programs offered by for-profit institutions and public institutions would be treated differently under the regulations, we note that this reflects the treatment of these programs under the HEA and a policy decision made by Congress. We firmly believe that implementing this policy decision through these regulations is necessary and appropriate and that students, prospective students, their families, the public, taxpayers, and the Government will benefit from these efforts.

Regarding the commenters’ request that we evaluate GE programs at the campus level, we do not agree that it would be beneficial to break down the definition of “GE program” beyond CIP code and credential level. A GE program’s eligibility for title IV, HEA program funds is determined at the institutional level, not by location; thus a program’s eligibility applies to each of the locations at which the institutions offers the program. We note also that §668.412 permits institutions offering a GE program in more than one location or format to create separate disclosure templates for each location or format. Thus, the institution has the discretion to provide information about its programs by location or format if it chooses to do so.

With respect to the commenter’s request that we exclude from the definition of “GE program” programs at public institutions that concurrently confer an associate degree and a certificate, we do not believe a specific exclusion is required. A degree program at a public institution is not a “GE program,” even though enrolled students may also earn a certificate as part of the degree program. Of course, if the student is separately enrolled in a certificate program that student is included in that GE program for purposes of the D/E rates measure and disclosures.

Changes: None.

Comments: One commenter suggested that the Department should exempt small businesses that offer GE programs or, if the regulations do not provide an exemption based on size, that the Department should consider an additional or alternate requirement that institutions must meet (such as spending 2.5 times on instruction and student services than on recruitment). Another commenter stated that the Department should exempt institutions that have an enrollment of less than 2,000 students because of the burden that would be imposed on small institutions.

Discussion: We disagree that programs at institutions that might be considered small businesses or institutions with an enrollment of less than 2,000 students should be exempted from the regulations. In addition to the limitations in our statutory authority, an institution’s size has no effect on whether the institution is preparing students for gainful employment in a recognized occupation. We also see no basis for establishing an alternative metric based on the amount of revenues an institution spends on instruction compared to recruiting because it would not indicate when a program is resulting in high debt burden. We believe that any burden on institutions resulting from these regulations is outweighed by the benefits to students and taxpayers. We discuss the burden on small institutions in the Final Regulatory Flexibility Analysis.

Changes: None.

Comments: One commenter suggested that in the final regulations, the Department should commit to evaluating whether the regulations result in the cost savings for the government estimated in the NPRM and the impact of the regulations on Federal student aid funding. The commenter also suggested that the Department commit to reviewing the estimated costs of implementing the regulations, including costs for meeting the information collection requirements. The commenter said the Department should commit to measuring whether the certification criteria for new programs are effective at ensuring whether those programs will remain eligible and pass the accountability metrics. Additionally, the commenter suggested that the Department affirm that it will measure whether the disclosure and reporting requirements improve market information as evidenced by increased enrollment in passing GE programs and decreased enrollment in failing and zone programs.

Discussion: We appreciate the commenters’ suggestions and, as with all of our regulations, we intend to review the regulations as we implement them to ensure they are meeting their intended purposes and to evaluate the impact on students, institutions, and taxpayers.

Changes: None.

Comments: A number of commenters raised concerns about the definition of “student,” specifically the limitation of the term “students” to those individuals who receive title IV, HEA program funds for enrolling in the applicable GE program. These commenters believed that “student” should be defined, for all or some purposes of the regulations, more broadly.

Some commenters proposed that “student” be defined to include all individuals enrolled in a GE program, whether or not they received title IV, HEA program funds. These commenters argued that the purpose of the regulations should be to measure, and disclose, the outcomes of all individuals in a program. They argued that limiting the definition of “student” to students who receive title IV, HEA program funds is arbitrary and would present inaccurate and unrepresentative program outcomes, particularly for community colleges. According to these commenters, many of the individuals attending GE programs at community colleges do not receive title IV, HEA program funds and any accountability measures and disclosures that exclude their debt and earnings would not accurately reflect the performance of the GE program. They claimed that individuals who receive title IV, HEA program funds are disproportionally from underserved and low-income populations and tend to have higher debt and lower earnings outcomes.

Other commenters stated that the definition should include all students with a record in the National Student Loan Database System (NSLDS) because these individuals either filed a Free Application for Federal Student Aid (FAFSA) or have previously received title IV, HEA program funds for attendance in another eligible program. According to the commenters, including these individuals would more accurately reflect the title IV, HEA program population at an institution and provide more relevant information for both eligibility determinations and consumer information. In making these suggestions, commenters were mindful of the court’s interpretation in APSCU v. Duncan of relevant law regarding the Department’s authority to maintain records in its NSLDS. Under these alternative proposed definitions, the commenters suggested that the Department could collect and maintain data regarding these individuals in a manner consistent with APSCU v. Duncan as they would already have records in NSLDS for these individuals.

Some commenters requested that the term “student” include individuals who did not receive title IV, HEA program funds for only specific purposes of the regulations. Some commenters argued that the definition of “student” for the purpose of the D/E rates measure should include all individuals who completed the program, whether or not they received title IV, HEA program funds, on the grounds that earnings and debt levels at programs are to some extent derived from differences in student characteristics and borrowing behavior between students receiving title IV, HEA program funds and individuals who do not receive title IV, HEA program funds. One commenter suggested that individuals who do not receive title IV, HEA program funds should be included in the calculation of D/E rates because otherwise, according to the commenter, institutions would encourage students who do not otherwise plan to take out loans to do so in order to improve a program’s performance on the D/E rates measure.

Other commenters argued that the definition should be broadened only for certain disclosure requirements. For example, some of the commenters suggested that the completion and withdrawal rates and median loan debt disclosures should include the outcomes of all individuals enrolled in a GE program, both those who receive title IV, HEA program funds and those who do not in order to provide students, prospective students, and other stakeholders with a complete picture of a GE program’s performance.

Discussion: We continue to believe that it is necessary and appropriate to define the term “student” for the purposes of these regulations as individuals who received title IV, HEA program funds for enrolling in the applicable GE program for two reasons.

First, as discussed in more detail in the NPRM, this approach is aligned with the court’s interpretation in APSCU v. Duncan of relevant law regarding the Department’s authority to maintain records in its NSLDS. See APSCU v. Duncan, 930 F. Supp. 2d at 220. Second, by limiting the D/E rates measure to assess outcomes of only students who receive title IV, HEA program funds, the Department can effectively evaluate how the GE program is performing with respect to the students who received the Federal benefit that we are charged with administering. Because the primary purpose of the D/E rates measure is determining whether a program should continue to be eligible for title IV, HEA program funds, we can make a sufficient assessment of whether a program prepares students for gainful employment based only on the outcomes of students who receive those funds.

Although we appreciate the commenters’ interest in expanding the definition of “student” to consider the outcomes of all individuals enrolled in a GE program, our goal in these regulations is to evaluate a GE program’s performance for the purpose of continuing eligibility for title IV, HEA program funds. Our proposed definition of “students” is directly aligned with that goal. In addition, this approach is consistent with our goal of providing students and prospective students who are eligible for title IV, HEA program funds with relevant information that will help them in considering where to invest their resources and limited eligibility for title IV, HEA program funds. We understand that some GE programs may not have a large number of individuals receiving title IV, HEA program funds, but given the overall purpose of the regulations--determining a GE program’s eligibility for title IV, HEA program funds--we do not believe it is necessary to measure the outcomes of individuals who do not receive that aid. For the same reasons, we do not believe it is necessary to include individuals who do not receive title IV, HEA program funds in the calculation of D/E rates or in the disclosures the Department calculates for a program.

Finally, the Department does not agree that limiting its analysis to only students receiving title IV, HEA program funds would create an incentive for institutions to encourage more students to borrow. We do not think it would be common for a student to take out a loan that the student did not otherwise plan to take on.

Changes: None.

Comments: One commenter stated that the Department had not adequately explained its departure from the approach taken in the 2011 Final Rules, which considered the outcomes of all individuals enrolled in a GE program rather than just individuals receiving title IV, HEA program funds.

Discussion: We have adequately justified the Department’s decision to base the D/E rates measure only on the outcomes of individuals receiving title IV, HEA program funds. Our analysis of this issue is described in the previous paragraphs, was set forth in considerable detail in the NPRM, and, additionally, as noted in the NPRM, is supported by the court’s decision in APSCU v. Duncan. The justifications presented meet the reasoned basis standard we must satisfy under the Administrative Procedure Act and relevant case law.

Changes: None.

Comments: We received a number of comments about the definition of “student” in the context of the mitigating circumstances showing in §668.406 of the proposed regulations. As proposed in the NPRM, an institution would be permitted to demonstrate that less than 50 percent of all individuals who completed the program during the cohort period, both those individuals who received title IV, HEA program funds and those who did not, incurred any loan debt for enrollment in the program. A GE program that could make this showing would be deemed to pass the D/E rates measure.

In this context, some commenters argued against allowing institutions to include individuals who do not receive title IV, HEA program funds for enrollment in the GE program. These commenters noted that including individuals who do not receive these loans is at odds with the legal framework that the Department established in order to align the regulations with the district court’s decision in APSCU v. Duncan. They suggested that permitting institutions to include individuals who do not receive loans under the title IV, HEA programs in a mitigating circumstances showing would be inconsistent with the court’s decision and as a result would violate the HEA.

Several commenters also asserted that permitting mitigating circumstances showings or providing for a full exemption would discriminate in favor of institutions, such as community colleges, where less than 50 percent of individuals enrolled in the program receive title IV, HEA program funds. According to these commenters, many of these public institutions have higher costs than institutions in the for-profit sector but have lower borrowing rates because the higher costs are subsidized by States. The commenters stated that if these institutions’ programs are considered exempt from the D/E rates measure, programs that perform very poorly on other measures like completion would continue merely because they are low cost even though they do not reflect a sound use of taxpayer funds.

Some commenters stated that permitting a mitigating circumstances showing would result in unfair and unequal treatment of similar institutions in different States. The commenter said that, for example, in some States, cosmetology programs are eligible for State tuition assistance grants, while in other States these programs are not eligible for such grants. Schools charging the same tuition and whose graduates are making the same amount in one State would pass the D/E rates measure while those in another would not. Finally, some commenters asserted that only a fraction of programs at public institutions would fail the D/E rates measure, and that this small number does not support an exemption or permitting a mitigating circumstances showing.

A number of commenters supported the proposed mitigating circumstances showing, and specifically the inclusion of individuals who do not receive title IV, HEA program funds. As noted previously, commenters argued that these individuals should be considered because the number of students receiving title IV, HEA program funds and incurring debt to enroll in many community college programs is typically very small and these students do not represent the majority of individuals who complete the program. According to these commenters, a program in which at least 50 percent of individuals enrolled in the program have no debt is unlikely to produce graduates whose educational debts would be excessive because tuition and costs are likely to be low and require little borrowing. Commenters further noted that including these individuals in the calculation would be consistent with the 2011 Prior Rule, where a program with a median loan debt of zero passed the debt-to-earnings measures based on the borrowing activity of individuals who receive title IV, HEA program funds and those who do not. These commenters stated that even though the Department is largely limiting the accountability measures to an analysis of the earnings and debt of students receiving title IV, HEA program funds due to the concerns expressed by the district court in APSCU v. Duncan, a program with a median loan debt of zero, whether or not the calculation is limited to students receiving title IV, HEA program funds, should still pass the D/E rates measure.

Finally, these commenters noted that the D/E rates measure is designed to help ensure that students are receiving training that will lead to earnings that will allow them to pay back their student loan debts after they complete their program. According to these commenters, many GE programs, including many programs offered by community colleges, have low tuition and many of their students can pay the costs of the program solely through a Pell Grant, rather than incurring debt.

Some of the commenters who supported allowing an institution to make a showing of mitigating circumstances under §668.406 of the proposed regulations also argued that, instead of requiring such a showing, the Department should completely exempt from the D/E rates measure any GE program for which less than 50 percent of the individuals who completed the program incurred loan debt for enrollment in the program. The commenters proposed several methodologies the Department could use to determine which programs qualify for the exemption. These commenters made similar arguments to those discussed previously--that these programs should not be subject to the administratively burdensome process for calculating the D/E rates, when ultimately these programs will have a median loan debt of zero and therefore will be determined to be passing the D/E rates measure. One of these commenters suggested that, if a program is failing or in the zone with respect to the D/E rates measure, the institution should have the ability to recalculate its median loan debt based on all graduates, to evaluate the overall quality of a program. The commenter proposed that, if the program passes on the basis of that recalculation, the notice of determination issued by the Department would be annotated to reflect that the institution made a showing of “mitigating circumstances” and the program would be deemed passing. Some of the commenters also argued that an exemption based on a borrowing rate of less than 50 percent should apply across the board to all GE program requirements, including the reporting and disclosure requirements.

Commenters asserted that, absent an exemption, many low-cost programs with a low borrowing rate would be inclined to leave the Direct Loan program or close their programs, even those programs that were effective. The commenters further stated that these closures would disproportionately affect minority and economically disadvantaged students, many of whom enroll in these programs, and that without these programs, these students would not have available economically viable options for furthering their education.

Discussion: We appreciate the commenters’ responses to our request for comment on the definition of “student” and the mitigating circumstances provision in proposed §668.406. None of the commenters, however, presented an adequate justification for us to depart from our proposed definition of “students” and the purpose of the regulations, which is to evaluate the outcomes of individuals receiving title IV, HEA program funds and a program’s continued eligibility to receive title IV, HEA program funds based solely on those outcomes. We do not agree that a borrowing rate below 50 percent necessarily indicates that a program is low cost or low risk. A program with a borrowing rate of under 50 percent, particularly a large program, could still have a substantial number of students with title IV loans and, additionally, those students could have a substantial amount of debt or insufficient earnings to pay their debt. We also note that, if a GE program is indeed “low cost” or does not have a significant percentage of borrowers, which commenters claimed is the case with many community college programs, it is very likely that the program will pass the D/E rates measure because most students will not have any debt. NPSAS data show that, of all students completing certificate programs at two-year public institutions who received title IV, HEA program funds, 77 percent received only Pell Grants and only 23 percent were borrowers.⁵ Program results in the 2012 GE informational D/E rates data set reflects the findings of the NPSAS analysis. Of the 824 programs at two-to-three-year public institutions in the 2012 GE informational D/E rates data set, 823 pass under the D/E rates measure. Further, of the 824 total programs at two-to-three-year public institutions, 504 (61 percent) have zero median debt, which means that, for these programs, less than half of the students completing the program are borrowers and that the majority of their students completing the program received title IV, HEA program funds in the form of Pell Grants only. Accordingly, we do not believe there is adequate justification to depart from our definition of “student,” by permitting a showing of mitigating circumstances based on individuals who do not receive title IV, HEA program funds for enrollment in a program, or to make a greater departure from our accountability framework, by permitting a related up-front exemption.

Changes: We have revised the regulations to remove the provisions in §668.406 that would have permitted institutions to submit a mitigating circumstances showing for a GE program that is not passing the D/E rates measure.

Comments: A number of commenters recommended revisions to the definition of “prospective student.” One commenter recommended that the Department use the definition of “prospective student” in §668.41(a), which provides that a “prospective student” is an individual who has contacted an eligible institution for the purpose of requesting information concerning admission to that institution. The commenter argued that using this definition would maintain consistency across the title IV, HEA program regulations.

Some of the commenters stated that the proposed definition is too broad. Specifically, they noted that an institution would not be able to identify, for example, to whom it was required to deliver disclosures and student warnings if anyone who had passive contact with an institution’s advertising constituted a “prospective student” under the regulations. They suggested that if “prospective student” is defined that broadly, they would not be able to meet their obligations with respect to these students under the regulations or that compliance would be very burdensome, potentially requiring the development of new admissions and marketing materials annually. These commenters recommended that we revise the definition of “prospective student” to include only individuals who actively seek information from an institution about enrollment in a program. Another commenter expressed concern about the definition because, according to the commenter, a prospective student would include anyone who has access to the Internet.

Other commenters stated that the definition is too narrow and recommended that the term include anyone in contact with an institution about “enrollment,” rather than “enrolling.” According to these commenters, with this change, the definition would include family members, counselors, and others making enrollment inquiries on behalf of someone else.

Discussion: We believe that it is appropriate to establish a definition of “prospective student” that is tailored to the purpose of these specific regulations. In that regard, the definition will account for the various ways that institutions and prospective students commonly interact and target interactions that are specific to enrollment in a GE program, rather than more general contact about admission to an institution. Specifically, unlike the existing definition of “prospective student” in §668.41(a), the definition in the GE regulations applies without regard to whether an individual or the institution initiates contact.

We agree, however, that an individual’s passive interaction with an institution’s advertising should not result in that individual being considered a “prospective student” for the purposes of the regulations. Accordingly, we are removing the reference to indirect contact through advertising from the definition of “prospective student.” Recognizing that institutions sometimes engage third parties to recruit students, we have also revised the definition to capture this type of direct contact with prospective students.

The commenters’ proposed alternative definition, which would include individuals other than those in contact with the institution about enrolling in a program, is too broad for each of the purposes for which the definition is used. However, as we discuss in “§668.410 Consequences of the D/E Rates Measure,” we agree that, where an initial inquiry about enrolling in a program is made by a third party on behalf of a prospective student, the third party, as a proxy for the prospective student, should be given the student warning, as that is when a decision is likely to be made about whether to further explore enrolling in that program. We do not believe that the same reasoning applies, for example, with respect to the requirement in §668.410 that a written warning be given to a prospective student at least three, but not more than 30, days before entering into an enrollment agreement.

Thus, the changes to the definition and to the related requirements that we have described balance the need to provide prospective students with critical information at a time when they can most benefit from it with ensuring that the administrative burden for institutions is not unnecessarily increased.

Changes: We have revised the definition of “prospective student” to exclude indirect contact through advertising and to include contact made by a third party on an institution’s behalf.

Comments: One commenter asked that we clarify whether credential level is determined by academic year or calendar year.

Discussion: After further review of the proposed regulations, we have made several changes to the definition of “credential level” that make the commenter’s concern moot. First, we are revising the definition to accurately reflect the treatment of a post-baccalaureate certificate as an undergraduate credential level under the title IV, HEA programs. This certificate was inappropriately listed as a graduate credential level in the proposed regulations.

We also are simplifying the definition by treating all of an institution’s undergraduate programs with the same CIP code and credential level as one “GE program,” without regard to program length, rather than breaking down the undergraduate credential levels according to the length of the program as we proposed in the NPRM. To do so would be inconsistent with other title IV, HEA program reporting procedures and would unnecessarily add complexity for institutions. We note that, under §668.412(f), an institution that offers a GE program in more than one program length must publish a separate disclosure template for each length of the program. Although D/E rates will not be separately calculated, several of the other required disclosures, including the number of clock or credit hours or equivalent, program cost, placement rate, and percentage of students who borrow, must be broken down by length of the program. Thus, students and prospective students will have information available to make distinctions between programs of different lengths.

Changes: We have revised the definition of “credential level” to include post-baccalaureate certificates as an undergraduate, rather than graduate, credential level and to specify that undergraduate credential levels are: undergraduate certificate or diploma, associate degree, bachelor’s degree, and post-baccalaureate certificate.

§668.402 Definitions

Comments: We received a number of comments regarding defined terms in the proposed regulations.

Discussion: Consistent with our organizational approach in the NPRM, we describe the comments received relating to a specific defined term in the section in which the defined term is first substantively used.

Changes: We have made changes to the following defined terms. The changes are described in the section or sections indicated after the defined term.

Credential level (§668.401)

Classification of instructional program (CIP) code and, within that definition, the term “substantially similar” (§§668.410 and 668.414)

Cohort period (§668.404)

GE measures (§668.403)

Program cohort default rate (§668.403)

Prospective student (§668.401)

§668.403 Gainful Employment Program Framework

Impact on For-Profit Institutions

Comments: Some commenters asserted that the poor outcomes identified by the D/E rates measure--high debt and low earnings--are problems across higher education and that, as a result, it would be unfair to hold only GE programs accountable under the D/E rates measure. Commenters cited data that, they argued, showed that this is the case for a large fraction of four-year programs operated by public and non-profit institutions. One commenter contended that between 28 percent and 54 percent of programs operated by the University of Texas would fail the Department’s accountability metrics.⁶

Several commenters alleged that the regulations are a Federal overreach into higher education. A number of these commenters believed that the regulations unfairly target for-profit institutions. They stated that while a degree program at a for-profit institution must meet the D/E rates measure to remain eligible for title IV, HEA program funds, a comparable degree program at a public or private non-profit institution, which may have low completion rates or other poor outcomes, would not be subject to the regulations.

Some commenters asserted that for-profit institutions play an important role in providing career training for students to enter into jobs that do not require a four-year bachelor’s degree. In that regard, one commenter contended that, because the regulations apply only to GE programs offered primarily by for-profit institutions, the regulations reflect a bias in favor of traditional four-year degree programs not subject to the regulations. This bias, the commenter argued, cannot be justified in light of BLS data showing that nearly half of bachelor’s degree graduates are working in jobs that do not require a four-year degree. These degree-holders, according to the commenter, are actually employed in what can be described as “middle-skill” positions, for which the commenter believed for-profit institutions provide more effective preparation. These commenters all asserted that traditional institutions are ill-suited to provide students with training for middle-skill jobs compared to for-profit institutions. Other commenters argued that enrollment growth at non-profit and public institutions has not kept up with demand from students and for-profit institutions have responded to this need by offering opportunities for students. One commenter presented data showing that a majority of degrees in the fastest growing occupations are awarded by for-profit institutions.

Several commenters asserted that the regulations would have a substantial and disproportionate impact on programs in the for-profit sector and the students they serve. Commenters cited an analysis by Mark Kantrowitz claiming that, of GE programs that would not pass the D/E rates measure, a large and disproportionate portion are operated by for-profit institutions compared to programs operated by non-profit and public institutions, while other commenters relied on Department data to draw the same conclusion.⁷

Commenters said the Department is targeting for-profit programs because of an incorrect assumption that student outcomes are worse at for-profit institutions. They said the Department has ignored studies showing that, when compared to institutions that serve similar populations of students, for-profit institutions achieve comparable outcomes for their students. Another commenter cited a study that showed that first-time enrollees at for-profit schools experience greater unemployment after leaving school, but among those working, their annual earnings are statistically similar to their counterparts at non-profit institutions.

Several commenters asserted that the student body profiles at for-profit institutions could significantly affect program performance under the D/E rates measure. Charles River Associates analyzed NPSAS:2012 data and found that for-profit institutions serve older students (average age of 30.0 years compared to 24.6 years at private non-profit and 26.0 years at public institutions), veterans (7 percent of students compared to 3 percent at private non-profit and public institutions), students that are not exclusively full-time (30 percent of students compared to 29 percent at private non-profit and 57 percent at public institutions), independent students (80 percent at private for-profit institutions to 34 percent at private non-profit institutions and 49 percent at public institutions), single parents (33 percent at private for-profit institutions to 9 percent at private non-profit institutions and 13 percent at public institutions), students with dependents (51 percent at private for-profit institutions to 18 percent at private non-profit institutions to 25 percent at public institutions), students working more than 20 hours per week (48 percent at private for-profit institutions to 29 percent at private non-profit institutions to 44 percent at public institutions), students who consider their primary role to be an employee rather than a student (52 percent at private for-profit institutions to 23 percent at private non-profit institutions to 31 percent at public institutions), and students less likely to have a parent with at least a bachelor’s degree (22 percent at private for-profit institutions to 52 percent at private non-profit institutions to 37 percent at public institutions).⁸ They also found that minority students make up a higher percentage of the student body at for-profit institutions, with African-Americans making up 26 percent of students compared to 15 percent at public institutions and 14 percent at private non-profit institutions and Hispanic students comprising 19 percent of students at for-profit institutions, similar to the 17 percent at public institutions but higher than the 10 percent at private non-profit institutions. Additionally, commenters stated that 65 percent of students at for-profit institutions receive Pell Grants, while at private non-profit and public institutions, the percentage of Pell Grant recipients averages 36 percent and 38 percent, respectively. In addition, one commenter suggested that the Department should have considered that for-profit institutions are more likely to be open-enrollment institutions.

Commenters asserted that for-profit institutions do not in fact cost more for students and taxpayers than public institutions, particularly community colleges, when State and local appropriations and other subsidies received by public institutions are taken into account. One commenter said that for-profit two-year institutions cost less per student than public two-year institutions and that completion rates are somewhat higher at for-profit institutions. Commenters pointed to a number of studies estimating taxpayer costs across types of institutions. One found that associate degree programs at public institutions cost $4,000 more per enrollee and $35,000 more per graduate than associate degree programs at for-profit institutions, while another found that the direct cost to taxpayers on a per-student basis is $25,546 lower at for-profit institutions than at public two-year institutions, and a third found that taxpayer costs of four-year public institutions averaged $9,709 per student compared to $99 per student at for-profit institutions. Another study estimated that public institutions receive $19.38 per student in direct tax support and private non-profit institutions receive $8.69 per student for every $1 received by for-profit institutions per student. Commenters also referenced research estimating the total costs to State and local governments if students affected by the regulations shift to public institutions, with results ranging from $3.6 to $4.7 billion to shift students from nine for-profit institutions in four States to public two-year or four-year institutions. Similarly, one commenter referenced a study estimating the total cost of shifting students to public institutions among all States would be $1.7 billion in State appropriations to support one cohort of graduates from failing or zone programs at public 2-year or least selective four-year institutions.

Other commenters referred to budget data related to the title IV, HEA programs to state that student loans do not constitute costs to taxpayers because the recovery rate for these loans is over 100 percent, and asserted that any cost reductions in the title IV, HEA programs would be offset by reduced tax revenues at all levels of government and increased demand for capacity in the public sector. Others noted a GAO Report indicating Federal student loans originated between 2007 and 2012 will bring in $66 billion in revenue and that Congressional Budget Office projections from 2013 indicate that loans originated in the next ten-year period would generate $185 billion. Whether approaching the issue on a per-student, per-graduate, or overall taxpayer cost basis, the commenters stated that the rationale that the regulations will protect taxpayer interests does not withstand scrutiny.

One commenter said that the NPRM overstated the cost of for-profit institutions relative to public two-year institutions, because many programs at for-profit institutions offer advanced degrees and their students accrue more debt. Other commenters said the Department ignores the comparable tuition costs of non-profit private institutions, which, like for-profit institutions, generally do not benefit from direct appropriations from State governments.

One commenter asserted that the 150 percent of normal time graduation rate for public and private non-profit open‐enrollment colleges is 28.3 percent and 39.7 percent respectively while for‐profit colleges graduated 35.2 percent of students within 150 percent of normal time. Additionally, the commenter contended, more than half (55.7 percent) of for-profit colleges were open enrollment institutions in 2011-12, compared to less than 18 percent of public and 12 percent of private not‐for‐profit schools. Based on these findings, the commenter argued that while the for‐profit graduation rate is lower than the average of all public and private nonprofit institutions, it is higher than the average of all open‐enrollment public and private nonprofit institutions, which the commenter stated is likely to be a more appropriate comparison group.

Several commenters claimed that the Department’s reference in the NPRM to qui tam lawsuits and State Attorneys General investigations into for-profit institutions evidence bias. In particular, commenters suggested such investigations were politically driven, based on bad-faith attacks, and failed to produce evidence of wrongdoing.

Some commenters said the Department’s reference in the NPRM to a GAO report on the for-profit sector also demonstrates bias against for-profit institutions. Commenters asserted that the GAO investigation in particular contained errors and relied on false testimony, which required the GAO to correct and reissue its report.⁹ Commenters said it was also inappropriate for the Department to rely on what the commenters called a “deeply flawed” partisan report by the Senate HELP committee majority staff, because the report partially relied on evidence presented in the GAO report, was actually issued by the committee majority staff for the committee, and was not adopted by vote of the whole committee.¹⁰

On the other hand, several commenters suggested that the Department should focus regulatory efforts on for-profit institutions because they have been engaged in predatory recruitment practices that hurt students and divert taxpayer funds away from higher-quality education programs. One commenter said that for-profit institutions increased recruiting of veterans by over 200 percent in just one year. Many commenters described the disproportionate distribution of government benefits to the for-profit sector, contending that for-profit institutions enroll only 10 percent of students, but account for 25 percent of Pell Grants and Stafford loan volume and account for half of defaults; that for-profit schools collected more than one-third of all G.I. Bill funds, but trained only 25 percent of veterans, while public colleges and universities received only 40 percent of G.I. Bill benefits but trained 59 percent of veterans; and that for-profit colleges cost taxpayers twice the tuition as non-profits. Several commenters described the high proportion of students who drop out of or withdraw from programs at for-profit institutions--about half of students who enroll.

On the other hand, several commenters cited an analysis of IPEDS data by Charles River Associates that found that the difference in FY 2010 institutional cohort default rates (iCDR) among for-profit (22 percent), private non-profit (8 percent), and public (13 percent) institutions was significantly reduced when institutions were grouped into two categories of Pell Grant recipient concentration. The High Pell group had at least 50 percent of students receiving Pell Grants and the Low Pell group had less than 50 percent of students with Pell Grants. The Charles River Associates analysis found that among two-year institutions, in the High Pell Group, the iCDR at for-profit institutions is 20.6 percent compared to 24.2 percent at public institutions and, in the Low Pell Group, the iCDR is 16.6 percent at for-profit institutions and 20.4 percent at public institutions.

Several commenters asserted that the Department has clear justification for limiting application of the regulations to institutions in the for-profit sector and other institutions offering programs that purport to prepare students for gainful employment. One commenter cited a study that found that students at for-profit institutions were twice as likely to default on their student loans as students at other types of schools and another study that found that graduation rates at for-profit colleges were less than one-third the rates at non-profit colleges. By comparison, the commenter cited economic research that found that students in non-profit and public certificate programs had lower debt burdens, higher earnings, lower unemployment, and lower student loan default rates and were more satisfied with their programs, even after controlling for student demographic factors.

One commenter said the Department has a specific legislative mandate to regulate gainful employment programs, which include the programs offered by for-profit institutions, and, as a result, the Department is correct to apply the regulations to those programs. Some commenters added that for-profit institutions are subject to less regulation and accountability than non-profit institutions because for-profit institutions are not governed by an independent board composed of members without an ownership interest. Consequently, they argued, the Department should particularly regulate programs operated by for-profit institutions.

Discussion: The regulations do not target for-profit programs for loss of eligibility under the title IV, HEA programs. To the contrary, the Department appreciates the important role for-profit institutions play in educating students.

The for-profit sector has experienced tremendous growth in recent years,¹¹ fueled by the availability of Federal student aid funding and an increased demand for higher education, particularly among non-traditional students.¹² The share of Federal student financial aid going to students at for-profit institutions has grown from approximately 13 percent of all title IV, HEA program funds in award year 2000-2001 to 19 percent in award year 2013-2014.¹³

The for-profit sector plays an important role in serving traditionally underrepresented populations of students. For-profit institutions are typically open-enrollment institutions that are more likely to enroll students who are older, women, Black, Hispanic, or with low incomes.¹⁴ Single parents, students with a certificate of high school equivalency, and students with lower family incomes are also more commonly found at for-profit institutions than community colleges.¹⁵

For-profit institutions develop curriculum and teaching practices that can be replicated at multiple locations and at convenient times, and offer highly structured programs to help ensure timely completion.¹⁶ For-profit institutions “are attuned to the marketplace and are quick to open new schools, hire faculty, and add programs in growing fields and localities,”¹⁷ including occupations requiring “middle-skill” training.

At least some research suggests that for-profit institutions respond to demand that public institutions are unable to handle. Recent evidence from California suggests that for-profit institutions absorb students where public institutions are unable to respond to demand due to budget constraints.^{18 19} Additional research has found that “[c]hange[s] in for-profit college enrollments are more positively correlated with changes in State college-age populations than are changes in public-sector college enrollments.”²⁰

Other evidence, however, suggests that for-profit institutions are facing increasing competition from community colleges and traditional universities, as these institutions have started to expand their programs in online education. According to the annual report recently filed by a large, publically traded for-profit institution, “a substantial proportion of traditional colleges and universities and community colleges now offer some form of . . . online education programs, including programs geared towards the needs of working learners. As a result, we continue to face increasing competition, including from colleges with well-established brand names. As the online . . . learning segment of the postsecondary education market matures, we believe that the intensity of the competition we face will continue to increase.”²¹

These regulations apply not only to programs operated by for-profit institutions, but to all programs, across all sectors, that are subject to the requirement that in order to qualify for Federal student assistance, they must provide training that prepares students for gainful employment in a recognized occupation. Under the HEA, for these purposes, an eligible program includes non-degree programs, including diploma and certificate programs, at public and private non-profit institutions such as community colleges and nearly all educational programs at for-profit institutions of higher education regardless of program length or credential level. Our regulatory authority in this rulemaking with respect to institutional accountability is limited to defining the statutory requirement that these programs are eligible to participate in the title IV, HEA programs because they provide training that prepares students for gainful employment in a recognized occupation. The Department does not have the authority in this rulemaking to regulate other higher education institutions or programs, even if such institutions or programs would not pass the accountability metrics.

The regulations establish an accountability framework and transparency framework for GE programs, whether the programs are operated by for-profit institutions or by public or private non-profit institutions. However, we are particularly concerned about high costs, poor outcomes, and deceptive practices at some institutions in the for-profit sector.

With respect to comments that the NPRM overstates the cost of for-profit institutions relative to public two-year institutions because many for-profit programs offer advanced degrees, the data do not support this contention. A comparison of costs at institutions offering credentials of comparable levels shows that for-profit institutions typically charge higher tuition than do public postsecondary institutions. Among first-time full-time degree or certificate seeking undergraduates at title IV, HEA institutions operating on an academic calendar system and excluding students in graduate programs, average tuition and required fees at less-than-two-year for-profit institutions are more than double the average cost at less-than-two-year public institutions and average tuition and required fees at two-year for-profit institutions are about four times the average cost at two-year public institutions.^{22 23} Because less than two-year and two-year for-profit institutions largely offer certificates and associate degrees, rather than more expensive four-year degrees or advanced degrees,²⁴ it is unlikely to be the case that higher tuition at for-profit institutions is the result of advanced degree offerings as argued by some commenters.

Comparing tuition at for-profit institutions and private non-profit institutions reveals similar results. Although the differential between for-profit institutions and private non-profit institutions that offer similar credentials is smaller than the difference between for-profit institutions and public institutions, for-profit institutions still charge more than private non-profit institutions when comparing two-year and less-than-two-year institutions, which includes the majority of institutions offering GE programs within the non-profit sector.²⁵

The Department acknowledges that funding structures and levels of government support vary by type of institution, with public institutions receiving more direct funding and public and private non-profit institutions benefiting from their tax-exempt status. However, as detailed in “Discussion of Costs, Benefits, and Transfers” in the Regulatory Impact Analysis, we do not agree that the regulations will result in significant costs for State and local governments. In particular, we expect that many students who change programs as a result of the regulations will choose from the many passing programs at for-profit institutions or that State and local governments may pursue lower marginal cost options to expand capacity at public institutions.

With respect to revenues generated by the Federal student loan programs, we note that the estimates presented reflect a low discount rate environment and that returns vary across different segments of the portfolio. Currently, the Direct Loan program reflects a negative subsidy. Subsidy rates represent the Federal portion of non-administrative costs--principally interest subsidies and defaults--associated with each borrowed dollar over the life of the loan. Under Federal Credit Reform Act (FCRA) rules, subsidy costs such as default costs and in-school interest benefits are embedded within the program subsidy, whereas Federal administration costs are treated as annual cash amounts and are not included within the subsidy rate.

Annual variations in the subsidy rate are largely due to the relationship between the OMB-provided discount rate that drives the Government’s borrowing rate and the interest rate at which borrowers repay their loans. Technical assumptions for defaults, repayment patterns, and other borrower characteristics would also apply. The loan subsidy estimates are particularly sensitive to fluctuations in the discount rate. Even small shifts in economic projections may produce substantial movement, up or down, in the subsidy rate. While the Federal student loan programs, especially Unsubsidized loans and PLUS loans, generate savings in the current interest rate environment, the estimates are subject to change. In any event, although the regulations may result in reduced costs to taxpayers from the title IV, HEA programs, the primary benefits of the regulations are the benefits to students.

Because aid received from grants has not kept pace with rising tuition in the for-profit sector, in contrast to other sectors, the net cost to students who attend GE programs has increased sharply in recent years.²⁶ Not surprisingly, “student borrowing in the for-profit sector has risen dramatically to meet the rising net prices.”²⁷ Students at for-profit institutions are more likely to receive Federal student financial aid and have higher average student debt than students in public and private non-profit institutions, even taking into account the socioeconomic background of the students enrolled within each sector.²⁸

In 2011-2012, 60 percent of certificate students who were enrolled at for-profit two-year institutions took out title IV student loans during that year compared to 10 percent at public two-year institutions.²⁹ Of those who borrowed, the median amount borrowed by students enrolled in certificate programs at two-year for-profit institutions was $6,629, as opposed to $4,000 at public two-year institutions.³⁰ In 2011-12, 20 percent of associate degree students who were enrolled at for-profit institutions took out student loans, while only 66 percent of associate degree students who were enrolled at public two-year institutions did so.³¹ Of those who borrowed in 2011-12, for-profit two-year associate degree enrollees had a median amount borrowed during that year of $7,583, compared to $4,467 for students at public two-year institutions.³²

Although student loan default rates have increased in all sectors in recent years, they are highest among students attending for-profit institutions.^{33 34} Approximately 19 percent of borrowers who attended for-profit institutions default on their Federal student loans within the first three years of repayment as compared to about 13 percent of borrowers who attended public institutions.³⁵ Estimates of “cumulative lifetime default rates,” based on the number of loans, rather than borrowers, yield average default rates of 24, 23, and 31 percent, respectively, for public, private, and for-profit two-year institutions in the 2007-2011 cohort years. Based on estimates using dollars in those same cohort years (rather than loans or borrowers, to estimate defaults) the average lifetime default rate is 50 percent for students who attended two-year for-profit institutions in comparison to 35 percent for students who attended two-year public and non-profit private institutions.³⁶ Although we included a regression analysis on pCDR and student demographic characteristics, including the percentage of Pell students attending each program, in the NPRM, we do not respond to comments on this subject because the regulations no longer include pCDR as an accountability metric to determine eligibility for title IV, HEA program funds.

There is evidence that many programs at for-profit institutions may not be preparing students as well as comparable programs at public institutions. A 2011 GAO report reviewed results of licensing exams for 10 occupations that are, by enrollment, among the largest fields of study and found that, for 9 out of 10 licensing exams, graduates of for-profit institutions had lower rates of passing than graduates of public institutions.³⁷

Many for-profit institutions devote greater resources to recruiting and marketing than they do to instruction or to student support services.³⁸ An investigation by the U.S. Senate Committee on Health, Education, Labor & Pensions (Senate HELP Committee) of 30 prominent for-profit institutions found that almost 23 percent of revenues were spent on marketing and recruiting but only 17 percent on instruction.³⁹ A review of useable data provided by some of the institutions that were investigated showed that they employed 35,202 recruiters compared with 3,512 career services staff and 12,452 support services staff.⁴⁰

We disagree with the commenters who asserted that the Department’s reference to the findings presented in the GAO and Senate HELP Committee staff reports are inappropriate because the GAO report (on which the Senate HELP Committee report partially relied) contained errors and misleading testimony. We rely upon available data presented in the re-released version of the GAO report. Because GAO included these data and conclusions on licensure passage rates in their re-released version, we believe this evidence is reliable and appropriate to reference in support of the regulations. Also, we note that the evidence we use from the Senate HELP Committee report⁴¹ is reliable because the data the report is based on are readily available and has been subject to public review. We do not rely upon qualitative testimony presented by the Committee. We referenced in the NPRM some descriptions and characterizations from the HELP and GAO reports of abusive conduct by for-profit institutions, but those descriptions and characterizations were incidental to our discussion and rationale.⁴² We make clear in the NPRM our “primary concern”--that a number of GE programs are not providing effective training and are training for low-paying jobs that do not justify costs of borrowing. 79 FR 16433. We stated that the causes of these problems are “numerous;” we listed five causes, the last of which is the deceptive marketing practices on which the two reports focus.⁴³ Moreover, the two reports were hardly the only evidence we cited of such practices. 79 FR 16435. More pertinent to the commenter’s objection, these regulations are not adopted to impose sanctions on schools that engage in misrepresentations; the Department has already adopted rules to address enforcement actions for misrepresentations by institutions regarding, among other things, their educational programs and the employability of their graduates. See 34 CFR part 668, subpart F. Rather, we concluded that these regulations are needed based on our analysis of the data and literature, and our objectives in these regulations are to establish standards to determine whether a GE program is an eligible program and to provide important disclosures to students and prospective students. We need not rely on reports that indicate predatory and abusive behavior in order to conclude that a test is needed to determine whether a program is in fact one that prepares students for “gainful employment.”

Lower rates of completion at many for-profit institutions are a cause for concern. The six-year degree/certificate attainment rate of first-time undergraduate students who began at a four-year degree-granting institution in 2003-2004 was 34 percent at for-profit institutions in comparison to 67 percent at public institutions.⁴⁴  However, it is important to note that, among first-time undergraduate students who began at a two-year degree-granting institution in 2003-2004, the six-year degree/certification attainment rate was 40 percent at for-profit institutions compared to 35 percent at public institutions.⁴⁵ We note that, as suggested by a commenter, completion rates for only open-enrollment institutions may be different than those discussed here.

The slightly lower degree/certification attainment rates of two-year public institutions may at least be partially attributable to higher rates of transfer from two-year public institutions to other institutions.⁴⁶ Based on available data, it appears that relatively few students transfer from for-profit institutions to other institutions. Survey data indicate about 5 percent of all student transfers originate from for-profit institutions, while students transferring from public institutions represent 64 percent of all transfers occurring at any institution (public two-year institutions to public four-year institutions being the most common type of transfer).⁴⁷ Additionally, students who transfer from for-profit institutions are substantially less likely to be able to successfully transfer credits to other institutions than students who transfer from public institutions. According to a recent NCES study, an estimated 83 percent of first-time beginning undergraduate students who transferred from a for-profit institution to an institution in another sector were unable to successfully transfer credits to their new institution. In comparison, 38 percent of first-time beginning undergraduate students who transferred between two public institutions were not able to transfer credits to their new institution.⁴⁸

The higher costs of for-profit institutions and resulting greater amounts of debt incurred by their former students, together with generally lower rates of completion, continue to raise concerns about whether some for-profit programs lead to earnings that justify the investment made by students, and additionally, taxpayers through the title IV, HEA programs.

In general, we believe that most programs operated by for-profit institutions produce positive educational and career outcomes for students. One study estimated moderately positive earnings gains, finding that “[a]mong associate’s degree students, estimates of returns to for-profit attendance are generally in the range of 2 to 8 percent per year of education.”⁴⁹ However, recent evidence suggests “students attending for-profit institutions generate earnings gains that are lower than those of students in other sectors.”⁵⁰ The same study that found gains resulting from for-profit attendance in the range of 2 to 8 percent per year of education also found that gains for students attending public institution are “upwards of 9 percent.”⁵¹ But, other studies fail to find significant differences between the returns to students on educational programs at for-profit institutions and other sectors.⁵²

Analysis of data collected on the outcomes of 2003-2004 first-time beginning postsecondary students in the Beginning Postsecondary Students Longitudinal Study shows that students who attend for-profit institutions are more likely to be idle--neither working nor still in school--six years after starting their programs of study in comparison to students who attend other types of institutions.⁵³ Additionally, students who attend for-profit institutions and are no longer enrolled in school six years after beginning postsecondary education have lower earnings at the six-year mark than students who attend other types of institutions.⁵⁴

The commenters’ claims that the Department’s reference in the NPRM to qui tam lawsuits and State Attorneys General investigations into for-profit institutions demonstrates bias by the Department against the for-profit sector are simply unfounded. The evidence derived from these actions shows individuals considering enrolling in GE programs offered by for-profit institutions have in many instances been given such misleading information about program outcomes that they could not effectively compare programs offered by different institutions in order to make informed decisions about where to invest their time and limited educational funding.

The GAO and other investigators have found evidence that high-pressure and deceptive recruiting practices may be taking place at some for-profit institutions. In 2010, the GAO released the results of undercover testing at 15 for-profit colleges across several States.⁵⁵ Thirteen of the colleges tested gave undercover student applicants “deceptive or otherwise questionable information” about graduation rates, job placement, or expected earnings.⁵⁶ The Senate HELP Committee investigation of the for-profit education sector also found evidence that many of the most prominent for-profit institutions engage in aggressive sales practices and provide misleading information to prospective students.⁵⁷ Recruiters described “boiler room”-like sales and marketing tactics and internal institutional documents showed that recruiters are taught to identify and manipulate emotional vulnerabilities and target non-traditional students.⁵⁸

There has been growth in the number of qui tam lawsuits brought by private parties alleging wrongdoing at for-profit institutions, such as misleading consumers about their effectiveness by inflating job placement rates.⁵⁹ Such conduct can reasonably be expected to cause consumers to enroll and borrow, on the basis of these representations, amounts that they may not be able to repay.

In addition, a growing number of State and Federal law enforcement authorities have launched investigations into whether for-profit institutions are using aggressive or even deceptive marketing and recruiting practices that will likely result in the same high debt burdens. Several State Attorneys General have sued for-profit institutions to stop these fraudulent marketing practices, including manipulation of job placement rates. In 2013, the New York State Attorney General announced a $10.25 million settlement with Career Education Corporation (CEC), a private for-profit education company, after its investigation revealed that CEC significantly inflated its graduates’ job placement rates in disclosures made to students, accreditors, and the State.⁶⁰ The State of Illinois sued Westwood College for misrepresentations and false promises made to students enrolling in the company’s criminal justice program.⁶¹ The Commonwealth of Kentucky has filed lawsuits against several private for-profit institutions, including National College of Kentucky, Inc., for misrepresenting job placement rates, and Daymar College, Inc., for misleading students about financial aid and overcharging for textbooks.⁶² And most recently, a group of 13 State Attorneys General issued Civil Investigatory Demands to Corinthian Colleges, Inc. (Corinthian), Education Management Co., ITT Educational Services, Inc. (ITT), and CEC, seeking information about job placement rate data and marketing and recruitment practices.⁶³ The States participating include Arizona, Arkansas, Connecticut, Idaho, Iowa, Kentucky, Missouri, Nebraska, North Carolina, Oregon, Pennsylvania, Tennessee, and Washington.

Federal agencies have also begun investigations into such practices. For example, the Consumer Financial Protection Bureau (CFPB) issued Civil Investigatory Demands to Corinthian and ITT in 2013, demanding information about their marketing, advertising, and lending policies.⁶⁴ The Securities and Exchange Commission also subpoenaed records from Corinthian in 2013, seeking student information in the areas of recruitment, attendance, completion, placement, and loan defaults.⁶⁵ And, the Department itself is gathering and reviewing extensive amounts of data from Corinthian regarding, in particular, the reliability of its disclosures of placement rates.⁶⁶

This accumulation of evidence of misrepresentations to consumers by for-profit institutions regarding their outcomes provides a sound basis for the Department to conclude that a strong accountability framework for assessing outcomes by objective measures is necessary to protect consumers from enrolling and borrowing more than they can afford to repay. The same accumulation of evidence demonstrates the need for requiring standardized, readily comparable disclosures of outcomes to consumers, to enable consumers to compare programs and identify those more likely to lead to positive results.

Commenters’ claims of bias are further belied by the Department’s own data estimates. We expect that the great majority of programs, including those in the for-profit sector, will pass the D/E rates measure and comply with the other requirements of the regulations. Further, we believe that the estimated data likely overstate the number of failing and zone programs because many programs will improve outcomes during the transition period.

Of the minority of programs that we expect will not pass the D/E rates measure, a disproportionate percentage may be operated by for-profit institutions. However, since a great many more for-profit programs will in fact pass the measure, we expect students to continue to have access to GE programs operated by for-profit institutions in addition to educational options offered by public and non-profit institutions. With respect to comments that a disproportionate percentage of programs operated by for-profit institutions will not pass the D/E rates measure because they provide open enrollment admissions to low-income and underrepresented populations of students, we do not expect student demographics to overly influence the performance of programs on the D/E rates measure. Please see “Student Demographic Analysis of Final Regulations” in the Regulatory Impact Analysis for a discussion of student demographics.

Finally, we disagree with the commenters that claimed the regulations unfairly assess for-profit institutions because programs operated by for-profit institutions are in fact less expensive than programs operated by public institutions, once State and local subsidies are taken into account. While for-profit institutions may need to charge more than public institutions because they do not have the State and local appropriation dollars and must pass the educational cost onto the student, there is some indication that even when controlling for government subsidies, for-profit institutions charge more than their public counterparts. To assess the role of government subsidies in driving the cost differential between for-profit and public institutions, Cellini conducted a sensitivity analysis comparing the costs of for-profit and community college programs. Her research found the primary costs to students at for-profit institutions, including foregone earnings, tuition, and loan interest, amounted to $51,600 per year on average, as compared with $32,200 for the same primary costs at community colleges. Further, Cellini’s analysis estimated taxpayer contributions, such as government grants, of $7,600 per year for for-profit institutions and $11,400 for community colleges.⁶⁷

These regulations will help ensure that students are receiving training that prepares them for gainful employment, regardless of the financial structure of the institution they attend. Although the regulations may disproportionately affect programs operated by for-profit institutions, we believe evidence on the performance, economic costs, and business practices of for-profit institutions shows that these regulations are necessary to protect students and safeguard taxpayer funds.

Changes: None.

Comments: A few commenters suggested that, in lieu of the gainful employment regulations, the Department adopt the college ratings system and College Scorecard to apply equally across all programs.

Discussion: In addition to these regulations, the Department publishes the College Scorecard, which includes data on institutional performance that can inform the enrollment decisions of prospective students. We also plan to release the college ratings system to provide additional information for students and develop the data infrastructure and framework for linking the allocation of title IV, HEA program funds to institutional performance. Because the College Scorecard and the proposed ratings system focus on institution level performance, rather than program level performance, we do not believe it is appropriate to consider them as alternatives to these regulations for purposes of public disclosure or accountability. Further, neither of these initiatives allow for determinations of eligibility for the title IV, HEA programs as provided for in these regulations.

Changes: None.

Impact on Students

Comments: One commenter asserted that the regulations would harm millions of students who attend private sector, usually for-profit, colleges and universities and requested that the Department withdraw the proposed regulations and instead engage in meaningful dialogue with stakeholders to reach shared goals. Numerous commenters contended that the regulations are biased against programs that serve a significant number of non-traditional, underserved, low-income, and minority students and, as a result, will reduce opportunities for these students. One commenter estimated that, by 2020, the regulations will restrict the access to education of between one and two million students, and nearly four million within the next decade.

The commenters argued that students from underserved populations have greater financial need, causing them to borrow more, and typically start with lower earnings, and so will also have relatively lower earnings after completion. Several commenters submitted data or information that they believed supported this point. One commenter asserted that Pell Grant recipients are 3.8 to 5 times more likely to borrow as those who do not have Pell Grants and that, among students who complete GE programs, African-Americans and Hispanics are 22 to 24 percent more likely to borrow than whites. Another commenter referenced NCES Baccalaureate and Beyond 2008/09 data to argue that socioeconomic status at the time of college entry affects a student's debt-to-earnings ratio one year after college and that only students at public institutions in the highest quartile of income before college had debt-to-earnings ratios below 8 percent while students in the lowest quartile had debt-to-earnings ratios of about 12 percent in all types of institutions. The commenters reasoned that as a result, the programs that serve students from these populations are disproportionately likely to be failing programs. Several commenters referred to the Department’s analysis in the NPRM that the commenters believed demonstrates that a large subset of students in failing and zone programs will be female, African-American, and Hispanic. Some commenters provided additional analyses conducted at the direction of an association representing for-profit institutions asserting that much of the variance in D/E rates is associated with student demographic characteristics.⁶⁸ The commenters contended that a substantial body of research exists demonstrating a strong correlation between student characteristics and outcomes including graduation, earnings, and loan default. One commenter posited that a multivariate regression analysis conducted by the Department in 2012 showed that race, gender, and income were all significant characteristics in predicting degree completion, with the odds of completing a degree 32 percent lower for male students, 43 percent lower for Black students, and 25 percent lower for Hispanic students. Other commenters pointed to an article noting that the overall B.A. graduation rate at private non-profit colleges in 2011 was 52 percent, but for institutions with under 20 percent of students receiving Pell Grants, the graduation rate was 79 percent, while for institutions with more than 60 percent of students receiving Pell Grants, the B.A. graduation rate was 31 percent.

According to the commenters, as a result of the regulations, students from underserved populations would be forced to either forego postsecondary education or instead attend passing programs, and the performance of those passing programs would be harmed by the increases in debt and decreases in earnings due to the shift in the composition of enrolling students. They also argued that educational opportunities for low-income and minority students would be reduced because both the Department’s and third-party analyses project that most of the programs that would lose eligibility for title IV, HEA program funds under the regulations would be programs offered by for-profit institutions, which serve a large number of these students. One commenter estimated the racial and ethnic composition of students in ineligible programs: between 25 and 40 percent of African‐American students, and between 21 and 39 percent of Hispanic students who are enrolled in GE programs would be in ineligible programs. Similarly between 24 and 41 percent of female students, between 32 and 46 percent of veteran students, and between 26 and 46 percent of Pell-eligible students would be in ineligible programs. Two commenters referred to the impact on the Latino community in particular, claiming that nearly 840,000 Latinos in Orange and Los Angeles counties alone will be denied access to community colleges over the next ten years because there are not enough programs to address growing demand in the Los Angeles metropolitan area.

Some commenters expressed concern that the regulations would create incentives for for-profit institutions to decrease access to low-income and minority students. At the same time, they argued, community colleges would not by themselves have the capacity to meet the increased demand resulting from this decreased access, and from programs that become ineligible, at for-profit institutions. The commenters suggested that community colleges are not flexible enough in course scheduling and other areas to accommodate many non-traditional and adult students and are not nimble enough to quickly adjust to labor market changes. Accordingly, they said, the regulations run counter to the goal of increasing educational opportunities for all students, not just those in socioeconomic and demographic groups that tend to enter into high-earning occupations, and, over time, the regulations would not improve the situations of students from underserved populations.

Commenters argued that the regulations, and the accountability metrics in particular, should factor in the effect of these and other student characteristics on outcomes. Some commenters suggested that the Department estimate earnings gains using regression-based methods that take into account student characteristics, while others suggested applying different D/E rates thresholds to each program, based on student characteristics, such as the percentage of students receiving a Pell Grant. Commenters cited an analysis conducted at the direction of an association representing for-profit institutions that focused on a subset of programs providing training for healthcare-related professions that they claimed showed student demographics are stronger predictors of GE program outcomes on the D/E rates and pCDR measures than the quality of program instruction.⁶⁹ The commenters said that these findings contradicted the analysis conducted by the Department. Other commenters said minority status and Pell Grant eligibility, in particular, are factors that significantly affect completion, borrowing, and default outcomes. Another commenter argued that the statutory provisions that allow an institution with high cohort default rates to appeal the determination of ineligibility if it serves a high number of low-income students are evidence that Congress intended to recognize that student demographics are unrelated to program quality. As such, the commenter suggested that student demographics should be taken into account in the regulations.

Specifically with respect to the pCDR measure, commenters argued that its use as an eligibility metric would hold institutions and programs accountable for factors beyond their control, including the demographics of their students and the amounts they borrowed. The commenters argued that, in the context of iCDR, data publically available through FSA and NCES show a strong relationship between a failing iCDR and high usage of Pell Grants (an indicator of students’ low-income status), and demonstrate a strong relationship between a failing iCDR and minority status. The commenters believed that outcomes under the pCDR measure would similarly be tied to students’ socioeconomic and minority statuses, resulting in less institutional willingness to enroll minority, low-income students or students from any subgroup that shows increased risk of student loan defaults.

One commenter stated that the regulations would have a negative effect on minority students because, on average, they do not have the existing financial resources to pay for more expensive programs and, thus, rely on debt to pay for programs leading to well-paying jobs such as medical programs. The commenter asserted that the regulations would restrict access to those programs for minority students and therefore increase disparities in economic opportunity between whites and minorities. Another commenter said the regulations are biased against institutions enrolling more first-generation college students, because these students, on average, have fewer financial resources, rely more on borrowing, and are less likely to complete the program.

On the other hand, several commenters argued that the regulations would help increase access to high-quality postsecondary education for underserved students. Based on the experience of financial aid programs--such as the Cal Grants program in California--that have tightened standards for institutions receiving State-funded student aid, commenters believed that the regulations are likely to direct more funds to programs producing positive student outcomes. They predicted that the redirection of public funding will encourage programs with strong performance to expand enrollment to meet the demands of students who would otherwise attend programs that are determined ineligible under the D/E rates measure or are voluntarily discontinued by an institution. They also argued that the regulations would encourage low-quality programs to take steps to improve outcomes of non-traditional students. One commenter predicted large financial gains for low-income and minority students who enroll in better performing programs.

Discussion: We do not agree that the regulations will substantially reduce educational opportunities for minorities, economically disadvantaged students, first-generation college students, women, and other underserved groups of students. We further disagree that the available evidence suggests that the D/E rates measure is predominantly a measure of student composition, rather than program quality. As provided in the Regulatory Impact Analysis, the Department’s analysis indicates that the student characteristics of programs do not overly influence the performance of programs on the D/E rates measure. See “Student Demographic Analysis of Final Regulations” in the Regulatory Impact Analysis for a discussion of the Department’s analysis.

For these regulations, the Department modified the two regression analyses it developed for the NPRM to better understand the extent to which student demographic factors may explain program performance under the regulations. As with the NPRM, the regression analyses are based on the 2012 GE informational D/E rates. We summarize the regression analysis for the annual earnings rate here.

For the annual earnings rate regression analysis, we explored the influence of demographic factors such as those cited by commenters. These were measured at the program level for the percentage of students who completed a program and have the following demographic characteristics: zero expected family contribution estimated by the FASFA; race and ethnicity status (white, Black, Hispanic, Asian or Pacific Islander, American Indian or Alaska Native); female; independent status; married; had a mother without a bachelor’s degree. ⁷⁰ We held the effects of credential level and institutional sector of programs constant. The regression analysis shows that annual earnings rates results do not have a strong association with programs serving minorities, economically disadvantaged students, first-generation college students, women, and other underserved groups of students. Descriptive analyses, also provided in the RIA, further indicate that the characteristics of students attending GE programs are not strong predictors of which programs pass the D/E rates measure, further suggesting the regulations do not disproportionately negatively affect programs serving minorities, economically disadvantaged students, first-generation college students, women, and other underserved groups of students.

Although we included a regression analysis on pCDR in the NPRM, we do not respond to comments on this analysis because the regulations no longer include pCDR as an accountability metric to determine eligibility for title IV, HEA program funds.

Changes: None.

Comments: Several commenters asserted that the problems associated with low completion rates and churn would not be resolved if low-income and minority students who are attending failing programs at for-profit institutions transfer to programs at community colleges. According to these commenters, completion rates are lower at public two-year institutions than at for-profit two-year institutions.

Discussion: We disagree that the regulations will negatively affect the completion rates of low-income and minority students if, as a result of the regulations, more of these students transfer to public two-year institutions. As stated previously in this section, we acknowledge six-year certificate/degree attainment rates may be slightly lower at public two-year institutions compared to for-profit two-year institutions. However, we believe this slight difference in attainment rates is too small to provide compelling evidence that these regulations will harm low-income or minority students due to a possible shift in enrollment to public institutions. Further, as also discussed previously, one possible factor that contributes to graduation rates at public two-year institutions being lower than graduation rates at for-profit two-year institutions is that a goal of many community college programs is to prepare students to transfer from public two-year institutions into programs offered at other institutions, particularly public four-year institutions. Without taking into account transfer outcomes, differences in graduation rates among for-profit two-year institutions and public two-year institutions do not provide convincing evidence that the regulations will negatively affect completion rates.

Further, the Department would not expect that the regulations would disproportionately harm low-income or minority students, particularly where institutions raise quality to provide better outcomes for students, or where they are more selective in their admissions. Research shows that when challenged to attend more selective institutions, minority and low-income students have increased attainment, and that characteristics of institutions play a bigger role in determining student outcomes than do individual characteristics of attendees.^{71 72}

Regardless of the distinctions between programs operated by public and for-profit institutions, our estimates indicate that the substantial majority of programs at for-profit institutions will pass the D/E rates measure and we believe the net effect of the D/E rates measure will be that students will have the opportunity to enroll in programs at both public and for-profit institutions with better performance than programs that do not pass the D/E rates measure. In addition, students leaving a failing program at a for-profit institution may transfer to another for-profit program, but one that is performing well on the D/E rates measure.

Changes: None.

Comments: One commenter claimed the regulations do not adequately protect veteran students from deceptive practices by for-profit institutions that result in enrollment in low-quality programs. One commenter said that for-profit institutions increased recruiting of veterans by over 200 percent in just one year. Another commenter contended that 500,000 veterans dropped out of the top eight for-profit schools over the course of just one year.

Discussion: We appreciate the commenters’ concerns with respect to protecting students from deceptive practices by for-profit institutions. As discussed in the NPRM, the Senate HELP Committee recently investigated deceptive practices targeted at military veterans, particularly within the for-profit sector. In its report, it noted finding extensive evidence of aggressive and deceptive recruiting practices, high tuition, and regulatory evasion and manipulation by for-profit colleges in their efforts to enroll service members, veterans, and their families.⁷³

We believe that the regulations will help protect all prospective students, including veterans, from unscrupulous recruiting practices. As discussed in “§668.410 Consequences of the D/E Rates Measure” and in “§668.412 Disclosure Requirements for GE Programs,” prospective students will have the benefit of a fulsome set of disclosures about a program and its students’ outcomes to inform their educational and financial decision making. Further, prospective students will be warned under §668.410 if the program in which they intend to enroll could become ineligible based on its D/E rates for the next award year. By requiring that at least three days pass after a warning is delivered to a prospective student before the prospective student may be enrolled, the prospective student will benefit from a “cooling-off period” for the student to consider the information contained in the warning without direct pressure from the institution, and for the prospective student to consider alternatives to the program either at the same institution or another institution. Moreover, the accountability framework is designed to improve the quality of GE programs available to prospective students by establishing measures that will assess whether programs provide quality education and training that allow students obtain gainful employment and thereby to pay back their student loan debt. The certification requirements in §668.414 will ensure that a program eligible for title IV, HEA program funds meets certain basic minimum requirements necessary for students to obtain gainful employment in the occupation for which the program provides training. Finally, by conditioning a program’s continuing eligibility for title IV, HEA program funds on its leading to acceptable student outcomes, we believe that the D/E rates measure will help ensure that prospective students, including veterans, will be less likely to enroll in a low-quality GE program.

Changes: None.

Accountability Metrics

Comments: A number of commenters opposed the Department’s proposal to use the D/E rates measure and the pCDR measure for accountability purposes. These commenters also offered suggestions for alternative metrics the Department should consider adopting in the final regulations.

D/E Rates Measure

Many commenters stated that the Department should not use the D/E rates measure as an accountability metric because it is flawed and, more specifically, would not capture the lifetime earnings gains that arise from attending a GE program. Without knowing lifetime earnings, these commenters contended, it is difficult to assess what an appropriate amount of debt is or whether a program is providing value to students. They asserted that the standard way to evaluate whether it is worthwhile to attend a postsecondary education program is to compare the full benefits against the cost. Consequently, they reasoned that the D/E rates measure is faulty because it only captures earnings after a short window of time.

Several commenters offered studies that show that a college degree leads to an annual increase in wages of somewhere between 4 to 15 percent. One commenter stated that the earnings premium between a high school graduate and a college graduate is lowest from ages 25-29 but peaks from ages 45-54. One commenter asserted that, based on an institutional survey of students five years after their graduation from associate and bachelor’s degree programs that compared the students’ initial 2009 median salaries to their 2014 median salaries, the students’ salaries increased about 50 percent over the first five years after graduation. Thus, the commenter suggested that the regulations consider earnings no less than five years after graduation for the calculation of D/E rates.

Commenters also expressed concern that the earnings assessed by the D/E rates measure do not include other returns from higher education, such as fringe benefits, contributions to retirement accounts, subsidized insurance, paid vacations, and employment stability. Further, they contended that the D/E rates measure does not account for the social benefits that accrue to students, in addition to the pecuniary benefits. Other commenters posited that the benefits of higher education have generally trended upwards over time and so the D/E rates measure understates the future benefits of programs that provide training for occupations in growing fields, such as health care.

One commenter suggested that as a result of differences in what for-profit institutions, as opposed to community colleges, receive in the form of State subsidies, and because for-profit institutions pay taxes, any accountability metrics should be divorced from the tuition charged, and should instead focus on the earnings increase resulting from increased education, completion rates at institutions, or job or advanced degree placement rates.

Finally, one commenter claimed the D/E rates measure is not valid because it is not predictive of default outcomes. Based on the 2012 GE informational rates, the commenter claimed students in programs in the lowest performing decile under the D/E rates measure were still more than four times as likely to be in repayment than in default. The commenter stated that, if the D/E rates measure were truly an indicator of affordability, there would have been much higher default and forbearance rates for students in programs with the highest D/E rates.

pCDR Measure

A number of commenters also opposed the Department’s proposal to include pCDR as an accountability metric, arguing that this metric is largely unrelated to whether a program prepares students for gainful employment. Several commenters argued that the Department lacks the legal authority to adopt pCDR to determine GE program eligibility and contended that the use of cohort default rates to assess program eligibility is contrary to the intent of Congress, because Congress never explicitly authorized the Department to use cohort default rates to assess program eligibility. The same commenters further contended that the history of congressional attention to the iCDR eligibility standard over the years, applied with periodic amendments, reflected Congress’s intent that cohort default rates be used only for institutional eligibility determinations, and left no room for the Department to apply that test for programmatic eligibility. Similarly, they contended that Congress’s choice to apply cohort default rates as an eligibility standard for all institutions receiving title IV, HEA program funds indicated a congressional intent that such a test should not be applied only to a subset of institutions--chiefly, for-profit schools.

Some commenters contended that the ruling by the court in APSCU v. Duncan requires the Department to base any program eligibility standard on expert studies or industry practice, or both. Because the Department did not cite to such support in the NPRM for adopting the iCDR methodology and the institutional eligibility threshold to determine program eligibility, these commenters believed the Department was barred from using cohort default rates to determine programmatic eligibility. Commenters contended that the Department provided no reasoned explanation in the NPRM for the proposed use of cohort default rates at the program level.

Commenters also asserted that the Department provided no reasoned basis for adopting a 30 percent cohort default rate as the threshold for program eligibility for title IV, HEA program funds. They asserted that the Department failed to consider the bases on which Congress, in its 2008 amendments to the HEA, increased the iCDR threshold rate from 25 percent to 30 percent. They argued that Congress, in amending the HEA to count defaults over a three-year term and raising the iCDR eligibility standard to 30 percent, recognized that setting a lower standard would deter institutions from enrolling “minority, low-income students, or any subgroup that shows any risk of more defaults on student loans.” The commenters conceded that iCDR was an important way to protect the Federal fiscal interest, but asserted that Congress did not consider iCDR to be a measure of educational quality, and that Congress did not consider rates greater than 30 percent to be evidence that institutions were not preparing their students adequately.

Several commenters asserted that measures of default like pCDR reflect personal decisions by individual borrowers, specifically whether or not to repay their debt, and not the performance of a program. Other commenters stated that institutions cannot control how much students borrow, or need to borrow. In this regard, commenters noted that, although institutions can control the cost of attendance, they cannot control other factors contributing to borrowing behavior, such as living expenses and the student’s financial resources at the time of enrollment, and that institutions have only a limited ability to affect repayment once a student has left the institution.

Some commenters contended that the proposed pCDR measure would impose a stricter standard than the iCDR standard on which it was based, because the iCDR standard allows offset of poor results of some programs against the more successful rates achieved by other programs offered by the institution.

While some commenters considered pCDR a poor metric for the reasons described, others expressed concern that pCDR would be a poor measure of performance because institutions could encourage students struggling to repay their debt to enter forbearance or deferment in order to evade the consequences of failing the pCDR measure. They stated that programs would not be held accountable for the excessive debt burden of these students because, by pushing students into deferment or forbearance during the three-year period that the pCDR measure would track defaults, any default would occur after the time during which the program would be held accountable under the proposed regulations. Several commenters expressed concern that, because the metric is subject to manipulation, the 30 percent threshold would be too lenient and should be lower, with some commenters suggesting a 15 percent threshold.

Alternative Metrics

Commenters proposed a number of alternatives to the D/E rates and pCDR measures to assess the performance of gainful employment programs. A number of commenters, arguing that both the D/E rates and pCDR measures are too tenuously linked to what institutions do to affect the quality of training students receive, encouraged the Department to consider metrics more closely linked to student academic achievement, loan repayment behavior, or employment outcomes like job placement rates. Commenters proposed alternative metrics that they felt better account for factors that are largely outside of programs’ control, such as fluctuations in local labor market conditions. Some commenters suggested that alternative metrics should be tailored to measure student outcomes in specific occupational fields, such as cosmetology or medical professions. For example, several commenters said the Department should use licensure exam pass rates and residency placement rates in tandem to evaluate medical schools. They said these metrics would take into account occupational preparedness and are better metrics than the D/E rates measure because earnings rise steadily across long periods of time among students completing medical degrees. On the other hand, one commenter expressed concern about job placement rates as a metric because there are no standard definitions of placement, national accreditation agencies each have different methodologies, and regional accreditation agencies do not require rates be reported.

A few commenters said programs should be evaluated according to metrics focusing on student success in a program. Commenters suggested the Department consider retention and graduation rates as alternative metrics, as completion of a degree or certificate program is closely linked to whether students obtain employment. One commenter criticized the Department for not including a graduation rate metric in the regulations because, based on GE informational rates, for-profit institutions with default rates higher than graduation rates have a very large percentage of programs that do not graduate enough students to meet the n-size requirements for D/E rates to be calculated. The commenter noted a similar pattern among some community colleges with very low graduation rates. The commenter also arrived at the same conclusion based on a study conducted by College Measures, a non-profit organization, which examined GE programs at 1,777 two-year public and for-profit institutions. The study referenced indicated that, among the 724 public and 24 for-profit institutions that had graduation rates below 30 percent, 29 percent of the for-profit programs with low graduation rates failed the D/E rates measure, while only 2 percent of the public institutions with low graduation rates failed the D/E rates measure. Based on this analysis, the commenter further asserted that the regulations are biased toward passing programs operated by public institutions because they do not include a graduation rate metric. According to the commenter, any program with a starting class that has fewer than 70 students and less than a 10 percent graduation rate would be automatically exempt from the regulation, even counting four years of graduates. Another commenter said the Department should focus on each program’s curriculum and other aspects of the program controlled by the institution rather than the proposed metrics.

Several commenters said the Department should include a repayment rate or a negative amortization test instead of pCDR, which they viewed as unreliable and easily manipulated by institutions. Some commenters favored using a repayment rate rather than pCDR because the former would hold programs accountable for students who go into forbearance and are unable to reduce the principal balances on their loans. Other commenters asserted that a repayment rate is a preferable metric for students who choose income-based repayment plans because under such plans, students with low incomes can avoid default even though their loans are in negative amortization, making pCDR a less reliable metric than repayment rate.

Several commenters suggested specific ways in which the Department could set a passing threshold for a repayment rate or negative amortization test. Some commenters stated that the regulations should provide that programs with more than half of loans in negative amortization would be considered failing. Several other commenters said the Department should invert the pCDR measure by failing programs with less than 70 percent of students reducing the balance on their debt. One commenter asserted that the Department should include a repayment rate metric based on the repayment definition from the 2011 Prior Rule. The commenter suggested that 45 percent would be an appropriate passing threshold for a repayment rate based on Current Population Survey (CPS) census data that estimates that 46.2 percent of young adults with a high school diploma could possibly afford student debt payments.

Some commenters argued the Department has adequate expertise and authority, as the issuer of all Federal Direct Loans, to set a loan repayment threshold appropriate for its own loan portfolio without needing to rely on an unrelated external standard. Additionally, commenters suggested the Department convene a panel of experts to set a repayment rate threshold for the regulations. One commenter said the Department should use available data to set a repayment rate threshold that would be difficult for programs to manipulate.

A few commenters offered what they believed are limitations of relying on a repayment rate metric. One commenter said the regulations should not include a repayment rate metric because such a standard would disproportionately affect programs providing access to low-income and minority students. Another commenter suggested that if the Department includes a repayment rate metric in the regulations, it should prohibit institutions from making loan payments on students’ behalf in an attempt to increase the proportion of students counted as successfully in repayment. As an alternative to pCDR or a repayment rate metric, one commenter proposed that the regulations evaluate iCDR and the percentage of enrolled students borrowing to set an eligibility standard that would identify and curtail abuses in the short run and suspend program participation if both iCDR and borrowing rates are high.

Some commenters believed that, if the 90/10 provisions in section 487(a)(24) of the HEA limiting the percentage of revenue for-profit institutions may receive from title IV, HEA programs were eliminated, there would be no need for the D/E rates measure. Several commenters said the 90/10 provisions should be modified to include GI benefits and other Federal sources of aid. Some commenters argued that the 90/10 provisions should be modified to provide for an 85/15 ratio such that a for-profit institution receiving more than an 85 percent share of revenue from title IV, HEA programs and other Federal programs would be determined ineligible to participate in the title IV, HEA programs.

Other commenters asked the Department to set standards that would cap the prices charged or amount of loans disbursed for different kinds of programs. For instance, one commenter proposed that loan disbursements could be capped for all cosmetology programs based on the average earnings of individuals who enter the field.

Several commenters contended the Department should use risk-adjusted lifetime earnings gains net of the average cost of program attendance as an alternative metric. One commenter suggested that the regulations consider earnings before and after attendance in a program in order to measure program success. The commenter also argued that the amount of debt incurred should not be used to measure the success of a program.

Discussion:

D/E Rates Measure

Although the creation of a program “value added” measure using some function of earnings gains may provide some information on program quality, we disagree that it is more appropriate than the D/E rates measure as a basis for an eligibility standard. We do not believe it is aligned with the accountability framework of the regulations, which is based on discouraging institutions from saddling students with unmanageable amounts of debt. Furthermore, the commenters have failed to establish an appropriate standard supported in the research that demonstrates how such a measure could be used to determine whether a program adequately prepares students for gainful employment in a recognized occupation.

The accountability framework of the regulations focuses on whether students who attend GE programs will be able to manage their debt. As we discussed in the NPRM, the gainful employment requirements are tied to Congress’ historic concern that vocational and career training offered by programs for which students require loans should equip students to earn enough to repay their loans. APSCU v. Duncan, 870 F.Supp.2d at 139; see also 76 FR 34392. Allowing students to borrow was expected to neither unduly burden the students nor pose “a poor financial risk” to taxpayers. In authorizing federally backed student lending, Congress considered expert assurances that vocational training would enable graduates to earn wages that would not pose a “poor financial risk” of default.

Congress’ decision in this area is supported by research that shows that high levels of debt and default on student loans can lead to negative consequence for borrowers. There is some evidence suggesting that high levels of student debt decrease the long-term probability of marriage.⁷⁴ For those who do not complete a degree, greater amounts of student debt may raise the probability of bankruptcy.⁷⁵ There is also evidence that high levels of debt increase the probability of being denied credit, not paying bills on time, and filing for bankruptcy--particularly if students underestimate the probability of dropping out.⁷⁶ Since the Great Recession, student debt has been found to be associated with reduced home ownership rates.⁷⁷ And, high student debt may make it more difficult for borrowers to meet new mortgage underwriting standards, tightened in response to the recent recession and financial crisis.⁷⁸

Further, when borrowers default on their loans, everyday activities like signing up for utilities, obtaining insurance, and renting an apartment can become a challenge. Such borrowers become subject to losing Federal payments and tax refunds and wage garnishment.⁷⁹ Borrowers who default might also be denied a job due to poor credit, struggle to pay fees necessary to maintain professional licenses, or be unable to open a new checking account.⁸⁰ As a responsible lender, one important role for the Department is to hold all GE programs to a minimum standard that ensures students are able to service their debt without undue hardship, regardless of whether students experience earnings gains upon completion.

Research has consistently demonstrated the significant benefits of postsecondary education. Among them are private pecuniary benefits⁸¹ such as higher wages and social benefits such as a better educated and flexible workforce and greater civic participation.^{82 83 84 85} Even though the costs of postsecondary education have risen, there is evidence that the average financial returns to graduates have also increased.⁸⁶

We recognize the value of programs that lead to earnings gains and agree that gains are essential. However, we believe that the D/E rates measure, rather than a measure of earnings gains, better achieves the objectives of these regulations because it assesses earnings in the context of whether they are at a level that would allow borrowers to service their debt without serious risk of financial or emotional harm to students and loss to taxpayers.

We also disagree with commenters who claim a low correlation between D/E rates and default undermines D/E rates as an indicator of financial risk to students. As our discussion of the D/E rates thresholds provides in more detail, our analyses indicate an association between ultimate repayment outcomes, including default, and D/E rates. Based on the best data available to the Department, graduates of programs with D/E rates above the passing thresholds have higher default rates and lower repayment rates than programs below the thresholds. Although many other factors may contribute to default outcomes, we believe high D/E rates are an important indicator of financial risk and possibility of default on student loans. In addition to addressing Congress’ concern of ensuring that students’ earnings would be adequate to manage their debt, research also indicates that debt-to-earnings is an effective indicator of unmanageable debt burden. An analysis of a 2002 survey of student loan borrowers combined borrowers’ responses to questions about perceived loan burden, hardship, and regret to create a “debt burden index” that was significantly positively associated with borrowers’ actual debt-to-income ratios. In other words, borrowers with higher debt-to-income ratios tended to feel higher levels of burden, hardship, and regret.⁸⁷

Further, although annual earnings may increase for program graduates over the course of their lives as a result of additional credentialing, the Department disagrees that this fact undermines the appropriateness of determining eligibility based on the D/E rates measure. Borrowers are still responsible for managing debt payments, which begin shortly after they complete a program, even in the early stages of their career.

Repayment under the standard repayment plan is typically expected to be completed within 10 years; the return on investment from training may well be experienced over a lifetime, but benefits ultimately available over a lifetime may not accrue soon enough to enable the individual to repay the student loan debt under and within the schedules available under the title IV, HEA programs. These regulations evaluate debt service using longer repayment terms than the typical 10-year plan, taking into account our experience with the history of actual borrower repayment and the use of forbearances and deferment. However, even the extended repayment expectations we use to amortize debt under the D/E rates measure (10, 15, and 20 years for non-baccalaureate credentials, baccalaureate and master’s degrees, and doctoral or professional degrees, respectively) do not encompass a lifetime of benefits. Rather, we believe it is important to measure whether the ratio of debt to earnings indicates whether a student is able to manage debt both in the early years after completion, and in later years, since students must be able to sustain loan payments at all stages, regardless of the benefits that may accrue to them over their entire career.

pCDR Measure

As we discussed in the NPRM, the Department’s proposal to include pCDR as a measure of whether a program prepares students for gainful employment in a recognized occupation is, like the D/E rates measure, grounded both in statute and legislative history. We included the pCDR measure as an accountability metric in the proposed regulations because it would measure actual repayment outcomes and because it would assess the outcomes of both students who completed a GE program and those who had not. Both reasons are responsive to the concerns of Congress in making the student aid loan programs available to students in career training programs. As previously discussed, the legislative history regarding GE programs shows that Congress considered these programs to warrant eligibility on the basis that they would produce skills and, therefore, earnings at a level that would allow students to manage their debt. This concern extended not only to students who completed a program, but also to those who transferred or dropped out of a program. Accordingly, to measure whether a program is leading to unmanageable debt for both students who complete a program and those who do not, we proposed adopting the identical eligibility threshold for pCDR that Congress established for iCDR.

The Department strongly believes in the importance of holding GE programs accountable for the outcomes of students who do not complete a program and ensuring that institutions make strong efforts to increase completion rates. As previously discussed, many commenters offered alternate metrics for the Department to consider adopting, including those that would measure the outcomes of students who do not complete their programs. Given the wealth of feedback we received on this issue through the comments, we believe further study is necessary before we adopt pCDR or another accountability metric that would take into account the outcomes of students who do not complete a program. We also believe further study is necessary before adopting other metrics based on CDR, including “borrowing indices” that take into account iCDR and the percentage of students who take out loans at the institution. Using the information we will receive from institutions through reporting, we will continue to develop a robust measure of outcomes for students who do not complete a program, which may include some measure based on repayment behavior. Because pCDR has been removed as an accountability metric, we do not specifically address the comments related to its operation for accountability purposes.

Despite our decision not to use pCDR as an accountability metric, we continue to believe in the importance of holding GE programs accountable for the outcomes of students who do not complete a program and ensuring that institutions make strong efforts to increase completion rates. Default rates are important information for students to consider as they decide where to pursue, or continue, their postsecondary education and whether or not to borrow to attend a particular program. Accordingly, we are retaining pCDR as one of the disclosures that institutions may be required to make for GE programs under §668.412. We believe that requiring this disclosure, along with other potential disclosures such as completion, withdrawal, and repayment rates, will bring accountability and transparency to GE programs with high rates of non-completion.

Alternative Metrics

We appreciate the suggestions to use retention rates, employment or job placement rates, and completion rates as alternative measures to the D/E rates measure. While these are all valid and useful indicators for specific purposes, there is no evidence that any of these measures, by themselves, indicates whether a student will be likely to repay his or her debt. For example, placing a student in a job related to the training provided by a program is a good outcome, but without considering any information related to the student’s debt or earnings, it is difficult to say whether the student will be able to make monthly loan payments. We also disagree that the D/E rates measure is tenuously linked to the performance of programs because it does not take into account these alternative metrics. We believe the measure appropriately holds programs accountable for whether students earn enough income to manage their debt after completion of the program.

We do not agree that, without a graduation rate metric, poorly performing programs will not be held accountable under the regulations due to having an insufficient number of students who complete the programs to be evaluated under the D/E rates measure. First, in order to address this concern, we calculate the D/E rates measure over a four-year cohort period for small programs in order to make it more likely that programs with low graduation rates are evaluated. Second, although the regulations do not include pCDR as an accountability metric, they will require programs to disclose completion rates and pCDR to students and we believe these disclosure items will help students and families make more informed enrollment decisions. Third, as previously stated, the focus of the D/E rates measure is to hold programs accountable for whether students are able to manage their debt after completion, and we do not believe it is appropriate to base eligibility for title IV, HEA program funding on a metric, such as graduation rate, that does not indicate whether a student will be likely to repay his or her debt.

We disagree with comments suggesting we tailor alternative metrics to measure student outcomes in specific occupational fields, such as cosmetology or medical professions. It is neither feasible nor appropriate to apply different metrics to different kinds of programs. By itself, the occupation an individual receives training for does not by itself determine whether debt is manageable. Rather, it is related to the debt that the individual accumulates and the earnings achieved as a result of the program’s preparation--exactly what the D/E rates measure assesses.

Similarly, we believe it is inappropriate to rely on licensure exam pass rates and residency placement rates to evaluate medical programs and other graduate programs. There is no evidence that any of these measures, by themselves, would indicate whether a student will be likely to be able to repay his or her debt.

We also disagree that programs should be evaluated according to each program’s curriculum and other aspects of the program controlled by the institution rather than under the D/E rates measure. Although factors such as program curriculum and quality of instruction may contribute to the value of the training students receive, other factors such as earnings and student debt levels affect whether students are able to manage their debt payments after completion. Accordingly, we believe it is more appropriate to evaluate programs based on the outcomes of their students after completion, rather than the curricular content or educational practices of the institutions operating the programs.

We continue to believe that a repayment rate metric is an informative measure of students’ ability to repay their loans and an informative measure of outcomes of both students who do and do not complete a program. However, as discussed in the NPRM, we have been unable to determine an appropriate threshold for distinguishing whether a program meets the minimum standard for eligibility. We have not identified any expert opinion, nor has any statistical analysis demonstrated, that a particular level of repayment should serve as an eligibility standard. We appreciate suggestions for repayment rate thresholds of 70 percent and 45 percent. Commenters indicated 70 percent may be appropriate because it seems to correspond to 100 percent minus 30 percent, the threshold for iCDR. We do not believe this rationale is sufficient as repayment rate reflects the percentage of students reducing the principal on their loans, rather than the percentage of students avoiding default. The commenter who recommended 45 percent relied on Census data for justification. However, we have been unable to identify any specific support in the Census data for this proposition.

The Department’s status as lender does not eliminate the need to support any standard adopted to define eligibility. As a result, we decline to adopt a repayment-based eligibility metric at this time.

Similarly, we lack expert opinion or statistical analysis that would support other metrics and thresholds based on borrower repayment. For example, we are unable to identify expert opinion or statistical analysis that supports negative amortization as a metric, or the proposed 50 percent threshold, as an appropriate measure for whether students are able to manage their debt. Some students who have chosen income-based or graduated repayment plans may be able to manage their debt payment, but are observed as being in negative amortization. On the other hand, students who reduce the principal on their debt may be earning too little to manage their debt without experiencing financial hardship.

Finally, with respect to suggestions that the 90/10 provisions should be modified, we note that such changes are beyond the Department’s regulatory authority because the 90/10 requirements are set in statute. Moreover, even if the Department had authority to change the 90/10 provisions, we do not believe doing so would serve the purposes of these regulations. First, the 90/10 provisions measure the revenues of institutions, not students’ ability to repay debt accumulated as a result of enrolling in a GE program. Second, the provisions apply only to for-profit institutions and could not be equally applied to GE programs in other sectors.

Changes: We have removed pCDR as an accountability metric. Other changes affecting the use of pCDR as a disclosure item are discussed in “§668.413 Calculating, Issuing, and Challenging Completion Rates, Withdrawal Rates, Repayment Rates, Median Loan Debt, Median Earnings, and Program Cohort Default Rate.”

Because the final regulations include only the D/E rates measure as an accountability metric, we have removed the term and definition of “GE measures” from §668.402.

Comments: Commenters posited that because the D/E rates measure does not measure actual benefits, it would have the effect of artificially reducing program prices and, as a result, lowering quality and academic standards.

Discussion: The Department disagrees that the D/E rates measure will result in GE programs with lower educational quality or less rigorous academic standards than they would have in the absence of the regulations. According to our data, the great majority of GE programs in all sectors will pass the D/E rates measure. Hence, most programs will not have to lower their prices as a result of the D/E rates measure.

Programs with high D/E rates will have several ways to ensure that the performance of their programs meet the standards of the regulations while maintaining or improving the quality of the training they provide, such as: providing financial aid to students with the least ability to pay in order to reduce the number of students borrowing and the amount of debt that students must repay upon completion; improving the quality of the vocational training they offer so that students are able to earn more and service a larger amount of debt; and decreasing prices for students and offsetting any loss in revenues by reducing institutional or program expenditures in areas not affecting programs quality, such as administrative overhead, recruiting, and advertising.

Changes: None.

Comments: One commenter asserted that short periods of attendance at GE programs may provide students with benefits not measured by the D/E rates or pCDR measures because underserved students can still acquire some skills even if they do not complete their program. The commenter argued that the regulations should recognize the benefits associated with partial completion of a program as a positive outcome by relying on a metric that measures incremental increases in the net present value of earnings. The commenter stated that the proposed regulations would not accomplish this because the D/E rates measure does not include the outcomes of students who do not complete a program and the pCDR measure punishes all “churn,” regardless of whether partial completion may have some positive benefits.

Discussion: We do not agree that the regulations should specifically recognize partial completion. Although students, including those from underserved backgrounds, may gain some benefit from attending a GE program even if they do not complete, we do not believe that some other negative outcome, such as high debt burden in the case of the D/E rates measure, should be ignored. Further, these students would presumably benefit even more by reaching completion.

Changes: None.

Comments: A few commenters said the D/E rates measure is flawed because it treats short-term certificate programs the same as graduate programs. The commenters said certain programs, such as certificate programs, are designed to leave graduates with little debt, but more short-term earnings gains, while graduate programs may produce larger debt levels, but have larger increases in lifetime earnings. Commenters suggested that the Department establish an alternative metric that takes into account the fact that students in professional graduate programs take out large amounts of debt but earn high enough lifetime earnings to service that debt.

Discussion: We believe that the D/E rates measure is an appropriate metric to assess all GE programs, including graduate professional programs. These regulations will help ensure that students who attend GE programs are able to manage their debt. Although graduates of professional programs may experience increased earnings later, as discussed previously, earnings must be adequate to manage debt both in the early years after entering repayment and in later years, regardless of what an individual’s lifetime earnings may be.

Further, as discussed later in this section, the discretionary income rate will help accurately assess programs that may result in higher debt that may take longer to repay but also provide relatively higher earnings. Also, as discussed in “§668.404 Calculating D/E Rates,” the regulations apply a relatively longer 20-year amortization period to the D/E rates calculation for graduate programs, and assess earnings for medical and dental programs at a later time after completion to account for time in a required internship or residency.

Changes: None.

D/E Rates Thresholds

Comments: Some commenters suggested that the D/E rates thresholds should be those established in the 2011 Prior Rule--a discretionary income rate threshold of 30 percent and an annual earnings rate threshold of 12 percent. Commenters suggested that because the D/E rates thresholds in these regulations differ from those in the 2011 Prior Rule, the D/E rates thresholds are arbitrary.

Other commenters cited studies and data in support of alternative thresholds and stated that the Department’s choice of thresholds more stringent than those they believed were supported by the studies is arbitrary and capricious, particularly in their application to the for-profit industry.

Commenters argued the 12 percent threshold for the annual earnings rate is inappropriate because, based on an NCES study, a substantial percentage of first-time bachelor’s degree recipients have an annual income rate greater than 12 percent.⁸⁸ The study analyzed earnings and debt levels collected by NCES in its 1993/94, 2000/01, and 2008/09 Baccalaureate and Beyond Longitudinal Studies Survey. According to the study, in 2009, 31 percent of bachelor’s degree recipients who borrowed and entered repayment had an annual income rate greater than 12 percent one year after graduation. Commenters noted 26 percent of recipients who borrowed at public institutions and 39 percent of recipients who borrowed at private, non-profit institutions exceeded the 12 percent threshold, suggesting the threshold for the annual earnings rate is too low. Commenters also contended the annual earnings rate threshold is inappropriately low because the same study indicated the average monthly loan payment as a percentage of income among bachelor’s degree recipients who borrowed, were employed, and were repaying their loans one year after graduation was about 13 percent in 2009.

One commenter reached similar conclusions based on a study that used Beginning Postsecondary Students Longitudinal Study (BPS) data to indicate annual earnings rates are, on average, about 10.5 percent among all bachelor’s degree recipients six years after enrollment.⁸⁹

According to some commenters, a 2010 study conducted by Mark Kantrowitz indicates that the majority of personal finance experts believe that an acceptable annual debt-to-earnings ratio falls between 10 percent and 15 percent.⁹⁰ These commenters suggested that the Department’s reliance on research conducted by Sandy Baum and Saul Schwartz in 2006 in establishing the 8 percent annual earnings rate threshold is arbitrary. The commenters stated that Baum and Schwartz acknowledge that the 8 percent threshold is based on mortgage underwriting practices, and they believe that there is not sufficient research to justify using an 8 percent annual earnings rate in the context of the regulations. Specifically, the commenters stated that Baum and Schwartz criticized the 8 percent threshold as not necessarily applicable to higher education loans because the 8 percent threshold (1) reflects a lender’s standard of borrowing, (2) is unrelated to individual borrowers' credit scores or their economic situations, (3) reflects a standard for potential homeowners rather than for recent college graduates who generally have a greater ability and willingness to maintain higher debt loads, and (4) does not account for borrowers’ potential to earn a higher income in the future. Commenters emphasized that Baum and Schwartz believe that using the difference between the front-end and back-end ratios historically used in the mortgage industry as a benchmark for manageable student loan borrowing has no particular merit or justification. The commenters believed the Department should recognize that borrowing for education costs is different from borrowing for a home mortgage because education tends to cause earnings to increase. As a result, the commenters believed the Department should increase the threshold.

Some commenters contended that the research by Baum and Schwartz also suggests that increased burden beyond the 8 percent annual earnings rate may be a conscious choice by those early in a career to take on increased burden and that the research justifies an annual earnings rate threshold of 12 to 18 percent, and a discretionary income rate threshold of 30 to 45 percent as “reasonable.”⁹¹One commenter said the Department could arrive at an annual earnings rate threshold higher than 8 percent using a methodology similar to the one cited by the Department in the NPRM. Specifically, the commenter said a higher threshold is justified by regulations issued by the Consumer Financial Protection Bureau (CFPB) that became final on January 10, 2014, defining the total debt service-to-earnings ratio at 43 percent for the purpose of a qualified mortgage. Moreover, the commenter cited the 2008 consumer expenditures survey showing that, on average, associate degree recipients pay 27 percent of income and bachelor’s degree recipients pay 25 percent of income toward housing costs, including mortgage principal and interest. Thus, the commenter said this would yield 16 percent and 18 percent of income available to pay for other debt, such as education-related loans. The commenter also asserted a higher annual earnings rate threshold is warranted because some mortgage lenders use a 28 percent to 33 percent threshold for mortgage debt, which still leaves 10 percent to 15 percent of income available for other debt.

Some commenters suggested that the Department should base the annual earnings rate threshold on a 2003 GAO study “Monitoring Aid Greater Than Federally Defined Need Could Help Address Student Loan Indebtedness” (GAO-03-508).⁹² Commenters said that the GAO study indicated that 10 percent of first-year income is the generally agreed-upon standard for student loan repayment and that the Department itself established a performance indicator of maintaining borrower indebtedness and average borrower payments for Federal student loans at less than 10 percent of borrower income in the first repayment year in the Department’s “FY 2002 Performance and Accountability Report.”⁹³

One commenter suggested that title IV, HEA program funds that students use to pay room and board costs should be factored into the D/E rates calculations because these funds are allowed to be used for those purposes and schools may be tempted to shift costs between tuition and room and board in order to create more favorable D/E rates. The commenter proposed that if these costs are factored into the D/E rates calculations, the passing thresholds should be increased from 8 percent to 15 percent for the annual earnings rate and from 20 percent to 30 percent for the discretionary income rate.

One commenter criticized the D/E rates measure and the thresholds of 8 and 20 percent because they would be sensitive to changes in the interest rate. The commenter explained that an increase in the interest rate would yield a lower maximum allowable total annual debt service amount as a percentage of annual earnings, since the monthly payment will be higher. For example, the commenter noted that an increase in the loan interest rate to 6.8 percent would increase the annual debt service amount, and therefore the debt-to-annual earnings ratio of a program, significantly, making it more difficult for institutions to pass the D/E rates measure.

Some commenters suggested that the 8 percent annual earnings rate and 20 percent discretionary income rate are too high to support sustainable debt levels. Commenters suggested that the annual earnings rate threshold is too high because, as Baum and Schwartz explained, a supportable annual earnings rate of 8 percent assumes that all non-housing debts do not exceed 8 percent of annual income. Commenters suggested that all other debts, including, but not exclusively, student loan debts, should be included in that 8 percent threshold, and, thus, the Department should provide a buffer to borrowers with other debts and investments to ensure sustainable debt levels. Other commenters suggested that the D/E rates thresholds are too high because they do not account for other educational costs (beyond tuition, fees, books, supplies, and equipment) which may limit students’ ability to repay debt.

In recommending that the annual earnings rate threshold be strengthened, some commenters noted that allowing a passing threshold of up to 8 percent for student loan debt alone already fails to account for a student’s other debts, but allowing up to 12 percent before a program is failing the D/E rates measure is without a sound rationale and should be eliminated from the regulations after a phase-in period.

Commenters also noted that a student’s debt is likely to be understated because the same interest rate that is used for calculating the annual debt service for Federal Direct Unsubsidized loans would also be used to calculate the debt service of private education loans, which are used more by students attending for-profit institutions, and which typically have rates equal to, or higher, than the Direct Unsubsidized loan rate. For these reasons, the commenters argued that the Department should avoid using any threshold higher than 8 percent of annual earnings.

With respect to the discretionary income rate threshold, commenters suggested that changes made by section 2213 of the Student Aid and Fiscal Responsibility Act (SAFRA) to lower the cap on allowable income-based repayments from 15 percent to 10 percent of discretionary income support a lower discretionary income rate threshold.⁹⁴ Furthermore, commenters stated that the 20 percent discretionary income rate threshold recommended by Baum and Schwartz provides an absolute maximum discretionary income rate that anyone could reasonably pay and that should never be exceeded. Accordingly, the commenters contended that the discretionary income rate thresholds for the D/E rates measure are far too high.

Discussion: We do not agree with the commenters that argued for passing D/E rates thresholds of 12 percent of annual earnings and 30 percent of discretionary income, rather than 8 percent and 20 percent. Instead, we establish 12 percent and 30 percent as the upper boundaries of the zone. Although these thresholds differ from those established in the 2011 Prior Rule, they are supported by a reasoned basis as we outlined in the NPRM and in the following discussion.

We first clarify the difference between the term “debt” as used in the D/E rates measure and as used in the literature and opinions on which those commenters who consider the D/E rates thresholds too strict rely. In connection with the 2011 Prior Rule and during the negotiated rulemaking process for these regulations, institutional representatives repeatedly stressed the inability of institutions to control the amount of debt that their students incurred.⁹⁵ In response to that concern, in §668.404(b)(1) of the regulations, the Department limits the amount of debt that will be evaluated under the D/E rates measure to the amount of tuition and fees and books, supplies, and equipment, unless the actual loan amount is smaller--in which case the Department evaluates the actual loan amount, including any portion taken out for living expenses. Thus, the D/E rates measure will typically capture, as a commenter noted, not the actual total student debt, but only a portion of that debt—up to the amount of direct charges. The commenters cite analysis and authority opining that the appropriate levels of student loan debt that borrowers can manage are in the range of 10 percent to 15 percent of annual income.⁹⁶ That position is not inconsistent with the standard we adopt here because those opinions address the actual student loan debt that borrowers must repay--what could be called the borrower’s real debt burden. That approach is reasonable when addressing actual borrower debt burden, and it is the Department’s approach when calculating the debt burden for an individual student borrower in other regulations. See, e.g., section 2213 of the SAFRA and 34 CFR 685.209. In contrast, the D/E rates measure assesses aggregate debt burden for a cohort of borrowers, and does so using a formula that holds the institution accountable only for the borrowing costs under its control--tuition, fees, books, equipment, and supplies. Accordingly, we decline to raise the annual earnings rate threshold to 12 percent and discretionary income rate threshold to 30 percent to capture the total amount borrowed; and we also decline to lower the rates to below 8 percent and 20 percent, respectively, to account for the exclusion of other debt.

In reference to the comment suggesting that title IV, HEA program funds that students use to pay room and board costs should be factored into the D/E rates calculations, we continue to believe that, for the purpose of the D/E rates measure, loan debt should be capped at the amount charged for tuition and fees and books, supplies, and equipment, because those costs are within an institution’s control. We do not believe that it is reasonable to include room and board charges in the amount at which loan debt is capped. Unlike tuition and fees, books, equipment, and supplies, costs which all students must pay for, room and board are within the choice of the student, and their inclusion runs counter to the general position that we hold schools accountable under these metrics for those costs that are under their control. Costs of room and board – or allowance for room and board, for students not in institutional housing – vary from institution to institution, depend on the housing choices actually available to, as well as the choices within those options of, individuals, and even the locale of the available housing choices. Including room and board would not only appear impracticable but difficult to implement in a manner that treats similar or identical programs in an evenhanded manner for accountability purposes as well as disclosure purposes.

We also disagree with the commenters who believe the failing thresholds should be lower because the debt payment calculations do not take into account debt other than student loan debt. Because of the substantial negative consequences associated with a program’s loss of title IV, HEA program eligibility, we believe it is appropriate to maintain the failing thresholds at 12 percent and 30 percent. Some programs may enroll students with very little debt other than the debt they accrue to attend their program. Decreasing the failing thresholds on the basis that students, on average, accrue non-educational debt would risk setting an overly strict standard for some programs.

We also clarify that, as discussed in “§668.404 Calculating D/E Rates,” we calculate interest rates for the annual debt payment using a sliding scale average based on the credential level of a program and, for most students, these interest rates are below the actual interest payments made by students. Although we agree the interest rates used in the calculation of D/E rates, as discussed in “§668.404 Calculating D/E Rates,” for most programs, result in debt calculations that are conservatively low estimates of the actual debt payments made by students, we disagree with the commenters arguing that we should set the failing thresholds for the D/E rates below 12 percent and 30 percent because of our interest rate assumptions. Since the interest rates used in the calculation of the D/E rates measure are conservatively low estimates of the actual debt payment made by students, we also disagree with the commenters who believe the D/E rates thresholds are too low because they are sensitive to interest rates.

As we stated in the NPRM, the passing thresholds for the discretionary income rate and the annual earnings rate are based upon mortgage industry practices and expert recommendations. The passing threshold for the discretionary income rate is set at 20 percent, based on research conducted by economists Sandy Baum and Saul Schwartz, which the Department previously considered in connection with the 2011 Prior Rule.⁹⁷ Specifically, Baum and Schwartz proposed a benchmark for a manageable debt level of not more than 20 percent of discretionary income. That is, they proposed that borrowers have no repayment obligations that exceed 20 percent of their income, a level they found to be unreasonable under virtually all circumstances.⁹⁸ The passing threshold of 8 percent for the annual earnings rate has been a fairly common mortgage-underwriting standard, as many lenders typically recommend that all non-mortgage loan installments not exceed 8 percent of the borrower’s pretax income.⁹⁹

Additionally, the 8 percent cutoff has long been referred to as a limit for student debt burden. Several studies of student debt have accepted the 8 percent standard.^{100 101 102 103} Some State agencies have established guidelines based on this limit. In 1986, the National Association of Student Financial Aid Administrators identified 8 percent of gross income as a limit for excessive debt burden.¹⁰⁴ Finally, based on a study that compared borrowers’ perception of debt burden versus their actual debt-to-earnings ratios, Baum and O’Malley determined that borrowers typically feel overburdened when that ratio is above 8 percent.¹⁰⁵

We note that we disagree with the characterization of some commenters that the paper by Baum and Schwartz that we rely on for support of the 20 percent discretionary income rate threshold rejects the 8 percent annual earnings rate threshold and that for this reason, a higher threshold for the annual earnings rate is more appropriate.¹⁰⁶ In their review of relevant literature, Baum and Schwartz specifically acknowledge the widespread acceptance of the 8 percent standard and conclude that, although it is not as precise as a standard based on a function of discretionary earnings, it is “not . . . unreasonable.”¹⁰⁷ Further, drawing from their analysis of manageable debt in relation to discretionary earnings, Baum and Schwartz recommend a sliding scale limit for debt-to-earnings, based on the level of discretionary earnings, that results in a “maximum Debt-Service Ratio” standard generally stricter than 8 percent.¹⁰⁸

More recently, financial regulators released guidance that debt service payments from all non-mortgage debt should remain below 12 percent of pretax income. In particular, current Federal Housing Administration (FHA) underwriting standards set total debt at an amount not exceeding 43 percent of annual income, a standard that, as noted by a commenter, was adopted by the CFPB in recently published regulations, with housing debt comprising no more than 31 percent of that total income, leaving 12 percent for all other debt, including student loan debt, car loans, and all other consumer debt.¹⁰⁹ That 12 percent is consumed by credit card debt (2.25 percent) and by other consumer debt (9.75 percent), which includes student loan debt. ¹¹⁰ The 2010 Federal Reserve Board Survey of Consumer Finances found that student debt comprises “among families headed by someone less than age 35, 65.6 percent of their installment debt was education related in 2010.”¹¹¹ Eight percent is an appropriate minimum standard because it falls reasonably within the 12 percent of gross income allocable to non-housing debt under current lending standards as well as the 9.75 percent of gross income attributable to non-credit card debt.¹¹² Thus, we disagree with commenters that state current FHA underwriting standards provide strong support for a threshold greater than 8 percent for the annual earnings rate.

In the 2011 Prior Rule, the passing thresholds for the debt-to-earnings ratios were based on the same expert recommendations and industry practice, but were increased by 50 percent to 30 percent for the discretionary income rate and 12 percent for the annual earnings rate to “provide a tolerance over the baseline amounts to identify the lowest-performing programs, as well as to account for former students . . . who may have left the workforce voluntarily or are working part-time.” 76 FR 34400. As we explained in the NPRM, we continue to believe that the stated objectives of the 2011 Prior Rule--to identify poor performing programs, to build a “tolerance” into the thresholds, and to ensure programs are accurately evaluated as to whether they produce graduates with acceptable levels of debt--are better achieved by setting 30 percent for the discretionary income rate and 12 percent for the annual earnings rate as the upper boundaries for a zone, or as failing thresholds, rather than as the passing thresholds.
We base this change on our evaluation of data obtained after the 2011 Prior Rule. We conclude that even though programs with D/E rates exceeding the 20 percent and 8 percent thresholds may not all be resulting in egregious levels of debt in relation to earnings, these programs still exhibit poor outcomes and unsustainable debt levels. For the following reasons, our analysis of the programs we evaluated using data reported by institutions after the 2011 Prior Rule went into effect indicates that the stricter thresholds would more effectively identify poorly performing programs.

First, we examined how debt burden that would have passed the 2011 Prior Rule thresholds would affect borrowers with low earnings. Students who completed programs that passed the 2011 Prior Rule thresholds (12 percent/30 percent) but would not pass the 8 percent/20 percent thresholds adopted in these regulations had average earnings of less than $18,000.¹¹³ Graduates of programs that would pass the thresholds of the 2011 Prior Rule (12 percent/30 percent) could be devoting up to almost $2,200, or 12 percent, of their $18,000 in annual earnings toward student loan payments. We believe it would be very difficult for an individual earning $18,000 to manage that level of debt, and we establish lower passing thresholds to help ensure programs are not leading to such results.

Next, we compared repayment outcomes for programs that meet the 8 percent/20 percent thresholds with those that did not, and that comparison also supports lowering the passing thresholds. Specifically, we examined data showing how borrowers default on, and repay, Federal loans through the first three years of repayment. We compared borrower performance among three groups of programs: programs that pass the 8 percent/20 percent thresholds, programs that do not pass the 8 percent/20 percent thresholds, but would pass the 2011 Prior Rule 12 percent/30 percent thresholds (programs in the zone under these regulations), and programs that fail under the 12 percent/30 percent thresholds of both the 2011 Prior Rule and these regulations. Borrowers in the first group (passing programs under these regulations), from programs that pass the 8 percent/20 percent thresholds, have an average default rate of 19 percent, and an average repayment rate of 45 percent.¹¹⁴ Borrower performance for the other two groups is different than those in the passing group: borrowers in the second group (zone programs under these regulations)--those from programs that met the 2011 Prior Rule passing thresholds (12 percent/30 percent) but would not meet the 8 percent/20 percent thresholds--have a default rate of 25 percent and only a 32 percent average repayment rate.¹¹⁵ Borrowers in the third group (failing programs under these regulations), from programs that fail even the 2011 Prior Rule thresholds (12 percent/30 percent), have rates like those in the zone group: about a 28 percent default rate and an average repayment rate of about 32 percent.¹¹⁶ Together, these results indicate that zone programs are much more similar to their failing counterparts than their passing counterparts. Accordingly, although zone programs are allowed additional time before ineligibility in comparison to failing programs, programs in both groups are ultimately treated the same if their results do not change because expert recommendations, industry practice, and the Department’s analysis all indicate that they are both resulting in similarly poor student outcomes and not resulting in gainful employment. By reducing the passing thresholds for the D/E rates measure to 8 percent and 20 percent, we treat as unacceptable those programs that exceed these thresholds, but allow a limited time to evaluate whether the unacceptable performance persists before revoking eligibility.

With regard to the stated intention to adopt a rate that includes a tolerance to reduce the likelihood that a program will be mischaracterized, we believe that the three-tier pass, zone, fail construction and the corresponding thresholds for these categories make it unnecessary to create buffer by raising the passing thresholds as was done in the 2011 Prior Rule. As discussed in the NPRM, setting the failing thresholds at 12 percent and 30 percent lower the probability to close to zero that passing programs will lose eligibility because they are mischaracterized, due to atypical factors associated with a non-representative cohort of students, as failing. Likewise, creating a buffer between the passing and failing thresholds, where programs in the zone have a longer time to loss of eligibility than those that fail the thresholds, lowers the probability to close to zero that passing programs will lose eligibility because they are mischaracterized as being in the zone as a result of atypical factors.

Further, a four year zone makes it unlikely that fluctuations in labor market conditions could cause a passing program to become ineligible. According to the National Bureau of Economic Research, recessions have, on average, lasted 11.1 months since 1945.¹¹⁷ An otherwise passing program is unlikely to fall in the zone for four consecutive years due to an economic downturn or fluctuations within the local labor markets.

Under the regulations, programs can satisfy the D/E rates measure in one of two ways. Programs whose graduates have low earnings relative to debt would benefit from the calculation based on total income, and programs whose graduates have higher debt loads that are offset by higher earnings would benefit from the calculation based on discretionary income. Even for programs where the average annual earnings rate for students who complete the program exceeds 8 percent, as long as the average discretionary income rate is below the 20 percent threshold, the program will be deemed passing.

We adopted a buffer in the 2011 Prior Rule in part to avoid mischaracterization of a program and in part to account for students who completed the program who are working part-time or who are not employed. As discussed in this section, because the D/E rates measure assesses whether students who complete a GE program will earn enough to manage the debt they incur, that assessment must take into account the outcomes of students who are not working or are not working full time, either by choice or involuntarily, without regard to whether such outcomes are typical. As stated previously, where such outcomes are atypical, several aspects of the regulations, including the pass, zone, and fail thresholds, use of mean and median earnings, use of a multi-year cohort period with a minimum n-size, and allowing several years of non-passing results before a program loses eligibility for title IV, HEA program funds reduce the likelihood to close to zero that a typically passing program will be made ineligible by being mischaracterized as failing or in the zone due to an atypical cohort of students who complete the program such as those identified by the commenter. Where it is typical for students to work time or regularly leave the labor force for long periods, institutions should adjust their costs and other features of their programs to ensure that these students can manage their debt.

Accordingly, for the reasons provided, a buffer is unnecessary. We revise the passing D/E rates in these regulations because we conclude that the 50 percent buffer in the 2011 Prior Rule is unnecessary. We instead establish a zone to identify programs that exceed the 8 percent and 20 percent thresholds, and use the 12 percent and 30 percent measures as the upper limits. This approach accounts for the reasons that a buffer was added in the 2011 Prior Rule, to make accurate and fair assessments of programs, while ensuring that once there is certainty that an accurate and fair assessment is being made, programs with sustained poor outcomes are not allowed to remain eligible and harm students.

We do not agree that alternative thresholds--including annual earnings rates thresholds of 10 percent, 13 percent, and 15 percent, as suggested by commenters--would be more appropriate for determining eligibility under the title IV, HEA programs. We recognize that some research points to these as reasonable thresholds. Likewise, some research may even point to thresholds below 8 percent for the annual earnings rate.¹¹⁸ However, we believe that 8 percent for education-related debt is well within the range of acceptable debt levels identified by researchers and the standard that is generally most supported.^{119 120 121 122} Based on the best available evidence, students whose annual earnings rate exceeds 8 percent are substantially more likely to default on their loans or experience serious financial or emotional harm.

Similarly, we disagree with the commenters that suggested that annual earnings rates be set between 10 and 15 percent because the majority of personal finance experts believe that an acceptable annual debt-to-earnings ratio falls within this range.¹²³ As stated previously, in the sources cited by the commenters, the personal finance experts often refer to the amount of total debt that individuals can manage, whereas the focus of the D/E rates measure, and the basis for the thresholds, is the acceptable level of debt incurred for enrollment in a GE program. Moreover, such expert advice does not take into consideration that the discretionary income rates allow some programs with annual income rates above 8 percent to pass, if their students earn enough to manage their debt, based on the best available evidence.

We also disagree with the contention made by some commenters that a recent NCES study shows the thresholds to be inappropriately low because a large fraction of graduating undergraduate students have debt-to-earnings ratios above 12 percent, suggesting many non-GE programs in the public and non-profit sector would fail the annual earnings rate if they were subject to the regulations.¹²⁴ The NCES methodology for calculating student debt-to-earnings ratios is not comparable to the methodology for calculating D/E rates at the program level under these regulations. Specifically, the NCES methodology for calculating each of loan debt, earnings, and the debt-to-earnings ratios results in higher estimates of debt burden than is observed under the D/E rates methodology. For example: First, the NCES study does not include students who only receive Pell Grants, while these students are included in the D/E rates calculations as having zero debt, which substantially lowers the median loan debt for each program. Also, while the NCES study includes all students paying loans for any reason, the D/E rates exclude students who are still enrolled in school, are serving in the military, have a total and permanent disability, or are deceased, the overall effect of which is to, again, lower the D/E rates for each program. Second, the NCES study measures actual amount borrowed, not the amount borrowed capped at the total of tuition, fees, books, equipment and supplies, as is the case under these regulations. As discussed earlier, in every instance in which the actual amount borrowed exceeds tuition, fees, books and supplies, the D/E rates will be capped at that tuition, fees, books and supplies – not the actual (larger) loan amount. In every one of those instances, the D/E rates calculated under these regulations will necessarily be lower than the amount of loan debt calculated in conventional studies, such as the NCES study (which includes no indication that the term “debt” had any special, restricted meaning) and the literature addressing this issue. Third, the NCES study measures earnings only one year after completion, but under the D/E rates measure, earnings are measured about three years after completion. Since earnings tend to increase after completion of postsecondary programs as students gain more experience in the workforce, D/E rates under the regulations will tend to be lower than those reflected in the NCES study. Fourth, the NCES study does not include a discretionary income rate. We believe some programs with relatively high annual earnings rates will pass the discretionary income rate metric because they have graduates who have higher earnings even though they have large amounts of debt. Fifth, under the D/E rates measure, we use the higher of mean and median of earnings and the median of debt, rather than just means. We believe this aspect of the regulations will also lead to lower D/E rates than those reflected in the NCES study because it makes the D/E rates measure less sensitive in extreme cases of high debt and low earnings among students who complete a program at each institution. These differences in methodology reflect policy goals that have been incorporated into the regulations, including goals relating to the accessibility and affordability of GE programs, as well as Department interests in ensuring the equitable application of these regulations to institutions in different sectors and the coordination of these regulations with other Federal student aid programs. As a result, the results of the NCES study do not provide a useful basis for evaluating the D/E rates thresholds.

Similarly, we disagree with commenters who argued that BPS data showing that, on average, graduating bachelor’s degree students have annual earnings rates above 8 percent indicate the thresholds are inappropriate. The data cited by the commenters exclude graduates who graduated with zero debt, which comprise about one-third of students graduating with a bachelor’s degree.¹²⁵ Also, earnings levels in BPS are reported six years after enrollment, while the D/E rates measure earnings about three years after completion. Another limitation of BPS survey data is that they only measure income from the student’s primary job, while the D/E rates include all sources of income reported to the Social Security Administration (SSA).

Changes: None.

Comments: Commenters said the D/E rates measure lacks a rational basis as an accountability metric. They contended that, in adopting the D/E rates measure, the Department places too much weight on the study by Baum and Schwartz and mortgage underwriting standards in identifying thresholds. Commenters said the Department disregards other studies and data sources showing that most programs would not pass the D/E rates measure if it were applied to all postsecondary programs. The commenters asserted the Department should be applying a metric supported by other data studies, relying on data from NPSAS, along with studies conducted by NCES and the American Enterprise Institute, on debt and earnings levels of college graduates.

Commenters also asserted that the data the Department used to analyze the proposed regulations was biased and weak because it only included a small fraction of all GE programs. For this reason, they argued the Department should have considered additional data sources that would have provided more accurate information about the impact of the regulations.

Discussion: The Department considered a number of data and research sources and authorities in formulating the D/E rates measure. In addition to the analysis and recommendation of Baum and Schwartz, we considered research on earnings gains by other scholars, including Cellini and Chaudhary,¹²⁶ Kane and Rouse,¹²⁷ Avery and Turner,¹²⁸ and Deming, Goldin, and Katz.¹²⁹ We also took into account lending ratios currently set by the FHA and the CFPB, as they estimate sustainable levels of non-housing debt. As stated previously, we do not believe that the NCES study and the other studies suggested by commenters use a comparable methodology, and further, we do not agree with the conclusions the commenters draw from these studies.

In analyzing the potential impact of the D/E rates measure, we relied primarily on data from NSLDS because it contains a complete record of all students receiving title IV, HEA program funds from each program. Although we also have access to data from sample surveys, such as BPS and NPSAS, we did not rely on such data because we had access to a full data set of students in GE programs. NPSAS data also do not allow for the calculation of D/E rates that are comparable to the D/E rates being evaluated under this regulation. Because NCES and NPSAS data focus on studying all undergraduate students rather than just students who attend GE programs, NCES and NPSAS data provide information on a different population of students than those we expect to be evaluated under the D/E rates measure. Additionally, NCES survey data do not provide earnings information about students three to four years after graduation, which is the timeframe for calculating D/E rates.

We do not agree that our analyses did not sufficiently consider data presented by the American Enterprise Institute.¹³⁰ As noted earlier in the summary of comments about the impact of the regulations on for-profit institutions, the American Enterprise Institute data suggest, based on data from the University of Texas, that a large fraction of programs operated by University of Texas would fail the D/E rates measure. These data are not appropriate for analyzing these regulations. First, as with the data used for the NCES report, the University of Texas data do not allow for calculation of D/E rates using a comparable methodology. Second, the American Enterprise Institute only considered data for a small subset of programs and students--that is, those who attended programs in the University of Texas system. We believe considering such a small subset of gainful employment programs has limited analytical value, and, thus, we relied on the data we had available on all gainful employment programs.

We disagree with claims that our analyses are unreliable and biased because we included only a fraction of gainful employment programs. Using our data, we analyzed all programs that we estimate would meet the minimum “n-size” requirement to be evaluated under the D/E rates measure--that is, all programs for which 30 students completed the program--for the cohort of students we evaluated.

Changes: None.

Comments: Some commenters recommended raising the D/E rates thresholds to account for longer-term earnings benefits from earned program credentials. Commenters offered research demonstrating that increased benefits from program completion, including non-pecuniary benefits, may not be immediately apparent and may increase over time in a way that the proposed regulations would not take into account.

Discussion: While we agree that gross earnings and earnings gains as a result of obtaining additional credentials will increase for program graduates over the course of their lives, and gains for some occupations may be more delayed than others, we do not believe that this merits increasing the D/E rates thresholds for the purpose of program accountability. As stated previously, these regulations will help ensure program graduates have sustainable debt levels both in the early part of their careers and in later years so loan payments are kept manageable and do not interfere with individuals’ ability to repay other debts or result in general over-indebtedness.

Further, our analysis indicates that the passing thresholds for the D/E rates measure are set at a level that reflects repayment outcomes. The Department’s data indicate the average volume-based repayment rate, measured at about the third year of repayment, of programs in the zone is comparable to those above the failing thresholds, while passing programs, on average, have a substantially higher average repayment rate. Average cohort default rates, measured within the first three years of repayment, are similar for zone and failing programs and substantially higher than the average default rate of passing programs.

Changes: None.

Comments: A number of commenters suggested that different thresholds for the D/E rates measure should be applied to institutions or programs that serve students with backgrounds that may increase their risk factors for over-indebtedness. Some commenters suggested that the thresholds be adjusted on a sliding scale based on the number of students served by a program who are eligible for Pell Grants.

One commenter also suggested that different D/E rates thresholds be applied to programs, such as those in the cosmetology sector, that serve mostly women, who the commenter suggested are more likely to choose part-time employment or to not work in order to raise children. This same commenter suggested that programs serving a high proportion of single parents are unfairly punished by the thresholds for the D/E rates measure because single parents would have an incentive to earn limited incomes in order to continue to qualify for various assistance programs.

Discussion: We do not agree that alternative metrics or thresholds should be applied to different types of programs or institutions or to programs serving different types of students, such as minority or low-income students. As described in greater detail in the Regulatory Impact Analysis, the Department has examined the effects of student demographic characteristics on results under the annual earnings rate measure and does not find evidence to indicate that the composition of a GE program’s students is determinative of outcomes. While the Department recognizes that the background of students has some impact on outcomes and that some groups may face greater obstacles in the labor market than others, we do not agree that the appropriate response to those obstacles is to set alternative standards based on them. As discussed previously, we seek to apply the same set of minimum standards across all GE programs, regardless of their sector, location, or the students they serve. As our analysis shows, the substantial majority of programs will meet these minimum standards, even when comparing programs with higher proportions of students with increased “risk factors.” The regulations will help ensure that programs only remain eligible for title IV, HEA program funds if they meet these minimum standards that define maximum levels of indebtedness that are acceptable for any student. We intend for the regulations to allow these successful programs to grow, and for institutions to establish new programs that achieve and build upon these results, so that all students, regardless of background or occupational area, will have options that will lead to positive results.

Changes: None.

Comments: One commenter suggested that the D/E rates thresholds are punitive, as more programs would fail under these regulations than would have failed under the 2011 Prior Rule.

Discussion: While the Department acknowledges that it is possible that more programs would not meet the passing thresholds under these regulations as compared to those in the 2011 Prior Rule, as previously discussed, the Department must ensure an appropriate standard is established to protect students from unmanageable levels of debt. As stated previously, we believe the D/E rates thresholds in these regulations appropriately define the maximum levels of indebtedness that are acceptable for all students.

Changes: None.

Comments: One commenter suggested that the Department include the outcomes of students who do not borrow in a program’s D/E rates calculation and suggested that the thresholds be increased to account for this change.

Discussion: The regulations provide for the consideration of the outcomes of students who have completed a program and have only received Pell Grants and, therefore, have no debt for the D/E rates calculation. Further, we assess debt as a median when calculating the D/E rates, so that programs in which a majority of the students who have completed the program but do not have any title IV loans would have D/E rates of zero and would pass the D/E rates measure.

As discussed in “§668.401 Scope and Purpose,” we are not including individuals who did not receive title IV, HEA program funds in the calculation of the D/E rates measure. We disagree, however, that this warrants adjustments or increases to the D/E rates thresholds. The expert research, industry practices, and internal analysis that we relied on in determining the thresholds apply to all students.

Changes: None.

Zone

Comments: Multiple commenters suggested that the addition of the zone results in unnecessarily complex and burdensome regulations that will confuse borrowers and institutions. One commenter suggested that the zone would create undue burden on State agencies and their monitoring responsibilities. Some commenters expressed concern that the zone yields additional uncertainty for institutions and students regarding the future of a program. Commenters also argued that the zone should be adjusted for student characteristics.

Some commenters suggested removing the zone and returning to the 2011 Prior Rule thresholds of 12 percent for the earnings rate and 30 percent for the discretionary income rate. Other commenters suggested that despite the presence of a zone, the regulations do not allow sufficient time for programs to take corrective actions and improve so that they can move from the zone to passing under the D/E rates measure, making the zone tantamount to failure. One of these commenters, using the 2012 GE informational D/E rates, calculated the aggregate failure rate, counting the zone as a failure, near 31.0 percent--about a five-fold increase in the number of programs ultimately losing eligibility for title IV, HEA program funds, as compared with the 2011 Prior Rule. The commenter also said about 42 percent of programs at for-profit colleges will be failing or in the zone, when weighted by program enrollment, including more than one-third of certificate programs, three-quarters of associate degree programs, one-fifth of bachelor’s degree programs, and one-third of professional degree programs. The commenter posited that more than 1.1 million students are enrolled in programs that will lose eligibility for title IV, HEA program funds under the proposed regulations.

Other commenters agreed with the Department’s proposal for a zone but argued that the length of time that a program could be in the zone before being determined ineligible is arbitrary. Some of the commenters said that the length of the zone is insufficient to measure programs where there is a longer time after completion before a student is employable, such as with medical programs. Some of the commenters complained that the four-year zone period, when taken together with the transition period, is too long, and would initially allow failing programs to have operated for eight years without relief to students who are enrolled during that time. Some of these commenters suggested a three-year zone as an alternative.

Some commenters suggested that the Department should provide for a zone only in the first few years after the regulations are implemented and then eliminate the zone. The commenters stated that this approach would help to remove the worst performing programs relatively quickly and allow poor performers that are closer to passing the D/E rates measure time to improve. The commenters said that eliminating the zone after a few years would prevent taxpayers from subsidizing low-performing programs that would otherwise be allowed to continue to enroll unlimited numbers of students while in the zone.

Other commenters suggested that the zone is insufficient because it provides minimal protection while potentially confusing students about the riskiness of a program they may be attending or considering for enrollment. Some of these commenters stated that the zone provides limited transparency, as institutions with potentially failing programs are required to warn students of potential loss of eligibility only in the year before they might be deemed ineligible. Some commenters suggested the Department eliminate the zone to ensure that students are not attending programs in which students who complete the program have a discretionary income rate above 20 percent, an unacceptable outcome.

Other commenters proposed that, while a zone may be necessary, the regulations should include a firm upper threshold by which, should a program’s D/E rates exceed the threshold, the program would immediately lose eligibility. Commenters suggested that there are cases in which outcomes for students are so egregious that programs need to lose eligibility immediately to protect students from additional harm.

Discussion: The Department disagrees that the zone should be eliminated or phased out. The zone under the D/E rates measure serves several important purposes.

First, as stated previously, a four-year zone provides a buffer to account for statistical imprecision due to random year-to-year variations, virtually eliminating the possibility that a program would mistakenly be found ineligible on the basis of D/E rates for students who completed the program in any one year. As discussed in the NPRM, our analysis shows that the chances that an unrepresentative population of students who completed a program could occur in four out of four consecutive years such that a program’s D/E rates exceed the 8 percent and 20 percent thresholds four years in a row when in fact its D/E rates are on average less than 8 percent and 20 percent for a typical year is close to zero percent.

As also stated previously, we believe that programs with an annual earnings rate above 8 percent and discretionary income rate above 20 percent are producing poor outcomes for students. A permanent four-year zone holds all of these programs accountable while ensuring that the Department is making an accurate assessment. In comparison, raising the passing thresholds to 12 percent and 30 percent to create a buffer for accuracy would allow many poorly performing programs to evade accountability.

With a shorter zone period, programs would be at risk of mischaracterization. Similarly, it is necessary to have a two out of three year time period to ineligibility for failing programs in order to ensure that an accurate assessment is made. Our analysis indicates the probability of mischaracterizing a program that is typically in the zone as failing in a single year could be as high as 4.1 percent. By allowing programs to remain eligible after a single failing result, we believe we are providing programs near the borderline of the 12 percent threshold a reasonable opportunity to remain eligible until we confirm that our assessment is accurate. Accordingly, we do not agree that programs with an annual earnings rate above 12 percent and discretionary income rate above 30 percent should immediately lose eligibility. We believe that the program disclosures and warnings mitigate the need to establish any threshold where a one-year outcome would immediately trigger a loss of eligibility.

While the zone may lead to at least some additional uncertainty for institutions and students, we believe this concern is outweighed by our interest in ensuring that all poorly performing programs are held accountable. To provide at least some level of protection to students, as discussed in “§668.410 Consequences of the D/E Rates Measure,” an institution will also be required to issue warnings to current and prospective students for a program in any year in which the program faces potential ineligibility based upon its next set of final D/E rates.

Second, the four-year zone helps to ensure that programs with rates that are usually passing or close to meeting the passing threshold are not deemed failing or made ineligible due to economic fluctuations. As stated previously, recessions have, on average, lasted 11.1 months since 1945.¹³¹ It is implausible that a program would fall in the zone for four consecutive years due to an economic downturn or fluctuations within the local labor markets.

Third, a four-year zone, coupled with the transitional D/E rates calculation, described in more detail in “§668.404 Calculating D/E Rates,” will provide institutions with more time to show improvement in their programs after the regulations become effective. Programs will have several years after these regulations take effect to improve and achieve passing rates. During the transition period, an alternative D/E rates calculation will be made so that institutions can benefit from any immediate reductions in cost they make. As discussed in “§668.404 Calculating D/E Rates,” we have changed the transition period by extending the length to ensure that institutions that make sufficient reductions in tuition and fees are able to benefit from such efforts. Because institutions have the ability to affect the debt that their students accumulate by lowering tuition and fees, we believe it is possible for zone and failing programs to improve as a result of the transitional D/E rates calculation. Analysis of the zone programs in the 2012 GE informational D/E rates data set suggests that zone programs would need to reduce their median annual loan payment by roughly 16 percent in order to pass.

While we acknowledge that the zone may add some additional level of complexity to the regulations, we believe it is necessary to ensure that programs that lead to poor outcomes are held accountable. With respect to the commenter who believed the zone would create additional burden for State regulators, we are unable to identify a reason for why this would be the case.

Changes: None.

Time Period to Ineligibility

Comments: Some commenters contended that the Department should revise the regulations to provide for a longer time before which a program that is failing the D/E rates measure would be determined ineligible under the title IV, HEA programs. The commenters stated that the time period should be longer because improvement would be impossible over the two out of three year period proposed. They argued that the Department should adopt the ineligibility time period from the 2011 Prior Rule, where programs would not be determined ineligible unless they failed the metrics in three out of four years.

Other commenters asserted that the two out of three year timeframe is not justified and is designed to deny eligibility to for-profit institutions before they have an opportunity to improve. A few commenters said the proposed period before ineligibility is particularly short for programs with longer lengths, such as advanced degree programs, because these programs would have even less opportunity to improve than would short-term certificate programs based on the fact that students completing these programs would have started attending the program in years even further before the implementation of the regulations.

In contrast, other commenters believed that even two out of three years is too long because allowing these programs to remain eligible for that period of time would harm too many students. They argued that failing programs already produce unacceptably poor outcomes and that allowing them to continue to operate will lead to more students taking out high amounts of debt with little benefit. The commenters proposed that failing programs should become immediately ineligible once the regulations are effective should they fail to pass the D/E rates measure.

Discussion: Institutions should already be striving to improve program outcomes for their students, and the outcomes for graduates every year may be influenced by prior changes an institution made to its program. Based on our analysis, we expect that 74 percent of programs will pass the D/E rates measure, and 91 percent will either pass or be in the zone. Any program with a discretionary income rate above 30 percent and an annual earnings rate above 12 percent is producing poor outcomes for its students and should, in order to minimize the program’s negative impact on students, be given as limited a period as is necessary to ensure statistical accuracy of program measurement before it loses its eligibility. Accordingly, we will allow programs to operate until they have failed twice within three years to be certain we are only making ineligible those programs that consistently do not pass the D/E rates measure. Because, as discussed in the NPRM, the probability that a passing program is determined ineligible due to statistical imprecision is nearly non-existent with a two out of three year period, we believe that this is an appropriate length of time to ineligibility for failing programs and that the longer three out of four year period of the 2011 Prior Rule is unnecessary.

Because of the 2011 Prior Rule and informational rates, institutions have had relevant information for a sufficient amount of time to make improvements. Further, the transition period will allow institutions to continue to improve their programs even after the regulations take effect. Even institutions that only begin to make improvements after the regulations take effect, or those that did not have informational rates for programs that were not in existence or are medical or dental programs, will get substantial, if not full, benefit of the transition period. Institutions that make immediate changes that at minimum move a failing program into the zone will then have additional years of the transition period coupled with the zone to continue to improve.

We are revising §668.403(c)(4) to state more clearly the circumstances in which a program becomes ineligible under the D/E rates measure.

Changes: We have revised the language in §668.403(c)(4) to clarify that a GE program becomes ineligible if the program either is failing the D/E rates measure in two out of any three consecutive award years for which the program’s D/E rates are calculated; or has a combination of zone and failing D/E rates for four consecutive award years for which the program’s D/E rates are calculated.

Other Issues Regarding the D/E Rates Measure

Comments: Some commenters suggested that programs should be required to pass both the annual earnings rate and discretionary income rate metrics in order to pass the D/E rates measure. These commenters argued that programs should be expected to generate sufficient income for graduates to cover basic living expenses and pay back their student loans. They expressed concern that many programs pass the annual earnings rate metric even though their students have to spend more than their entire discretionary income on debt service. Similarly, some commenters suggested that the regulations include a minimum earnings level below which a program would automatically fail both the annual earnings rate and discretionary income rate metrics, arguing that there is a baseline income below which any required debt payments would result in unmanageable debt. Multiple commenters made a related suggestion to base the D/E rates measure only on discretionary income, and eliminate the annual earnings rate, so that programs would be deemed failing if their students have earnings below the poverty line.

On the other hand, some commenters argued that the discretionary income rate metric is unnecessary because very few programs would be affected by it.

Discussion: The annual earnings rate and the discretionary income rate, which comprise the D/E rates measure, serve distinct and important purposes in the regulations. The annual earnings rate more accurately assesses programs with graduates that have low earnings but relatively low debt. The discretionary income rate will help capture programs with students that have higher debt but also relatively higher earnings.

The annual earnings rate by itself would fail to properly assess many programs that, according to expert recommendations, meet minimum standards for acceptable debt levels. As a result, the Department disagrees with those commenters who suggested that including the discretionary income rate is of limited value. Without the discretionary income rate, programs where students have high levels of debt, but earnings adequate to manage that debt, would not pass the D/E rates measure. While there may be a more limited universe of programs that would pass the D/E rates measure based on the discretionary income rate threshold, the Department believes it is important to maintain this threshold to protect those programs that may be producing good outcomes for students.

Requiring programs to pass both the annual earnings rate and discretionary income rate, removing the annual earnings rate altogether, or establishing a minimum earnings threshold for the D/E rates measure would all have the same impact--making ineligible programs that, based on expert analysis, leave students with manageable levels of debt. In some cases, programs may leave graduates with low earnings, but these students may also have minimal debt that experts have deemed manageable at those earnings levels. For other programs, students may be faced with high levels of debt, but also be left with significantly higher earnings such that high debt levels are manageable. In both cases, the discretionary income rate and the annual earnings rate, respectively, ensure programs meet a minimum standard while also being allowed to operate when providing acceptable outcomes for graduates. We provide an analysis in the Regulatory Impact Analysis of how many programs passed, failed, or were in the zone under the 2011 GE informational D/E rates.

Changes: None.

Comments: Many commenters contended that the D/E rates measure is flawed because (1) students’ earnings are affected by economic conditions beyond the control of the institution, such as fluctuations in the national or regional economy, and (2) earnings vary by regional or geographic location, particularly between rural and urban areas. A few commenters believed it would be difficult for institutions to predict local labor market conditions with enough reliability to set tuition and fees sufficiently low to ensure their programs pass the D/E rates measure.

Discussion: We believe that institutions should be responsive to regional labor market needs and should only offer programs if they reasonably expect students to be able to find stable employment within that occupation. We do not agree that institutions cannot assess their graduates’ employment and earnings prospects in order to price their programs appropriately. Indeed, it is an institution’s responsibility to conduct the due diligence necessary to evaluate the potential outcomes of students before offering a program. We do not believe that this is an unreasonable expectation because some accreditors and State agencies already require institutions to demonstrate that there is a labor market need for a program before it is approved.

However, we agree that a program should not be determined ineligible under the D/E rates measure due to temporary and unanticipated fluctuations in local labor market conditions. We believe that several components of the accountability framework will help ensure that passing programs do not become ineligible due to such fluctuations.

The regulations provide for a zone that allows programs to remain eligible for up to four years despite not passing the D/E rates measure in any of those years. The zone protects passing programs from losing their eligibility for title IV, HEA program funds where their increase in D/E rates was attributable to temporary fluctuations in local labor market conditions. Most economic downturns are far too short to cause a program that would otherwise be passing to have D/E rates in the zone for four consecutive years due to fluctuations in the local labor market. As stated previously, recessions have, on average, lasted 11.1 months since 1945--far shorter than the four years in which programs are permitted to remain in the zone.¹³²

Sensitivity to temporary economic fluctuations outside of an institution’s control is also reduced by calculating the D/E rates based on two-year and four-year cohorts of students, rather than a single-year cohort, and calculating a program’s annual earnings as means and medians. Calculating D/E rates based on students who completed over multiple years reduces the impact of short term fluctuations in the economy that may affect a particular cohort of graduates but not others. Similarly, means and medians mitigate the effects of economic cycles by measuring central tendency and reducing the influence of students who may have been most impacted by a downturn.

Changes: None.

Comments: Some commenters argued that the D/E rates measure is flawed because for some occupations, such as cosmetology, earnings may be depressed because a significant number of program graduates tend to leave but then return to the workforce, sometimes repeatedly, or to work part-time. According to the commenters, this is particularly the case in occupations in which workers are predominately women, who may leave and return to the workforce for family purposes more frequently than workers in other occupations. The commenters contended that, for students entering such occupations, earnings will be low, so that the regulations will be biased against programs providing training in these occupations.

Discussion: In examining programs generating an unusually large number of graduates without full-time employment, the Department believes it is reasonable to attribute this outcome less to individual student choices than to the performance of the program itself. The D/E rates measure will identify programs where the majority of program graduates are carrying debts that exceed levels recommended by experts. If an institution expects a program to generate large numbers of graduates who are not seeking employment or who are seeking only part-time employment, it should consider reducing debt levels rather than expecting students to bear even higher debt burdens. Regardless of whether a student works full-time or part-time or intermittently, the student is still burdened in the same way by the loans he or she received in order to attend the program.

Changes: None.

Comments: Commenters argued that the D/E rates measure is inequitable across programs in different States because, according to the commenters, some States provide more financial aid grants to students and greater financial support to institutions, requiring students to acquire less debt. Commenters said the regulations should take State funding into account because, otherwise, programs in States with less funding for higher education would be adversely affected by the D/E rates measure.

Discussion: While we recognize that there may be differences in support for higher education among States, such that borrowers’ debt levels may depend on the State in which they reside, those differences are not relevant to address the question of whether students are overburdened with debt as a result of enrolling in a particular program. Some States’ investments in higher education may permit students who benefit from that support to borrow less, in which case programs in that State may have an easier time passing the D/E rates measure, but it would not change the need to ensure borrowers are protected from being burdened in other States that do not provide as much support for higher education. Accordingly, we decline to adjust the D/E rates measure to account for State investment in higher education.

Changes: None.

Comments: Many commenters did not support the Department's proposal in the NPRM that a program must pass both the D/E rates measure and pCDR measure to remain eligible for title IV, HEA program funds. The commenters stated that this approach is inconsistent with the position the Department took under the 2011 Prior Rule, under which a program would remain eligible if it passed either the debt-to-earnings ratios or the second debt measure in that regulation, the loan repayment rate. The commenters contended that the Department did not justify this departure from the 2011 Prior Rule. They suggested that programs should remain eligible for title IV, HEA program funds if they pass either the D/E rates measure or the pCDR measure. They asserted that there is a lack of overlap between programs that fail the D/E rates measure and programs that fail the pCDR measure and this indicates that the two metrics set different and conflicting standards.

We also received a number of comments in support of the Department’s proposal to require that programs pass both the D/E rates and pCDR measures. A few of these commenters were concerned that the pCDR measure does not adequately protect students, citing concerns about the validity of the metric and its susceptibility to manipulation. As a result, they argued that programs should be required to pass both measures if pCDR is included in the final regulations. Some commenters argued that the lack of overlap between the measures supports requiring programs to pass both because it indicates that they assess two distinct and important aspects of program performance. Other commenters were concerned that allowing programs to remain eligible solely on the basis of passing the D/E rates measure would harm students because the D/E rates measure assesses only the outcomes of students who complete a program and does not hold programs accountable for low completion rates.

Similarly, a few commenters suggested the independent operation of pCDR undermines the validity of the D/E rates measure because there are many programs with high D/E rates but low pCDR rates or where fewer than 30 percent of students default, which, in their view, showed that the D/E rates measure does not provide a reasonable basis for eligibility determinations. They contended that because such programs would be ineligible under the proposed regulations, the independent operation of the metrics would result in the application of an inconsistent standard.

Other commenters believed that the pCDR measure by itself is a sufficient measure of whether a program prepares students for gainful employment. Some of these commenters argued that a cohort default rate measured at the program level, as set forth in the NPRM, with a three-year period before ineligibility and with time limits on deferments and forbearances would sufficiently address concerns about the validity of the metric and its susceptibility to manipulation. The commenters contended that the three-year cohort default window is longer than any combination of deferments or forbearances, and that using a three-year default rate measure would ensure borrowers are counted as being in default on a loan if they consistently do not make minimum payments during the three-year window. One commenter said the pCDR measure would protect taxpayers better than the D/E rates measure by ensuring fewer defaults, and, accordingly, this commenter asserted, passing the pCDR measure should be sufficient to remain eligible.

Discussion: As discussed elsewhere in this section, we have not included the pCDR measure as an accountability metric in the final regulations. The Department will assess program performance using only the D/E rates measure. Accordingly, we do not address comments regarding whether the measures should operate independently or whether pCDR is a reasonable measure of continuing eligibility for title IV, HEA program funds.

We do not agree that the D/E rates measure by itself is an improper measure of whether a program prepares students for gainful employment simply because some programs have high D/E rates but a low pCDR. These results are not surprising for two reasons. First, the measures use different approaches to assess the outcomes of overlapping, but disparate groups of students. The D/E rates measure certain outcomes of students who completed a program, while pCDR measures certain outcomes of both students who do, and do not, complete a program. Second, the measures assess related, but different aspects of repayment behavior. While the pCDR measure identifies programs where a large proportion of students have defaulted on their loans, it does not recognize programs where too many borrowers are experiencing extreme difficulty in making payments and reducing loan balances but have not yet defaulted as the D/E rates measure does.

Changes: None.

Comments: Some commenters said the D/E rates measure is unfair in its application to medical programs. One commenter noted that some medical degree programs in the non-profit sector would not be subject to the regulations, while the same medical programs in the for-profit sector would be. Another commenter compared the earnings outcomes of medical programs subject to the regulations to those of some social work degree programs operated by non-profit institutions that are not subject to the regulations. The commenter claimed the regulations are inequitable because D/E rates are generally higher among social workers than those students completing medical certificate programs.

Discussion: As discussed in “§668.401 Scope and Purpose,” the Department’s regulatory authority in this rulemaking is limited to defining statutory requirements under the HEA that apply only to GE programs. The Department does not have the authority in this rulemaking to regulate those higher education institutions or programs that do not base their eligibility on the offering of programs that prepare students for gainful employment, even if such institutions or programs would not pass the D/E rates measure. Further, the regulations establish minimum standards regarding reasonable debt levels in relation to earnings for all GE programs, regardless of how programs that provide training for occupations in different fields, such as social work and medicine, compare to one another.

Changes: None.

Comments: We received a number of comments on how the Department should treat GE programs for which D/E rates are calculated in some years but not others. Some commenters asserted that the Department should not disregard years for which D/E rates are not calculated for a program and instead should treat the program as if it had passed the D/E rates measure for that year. They argued that any other result would be unfair because a program could be determined ineligible as a result of failing the D/E rates measure in two out of three consecutive years for which rates were calculated, even though those assessments had been made very far apart in time from one another.

One commenter suggested using the most recent five award years regardless of whether D/E rates were calculated during any or all of the years. Another commenter supported resetting a program’s results under the D/E rates measure after two consecutive years in which D/E rates are not calculated.

Discussion: We do not believe that it is unfair or invalid to use a program’s D/E rates for non-consecutive years in determining the program’s continuing eligibility for title IV, HEA program funds. The probability of mischaracterizing a program as failing or in the zone due to an unusual cohort of students or other anomalies does not increase if D/E rates are calculated during non-consecutive years.

In determining a program’s continuing eligibility, rather than making assumptions about a program’s D/E rates in years where less than 30 students complete the program, we believe it is important to use the best available evidence as to whether a program produces positive student outcomes, which is the program’s most recent actual results. If the program has in fact improved since a prior result under the D/E rates measure, its improved performance will be apparent once it has enough students who completed the program to be assessed under the D/E rates measure again.

We agree, however, that the longer the hiatus between years for which rates are calculated, the less compelling the inference becomes that a prior result is reflective of current performance. Accordingly, we are revising §668.403 to provide that, in making an eligibility determination, we will not consider prior D/E rates after four consecutive years in which D/E rates are not calculated. A four-year limitation aligns with the general operation of the D/E rates measure which, under the zone, finds outcomes over a four-year period as relevant. We are also clarifying that, generally, subject to the four-year “reset,” if a program’s D/E rates are not issued or calculated for an award year, the program receives no result under the D/E rates measure for that award year and the program’s status under the D/E rates measure is unchanged from the last year for which D/E rates were calculated. For example, where a program receives its first failing result and the institution is required to give student warnings as a result, the program will still be considered to be a first time failing program and the institution will continue to be required to give student warnings in the next award year even if the program’s next D/E rates are not calculated or issued because it did not meet the minimum n-size requirement.

Changes: We have revised §668.403 to add new paragraph (c)(5), which provides that, if a program’s D/E rates are not calculated or issued for an award year, the program receives no result under the D/E rates measure for that award year and the program’s status under the D/E rates measure is unchanged from the last year for which D/E rates were calculated, provided that, if the Secretary does not calculate D/E rates for the program for four or more consecutive award years, the Secretary disregards the program’s D/E rates for any award year prior to the four –year period in determining whether the program is eligible for title IV, HEA program funds.

We have also revised §668.404(f) to make a corresponding technical change that the Secretary will not issue draft or final D/E rates for a GE program that does not meet the n-size requirements or for which SSA does not provide earnings data.

Comments: Some commenters recommended that the Department’s accountability framework recognize, or exempt from the regulations in whole or in part, programs with exceptional performance under the accountability metrics. A few commenters suggested that institutions or programs with low default rates should be exempt from assessment under the D/E rates measure. Several commenters proposed 15 percent as the appropriate threshold to identify exceptional performance under iCDR, while a few commenters suggested that programs with a pCDR below 30 percent should be exempt from the D/E rates measure. Similarly, a few commenters suggested exemptions for programs or institutions with low rates of borrowing. Specifically, commenters said a program should be deemed to be passing the D/E rates measure if the majority of students who complete the program do not have any debt at the time of graduation.

Other commenters suggested the Department exempt programs with high completion or job placement rates from both the pCDR measure and D/E rates measure. They said high performance on these alternative metrics would demonstrate that programs are successfully preparing students for gainful employment in a recognized occupation. Several commenters contended that a program that provides the highest lifetime net benefits to students who complete the program is an exceptional performer. The commenters proposed that this would be established by subtracting average costs of program attendance from average graduate earnings after factoring in low-income and subgroup characteristics of graduates.

One commenter recommended the Department apply a higher annual earnings rates passing threshold of 13 percent for programs operated by for-profit institutions that adopt programs similar to trial enrollment periods, which would allow students to tryout a program for short period of time with the option of withdrawing from the program without paying any tuition or fees. The commenter also suggested the Department should provide that institutions that implement trial enrollment periods are eligible under the title IV, HEA programs if their programs satisfy the pCDR requirements alone, as the 2011 Prior Rule provided with respect to repayment rate.

Discussion: We appreciate the suggestions for recognizing GE programs that exhibit exceptional performance. There are exemplary programs at institutions across all sectors, including at for-profit institutions and community colleges. We also believe that it is important to identify these programs to recognize their achievements and so that they can be emulated.

However, we disagree with the commenters who suggested that programs or entire institutions should be exempted from some or all parts of the regulations as a reward for exceptional performance. The Department must apply the same requirements to all programs under these regulations and assess all programs equally. Accordingly, we decline to adopt the commenters’ suggestions.

We also disagree with the commenter who recommended we apply an annual earnings rate threshold of 13 percent for programs operated by for-profit institutions that offer tuition- and fee-free enrollment trial periods. The calculation of the D/E rates measures does not evaluate students who withdraw before completing a program, and we accordingly, do not believe an enrollment trial period is pertinent to the thresholds for the D/E rate measures. Institutions may, of course, offer enrollment trial periods for their programs and we encourage them to do so.

We will continue to consider ways to recognize exceptional programs. In the meantime, we expect that the disclosure requirements of the regulations will help students identify programs with exceptional performance. We also expect that the disclosures will allow institutions to identify these programs for the purpose of adopting successful practices that lead to exceptional results for students. Finally, we note that programs that are performing at an exceptional level will pass the D/E rates measure and this will be reflected in their disclosures and promotional materials.

Changes: None.

§668.404 Calculating D/E Rates

Including Students Who Do Not Complete the Program in the D/E Rates Measure

Comments: We received a number of comments responding to the Department’s question about whether we should include students who do not complete a GE program in calculating D/E rates.

Several commenters urged the Department to hold institutions accountable for students who do not complete GE programs, arguing that these students often accumulate large amounts of debt, even in short periods of time, that they struggle to repay. Some commenters believed students who do not complete a program should be included in the D/E rates calculations to avoid allowing poor-quality programs to remain eligible for title IV, HEA program funds. Other commenters argued it would be inappropriate to include the debt and earnings of students who do not complete because the earnings of those students and their ability to repay their loans do not reflect the quality of the program they attended. These commenters believed that if students do not complete a GE program, they cannot benefit from the training the program offers. The commenters reasoned that students who do not complete a program are much less likely to qualify for the types of jobs for which the program provides training, and far more likely to obtain employment in completely different fields. One commenter that favored excluding students who do not complete a program stated that the reasons a student drops out of a program are correlated with socioeconomic factors (e.g., the student is a single parent, is unprepared for college work, or is a first-generation college student) that are also correlated with low earnings. The commenter cited a study conducted by Charles River Associates, commissioned by APSCU, showing that, of the students who do not complete a program, 50 percent drop out within the first six months of enrolling in the program and 75 percent drop out within the first year. The commenter asserted that the debt these students accumulate is relatively low, and, accordingly, churn is not necessarily a negative outcome and institutions should not be discouraged from allowing non-traditional students to explore different options.

Some commenters, however, did not support including students who do not complete a program because programs with high drop-out rates may have low D/E rates as many students would not remain enrolled long enough to accumulate large amounts of debt.

Discussion: As discussed in “§668.403 Gainful Employment Program Framework,” we agree it is important to hold institutions accountable for the outcomes of students who do not complete a GE program. However, we do not believe that the D/E rates measure is an appropriate metric for this purpose for some of the reasons noted by the commenters. In addition, we agree that including students who do not complete a program in the D/E rates measure could have the perverse effect of improving the D/E rates of some of those programs because students who drop out early may accrue relatively lower amounts of debt than students who complete the program.

Changes: None.

Comments: One commenter recommended that the Department determine which students to include in the calculation of D/E rates based on the amount of debt that a student accumulates, rather than only on whether or not a student completed the program. The commenter agreed with others that an institution should not be held accountable in situations where students incur a minimal amount of debt before dropping out of a GE program, acknowledging that students who do not complete a program will likely have lower earnings than those who complete the program. However, the commenter argued that, at the same time, institutions should be accountable for students who accumulate a significant amount of debt to attend a GE program but ultimately do not complete that program. The commenter believed that, at a certain point, if a student has accrued high levels of debt for attending a program, then the program should have prepared the student for gainful employment in that field to some extent. As an example, the commenter offered that all students who borrow more than $15,000 should be included in the calculation of D/E rates.

Discussion: The Department appreciates but cannot adopt this suggestion. First, we lack sufficient data and evidence to set a threshold for the amount of debt that would be considered sufficiently excessive to warrant including a student in the calculation. Second, as previously discussed, we do not believe it is appropriate to include in the D/E rates measure students who did not complete a GE program. Finally, the notion that including in the D/E rates measure only those students with significant or high levels of debt would not account for the students who incur less debt but are having difficulty repaying their loans because of low earnings.

Changes: None.

Two-Year Cohort Period

Introduction: We received a number of comments on the two-year cohort period that the Department uses in calculating the D/E rates. To aid readers in their review of the comment summaries and our responses, we provide the following context.

Under the regulations, the two-year cohort period covers the two consecutive award years that are the third and fourth award years prior to the award year for which the D/E rates are calculated or, for programs whose students are required to complete a medical or dental internship or residency, the sixth and seventh award years prior to the award year for which D/E rates are calculated. The Department will calculate the D/E rates for a GE program by determining the annual loan payment for the students who completed the program during the two-year cohort period and obtain from SSA the mean and median aggregate earnings of that group of students for the most recently available calendar year. Because the earnings data we obtain from SSA are for a calendar year, and because students included in the two-year cohort period may complete a program at any time during the cohort period, the length of time that a particular student could potentially be employed before the year for which we obtain earnings data from SSA varies from 18 to 42 months. Counting the year for which we obtain earnings data (earnings year) would extend this period of employment to 30 to 54 months. For example, for D/E rates calculated for the 2015 award year (July 1, 2014 to June 30, 2015), the two-year cohort period is award years 2011 (July 1, 2010 to June 30, 2011) and 2012 (July 1, 2011 to June 30, 2012). We will obtain the annual earnings of students who completed the program during this two-year cohort period from SSA for the 2014 calendar year. So, a student who completes the program at the very beginning of the two-year cohort period, on July 1, 2010, and is employed immediately after completion could be employed for up to 42 months--from July 2010 through December 2013--before the year for which earnings are used to calculate the D/E rates, and up to 54 months if the earnings year itself is included. A student who completes the program at the very end of the two-year cohort period, on June 30, 2012, and is employed immediately after completing the program could be employed for up to 18 months--July 1, 2012 through December 2013--before the year for which earnings data are obtained, and up to 30 months if the earnings year itself is included. Accordingly, although in the NPRM we, and many of the commenters, referred to a three-year employment period, there is a range of possible employment periods for students who complete a program in a two-year cohort period.

Comments: Several commenters requested that the Department clarify which year is the “most currently available” year for SSA earnings data in §668.404(c).

Discussion: The following chart provides the earnings calendar year that corresponds to each award year for which D/E rates will be calculated.

	Award year for which the D/E Rates are calculated	2014-2015	2015-2016	2016-2017	2017-2018
Medical and Dental Programs	Two-Year Cohort Period	2007-2008	2008-2009	2009-2010	2010-2011
		2008-2009	2009-2010	2010-2011	2011-2012
	Four-Year Cohort Period			2007-2008	2008-2009
				2008-2009	2009-2010
				2009-2010	2010-2011
				2010-2011	2011-2012
All Other GE Programs	Two-Year Cohort Period	2010-2011	2011-2012	2012-2013	2013-2014
		2011-2012	2012-2013	2013-2014	2014-2015
	Four-Year Cohort Period	2008-2009	2009-2010	2010-2011	2011-2012
		2009-2010	2010-2011	2011-2012	2012-2013
		2010-2011	2011-2012	2012-2013	2013-2014
		2011-2012	2012-2013	2013-2014	2014-2015
	SSA Earnings Year (Jan. 1-Dec. 31)	2014	2015	2016	2017
	Department Receives Mean and Median Earnings from SSA (anticipated dates)	Feb. 2016	Feb. 2017	Feb. 2018	Feb. 2019

Changes: None.

Comments: Commenters raised various concerns regarding the definition of the “two-year cohort period.”

Some commenters believed that evaluating earnings after three years is arbitrary, will lead to underestimating how much borrowing is reasonable for education, and will not adequately account for the long-term benefits of completing a program. These commenters asserted that many students experience substantial increases in earnings later in their careers as they gain experience or various licensures, and that using earnings after only three years would therefore understate the value of the program. Similarly, some commenters asserted that many individuals experience significant income fluctuations in the initial years of their careers.

Some commenters expressed concern that evaluating programs using graduates’ earnings three years after graduation will cause institutions to stop offering programs with strong long-term salary growth potential but with low starting salaries. Along these lines, other commenters believed that this approach will lead institutions to offer a disproportionate number of programs in higher-paying fields like business and information technology rather than programs in less lucrative fields like teaching and nursing. To address these concerns, several commenters recommended modifying the proposed regulations to evaluate programs based on graduates’ earnings at a later time in their careers. The commenters suggested different points in time that would be appropriate, varying from three to 10 years after completion. Other commenters recommended using a rolling average of graduates’ earnings over several years, rather than a snapshot at three years.

Some commenters asserted that, in some cases, the Department will be obtaining earnings data for graduates who were employed for just 18 months. They suggested that students’ ultimate earnings, particularly for professional school graduates, would be better reflected by allowing for a longer period after graduation or after the completion of residency training or fellowships for medical or dental school graduates before D/E rates are calculated.

Discussion: We believe that measuring earnings for the employment range covered by the two-year cohort period strikes the appropriate balance between providing ample time for students to become employed and increase earnings past entry level and yet not letting so much time pass that the D/E rates are no longer reflective of the current or recent performance of the program.

The D/E rates measure primarily assesses whether the loan debt incurred by students actually “pay[s] dividends in terms of benefits accruing from the training students received,” and whether such training has indeed equipped students to earn enough to repay their loans such that they are not unduly burdened. H.R. Rep. No. 89-308, at 4 (1965); S. Rep. No. 89-758, at 7 (1965). As discussed in “§668.403 Gainful Employment Program Framework,” high D/E rates indicate that the earnings of a program’s graduates are insufficient to allow them to manage their debt. The longer the Department waits to assess the ability of a cohort of students to repay their loans, the less relevant that assessment becomes for prospective students, and the more likely it is that new students will attend a program that is later determined to be ineffective at preparing students for gainful employment. Assessing the outcomes of less recent graduates would also make it more difficult for institutions to improve student and program outcomes under the D/E rates measure as it would take many years before subsequently enrolled students who complete the program would be included in the D/E rates calculation.

There is no evidence that relying on earnings during the employment range used in the regulations would actually create the disincentives or result in the harms that commenters suggest. Specifically, many programs training future nurses, teachers, and other modest-earning professions, as characterized by the commenters, would successfully pass the D/E rates measure. For example, of the 497 licensed practical/vocational nurse training programs in the 2012 GE informational D/E rates data set, 493 (99 percent) passed, 4 (1 percent) fell in the zone, and none of the programs failed. In addition, of the 113 programs categorized as education programs by the two-digit CIP code,¹³³ 109 (96 percent) passed, 3 (3 percent) were in the zone, and only 1 (1 percent) failed. This suggests that programs preparing students for “less lucrative” occupations or occupations with delayed economic benefits are not problematic as a class--many programs in these categories succeed in ensuring that the debt of their students is proportional to earnings.

Changes: None.

Comments: Some commenters believed that using both two-year and four-year cohort periods would be confusing, make it difficult to compare programs, and result in misleading comparisons. The commenters reasoned that because economic conditions may vary markedly from year to year, including earnings of graduates who are employed for an additional two years under a four-year cohort period would inflate the earnings used in calculating the D/E rates. Consequently, the commenters suggested that the Department use only a two-year cohort period. In cases where fewer than 30 students complete a program during the two-year cohort period, the commenters suggested that the Department treat the program as passing the D/E rates measure.

Some commenters argued that the Department did not provide any data showing the effect of the four-year cohort period on GE programs or otherwise adequately justify the use of a four-year cohort period. These commenters suggested removing the four-year cohort period provisions until the Department completes a more thorough assessment.

Some commenters believed that the proposed regulations did not adequately specify when and how the Department intends to use the two-year cohort period and four-year cohort period, specifically taking issue with what they believed was the repetitious use of the reference to “the cohort period.” The commenters opined that the Department should specify when the two-year cohort period and four-year cohort period are used, in the same manner in which proposed §668.502(a)(1) of subpart R describes how the Department would determine the cohort for the pCDR measure. Similarly, the commenters were concerned that institutions would be confused by the language used in proposed §668.404(f)(1) to describe the circumstances under which the Department would not calculate D/E rates if fewer than 30 students completed the program.

Discussion: We agree that using the four-year cohort period may add some complexity, but believe that this concern is outweighed by the benefits of evaluating more programs under the D/E rates measure as some programs that do not meet the minimum n-size of 30 students who complete the program over the two-year cohort period would do so when the four-year cohort period is applied.

With respect to the commenters who argued that the Department did not adequately justify using a four-year cohort period, we disagree. In the NPRM, the Department acknowledged that one of the limitations of using an n-size of 30 as opposed to an n-size of 10 is that use of a larger n-size results in significantly fewer GE programs being evaluated. We estimated that, at an n-size of 30, the programs that will be evaluated under the D/E rates measure account for 60 percent of the enrollment of students receiving title IV, HEA program funds in GE programs. Using the four-year cohort period will help to increase the number of students in programs that are accountable under the D/E rates measure.

In response to comments regarding how the Department intends to use the two- and four-year cohort periods, we note that the preamble discussion in the NPRM under the heading “Section 668.404 Calculating D/E rates,” 79 FR 16448-16449, contains a thorough explanation. In short, the calculations for both D/E rates would be based on the debt and earnings outcomes of students who completed a program during a cohort period. As with the 2011 Prior Rule, for D/E rates to be calculated for a program, a minimum of 30 students would need to have completed the program, after applying the exclusions in §668.404(e), during the cohort period. If 30 or more students completed the program during the third and fourth award years prior to the award year for which D/E rates are calculated, then the cohort period would be that “two-year” cohort period. If at least 30 students did not complete the program during the two-year cohort period, then the cohort period would be expanded to include the previous two years, the fifth and sixth award years prior to the award year for which the D/E rates are being calculated, and rates would be calculated if 30 or more students completed the program during that “four-year cohort period.” If 30 or more students did not complete the program over the two-year cohort period or the four-year cohort period, then D/E rates would not be calculated for the program.

The two- and four-year cohort periods as described would apply to all programs except for medical and dental programs whose students are required to complete an internship or residency after completion of the program. For medical and dental programs, the two-year cohort period would be the sixth and seventh award years prior to the award year for which D/E rates are calculated. The four-year cohort period for these programs would be the sixth, seventh, eighth, and ninth award years prior to the award year for which D/E rates are calculated.

Changes: We have revised the definition of “cohort period” in §668.402 to clarify that we use the two-year cohort period when the number of students completing the program is 30 or more. We use the four-year cohort period when the number of students completing the program in the two-year cohort period is less than 30 and when the number of students completing the program in the four-year cohort period is 30 or more.

Comments: Another commenter suggested that the Department replace the term “cohort period” with the term “GE cohort period” to avoid confusion with the iCDR regulations.

Discussion: We appreciate the commenter’s concern but we do not believe that the regulations are confusing with respect to the term “cohort period.” While “cohort” is a defined term under the iCDR regulations, those regulations do not use the term “cohort period.” The term “cohort period” appears only in these regulations.

Changes: None.

Comments: One commenter raised concerns about calculating D/E rates for graduates of veterinary or medical school using earnings after only three years following completion of the program. Using the example of a student graduating during the 2011-2012 award year from a veterinary program, whose earnings the commenter believed would be measured based upon SSA earnings data for calendar year 2014, the commenter asserted that the D/E rates would not be an accurate reflection of the student’s ability to earn an income or be gainfully employed.

Discussion: We believe that the commenter may have misunderstood the D/E rates calculation for graduates of medical and dental programs whose students are required to complete a period of internship or residency. The regulations do, in fact, consider the resulting delay between when such students complete their respective programs and when they may begin professional practice. For medical and dental programs, the two-year cohort period would be the sixth and seventh award years prior to the award year for which D/E rates are calculated. The four-year cohort period would be the sixth, seventh, eighth, and ninth award years prior to the award year for which D/E rates are calculated. In the example given by the commenter, SSA earnings for the 2014 calendar year would be used in the D/E rates calculations for the 2014-2015 award year. The two-year cohort period for a medical program would be 2007-2008 and 2008-2009.

Veterinarians, on the other hand, do not have a required internship or residency. They can begin practice immediately following graduation from veterinary school. As with other types of training programs that do not require an internship or residency after program completion, we believe that graduates of veterinary programs will have sufficient time after completion of their program to become employed and increase earnings beyond an entry level in order for the program they attended to be accurately assessed under the D/E rates measure.

Changes: None.

Comments: One commenter said that since there has been no informational rate data provided for medical school programs, institutions with these types of programs would be at a greater disadvantage under accountability metrics that determine a program’s continuing eligibility for title IV, HEA program funds based on historical program performance.

Discussion: The Department did not provide informational rate data for medical school programs because we do not have such data. However, an institution can reasonably be expected to know about the borrowing patterns of its students, because the institution’s financial aid office typically “packages” financial aid, including loans, in arranging financial aid for students. All institutions should also be conducting the necessary local labor market research, including engaging with potential employers, to determine the typical earnings for the occupations for which their programs provide training. Institutions may use this information to estimate their results under the D/E rates measure. Additionally, we believe that the “zone” provisions described under “§668.403 Gainful Employment Program Framework,” together with the transition period in §668.404(g) described later in this section, will provide programs with an adequate opportunity to make adjustments and improvements to their programs as needed.

Changes: None.

Use of Mean and Median Earnings

Comments: Some commenters supported the proposal in §668.404(c)(2) to use the higher of the mean or median annual earnings to calculate the D/E rates, arguing that using the higher of the two would better reflect the earnings of students who complete programs and would therefore be fairer to institutions than using only the mean or only the median.

Other commenters recommended using either the mean or the median earnings to calculate D/E rates, rather than the higher of the two. These commenters believed that the proposed approach would make it difficult for consumers, schools, researchers, policymakers, and others to understand the D/E rates. The commenters also said that the informational rates released by the Department in 2010, which were calculated using the higher of the mean or median earnings, were confusing. The commenters expressed further concern that, in addition to causing confusion, the use of either the mean or the median annual earnings would undermine the public’s ability to compare D/E rates across GE programs. These commenters did not believe that the Department presented a reasoned basis for using the higher of the mean or median earnings and argued that the Department’s proposed approach would weaken the D/E rates measure.

Some commenters believed that the Department should use the mean in all cases, but they did not elaborate on their reasons for that approach. Other commenters recommended using the median in all cases because they believed that it would be inconsistent to use median loan debt in the numerator of the D/E rates but the mean earnings in the denominator. They also argued that using the median would guarantee that the earnings data reflect the outcomes of at least 50 percent of the students who complete a program and that the earnings of one outlier student would not skew the calculation.

Discussion: We agree with commenters that it is important that consumers and other stakeholders receive clear, useful information about program outcomes. By using the higher of the mean or median earnings, the regulations strike a balance between providing stakeholders information that is easy to use and comprehend and ensuring an accurate assessment of program performance.

Because using the mean or median earnings may affect a particular program, we use the higher of the mean or median earnings to account for the following circumstances:

• In cases where mean earnings are greater than median earnings, we use the mean because the median may be sensitive to zero earnings. For example, if the majority of the students on the list submitted to SSA have zero earnings, the program would fail the D/E rates measure even if most of the remaining students had relatively high earnings. In other words, when the median is less than the mean, there may be a large number of students with zero earnings. So, we use the mean earnings to diminish the sensitivity of the D/E rates to zero earnings and better reflect the central tendency in earnings for programs where many students have extremely low and extremely high earnings.

• In cases where median earnings are greater than mean earnings, we use the median because it is likely that there are more students who completed a program with relatively high earnings than with relatively low earnings. For these cases, we believe that median earnings are a more representative estimate of central tendency than mean earnings. Relatively high median earnings indicate higher employment rates, and by using the median when it is higher than the mean, we reward programs where a high fraction of students who complete a program obtain employment.

Changes: None.

Comments: A few commenters suggested that, if the Department calculates the D/E rates using the higher of mean and median earnings, the Department should publish both the mean and median earnings data for each GE program and indicate which figure was used in the D/E rates calculation. These commenters argued that disclosing this information would mitigate some of the concerns about difficulties comparing and conducting analyses across programs.

Discussion: As an administrative matter, we agree to post the mean and median earnings for all GE programs on the Department’s Web site, and we will identify whether the mean or the median earnings were used to calculate the D/E rates for any particular program.

Changes: None.

Comments: A commenter suggested that, in calculating the D/E rates, we use the earnings of the student’s household, and not just the earnings of the student.

Discussion: We do not believe it would be appropriate to use household earnings in the calculation of D/E rates. The earnings of other members of the household have no relation to the assessment of the effectiveness of the program in which the student was enrolled.

Changes: None.

Comments: One commenter recommended using the earnings of the top 10 percent of earners in the cohort in the denominator of the D/E rates calculations, rather than the higher of the mean or median earnings of all students who completed the program in the cohort period (other than those excluded under §668.404(e)). The commenter believed that using the top 10 percent of earners would best represent the earnings potential of students who complete the program and would mitigate the effects of students who opt to leave the workforce, work other than full-time, work in a different field, or are not top performers at work.

Discussion: The regulations seek to measure program-level performance, which we believe is best accomplished by including the outcomes of all students who completed a program. An assessment of just the top 10 percent of earners may provide information on how those particular students are faring, but would say little about actual overall program performance. For example, if the other 90 percent of students were unable to secure employment, then reviewing the outcomes of just the top 10 percent would result in a substantially inaccurate assessment. Further, as discussed in this section and in “§668.403 Gainful Employment Program Framework,” we believe several aspects of the regulations, including use of mean and median earnings, use of a multi-year cohort period with a minimum n-size, and allowing several years of non-passing results before a program loses eligibility for title IV, HEA program funds reduce the likelihood to close to zero that a typically passing program will be mischaracterized as failing or in the zone due to an atypical cohort of students who complete the program such as those identified by the commenter.

Changes: None.

Comments: One commenter argued that the Department should consider policies that would help students succeed in the recovering labor market, rather than examine average graduate earnings.

Discussion: We agree with the commenter that policies should be designed to help students succeed in the job market. These regulations are intended to accomplish this very objective, at least partly by measuring student earnings outcomes. As a result of the disclosure requirements, which will include earnings information, students and prospective students will have access to more and better information about GE programs so that they can choose a program more likely to lead to successful employment outcomes. The minimum certification requirements will ensure that all GE programs provide students who complete programs with the basic academic qualifications necessary for obtaining employment in their field of training. And, because programs will be held accountable for the outcomes of their students under the D/E rates measure, which requires an assessment of earnings, we expect that, over time, institutions will offer more high-quality programs in fields where students can secure employment at wages that allow them to repay their debt.

Changes: None.

Poverty Guideline

Comments: Some commenters noted that in calculating the discretionary income rate under the proposed regulations, the Department would use the most currently available annual earnings and the most currently available Poverty Guideline, but those items would correspond to different years. The commenters provided an example where the most currently available annual earnings year might be the 2014 tax year, but the Poverty Guideline used to calculate the rate could be for the 2015 year. According to the commenter, this discrepancy could negatively affect a program’s discretionary income rate because the benefit of obtaining the education would not be observed if historical earnings are used. The commenters suggested that, to the extent possible, the Department should use the Poverty Guideline for the same year that the Department obtains SSA earnings data.

Discussion: Under the discretionary income rate, a portion of annual earnings, the amount equal to 150 percent of the Poverty Guideline for a family size of one, is considered to be protected or reserved to enable students to meet basic living costs. Only the remaining amount of annual earnings is considered to be available to make loan payments.

As explained by the Department of Health and Human Services (HHS), the Poverty Guidelines issued at the beginning of a calendar year reflect price changes for the most recently completed calendar year.¹³⁴ In the example provided by HHS, the Poverty Guidelines issued in January 2014 take into account the price changes that occurred during the entire 2013 calendar year. Because the HHS process typically results in higher Poverty Guidelines from year to year, we agree with the commenters that the Poverty Guideline used to calculate the discretionary income rate should correspond with the year for which we obtain earnings data from SSA. Otherwise, earnings would be over-protected. For example, as shown in the chart under “Two-Year Cohort Period,” we will not obtain earnings data from SSA for the 2014 calendar year until early 2016. So, under the proposed regulations we would have calculated the discretionary income rate using 2014 calendar year earnings and the Poverty Guideline published by HHS in 2016, which would reflect price changes in 2015. It would be more appropriate to use the Poverty Guideline that reflects the price changes during the calendar year for which we obtained earnings, 2014, which would be the Poverty Guideline published in 2015 by HHS.

Changes: We have revised §668.404(a)(1) to specify that in calculating the discretionary income rate, the Department will use the Poverty Guideline for the calendar year immediately following the calendar year for which the Department obtains earnings data from SSA.

Comments: One commenter stated that according to 2011-2012 NPSAS data, of students attending for-profit institutions, 50 percent have dependent children and 30 percent have at least two dependent children. In view of this information, the commenter concluded that because the discretionary income rate is calculated based on an assumed family size of one, student debt burden is understated.

Similarly, other commenters suggested that the Department use the Poverty Guideline for families. The commenters believed that institutions should be sensitive to students with dependents who are seeking to improve their credentials and earnings by enrolling in GE programs and that using the appropriate Poverty Guideline would provide that incentive to institutions.

Discussion: Although we agree that applying the Poverty Guideline based on actual family size would result in a more precise assessment of loan burden, it would be difficult and highly burdensome, if not impossible, to adopt this approach. There is no apparent way for either institutions or the Department to collect information about the family size of students after they complete a program. At or before the time students enroll in a GE program, they may have reported the number of dependents on the FAFSA, but that information may change between the time students completed the program and when the Department calculates the D/E rates. Even if we were able to collect accurate information, applying a different Poverty Guideline for each student who completed a program, or otherwise accounting for differences in family size, would not only complicate the calculation but result in D/E rates that may not be comparable as there would be different assumptions for discretionary income for different programs. The rate for a program with an average family size of two would be different than the rate for the same program with an average family size of four, creating situations where the Department would not be uniformly assessing the performance of programs and making it difficult for students and prospective students to compare programs.

Changes: None.

Loan Debt

Comments: Several commenters were critical of the Department’s proposal to calculate a program’s loan debt only as a median. The commenters recommended that we apply the lower of the mean or median loan debt to the D/E rates calculation. Some of these commenters argued that using the median loan debt would create distorted assessments of debt burden for programs that have a small number of students who completed.

A number of commenters stated that using median loan debt would unfairly benefit low-cost programs offered by community colleges because the regulations cap loan debt at the lesser of the student’s tuition and fees and books, supplies, and equipment or the amount of debt the students incurred for enrollment in the program. Other commenters suggested that instead of using the lesser of these amounts to calculate the median loan debt, the Department should use the total amount of loan funds that a student used to pay direct charges after taking into account any grants or scholarships the student received to pay for these charges. The commenters argued that if the D/E rates measure is designed to hold institutions accountable for how much they assess students for direct charges, the amount assessed should be the amount of direct costs net of institutional aid. Otherwise, the student’s actual costs for direct charges would be overstated.

Some commenters asserted that because independent students may be able to borrow larger amounts than dependent students, a program for which the majority of students who completed the program were independent students would tend to have a higher median loan debt. For this reason, the commenters opined that institutions might be inclined to discourage independent students from enrolling or avoid enrolling other students that are more likely to borrow.

Discussion: We elected to use the median loan debt because a median, as a measure of central tendency of a set of values, is less affected by outliers than a mean. Means are generally more sensitive to extremely high and low values compared to values that do not fall on either extreme, while medians are more sensitive to the values near the 50th percentile of a population being sampled.¹³⁵ We also elected to use median loan debt, as opposed to the mean, to reward programs that keep costs sufficiently low such that the majority of students do not have to borrow. For example, if a majority of students in a program only receive Pell Grants and do not borrow, the median loan debt will be zero for that program. Taking into consideration the same logic, we elected to use the mean for earnings because, although the mean is more sensitive to extreme values, it is also less sensitive to zero earnings values. For example, if a majority of students in a program earn zero dollars, the median would be zero, but the mean may still be a substantially greater number than zero if some students have high levels of earnings. We believe it is appropriate to credit such programs for the minority of students who have high earnings and that such a calculation more accurately reflects the central tendency in the earnings of the students who completed the program.

With regard to programs with a small number of students completing the program, as discussed in this section, we mitigate the potential for distorted outcomes by requiring a minimum n-size of 30 students who completed the program in the cohort period for D/E rates to be calculated.

We do not agree with the comment that programs offered by community colleges would benefit more from the capping of a student’s loan amount to tuition and fees, and books, equipment, and supplies, because many students at community colleges do not borrow or borrow amounts less than the total amount of tuition and fees and books, equipment, and supplies. For these students, the loan cap would not be applied in determining a program’s median loan debt.

With regard to the suggestion that median loan debt should be based on the total amount of loans used to pay direct charges, the commenter is referring to situations where grant or scholarship funds are used ahead of loan funds to pay for direct costs. In these situations the grants and scholarships may be designated to pay direct costs so the amount of loan debt would be no more than the amount of direct costs that were not paid by the grant and scholarships funds. Whereas the suggestion would reduce the amount of the loan debt used to calculate the D/E rates by effectively replacing loan funds with grant or scholarship funds, we believe doing so is contrary to the intent of these regulations to evaluate whether students are able to service the amount of loan debt for the amount up to the direct charges assessed by the institution.

In response to the concerns that an institution might alter its admissions policies based on a student’s dependency status or need to borrow, we note that because the loan cap limits the amount of debt on a student-by-student basis to the total amount of direct charges (tuition and fees, and books, supplies, and equipment), the principal factor influencing a program’s median loan debt may be tied more to the amount of the direct charges than to the amount that individual students borrow. In addition, as discussed in the Regulatory Impact Analysis, our analysis shows that dependency status or socioeconomic background are not determinative of results and so we do not believe the regulations create this incentive.

Changes: None.

Comments: A few commenters asked the Department to clarify how it will calculate a program’s median loan debt. They argued that the proposed methodology could be interpreted in two ways, each likely yielding a different result. Under one reading, the Department would determine student by student the lesser of the loan debt and the total program costs assessed to that student, and then calculate the median of all of those amounts. Under another reading, the Department would determine the median amount of all students’ loan debts and the median amount of all students’ total program costs and use the lesser amount.

Discussion: The commenters’ first reading is correct. We will determine individually, for each student who completes a program, the lesser of the total amount of a student’s loan debt and the total costs assessed that student for tuition and fees and books, supplies, and equipment, and use whichever of these amounts is lower to calculate the median loan debt for the program.

Changes: We have revised §668.404(b)(1) to more clearly describe how the Department will calculate the median loan debt for a program. We have also revised §668.404(d)(2) to clarify that for the purpose of determining the lesser amount of loan debt or the costs of tuition and fees and books, supplies, and equipment, we attribute these costs to a GE program in the same way we attribute the loan debt a student incurs for attendance in other GE programs.

Comments: One commenter stated that loan debt incurred by a medical school graduate increases because interest accrues while the student is in a residency period and that this additional debt would affect D/E rates.

Discussion: In determining a student’s loan debt, the Department uses the total amount of loans the student borrowed for enrollment in a GE program, net of any cancellations or adjustments made on those loans. Any interest that accrues on those loans or that is subsequently capitalized is not considered loan debt for the purpose of calculating a program’s D/E rates.

Changes: We have revised §668.404(d)(1)(i) to clarify that the total amount borrowed by a student for enrollment in a GE program is the total amount disbursed less any cancellations or adjustments.

Comments: Some commenters recommended that the Department clarify the timing and conditions under which the Department would remove loan debts for students for whom SSA does not have earnings information.

Discussion: As explained more fully in “§668.405 Issuing and Challenging D/E Rates,” at the time that SSA provides the Department with the mean and median earnings of the students who completed a program, SSA will also provide a count of the number of students for whom SSA could not find a match in its records, or who died. Before calculating the program’s median loan debt, we will remove the number of highest loan debts equal to the number of students SSA did not match. Since we do not have information on each individual student who was not matched with SSA data, we remove the highest loan debts to provide a conservative estimate of median loan debt that ensures we do not overestimate the amount of debt borrowed by students who were successfully matched with SSA data.

Changes: None.

Comments: Several commenters stated that the proposed regulations do not clearly show how debt is attributed in situations where students are enrolled in multiple GE programs simultaneously at the same or different credential levels.

Discussion: Under §668.411(a), an institution is required to report a student’s enrollment in each GE program even when the student was enrolled in more than one program, either at different times, at the same time, or for overlapping periods. The institution reports information about each enrollment (dates, tuition and fees, books, supplies, and equipment, amounts of private student loans and institutional financing, etc.) separately for each program. The Department uses the reported enrollment dates to attribute a student’s loan amounts to the relevant GE program. In instances where a student was enrolled in more than one GE program during a loan period, we attribute a portion of the loan to each program in proportion to the number of days the student was enrolled in each program.

In attributing loans, we exclude those loans, or portions of loans, that were made for a student’s enrollment in a non-GE program (e.g., a degree program at a public or not-for-profit institution). In instances where a loan was made for a period that included enrollment in both a GE program and in a non-GE program, the loan will be attributed to the GE program under the assumption that the student would have taken out the loan if the student was enrolled only in the GE program.

Changes: None.

Comments: Some commenters stated that many students enter for-profit schools after accumulating loan debt from traditional colleges, and that the added debt may severely affect the students’ ability to repay their loans.

Discussion: We agree that increasing amounts of debt, regardless of where that debt was incurred, will affect a student’s ability to repay his or her loans. However, the D/E rates are calculated based only on the amount a student borrowed for enrollment in GE programs at the institution, and are not based on any debt accumulated at other institutions the student previously attended, except where the student incurred debt to attend a program offered by a commonly owned or controlled institution, and where disregarding the common ownership or control would allow manipulation of D/E rates, as provided under §668.404(d)(3).

Changes: None.

Comments: One commenter suggested that the Department revise §668.404(d)(1)(iii) to clarify that the amount of any obligation that a student owes the institution is the amount outstanding at the time the student completes the program. The commenter provided the following language: “The amount outstanding, as of the date the student completes the program, on any credit extended by or on behalf of the institution for enrollment in the GE program that the student is obligated to repay after program completion, even if that obligation is excluded from the definition of a ‘private education loan,’ in 34 CFR §601.2.”

Other commenters opined that total loan debt should not include any funds a student owes to an institution unless those funds are owed pursuant to an executed promissory note.

Discussion: We believe that any amount owed to the institution resulting from the student’s attendance in the GE program should be included, regardless of whether it is evidenced by a promissory note or other agreement because the amount owed is the same as any other debt the student is responsible to repay. For this reason, we clarify that, in addition to an obligation stemming from extending credit, an obligation includes any debts or unpaid charges owed to the institution. In addition, we adopt the commenter’s suggestion to specify that the amount included in determining the student’s loan debt is the amount of credit extended (not from private education loans) by or on behalf of the institution, including any unpaid charges, that are outstanding at the time the student completed the program.

Changes: We have revised the regulations to clarify, in §668.404(d)(1)(iii), that loan debt includes any credit, including for unpaid charges, extended (other than private education loans) by or on behalf of an institution, that is owed to the institution for any GE program attended at the institution, and that the amount of this institutional credit includes only those amounts that are outstanding at the time the student completed the program.

Comments: One commenter asked the Department to clarify if institutional debt would include amounts owed to the institution resulting from the institution’s return of unearned title IV aid under the return to title IV aid regulations.

Discussion: The situation described by the commenter results where a student enrolls at an institution, the student withdraws at a point where the institution returns the unearned portion of the student’s title IV, HEA program funds and the student is required to pay the institution at least a portion of the charges that would have been paid by those unearned funds, and the student subsequently completes a GE program at the same institution before paying those charges from the prior enrollment. We confirm that the institutional debt for the program the student completes includes the student debt from the prior enrollment at the institution. We do not believe this series of events will happen often, and it is unlikely that it would significantly change the median loan debt calculated for a program.

Changes: None.

Comments: Some commenters opined that the regulations do not provide for an accurate assessment of debt burden because, in addition to title IV loans and private loans, students use other financing options, such as credit cards and home equity loans, to cover educational expenses. They argued that the Department should not ignore these other forms of credit because doing so would understate the debt burden of students.

Discussion: While we agree that there may be instances where counting debt incurred through various financing options may provide a better assessment of total debt, the information needed to include that debt in calculating the D/E rates is generally not available and may not be useable if the debt is not tied directly to a student. For example, an institution would not typically know or inquire whether a student or the student’s family obtained an equity loan or used a portion of that loan to pay for educational expenses. For a credit card, even when an institution knows that it was used to pay for educational expenses, the institution does not typically know or inquire whether the amount charged on the credit card was paid in full shortly thereafter or created a longer-term obligation similar to a student loan.

Changes: None.

Comments: Some commenters argued that the Department did not clarify how an institution might “reasonably be aware of” a student who has a private student loan and that, as a result, some borrowing will go unreported, perhaps intentionally. One of the commenters noted that Federal law does not currently require an institution to certify that a borrower has demonstrated need to receive a private student loan. As noted in a 2012 study conducted by the CFPB and the Department, according to the commenter, private student lenders have directly originated loans to students, sometimes without the school’s knowledge. The commenters encouraged the Department to clarify the phrase “reasonably aware” to reduce the likelihood that institutions will engage in tactics to arrange credit from private lenders for students in an attempt to circumvent the requirements of the regulations.

Similarly, other commenters argued that the “reasonably aware” provision gives too much discretion to institutions to report private loans. The commenters stated that private loans are an expensive form of financing that is used by students attending for-profit institutions at twice the rate as students attending non-profit institutions and that, in some cases, for-profit institutions use private loans to evade the 90/10 provisions in section 487(a)(24) of the HEA. For these reasons, the commenters suggested that the Department require institutions to affirmatively assess whether their students have private loans.

Discussion: The HEOA requires private education lenders to obtain a private loan certification form from every borrower of such a loan before the lender may disburse the private education loan. Under 34 CFR 601.11(d), an institution is required to provide the self-certification form and the information needed to complete the form upon an enrolled or admitted student applicant’s request. An institution must provide the private loan self-certification form to the borrower even if the institution already certifies the loan directly to the private education lender as part of an existing process. An institution must also provide the self-certification form to a private education loan borrower if the institution itself is the creditor. Once the private loan self-certification form and the information needed to complete the form are disseminated by the institution, there is no requirement that the institution track the status of the borrower’s private education loan.

The Federal Reserve Board, in 12 CFR 226.48, built some flexibility into the process of obtaining the self-certification form for a private education lender. The private education lender may receive the form directly from the consumer, the private education lender may receive the form through the institution of higher education, or the lender may provide the form, and the information the consumer will require to complete the form, directly to the borrower. However, in all cases the information needed to complete the form, whether obtained by the borrower or by the private education lender, must come directly from the institution.

Thus, even though an institution is not required to track the status of its student borrowers’ private education loans, the institution will know about all the private education loans a student borrower receives, with the exception of direct-to-consumer private education loans, because as previously , the institution’s financial aid office “packages” most private education loans in arranging financial aid for students. We consider the institution to be reasonably aware at the very least of private education loans that its own offices have arranged or helped facilitate, including by providing the certification form. The institution must report these loans. Direct-to-consumer private education loans are disbursed directly to the borrower, not to the school. An institution is not involved in a certification process for this type of loan. Nothing prevents an institution from asking students whether they obtained direct-to-consumer private loans, and we encourage institutions to do so. However, we are not persuaded that requiring institutions to affirmatively assess whether students obtain direct-to-consumer private education loans through additional inquiry, as suggested by some commenters, will be helpful or result in reporting of additional loans that would materially impact the median loan debt of a program.

Changes: None.

Comments: A few commenters argued that loan debt should include all loans held by each student, not just loans attributed to the relevant program. The commenters suggested that by including debt previously received for attendance at prior institutions, the metric would better take into account previous educational and job experience, factors not currently reflected in the D/E rates measure.

Discussion: The Department is adopting the D/E rates measure as an accountability metric because we believe that comparing debt incurred for completing a GE program with earnings achieved after that training provides the most appropriate indication of whether students can manage the debt they incurred. We attribute loan debt to the highest credentialed program completed by a student for two reasons: earnings most likely stem from the highest credentialed program and some or all of the coursework from a lower credentialed program may apply to the higher credential program. For these reasons, in cases where a student completes a lower credential program but previously enrolled in a higher credentialed program, we do not believe it is appropriate to include the loan debt from the higher credentialed program.

Changes: None.

Comments: One commenter noted that the reference in §668.404(b)(1)(ii) to the reporting requirements relating to tuition and fees and books, equipment, and supplies is incorrect.

Discussion: The commenter is correct.

Changes: We have relocated and corrected the reference in §668.404(b)(2) to the tuition and fees and books, equipment, and supplies reported under §668.411(a)(2)(iv) and (v).

Tuition and Fees

Comments: A number of commenters agreed with the Department’s proposal to cap the loan debt for a student at the amount assessed for tuition and fees but disagreed with the proposal in §668.404(b)(1)(i) and (ii) to include books, supplies, and equipment as part of the cap. Some of the commenters stated that institutions include the costs of books, “kits,” and supplies as part of the tuition for many programs as a way to limit student out-of-pocket costs and, accordingly, did not believe they should be held accountable for those costs. A few of these commenters suggested that the Department exclude from the cap the costs of books, supplies, and equipment if an institution can show that it reduced the price of these items to the student through direct purchasing. Other commenters believed that since students may purchase the supplies they want, but not necessarily need, and because the prices for books, supplies, and equipment may vary greatly, the loan cap should include only tuition and fees.

Some commenters supported the proposed tuition and fees and books, equipment, and supplies cap, opining that because the title IV, HEA programs permit students to borrow in excess of direct educational costs, calculating the loan debt without a cap would unfairly hold institutions accountable for portions of debt unrelated to the direct cost of the borrower’s program. The commenters reasoned that inasmuch as institutions are not permitted to limit borrowing (other than on a case-by-case basis), it would be unfair to allow decisions by students to borrow above the cost of the program to affect a program’s eligibility. Some of these commenters requested that the Department give institutions more tools or the authority to reduce over-borrowing if they are to be held accountable for debt above tuition and fees.

On the other hand, some commenters objected to the cap. They asserted that limiting loan debt would invalidate the D/E rates as an accountability metric because a portion of a student’s debt (debt incurred for living expenses and other indirect costs) would not be considered.

A few commenters disagreed with the Department’s position that tuition, fees, books, supplies, and equipment are the only costs over which an institution exercises direct control. These commenters argued that an institution has control over the cost of attendance elements that enable students to borrow for indirect expenses such as room and board.

Other commenters opined that costs for books, supplies, and equipment are largely determined by students and that, even for students in the same program, costs may vary depending on whether students purchase new or used materials, rent materials, or borrow the materials. Given this variability, the commenters noted that it could be difficult for an institution to establish an appropriate amount for these items in a student’s cost of attendance budget, and were concerned that less reputable institutions may misreport data for books, supplies, and equipment to lower the amount at which the Department would cap loan debt for a program. The commenters concluded that including books, supplies, and equipment in the loan cap may hurt institutions that truthfully report information to the Department.

Discussion: We believe that an institution has control over the costs of books, supplies, and equipment, either by including those costs in the amount it charges for tuition and fees, as noted by some of the commenters, or through a process where a student purchases those items from the institution. To account for instances where the student purchases, rents, or otherwise obtains books, supplies, and equipment from an entity other than the institution, §668.411(a)(2)(v) requires the institution to report the total amount of the allowances for those items that were used in the student’s title IV Cost of Attendance (COA). As explained more fully in volume 3, chapter 2 of the FSA Handbook, section 472 of the HEA specifies the items or types of costs, like the costs for books and supplies, that are included in the COA, but the institution is responsible for determining the appropriate and reasonable amounts of those items.¹³⁶ The COA is a longstanding statutory provision with which institutions have had to comply, so we do not agree that it would be difficult for institutions to establish reasonable allowances for COA items. In any event, to comply with the reporting requirements, an institution simply reports the total amount of the COA allowances for books, supplies, and equipment or the amount of charges assessed the student for obtaining or purchasing these items from the institution, whichever amount is higher. Under this approach, it does not matter where a student purchased books or supplies or how much they paid, or whether he or she needed or wanted the supplies. The institution controls the COA allowances and controls the cost of these items.

Although we encourage institutions to reduce the costs of books and supplies, those actions have no bearing on the central premise of capping loan debt--that an institution is accountable under these regulations for the amount of debt a student incurs to pay for direct costs that the institution controls. In this regard, we limit the direct costs for items under the cap to those that are the most ubiquitous--books, supplies, and equipment. As noted in the comments, room and board is a COA item that could be included in the cap, but many GE program students enroll in distance education or online programs or attend programs at institutions that do not have or offer campus housing or meal plans.

Although we agree that it would be appropriate for research and consumer purposes to recognize all educational loan debt incurred by students attending GE programs, we disagree with the comment that limiting loan debt under the cap would invalidate the D/E rates measure. In the context of an eligibility requirement related to program performance, we believe it is appropriate to hold an institution accountable for only those program charges over which it has control, and could exercise that control to comply with the thresholds under the D/E rates measure. However, students and prospective students should have a complete picture of program outcomes, including information about the total amount of loan debt incurred by a typical student who completed the program. Accordingly, the median loan debt for a program that is disclosed under §668.412 is not limited to the amount assessed for tuition and fees and books, equipment, and supplies.

With respect to the comment that the Department should give institutions more flexibility to control student borrowing, we do not have the authority to change rules regarding loan limits because these provisions are statutory. See section 454(a)(1)(C) of the HEA, 20 U.S.C. 1087d(a)(1)(C).

Finally, we do not believe that including books, supplies, and equipment in the loan cap would encourage an institution to misreport the COA allowances for these items to the Department. We note that institutions that submit reports to the Department are subject to penalty under Federal criminal law for making a false statement in such a report. See, e.g., 18 U.S.C. 1001, 20 U.S.C. 1097(a).

Changes: None.

Comments: Some commenters were concerned that capping loan debt may inappropriately benefit GE programs with low reported direct costs. For example, a GE program may appear to have better D/E rates if an institution keeps tuition and fees low by shifting costs, and loan debt related to those costs, to housing or indirect costs that are not included in calculating the D/E rates. Consequently, the commenters believed it was unfair for some GE programs to benefit from a cap because these programs could have the same total loan debt as GE programs where the cap would not apply. The commenters concluded that lower direct costs are not necessarily indicative of lower debt and may actually serve to hide the true balance of the loan debt, an outcome that would lead the public, students, and prospective students to draw erroneous conclusions about a program’s D/E rates.

Discussion: We do not agree there is a material risk that an institution would shift costs in the manner described by the commenters to take advantage of the cap, but we will know about any changes in program costs through the reporting under these regulations and may require an institution to explain and document those changes.

Changes: None.

Comments: A commenter stated that foreign veterinary schools do not control the amount of tuition assessed for the clinical year of instruction. The commenter noted that under 34 CFR 600.56(b)(2)(i), students of foreign veterinary schools that are neither public or non-profit must complete their clinical training at veterinary schools in the United States. For the fourth or clinical year of study, the U.S. veterinary school, which is not subject to the GE regulations, charges the foreign school an amount for tuition that is typically the out-of-state tuition rate. In the case cited by the commenter, approximately 77 percent of the tuition amount the foreign veterinary school assesses its students is paid to the U.S school. Because foreign veterinary schools have no control over the tuition charged by U.S. schools that its students are required to attend, the commenter suggests that the Department allow foreign veterinary schools to exclude from total direct costs the portion of tuition that is charged by U.S schools.

Discussion: We do not agree that it would be appropriate to ignore loan debt that students incur for completing coursework provided by other institutions. For foreign veterinary schools and home institutions that enter into written arrangements under 34 CFR 668.5 to provide education and training, the veterinary school, or the home institution considers that coursework in determining whether to confer degrees or credentials to those students in the same way as if they provided the coursework themselves and the students are responsible for the debt accumulated for that coursework. Furthermore, in arranging for other institutions to provide coursework, the veterinary school or the home institution may be able to negotiate the cost of that coursework, but at the very least accepts those costs. For these reasons, we view the veterinary school or home institution as the party responsible for the loan debt students incur for completing coursework at other institutions.

Changes: None.

Amortization

Comments: Several commenters supported the Department’s proposal to amortize the median loan debt of students completing a GE program over 10, 15, or 20 years based on the credential level of the program, as opposed to a fixed amortization period of 10 years for all programs. These commenters believed that this amortization schedule more fairly accounts for longer and higher credentialed programs where students take out greater amounts of debt, better reflects actual student repayment patterns, and appropriately mirrors available loan repayment plans.

Some commenters supported the proposed amortization schedule based on credential level but suggested longer amortization periods than those proposed. For instance, some commenters recommended increasing the minimum amortization period from 10 years to 15 or 20 years.

One commenter suggested that we extend the amortization period from 10 years to 20 years because the commenter believed a 20‐year amortization schedule would more accurately reflect the actual time until full repayment for most borrowers. The commenter cited to the Department’s analysis in the NPRM that showed that within 10 years of entering repayment, about 58 percent of undergraduates at two‐year institutions, 54 percent of undergraduates at four‐year institutions, and 47 percent of graduate students had fully repaid their loans; within 15 years of entering repayment, about 74 percent of undergraduates at two‐year institutions, 76 percent of undergraduates at four‐year institutions, and 72 percent of graduate students had fully repaid their loans; and within 20 years of entering repayment, between 81 and 83 percent of students, depending on the cohort year, fully repaid their loans. The commenter also contended that far more bachelor’s degree programs would pass the D/E rates measure if we adopted a 20-year amortization period.

Other commenters agreed with using 10 years for certificate or diploma programs, but argued for extending the amortization period to 25 years for graduate, doctoral, and first professional degree programs. They asserted that students in graduate-level programs would likely have higher levels of debt that might take longer to repay. Some commenters were particularly concerned that some programs in high-debt, high-earnings fields would not be able to pass the D/E rates measure absent a longer amortization period. One commenter expressed concern that, even with a 20-year amortization period, medical programs, including those preparing doctors for military service and service in areas that have critical shortages of primary care physicians, would fail to pass the annual earnings rate despite successfully preparing their graduates for medical practice.

Other commenters advocated using a single 10-year amortization period regardless of the credential level. These commenters argued that a 10-year amortization period would best reflect borrower behavior, observing that most borrowers repay their loans under a standard 10-year repayment plan. The commenters referred to the Department’s analysis in the NPRM, which they believed showed that 54 percent of borrowers who entered repayment between 1993 and 2002 had repaid their loans within 10 years, and about 65 percent had repaid their loans within 12 years, despite economic downturns during that period. In view of this analysis, the commenters believed that the proposed 15- and 20-year amortization periods are too long and would allow excessive interest charges. These commenters also argued that longer repayment plans, like the income-based repayment plan, are intended to help struggling borrowers with unmanageable debts and should not become the expectation or standard for students repaying their loans. They asserted that the income-driven repayment plans result in considerably extending the repayment period, add interest cost to the borrower, and allow cancellation of amounts not paid at potential cost to taxpayers, the Government, and the borrower.

Discussion: Under these regulations, the Department determines the annual loan payment for a program, in part, by applying one of three different amortization periods based on the credential level of the program. As noted by some of the commenters, the amortization periods account for the typical outcome that borrowers who enroll in higher-credentialed programs (e.g., bachelor’s and graduate degree programs) are likely to have more loan debt than borrowers who enroll in lower-credentialed programs and, as a result, are more likely to take longer to repay their loans.

Based on our analysis of data on the repayment behavior of borrowers across all sectors who entered repayment between 1980 and 2011 that was provided in the NPRM, we continue to believe that 10 years for diploma, certificate, and associate degree programs, 15 years for bachelor’s and master’s degree programs, and 20 years for doctoral and first professional degree programs are appropriate amortization periods. We restate the relevant portions of our analysis here.

Of borrowers across all sectors who entered repayment between 1993 and 2002, we found that within 10 years of entering repayment, the majority of undergraduate borrowers, about 58 percent of borrowers from two-year institutions and 54 percent of undergraduate borrowers from four-year institutions, had fully repaid their loans. In comparison, less than a majority of graduate student borrowers had fully repaid their loans within 10 years. Within 15 years of entering repayment, a majority of all borrowers regardless of credential level had fully repaid their loans: about 74 percent of borrowers from two-year institutions, 76 percent of undergraduate borrowers from four-year institutions, and 72 percent of graduate student borrowers.¹³⁷

For more recent cohorts, the majority of borrowers from two-year institutions continue to fully repay their loans within 10 years. For example, of undergraduate borrowers from two-year institutions who entered repayment in 2002, 55 percent had fully repaid their loans by 2012. We believe this confirms that a 10-year amortization period is appropriate for diploma, certificate, and associate degree programs.

In contrast, recent cohorts of undergraduate borrowers from four-year institutions and graduate student borrowers are repaying their loans at slower rates than similar cohorts. Of borrowers who entered repayment in 2002, only 44 percent of undergraduate borrowers from four-year institutions and only 31 percent of graduate student borrowers had fully repaid their loans within 10 years. Even at this slower rate of repayment, given that 44 percent of undergraduate borrowers at four-year institutions fully repaid within 10 years, we believe it is reasonable to assume that the majority, or more than 50 percent, of borrowers from this cohort will reach full repayment by the 15-year mark. Accordingly, we believe that a 15-year amortization period is appropriate for bachelor’s degree programs and additionally master’s degree programs where students are likely to have less debt than longer graduate programs. Given the significantly slower repayment behavior of recent graduate student borrowers and the number of increased extended repayment periods available to borrowers, however, we do not expect the majority of these borrowers to fully repay their loans within 15 years as graduate student borrowers have in the past. But even at this slower rate of repayment, we believe it is likely that the majority of graduate student borrowers from this cohort will complete their repayment within 20 years. As a result, we see no reason to apply an amortization period longer than 20 years to doctoral and first professional degree programs.

We agree with the commenters who argued that the Department has made income-driven repayment plans available to borrowers who have a partial financial hardship only to assist them in managing their debt--and that programs should ideally lead to outcomes for students that enable them to manage their debt over the shortest period possible. As we noted in the preamble to the 2011 Prior Rule, an educational program generating large numbers of borrowers in financial distress raises troubling questions about the affordability of those debts. Moreover, the income-driven repayment plans offered by the Department do not provide for a set repayment schedule, as payment amounts are determined as a percentage of income. Accordingly, we have not relied on these plans for determining the amortization schedule used in calculating a program’s annual loan payment for the purpose of the D/E rates measure.

Changes: None.

Comments: One commenter suggested that instead of amortizing the median loan debt over specified timeframes, we should use the average of the actual annual loan amounts of the cohort that is evaluated. The commenter argued that by providing income-driven repayment plans, the Department acknowledges that recent graduates may not be paid well but need a way to repay their loans. As these graduates gain work experience, their earnings will increase. The commenter suggested that using the actual average of the cohort would allow for programs that provide training for occupations that require experience before earnings growth and motivate institutions to work with graduates who would be better off in an income-driven repayment plan than defaulting on their loans.

Discussion: We cannot adopt this suggestion because we do not have all the data needed to determine the actual annual loan amounts, particularly for students who received FFEL and Perkins Loans. But even if we had the data, adopting this suggestion would have the perverse effect of overstating the performance of a program where, absent adequate employment, many students who completed the program have to rely on the debt relief provided by income-driven repayment plans--an outcome that belies the purpose of these regulations.

Changes: None.

Interest Rate

Comments: Several commenters opposed the Department’s proposal to apply an interest rate that is the average of the annual interest rate on Federal Direct Unsubsidized Loans over the six-year period prior to the end of the cohort period. Some commenters asserted that a six-year average rate would inappropriately place greater emphasis on the predictability of the rate than on capturing the actual rates on borrowers’ loans. They argued that, particularly in the case of shorter programs, the six-year average interest rate might bear little resemblance to the actual interest rate that students received on their loans. One commenter stated that the average rate could obscure periods of high interest rates during which borrowers would still have to make loan payments. Referring to qualified mortgage rules that instruct lenders to assess an individual’s ability to repay using the highest interest rate a loan could reach in a five-year period, the commenter recommended that we likewise calculate the annual loan payment based on the highest interest rate during the six-year period.

Many commenters urged the Department to use an interest rate closer to the actual interest rate on borrowers’ loans. Specifically, commenters recommended calculating each student’s weighted average interest rate at the time of disbursement so that the interest rate applied for each program would be a weighted average of each student’s actual interest rate. However, acknowledging the potential burden and complexity of this approach, some commenters alternatively suggested varying the time period for determining the average interest rate by the length of the program. Although they suggested different means of implementing this approach (e.g., averaging the interest rate for the years in which the students in the cohort period received loans, or using the interest rates associated with the median length of time it took for students to complete the program), the commenters argued that determining an average interest rate based on the length of a program would provide more accurate calculations than using a six-year average interest rate for all GE programs. In particular, they believed that this approach would avoid situations in which a six-year average interest rate would be applied to a one-year certificate program, potentially applying an interest rate that would not reflect students’ repayment plans.

Some commenters suggested modifying proposed §668.404(b)(2)(ii) to add a separate interest rate for private education loans. These commenters argued that applying the average interest rate on Federal Direct Unsubsidized Loans to an amount that includes private loans would likely understate the amount of debt that a student incurred. They suggested that the Department could determine an appropriate interest rate to apply to private education loans by obtaining documentation of the actual interest rate for institutional loans and, for private education loans, surveying private student loan rates and using a rate based on that survey.

One commenter supported the Department’s proposal to use the average interest rate on Federal Direct Unsubsidized Loans during the six-year period prior to the end of the cohort period but suggested that the Department use the lower of the average or the current rate of interest on those loans. The commenter asserted that this approach would ensure that institutions are not penalized for economic factors they cannot control.

Finally, one commenter offered that Federal student loan interest rates, a significant predictor and influencer of borrowing costs, are now pegged to market rates and, as a result, exposed to rate fluctuations. Accordingly, different cohorts of students amassing similar levels of debt will likely see vastly different costs associated with their student loans depending upon when those loans were originated. This, the commenter suggests, will affect default rates and debt-to-earnings measurements, even if program quality and outcomes remain constant.

Discussion: We generally agree with the commenters that the interest rate used to calculate the annual loan payment should reflect as closely as possible the interest rates on the loans most commonly obtained by students. In particular, we agree that using the average interest rate over a six-year period for programs of all lengths might not accurately reflect the annual loan payment of students in shorter programs. However, we cannot adopt the suggestion made by some commenters to use the weighted average of the interest rates on loans at the time they were made or disbursed because we do not have the relevant information for every loan. However, we are revising §668.404(b)(2)(ii) to account for program length and the interest rate applicable to undergraduate and graduate programs. Specifically, for programs that are typically two years or less in length we will use the average interest rate over a shorter three-year “look-back” period, and use the longer six-year “look-back” period for programs over two years in length. In calculating the average interest rate for a graduate program, we will use the statutory interest rate on Federal Direct Unsubsidized loans applicable to graduate programs. Similarly, we will use the undergraduate interest rate on Federal Direct Unsubsidized loans for undergraduate programs. For example, for an 18-month certificate program, we will use the average of the rates for undergraduate loans that were in effect during the three-year period prior to the end of the cohort period.

Finally, we do not see a need to establish separate interest rates for private education loans. The Department does not collect, and does not have ready access to, data on private loan interest rates. The Department could calculate a private loan interest rate only if a party with knowledge of the rate on a loan were to report that data. The institution may be well aware that a student received a private education loan, but would not be likely to know the interest rate on that loan, and could not therefore be expected to provide that data to the Department. The Department could not readily calculate a rate from other sources because lenders offer private loans at differing rates depending on the creditworthiness of the applicant (and often the cosigner).¹³⁸ Although some lenders offer private loans for which interest rates are comparable to those on Federal Direct Loans, more commonly private loan interest rates are higher than rates on Federal loans; lenders often set rates based on LIBOR, but use differing margins to set those rates.¹³⁹ Thus, we could not determine from available data the terms of private loans obtained by a cohort of borrowers who enrolled in a particular GE program.

The CFPB rule to which the commenter refers does not appear to be relevant to the issue of the interest rate that should be used to calculate loan debt. The CFPB rule defines a “qualified mortgage” that is presumed to meet the ability to repay requirements as one “for which the ‘creditor’ underwrites the loan, taking into account the monthly payment for mortgage-related obligations, using: the maximum interest rate that may apply during the first five years after the date on which the first regular periodic payment will be due.” 12 CFR 1026.43(e)(2)(iv). Interest rates during the repayment period on title IV, HEA loans (FFELP and Direct Loans) made on or after July 1, 2006 have been fixed, rather than variable, and therefore the interest rate on a FFELP or Direct Loan made since 2006 remains fixed during the entire repayment term of the loan. 20 U.S.C. 1077A(i); 1087e(b)(7). Because these rates do not change, we see no need to adopt a rule that would cap interest rates for calculation of loan debt at a rate that would vary during the first five years of the repayment period.

Changes: We have revised §668.404(b)(2) to provide that the Secretary will calculate the annual loan payment for a program using the average of the annual statutory interest rates on Federal Direct Unsubsidized Loans that apply to loans for undergraduate and graduate programs and that were in effect during a three- or six-year period prior to the end of the cohort period.

Comments: One commenter expressed concern that independent, nonprofit, and for-profit institutions that do not charge interest as part of a student’s payment plan, either during the time the student is attending the institution or later after the student completes the program, would be discouraged from continuing this practice because the debt burden used to calculate the D/E rates would be overestimated. The commenter suggested that the Department either allow institutions to separate debt on interest-bearing accounts from debt on non-interest bearing accounts so the total loan debt and annual payment amounts are more accurate, or provide that institutions may appeal the loan debt calculation.

Discussion: The Department has crafted the D/E rates measure to assess programs based on the actual outcomes of students to the extent feasible. However, the Department has balanced this interest against the need for uniformity and consistency to minimize confusion and administrative burden. As there is no evidence that interest-free loans are a common practice, we do not believe the interest rate provisions of the regulations will significantly misstate debt burden if they do not specifically recognize interest-free institutional payment plans. Given the low chance of a materially unrepresentative result, simplicity and uniformity outweigh the commenter’s concerns.

Changes: None.

Comments: Some commenters disagreed with the Department’s proposal to apply the interest rate on Federal Direct Unsubsidized Loans, arguing that this approach would not account for whether students were undergraduate or graduate students, or for the percentage of students who received Subsidized Loans instead of Unsubsidized Loans. Some commenters also asserted that using the Unsubsidized Loan rate would artificially increase the annual loan payment amount used to calculate the D/E rates for a program.

Discussion: We will use the interest rate on Federal Direct Unsubsidized Loans to calculate the annual debt payment for the D/E rates measure for several reasons. First, the majority of students in GE programs who borrow take out Unsubsidized Loans. Second, the rate is one that will be used to calculate debt service on private education loans received by GE students, the most favorable of which are made at rates, available to only a small group of borrowers, that are comparable to the rate on Direct Plus loans (currently 7.21 percent).¹⁴⁰ Third, the rate we choose will be used to calculate debt service not on the entire loan, but, in every instance in which the loan amount is “capped” at tuition fees, books, equipment, and supplies, on a lesser amount. This tends to offset the results of a mismatch between the Unsubsidized Loan rate and a lower applicable loan rate.

Changes: None.

Bureau of Labor Statistics (BLS) Data

Comments: A number of commenters urged the Department to base the annual earnings component of the D/E rates on annualized earnings data from BLS, rather than on actual student earnings information from SSA. These commenters were concerned that the lack of access to SSA individual earnings data would hinder an institution’s ability to manage the performance of its programs under the D/E rates measure, and therefore advocated for using a publically available source of earnings data, such as BLS.

Other commenters who suggested using BLS data asserted that BLS data are more objective than income data from SSA because of the way that BLS aggregates and normalizes income information to smooth out anomalies.

Discussion: As we stated in the NPRM, we believe that there are significant difficulties with the use of BLS data as the basis for calculating annual earnings. First, as a national earnings data set that aggregates earnings information, BLS earnings data do not distinguish between graduates of excellent and low-performing programs offering similar credentials.

Second, BLS earnings data do not relate directly to a program. Rather, the data relate to a Standard Occupational Classification (SOC) code or a family of SOC codes based on the work performed and, in some cases, on the skills, education, or training needed to perform the work at a competent level. An institution may identify related SOC codes by using the BLS CIP-to-SOC crosswalk that lists the various SOC codes associated with a program, or the institution may identify through its placement or employment records the SOC codes for which students who complete a program find employment.

In either case, the BLS data may not reflect the academic content of the program, particularly for degree programs. Assuming the SOC codes can be properly identified, the institution could then attempt to associate the SOC codes to BLS earnings data. However, BLS provides earnings data at various percentiles (10, 25, 50, 75, and 90), and the percentile earnings do not relate in any way to the educational level or experience of the persons employed in the SOC code.

Accordingly, it would be difficult for an institution to determine the appropriate earnings for a program’s students, particularly for students who complete programs with the same CIP code but at different credential levels. For example, BLS data would not show a difference in earnings in the SOC codes associated with a certificate program and an associate degree program with the same CIP code.

Moreover, because BLS percentiles simply reflect the distribution of earnings of individuals employed in a SOC code, selecting the appropriate percentile is somewhat arbitrary. For example, the 10th percentile does not reflect entry-level earnings any more than the 50th percentile reflects earnings of persons employed for 10 years. Even if the institution could reasonably associate the earnings for each SOC code to a program, the earnings vary, sometimes significantly, between the associated SOC codes, so the earnings would need to be averaged or somehow weighted to derive an amount that could be used in the denominator for the D/E rates.

Finally, and perhaps most significantly, BLS earnings do not directly show the earnings of those students who complete a particular program at a particular institution. Making precisely such an assessment is essential to the GE outcome evaluation. Instead, BLS earnings reflect the earnings of workers in a particular occupation, without any relationship to what educational institutions those workers attended. While it is reasonable to use proxy earnings for research or consumer information purposes, we believe a direct measure of program performance must be used in determining whether a program remains eligible for title IV, HEA program funds. The aggregate earnings data we obtain from SSA will reflect the actual earnings of students who completed a program without the ambiguity and complexity inherent in using BLS data for a purpose outside of its intended scope.

Recognizing these shortcomings, in the 2011 Prior Rule, the Department permitted the use of BLS data as a source of earnings information only for challenges to debt-to-earnings ratios calculated in the first three years of the Department’s implementation of §668.7(g). This was done to address the concerns of institutions that they would be receiving earnings information for the first time on students who had already completed programs. In order to confirm the accuracy of the data used in a BLS-based alternate earnings calculation, §668.7(g) of the 2011 Prior Rule also required an institution to submit, at the Department’s request, extensive documentation, including employment and placement records.

We believe that the reasons for previously permitting the use of BLS data for a limited period of time, despite its shortcomings, no longer apply. Most institutions have now had experience with SSA earnings data, through the 2011 GE informational rates and 2012 GE informational rates; thus, for many programs, institutions are no longer in the situation where they would be receiving earnings data for the first time under the regulations.

Changes: None.

Debt Roll-Up

Undergraduate and Graduate Programs

Comments: Some commenters supported proposed §668.404(d)(2), under which the Department would attribute all undergraduate loan debt to the highest undergraduate credential that a student completed, and all graduate loan debt to the highest graduate credential that a student completed, when calculating the D/E rates for a program. They believed that this would address concerns raised by the 2011 Prior Rule that an institution’s graduate programs would be disadvantaged if a student pursued a graduate degree after completing an undergraduate program at the same institution. They explained that, under the 2011 Prior Rule, all of a student’s loan debt for an undergraduate program would have been attributed to the graduate program, which could have put the graduate program at a disadvantage and, as a result, might have deterred institutions from encouraging students to pursue further study. Although supportive of the Department’s proposal, one commenter suggested that the Department should go further by distinguishing between loan debt incurred for master’s and doctoral programs. The commenter argued that it is difficult to justify attributing debt from a shorter master’s program to a longer doctoral program and that institutions would be deterred from encouraging students to pursue doctoral-level study.

Another commenter believed that loan amounts should be attributed to a higher credentialed program only if the student was enrolled in a program in the same field. The commenter questioned the Department’s authority to use debt from two unrelated programs and attribute it to only one of them. The commenter opined that in some cases, students might enroll in one institution to earn an associate degree in a particular field, and then subsequently enroll in a higher credentialed program in a different field and may have to take additional coursework to fulfill the requirements of the second degree program. The commenter was concerned that the outcomes for these students would skew the D/E rates calculation for the higher credentialed program, resulting in inaccurate information for the public about the cost of completing the program.

Other commenters disagreed with the Department’s proposal to attribute a student’s loan debt to the highest credential subsequently completed by the student. These commenters believed that this approach would inflate and double-count loan debt of students who pursue multiple degrees at institutions because an institution would report and disclose debt at a lower credential level and then report the combined debt at a higher credential level. They were also concerned that attributing loan debt incurred for multiple programs to just the highest credentialed program would be confusing and misleading for prospective students and the public and would discourage students from enrolling in higher credentialed programs. The commenters recommended that the Department attribute loan debt and costs to each completed program separately instead of combining them.

Discussion: Although we appreciate the general support for our proposal to disaggregate the loan debt attributed to the highest credential completed at the undergraduate and graduate levels, we are not persuaded that further disaggregating loan debt between masters and doctoral-level programs is needed or warranted. As noted by some of the commenters, our proposal was intended to level the playing field between institutions that offer only graduate-level programs and institutions that offer both undergraduate and graduate programs. Without this distinction, the loan debt for students completing a program at a graduate program-only institution would be less than the loan debt for students who completed their undergraduate and graduate programs at the same institution because the student’s undergraduate loan debt would be attributed to the graduate-level program in the latter scenario.

Although we acknowledge that one student may take a different path than another student in achieving his or her educational objectives and that some coursework completed for a program may not be needed for, or transfer to, a higher-level program, we believe that the loan debt associated with all the coursework is part and parcel of the student’s experience at the institution in completing the higher-level program. Moreover, since the student’s earnings most likely stem from the highest credentialed program completed, we believe our approach will result in D/E rates that more closely tie the debt incurred by students for their training to the earnings that result from that training.

We note that the commenters’ description of how loan debt would be reported for students enrolled in a lower credentialed program who subsequently enroll in a higher credentialed program at the same institution is not entirely accurate. Though it is correct that loan debt from the lower credentialed program will be attributed to the completed higher credentialed program, the loan debt associated with that higher program prior to the amounts being “rolled-up” does not, as is suggested by the commenter, include loan debt from the lower credentialed program.

Changes: None.

Comments: One commenter asserted that students frequently withdraw from a higher credentialed program and subsequently complete a lower credentialed program at the same institution and was concerned that proposed §668.404(d)(2) would not adequately account for the total debt that a student has accumulated for both programs and must repay. Specifically, the commenter believed that a student’s loan debt from a higher credentialed program that the student did not complete would not be included in the D/E rates calculation for either that program or in the calculation for the lower credentialed program that the student completed. The commenter recommended that institutions be required to report the total debt that a student incurs while continuously enrolled, as well as the debt incurred in each program, for a more accurate picture of how much debt students have accumulated and their ability to repay their loans. The commenter also argued that this approach would provide an incentive for institutions to monitor students who are not meeting the academic requirements for a higher credentialed program and to counsel them on alternatives such as completing a lower credentialed program before they have taken on too much debt.

Discussion: The commenter is correct that the loan debt incurred for a higher credentialed program from which the student withdrew will not be attributed to a lower credentialed program that the student subsequently completed at the same institution. While we appreciate the commenter’s concerns, as we noted previously in this section, the loan debt associated with the student’s prior coursework at the institution is only counted if the student completes a higher-credentialed program because earnings most likely stem from that program. In this case, the only program completed is the lower credentialed program so only loan debt associated with that program is included in the D/E rates measure.

Changes: None.

Comments: One commenter requested that the Department clarify how loan debt incurred by a student for enrollment in a post-baccalaureate GE program, graduate certificate GE program, and graduate degree GE program would be attributed under proposed §668.404(d)(2)(ii) and (iii) and asked whether both of these provisions were needed.

Discussion: First, we note that loan debt incurred for enrollment in a post-baccalaureate program would be attributed to the highest credentialed undergraduate GE program subsequently completed by the student at the institution, rather than to the highest graduate GE program. This treatment is consistent with the definition of “credential level” in §668.402, which specifies that a post-baccalaureate certificate is an undergraduate program. Second, we agree with the commenter that the provisions in §668.404(d)(2)(ii) and (iii) are redundant.

Changes: We have removed §668.404(d)(2)(iii).

Common Ownership/Control

Comments: Some commenters warned that including loan debt incurred by a student for enrollment in programs at institutions under common ownership or control only at the Department’s discretion under proposed §668.404(d)(3) created a loophole. They believed that bad actors would exploit this loophole to manipulate the D/E rates for their programs by setting up affiliated institutions and encouraging students to transfer from one to the other. They were concerned that the Department would be unable or unwilling to apply loan debt incurred at an affiliated institution without specific criteria as to what would trigger a decision to include loan debt incurred at an affiliated institution in the D/E rates calculation for a particular program. To address this risk, these commenters recommended that the Department always include in a program’s D/E rates calculation loan debt that a student incurred for enrollment in a program of the same credential level and CIP code at another institution under common ownership or control, as proposed in the NPRM for gainful employment published in 2010. Short of this recommendation, they suggested that, at a minimum, the Department clarify the circumstances in which the Department would exercise its discretion in proposed §668.404(d)(3) to attribute loan debt from other institutions under common ownership or control.

Other commenters acknowledged the Department’s concern that some bad actors might try to manipulate the D/E rates calculations for their GE programs by encouraging students to transfer to affiliated institutions, but they did not believe that the Department should always attribute loan debt incurred at another institution under common ownership or control to the D/E rates calculation for the program. They suggested that institutions should not be held responsible for a student’s individual choice to move to an affiliated institution to pursue a more advanced degree simply because the institutions share a corporate ownership structure. They recommended that the Department specify that it would only attribute debt incurred at an institution under common ownership or control if the two institutions do not have separate accreditation or admission standards.

One commenter similarly requested clarification about the circumstances in which the Department would include loan debt incurred at another institution, but also suggested that the provision allowing the Department to include loan debt incurred at an institution under common ownership or control was unnecessary, given the proposed changes in §668.404(d)(2). They believed that requiring institutions to attribute loan debt to the highest credentialed program completed by the student provides adequate information on the outcomes of students at each institution.

Some commenters argued that the Department should never include loan debt that a student incurred at another institution, even if the institutions are under common ownership and control. One of these commenters argued that this provision would unfairly target for-profit institutions, noting that some public institutions, while not owned by the same corporate entity, are coordinated through a single State coordinating board or system tasked with developing system-wide policies. The commenter believed that the Department had not provided sufficient justification for treating proprietary institutions under common ownership or control differently from State systems with, in their view, parallel governance structures. Further, the commenter noted that institutions under common ownership or control might have different institutional missions and academic programs, and that it would therefore not be fair to attribute loan debt incurred for a program at one institution to a program at another.

Other commenters believed that it would be unfair to combine loan debt from institutions under common ownership or control, arguing that it could skew a program’s D/E rates. They were concerned that, in cases in which two students complete the same credential at the same institution, and one student goes on to complete a higher credential at an affiliated institution but the other completes a similar program at an unaffiliated institution, the D/E rates for the programs would not provide prospective students with a clear picture of the debt former students incurred to attend.

Discussion: We acknowledge the concerns of commenters who urged the Department to always include loan debt incurred at an affiliated institution in the D/E rates calculation for a particular program. We clarified in the NPRM that because this provision is included to ensure that institutions do not manipulate their D/E rates, it should only be applied in cases where there is evidence of such behavior. In such cases, the Secretary has the discretion to make adjustments. We believe this authority is adequate both to deter the type of abuse warned of by the commenters and act on instances of such abuse where necessary.

We remind those commenters who suggested that the Department should never include loan debt incurred at another institution, even if the institutions are under common control, that, except for loan debt associated with education and training provided by another institution under a written arrangement between institutions as discussed in “Tuition and Fees” in this section, we generally would not include loan debt from other institutions students previously attended, including institutions under common ownership or control.

We do not agree that this provision unfairly targets for-profit institutions subject to common ownership or control by not treating public institutions operating under the aegis of a State board or system in the same way. First, in the normal course of calculating D/E rates, programs at both types of institutions will be treated the same and the debts would not be combined. The debts would only be combined at institutions under common ownership and control in what we expect to be rare instances of the type of abuse described in this section. Second, since loan debt is “rolled-up” to the highest credentialed program completed by the student, any student who transferred into a degree program at a public institution would be enrolling in a program that is not a GE program, and therefore not subject to these regulations. The potential abuse is unlikely to arise when student debt from a certificate program at one institution would be rolled up to a certificate program that a student completed at another institution under the same ownership and control.

Changes: None.

Exclusions

Comments: Several commenters expressed concerns about the provisions in §668.404(e) under which the Department would exclude certain categories of students from the D/E rates calculation. Commenters argued that, because the Department would exclude students whose loans were in deferment, or who attended an institution, for as little as one day during the calendar year, institutions would not be held accountable for the outcomes of a significant number of students. Some commenters suggested that the Department should not exclude these students unless their loans were in a military-related deferment status for 60 consecutive days or they attended an eligible institution on at least a half-time basis for 60 consecutive days. The commenters cited as a basis for the 60 days the provisions for returning title IV, HEA program funds under §668.22 and reasoned that 60 percent of a three- to four-month term is about 60 days. In addition, they noted that to qualify for an in-school deferment, a student must be enrolled on at least a half-time basis and asserted that this provision provides a reasonable basis for excluding from the D/E rates calculation only students enrolled at least half-time.

Some commenters argued that students whose loans are in a military-related deferment status should not be excluded because these individuals made a valid career choice. The commenters also argued that because those students have military-based earnings, excluding them could have a significant impact on the earnings for the D/E rates calculations, as well as on the number of students included in the cohort. The commenters said that if the Department retains the military deferment exclusion, all individuals in military service should be excluded, based on appropriate evidence, not just those who applied for a deferment.

Some commenters supported the proposed exclusions, stating there is no evidence that supports establishing a time period or minimum number of days after which earnings should be excluded and that attempting to do so would be arbitrary and overly complex.

Discussion: While we appreciate the commenters’ recommendation that a student must attend an institution or have a loan in a military-deferment status for minimum number of days in the earnings year before these exclusions would apply, we do not believe there is a sound basis for designating any particular number of minimum days. Accordingly, we will apply the exclusions if a student was in either status for even one day out of the year.

We do not agree that the regulations regarding the return of title IV, HEA program funds provide a basis to set 60 days as the minimum. Students with military deferments or who are attending an institution during the earnings year are excluded from the D/E rates calculations because they could have less earnings than if they had chosen to work in the occupation for which they received training. The 60 percent standard in the regulations regarding the return of title IV, HEA program funds is unrelated to this rationale and, as a result, not applicable. With regard to the suggestion that a student must be enrolled on at least a half-time basis, we continue to believe that it is inappropriate to hold programs accountable for the earnings of students who pursue additional education because, regardless of course load, those students could have less earnings than if they chose to work in the occupation for they received training.

As previously discussed, the earnings of a student in the military could be less than if the student had chosen to work in the occupation for which they received training. Further, a student’s decision to enlist in the military is likely unrelated to whether a program prepares students for gainful employment. Accordingly, it would be unfair to assess a program’s performance based on the outcomes of such students. We believe that this interest in fairness outweighs any potential impact on the mean and median earnings calculations and number of students in the cohort period.

The military deferment exclusion would apply only to those individuals who have actually received a deferment. To the extent that borrowers serving in the military request such deferments, they are asking for assistance in the form of a period during which repayment of principal and interest is temporarily delayed. Borrowers who qualify for a military deferment, but do not request one, have made the determination that their income is sufficient to permit continued repayment of student loan debt while they are serving in the military. The Department confirms whether a borrower is enlisted in the military as part of the deferment approval process. Relying on this determination will be much more efficient and accurate than making individual determinations as to military status solely for the purposes of these regulations.

Changes: None.

Comments: Some commenters suggested that the Department exclude students who become temporarily disabled during the earnings year, opining that any earnings used for these students would distort the D/E rates. Other commenters suggested that a student with a loan deferment for a graduate fellowship or for economic hardship related to the student’s Peace Corps service at any point during the calendar year for which the Secretary obtains earnings information should be excluded from the D/E rates calculation. The commenters reasoned that graduate fellowships and Peace Corps service are competitive opportunities, and that only individuals who received a quality education would have been accepted. They concluded that a GE program’s D/E rates should not be affected by students who are accepted into these programs because their low wages would not be indicative of the quality of the program.

Discussion: As a general matter, we believe the additional exclusions mentioned by the commenters are rare and would not materially affect the D/E rates, so it would not be cost effective to establish reporting streams for gathering and verifying the information needed to apply these exclusions. We note that there are currently no deferments for students in the Peace Corp or who are temporarily disabled, but students with graduate fellowships may be excluded if they are attending an institution during the earnings year.

Changes: None.

Comments: Some commenters argued that students who are not employed for a portion of the earnings year should be excluded from the D/E rates calculation.

Discussion: We disagree that we should exclude from the D/E rates calculation students who are not employed for a portion of the earnings year. As discussed under “§668.405 Issuing and Challenging D/E Rates,” if graduates are unemployed during the earnings year, it is reasonable to attribute this outcome to the performance of the program, rather than to individual student choices.

Changes: None.

Comments: One commenter suggested that institutions should be provided access to Department databases to obtain the information necessary to determine whether students who complete a program satisfy any of the exclusion criteria.

Discussion: If a student has attended a particular institution, that institution already has access to NSLDS information for the student. In addition, the data provided to institutions with the list of students who completed the program will have information on which students were excluded from the calculation and which exclusions were applied. If an institution has evidence that the data in NSLDS are incorrect, it may challenge that information under the procedures in §§668.405 and 668.413.

Changes: None.

N-Size

Comments: Several commenters recommended that the Department use a minimum n-size of 10 students, instead of 30, when calculating the D/E rates. The commenters argued that an n-size of 30 is unnecessarily large in view of the Department’s analysis in the NPRM showing that an n-size of 10 adequately provides validity, and that there would be only a small chance that a program would erroneously be considered to not pass the D/E rates measure. One of these commenters expressed concern that increasing the n-size from 10 to 30 would leave unprotected many students enrolled in GE programs and did not believe this was sufficiently emphasized in the NPRM. Specifically, the commenter pointed to analysis in the NPRM showing that, using an n-size of 30, more than one million students would enroll in GE programs that would not be evaluated under any of the proposed accountability metrics.

Another commenter similarly urged the Department to select the smallest n-size needed for student privacy and statistical validity, and design the final regulations so that programs that capture the vast majority of career education program enrollment are assessed under the accountability metrics. The commenter was concerned in particular that the provision in the NPRM to disaggregate undergraduate certificates into three credential levels based on their length would result in many programs falling below the minimum n-size of 30 and therefore not being evaluated under these regulations.

One commenter contended that the Department’s statistical analysis showed that the probability of a program that is near failing actually losing eligibility under the regulations is 1.4 percent. The commenter argued that, because this probability was only for programs on the margin, the chance that a randomly chosen program could lose eligibility when it was actually passing approached zero. The commenter believed that an n-size of 30 would be a weaker standard and that the data demonstrated accuracy of the metrics at an n-size of 10. As a result, the commenter concluded that there is little justification for an n-size of 30 and allowing hundreds of failing programs to remain eligible for title IV, HEA program funds.

Other commenters also believed that the larger n-size would allow some failing programs to pass the accountability metrics. One of these commenters cited the Department’s analysis, which stated that using an n-size of 10 will cover 75 percent of all students enrolled in GE programs while using an n-size of 30 would only cover 60 percent of students enrolled in GE programs. The commenter said that by moving to a larger n-size, the Department estimates that over 300 programs that would fail the D/E rates measure would no longer be held accountable and that an additional 439 programs in the “zone” would not be subject to the D/E rates measure. The commenter concluded that the larger n-size creates a loophole that will allow hundreds of failing programs to continue to receive title IV, HEA program funds. Other commenters similarly concluded that an n-size of 30 creates a loophole where institutions would have the ability to adjust their program size to evade the regulations.

On the other hand, several commenters supported the Department’s proposal to use a minimum n-size of 30. These commenters stated that the substantial majority of students in GE programs would be captured using this n-size. These commenters believed that an n-size of 10 is too small and not statistically significant, and that with an n-size of 10, the results of a small number of students would sway outcomes from year to year and outcomes would be more sensitive to economic fluctuations. The commenters asserted that when compared with outcomes under an n-size of 10, outcomes under an n-size of 30 will always have a lower standard error and are therefore likely to lead to more accurate results. The commenters argued that a larger sample size will have less variability and yield more reliable results than a smaller one taken from the same population. One commenter referred to Roscoe, J.T., Fundamental Research Statistics for the Behavioral Sciences, 1975, which, according to the commenter, cites as a rule of thumb that sample sizes larger than 30 and less than 500 are appropriate for most research. The commenter suggested that the Department’s analysis showed that the average probability that a passing program would be mischaracterized as a zone program in a single year drops from 6.7 percent to 2.7 percent when the n-size changes from 10 to 30.

Another commenter argued that a minimum n-size of 10 increases the potential that a particular student in a cohort could be identified, putting student privacy at risk. Other commenters also asserted that an n-size of 10 might result in the disclosure of individually identifiable information, especially at the extremes of high and low earners.

One commenter believed that volatility resulting from too small of a sample size would create uncertainty that would chill efforts to launch new programs.

Discussion: We believe that an n-size of 30 strikes an appropriate balance between accurately measuring D/E rates for each program and applying the accountability metric to as many gainful employment programs as possible. Although a number of commenters supported our proposal to use an n-size of 30, in general we do not agree with their reasoning for doing so.

We disagree that mitigating the impact of economic fluctuations on D/E rates provides a direct rationale for choosing a higher minimum n-size. The Department has not found any evidence that D/E rates for smaller programs are more sensitive to economic fluctuations than larger programs. N-size affects the variability of D/E rates from year to year due to statistically random differences in the D/E rates of individual students. The greater the n-size, the less these year-to-year differences will affect measures of central tendency, such as those used to calculate the D/E rates. As discussed in “§668.403 Gainful Employment Program Framework,” we believe the impact of economic fluctuations on program performance is mitigated because programs must fall in the zone for four consecutive years before becoming ineligible for title IV, HEA program funds. We also include multiple years of debt and earnings data in our D/E rates calculation to smooth out fluctuations in the economic business cycle, along with fluctuations in the local labor market.

We also disagree that a minimum n-size of 30 is preferable to an n-size of 10 in order to minimize year-to-year fluctuations, per se. A program’s D/E rates may change from year to year due to changes in educational quality provided to students, prices charged by the institution, or other factors. These fluctuations are likely to occur regardless of n-size and we view them as accurate indications of changes in programmatic performance under the D/E rates measure.

We further disagree that a minimum n-size of 30 is necessary to protect the privacy of students. Based on NCES standards, an n-size of 10 is sufficient to protect the privacy of students on measures of central tendency such as the D/E rates measure.

Finally, we disagree that our data analysis indicates that a D/E rates measure with a minimum n-size of 10 is statistically unreliable. Our analysis indicates that the probability of mischaracterizing a program as zone or failing due to statistical imprecision when the n-size is 10 is 6.7 percent. By most generally accepted statistical standards, this probability of mischaracterization is modest. For this reason that we believe a minimum n-size of 10 produces D/E rates, and additionally median loan debt and mean and median earnings calculations, sufficiently precise for disclosure.

As discussed in the NPRM, we believe a minimum n-size of 30 is a more appropriate threshold for the D/E rates measure when it is used as an accountability metric--not because it would be invalid at a minimum n-size of 10, but because even slight statistical imprecision could lead to mischaracterizing a program as zone or failing which would precipitate substantial negative consequences, such as requiring programs to warn students they could lose eligibility for title IV, HEA program funds. Given these consequences, we believe it is more appropriate to set the minimum n-size at 30 for accountability determinations.

So, even though an n-size of 10 would provide a sufficiently precise measure of D/E rates, our analysis shows an n-size of 30 is more appropriate because it reduces the possibility of mischaracterizing a program as zone or failing in a single year. It also reduces the possibility of a program becoming ineligible as a result of multiple mischaracterizations over time.

As provided in the NPRM, if the minimum number of students completing a program necessary to calculate the program’s D/E rates is set at 30, the expected or average probability that a passing program would be mischaracterized as a zone program in a single year is no more than 2.7 percent. Because this is an average across all programs with passing D/E rates, the probability is lower the farther a program is from the passing threshold and higher for programs with D/E rates closer to the passing threshold. At an n-size of 10, the probability that a passing program would be mischaracterized as a zone program in a single year would be no more than 6.7 percent.

Although the difference in the precision of the D/E rates with n-sizes of 10 and 30, respectively, may seem modest, there are substantial benefits in reducing the probability of mischaracterization of being in the zone from 6.7 percent to 2.7 percent. While a program will not lose eligibility if it is mischaracterized in the zone for a single year, it will face some negative consequences because the institution could lose eligibility for title IV, HEA program funds within four years. Further, the program’s D/E rates will be published by the Department and potentially subject to disclosure by the institution.

Additionally, there are benefits to ensuring that the probability of a passing program being mischaracterized as a failing program in a single year is close to zero. At an n-size of 10, the probability is as high as 0.7 percent, while at an n-size of 30 it is close to 0 percent. By setting the n-size at 30, it is a virtual certainty that passing programs will not mischaracterized as failing the D/E rates measure due to statistical imprecision. In this case, reducing imprecision is particularly important because programs would be required to warn students they could lose eligibility as soon as the next year for which D/E rates are calculated.

In addition to reducing the probability of single-year mischaracterizations, it is appropriate to set an n-size of 30 to reduce the probability of a passing program losing eligibility due to statistical imprecision and anomalies. Because the consequences are substantial, it is important we set the minimum n-size at 30 in order to reduce the probability of statistical mischaracterization to near zero. As stated in the NPRM, because no program would be found ineligible after just a single year, it is important to look at the statistical precision analysis across multiple years. These probabilities drop significantly for both an n-size of 30 and 10 when looking across the four years that a program could be in the zone before being determined ineligible. The average probability of a passing program becoming ineligible as a result of being mischaracterized as a zone program for four consecutive years at an n-size of 30 is close to 0 percent. At an n-size of 10, the average probability is as high as 1.4 percent. Although we are unable to provide precise probabilities for the scenario in which a program fails the D/E rates measure in two out of three years due to limitations in our data, our analysis indicates the probability of a passing program becoming ineligible due to failing the D/E rates measure two out of three years could be as high as 0.7 percent with a minimum n-size of 10.¹⁴¹ In contrast, the probability of mischaracterization due to failing the D/E rates measure in two out of three years is close to zero percent with a minimum n-size of 30.

Although setting a minimum n-size of 30 reduces the percentage of programs that are evaluated by the D/E rates measure, which may result in more programs with high D/E rates remaining eligible than with a minimum n-size of 10, we believe the consequences of mischaracterizing programs due to statistical imprecision outweighs this concern.

We also do not believe that the possibility of increased “churn” due to programs attempting to decrease the number of students who complete a program to below 30 outweighs the benefits of greater statistical precision. First, if the minimum n-size is 10, it is unclear that we would reduce the possibility of “churn.” Programs, particularly programs near an n-size of 10, could still attempt to lower the number of students completing the program to avoid being evaluated. Second, we have included several provisions in the regulations to discourage programs from increasing non-completion among students. As discussed in “§668.403 Gainful Employment Program Framework,” among the items institutions may be required to disclose are completion rates and pCDR, which will provide prospective students with information to avoid enrollment in high “churn” programs.

Changes: None.

Comments: One commenter noted it is difficult to evaluate the impact of the n-size provision of the regulations because the Department changed how it defines a program by proposing to break out undergraduate certificates into three credential levels based on program length.

Discussion: As noted previously, we are no longer classifying certificate programs based on program length.

Changes: None.

Transition Period

Comments: A number of commenters expressed concern that the proposed transition period would not provide sufficient time for programs to improve after the regulations go into effect. Specifically, commenters questioned whether an institution would be able to improve a program’s D/E rates in the years following an initial failure, because the students included in calculating the D/E rates for the first several years will have already graduated from the program. These commenters asserted that, as a result, it will be too late for institutions to improve program performance through changing the program’s admissions standards or improving financial literacy training, debt counseling, and job placement services. One of these commenters contended that the data that will be used to calculate D/E rates in 2015 is already fixed and cannot be affected by any current program improvement efforts.

Another commenter asserted that the Department’s proposal to consider only the debt of students graduating in the current award year during the transition period would not adequately address the challenge faced by programs longer than one year because, regardless of any recent reduction in program cost, students’ debt loads would initially be affected by debt undertaken to support earlier, potentially more costly, years in the program. Consequently, institutions would find it very difficult to improve program outcomes for longer programs during the transition period.

One commenter suggested that the Department defer the effective date of the regulations and revise the transition period so that institutions could affect the borrowing levels for all students in a cohort period throughout their period of enrollment before the program would be evaluated under the D/E rates.

One commenter contended that SSA earnings data would not be released until 2016 when the first D/E rates are issued. This commenter suggested eliminating the transition period in favor of four years of informational rates. Another commenter suggested there should be two years of informational rates before sanctions begin.

Some commenters proposed limiting the impact of the regulations during the transition period by reinstituting a cap on the number of programs that could become ineligible in the early years of implementation in order to give failing programs another year to improve. Several commenters recommended including the five percent cap on ineligible programs that was included in the 2011 Prior Rule.

Some commenters stated that the proposed transition period was better than the five percent cap in the 2011 Prior Rule, but were skeptical that institutions would use the transition period to make changes to poorly performing programs. Instead, they argued that institutions will give scholarships or tuition discounts to students completing programs, which would result in improved D/E rates but not lower tuition for all students.

Discussion: In view of the comments that the proposed four-year transition period did not provide sufficient time for programs to improve, we are extending the transition period. As illustrated in the following chart, the transition period is now five years for programs that are one year or less, six years for programs that are between one and two years, and seven years for programs that are longer than two years.

Award year for which the D/E rates are calculated	2014-2015	2015-2016	2016-2017	2017-2018	2018-2019	2019-2020	2020-2021	2021-2022
Two-year cohort	2010-2011 & 2011-2012	2011-2012 & 2012-2013	2012-2013 & 2013-2014	2013-2014 & 2014-2015	2014-2015 & 2015-2016	2015-2016 & 2016-2017	2016-2017 & 2017-2018	2017-2018 & 2018-2019
Transition year	1	2	3	4	5	6	7	8
Programs less than one year	2014-2015	2015-2016	2016-2017	2017-2018	2018-2019	2015-2016 & 2016-2017
Programs between one and two years	2014-2015	2015-2016	2016-2017	2017-2018	2018-2019	2019-2020	2016-2017 & 2017-2018
Programs more than two years	2014-2015	2015-2016	2016-2017	2017-2018	2018-2019	2019-2020	2020-2021	2017-2018 & 2018-2019

For a GE program that is failing or in the zone for any award year during the transition period, in addition to calculating the regular D/E rates the Department will calculate alternate, or transitional, D/E rates using the median loan debt of the students who completed the program during the most recently completed award year instead of the median loan debt for the two-year cohort. For example, as shown in the chart, in calculating the transitional D/E rates for the 2014-2015 award year, we will use the median loan debt of the students who completed the program during the 2014-2015 award year instead of the median loan debt of the students who completed the program in award years 2010-2011 and 2011-2012. For programs that are less than one year, we will calculate transitional D/E rates for five award years--2014-2015 through 2018-2019. After the transitional D/E rates are calculated for those award years, the transition period expires and the Department uses only the median loan debt of the students in the cohort period to calculate the D/E rates for subsequent award years. The first D/E rates the Department will calculate after the transition period will be for award year 2019-2020. As shown in the chart, the two-year cohort period for that award year includes the students who completed the program during the 2015-2016 and 2016-2017 award years. So, for programs that are less than one year in length, the five-year transition period ensures that most of the students in the two-year cohort period began those programs after these final regulations are published. We applied the same logic in determining the transition periods for programs that are between one and two years, and for programs that are over two years long. Consequently, institutions will be able to make immediate reductions in the loan debt of students enrolled in its GE programs, and those reductions will be reflected in the transitional D/E rates.

We note that the transitional D/E rates would operate in conjunction with the zone to allow institutions to make improvements to their programs in the initial years after the regulations go into effect in order to pass the D/E rates measure. That is, an institution with a program in the zone will have four years to lower loan debt in an effort to achieve passing results for that program. For a failing program, an institution that lowers loan debt sufficiently at the outset of the transition period could move the program into the zone and thereby avoid losing eligibility. The institution would then have additional transition and zone years to continue to improve the program. Moreover, because the Department will provide the regular D/E rates to institutions during the transition period, institutions will be able to gauge the amount of the loan reduction needed for their programs to pass the D/E rates measure once the transition period concludes.

The transition period runs from the first year for which we issue D/E rates under these regulations. The length of the transition period is determined by the length of the program and the number of years we have issued D/E rates under this subpart--not the number of years that we have issued D/E rates for the particular GE program. We may not issue D/E rates for a particular GE program for a particular year for several reasons, such as insufficient n-size, but each year we issue any D/E rates for the regulations is included in any transition period whether or not we issued D/E rates for a specific program in a given year.

We believe that extending the number of years that the transition period will remain in effect is not only responsive to concerns raised by the commenters about the time that institutions need to improve program performance but that doing so will result in tangible benefits for students.

We believe that this option better serves the purposes of the regulations than the provision in the 2011 Prior Rule setting a cap on the number of programs that could be determined ineligible. The cap afforded institutions an opportunity to avoid a loss of eligibility without taking any action to improve their programs. The transition period provisions in these regulations provide institutions an incentive to improve student outcomes as well as an opportunity to avoid ineligibility.

We do not agree that delaying implementation of the regulations or providing informational rates for a set period of time before imposing consequences will be as effective as the revised transition period. The purpose of the transition period is to provide institutions with an incentive to make improvements in their programs so that students will see improved outcomes. Delaying implementation or only providing informational rates the first few years the regulations are in effect would likely create a disincentive for programs to make improvements, which in turn would negatively affect students.

With the changes we are making in these final regulations, we believe that institutions will have a significant incentive to make improvements. It is possible that an institution may also seek to improve its D/E rates by giving scholarships or tuition discounts to students completing the program. A scholarship or tuition discount benefits the student by reducing debt burden, and therefore we would not discourage an institution from offering that type of benefit to its students.

Changes: We have revised the regulations in §668.404(g) to provide that the transition period is five award years for a program that is one year or less in length; six award years for a program that is between one and two years in length; and seven award years for a program that is more than two years in length.

90/10 Rule

Comments: Several commenters argued that the proposed definition of “gainful employment,” as reflected in the D/E rates measure, conflicts with the 90/10 provisions in section 487(a)(24) of the HEA, under which for-profit institutions must derive at least 10 percent of their revenue from sources other than the title IV, HEA programs.

Some of these commenters opined that the regulations would limit the ability of for-profit institutions to increase tuition since increases in tuition correlate strongly with increases in Federal and private student loan debt. The commenters stated that increasing tuition beyond the total amount of Federal student aid available to students is the principal means available to for-profit institutions for complying with the 90/10 provisions. Consequently, the commenters reasoned that it would be extremely difficult for institutions to comply with both the GE regulations and the 90/10 provisions, particularly for institutions that are at or near the 90 percent limit, that enroll predominately students who are eligible for Pell Grants, or that are located in States where grant aid is not available to for-profit institutions. One of these commenters asked the Department to refrain from publishing any final regulations addressing student debt until the Department works with Congress to modify the 90/10 provisions to address this conflict.

Other commenters contended that the proposed regulations are contrary to the 90/10 provisions because as tuition decreases, the chances increase that institutions will not be able to comply with the 90/10 provisions because the percentage of tuition that students pay with title IV, HEA program funds will remain constant or increase. Some commenters concluded that as institutions attempt to balance the requirements of these regulations with their 90/10 obligations, opportunities for students who rely heavily on title IV, HEA program funds will be curtailed, particularly because the Department interprets the HEA to prohibit institutions from limiting the amount students may borrow on an across-the-board or categorical basis.

Other commenters argued that if one of the objectives of these regulations is to reduce tuition (and by implication, student loan debt), this objective conflicts directly with the 90/10 provisions, which often lead to tuition increases resulting from mathematical expediency. The commenters stated that because institutions are prohibited from capping the amount students may borrow, but are effectively given incentives to maintain tuition at amounts higher than the Federal loan limits, these regulations would place institutions at risk of violating the 90/10 provisions.

Similarly, other commenters stated that for-profit institutions are often prevented from reducing tuition because they must satisfy the 90/10 provisions and because they are prohibited from reducing borrowing limits for students in certain programs. The commenters suggested that the Department use its Experimental Sites authority as a way to develop a better approach for making programs more affordable. Specifically, the commenters proposed that institutions participating in an approved experiment could be exempt from the 90/10 provisions in order to reduce the cost of a program to a level aligned with the cost of delivering that program and the expected wages of program graduates. The commenters offered that under this approach, an institution could be required to submit a comprehensive enrollment management and student success plan and annual tuition increases would be indexed to annual rates of inflation. Or, at a minimum, the commenters suggested that the Department exempt institutions that would otherwise fail the 90/10 revenue requirement by lowering tuition amounts to pass the D/E rates measure. In addition, the commenters offered other suggestions, such as exempting from the 90/10 provisions institutions that serve a majority of students who are eligible for Pell Grants or, instead of imposing sanctions on programs that fail the D/E rates measure, using the D/E rates calculations to set borrowing limits in advance to prevent students from taking on too much loan debt.

Another commenter believed that if the 90/10 provisions were eliminated, there would be no need for the D/E rates measure.  The commenter opined that the 90/10 provisions place constraints on market forces that, absent these provisions, would lead to reductions in tuition at for-profit institutions, shorten vocational training, reduce student indebtedness, and eliminate the need for funding above the Federal limits.

Discussion: The 90/10 provisions are statutory and beyond the scope of these regulations. However, we are not persuaded that the 90/10 provisions conflict with the D/E rates measure. In a report published in October 2010,¹⁴² GAO did not find any relationship between an institution’s tuition rate and its likelihood of having a very high 90/10 rate. GAO’s regression analysis of 2008 data indicated that schools that were (1) large, (2) specialized in healthcare, or (3) did not grant academic degrees were more likely to have 90/10 rates above 85 percent when controlling for other characteristics. Other characteristics associated with higher than average 90/10 rates included (1) high proportions of low-income students, (2) offering distance education, (3) having a publicly traded parent company, and (4) being part of a corporate chain. GAO defined “very high” as a rate between 85 and 90 percent, and about 15 percent of the for-profit institutions were in this range. GAO found that, in general, there was no correlation between an institution’s tuition rate and its average 90/10 rate. In one exception, GAO found that institutions with tuition rates that did not exceed the 2008-2009 Pell Grant and Stafford Loan award limits (the award amounts were for first-year dependent undergraduates) had slightly higher average 90/10 rates than other institutions, at 68 percent versus 66 percent.

The Department’s most recent data on 90/10, submitted to Congress in September 2014,¹⁴³ show that only 27 of 1948 institutions had ratios over 90 percent, and that about 21 percent had ratios in the very high range of 85 to 90 percent. The GAO report and the Department’s data suggest that most institutions could reduce tuition costs without the consequences envisioned by the commenters.

Several other factors also suggest that any tension between the 90/10 provisions and the GE regulations can be managed by most institutions. First, some of the 90/10 provisions that are not directly tied to the title IV, HEA program funds received to pay institutional charges for eligible programs, such as allowing an institution to count income from programs that are not eligible for title IV, HEA program funds, count revenue from activities that are necessary for the education and training of students, or count as revenue payments made by students on institutional loans, make it easier for institutions to comply with the 90/10 provisions. Second, institutions have opportunities to recruit students that have all or a portion of their costs paid from other sources. In addition, as a result of the changes to the HEA in 2008, an institution may fail the 90/10 revenue requirement for one year without losing eligibility, and the institution can retain its eligibility so long as it does not fail the 90/10 revenue requirement for two consecutive years. Furthermore, institutions that have students who receive title IV, HEA program funds to pay for non-tuition costs, such as living expenses, are already in the situation described by the commenters in which the amount of title IV, HEA program funds may exceed institutional costs. These institutions are presumably managing their 90/10 ratios using a combination of other resources, and this result would also be consistent with the GAO report.

We appreciate the suggestions made by some of the commenters that we use our authority under Experimental Sites to exempt from the 90/10 provisions institutions that would make programs more affordable. At this time, however, we are not prepared to establish experiments that could test whether exemptions from the 90/10 provisions would lead to reductions in program costs but will take the suggestion under consideration.

Changes: None.

Comments: Some commenters stated that it is unfair that the 90/10 requirements ostensibly encourage institutions to recruit students who can pay cash but the D/E rates measure would not take into account cash payments made by those students.

Discussion: We do not agree that the D/E rates measure disregards out-of-pocket payments made by students. Students who pay for some tuition costs out of pocket may have lower amounts of debt, which may be reflected in the calculation of median loan debt for the D/E rates measure.

Changes: None.

Comments: Some commenters believed that allowing G.I. Bill and military tuition assistance to be counted as non-Federal revenue creates a loophole that some for-profit institutions exploit to comply with the 90/10 requirements by using deceptive and aggressive marketing practices to enroll veterans and service members. The commenters stated that the GE regulations would help to protect veterans and service members by eliminating poorly performing programs that would otherwise waste veterans’ military benefits and put them further into debt.

Discussion: Section 487(a)(24) of the HEA directs that only “funds provided under this title [title IV] of the HEA” are included in the 90 percent limit. 20 U.S.C. 1094(a)(24). Other Federal assistance is not included in that term. We agree that these regulations are designed and are expected to protect all students, including veterans and service members, from poorly performing programs that lead to unmanageable debt.

Changes: None.

Effect of the Affordable Care Act

Comments: Some commenters believed that the Affordable Care Act has caused some employers to limit new employees to less than 30 hours of work per week to avoid having to provide health insurance benefits. These commenters were concerned that, as a result, institutions with programs in fields where most employees are paid by the hour would be unfairly penalized for these unintended consequences of the law because students who completed their program might be unable to find full-time positions.

Discussion: Employers often change their hiring practices and wages paid to account for changes in the workforce and market demand for certain jobs and occupations. In these circumstances, we expect that institutions will make the changes needed for their programs to pass the D/E rates measure.

Changes: None.

§668.405 Issuing and Challenging D/E Rates

Comments: Several commenters asked the Department to clarify, and specify in the regulations, what would constitute a “match” with the SSA earnings data and how “zero earnings” are treated for the purpose of calculating the D/E rates.

Discussion: Using the information that an institution reports to the Secretary under §668.411, the Department will create a list of students who completed a GE program during the cohort period. For every GE program, the list identifies each student by name, Social Security Number (SSN), date of birth, and the program the student completed during the cohort period. After providing an opportunity for the institution to make any corrections to the list of students, or information about those students, the Department submits the list to SSA. SSA first compares the SSN, name, and date of birth of each individual on the list with corresponding data in its SSN database, Numident. SSA uses an Enumeration Verification System to compare the SSN, name, and date of birth as listed by the Department for each individual on its list against those same data elements recorded in Numident for SSN recipients. A match occurs when the name, SSN, and date of birth of a student as stated on the Department’s list is the same as a name, SSN, and date of birth recorded in Numident for an individual for whom an SSN was applied. SSA then tallies the number of individuals whose Department-supplied identifying data matches the data in Numident. The system also identifies SSNs for which a death has been recorded, which will be considered to be “unverified SSNs” for purposes of this calculation. Unverified SSNs will be excluded from the group of matched individuals, or “verified SSNs,” and therefore no earnings match will be conducted for those SSNs. If the number of verified SSNs is fewer than 10, SSA will not conduct any match against its earnings records, and will notify the Department. As noted in the NPRM, the incidence of non-matches has proven to be very small, less than two percent, and we expect that experience to continue.

If the number of verified SSNs is 10 or more, SSA will then compare those verified SSNs with earnings records in its Master Earnings File (MEF). The MEF, as explained later in this section, is an SSA database that includes earnings reported by employers to SSA, and also by self-employed individuals to the Internal Revenue Service (IRS), which are in turn relayed to SSA. SSA then totals the earnings reported for these SSNs and reports to the Department the mean and median earnings for that group of students, the number of verified individuals and the number of unverified individuals in the group, the number of instances of zero earnings for the group, and the earnings year for which data is provided. SSA does not provide to the Department any individual earnings data or the identity of students who were or were not matched. Where SSA identifies zero earnings recorded for the earnings year for a verified individual, SSA includes that value in aggregate earnings data from which it calculates the mean and median earnings that it provide to the Department, and we use those mean and median earnings to calculate the earnings for a program. As reflected in changes to §668.404(e), we do not issue D/E rates for a program if the number of verified matches is fewer than 30. If the number of verified matches is fewer than 30 but at least 10, we provide the mean and medium earnings data to the institution for disclosure purposes under §668.412.

This exchange of information with SSA and the process by which SSA matches the list of students with its records is conducted pursuant to one or more agreements with SSA. The agreements contain extensive descriptions of the activities required of the two agencies, and those terms may be modified as the agencies determine that changes may be desirable to implement the standards in these regulations. The Department engages in a variety of data matches with other agencies, including SSA, and does not include in pertinent regulations either the agreements under which these matches are conducted, or the operational details included in those agreements, and is not doing so here. The agreements are available to any requesting individual under the Freedom of Information Act, and commenters have already obtained and commented on their terms in the course of providing comments on these regulations.

Changes: We have revised §668.405(e) to clarify that the Secretary does not calculate D/E rates if the SSA earnings data returned to the Department includes reports for records of earnings on fewer than 30 students.

Comments: Several commenters criticized the Department’s reliance on SSA earnings data in calculating the earnings of students who complete a GE program on several grounds. The commenters contended that SSA data are not a reliable source for earnings because the SSA database from which earnings data will be derived--the MEF--does not contain earnings of those State and local government employees who are employed by entities that do not have coverage agreements with SSA.

Discussion: We think there may be some confusion regarding the data contained in the SSA MEF and used by SSA to compute the aggregate mean and median earnings data provided to the Department and used by the Department to calculate D/E rates, and in particular the reporting and retention of earnings of public employees. As explained by SSA¹⁴⁴:

The Consolidated Omnibus Budget Reconciliation Act of 1985 (COBRA) imposed mandatory Medicare-only coverage on State and Local employees.  All employees, with certain exceptions, hired after March 31 1986, are covered for Medicare under section 210(p) of the Act (Medicare Qualified Government Employment).  Employees covered for Social Security under a Section 218 Agreement have Medicare coverage as a part of Social Security, therefore they are excluded from mandatory Medicare.  However, COBRA 85 also contained a provision allowing States to obtain Medicare-only coverage for employees hired before April 1, 1986 who are not covered under an Agreement.  Authority for Medicare-only tax administration was placed in the Code [26 U.S.C. 3121(u)(2)(C)] as the responsibility of IRS.

Regardless of whether State and local government employees participate in a State retirement system or are covered or not covered by Section 218, all earnings of public employees are included in SSA’s MEF and included in the aggregate earnings data set provided to the Department. In addition, earnings from military members are included in the MEF.

Changes: None.

Comments: Commenters contended that the earnings in the MEF are understated because the amount recorded in the MEF is capped at a set figure ($113,700 in 2013), and that earnings accurately reported but exceeding that amount are disregarded and not included in the aggregate earnings data set provided to the Department by SSA.

Discussion: The commenter is incorrect. Total earnings are included in MEF records without limitation to capped earnings. As explained in greater detail below, SSA uses total earnings for the matched individuals to create the aggregate data set provided to the Department.

Changes: None.

Comments: Commenters contended that other earnings are not reported to SSA and retained in the MEF, including deferred compensation. Commenters claimed that aggregate earnings does not include earnings contributed to dependent care or health savings accounts, and therefore aggregate earnings data reported by SSA to the Department understate the earnings of students who completed programs. Commenters also asserted that reported earnings would not include such compensation as deductions for deferred earnings and 401(k) plans and similarly understate earnings. Commenters stated that an individual’s SSA earnings do not include sources of income such as lottery winnings, child support payments, or spousal income.

Discussion: Other earnings of the wage earner, such as deferred compensation, must be reported, are included in the MEF, and are used to create the aggregate earnings data set provided by SSA to the Department. Not all earnings are included as earnings reported to SSA. However, reported earnings include those earnings reported under the following codes on the W2 form:

Box D: Elective deferrals to a section 401(k) cash or deferred arrangement plan (including a SIMPLE 401(k) arrangement);

Box E: Elective deferrals under a section 403(b) salary reduction agreement;

Box F: Elective deferrals under a section 408(k)(6) salary reduction SEP;

Box G: Elective deferrals and employer contributions (including nonelective deferrals) to a section 457(b) deferred compensation plan;

Box H: Elective deferrals to a section 501(c)(18)(D)

tax-exempt organization or organization plan; and

Box W: Employer contributions (including employee contributions through a cafeteria plan) to an employee's health savings account (HSA).¹⁴⁵

Institutions that contend that the omission of earnings not included in those that must be reported to IRS and SSA significantly and adversely affects their D/E rate can make use of alternate earnings appeals to capture that earnings data. The commenters are correct that lottery winnings, child support, and spousal income are not included in the aggregate earnings calculation prepared by SSA for the Department. Funds from those sources do not constitute evidence of earnings of the individual recipient, and their exclusion from aggregate earnings is appropriate.

Changes: None.

Comments: A commenter contended that our process for gathering earnings data disregards actual earnings, unless the wage earner has earnings subject to the Federal Insurance Contribution Act (FICA). The commenter cites a response from SSA to an inquiry posed by the commenter, in which SSA advised that SSA would record earnings for an individual only if those earnings, or other earnings reported for the same individual, were subject to FICA. The commenter contended that aggregate earnings data provided to us by SSA would therefore erroneously treat that individual as having no earnings at all. Because the commenter contended that earnings of public employees in States that do not have section 218 agreements with SSA are not subject to FICA, and are excluded from the MEF, the commenter contended that this results in zero earnings in MEF records of many public employees, and incorrect wage data being provided in the aggregate earnings data SSA provides to the Department.

Discussion: As previously explained, all public employers are now subject to Medicare, and their earnings are now reported to SSA, included in SSA’s MEF, and included by SSA in calculating the aggregate earnings data provided to the Department.

Instances in which an individual may have zero amounts in one or more fields reported to IRS, SSA, or both are handled as follows:

Self-Employment Data:

IRS sends SSA Self-Employment data.  IRS does not send Self Employment records with all zero money fields.  SSA posts the information that is received from IRS to the MEF.

The only time the Social Security Self-Employment Income field is zero on the file received from IRS is when the taxpayer has W-2 earnings at the Social Security maximum.  In this case the Total Net Earnings from Self-Employment is reported in the Self-Employment Medicare Income field on the file received from IRS. 

W-2 Data:

If a form W-2 has a nonzero value in any of the following money fields (and the employee name matches SSA’s records for the SSN) SSA posts the nonzero amount(s) to the MEF:

Box 1 – Wages, tips, other compensation 

Box 3 – Social Security Wages

Box 5 – Medicare wages and tips

Box 7 – Social Security tips

Box 11 – Nonqualified plans

Box 12 code D - Elective deferrals to a section 401(k) cash or deferred arrangement

Box 12 code E - Elective deferrals under a section 403(b) salary reduction arrangement

Box 12 code F - Elective deferrals under a section 408(k)(6) salary reduction SEP

Box 12 code G - Elective deferrals and employer contributions (including non-elective deferrals) to a section 457(b) deferred compensation plan

Box 12 code H - Elective deferrals to a section 501(c)(18)(D) tax-exempt organization plan

Box 12 code W - Employer contributions to your Health Savings Account

If a W-2 has zeroes in all of the above money fields SSA still processes the W-2 for IRS purposes, but does not post the W-2 to the MEF.

In creating the file to send for the Dept. of Education Data Exchange:

(1) If any of the following W-2 Boxes are greater than zero:

• Box 3 (Social Security wages)

• Box 5 (Medicare wages and tips)

• Box 7 (Social Security tips),

the data exchange summary amount includes the greater of the following:

• The sum of Box 3 (Social Security wages) and Box 7  (Social Security tips), or

• Box 5 (Medicare wages and tips).

(2) If:

• Boxes 3, 5, and 7 are all zero, and

• Box 1 (Wages, tips and other compensation) is greater than zero,

the data exchange summary amount includes Box 1 (Wages, tips and other compensation).

(3) In addition to the above, the data exchange summary amount also includes:

• W-2 Box 11 (Nonqualified plans) and

• W-2 Box 12 codes:

  • D (Elective deferrals to a section 401(k) cash or deferred arrangement)

  • E (Elective deferrals under a section 403(b) salary reduction arrangement)

  • F (Elective deferrals under a section 408(k)(6) salary reduction SEP)

  • G (Elective deferrals and employer contributions (including non-elective deferrals) to a section 457(b) deferred compensation plan)

  • H (Elective deferrals to a section 501(c)(18)(D) tax-exempt organization plan)

  • W (Employer contributions to your Health Savings Account)

• For SE the data exchange summary amount includes the amount of Self-Employment income as determined by IRS.

• Earnings adjustments that were created from a variety of IRS and SSA sources.¹⁴⁶

Changes: None.

Comments: One commenter challenged the sufficiency of the SSA MEF data on the ground that many professionals--such as graduates of medical and veterinary schools and perhaps other professional programs--work through subchapter S corporations which do not report earnings through Schedule SE. The commenters stated that the earnings of these individuals would not be included in the MEF. A commenter was concerned that such professionals receive distributions as well as payments labeled compensation, and income for such individuals as captured in SSA data would not reflect the amount earned that was characterized as distributions rather than as salaries.

Discussion: According to IRS guidance, a payment made by a subchapter S corporation for the performance of services is generally considered wages. This is the case regardless of whether the person receiving the payment for the performance of services is an officer or shareholder of a subchapter S corporation.¹⁴⁷ Accordingly, these payments are required to be reported by the subchapter S corporation employer on a Form W-2 filed with the SSA and, therefore, are included in SSA’s MEF.

Changes: None.

Comments: One commenter stated that SSA data do not include earnings information for graduates who secure employment between the end of the calendar year for which earnings are measured and the start of the next award year, nor do the data include a methodology for annualizing earnings of borrowers who secure employment toward the end of the calendar year for which earnings are being measured.

Discussion: In order to measure earnings, one must select a time period for which earnings are counted. Any earnings measurement period, therefore, must include some earnings and exclude others. The objection posed by the commenter is not solved by modifying the earnings measurement period, because any modification would necessarily exclude some other earnings. If students who complete a program have no earnings for some part of the earnings measurement year selected, we see no reason why that period of unemployment should be disregarded in gathering the earnings data used to assess programs under the D/E rates measure. This exercise is not only impracticable, but we believe contrary to the objective of the assessment, which is to take into account periods of unemployment in assessing the outcomes for a GE program. Annualizing earnings--attributing to a student earnings that the individual did not actually receive or otherwise ignoring periods of unemployment--would contravene the Department’s goal to assess the actual outcomes of students who complete a GE program.

Changes: None.

Comments: A commenter objected that §668.405(c) improperly imposed on the institution the burden of identifying those students completing a program who can be excluded under §688.404(e), although the institution would have limited information available to contest their inclusion.

Discussion: The objection misstates the process the Department will follow. Section 668.405(b)(1)(ii) states that the Department compiles and sends to the institution the list of students who completed a program during the cohort period to be assessed, and indicates on that list those students whom the Department considers likely to qualify for exclusion. The institution is free to contend that any of those individuals should be removed for any reason, including qualifying for exclusion under §668.404(e); that an individual designated to be excluded from the list should be included; and that an individual not on the list should be included. The institution has access to NSLDS to gather information relevant to the challenges, and can use information gathered directly from students completing the program and its own records to support a challenge. We note that the assessment occurs at the end of an institutional cohort default rate period, during which an institution is expected to maintain sufficient contact with all of its former students so that it can assist those who may not be meeting their loan repayment obligations. Using those contacts to gather relevant information on those who may qualify for exclusion poses little added burden on the institution.

Changes: None.

Comments: Some commenters contended that using SSA earnings data contravenes the stated objective of the regulations because SSA earnings data capture all earnings regardless of whether the earnings were in an occupation related to the training provided by the program.

Discussion: While we appreciate the commenter’s interest in understanding whether the earnings of students who have completed a program are linked with the training provided by their respective programs, the Department has no way of obtaining this information because SSA cannot disclose the kind of individual tax return data that would identify even the employer who reported the earnings, much less the occupation for which the wages were paid. The regulations are built on the inference that earnings in the period measured are reasonably considered to be the product of the quality of the GE program that the wage earner completed. The training is presumed to prepare an individual for gainful employment in a specific occupation, but it is not unreasonable to attribute gainful employment achieved in a different occupation so shortly after completion of a GE program to be the product of that training. Although there is no practical way to directly connect a particular GE program with earnings achieved relatively soon after completion, the inference that the earnings are the outcome of the training is sufficiently compelling that we do not consider further efforts, even if data were available, to be warranted.

Changes: None.

Comments: Commenters also criticized the Department’s proposal to use SSA data because SSA assigns (“imputes”) zero earnings to all those individuals for whom it does not receive an earnings report that correctly identifies the wage earner and correctly lists the individual’s SSN. The commenters said that earnings reported for these individuals are placed in a suspense file. The commenters cited various reports critiquing the adequacy of efforts to eliminate these mistakes and stated that the scale of these errors suggests that a significant amount of actual earnings would be disregarded because of mistakes by employers on earnings reports.

Discussion: We acknowledge that some earnings are reported but cannot be associated with individuals whose accounts are included in the MEF database, but do not consider the magnitude of the omitted earnings to vitiate the general accuracy of the earnings data contained in the MEF. ¹⁴⁸ The HHS OIG report to which the commenter refers regarding these mismatches cites the employment of unauthorized non-citizens as a major cause of mismatches.¹⁴⁹ Unauthorized non-citizens are not eligible for Federal student financial assistance, and the Department routinely scrutinizes applicants’ immigration status to reduce the likelihood that such individuals will receive title IV, HEA program funds. See 20 U.S.C. 1091(g). Institutions themselves are in a position to identify instances in which unauthorized non-citizens may seek aid. While we recognize that mismatching of earnings occurs, we believe that these restrictions on student eligibility reduce the likelihood that mismatches will affect the accuracy of the MEF earnings data on the population of students who have enrolled in GE programs and whose earnings data are provided to the Department by SSA.

In addition, we believe that the frequency and amount of mismatched earnings are decreasing. SSA moves reported earnings into the suspense file when the individual’s name and SSN combination do not match against SSA’s Numident file. The suspense file does grow over the years; however, SSA performs numerous reinstate processes throughout the tax year that matches previously unmatched records to record the earnings on the proper record. These efforts have resulted in a substantial decrease in the outstanding amounts in the suspense file over the most recent five years for which complete data are available from SSA, as indicated by the following chart.¹⁵⁰

	Earnings Suspense File	Number of Mismatched W-2 Reports
2007	$90,696,742,837.94	10,842,269
2008	87,571,814,470.22	9,580,201
2009	73,380,014,667.81	7,811,295
2010	70,650,921,709.94	7,356,265
2011	70,122,804,272.37	7,128,598

Changes: None.

Comments: Commenters criticized what they described as an assumption of “zero earnings” by SSA for individuals included in the MEF, and contended that this practice suggests that the aggregate earnings data provided by SSA to the Department is not accurate. Commenters further noted that available data indicate that the percentage of zero earnings reported in the 2011 and 2012 GE informational rates showed what the commenters considered to be an unacceptably high percentage of instances of reports of zero earnings, ranging from nine percent for earnings data obtained in July 2013 to as much as 12.5 percent for earnings data obtained in December 2013.

Discussion: There is only one situation in which SSA assumes that an individual has zero earnings. For wage earners with earnings reported for employment type “Household,” the so-called “nanny tax” edit in employer balancing changes to zero the amounts of earnings for Social Security and Medicare covered earnings that fall below the yearly covered minimum amount. If the earnings reported by the employer for such an individual is successfully processed, SSA posts the earnings to the MEF as zero. SSA plans to discontinue this practice next year and will reject the report and have the employer make the correction. These amounts are so low (for 2014, this amount affects only annual earnings less than $1,900) that it is implausible to contend that these assumptions affect the accuracy of the aggregate earnings data provided by SSA to the Department.¹⁵¹

The Department has secured aggregate earnings data from SSA in five instances, as shown in the table below.¹⁵²

	Date Received from SSA	# ED sent to SSA	# SSA verified	# SSA did not verify	# with Earnings	# with Zero Earnings
2011 GE informational rates – includes non- Title IV	3/5/12	811,718	797,070	14,708	699,024	98,046 [12.3% of verified]
2012 GE informational rates for reg neg Title IV only	7/18/13	255,168	252,328	2845	232,006	20,317 [7.96% of verified]
2012 GE post reg neg - Title IV only	8/14/13	923,399	917,912	8,487	798,952	115,960 [12.6% of verified]
For College Scorecard - Title IV only derived from ED data on borrowers in FY 2007 iCDR cohort for selected institutions of higher education	9/13/13	900,419 901,719 902,380 921,749	892,796 894,260 892,840 909,613	7623 7459 9540 12,136	809,204 819,542 787,223 772,574	83,592 74,718 105,617 137,039
	Totals	3,626,267	3,589,509	36,758	3,188,543	400,966 [11.1% of verified]
For College Scorecard - Title IV only derived from ED data on borrowers in FY 2008 iCDR cohort for selected institutions of higher education	12/13/13	969,145 985,742 490,305	954,728 970,742 480,421	14,417 15,000 9884	857,539 865,060 411,917	97,189 105,682 68,504
	Totals	2,445,192	2,405,891	39,301	2,134,516	271,375 [11.3% of verified]
Grand Totals		8,061,744	7,959,705	102,099	7,053,041	906,664 [11.4% of verified]

The commenter asserts that on average, the percentage of verified (matched) individuals who were reported as having zero earnings was 12 percent; in fact, the average was 11.4 percent. We note that the universes of individuals on which SSA provided aggregate earnings data were different: the GE earnings data was obtained for individuals who completed a GE program; the Scorecard data was obtained on all FFEL and Direct Loan borrowers who entered repayment in fiscal years 2007 and 2008, respectively, regardless of the institution or type of program in which they had enrolled, and therefore including borrowers who had been enrolled in GE programs and those who had been enrolled in other programs. Nevertheless, the incidence of zero earnings is similar for both groups.

We note that the 2011 GE informational rates were based on earnings for calendar year 2010; the annual unemployment rate for calendar year 2010 was 9.6 percent.¹⁵³ Those counted as “unemployed” in the published rate do not account for all those who are in fact not employed and earned no reported income; BLS includes as unemployed only those who “do not have a job, have actively looked for work in the prior 4 weeks, and are currently available for work.”¹⁵⁴ Those not included in this group can reasonably be expected to include those students included in a program’s D/E rates calculation who not only do not have a job, but have ceased actively looking for work in the prior month. For this group of students, the SSA data showed zero earnings for 8 percent of the verified individuals included in the rate calculation. Unemployment rates for 2010 for two age groups likely to include most students were higher: for the group ages 20-24, the annual unemployment rate for 2010 was 18.8 percent, and for the group ages 25-34, the annual unemployment rate for 2010 was 10.8 percent.¹⁵⁵ As at least one commenter observed, these results are consistent with high unemployment rates.¹⁵⁶

The 2012 GE informational rates the Department disseminated after the negotiation sessions were based on students’ earnings in calendar year 2011, for which the annual unemployment rate was 8.9 percent, and the annual unemployment rate was 18.1 percent for individuals in the 20-24 age group and 10 percent for individuals in the 25-34 age group. The SSA data for this group of students in GE programs included a 12.6 percent incidence of zero earnings.

In light of the unemployment rates reported for 2010 and 2011, and particularly the rates for the two age groups that likely include the great majority of students completing a GE program, the incidence of zero earnings in the SSA records is neither unexpected nor of such a magnitude with regard to the number of wage earners as to demonstrate that the SSA MEF database is unreliable as a data source for determining D/E rates.¹⁵⁷

Changes: None.

Comments: Commenters asserted that by considering all zero earnings data to evidence no earnings for an individual, the Department treats each such individual as having no earnings during that year, although the individual may in fact have significant but misreported earnings. The commenters cited as a significant example of such earnings omissions the earnings of public employees whom the commenters consider as good examples of individuals with significant earnings, but whose SSA earnings would show zero earnings. The commenters criticized this as producing a bias that understates earnings. The commenters contended that the D/E rates should be adjusted, based on assumptions that the missing earnings are actually distributed throughout a program’s cohort of earners. The commenters asserted that if earnings of failing GE programs were to be adjusted on that assumption, 19 percent of programs that failed the annual earnings rate would pass that threshold, and 9 percent of programs that failed the discretionary income rate would pass that threshold.

Discussion: As explained earlier, the commenter’s assertion that the earnings of public employees are often, even typically, not reported to SSA is not correct. The earnings of public employees are reported to SSA, public employees are not “deemed” by SSA to have “zero earnings,” and SSA includes actual earnings reported for public employees in the aggregate earnings data SSA provides to the Department. Accordingly, it is not reasonable to conclude that public employees with actual earnings account for any appreciable number of “zero earnings” records.

The commenters argue that in those instances in which actual earnings are missing from the MEF, those missing wages include earnings in amounts spread throughout the cohort of students who completed a program. Thus, the commenters contend, our practice that considers all instances of “zero earnings” to be evidence that the individual in fact had no earnings during that year causes the earnings for the cohort to be significantly understated. Some “zero earnings” records result from misreported earnings or unreported earnings. However, other individuals will in fact have zero earnings, and the contention that the missing earnings belong to individuals with significant earnings appears to rest in large part on the misconception that earnings of public employees are not included in MEF, and thus appear as “zero earnings.”

We recognize that misreported and underreported earnings can have some effect on the earnings data we use, but those same issues would affect any alternative data source that might be available. The commenters suggest no practicable alternative that would eliminate these issues and provide more reliable data sufficient to accomplish our objective here--determining earnings of individuals who completed a particular GE program offered by a particular institution. We note that an institution that believes that incidents of mismatches significantly and adversely affect SSA aggregate earnings data for the students completing a program may appeal its zone or failing D/E rates by submitting an alternate earnings appeal based on State earnings database records or a survey.

Changes: None.

Comments: Commenters contended that the Department’s earnings assessment process is flawed with regard to information on self-employed individuals because the source of data on their earnings is the individual, who may fail to report or significantly underreport earnings, or who may have relatively significant business expenses that offset even substantial income. According to the commenters, barbering, cosmetology, food service, and Web design are examples of occupations in which significant numbers of individuals are self-employed and tend to underreport earnings, particularly earnings from tips, which a commenter states account for about half of earnings in service occupations such as cosmetology. Another commenter believed that employers may often fail to report payments to independent contractors whom they have retained for relatively short periods, which would further depress the amount of earnings shown for the contractors in SSA records. One commenter provided an alternate analysis that imputes certain values derived from the CPS conducted by the Census Bureau on behalf of BLS. The commenter proposed to adjust the calculation of D/E rates to take into account what the commenter considered bias in the income data reported to SSA for workers in several occupations that the CPS shows involve both significant tip income and a high percentage of income from self-employment. The commenter contended that these adjustments would significantly augment the SSA aggregate earnings reported for these occupations, increasing the median earnings by 19 percent and the mean earnings by 24 percent.

Discussion: We do not agree that our reliance on reported earnings is flawed because of its treatment of self-employment earnings and tips, or that the suggested methods for remedying the claimed flaws would be effective in achieving the goals of these regulations, for several reasons. We acknowledge that some self-employed individuals may fail to report, or underreport, their earnings. However, section 6017 requires self-employed individuals to file a return if the individual earns $400 or more for the taxable year. 26 U.S.C. 6017. Underreporting subjects the individual to penalty or criminal prosecution. See, e.g., 26 U.S.C. 6662, 7201 et seq.

Some self-employed individuals have significant income but substantial and offsetting business expenses, such as travel expenses and insurance, but our acceptance of net reported earnings for these individuals is not unreasonable. These individuals must use available earnings to pay their personal expenses including repaying their student loan debt. The fact that an individual used some revenue to pay business expenses does not support an inference that the individual had those same funds actually available to pay student loan debt.

With respect to the earnings of workers who regularly receive tips for their services, section 6107 of the Code requires individuals to report to IRS their tip earnings for any month in which those tips exceeded $20, and individuals who fail to do so are subject to penalties. 26 U.S.C. 6107, 6652(b).¹⁵⁸

As to the concern that some businesses may fail to report payments to contractors, the individual contractor remains responsible for reporting those payments as with other self-employment earnings, whether or not the payments were reported by the party that engaged the individual.

Imputing some percentage of added earnings to account for underreported tips and other compensation could only be done by generalizations drawn from some source of data on earnings, but none has been suggested that would permit doing so in a way that would distinguish between programs.

To assess the bias that the commenter asserted arises from what the commenter calls “imputing” zero earnings to individuals with no reported earnings in the MEF, the commenter relies on earnings data from the CPS, which is derived from surveys of households. The survey samples data on a selection of all households, and relies on earnings data as provided by the individuals included in the survey. As the commenter noted, there are no data in the CPS that allow one to associate a particular respondent with a particular GE program.

Unlike the approach taken in these regulations, which captures all earnings of the cohort of students completing a program and credits those earnings to the program completed by the wage earners, the analysis proposed by the commenter does the reverse: it extrapolates from earnings reported by those survey recipients who identify their occupation as one that appears related to GE programs of that general type, and then projects an increase in aggregate earnings for all GE programs in the category of programs that appears to include that occupation.. In fact, even if the respondents were all currently employed in occupations for which a category of GE programs trains students, the respondents’ earnings will almost certainly have no connection with a particular GE programs we are assessing. Because any inference drawn from CPS respondents’ earnings could only benefit a whole category of programs--improving the D/E rates for every program in that category--using such inferences would mask poorer performing programs and thwart a major purpose of the GE assessment.

In addition, by the time the survey is conducted, the respondent can be expected to identify his or her current or most recent job, which may be different than the occupation for which training was received years before in a GE program. Thus, to draw a usable inference about D/E earnings from data gathered in the CPS one must connect a particular GE program now being offered and evaluated with earnings and occupations disclosed by the CPS respondents years, even decades, into their careers, during which they may have worked in different kinds of occupations.

For these reasons, we do not agree with the commenters’ assertion that aggregate earnings data provided by SSA from MEF are unreliable with respect to workers in occupations that involve significant tip income or a high percentage of income from self-employment. More importantly, the critique fails to demonstrate either that a different and more reliable source of earnings data is available and should reasonably be used instead of the SSA data, or that adjustments must be made based on CPS data. Moreover, the regulations allow an institution to submit an alternate earnings appeal using State databases or a survey.

Changes: None.

Comments: For the various reasons stated in the comments summarized here, commenters contended that the SSA MEF data is not the “most reliable data available” for the Department to use in calculating D/E rates for GE programs, and does not “produce figures that can be considered sufficiently accurate.” They asserted that the Department has not met its obligation to use the “best available data” to calculate the D/E rates.

Discussion: The commenter’s argument that the Department failed to use the “most reliable data available” is based on cases in which parties claimed that an agency chose to rely on incomplete or outdated data at the time it made a determination, rather than more accurate data available to the agency at that time. In the relevant cases, the court considered whether the agency reasonably relied on the data available to the agency at the time of determination.¹⁵⁹ An agency may not disregard data actually available to it, as where, for example, data are available from a component of the same agency as the component of that agency that makes the determination. The data required to calculate the earnings component of the D/E rates is not available within components of the Department.

Similarly, an agency may not ignore or fail to seek data actually held by an agency with which it has a “close working relationship.” See Baystate, 545 F.Supp.2d at 44-45. SSA and the Department have a close working relationship, and the Department has, in fact, sought and obtained the relevant data available from SSA. The commenter does not identify any source other than SSA for the aggregate earnings data needed to calculate D/E rates. Rather, the commenter focuses on the lack of better data from SSA. We have confirmed with SSA that it does not have better data available to share with the Department, and, therefore, the Department uses the best data available from SSA to calculate earnings. Accordingly, the Department has satisfied the requirement to use the most reliable data available.

The case law establishing the requirement that an agency use the best available data does not require that the data be free from errors. The case law “amply supports the proposition that the best available data standard leaves room for error, so long as more data did not exist at the time of the agency decision.” Baystate, 545 F.Supp.2d at 49. As discussed, the commenter does not identify, and the Department is not aware of, any other source of earnings data available to the Department to calculate D/E rates for a GE program. As we recognize that there are shortcomings in the D/E rates data-gathering process, we provide for a process under §668.405(c) for institutional corrections to the information submitted to SSA, and, to address any perceived flaws in the SSA aggregate earnings data, in §668.406, we provide institutions an opportunity to appeal their final D/E rates using alternate earnings data obtained from a student survey or State-sponsored data system. For these reasons, by using aggregate earnings data provided by SSA from its MEF, the Department has satisfied the requirement to use the best available data.

Changes: None.

Comments: Several commenters contended that the Department’s use of SSA aggregate earnings data to determine the D/E rates violates the institution’s due process rights because the regulations prohibit the institution from examining and challenging the earnings data the SSA uses to calculate the mean and median earnings. The commenters argued that the regulations deprive the institution of the right to be apprised of the factual material on which the Department relies so that the institution may rebut it. Commenters further contended that appeal opportunities available under the regulations are not adequate, and that the regulations impermissibly place burdens of proof on the institution in exercising challenges available under the regulations.

Discussion: As previously explained, SSA is barred from disclosing the kind of personal data that would identify the wage earners and from disclosing their reported earnings because section 6103(a) of the Internal Revenue Code (Code) bars a Federal agency from disclosing tax return information to any third party except as expressly permitted by the Code. 26 U.S.C. 6103(a). Return information includes taxpayer identity and source or amount of income. 26 U.S.C. 6103(b)(2)(A). No provision of the Code authorizes SSA to disclose return information to the Department for the purpose of calculating earnings, and therefore we cannot obtain this information from SSA (or IRS itself).

We disagree that the limits imposed by law on SSA’s release of tax return information on the students comprising a GE cohort deprives the institution of a due process right. One commenter’s contention that the failure to make return information available violates the institution’s right to meaningful disclosure of the data on which the Department relies is not supported by the case law. Indeed, the case law to which the commenter refers simply states that an agency must provide a party with--

[E]nough information to understand the reasons for the agency’s action. . . . Claimants cannot know whether a challenge to an agency’s action is warranted, much less formulate an effective challenge, if they are not provided with sufficient information to understand the basis for the agency's action.

Kapps v. Wing, 404 F.3d 105, 123-24 (2d Cir. 2005) (emphasis added). Similarly another commenter cites to Bowman Transp., Inc. v. Arkansas-Best Freight Sys., Inc., 419 U.S. 281 (1974) to support a claim that failure to provide the completers’ tax return data denies the institution a right to due process, but the Court there held that--

A party is entitled, of course, to know the issues on which decision will turn and to be apprised of the factual material on which the agency relies for decision so that he may rebut it. Indeed, the Due Process Clause forbids an agency to use evidence in a way that forecloses an opportunity to offer a contrary presentation.

Bowman Transp., Inc. v. Arkansas-Best Freight Sys., Inc., 419 U.S. at 289, fn.4. The procedure we use here apprises the institution of the factual material on which we base our determination, and more importantly in no way forecloses an opportunity to offer a “contrary presentation.”

The regulations establishing the procedure we use to calculate a program’s D/E rates provide not merely an opportunity to challenge the accuracy of the list of students who completed the program and the debts attributed to the cohort, but also two separate kinds of “contrary presentations” regarding earnings themselves--a survey of students who completed the program and their earnings, and data on their earnings from State databases. An institution may make either or both such presentations. Under the Mathews v. Eldridge test, an agency must provide procedures that are “tailored, in light of the decision to be made, to ‘the capacities and circumstances of those who are to be heard,’ . . . to insure that they are given a meaningful opportunity to present their case.” Mathews v. Eldridge, 424 U.S. 319, 349 (1976) (citations omitted). The circumstances in which the Department determines D/E rates include several facts that bear on the fairness of the opportunity given the institution to contest the determination. First, SSA is legally barred by section 6103 of the Code from providing the Department or the institution with individualized data on the members of the program cohort. Second, SSA MEF data is the only source of data readily and generally available on a nationwide basis to obtain the earnings on these cohorts of individuals. Third, parties who report to SSA the data maintained in the MEF do so under penalty of law. Fourth, millions of taxpayers, as well as the government, rely on the SSA MEF data as an authoritative source of data that controls annually hundreds of billions of dollars in Federal payments and taxpayer entitlement to future benefits.¹⁶⁰ Fifth, the entities directly affected by the determinations--businesses that offer career training programs, many of which derive most of their revenue from the title IV, HEA programs--are sophisticated parties. Lastly, institutions are free to present, and have us consider, alternative proofs of earnings. As previously discussed in the context of the requirement to provide the “best available data,” the agency’s determination “cannot be weighed in a vacuum, but must be evaluated by reference to the data available to the agency at the relevant time.” Baystate, 545 F.Supp.2d at 41. Under these circumstances, the regulations provide institutions sufficient opportunity to understand the evidence on which the Department determines D/E rates and a meaningful opportunity to contest and be heard on a challenge to that determination. No more is required.¹⁶¹ And, although State earnings databases may not be readily available to some institutions because of their location or the characteristics of the data collected and stored in the database, an institution has the option of conducting a survey of its students and presenting their earnings in an alternate earnings appeal.

Changes: None.

Comments: A commenter contended that the Department’s practice of treating a “zero earnings” instance in SSA’s MEF data as no earnings for the individual is improper, contrary to the practice of other Federal and State agencies, and in violation of acceptable statistical methods. According to the commenter, the U.S. Census Bureau, BLS, the Federal Economic Statistical Advisory Committee, and the Bureau of Justice Statistics all replace zero values with imputed values derived, for example, from demographically similar persons for whom data are available.

Specifically, the commenter cited the following examples in which agencies impute positive values where data are missing:

The United States Census Bureau (The Federal Economic Statistical Advisory Committee) uses the following imputation methods¹⁶²:

• Relational imputation: Infers the missing value from other characteristics on the person’s record or within the household (i.e., if other members of household report race, then census will infer race based on household data).

• Longitudinal edits: Data entered based on previous entries (from past reporting periods) from the same individual or household.

• Hot Deck edits: A record with similar characteristics (race, age, sex, etc.) is a hot deck. Uses data from hot deck entries to impute missing values.

BLS¹⁶³ and the Department of Education, National Center for Education Statistics¹⁶⁴ also use hot deck imputation (or a similar method based on demographics).

The Bureau of Justice Statistics uses the median value of an item reported in a previous survey by other agencies in the same sample cell.

Similarly, the commenter noted that State child support enforcement agencies typically impute earnings values when calculating the amount of child support required from a parent for whom no earnings data are available. The commenter stated that the Department’s failure to impute earnings values in instances in which SSA data show no earnings can be expected to result in underestimation of mean and median earnings.

Discussion: The Department recognizes that other agencies, and the Department itself, may in some circumstances impute values for missing data in various calculations. Surveys conducted to discern and evaluate economic and demographic characteristics of broad populations can and are regularly made without the need for complete values for each individual data element included in the survey or analysis. In these assessments, the objective is determining characteristics of broad groups of entities or individuals. These surveys or studies typically involve universes comprising a great number of entities or individuals, about which the survey conductor has a considerable amount of current and older data available both from the entity for which data are missing and from others in the universe. Where such data are available, the survey conductor can identify both entities that sufficiently resemble the entity for which data are missing, and what data were actually provided by that entity in the past, to allow the surveyor to impute values from the known to the unknown. Where sufficient data exist, the agency can control the effect of imputing values by limiting the extent to which values will be imputed. Whether the imputation provides precisely accurate values for those values missing in the data is irrelevant to the accuracy of the overall assessment. In calculating D/E rates for a particular program, the opposite is the case; measuring the earnings of a particular cohort of graduates of a GE program offered by a particular institution requires that the Department use data that allow it to differentiate among the outcomes of identical GE programs offered by separate institutions.¹⁶⁵

Imputation of income in the context of establishing child support obligations is a completely different enterprise: income is imputed to a non-custodial parent only in an individual judicial or administrative proceeding in which the non-custodial parent is a defendant, and has failed to produce earnings evidence or is either unemployed or considered to be underemployed.¹⁶⁶ Imputed income is used when the court believes the parent’s testimony regarding reported income is false; the evidence of the parent’s income and the parent’s actual income does not meet his or her demonstrated earnings; or a decrease in income is voluntary. At a minimum, income is imputed to equal the amount earned from a full-time job earning minimum wage.¹⁶⁷ The objective of the child support determination process is to ensure that the defendant parent is contributing to the support of the child, and not shirking that responsibility by failing to find employment or failing to maximize earnings. Thus, the parent is expected to find appropriate employment to meet this obligation, and can object by demonstrating a “good faith reason” why he or she cannot do so.¹⁶⁸ In each instance, income is imputed only on a particularized assessment of the individual and his or her circumstances.

Because of these differences in procedure and objective, child support practice offers no useful model for imputing earnings to those graduates of a GE program whose MEF records show no reported earnings. The objective of calculating the mean and median earnings for graduates of a GE program--to assess the actual outcomes of that program for a specific group of students who completed the program--is very different. The assessment assumes that those graduates enrolled and persisted in order to acquire the skills needed to find gainful employment, and had no reason--such as a desire to minimize a child support obligation--to decline gainful employment that they could otherwise achieve using the skills acquired in a GE program. Because the Department receives no data that would identify an individual whose MEF record shows no reported earnings, the Department is not able to determine whether an individual was making full use of the skills for which the individual enrolled in a GE program to acquire.

Changes: None.

Comments: One commenter objected to the language in §668.405(c)(1) that provides that the Secretary presumes that the list of students who completed a program and the identity information for those students is correct. The commenter was concerned that, through this presumption, the Department would limit its ability to reject an inaccurate or falsified list of students. For example, this commenter explained, an institution could falsely report that fewer than 30 students completed a program so as to avoid a D/E rates calculation under the n-size provisions of the regulations. The commenter recommended modifying §668.405(c)(1) to state “the Secretary may presume” that the list is correct, in order to clarify that the presumption is at the Secretary’s discretion.

Discussion: Because the list of students who completed a program is created by the Department from data reported by the institution, we presume that it is correct. We do not agree that this presumption is a limitation on the Department. Rather, it confirms that the burden of proof to demonstrate that the list is incorrect resides with the institution. The list is created using data originally reported to the Department by the institution.

We note that institutions that submit reports to the Department are subject to penalty under Federal criminal law for making a false statement in such a report. See, e.g., 18 U.S.C. §1001, 20 U.S.C. §1097(a). Because the Department can take enforcement action under these statutes, the Department need not, and typically does not, include in procedural regulations explicit provisions explaining that the Department can take enforcement action when we determine that an institution has submitted untruthful statements.

Changes: None.

Comments: Many commenters objected to the proposal that earnings data could be obtained from SSA “or another Federal agency” because it was not transparent as to which other agency the Department may rely on to provide earnings data. The commenters objected to not being able to provide informed comment during the rulemaking process on the data source. The commenters also questioned the quality of the data that the Department would receive from another Federal agency.

Discussion: This clause was included in the proposed regulations so that, if a future change in law or policy precluded SSA from releasing earnings data, the Department would have the option to obtain this information from another Federal agency. However, in response to the commenters’ concerns, we will designate any new source of earnings data through a change in regulations through the rulemaking process so that the public has an opportunity to understand any proposed change and offer comments.

Changes: The clause “or another Federal agency” has been removed from §§668.404(c)(1), 668.405(a)(3), 668.405(d), 668.413(b)(8)(i)(C), and 668.413(b)(9)(i)(C).

Comments: One commenter urged the Department to create a mechanism for institutions to monitor and evaluate the student data used to calculate the D/E rates on a continuous basis so that they can make operational adjustments to ensure that programs pass the D/E rates measure.

Discussion: There are several factors that preclude institutions from using real-time data to estimate the D/E rates for a GE program on a continuous basis. First, the Department may only request mean and median earnings for a cohort of students from SSA once per year. As a result, we would not be able to provide institutions with updated earnings information at multiple points during the year. Second, any estimate of the amount of debt a student will have incurred upon completion of a GE program would involve too many assumptions to make the estimate meaningful. For example, any estimate would have to make assumptions regarding how many loan disbursements a student received and whether and when the student completed the program. Further, the estimate would have to make assumptions as to whether a student would be excluded from the calculation for any of the reasons listed in §668.404(e).

Changes: None.

§668.406 D/E Rates Alternate Earnings Appeals

Comments: We received a number of comments requesting clarification regarding the cohort of students on whom an alternate earnings appeal would be based. Although the proposed regulations provided that an appeal would be based on the annual earnings of the students who completed the program during the same cohort period that the Secretary used to calculate the final D/E rates, commenters suggested that we specify the calendar year for that period. One commenter suggested that we specify that the cohort period is the calendar year that ended during the award year for which D/E rates were calculated. Another commenter recommended that, where the most recently available earnings data from SSA are not from the most recent calendar year, institutions should be permitted to use alternate earnings data from the most recent calendar year.

Some commenters asked that we specify that the students whose earnings are under consideration are the same students on the final list submitted to SSA under §668.405(d). In that regard, a number of commenters suggested that institutions should be able to apply the exclusions in §668.404(e), in determining the students in the cohort period.

One commenter asked the Department to permit institutions to modify the cohort of students to increase the availability of an alternate earnings appeal. Other commenters asked the Department to permit institutions to expand the cohort period if necessary to meet the survey standards or the corresponding requirements of an appeal based on earnings information in State-sponsored data systems.

Discussion: We believe the regulations sufficiently describe the relevant period for which earnings information is required in an alternate earnings appeal. As discussed in “§668.404 Calculating D/E Rates,” because D/E rates are calculated for the award year, rather than the calendar year, and because of the timeline associated with obtaining earnings data from SSA, we state that the earnings examined for an alternate earnings appeal must be from the same calendar year for which the Department obtained earnings from SSA under §668.405(c). The purpose of the appeal is to demonstrate that, using alternate earnings for the same cohort of students, the program would have passed the D/E rates measure. Accordingly, it would not be appropriate to use data from a year that is different from the one used in calculating the D/E rates. In “§668.404 Calculating D/E Rates,” we provide an example that illustrates how the period will be determined.

Under this approach, because an institution will know in advance the cohort of students and calendar year for earnings that will be considered as a part of an appeal, the institution can begin collecting alternate earnings data well before draft D/E rates are issued in the event that the institution believes its final D/E rates will be failing or in the zone and plans to appeal those D/E rates.

We agree that institutions should be able to exclude students who could be excluded under §668.404(e) in their alternate earnings appeal. We recognize that in order to maximize the time that an institution has to conduct a survey or database search, the institution may elect to begin its survey or search well before the list of students is submitted to SSA, and the exclusions from the list under §668.404(e), are finalized.

We also agree that there may be instances where a minor adjustment to the cohort period may make available an alternate earnings appeal that would not otherwise meet the requirements of the regulations. For example, for an appeal based on earnings information in State-sponsored data systems, the information may not be collected or organized in a manner identical to the way in which earnings data are collected and organized by SSA, and a minor adjustment to the cohort period may be necessary to meet the matching requirements. In this regard, we note that an institution would not be permitted, however, to present annualized, rather than annual, earnings data in an alternate earnings appeal, even if that is how the data are maintained in a State-sponsored data system.

In accordance with instructions on the survey form, an institution may exclude from its survey students that are subsequently excluded from the SSA list. For a State data system search, the institution may exclude students that are subsequently excluded as long as it satisfies the requirements under §668.406(d)(2). Under those requirements the institution must obtain earnings data for more than 50 percent of the students in the cohort, after exclusions, and that number of students must be 30 or more.

Changes: We have revised the provisions of §668.406(c)(1) and (d)(1) in the final regulations (§668.406(a)(3)(i) and (a)(4)(i) in the proposed regulations), and added §668.406(b)(3), to permit institutions to exclude students who are excluded from the D/E rates calculation under §668.404(e). If the institution chooses to use an alternate earnings survey, the institution may, in accordance with the instructions on the survey form, exclude students that are excluded from the D/E rates calculation. If the institution obtains annual earnings data from one or more State-sponsored data systems, it may, in accordance with §668.406(d)(2), exclude from the list of students submitted to the administrator of the State-administered data system students that are excluded from the D/E rates calculation. We have also included in §668.406(d)(2) that an institution may exclude these students with respect to its appeal based on data from a State-sponsored data system.

We have also provided in §668.406(b)(3) that an institution may base an alternate earnings appeal on the alternate earnings data for students who completed the program during a cohort period different from, but comparable to, the cohort period that the Secretary used to calculate the final D/E rates.

Comments: We received comments in support of permitting institutions in an alternate earnings appeal to include the earnings of individuals who did not receive title IV, HEA program funds for enrollment in the program and, also, a comment opposing the inclusion of those individuals. Those commenters in support argued that the earnings of students who receive title IV, HEA program funds for enrollment in a program are not representative of the earnings of all their program graduates and therefore the earnings of all individuals who complete a program should be considered on appeal. On the other hand, one commenter recommended that the basis for an alternate earnings appeal be limited to the earnings of students who received title IV, HEA program funds for enrollment to align the regulations with the district court’s decision in APSCU v. Duncan.

Discussion: We agree with the commenter who recommended that the basis for an alternate earnings appeal be limited to the earnings of students who received title IV, HEA program funds for enrollment in the program. We believe this approach better serves the purpose of the alternate earnings appeal--to allow institutions, which are not permitted to challenge the accuracy of the SSA data used in the calculation of the D/E rates, to demonstrate that any difference between the mean or median annual earnings the Secretary obtained from SSA and the mean or median annual earnings from an institutional survey or State-sponsored data system warrants revision of the final D/E rates. The purpose of the appeal is to permit institutions to present evidence that the earnings data used to calculate the D/E rates may not capture the earnings outcomes of the students on whom the D/E rates were based, rather than to present evidence of the earnings of a different set of individuals who completed the program. As the commenter noted, the approach we take here, which considers only outcomes for individuals receiving title IV, HEA program funds, also aligns the regulations with the court’s interpretation of relevant law in APSCU v. Duncan that the Department could not create a student record system based on all individuals enrolled in a GE program, both those who received title IV, HEA program funds and those who did not. See APSCU v. Duncan, 930 F. Supp. 2d at 221. Further, because the primary purpose of the D/E rates measure is to determine whether a program should continue to be eligible for title IV, HEA program funds, we believe we can make a sufficient assessment of whether a program prepares students for gainful employment based only on the outcomes of students who receive title IV, HEA program funds, including in connection with an alternate earnings appeal of a program’s D/E rates. By limiting the alternate earnings appeal to an assessment of outcomes of only students who receive title IV, HEA program funds, the Department can monitor the Federal investment in GE programs. See the NPRM and our discussion in this document in “§668.401 Scope and Purpose” for a more detailed discussion regarding the definition of “student” in these regulations as an individual who receives title IV, HEA program funds for enrollment in a program.

Changes: None.

Comments: A number of commenters urged the Department to permit appeals based on current BLS earnings data, either as a standing appeal option or as an option only during the transition period.

Discussion: We do not believe that BLS data reflect program-level student outcomes, which are the focus of the accountability framework in the regulations. The average or percentile earnings gathered and reported by BLS for an occupation include all earnings gathered by BLS in its survey, but do not show the specific earnings of the individuals who completed a particular GE program at an institution and, therefore, would not provide useful information about whether the program prepared students for gainful employment in that occupation. Accordingly, we decline to include an option for alternate earnings appeals that rely on BLS data.

Changes: None.

Comments: One commenter recommended that an institution should be required to deliver any student warnings and should be subject to any other consequences under §668.410 based on a program’s final D/E rates while an appeal is pending. The commenter expressed concern that suspending any such requirements and consequences until resolution of an appeal, as we provide in §668.406(a)(5)(ii) of the proposed regulations (§668.406(e)(2) of the final regulations), would prevent students from receiving information that may be critical to their educational decision making. The commenter also proposed that an appeal, if successful, should not change a program’s results--that is, failing or in the zone--under the D/E rates measure, but should only preserve a program’s eligibility for title IV, HEA program funds for another year.

Discussion: Although we agree that it is important for students and prospective students to receive important information about a GE program’s student outcomes in a timely manner, we continue to believe that it is not appropriate to sanction an institution on the basis of D/E rates that are under administrative appeal. The purpose of the administrative appeal is to allow an institution to demonstrate that, based on alternate earnings data, a program’s final D/E rates, calculated using SSA earnings data, warrant revision. To make the administrative appeal meaningful, we do not believe that institutions should be subject to the consequences of failing or zone D/E rates during the limited appeal period. We also believe it could potentially be confusing and harmful to students and prospective students to receive student warnings from an institution that is ultimately successful in its administrative appeal. We note that, under §668.405(g)(3) and §668.406(e)(2) of the final regulations, the Secretary may publish final D/E rates once they are issued pursuant to a notice of determination, with an annotation if those rates are under administrative appeal. Accordingly, we expect that final D/E rates will be available to inform the decision making of students and prospective students, even during an administrative appeal.

In addition, we believe that a successful appeal should result in a change in a program’s final D/E rates. The purpose of the alternate earnings appeal process is to allow institutions to demonstrate that any difference between the mean or median annual earnings the Secretary obtained from SSA and the mean or median annual earnings from a survey or State-sponsored database warrants revision of the D/E rates. If an institution is able to demonstrate that, with alternate earnings data, a program would have passed the D/E rates measure, the program should have all benefits of a passing program under the regulations.

Changes: None.

Comments: Two commenters asked the Department to provide institutions a period longer than three business days after the issuance of a program’s final D/E rates to give notice of intent to file an alternate earnings appeal. One commenter proposed a period of 15 days after issuance of the final D/E rates. The commenters believed that the time provided in the proposed regulations is not sufficient to complete review of a program’s D/E rates.

Discussion: Section 668.406(a)(5)(i)(A) of the proposed regulations provided that, to pursue an alternate earnings appeal, an institution would notify the Secretary of its intent to submit an appeal no earlier than the date the Secretary provides the institution with the GE program’s draft D/E rates and no later than three business days after the Secretary issues the program’s final D/E rates. In other words, although an appeal is made based on a program’s final D/E rates, an institution can give notice of its intent to submit an appeal as soon as it receives draft D/E rates. Under §668.405, a program’s final D/E rates are not issued until the later of the expiration of a 45-day period in which an institution may challenge the accuracy of the loan debt information the Secretary used to calculate the median loan debt for the program and the date on which any such challenge is resolved. Accordingly, under the proposed regulations, the window during which an institution may submit notice of its intent to submit an alternate earnings appeal would not be, as suggested by the commenters, limited to the three-day period after the issuance of the final D/E rates. Rather, an institution would have, at a minimum, the 48-day period after draft D/E rates are issued. We believe that draft D/E rates provide an institution with sufficient information to determine whether to submit an alternate earnings appeal. We also believe that a 48-day minimum period to give notice of intent to submit an appeal adequately balances the Department’s interests in ensuring that a program’s final D/E rates are available to prospective students and students at the earliest date possible and providing institutions with a meaningful opportunity to appeal. Nonetheless, we appreciate that some institutions may not be able to give notice of intent to appeal until final D/E rates have been issued. To provide institutions with adequate time to decide whether to pursue an alternate earnings appeal, and if so, to communicate that intention, while still ensuring that the Department can promptly disclose the program’s final D/E rates to the public, we are revising the regulations to provide that, as in the 2011 Prior Rule, an institution has until 14 days after final D/E rates have been issued to notify the Department of its intent to submit an appeal.

Changes: We have revised the provision in §668.406(e)(1)(i) of the final regulations (§668.406(a)(5)(i)(a) of the proposed regulations), to require an institution to notify the Secretary of its intent to submit an alternate earnings appeal no later than 14 days after the Secretary issues the notice of determination.

Comments: Two commenters asked the Department to give institutions a period longer than 60 days after the issuance of a program’s final D/E rates to submit the documentation required for an alternate earnings appeal. One of the commenters proposed 120 days. The commenters believed that the time provided is not sufficient to meet the requirements of an appeal.

Discussion: Under §668.405, a program’s final D/E rates are not issued until the later of the expiration of a 45-day period after draft D/E rates are issued, during which an institution may challenge the accuracy of the loan debt information used to calculate the median loan debt for the program, and the date on which any such challenge is resolved. The period available to an institution to take all steps required to submit an alternate earnings appeal is not, as suggested by some of the commenters, limited to the 60-day period after the issuance of the final D/E rates. As we note previously, draft D/E rates should provide an institution with sufficient information to determine whether it intends to submit an alternate earnings appeal. Consequently, an institution has, at a minimum, the 45-day period after draft D/E rates are issued, together with the 60 days after issuance of final D/E rates, or 105 days in total to submit the documentation required for an alternate earnings appeal.

An institution also has the option to begin its alternate earnings survey or collection of data from State-sponsored data systems well before the Secretary provides the institution with its draft D/E rates. For example, assume that the first award year for which D/E rates could be issued is award year 2014-2015. Those rates would be based on the outcomes of students who completed a GE program in award years 2010-2011 and 2011-2012 for a two-year cohort period, and 2008-2009, 2009-2010, 2010-2011, and 2011-2012 for a four-year cohort period. SSA would provide to the Department data on the students’ earnings for calendar year 2014 in early 2016, approximately 13 months after the end of calendar year 2014. Those earnings data would be used to calculate the D/E rates for award year 2014-2015, and draft rates would be issued shortly after the final earnings data are obtained from SSA. Under our anticipated timeline, an institution that receives draft D/E rates that are in the zone or failing for award year 2014-2015 would receive those draft rates early in 2016. An institution that wished to conduct a survey to support a potential alternate earnings appeal of its D/E rates for award year 2014-2015 would base its appeal on student earnings during calendar year 2014. Students who completed the GE program would know by early 2015 how much they earned in 2014, and could be surveyed, as early as the beginning of 2015--more than a full year before the Department would issue final D/E rates for award year 2014-2015.

We believe the regulations provide sufficient time to permit an institution to conduct an earnings survey or collect State earnings data and submit an alternate earnings appeal. To permit more time would further delay the receipt by students and prospective students of critical information about program outcomes and unnecessarily increase the risk that more students would invest their time and money, and their limited eligibility for title IV, HEA program funds, in a program that does not meet the minimum standards of the regulations.

Changes: None.

Comments: None.

Discussion:  Section 668.406(a)(3)(i) of the proposed regulations provided that NCES will develop a valid survey instrument targeted at the universe of applicable students who complete a program.  We have determined that a pilot-tested universe survey, rather than a field-tested sample survey, as provided in the proposed regulations, is the appropriate vehicle to understand the appropriateness of the survey items and the order in which they are presented.  While a field test implies a large-scale, nationally representative survey that is the precursor to a full-scale survey administration, and evaluates the operational aspects of a data collection as well as the survey items themselves, a pilot test is smaller and is more geared towards evaluating the survey items, rather than the operational procedures, as is more appropriate for these purposes.

Although institutions are not required to use the exact Earnings Survey Form provided by NCES, we believe that institutions should use the same survey items and should present them in the same order as presented in the Earnings Survey Form to ensure that the pilot-tested survey items are effectively implemented.  We note that, as we stated in the NPRM, the NCES Earnings Survey Form will be made available for public comment before it is implemented in connection with the approval process under the Paperwork Reduction Act of 1995. 

Changes: We have revised the provision in §668.406(c)(1) of the final regulations (§668.406(a)(3)(i) of the proposed regulations), to specify that the Earnings Survey Form will include a pilot-tested universe survey and provide that, although an institution is not required to use the Earnings Survey Form, in conducting a survey it must adhere to the survey standards and present to the survey respondent in the same order and same manner the same survey items included in the Earnings Survey Form.

Comments:  Several commenters noted that they were unable to evaluate whether the standards for alternate earnings appeals based on survey data are appropriate because the NCES Earnings Survey Form that will include the standards will not be released until a later date.  These commenters also questioned the fairness and expense of requiring institutions to submit an independent auditor’s report with the survey results. Another commenter suggested that a survey-based alternate earnings appeal would be too costly for small institutions.

On the other hand, one commenter argued that less rigorous survey standards would not be appropriate and recommended that the Department institute additional measures to ensure that institutions do not improperly influence survey results. Specifically, the commenter suggested that the Department conduct audits of surveys to determine if there was improper influence and require an institution’s chief executive officer to include in the required certification a statement that no actions were taken to manipulate the survey results.

Discussion:  We appreciate the commenters’ concerns and expect that the survey standards developed by NCES will balance the need for reliable data with our intent to provide a meaningful opportunity for appeal that is economically feasible even for smaller institutions. As we stated in the NPRM, the NCES Earnings Survey Form, including the survey standards, will be made available for public comment before it is implemented as a part of the approval process under the Paperwork Reduction Act of 1995. At such time, the public will be able to comment on the standards and any associated burden.

NCES fulfills a congressional mandate to collect, collate, analyze, and report complete statistics on the condition of American education and develops statistical guidelines and standards that ensure proper fieldwork and reporting guidelines are followed.  NCES standards are established through an independent process so that outside organizations can rely on these guidelines. Although the standards have not been developed for public review and comment at this time, we are confident that NCES will provide a sufficient methodology under which accurate earnings can be reported and used in calculations for appeals.

To ensure that surveys are conducted in accordance with the standards set for the NCES Earnings Form, we are requiring that institutions submit in connection with a survey-based appeal an attestation engagement report prepared by an independent auditor, certifying that the survey was conducted in accordance with those standards. We note that independent auditor certification is required by section 435(a)(5) of the HEA in a similar context--the presentation of evidence that an institution is achieving academic or placement success for low-income students as proof that an institution’s failing iCDR should not result in loss of title IV, HEA program eligibility. 20 U.S.C. 1085(a)(5). Given NCES’ experience in developing survey standards and this independent auditor requirement, we do not think additional audit or certification requirements are necessary.

Although use of the Earnings Survey Form is not required, we believe use of the form will streamline the process for both the institution and the party preparing the attestation engagement report.

Changes:  None.

Comments:  Several commenters expressed support for the option to base an alternate earnings appeal on earnings data obtained from State-sponsored databases, noting that this option would increase the likelihood that an institution may successfully appeal a program’s D/E rates. One commenter suggested that this option was particularly useful for programs that prepare students for employment in industries where earnings are often underreported. However, another commenter questioned why the Department would include this appeal option given the flaws cited in the NPRM with this approach, such as the potential inaccessibility and incompleteness of these databases.

Discussion: As one commenter noted, and as described in more detail in the NPRM, we believe that there are limitations of State earnings data, notably relating to accessibility and the lack of uniformity in data collected on a State-by-State basis. However, as other commenters noted, the alternate earnings appeal using State earnings data provides institutions with a second appeal option. This option may be useful to those institutions that already have, or may subsequently implement, processes and procedures to access State earnings data. Further, we believe that the matching requirements of the State earnings appeal option will make it more likely that the earnings data on which the appeal is based are reliable and representative of student outcomes.

Changes: None.

Comments: We received a number of comments both in support of, and opposed to, our proposal to allow an institution to submit, for a program that is failing or in the zone under the D/E rates measure, a mitigating circumstances showing regarding the level of borrowing in the program. As proposed in the NPRM, an institution would show that less than 50 percent of all individuals who completed the program during the cohort period, both those individuals who received title IV, HEA program funds and those who did not, incurred any loan debt for enrollment in the program. A GE program that could make this showing successfully would be deemed to pass the D/E rates measure.

Commenters who supported the showing of mitigating circumstances argued that programs for which fewer than 50 percent of individuals enrolled in the program incur debt pose low risk to students and taxpayers. Further, these commenters urged the Department to go beyond a showing of mitigating circumstances and exempt such programs from evaluation under the accountability metrics altogether. A subset of these commenters proposed other requirements that a program would have to meet to qualify for an up-front exemption based on borrowing levels, for example, requiring that tuition and fees are set below the maximum Pell Grant amount. The commenters argued that an up-front exemption for “low risk” programs would lessen the burden on institutions and the Department. These commenters stated that low-cost, open-access institutions serve high numbers of low-income students and generally have the fewest resources to meet new administratively burdensome regulations.  Without up-front relief for these programs, the commenters suggested that many of these institutions would elect to close programs or cease to participate in the title IV, HEA loan programs. 

Other commenters opposed the proposed showing of mitigating circumstances based on borrowing levels. These commenters argued that such a showing, or the related exemption proposed by commenters, would inappropriately favor public institutions. These commenters suggested that, although GE programs offered by public institutions may have lower rates of borrowing, such programs are not necessarily lower cost. Rather, these commenters argued, public institutions, unlike for-profit institutions, benefit from State and local subsidies and do not pay taxes. In this regard, one commenter noted that the showing of mitigating circumstances would result in inequitable treatment among public institutions in different States, where there is varying eligibility for State tuition assistance grants. Another commenter argued that cost--as reflected in a low borrowing rate--should not be the only determinative factor of program quality, as it would permit programs with low completion rates, for example, to remain eligible for title IV, HEA program funds. Other commenters contended that, particularly when only a fraction of programs offered by public institutions would fail the accountability metrics, it would be unjust to include individuals who did not receive title IV, HEA program funds for enrollment in a program in a showing of mitigating circumstances based on borrowing levels when the Department otherwise evaluates GE programs based solely on the outcomes of students who receive title IV, HEA program funds. Some commenters noted that to do so would be at odds with the legal framework established by the Department in order to align the regulations with the court’s interpretation of relevant law in APSCU v. Duncan, 930 F. Supp. 2d at 221, regarding student record systems.

Discussion: As we discuss in detail in “668.401 Scope and Purpose,” in our discussion of the definition of “student,” we do not believe the commenters who supported a “low borrowing” appeal presented a sufficient justification for us to depart from the purpose of the regulations--to evaluate the outcomes of students receiving title IV, HEA program funds and a program’s continuing eligibility to receive title IV, HEA program funds based solely on those outcomes--even for the limited purpose of demonstrating that a program is “low risk.”

We agree with the commenters who suggested that a program for which fewer than 50 percent of individuals borrow is not necessarily low risk to students and taxpayers. Because the proposed showing of mitigating circumstances would be available to large programs with many students, and therefore there may be significant title IV, HEA program funds borrowed for a program, it is not clear that the program poses less risk simply because those students, when considered together with individuals who do not receive title IV, HEA program funds, compose no more than 49 percent of all students. We also note that, if a program is indeed “low cost” or does not have a significant number of borrowers, it is very likely that the program will pass the D/E rates measure.

For these reasons, we do not believe there is adequate justification to depart from the accountability framework established in the proposed regulations, by permitting consideration of the outcomes of individuals other than students who receive title IV, HEA program funds for enrollment in a program in determining whether a program has passed the D/E rates measure. For the same reasons, we do not think there is justification to make an even greater departure from the regulatory framework to allow for an upfront exemption from the accountability framework based on borrowing levels.

We appreciate the commenters’ concerns about administrative burden. As we discuss in more detail in “§668.401 Scope and Purpose,” in preparing these regulations, we have been mindful of the importance of minimizing administrative burden while also serving the important interests behind these regulations.

Changes: We have eliminated from §668.406 the provisions relating to showings of mitigating circumstances.

§668.407 [Reserved] (formerly §668.407 Calculating pCDR)

§668.408 [Reserved] (formerly §668.408 Issuing and Calculating pCDR)

Subpart R

Comments: Some commenters argued that the pCDR measure should take into account only individuals who received title IV, HEA program funds because the focus of the regulations is assessing the likelihood that a program will lead to gainful employment for those students. Others objected to limiting the pCDR measure to these students, other than in a challenge or appeal based on a program’s participation rate index or economically disadvantaged student population, because, according to the commenters, this would produce distorted assessments of program outcomes. These commenters argued that many of the students who receive title IV, HEA program funds are both first-time borrowers and first-generation postsecondary students, who have historically been more likely to default than other borrowers.

Discussion: As discussed in “§668.403 Gainful Employment Program Framework,” we have eliminated the pCDR measure as an accountability metric. However, we have retained program cohort default rate as a possible item on the disclosure template. Accordingly, we do not address the commenters’ concerns in the context of program eligibility. We discuss comments regarding program cohort default rates as a disclosure item in “§668.412 Disclosure Requirements for GE Programs” and “§668.413 Calculating, Issuing, and Challenging Completion Rates, Withdrawal Rates, Repayment Rates, Median Loan Debt, Median Earnings, and Program Cohort Default Rates.” Finally, as discussed in more detail in “§668.401 Scope and Purpose” and “§668.412 Disclosure Requirements for GE Programs,” the information that institutions must disclose about their programs will be based only on the outcomes of students who received title IV, HEA program funds so that students and prospective students who are eligible for title IV, HEA program funds can learn about the outcomes of other students like themselves. We believe that this information will be more useful to these students in deciding where to invest their resources, including, for certain types of title IV, HEA program funds, the limited funds that they may be eligible for, rather than information that is based partly on the outcomes of dissimilar students.

Changes: We have revised the regulations to remove pCDR as a measure for determining program eligibility. We have removed the proposed provisions of §§668.407 and 668.408 and reserved those sections.

§668.409 Final Determination of D/E Rates Measure

Comments: One commenter requested that we synchronize the timing of the D/E rates measure and pCDR measure calculations, notices of determination, and student warning requirements to reduce the complexity of compliance. The commenter proposed that the Secretary issue a single notice of determination that would include a program’s results under both measures.

Discussion: As discussed in “§668.403 Gainful Employment Program Framework,” we have eliminated the pCDR measure as an accountability metric but retained program cohort default rates as a possible item on the disclosure template. Accordingly, there is no reason to synchronize the D/E rates and program cohort default rates calculations because institutions will receive notices of determination under §668.409 with respect to the D/E rates measure only and there will be no student warning requirements tied to pCDR. The Secretary will notify institutions of the draft and official program cohort default rates of their programs, along with related information, under the procedures in §668.413.

Changes: We have revised §668.409 to eliminate references to the pCDR measure.

Comments: One commenter recommended that a notice of determination be issued no later than one year after the Department obtains the data necessary to determine a program’s results under the D/E rates measure. The commenter stated that such a requirement would allow sufficient time for challenges and appeals.

Discussion: The Department will issue a notice of determination under §668.409 when final D/E rates are determined under §§668.404 and 668.405 and, if a program’s D/E rates are recalculated after a successful alternate earnings appeal, under §668.406. It is not clear whether the commenter intended for the one-year time limit to apply to a notice of determination of final D/E rates or recalculated D/E rates. In either case, although we appreciate the concern, we do not believe that a time limit is necessary as the Department will work to issue notices of determination as quickly as possible but in some cases, resolution of an appeal may take longer than one year.

Changes: None.

§668.410 Consequences of the D/E Rates Measure

Comments: Commenters recommended that we eliminate the student warning requirement. They suggested that, if an institution is required to give the student a warning about a program, it would be difficult or impossible to recruit new students and current students would be encouraged to transfer into other programs or withdraw from their program. The commenters argued that, as a result, the student warning requirement effectively undermines the Department’s stated policy of permitting programs time and opportunity to improve. Another commenter proposed eliminating the student warning requirement on the grounds that, as a result of the warnings, States would be burdened with “unwarranted” consumer complaints against institutions from students concerned that their program is about to lose title IV, HEA program eligibility.

On the other hand, some commenters supported the proposed student warning requirements.

Discussion: A student enrolled in a program that loses its title IV, HEA program eligibility because of its D/E rates faces potentially serious consequences. If the program loses eligibility before the student completes the program, the student may need to transfer to an eligible program at the same or another institution to continue to receive title IV, HEA program funds. Even if the program does not lose eligibility before the student completes the program, the student is, nonetheless, enrolled in a program that is failing or consistently resulting in poor student outcomes and could be amassing unmanageable levels of debt. Accordingly, we believe it is essential that students be warned about a program’s potential loss of eligibility based on its D/E rates. The student warning will provide currently enrolled students with important information about program outcomes and the potential effect of those outcomes on the program’s future eligibility for title IV, HEA program funds. This information will also help prospective students make informed decisions about where to pursue their postsecondary education. Some students who receive a warning may decide to transfer to another program or choose not to enroll in such a program. Other students may decide to continue or enroll even after being made aware of the program’s poor performance. In either scenario students will have received the information needed to make an informed decision. We believe that ensuring that students have this information is necessary, even if it may be more difficult for programs that must issue student warnings to attract and retain students. Institutions may mitigate the impact of the warnings on student enrollment by offering meaningful assurances and alternatives to the students who enroll in, or remain enrolled in, a program subject to the student warning requirements.

As a result of the student warning requirements, we expect fewer students will make complaints with State consumer agencies about being misled and enrolling in a program that subsequently loses eligibility. We also believe any additional burden that might be imposed on State agencies due to an increased number of complaints is outweighed by the benefits of providing the warnings.

Changes: None.

Comments: One commenter recommended that we use data regarding GE program performance previously collected by the Department in connection with the 2011 Prior Rule to identify high-risk programs and require those programs to issue student warnings and make other disclosures, effective upon the implementation of the regulations.

Discussion: Although we appreciate the commenter’s interest in providing students with timely information, it is not feasible to implement the commenter’s proposal. In the interest of fairness and due process, we have provided for a challenge and appeals process in the regulations. The 2012 GE informational D/E rates are estimated results intended to inform this rulemaking that were not subject to institutional challenges or appeals. As a result, using these results for accountability purposes would present fairness and due process concerns. In addition, we would be unable to uniformly apply the commenter’s proposal because the Department does not have data for programs that were established after institutions reported information under the 2011 Prior Rule or for those programs that were in existence at that time but for which data were not reported because institutions lacked records for older cohorts, as may be the case with some medical and dental programs.

Changes: None.

Comments: One commenter suggested that an institution should not be required to deliver student warnings as a result of a failing program cohort default rate until the resolution of all related appeals.

Discussion: As discussed in “§668.403 Gainful Employment Program Framework,” we have eliminated the program cohort default rate measure as an accountability metric. Accordingly, the student warning requirements will apply only to programs that may lose eligibility based on their D/E rates for the following award year.

Changes: None.

Comments: Some commenters recommended that institutions be required to issue student warnings whenever a program fails or is in the zone under the D/E rates measure rather than just in the year before a program could become ineligible for title IV, HEA program funds, as provided in the proposed regulations. These commenters reasoned that students and prospective students should be alerted to poor program performance as early as possible.

Other commenters, however, agreed with the Department’s proposal to require student warnings only if a program could become ineligible based upon its next set of final D/E rates. They argued that it would be unfair to require student warnings based on only a single year’s results.

One commenter asserted that it takes a long time to build or rebuild a quality academic program because an institution must develop and maintain courses and curricula and find and retain qualified faculty. According to the commenter, requiring the student warning after one failing or zone result under the D/E rates measure would curtail enrollments, making it difficult to maintain program infrastructure and offerings and resulting in fewer GE programs available to students.

Discussion: We agree with the commenters who argued that students and prospective students should receive a warning when a program may lose eligibility in the following award year based on its D/E rates, rather than at any time the program is not passing under the D/E rates measure. We recognize that requiring an institution to provide the student warning after a program receives D/E rates that are in the zone for the first or second year may adversely affect the institution’s ability to improve the program’s performance. We also appreciate that a program’s D/E rates may be atypical in any given year, and deferring the warning until the program receives a failing rate or a third consecutive zone rate increases the likelihood that the warning is warranted. Until such time as the warning is required, information about the program’s performance under the D/E rates measure will, nonetheless, be available to students and prospective students. The Department will publish the final D/E rates, and a program’s disclosure template may include the annual earnings rates, as well as a host of other critical indicators of program performance.

We recognize that some students who receive a warning about a program may decide to transfer to another program or choose not to enroll in the program. Other students may decide to continue or enroll even after being made aware of the program’s poor performance. In either event, students will have the information necessary to make an informed decision. Further, as discussed in “§668.403 Gainful Employment Framework,” while some programs will be unable to improve, we believe that many will and that institutions with passing programs will expand them or establish new programs. Accordingly, we expect that most students who decide not to enroll or continue in a program will have other viable options to continue their education.

We are making a number of revisions to the proposed text of the student warning. In order to reduce complexity, we are revising §668.410(a) to provide for a single uniform warning for both enrolled and prospective students rather than, as was the case in the proposed regulations, warnings with varying language depending on whether the student is currently enrolled or a prospective student. We are also revising the text of the single warning to make it more broadly applicable, easier to understand, and limited to statements of fact.

First, we are revising the text of the warning to reflect that students to whom the warning is provided may complete their program before a loss of eligibility occurs. Second, we are revising the text to clarify that such a loss of eligibility by the program would affect only those students enrolled at the time a loss of eligibility occurs. Third, because a program loses eligibility if it fails in two out of three consecutive years, we are revising the text of the warning to reflect that a program that has failed the D/E rates measure in one year but passed the D/E rates measure in the following year still faces loss of eligibility based on its D/E rates for the next award year.

To convey a program’s status under the accountability framework to students and prospective students effectively, we are revising the text of the warning so that it is accurate for both current and prospective students, yet succinct and simply worded. We avoid, for example, any explanation as to why a program with D/E rates that are passing in the current year could nevertheless lose eligibility based on rates that are failing in the next year, or why a program that has received no failing D/E rates could lose eligibility based on rates for the next year that are in the zone for the fourth consecutive year. We therefore are revising the text of the warning to describe the current status of the program in a manner that is accurate in all circumstances in which the warning is required: that the program “has not passed” the standards (without identifying whether the statement refers to the current year or the immediately preceding year or years) and that loss of student aid eligibility may occur “if the program does not pass the standards in the future.” Finally, we are revising the text to simply describe the kind of data on which the D/E rates measure is based.

Changes: We have revised §668.410(a) to replace the separate warnings for enrolled students and for prospective students with a single warning for both groups. We have revised the text of the warning to reflect this change and to make the warning more broadly applicable, easier to understand, and limited to statements of fact.

Comments: One commenter contended that, for shorter programs, even if a program becomes ineligible for title IV, HEA program funds in the next year, a student may be able to complete the program without any effect on the student’s ability to continue receiving financial aid. The commenter recommended that in these circumstances, institutions should not be required to give a student warning or should be permitted to revise the content of the warning.

Discussion: We agree that at the time that a student receives the student warnings, loss of access to title IV, HEA program funds will be only a possibility rather than a certain result. Accordingly, as discussed above, we have revised the text of the student warnings to state that if the program does not pass Department standards in the future, “students who are then enrolled may” lose access to title IV, HEA program funds to pay for the program.

Changes: As previously discussed, we have revised §668.410(a) to clarify in the student warning that loss of eligibility may occur in the future, and students then enrolled may lose access to title IV, HEA program funds.

Comments: One commenter asserted that the student warnings in the proposed regulations incorrectly state that programs provide Federal financial aid, when it is the Department that provides title IV, HEA program funds.

Discussion: The commenter is correct that title IV, HEA program funds are not provided by a program.

Changes: We have revised the text of the student warning in §668.410(a) to clarify that title IV, HEA program funds are provided by the Department.

Comments: One commenter recommended that, with respect to warnings to enrolled students, institutions should be required to specify the options that will be available if the program loses its eligibility for title IV, HEA program funds.

Discussion: The proposed regulations required that the warning to enrolled students must:

• Describe the options available to students to continue their education at the institution, or at another institution, in the event that the program loses eligibility for title IV, HEA program funds; and

• Indicate whether the institution will allow students to transfer to another program at the institution; continue to provide instruction in the program to allow students to complete the program; and refund the tuition, fees, and other required charges paid to the institution by, or on behalf of, students for enrollment in the program.

We are revising the regulations to require the warning to enrolled students to include additional details. First, the institution must provide academic and financial information about transfer options available within the institution itself. Because there are often limitations on the transfer of credits from one program to another, institutions must also indicate which course credits would transfer to another program at the institution and whether the students could transfer credits earned in the program to another institution. Finally, we are requiring that all student warnings refer students and prospective students to the Department’s College Navigator or other Federal resource for information about similar programs. With this change, we have eliminated the obligation under proposed §668.410(a)(1)(ii) that the institution research, and advise the student, whether similar programs might be available at other institutions for a student who wishes to complete a program elsewhere.

Changes: We have revised §668.410(a) to require institutions to provide students with information about their available financial and academic options at the institution, which course credits will transfer to another program at the institution, and whether program credits may be transferred to another institution. For these programs we also have eliminated the requirement that institutions describe the options available to students at other institutions and, instead, have required that institutions include in all of their student warnings a reference to College Navigator for information about similar programs.

Comments: One commenter stressed the importance of consumer testing of the content of the student warning and recommended that we develop the text of the warning in coordination with the Consumer Financial Protection Bureau, Federal Trade Commission, and State attorneys general. Another commenter emphasized the importance of including students who are currently attending the programs most likely to be affected in any consumer testing, including students attending programs offered by for-profit institutions.

Discussion: The regulations include text for the student warnings. The Secretary will use consumer testing to inform any modifications to the text that have the potential to improve the warning’s effectiveness. As a part of the consumer testing process, we will seek input from a wide variety of sources, which may include those suggested by the commenter.

Changes: None.

Comments: Some commenters asserted that requiring an institution to give warnings to students and prospective students would violate the institution’s First Amendment rights and particularly its rights relating to commercial speech. These commenters argued that the required warning is not purely factual and uncontroversial, but rather is an ideological statement reflecting a Department bias against the for-profit education industry. Commenters stated that the Department should provide to students and prospective students any such warnings it considers necessary, rather than requiring the institution to do so.

Discussion: We do not agree that it is a violation of an institution’s First Amendment rights to require it to give warnings to students and prospective students. We discuss, first, the commenters’ objections to the content of the required warnings and, next, their objection to the requirement that the institution itself provide the warnings.

As acknowledged by the commenters who objected to the required warnings, these regulations govern commercial speech, which is “expression related solely to the economic interests of the speaker and its audience, . . . speech proposing a commercial transaction”; “material representations about the efficacy, safety, and quality of the advertiser’s product, and other information asserted for the purpose of persuading the public to purchase the product also can qualify as commercial speech.” APSCU v. Duncan, 681 F.3d 427, 455 (D.C. Cir. 2012) (citations omitted). As the commenters also acknowledged, the case law recognizes that the government may regulate commercial speech, and that different tests apply depending on whether the government prohibits commercial speech or, as is the case with these regulations, merely requires disclosures.¹⁶⁹

Courts have required that laws regulating commercial speech must directly advance a significant government interest and must do so in a manner narrowly tailored to that goal. Central Hudson Gas and Elec. Corp. v. Public Service Comm’n of N.Y., 447 U.S. 557, 564 (1980).

A government requirement that parties disclose “accurate, factual commercial information” does not violate the First Amendment if the requirement is “reasonably related” to a significant government interest, including not merely “preventing deception,” but other significant interests as well. Am. Meat Inst. v. U.S. Dep't of Agric., 76 F.3d 18(D.C. Cir. 2014). In the context of gainful employment programs, as discussed in the NPRM, the government does indeed have an interest in preventing deceptive advertising. Advertising that a service provides a benefit “without alerting consumers to its potential cost . . . is adequate to establish that the likelihood of deception . . . ‘is hardly a speculative one.’” Milavetz, Gallop & Milavetz, P.A. v. United States, 559 U.S. 229, 251 (2010) (quoting Zauderer, 471 U.S. at 652). However, the government has an interest in not just preventing deception, but an affirmative interest in providing consumers information about an institution’s educational benefits and the outcomes of its programs. This interest is well within the range of interests that justify requiring a regulated entity to make disclosures about its products or services. See Am. Meat Inst., 760 F.3d at 27.

The warnings will provide consumers with information of the kind that Congress has already determined necessary to make an “informed judgment about the educational benefits available at a given institution.” Pub. L. 101–542, §102, November 8, 1990, 104 Stat. 2381. Moreover, the government’s continued interest over time in disclosures of this nature evidence the significance of its interest. See Am. Meat Inst., 760 F.3d at 23-24.

The particular warnings in these regulations are new, but, for more than thirty years, Congress has required institutions that receive title IV, HEA program funds to make numerous disclosures to current and prospective students akin to the disclosures required under these regulations.¹⁷⁰ The statutory disclosure requirements were first enacted in 1980 and have been expanded repeatedly since then, most recently in 2013. The warning requirements in these regulations are based on the same Federal interest in consumer disclosures demonstrated over these past decades, demonstrating that the interest underlying these regulations is a significant governmental interest.

Courts have found that the requirement that the disclosure is “narrowly tailored” to the governmental interest is “self-evidently satisfied” when the government requires an entity to “disclose ‘purely factual and uncontroversial information’ about attributes of the product or service being offered.” Am. Meat Inst., 760 F.3d at 26 (citation omitted). The commenters contended that the required warnings and disclosures are not “factual, uncontroversial information” and noted that the court in APSCU v. Duncan indicated doubt that the language of the warning required under the 2011 Prior Rule would meet that test. APSCU v. Duncan, 870 F.Supp.2d at 155 n.7. They contended that the text of the warning proposed in §668.410(a) is similarly flawed.

We do not agree that the text of the proposed warning was not factual and uncontroversial. However, as discussed in this section, we have made a number of revisions to the proposed student warning text, and, accordingly, we consider here whether the student warning text in the final regulations is factual and not controversial.

The text of the student warning contains a mixture of fact and explanation. The purely factual component--that “this program has not passed standards established by the Department”--is not controversial at the time the warning is required because institutions will have had an opportunity to challenge or appeal the Department’s calculation of the relevant data.¹⁷¹ Similarly, the statement that “if in the future the program does not pass the standards, students who are then enrolled may not be able to use federal student grants or loans to pay for the program” and may have to find other ways to pay for the program is simply a statement of what might happen if a program does not meet the standards and cannot be considered inaccurate or controversial. The remainder of the warning text in the final regulations--which states that the Department based these standards on the amounts students borrow for enrollment in the program and their reported earnings--is also a factual statement. No part of the student warning text conveys an ideological message or bias against for-profit institutions, given that all GE programs, whether they are offered by for-profit institutions or by public institutions, must provide the warnings in accordance with the regulations, and the warnings are composed solely of factual statements.

In response to comments contending that the Department--rather than the institution--should issue warnings to the consumer on a Department Web site, such as College Navigator, or by direct mailings, we note that existing HEA disclosure requirements are based on congressional findings that having the institution disclose “timely and accurate data is essential to any successful student assistance system.” H. R. Rep. No. 733, 96th Cong., 2d Sess. (1980) at 52.¹⁷² These regulations similarly require the institution to disclose through the student warning the potential significance of a program’s D/E rates. The mandate that institutions deliver the message on their Web sites is tailored to deliver the message in an effective manner, and the content of the message is tailored to provide the kind of information that consumers need to evaluate an individual program that the institution promotes as preparing students for gainful employment.

Although the Department can post warnings for hundreds or even thousands of GE programs on a Department Web site, we do not consider such posting to be an effective means of reaching consumers. We note that Congress has reached the same conclusion by requiring that institutions make numerous disclosures not only in their publications but, more recently, through “electronic media,” 20 U.S.C. 1092(a)(1), a term already interpreted by the Department to include posting on Internet Web sites, 34 CFR 668.41(b), and posting to the institution’s Web site, 20 U.S.C. 1015a(h)(3) (net price calculator). These statutory requirements demonstrate a congressional determination that disclosure to the consumer by the institution itself is necessary to achieve the Federal objective of enabling consumers to make “informed choices.”¹⁷³ Because the student warnings required by these regulations target a similar and often identical audience as the disclosures already required by the HEA, we believe the congressional mandate provides a sound basis for requiring institutions themselves to make the warnings in order to achieve the purpose of the regulations.

The regulations require an institution to provide the warnings not only by including the warning on its Web site, but by delivering the warning directly to the consumer. The latter method is also tailored to the objective of giving effective and timely information. This is not the first instance in which regulations have required this kind of individual, direct communication by institutions with consumers about Federal aid: section 454(a)(2) of the HEA authorizes the Department to require institutions to make disclosures of information about Direct Loans, and Direct Loan regulations require detailed explanations of terms and conditions that apply to borrowing and repaying Direct Loans. The institution must provide this information in “loan counseling” given to every new Direct Loan borrower in an in-person entrance counseling session, on a separate form that must be signed and returned to the institution by the borrower, or by online or electronic delivery that assures borrower acknowledgement of receipt of the message. 34 CFR 685.304(a)(3).¹⁷⁴ The requirement in those regulations closely resembles the requirements here that the institution provide the warnings directly to the affected consumers.

Although we carefully considered the commenters’ concerns, we do not believe that there are any First Amendment issues raised by the student warning requirements in the final regulations. Further, we weighed the concerns against the significant government interest in providing consumers an effective warning regarding a program’s performance and eligibility status. In this situation, failure to disclose the potential for loss of eligibility and the consequences of that loss could be misleading and this information is critical to the informed educational decision making of students and prospective students.

Changes: None.

Comments: We received a number of comments about when student warnings must be delivered to prospective students and who constitutes a “prospective student.” First, commenters expressed concern that institutional obligations with respect to prospective students were unclear. As discussed under “§668.401 Scope and Purpose,” commenters were confused about when an individual would be considered a “prospective student” for the purpose of the student warning requirements and when student warnings were first and subsequently required to be given to prospective students. In this regard, commenters recommended that, to avoid undue administrative burden and compliance challenges, we eliminate the requirement that institutions provide student warnings upon first contact with a prospective student, given that student warnings are required before execution of an enrollment agreement and in connection with promotional materials. Commenters also expressed concern that the burden associated with giving repeated warnings may outweigh the benefits. Along these lines, some commenters recommended that we conduct consumer testing to determine the point at which student warnings would be most meaningful to prospective students.

As discussed in “§668.401 Scope and Purpose,” some commenters recommended that student warnings be given not just to “prospective students” as defined in the proposed regulations, but also to family members, counselors, and others making enrollment inquiries on their behalf.

Discussion: We agree that the proposed regulations were not clear about how the definition of “prospective student” and the student warning requirements interacted. As discussed under “§668.401 Scope and Purpose,” we have narrowed the definition of “prospective student.” However, we agree with the commenter that a third party who makes the first contact with an institution, such as a parent or counselor, may play a significant advisory role in the educational decision-making process for a prospective student. That individual should be given the student warning to convey to the student and we are revising the regulations accordingly. With these changes, we believe that it will be clear when and to whom student warnings must be delivered.

For prospective students, we continue to believe that student warnings should be required both upon first contact and prior to enrollment. Although there will be situations in which contact is first made and a prospective student indicates his or her intent to enroll within a relatively short period of time after that, we believe that any redundancy in requiring delivery of the student warnings at both of these junctures is outweighed by the value in ensuring prospective students have this critical program information at times when they may most benefit from it.

Changes: We have clarified in §668.410(a)(6)(i) (§668.410(a)(2)(i) in the proposed regulations) that first contact about enrollment in a program, triggering the obligation to deliver the student warning, may be between a prospective student and a third party acting on behalf of an institution. We have also clarified in the definition of “prospective student” in §668.402 that such first contact may be between a third party acting on behalf of a prospective student and an institution or its agent.

Comments: Some commenters were concerned about the manner in which student warnings may be delivered to students and prospective students. With respect to enrolled students, commenters expressed concern that institutions would bury the warning in a lot of other information to lessen the warning’s impact. These commenters believed that the permitted methods of delivery--hand-delivery, group presentations, and electronic mail--allow for institutional abuse. They suggested that the Department be more specific about the permitted methods of delivery, consider other ways in which student warnings could be delivered--for example, requiring posted warnings in classrooms and financial aid offices--and use consumer testing to determine the most effective means of delivery and format. One commenter recommended that we require institutions to obtain student acknowledgement of receipt of the warning.

Other commenters recommended changes to the student warning requirements to lessen institutional burden and give institutions more flexibility. Some of these commenters conflated the student warning and the disclosure template delivery requirements. One commenter noted their differences and requested that we collapse the requirements into a single requirement. For example, the proposed regulations require institutions to obtain written confirmation that a prospective student received a copy of the disclosure template; as noted by another commenter, there was no such requirement with respect to the student warning. Some commenters recommended that email confirmation that students have received the student warning should satisfy the student warning requirements. One commenter suggested that an institution should be able to meet the student warning requirements by delivering the disclosure template that includes the student warning to a prospective student as required under §668.412. One commenter was unsure how institutions would deliver the required written student warning to prospective students who contact the institution by telephone about enrollment in a program, and one commenter proposed that oral warnings be permitted.

Discussion: We agree with the commenter who suggested the Department should more clearly specify the manner in which student warnings may be delivered. To that end, we indicate in the final regulations the permitted methods of delivery of a student warning to each of: (1) enrolled students, (2) prospective students upon first contact, and (3) prospective students prior to entering into an enrollment agreement.

For enrolled students, as in the proposed regulations, the regulations permit delivery of the student warning in writing by hand-delivery or by email. To ensure that the student warning is prominently displayed, and not lost within an abundance of other information, we are revising the regulations to clarify that any warning delivered by hand must be delivered as a separate document, as opposed to one page in a longer document; and any warning delivered by email must be the only substantive content of the email. We recognize that student warnings delivered by email may go unread by students and that there is a significant benefit to taking steps to help ensure that warnings delivered by email are actually read by the students. Accordingly, as suggested by a commenter, we are revising §668.410(a) to require that, for a warning delivered by email, an institution must send the email to the primary email address used by the institution for communicating with the student about the program, and receive electronic or other written acknowledgement that the student has received the email. If an institution receives a response indicating the email could not be delivered, the attempted delivery is not enough to meet the requirement in the regulations, and the institution must send the information using a different address or method of delivery. An institution may satisfy the acknowledgement requirement through a variety of methods such as a pop-up window that requires students to acknowledge that they received the warning. Institutions must maintain records of their efforts to deliver the warnings required under the regulations. We believe that the burden on institutions to obtain this acknowledgement is outweighed by the increased likelihood that in the course of, or as a result of, acknowledging receipt, students will read the warning and take it into account when making educational and financial decisions. We note that the requirement to obtain this kind of acknowledgement is no more burdensome than the requirement that institutions do so with regard to entrance counseling requirements. See section 485(l)(2) of the HEA (20 USC 1092(l)(2)); 34 CFR 682.604(f)(3); 34 CFR 685.304(a)(3)(ii)-(iii) (requiring written or electronic receipt acknowledgment).

For the requirement that an institution or its agent provide the student warning upon first contact with a prospective student or a third party acting on behalf of a prospective student, we are clarifying that the warning may be delivered in the same manner as the warning is delivered to enrolled students--by hand-delivery or by email--in accordance with the same requirements that apply to the delivery of warnings to enrolled students. As proposed by a commenter, we are revising the student warning and disclosure template delivery requirements relating to prospective students to permit an institution to deliver the disclosure template with the student warning. In this regard, we are moving the requirement that an institution update its disclosure template to include the student warning from §668.412 to §668.410(a)(7) in order to consolidate all of the requirements related to student warnings in one section of the regulations, although we continue to reference this requirement in §668.412.

We recognize that the first contact between an institution or its agent and a prospective student or a third party acting on the prospective student’s behalf may be made by telephone. Although we continue to believe that a written warning is more effective than an oral warning, given that a prospective student will receive the student warning in writing prior to entering into an enrollment agreement, we are revising the regulations to permit an oral warning in these circumstances to lessen administrative burden for institutions, while at the same time ensuring that prospective students receive important information at a critical time in their decision-making process.

For the student warning that must be delivered to a prospective student at least three, but not more than 30, days prior to entering into an enrollment agreement, we are clarifying that all the written methods of delivery permitted for student warnings upon first contact--but not oral delivery--are also permitted in this circumstance. In this regard, we note that, in requiring that a written warning delivered by hand be in a separate document, an institution may not build the student warning into an enrollment or similar agreement where the information could be easily overlooked.

We believe that direct delivery of the warning to students and prospective students makes it most likely that students receive it and review it. While we encourage institutions to post the student warning in classrooms and financial aid offices, institutions will not be required to do so as it is unclear whether the additional benefits of this beyond the other delivery requirements would outweigh the added burden.

As suggested by a commenter, we intend to solicit feedback on the most effective delivery methods through consumer testing.

Changes: We have clarified the methods by which an institution may deliver the required warnings to students and prospective students in §668.410(a)(5) and (a)(6). In §668.410(a)(5), we have added the requirement that student warnings that are hand-delivered must be provided in a separate document. We have also required that student warnings that are delivered by email must be the only substantive content of the email and the institution must receive an electronic or other written acknowledgement from the student that the student received the warning. In addition, we have specified that if an institution receives a response that the email could not be delivered, the institution must use a different address or mode of delivery. Finally, the regulations have been revised to require that an institution maintain records of its efforts to deliver the warning.

In §668.410(a)(6), we have clarified that the methods of delivery specified for enrolled students, as revised, also apply to prospective students, and we have provided that student warnings may be delivered to a prospective student by providing the prospective student a disclosure template that has been updated to include the student warning. The same requirements with respect to email delivery and acknowledgment of receipt that apply to the warnings to enrolled students will also apply to warnings delivered to prospective students or a third party acting on behalf of the prospective student.

We also have revised §668.410(a) to specify that an institution may deliver any required warning orally to a prospective student or third party except in the case of a warning that is required to be given before a prospective student enrolls in, registers, or makes a financial commitment with respect to a program.

Comments: Some commenters contended that the requirement that student warnings be provided to the extent practicable in languages other than English for students for whom English is not their first language is unclear because the requirement does not indicate how a school would determine whether English is the first language of a student.

Discussion: Section 668.410(a)(4) (§668.410(a)(3) in the proposed regulations) requires that an institution provide, “if practicable,” “alternatives to English-language warnings” to those prospective students and currently enrolled students for whom English is not their first language. This requirement is not unconstitutionally vague. There are many ways in which an institution could practicably identify individuals for whom English may not be their first language. However, we note one simple test generally applicable to consumer transactions that could be used by institutions in determining whether alternatives to non-English warnings are warranted. That test is whether the language principally used in marketing and recruiting for the program was a language other than English.¹⁷⁵ Where institutional records show that a student responded to an advertisement in a language other than English, or was recruited by an institutional representation in an oral presentation conducted in a language other than English, an institution may readily and practicably identify that student or prospective student as one whose first language is not English. Other methods might also be practicable, but institutions should at a minimum already be familiar with their obligations when they advertise in languages other than English. In addition, institutions should be mindful that Federal civil rights laws (including title VI of the Civil Rights Act of 1964) require institutions to take appropriate measures to ensure that all segments of its community, including those with limited English proficiency, have meaningful access to all their programs and all vital information.

Changes: None.

Comments: With respect to the provision in proposed §668.412(b)(2) that would require institutions to update a program’s disclosure template to include the student warning, one commenter requested that institutions have 90 days from receipt of notice from the Secretary that student warnings are required to make the update, rather than 30 days as provided in the regulations.

Discussion: Because the student warning will include critical information that students will need to consider as a part of their educational and financial decision making, we believe that the student warning must be conveyed as quickly as possible once it has been determined that the program could become ineligible based on its D/E rates in the next award year. As the Department will provide the text of the warning, and institutions should already be aware of or have ready access to any required additional information, we believe that 30 days is a reasonable amount of time to update the disclosure template with the warning. Any burden that institutions might face in meeting this requirement is outweighed by the necessity that students receive this important information as promptly as possible.

Changes: We have moved the requirement that institutions update their disclosure templates to include any required student warning from §668.412(b)(2) to §668.410(a)(7), so that all of the requirements with respect to student warnings are in one place for the reader’s convenience.

Comments: Several commenters opposed the provision prohibiting an institution from enrolling a prospective student before expiration of a three-day period following delivery of a required student warning. The commenters argued that students are intelligent consumers who do not require a cooling-off period and that the provision is designed to discourage prospective students from enrolling by making enrollment inconvenient. For the same reasons, one of the commenters asked that, if the Department retains the cooling-off period in the final regulations, it eliminate the requirement that a student warning be provided anew before a prospective student may be enrolled, if more than 30 days have passed since the student warning was last given.

Discussion: There is evidence that some institutions use high-pressure sales tactics that make it difficult for prospective students to make informed enrollment decisions.¹⁷⁶ We believe that the three-day cooling-off period provided for in §668.410(a)(6)(ii) (§668.410(a)(2)(ii) of the proposed regulations) strikes the right balance between allowing sufficient time for prospective students to consider their educational and financial options outside of a potentially coercive environment, while ensuring that those prospective students who have had the opportunity to make an informed decision can enroll without having to wait an unreasonable amount of time. We further believe that students are more likely to factor the information contained in the student warning into their financial and educational decisions if the warning is delivered when the student is in the process of making an enrollment decision. We believe 30 days is a reasonable window before a student warning must be reissued.

Changes: None.

Comments: Many commenters stated that GE programs that do not pass the D/E rates measure should be subject to limits on their enrollment of students who receive title IV, HEA program funds. Commenters variously proposed that we limit enrollment of students who receive title IV, HEA program funds to the number of students enrolled in the program in the previous year or to an average enrollment of students receiving title IV, HEA program funds over the previous three years. These commenters argued that enrollment limits would provide institutions with the incentive to improve programs more quickly and limit the potential risks to students and taxpayers. According to these commenters, disclosures and student warnings do not provide sufficient protection for students and will not stop an institution from increasing the enrollment of a poorly performing program to maximize title IV, HEA program funds received before the program loses eligibility, at significant cost to students, taxpayers, and the Federal government.

We also received a number of comments opposing limits on enrollment for programs that do not pass the D/E rates measure. These commenters asserted that disclosures and student warnings are sufficient to provide students with the information they need to make their own educational decisions. One commenter cited economic theory as supporting the proposition that, if parties are fully informed, imposing quotas or limitations creates market inefficiencies. This commenter asked that we consider the costs to students who are not permitted to enroll in a program and compare those costs to the assumed benefits of not enrolling in a program that may or may not become ineligible. The commenters argued that enrollment limits would significantly hinder efforts by institutions to improve programs and could lead to the premature closing of programs.

Discussion: We agree that it is important to protect students from enrolling in poorly performing programs and to protect the Federal investment in GE programs. However, we believe that the accountability framework, in which the D/E rates measure is used to determine a program’s continuing eligibility for title IV, HEA program funds, adequately safeguards the Federal investment and students, while allowing GE programs the opportunity to improve. Further, we believe that the warnings to students and prospective students about programs that could become ineligible based on their D/E rates for the next award year, and the required disclosures, are meaningful protections that will enable students and their families to make informed decisions.

Changes: None.

Comments: Several commenters suggested that institutions should have the opportunity to pay down the debt of students and provide the students some relief while, at the same time, improving program performance under the accountability metrics. These commenters argued that a voluntary loan reduction plan would permit institutions a greater measure of control over program performance under the accountability metrics and benefit students, particularly those students who withdraw from, or fail, a program early in the program. The commenters proposed a number of specific terms for such a loan reduction plan, including giving institutions flexibility to determine the amount of institutional grants to be used to pay down student debt.

Discussion: We acknowledge the desire to ease the debt burden of students attending programs that become ineligible and to shift the risk to the institutions that are enrolling students in these programs. We also recognize that the loan reduction plan proposal would give institutions with the funds to institute such a program a greater measure of control over their performance under the D/E rates measure. However, as stated in the NPRM, the discussions among the negotiators made it clear that these issues are extremely complex, raising questions such as the extent to which relief would be provided, what cohort of students would receive relief, and whether the proposals made by negotiators would be sufficient. The comments we received confirm that this issue requires further consideration. Accordingly, the Department is not addressing these concerns in the final regulations, and will continue to explore ways to provide debt relief to students in future regulations.

Changes: None.

Comments: Many commenters urged the Department to directly offer debt relief to students enrolled in programs that lose eligibility for title IV, HEA program funds under the GE regulations, as well as to students enrolled in programs that are not passing under the D/E rates measure, so that students are not burdened with sole responsibility for debts accumulated at programs that did not prepare them for employment in their respective fields. They argued that affected students should be “made whole” through discharges of their title IV, HEA program loans from the Department and reinstatement of their lost Pell Grant eligibility. The Department, the commenters said, could then pursue from the institutions collection of the discharged funds. They reasoned that such relief would be fair to students, provide institutions with incentive for improvement, and reallocate risk from students to institutions, which are in a better position to assume it. The commenters asserted that students should not be subject to potentially severe financial consequences from borrowing title IV, HEA program funds to attend programs that the Department permitted to operate with its approval, despite not achieving program outcomes deemed acceptable under the D/E rates measure. According to the commenters, provisions for borrower relief would allow affected students to pursue educational opportunities that offered value, and institutions would be held accountable for the costs to taxpayers of poorly spent title IV, HEA program funds.

One commenter contended that, in the context of borrower relief, the Department was placing undue emphasis on supporting institutions and avoiding litigation, and not enough emphasis on protecting students and their families. The commenter proposed that the Department could phase in borrower relief for students over the transition period, with programs not passing the D/E rates measure subject only to student warnings in the first year after implementation of the regulations and enrollment limits and borrower relief provisions taking effect in subsequent years of the transition period.

Many of the commenters who supported full debt relief for borrowers in affected programs requested that, if full relief is not possible, student borrowers be provided partial relief, in the form presented by the Department during the negotiated rulemaking sessions where an institution with a program facing ineligibility in the next year would be required to make available to the Department, for example, through a letter of credit, sufficient funds to reduce the debt burden of students who attended the program during that year if the program became ineligible.

We also received general comments opposing any borrower relief provisions in the regulations.

Discussion: The Department acknowledges the concern that borrowers attending programs that are determined ineligible will remain responsible for the debt they accumulated. However, as explained in the NPRM, none of the circumstances under which the Department has the authority to discharge title IV, HEA loans under the HEA as a result of ineligibility are applicable to these regulations. 20 U.S.C. 1087(c)(1). This discharge authority does not extend to loans obtained by borrowers who met properly administered admission standards for enrollment in a program or at an institution that was not eligible.¹⁷⁷ We also acknowledge the commenters’ interest in excluding those periods in which a student may have received a Pell Grant for attendance at a GE program that did not pass the D/E rates measure from limits otherwise applicable to Pell Grant eligibility. However, section 401(c)(5) of the HEA provides that the period during which a student may receive Federal Pell Grants “shall not exceed 12 semesters.” 20 U.S.C. §1070a(c)(5). We read this provision as leaving the Department no authority to exclude specific time periods from that limit.

With respect to the other borrower relief proposals that commenters offered, as we have previously stated, these proposals raise important but complex issues that the Department will continue to consider outside of this rulemaking.

Changes: None.

Comments: One commenter recommended that we revise §668.410(b)(1), which generally prohibits disbursement of title IV, HEA program funds to a student enrolled in a program that has lost eligibility under the regulations, to permit disbursement of such funds until the student completes the program.

Discussion: We decline to adopt the commenter’s proposal. A GE program’s loss of eligibility is effective, under 34 CFR §668.409(b), on the date specified in the notice of final determination. Section 668.410(b)(1) adopts by explicit reference the general rule in §668.26(d), which the Department applies in all instances in which an institution’s participation in the title IV, HEA programs ends. Section 668.26(d)(1), consistent with §600.41(d), provides that after a GE program loses eligibility, an institution may make no new commitments for title IV, HEA program funds, but may fund the remainder of certain commitments of grant and loan aid. These provisions apply the loss of eligibility to students then enrolled in the program in a way that modestly defers the effect of that loss as it affects their ability to meet their financial commitments and provides some time to make alternative arrangements or transition to another program or institution. Students may therefore continue to receive title IV, HEA program funds for attendance at a program that has lost eligibility through the end of any ongoing loan period or payment period, which periods could include a full award year.¹⁷⁸ Even if we were to interpret the HEA to permit extending the period during which students could receive title IV, HEA program funds to attend an ineligible program beyond these long-established limits, we see no valid reason to do so. To further extend the period during which students may continue to receive title IV, HEA program funds to attend an ineligible program would encourage students to invest more time, money, and limited Pell Grant eligibility in programs that produce unacceptable student outcomes. The commenter offers no reason to treat a loss of eligibility under these regulations differently than any other loss of eligibility, and we see none.

Changes: None.

Comments: One commenter suggested that we revise §668.410(b)(2), which provides for a three-year period of ineligibility for programs that are failing or in the zone and that are voluntarily discontinued, to more clearly indicate when the period of ineligibility begins and ends. The commenter recommended revisions based on language in the iCDR regulations in 34 CFR 668.206.

Discussion: We appreciate the commenter’s suggestion and are revising the provision to indicate more clearly when the three-year period of ineligibility begins.

Changes: We have revised §668.410(b)(2) to clarify that the three-year period of ineligibility begins, as applicable, on the date specified in the notice of determination informing the institution of a program’s ineligibility or on the date the institution discontinued a failing or zone program.

Comments: One commenter suggested that we revise §668.410(b)(2) and (b)(3), which provide for a three-year period of ineligibility for programs that are failing or in the zone and that are voluntarily discontinued, to capture programs that are voluntarily discontinued after the institution receives draft D/E rates that would be failing or in the zone if they were final. In such cases, the commenter recommended that the Department should, despite the program’s discontinuance, calculate its final D/E rates and, if those final D/E rates are failing or in the zone, impose the three-year ineligibility period as provided in §668.410(b)(2) on that program and any substantially similar programs. The commenter suggested that, without the proposed revision, there would be a “loophole” that institutions could exploit to avoid the three-year ineligibility period.

Discussion: We agree with the commenter that we should not permit an institution to avoid the three-year ineligibility period by discontinuing a poorly performing program after the issuance of draft D/E rates that are failing or in the zone, but before the issuance of final D/E rates. Accordingly, the final regulations provide that, if an institution discontinues a program after receiving draft D/E rates that are failing or in the zone, the institution may not seek to reestablish that program, or establish a substantially similar program, until final D/E rates have been issued for that program, and only then if the final D/E rates are passing or the three-year period of ineligibility has expired. In the event there is a three-year period of ineligibility that is triggered by the final D/E rates, the period will begin on the date that the program was discontinued, and not the date the final D/E rates were issued, so that the ineligibility period is no longer than the three years that would apply to any other zone or failing program that is voluntarily discontinued.

Changes: We have revised §668.410(b)(2) to provide that a program that was discontinued after receiving draft D/E rates that are failing or in the zone, but before receiving final D/E rates, is ineligible, and the institution may not seek to establish a substantially similar program, unless the program’s final D/E rates are determined to be passing or, if its final D/E rates are also failing or in the zone, the three-year ineligibility period, dating from the institution’s discontinuance of the program, has expired. We also have revised this section to clarify that the provision regarding determination of the date a program is voluntarily discontinued applies to programs discontinued before their final D/E rates are issued.

Comments: We received a number of comments about the definition of “substantially similar” programs and the limitations on an institution’s ability to start a program that is substantially similar to an ineligible program.

Several commenters expressed concern that the definition of “substantially similar” is not broad enough to capture all of the similar programs that an institution may seek to establish in the place of a poorly performing program in order to avoid accountability. These commenters said that the definition should not require that programs share the same credential level in order to be considered substantially similar. These commenters were concerned that, for example, an institution could simply convert an ineligible certificate program into a new associate degree program, without complying with the three-year ineligibility period and taking any action to improve the program. Similarly, commenters were also concerned that the requirement that substantially similar programs share the first four digits of a CIP code is too narrow. They argued that there is sufficient overlap between four-digit CIP codes such that institutions could avoid the restriction on establishing a program that is substantially similar to a program that became ineligible within the most recent three years by using another four-digit CIP code that aligns with the same curriculum. These commenters suggested that we define programs as “substantially similar” if they share the same two-digit CIP codes. Alternatively, the commenters recommended that the Department evaluate on a case-by-case basis whether programs with the same two-digit CIP code are substantially similar, and require documentation that a new program within the same two-digit CIP code will meet the D/E rates measure.

Other commenters suggested that we treat programs as substantially similar only if they share the same four-digit CIP code and credential level. These commenters also recommended that we permit the establishment of programs that are substantially similar to an ineligible program if the institution has other substantially similar programs that are passing the D/E rates measure. For example, the commenters explained, if an institution offers multiple substantially similar programs and at least 50 percent of those programs are passing the D/E rates measure, an institution would be permitted to establish a substantially similar program.

Discussion: We agree with the commenters who recommended that programs should not be required to share the same credential level in order to be considered substantially similar and that a definition of “substantially similar” that considers credential level would permit institutions to avoid accountability by changing program length.

However, we do not agree that the definition of substantially similar should be broadened to encompass all programs within a two-digit CIP code as substantially similar or that it is necessary to establish a process to evaluate for each new program whether the assigned four-digit CIP code best represents the program content. We are removing the phrase “substantially similar” from the definition of CIP code and establishing in §668.410 that two programs are substantially similar to one another if they share the same four-digit CIP code. Institutions may not establish a new program that shares the same four-digit CIP code as a program that became ineligible or was voluntarily discontinued when it was in the zone or failing within the last three years. An institution may establish a new program with a different four-digit CIP code that is not substantially similar to an ineligible or discontinued program, and provide an explanation of how the new program is different when it submits the certification for the new program. We presume based on that submission that the new program is not substantially similar to the ineligible or discontinued program, but the information may be reviewed on a case by case basis to ensure a new program is not substantially similar to the other program.

We believe that these revisions strike an appropriate balance between preventing institutions from closing and restarting a poorly performing program to avoid accountability and ensuring that institutions are not prevented from establishing different programs to provide training in fields where there is demand.

We believe that it is appropriate to require an institution that is establishing a new program to provide a certification under §668.414 that includes an explanation of how the new program is not substantially similar to each program offered by the institution that, in the prior three years, became ineligible under the regulations’ accountability provisions or was voluntarily discontinued by the institution when the program was failing, or in the zone with respect to, the D/E rates measure. We also discuss this change in “§668.414 Certification Requirements for GE Programs.”

Changes: We have revised §668.410 to provide that a program is substantially similar to another program if the programs share the first four digits of a CIP code. We also have revised this section to provide that the Secretary presumes a program is not substantially similar to another program if the programs do not share a four-digit CIP code. The institution must submit an explanation of how the new program is not substantially similar to the ineligible or voluntarily discontinued program. In §668.410(b)(3), we have also corrected the reference to §668.414(b) to §668.414(c).

§668.411 Reporting Requirements for GE Programs

Comments: Numerous commenters asserted that institutions with low borrowing rates or low cohort default rates should be exempt from the reporting requirements, arguing that such programs do not pose a high risk to students or taxpayers. For example, some commenters recommended exempting a program from the reporting requirements where an institution certifies that: (1) less than fifty percent of the students in the program took out loans for the two most recent academic years, (2) fewer than 20 students receiving title IV, HEA program funds completed the program during the most recent two academic years, and (3) the default rate falls below a reasonable threshold for two consecutive years. These commenters proposed that a program should be subject to the reporting requirements for a minimum of two years at the point that it does not meet one of these three exemption requirements for two consecutive years. Other commenters proposed variations of this approach, such as exempting from the reporting requirements institutions with an institutional cohort default rate of less than fifteen percent. Similarly, one commenter said that foreign schools should be exempt from the reporting requirements, asserting that certificate programs at foreign institutions are of low risk to American taxpayers since those programs have relatively few American students compared to the entire enrollment in the program.

Discussion: We do not agree that a program, foreign or domestic, should be exempt from the reporting requirements because it has a low borrowing rate, low institutional cohort default rate, or low number of students who receive title IV, HEA program funds. The information that institutions must report is necessary to calculate the D/E rates and to calculate or determine many of the disclosure items as provided in §668.413. (See “§668.413 Calculating, Issuing, and Challenging Completion Rates, Withdrawal Rates, Repayment Rates, Median Loan Debt, Median Earnings, and Program Cohort Default Rates” for a discussion of the disclosure items that the Department will calculate.) Exempting some institutions from the reporting requirements, whether partially or fully, would undermine the effectiveness of both the accountability and transparency frameworks of the regulations because the Department would be unable to assess the outcomes of many programs. In addition, students would not be able to access relevant information about these programs and compare outcomes across multiple metrics. Further, a policy that allowed exemptions from reporting, accountability, and transparency, regardless of the basis, in some years but not others would be impossible to implement. Without consistent annual reporting, the Department would, in many cases, be unable to calculate the D/E rates or disclosures in non-exempted years as these calculations require data from prior years when the exemption may have applied.

Changes: None.

Comments: A few commenters recommended requiring institutions to report additional items to the Department. Specifically, some commenters argued that the Department should collect and make public job placement rates to enable the Department, States, researchers, and consumers to easily access this information to compare programs at different schools. The commenters also asserted that requiring institutions to report these rates at the student level would improve compliance at institutions that are currently required to calculate job placement rates but do not do so.

Other commenters recommended that institutions be required to report the SOC codes associated with their programs. These commenters disagreed with the Department’s assertion in the NPRM that it would not be appropriate to collect SOC codes at the student level. They argued that requiring institutions to report the SOC codes that they must disclose under §668.412 would strengthen the Department’s ability to monitor whether programs have the necessary accreditation or other requirements for State licensing and would support more accurate and realistic disclosure of the SOC codes associated with a program’s CIP code.

Discussion: We agree that allowing the Department, States, researchers, and consumers to access job placement information will be beneficial. Accordingly, we are adding a requirement for institutions to report job placement rates at the program level if the institution is required by its accrediting agency or State to calculate a placement rate for either the institution or the program using the State’s or agency’s required methodology and to report the name of the State or accrediting agency. For additional information about job placement rates, see the discussion under “§668.412 Disclosure Requirements for GE Programs.” While all other required reporting for the initial reporting period must be made by July 31, 2015, due to operational issues, institutions will report job placement rates at a later date and in such manner as prescribed by the Secretary in a notice published in the Federal Register.

The Department already identifies SOC codes for GE programs as part of each institution’s PPA. We will continue to consider requirements for updating and monitoring SOC codes to improve oversight while limiting the reporting burden on institutions.

Changes: We have added a requirement in §668.411(a)(3) that institutions must report to the Department a placement rate for each GE program, if the institution is required by its accrediting agency or State to calculate a placement rate for either the institution or the program, or both, using the methodology required by that accrediting agency or State, and the name of that accrediting agency or State. We have also renumbered the paragraphs that follow this reporting requirement. In §668.411(b)(1), we have clarified that the July 31 reporting deadline does not apply to the reporting of placement rates but rather that reporting on that item will be on a date and in a manner announced by the Secretary in a notice published in the Federal Register.

Comments: Several commenters raised concerns that the reporting requirements would be very burdensome for institutions and that the Department underestimated in the NPRM the burden and cost to implement these provisions. In particular, some commenters argued that the reporting requirements would duplicate reporting that institutions already provide and that the additional compliance burden and paperwork hours would lead to higher costs for students. Another commenter said that they would need to hire additional staff to comply with the reporting requirements.

Discussion: Any burden on institutions to meet the reporting requirements is outweighed by the benefits of the accountability and transparency frameworks of the regulations to students, prospective students, and their families. The Department requires the reporting under the regulations to calculate D/E rates, as provided in §§668.404 and 668.405, and to calculate or determine many of the disclosure items, as provided in §668.413. (See “§668.413 Calculating, Issuing, and Challenging Completion Rates, Withdrawal Rates, Repayment Rates, Median Loan Debt, Median Earnings, and Program Cohort Default Rate” for a discussion of the disclosure items that the Department will calculate.) Although there is some overlap with current enrollment reporting and reporting for the purposes of the 150 percent Direct Subsidized Loan Limits, those data do not include award years prior to 2014-2015, nor do they include several data elements required for the calculation of D/E rates, including institutional debt, private education loan debt, tuition and fees, and allowance for books and supplies.

We believe that our estimates of the burden of the reporting requirements are accurate. As an initial matter, the commenters did not submit any data to show that the Department’s estimates are inaccurate. The Department’s estimates are based on average anticipated costs and the actual burden may be higher for some institutions and lower for others. Various factors, such as the sophistication of an institution’s systems, the size of the institution and the number of GE programs that it has, whether or not the institution’s operations are centralized, and whether the institution can update existing systems to meet the reporting requirements will affect the level of burden for any particular institution. (See Paperwork Reduction Act of 1995 for a more detailed discussion of the Department’s burden estimates.)

We have not estimated whether or how many new personnel may be needed to comply with the reporting requirements.  Allocating resources to meet the reporting requirements is an individual institution’s administrative decision.  Some institutions may need to hire new staff, others will redirect existing staff, and still others will not need to make staffing changes because they have highly automated reporting systems. 

In order to minimize burden, the Department will provide training to institutions on the new reporting requirements, provide a format for reporting, and, so that institutions have sufficient time to submit their data for the first reporting period, enable NSLDS to accept reporting from institutions beginning several months prior to the July 31, 2015, deadline. Additionally, we will consider other ways to simplify our reporting systems.

Changes: None.

Comments: One commenter recommended that institutions should only be required to report data they have currently available in an electronic format. The commenter believed that some institutions may not have, in easily accessible formats, the older data that the Department would need to calculate rates in the first few years after implementation of the regulations due to migrations to new data systems and the rapid changes in student information systems in recent years.

Discussion: In accordance with the record retention requirements under §668.24(e), most institutions should have retained the information regarding older cohorts of students that must be reported in the initial years of the regulations, even if the data are maintained in multiple systems or formats. Further, many institutions may have a policy of retaining student records for longer periods, or do so as a result of State or accreditor requirements. Nonetheless, we understand that some institutions may no longer have records for years prior to the required retention period under §668.24(e). Pursuant to the 2011 Final Rules, institutions were similarly required to report information from several years prior to the reporting deadline. The vast majority of institutions were able to comply with the requirements of the 2011 Final Rules, and we again anticipate that cases where data are completely unavailable will be limited. In those instances, an institution may, under §668.411, provide an explanation acceptable to the Secretary for the institution’s inability to comply with part of the reporting requirements.

Changes: None.

Comments: Some commenters recommended adding an alpha-numeric program identifier as an optional reporting requirement so that institutions could report program information separately for individual program locations or formats (e.g. on-line program, part-time program, evening, or weekend program). The commenters asserted that calculating the disclosure items separately in this way would give students and prospective students more meaningful information about program outcomes for their particular location or format.

Discussion: Although we will permit an institution to disaggregate some disclosure items, such as tuition and fees and the percentage of students who borrowed to attend the program by program length, location or format, other disclosures, such as the D/E rates and the items that the Department calculates for institutions under §668.413, will be made at the six-digit OPEID, CIP code, and credential level and may not be disaggregated. Therefore, adding this optional reporting field is unnecessary. See “§668.412 Disclosure Requirements for GE Programs” for a more detailed discussion of whether and when an institution may disaggregate its disclosures.

Changes: None.

Comments: Some commenters requested clarification and additional information about how institutions should report and track students’ enrollment in GE programs. They noted that students often switch programs mid-course or enroll in multiple programs at once, particularly at community colleges.

Discussion: We intend to revise the GE Operations Manual and the NSLDS GE User Guide to reflect the regulations. In updating these resources, we will provide additional guidance on tracking student enrollment. Additionally, we will provide ongoing technical support to institutions regarding compliance with the reporting requirements.

Changes: None.

Comments: One commenter argued that the reporting requirements in §668.411 would violate section 134 of the HEA (20 U.S.C. 1015c), which prohibits the creation of new student unit record databases. The commenter asserted that the new requirements under the regulations for institutions to report private education loan data and other personal data on individuals who receive title IV, HEA program funds and for the Department to retain this newly required data in NSLDS would constitute such a drastic expansion of NSLDS as to constitute a new database in violation of the statutory prohibition against such an expansion of an existing database. APSCU v. Duncan, 930 F.Supp.2d at 220, 221. The commenter further contended that the Department has the burden of proving that gathering personally identifiable information pursuant to these regulations does not create a new database under section 134 of the HEA even if that collection were limited to data on individuals receiving title IV, HEA program funds.

Discussion: As explained previously, in response to the court’s interpretation of relevant law in APSCU v. Duncan, the Department has changed the reporting and accountability determinations in these regulations such that they pertain only to individuals receiving title IV, HEA program funds. The 2011 Prior Rule required institutions to report data on all individuals enrolled in a GE program, including those who did not receive title IV, HEA program funds; the Department retained that data in NSLDS. The court found that retaining data on individuals who did not receive title IV, HEA program funds was an improper creation of a new database.¹⁷⁹ Importantly, the court disavowed any view that it was ruling that 20 U.S.C. 1015c barred the Department from gathering and retaining in NSLDS new data not previously collected on individuals who received title IV, HEA program funds. Accordingly, the commenter’s assertion that the court considered 20 U.S.C. 1015c to bar the addition of new data to NSLDS on individuals receiving title IV, HEA funds is unsupportable.

The objection that the Department fails to demonstrate that adding to the NSLDS new data title IV, HEA program funds recipients does not create a new database disregards the essential purposes for gathering this added data: to determine GE program eligibility, and to provide “accurate and comparable information” to “students, prospective students, and their families.” 79 FR 16426, 16488. Each of these objectives is distinct, and therefore the Department intended each to operate if the other were found to be unenforceable. Id. Section 134 of the HEA allows us to use current NSLDS data, and to add data to NSLDS, for both purposes under section 134 because both are “necessary for the operation of programs authorized by . . . title.” 20 U.S.C. 1015c(b)(1). Section 134 does not define what uses are “necessary for operation of the title IV programs,” nor does the HEA statute articulate a list of those functions for which the Department can use NSLDS. Whether a use is “necessary” is left to the Department’s discretion, in light of statutory mandates, duly-authorized regulations, or simple practical necessity. For example, from its inception in 1993, the Department has used NSLDS as to determine institutional eligibility by reason of an institution’s CDR, a purpose almost identical to determining GE program eligibility. Nothing in section 435 of the HEA, which controls calculation of iCDR, mentions NSLDS or directs the Department to use NSLDS to calculate iCDR. Nevertheless, the Department has consistently used NSLDS to calculate iCDR for purposes of section 435(a). Similarly, the Department has by regulation since 1989 terminated eligibility of an institution with a single year iCDR exceeding 40 percent or more. 34 CFR 668.206(a)(1), 54 FR 24114, 24116 (June 5, 1989). The Department has used NSLDS for that regulatory eligibility determination as well. See Notice of a New System of Records, 59 FR 65532 (Dec. 20, 1994) 18-40-0039, Purpose (2), Routine Use (a)(2).¹⁸⁰ Accordingly, use of NSLDS data to determine programmatic eligibility under these regulations involves the identical kind of eligibility determination as the iCDR determination process used for NSLDS over the past 20 years. Section 485 of the HEA authorizes the Department to maintain in NSLDS information that “shall include (but is not limited to) . . . the eligible institution in which the student was enrolled. . .” 20 U.S.C. 1092b(a)(6). Because the court upheld the Department’s authority to determine whether a program in fact prepared students for gainful employment, the Department is adding data to the existing NSLDS database as needed to make a programmatic eligibility determination. Adding data regarding recipients of title IV student financial assistance in order to make this eligibility determination does not change NSLDS into a new database.

The Court further concluded that requiring disclosures was well within the Department’s authority. APSCU v. Duncan, 870 F.Supp.2d at 156. Doing so is, in the judgment of the Department, necessary for the operation of the title IV, HEA programs. Adding data on individuals who have received title IV, HEA program funds to NSLDS in order to facilitate these disclosures similarly does not change NSLDS into a new database.

Changes: None.

§668.412 Disclosure Requirements for GE Programs

General

Comments: Several commenters recommended that the Department include some but not all of the proposed disclosure items in the final regulations. They argued that including all of the information would overwhelm students. Although commenters identified varying disclosure items that they believed prospective and enrolled students would find most helpful, they generally agreed that the most critical information for students includes information about how long it takes to complete a program, how much the program costs, the likelihood that students would find employment in their field of study, and their likely earnings in that field. Another commenter suggested that the Department survey students about the types of information they would find helpful in choosing an academic program or college.

Discussion: We believe that all of the proposed disclosures would provide useful and relevant information to prospective and enrolled students. However, we agree with the commenters that it is critical to provide prospective and enrolled students with the information that they would find most helpful in evaluating a program when determining whether to enroll or to continue in the program. As we discussed in the NPRM, we do not intend to include all of the disclosure items listed in §668.412 on the disclosure template each year. We will use consumer testing to identify a subset of possible disclosure items that will be most meaningful for students.

Changes: None.

Comments: Several commenters supported having robust disclosures, and they recommended requiring additional disclosures on the disclosure template. In particular, commenters recommended requiring institutions to disclose the names and qualifications of a program’s instructors, the institution’s most recent accreditation findings (e.g., self-studies, accreditation visiting team action reports and action letters), compliance audits, financial statements, and the institution’s application for Federal funds to the Department. Commenters also recommended that the Department post each institution’s program participation agreement (PPA) online for public inspection or, at a minimum, require institutions to publicly post the GE-related portions of the institution’s PPA so that the public can review the information regarding its GE programs certified by the institution under §668.414. Some of the commenters argued that even robust disclosures would be inadequate to protect consumers and that the disclosures should work in conjunction with other substantive protections like strong debt metrics and certification requirements, provisions for borrower relief, and enrollment caps.

Discussion: In determining which pieces of information to require institutions to disclose, we have focused on identifying the information that will be most helpful to prospective and enrolled students, and we have built flexibility into the regulations to allow for modifications based on consumer testing and student feedback. Although access to accrediting agency documentation or Federal compliance audits of institutions is valuable and institutions may opt to disclose this information independently, including this information on the disclosure template may not be useful to prospective and enrolled students. Nonetheless, if consumer testing or other sources of evidence show that prospective and enrolled students would benefit from this information, we would consider adding these items to the disclosure template in the future through a notice published in the Federal Register.

As discussed under “§668.414 Certification Requirements for GE Programs,” institutions will be required to certify that the GE programs listed on their PPA meet applicable accreditation, licensure, and certification requirements. The PPA is a standardized document that largely mirrors the requirements in 34 CFR §668.14. Unless an institution has a provisional PPA, the PPA for one institution will be nearly identical to that of another except for the list of the institution’s GE programs. Because PPAs do not generally contain unique information about institutions, we do not believe that it would be helpful to consumers for the Department to begin publishing institutions’ PPAs or requiring institutions to publish the GE-related portions of their PPAs. We note, however, that we would provide a copy of an institution’s PPA upon request through the Freedom of Information Act process.

Lastly, as discussed in the NPRM and in these regulations, we believe that the disclosure requirements, combined with the accountability metrics, the certification requirements, and the student warnings, will be effective in supporting and protecting consumers. We address in “§668.410 Consequences of the D/E Rates Measure” comments suggesting we adopt enrollment limits and borrower relief provisions.

Changes: None.

Comments: A commenter stated that institutions should be allowed to disclose multiple SOC codes that match a program’s CIP code.

Discussion: We agree that a program may be designed to lead to several occupations as indicated by Department of Labor SOC codes. For this reason, allowing institutions to select one or multiple SOC codes for inclusion on the disclosure template is among the disclosures that were required under the 2011 Final Rules and the potential disclosures under these regulations.

Changes: None.

Comments: Several commenters compared the disclosure requirements of the proposed regulations to those of the current regulations. One commenter believed that adding new disclosures to the current requirements without coordinating them would be administratively burdensome for institutions and confusing for students. Some commenters noted that, as under the current regulations, some programs will have too few students to make some of the disclosures because of privacy concerns. These commenters recommended incorporating existing sub-regulatory guidance from the Department into the final regulations that directs institutions to refrain from disclosing information, such as median loan debt, where ten or fewer students completed the program. Some commenters argued that the current disclosures are adequate and should be retained in the final regulations without any changes. Lastly, one commenter noted that the NPRM did not describe the impact of the current disclosure requirements or whether they are achieving their purpose.

Discussion: Although the disclosures in §668.6(b) of the 2011 Final Rules are useful, the additional disclosures in these regulations will make additional valuable information available to students and prospective students. Further, the current disclosure requirements are limiting because §668.6(b) does not give the Department the flexibility to change the items as it learns more about the information students find most useful. We agree with the commenters that we must carefully consider how to transition from the current disclosure requirements to the requirements of the final regulations without confusing or overwhelming students, and we will use consumer testing to identify the best way to do this. We will also provide guidance and technical assistance to institutions to help them transition to the new disclosures. We will be evaluating the impact of the disclosures we are establishing in these regulations.

Because it will take some time for the Department to conduct consumer testing regarding the disclosure template and to seek comment on the disclosure template pursuant to the Paperwork Reduction Act of 1995, we are providing in the regulations that institutions must comply with the requirements in this section beginning on January 1, 2017. To ensure that institutions continue to disclose information about their GE programs, we are retaining and revising §668.6(b) to provide that institutions must comply with those disclosure requirements until December 31, 2016.

With respect to the privacy concerns raised by the commenters, for the 2011 Final Rules, the Department provided sub-regulatory guidance to institutions instructing them not to disclose median loan debt, the on-time completion rate, or the placement rate (unless the institution’s State or accrediting agency methodology requires otherwise) for a program if fewer than 10 students completed the program in the most recently completed award year. This guidance remains in effect. Further, we are revising §§668.412 to reflect this guidance.

Changes: We have revised §§668.412 to specify that an institution may not include on the disclosure template information about completion or withdrawal rates, the number of individuals enrolled in the program during the most recently completed award year, loan repayment rates, placement rates, the number of individuals enrolled in the program who received title IV loans or private loans for enrollment in the program, median loan debt, mean or median earnings, program cohort default rates, or the program’s most recent D/E rates if that information is based on fewer than 10 students.

We also have revised §668.412 to specify that institutions must begin complying with the disclosure requirements beginning on January 1, 2017. We also have revised §668.6(b) to provide that institutions must comply with those disclosure requirements through December 31, 2016.

Comments: Commenters raised general concerns about the burden associated with the disclosure requirements. In particular, some commenters were concerned that the potential for annual changes in the content and format of the disclosures would create uncertainty and significant administrative burden for institutions. One commenter recommended that the Department study how students use information before establishing the disclosure requirements. The commenter suggested that the Department calculate simple measures and publish relevant information on College Navigator while conducting this study. Other commenters objected that disclosure requirements were vague and burdensome by, for example, requiring disclosure of the total cost of tuition, fees, books, supplies, and equipment that would be incurred to complete the program within its stated term.

Discussion: We believe that the benefits of disclosure items for consumers outweigh the increase in institutional burden. In addition, the Department does not intend to require institutions to make all of the disclosures each year. The regulations allow the Department flexibility to adjust the disclosures as we learn more about what information will be most helpful to students and prospective students. However, we do not expect that the disclosure template will vary dramatically from year to year, and so in most years, there will be little added burden because of this provision. We will publish changes to the items to be disclosed in the Federal Register, providing an opportunity for the public, specifically institutions and consumers, to provide us with feedback about those changes.

Further, we have included provisions to minimize the burden associated with the disclosures as much as possible. We recognize that an institution may not know precisely the cost that a prospective student would incur to attend and complete a GE program, as must be disclosed but the institution must already gather much of the same data to comply with the disclosure obligations imposed by section 1132(h) of the HEA, and the solution adopted there is applicable here: if the institution is not certain of the amount of those costs, the institution shall include a disclaimer advising that the data are estimates.¹⁸¹

In addition, the Department, rather than institutions, will calculate the bulk of the disclosure items, as discussed under “§668.413 Calculating, Issuing, and Challenging Completion Rates, Withdrawal Rates, Repayment Rates, Median Loan Debt, Median Earnings, and Program Cohort Default Rates.” As we implement the regulations, we will continue to analyze the burden associated with the disclosure requirements and consider ways to minimize that burden.

Changes: None.

Comments: Some commenters raised concerns about how the proposed disclosure requirements would affect or be affected by other existing or planned efforts and initiatives such as the college ratings system, College Navigator, and College Scorecard. One commenter suggested that the disclosures should be coordinated with the planned college ratings system. Other commenters noted that institutions already disclose graduation rates, costs, and other information through College Navigator and the College Scorecard, and argued that requiring additional disclosures that use similar data points but measure different cohorts of students would not be helpful to prospective students. Some of the commenters suggested that modifying College Navigator and College Scorecard to provide students and families with meaningful information with respect to all programs and all institutions would be less burdensome and more effective.

In addition to these concerns, one commenter suggested that the Department utilize College Navigator, the College Scorecard, and the College Affordability and Transparency Center to disclose when an institution’s GE program is in the zone to ensure that students and other users have access to information about programs in jeopardy of losing their eligibility.

Discussion: The College Navigator and the College Scorecard are useful for consumers and we intend for the planned college ratings system to provide additional helpful information. But, we do not agree that they make the GE disclosures unnecessary. First, these three tools provide, or in the case of the college ratings system will provide, consumers with information at an institutional level. They do not provide information about the graduation rates, debt, or employment and earnings outcomes of particular GE programs. Second, College Navigator and the College Scorecard are, and the college ratings system will be, accessible through the Department’s Web site, whereas institutions will be required to publish the disclosures required by these regulations where students are not only more likely to see them, but also more likely to see them early in their search process--on the institutions’ own Web sites and additionally, in informational materials such as brochures. Accordingly, we believe that the disclosures required by these regulations will be more effective in ensuring that students and prospective students obtain critical information about program-level student outcomes. We note that this approach is consistent with long-standing provisions in the HEA requiring institutions to publish consumer information on their Web sites under the assumption that students and families are likely to look on those Web sites for that information.

With respect to the suggestion that the Department use College Navigator, the College Scorecard, and the College Affordability and Transparency Center to alert prospective students and families when an institution has a GE program in the zone under the D/E rates, the Department intends to make this information publicly available and may choose to use one of these or another vehicle to do so.

Changes: None.

Comments: Several commenters argued that all institutions participating in the title IV, HEA programs should be required to make the disclosures for all of their programs. They contended that it is unfair and discriminatory to apply the transparency framework only to GE programs. The commenters asserted that the disclosures would not be meaningful and could be misleading to students because of a lack of comparability across institutions in different sectors, noting that a program at a proprietary institution would be subject to the regulations while the same program at a public institution might not.

Discussion: As discussed under “§668.401 Scope and Purpose,” these regulations apply to programs that are required, under the HEA, to prepare students for gainful employment in a recognized occupation in order to be eligible to participate in the title IV, HEA programs. The regulations do not establish requirements for non-GE programs.

The disclosures will be valuable even though they do not apply to all programs at all institutions because, we believe, that information about program performance and student outcomes have value in and of themselves. Prospective students will be able to evaluate the information contained in a particular program’s disclosures against their own goals and reasons for pursuing postsecondary education regardless of whether they have comparable information for programs at other institutions. For example, they can consider whether a program will lead to the earnings they desire, and whether the debt that other students who attended that program incurred would be manageable for them. Further, students will have access to comparable information for all programs leading to certificates or other non-degree credentials since these programs will be subject to the disclosure requirements regardless of the institution’s sector. We acknowledge that students will have less ability to compare degree programs because only degree programs offered by for-profit institutions will be subject to these regulations. We do not believe this significantly diminishes the value of the disclosures as students will nonetheless have the ability to compare programs across the for-profit sector.

Changes: None.

Comments: Some commenters asserted that requiring an institution to make the disclosures required under §668.412 would violate the institution’s First Amendment rights. They made similar arguments to those made by some commenters in connection with the student warning requirements under §668.410.

Discussion: See “§668.410 Consequences of the D/E Rates Measure” for a discussion of the relevant law with respect to laws that mandate disclosures to consumers and potential consumers. As with the student warnings, the disclosure requirements directly advance a significant government interest--both preventing deceptive advertising about GE programs and providing consumers information about an institution’s educational benefits and the outcomes of its programs. The disclosure requirements too are based on the same significant Federal interest in consumer disclosures evidenced in more than thirty years of statutory disclosure requirements for institutions that receive title IV, HEA program funds akin to the disclosures required under these regulations.¹⁸² As with the student warnings, the disclosures required under §668.412 are purely factual and will not be controversial when disclosed, as institutions will have had the opportunity to challenge or appeal the disclosures calculated or determined by the Department.¹⁸³ Finally, the individual disclosure items listed in §668.412 have been narrowly tailored to provide students and prospective students with the information the Department considers most critical in their educational decision making, and the Department will use consumer testing to inform its determination of those items it will require on the disclosure template. As with the student warnings, we believe that requiring an institution to both include the disclosure template on its program Web site and directly distribute the template to prospective students is the most effective manner of advancing our significant government interest.

The fact that Congress has already required, in section 485 of the HEA, that institutions disclose data such as completion rates and cost of attendance does not mean that the disclosures required by these regulations would cause confusion. The HEA requires disclosures about the institution as a whole, for example, the completion, graduation, and retention rates of all its students, disaggregated by such characteristics as gender, race, and type of grant or loan assistance received, but not by program. 34 CFR 668.45(a)(6). Far from creating consumer confusion, the regulations here address a significant gap in those disclosures: the characteristics of individual GE programs. Particularly for consumers who enroll in a program in order to be trained for particular occupations, this program-level information can reasonably be expected to be far more useful than information on the institution as a whole.

Changes: None.

Specific Disclosures

Comments: Several commenters raised concerns that because the 100 percent of normal time completion rate disclosure is calculated on a calendar time basis, it does not align with the time period (award years) over which the D/E rates are calculated. One of these commenters also questioned how an institution that offers programs measured in clock hours would determine the length of the program in weeks, months, or years.

Discussion: We continue to believe that completion rates should be disclosed on a calendar time basis rather than on an academic or award year basis for the purposes of the disclosures. For example, for a program that is 18 months in length, an institution will disclose the percentage of students that completed the program within 18 months. This disclosure is intended to help prospective and enrolled students understand how long it might take them to complete a program. Consumers understand time in terms of calendar years, months, and weeks much more readily than they understand time in terms of an “academic year” or “award year” as defined under the title IV, HEA program regulations. Several title IV, HEA program regulations, including the disclosure provisions of the current regulations that have been in effect since July 1, 2011, already require that institutions determine the length of a program in calendar time. In addition, institutions must provide the program length, in weeks, months, or years, for all title IV, HEA programs to NSLDS for enrollment reporting.

Changes: None.

Comments: None.

Discussion: Section 668.412(a)(4) of the proposed regulations would have required institutions to disclose the number of clock or credit hours, as applicable, necessary to complete the program. However, in some cases, competency-based and direct-assessment programs are not measured in clock or credit hours for academic purposes. Accordingly, we are adding language that would allow an institution to disclose the amount of work necessary to complete such programs in terms of a unit of measurement that is the equivalent of a clock or credit hour.

Changes: In §668.412(a)(4), we have added the words “or equivalent.”

Comments: Numerous commenters urged the Department to develop a standardized placement rate that would apply to all GE programs, arguing that it would provide important information to students. The commenters criticized the approach in the proposed regulations of requiring an institution to calculate a placement rate only if required to do so by its accrediting agency or State, arguing that it would lead to inconsistent disclosures because not all programs would have placement rates and because institutions would use differing methodologies. The commenters believed that developing a national placement rate methodology, even if the rate itself is not verifiable, would allow students to compare placement rates across programs and would protect against manipulation and misrepresentation of placement rates. They believed that standardizing the rate by specifying, for example, how soon after graduation a student must be employed, how long a student must be employed, and whether a student must be working in the field for which he or she was trained to be considered “placed” would improve the reliability and comparability of the rates.

Some of the commenters suggested alternatives to developing a standardized placement rate methodology. For instance, a few commenters suggested that the Department use the placement rate under §668.513 for the purposes of the disclosures. Another commenter suggested that, if requiring all institutions to calculate placement rates using a standardized methodology for all of their programs would be overly burdensome for institutions not already required to calculate a placement rate, the Department should require only institutions already required to calculate a placement rate by their accrediting agency or State to disclose a placement rate calculated using a national methodology.

Discussion: We appreciate the commenters’ suggestion to develop a national placement rate methodology, and we agree that this would be a useful tool for prospective and enrolled students, researchers, policymakers, and the public. However, we are not prepared at this time to include such a methodology in these regulations. We will continue to consider developing a national placement rate methodology in the future.

Changes: None.

Comments: Some commenters argued that if the Department does not establish a uniform methodology, it should require institutions subject to existing placement rate disclosure requirements from their State or accrediting agency to disclose the lowest placement rate of the rates they are required to calculate. Other commenters suggested that the Department require institutions to disclose under these regulations each of the placement rates that they are required to disclose by other entities. These commenters believed that including all of the calculated rates on the disclosure template would provide prospective students and other stakeholders a more comprehensive picture of student outcomes.

Discussion: The regulations in §668.412 provide that job placement rates must be disclosed if an institution is required to calculate such rates by a State or accrediting agency. This requirement applies to all placement rate calculations that a State or accrediting agency may require.

We are revising §668.412(a)(8) to clarify that, as in the 2011 Final Rules, an institution is required to disclose a program’s placement rate if it is required by an accrediting agency or State to calculate the placement rate at the institutional level, the program level, or both. If the State or accrediting agency requirements apply only at the institutional level, under these regulations, the institution must use the required institution level methodology to calculate a program level placement rate for each of its programs. As in the 2011 Final Rules, a “State” is any State authority with jurisdiction over the institution, including a State court or a State agency, and the requirement to calculate a placement rate under these regulations may stem from requirements imposed by the authority directly or agreed to by the institution in an agreement with the State authority.

Changes: We have revised §668.412(a)(8) to clarify that an institution must disclose a program’s placement rate if it is required by an accrediting agency or State to calculate the placement rate either for the institution, the program, or both, using the required methodology of the State or accrediting agency.

Comments: Some commenters recommended requiring institutions to disclose the mean or the median earnings of graduates of the GE program.

Discussion: We agree with the commenters that either of the mean or median earnings of a program would be useful information for prospective students and enrolled students.

Changes: We have revised §668.412(a)(11) to add the mean, in addition to median, earnings as a possible disclosure item to be included on the disclosure template.

Comments: We received a number of comments regarding the requirement that institutions disclose whether a program satisfies applicable professional licensure requirements and whether the program holds any necessary programmatic accreditations. Commenters recommended that we require institutions to disclose the applicable educational prerequisites for professional licensure in the State in which the institution is located and in any other State included in the institution’s MSA rather than just whether the program satisfies them. Some commenters questioned the value of including disclosures regarding licensure, certification, and accreditation, noting that a program would not be eligible for title IV, HEA program funds if it could not certify under §668.414 that it meets the licensure, certification, and accreditation requirements. These commenters urged the Department to maintain and strengthen the certification requirements under “§668.414 Certification Requirements for GE Programs.” They also recommended that if the certification requirements are removed, then an institution should be required to clearly and prominently disclose if a GE program does not have the necessary programmatic accreditation. These commenters asserted that where a program does meet certain requirements, it is typically easy to find disclosures indicating this information, but that it is often much more difficult to find disclosures indicating that a program does not meet particular requirements and that provide information on the consequences of failing to do so.

Several commenters recommended that the disclosures be broadened to reflect the circumstances in the location where a prospective student lives, rather than the State in which the institution is located. (See the more detailed discussion of this issue under “§668.414 Certification Requirements for GE Programs.”)

Other commenters argued that the disclosure requirements are overly broad and that it would be extremely burdensome for institutions to determine whether a program holds proper programmatic accreditation. They believed that such a determination would be subjective and that it would be almost impossible to meet this requirement using a standardized template.

Some commenters asserted that, if a program does not meet the requirements in §668.414, for consistency purposes, the institution should be required to disclose that students are unable to use title IV, HEA program funds to enroll in the program.

Some commenters suggested that the Department use consumer testing, as well as consult with other agencies and parties such as the CFPB, FTC, accrediting agencies, and State attorneys general, to specify the text and format of the programmatic accreditation disclosure. Along these lines, some commenters were concerned that describing criteria as “required” or “necessary” would be ineffective without adding clarifying text to make it clear that the programmatic accreditation is needed to qualify to take an exam without additional qualifications such as a minimum number of years working in the field of study.

Discussion: We agree with the commenters that students and prospective students should know whether a program satisfies the applicable educational prerequisites for professional licensure required by the State in which the institution is located and in any other State within the MSA in which the institution is located and whether a program is programmatically accredited. Because students may seek employment outside of their State or MSA, however, we believe it would also be helpful to students to know of any other States for which the institution has determined whether the program meets licensure and certification requirements and those States for which the institution has not made any such determination. We are revising the regulations accordingly.

We decline to require institutions to disclose the actual licensure or certification requirements that are met given the burden this would impose on institutions. We believe that the more critical information for students is whether or not the program satisfies the applicable requirements.

The disclosure requirements regarding programmatic accreditation in the proposed regulations were not overly broad, burdensome, or subjective. However, we are simplifying these requirements to make the disclosures more effective for consumers and to facilitate institutional compliance. We are revising §668.412(a)(15) to require institutions to disclose, if required on the disclosure template, simply whether the program is programmatically accredited. Under §668.414, an institution is already required to certify that a program is programmatically accredited, if such accreditation is required by a Federal governmental entity or by a governmental entity in the State in which the institution is located or in which the institution is otherwise required to obtain State approval under 34 CFR 600.9. Accordingly, institutions should already have obtained these necessary programmatic accreditations. For any other programmatic accreditation, the regulations merely require disclosure of this information. It will be to an institution’s benefit to disclose any programmatic accreditation it has obtained beyond the accreditation required under §668.414. Finally, we do not agree that the proposed requirements were subjective, but we have, nonetheless, revised the requirement to avoid reference to “necessary” programmatic accreditation. As revised, institutions are required only to disclose whether they have the programmatic accreditation. We are also revising the regulations to require an institution to disclose the name of the accrediting agency or agencies providing the programmatic accreditation so that students have this important information.

It is not necessary to require institutions to disclose that students are unable to use title IV, HEA program funds to enroll in a program if the program does not meet the requirements in §668.414. If a program does not meet those requirements, then it is not considered a GE program and therefore would not be required to make any disclosures under these regulations.

As we have discussed, we will conduct consumer testing to learn more about how to convey information to students and prospective students. However, we believe that there is sufficient explanation within the description of the disclosure items for institutions to know what needs to be disclosed and when State or Federal licensing and certification requirements have been met or whether a program has been programmatically accredited.

Changes: We have revised §668.412(a)(14) to require that an institution indicate whether the GE program meets the licensure and certification requirements of each State within the institution’s MSA, and any other State in which the institution has made a determination regarding those requirements. We have also revised the regulations to require that the institution include a statement that the institution has not made a determination with respect to the licensure or certification requirements of other States not already identified. We have revised §668.412(a)(15) to simplify the required disclosure and to require institutions to disclose, in addition to whether the program is programmatically accredited, the name of the accrediting agency.

Comments: None.

Discussion: The proposed regulations provided that the disclosure template must include a link to the Department’s College Navigator Web sitehttp://nces.ed.gov/collegenavigator/, or its successor site, so that students and prospective students have an easy reference to a resource that permits easy comparison among similar programs. As the Department or another Federal agency may in the future develop a better tool that serves prospective students in this regard, we are revising §668.412(a)(16) to refer to College Navigator, its successor site, or another similar Federal resource, which would be designated by the Secretary in a notice published in the Federal Register.

Changes: We have added in §668.412(a)(16) a reference to other similar Federal resource.

Comments: None.

Discussion: For the readers’ convenience, we have consolidated the requirements relating to student warnings in §668.410(a), including the requirement that institutions include the student warning on the disclosure template. Although we are removing the substantive provisions of this requirement from §668.412(b)(2), we are adding a cross-reference to the requirement in §668.410(a).

Changes: We have revised §668.412(b)(2) to provide that an institution must update the disclosure template with the student warning as required under §668.410(a)(7).

Program Web Pages and Promotional Materials

Comments: A commenter objected to the provision that would require the institution to change its Web site if the Department were to determine that the required link to the disclosure template is not sufficiently prominent, on the ground that this restricted its First Amendment rights.

Discussion: Section 668.412(c) requires an institution to “provide a prominent, readily accessible, clear, conspicuous and direct link to the disclosure template for the program” on various Web pages, and to “modify” its Web site if the Department determines that the required link is not “prominent, readily accessible, clear, conspicuous and direct.” This provision does not, as the commenter suggests, give the Department free rein to dictate the content of the institution’s Web site in derogation of the institution’s First Amendment rights. The Department’s authority reaches no further than necessary to cure a failure by the institution to display the required link adequately. Requirements that consumer disclosures be “clear and conspicuous” are not unusual in Federal law, and the Federal Trade Commission (FTC), for example, has provided extensive guidance on how required disclosures are to be made in electronic form in a manner that meets a requirement that information be presented in a “clear and conspicuous” manner.¹⁸⁴ The Department would require any corrective action based on the kinds of considerations listed by the FTC in this guidance. We believe the regulations give the institution sufficient flexibility to design, manage, and modify, as needed, the content of its Web page as long as it makes the link sufficiently prominent. The regulations do not authorize the Department to require an institution to remove or modify any content included on the pertinent Web page. Rather, the institution is required to make only those changes needed to make the required link stand out to the consumer.

Changes: None.

Comments: Several commenters supported the provisions designed to ensure that the link to a program’s disclosure template is easily found and accessible from multiple access points on a program’s Web site. However, the commenters urged the Department to provide examples of a link that is “prominent, readily accessible, clear, conspicuous, and direct.” Some of the commenters advised conducting consumer testing with the types of individuals who are a part of the target audience for these templates, including prospective students and those advising them on which program to attend, as well as consulting with the Consumer Financial Protection Bureau (CFPB), FTC, and State attorneys general. They argued that these efforts would help to ensure that the template will be easily found and that complicated terms like “repayment rates” and “default” that consumers might not readily understand will be adequately explained. One commenter recommended that the disclosures be incorporated directly into program Web sites so that prospective students will be able to find them easily.

Discussion: We will test the format and content of the disclosure template with consumers and other relevant groups, and will provide examples of acceptable ways to make a link easy to find and accessible based on the results of that testing.

We agree that, in lieu of providing on a program’s Web page a link to the disclosure template, an institution should be able to include the disclosure template itself.

Changes: We have revised §668.412(c) to clarify that an institution may include the disclosure template or a link to the disclosure template on a program’s Web page.

We have also clarified in this section that the provisions relating to a program’s Web page apply without regard to whether the Web page is maintained by the institution or by a third party on the institution’s behalf. To improve the organization of the regulations, we have moved the provisions relating to providing separate disclosure templates for different program locations or formats to new paragraph (f).

Comments: A few commenters provided suggestions regarding the requirement that institutions include the disclosure template or a link to the disclosure template in all promotional materials. One commenter urged the Department to increase enforcement of these requirements, noting that many institutions are not in compliance with the current regulations in §668.6(b)(2)(i). The commenter recommended that the Department consult with the CFPB, FTC, and State attorneys general to identify effective enforcement mechanisms related to disclosures. Other commenters argued that the Department should specify that the links to the disclosure template from promotional materials should also be prominent, clear, and conspicuous, because predatory schools will hide this information by using illegible type. The commenters urged the Department to provide clear examples of links on promotional materials that would be considered acceptably prominent or clear and conspicuous, noting that the FTC and FCC have issued this type of information. Another commenter urged the Department to address situations where a student’s first point of contact with a program is through a lead generating company. The commenter recommended requiring institutions to post disclosures prominently in any venue likely to serve as a student’s first point of interaction with the institution, including lead generation outlets, and also to require lead generation companies that work with GE programs and institutions to provide clear and conspicuous notice to students that they should consult with the Department for information about GE programs, costs, outcomes, and other pertinent information.

Discussion: We appreciate the commenters’ suggestion to consult with and learn from other enforcement-focused agencies to improve and strengthen our enforcement efforts. We note that under §668.412(c)(2), the Secretary has the authority to require an institution to modify a Web page if it provides a link to the disclosure template that is not prominent, readily accessible, clear, conspicuous, and direct. This provision will strengthen our ability to enforce these provisions by giving us a way to prompt institutions to make changes without requiring a full program review.

We agree with the commenters who suggested requiring that links to the disclosure template from promotional materials be prominent, clear, and conspicuous. We believe that this will make it clear that institutions may not undermine the intent of this provision by including in their promotional materials a link in a size, location, or, in the case of a verbal promotion, speed that will be difficult to find or understand. We are revising the regulations to make this clear. We intend to issue guidance consistent with the guidance provided by the FTC on what we would consider to be a prominent, readily accessible, clear, conspicuous, and direct link to the disclosure template on promotional materials.

With regard to lead generating companies, we are clarifying in the regulations that institutions will be responsible for ensuring that all of their promotional materials, including those provided by a third party retained by the institution, contain the required disclosures or a direct link to the disclosure template, as required under §668.412(d)(1).

Changes: We have revised §668.412(d)(1) to make clear that the requirements apply to promotional materials made available to prospective students by a third party on behalf of an institution. We have also revised §668.412(d)(1)(ii) to require that all links from promotional materials to the disclosure template be prominent, readily accessible, clear, conspicuous, and direct.

Format and Delivery

Comments: None.

Discussion: As discussed in “§668.401 Scope and Purpose,” we are simplifying the definition of “credential level” by treating all of an institution’s undergraduate programs with the same CIP code and credential level as one “GE program,” without regard to program length, rather than breaking down the undergraduate credential levels according to the length of the program as we proposed in the NPRM. For the purpose of the accountability framework, we believe the benefits of reducing reporting and administrative complexity outweigh the incremental value that could be gained from distinguishing among programs of different length. For the purposes of the transparency framework, however, there are not the same issues of reporting and administrative complexity. Further, we believe that prospective students and students will benefit from having information available to make distinctions between programs of different lengths. We are revising §668.412(f) to require institutions to provide a separate disclosure template for each length of the program. The institution will be allowed to disaggregate only those items specified in §668.412(f)(3), which were discussed in connection with disaggregation by location and format.

Changes: We have revised §668.412(f)(1) to require institutions to provide a separate disclosure template for each length of the program, and specified in §668.412(f)(3) the disclosure items that may be disaggregated on the separate disclosure templates.

Comments: Some commenters argued that the Department should not permit institutions to disaggregate the disclosures that the Department calculates under §668.413 by location or format, as provided in §668.412(c)(2) of the proposed regulations. The commenters noted that this could undermine the Department’s intention to avoid inaccuracies and distortions in the relevant data. These commenters were also concerned that if more information is disaggregated by location or format, it will be very difficult for consumers to find and understand that information. The commenters recommended that the Department test whether disaggregated data would provide better, clearer, and more accessible information and, if testing shows positive results, revise the regulations in the future to provide this option.

Other commenters recommended that the Department calculate separate rates for the disclosures under §668.413 for different locations or formats of a program if an institution opted to distinguish its programs in reporting to the Department. As discussed under “§668.411 Reporting Requirements for GE Programs,” these commenters suggested allowing institutions to use an optional program identifier to instruct the Department to disaggregate the disclosure calculations based on different locations or formats of a program.

Discussion: Because there are several disclosure items that may vary significantly depending on where the program is located or how it is offered, allowing institutions to disaggregate some of their disclosures will provide consumers with a more accurate picture of program costs and outcomes. For example, a program that is offered in multiple States may be subject to placement rate requirements by more than one State or accrediting agency with differing methodologies.

However, we agree with the commenters who were concerned that allowing institutions to disaggregate the disclosures calculated by the Secretary could be counterproductive. We did not intend in the proposed regulations for institutions to be able to disaggregate the disclosure rates calculated under §668.413, and we have revised the regulations to make this more clear by specifying which of the disclosure items institutions may disaggregate. The following chart identifies the disclosure items that institutions must disaggregate if they provide separate disclosures by program, based on the length of the program or location or format, and those items that may not be disaggregated under any circumstances. We note that, regardless of whether institutions choose to disaggregate certain disclosure items, programs will still be evaluated at the six-digit OPEID, CIP code, and credential level.

If an institution disaggregates by length of the program or chooses to disaggregate by location or by format, the following disclosures must be disaggregated by length of the program, location, or format, as applicable.	Disclosure items institutions may not disaggregate by location or format under any circumstances
	§668.412(a)(1) – primary occupations program prepares students to enter
§668.412(a)(4) – number of clock or credit hours or equivalent, as applicable	§668.412(a)(2) – completion and withdrawal rates
§668.412(a)(5) – total number of individuals enrolled in the program during the most recently completed award year	§668.412(a)(6) – loan repayment rate
§668.412(a)(7) – total cost of tuition and fees and total cost of books, supplies, and equipment incurred for completing the program within the length of the program	§668.412(a)(10) – median loan debt
§668.412(a)(8) – program placement rate	§668.412(a)(11) – mean or median earnings
§668.412(a)(9) – the percentage of individuals enrolled during the most recently completed award year that received a title IV loan or a private loan for enrollment	§668.412(a)(12) – program cohort default rate
§668.412(a)(14) – whether the program satisfies applicable educational prerequisites for professional licensure or certification in States within the institution’s MSA or other States for which the institution has made that determination and a statement indicating that the institution has not made that determination for other States not previously identified	§668.412(a)(13) – annual earnings rate
	§668.412(a)(15) – whether the program is programmatically accredited and the name of the accrediting agency §668.412(a)(16) – link to College Navigator

Changes: We have renumbered the applicable regulations. The provisions permitting an institution to publish a separate disclosure template for each location or format of a program are in §668.412(f)(2) of the final regulations. In §668.412(f)(3), we have specified the disclosure items that an institution must disaggregate if it uses a separate disclosure template for the length of the program or if it chooses to use separate disclosure templates based on the location or format of the program.

Comments: We received a number of suggestions for how we could improve the disclosure template from an operational perspective. For example, some of the commenters recommended adding skip logic to the template application so that fields for which there is no information to disclose can be skipped. These commenters further suggested that the template instructions should clarify that institutions should enter only information for students enrolled in programs for a given CIP code, not for students enrolled in other non-GE credential level programs in the same CIP code. The commenters also recommended the template be designed to ensure cohorts are designated appropriately and to allow institutions to enter different time increments instead of weeks, months, or years. Additionally, some commenters recommended ensuring that the template is compliant with the Americans with Disabilities Act. Another commenter argued that the disclosure template should be a “fill-in-the-blank” document in a common Microsoft Word file for easy incorporation into Web sites.

Discussion: We will continue to improve the template to make it easier for institutions to complete and display it as well as to make it more useful for students and prospective students. We appreciate the suggestions offered by the commenters and will consider them as we revise the template to reflect these final regulations.

We note that we have already addressed several of the recommendations in the current template. For instance, we have incorporated skip logic so that institutions will not be asked to disclose certain information if fewer than 10 students completed the applicable GE program. The template also meets all accessibility requirements. Further, we have refined our Gainful Employment Disclosure Template Quick Start Guides and the instructions within the template to provide greater clarity.

With respect to the suggestion to allow institutions more flexibility to use different increments of time besides weeks, months, and years, we note that we have selected these units intentionally to match the program lengths used for the purposes of the 150 percent Subsidized Loan Limit and NSLDS Enrollment reporting requirements. Further, we believe that calendar time is most easily understood by consumers.

Regarding the suggestion to provide a fill-in-the-blank disclosure document for institutions to complete and incorporate into their Web site, we disagree that this approach would be appropriate. We believe that the disclosure template is effective because it is standardized in its appearance. Although a Word document may be easier to use, it would result in a lack of consistency in presentation across programs. We believe that requiring an easy-to-find link and description of the disclosures, combined with our ability to work with institutions to make changes to improve the placement and visibility of the disclosures, will offset any perceived disadvantage to using an application to create the template.

Changes: None.

Comments: We received suggestions from numerous commenters on proposed §668.412(e) regarding the direct distribution of disclosures to prospective students, particularly on how the disclosures must be provided, when the disclosures must be provided, and how institutions should document that students received the disclosures.

Several commenters provided feedback about how institutions should provide the disclosures. Specifically, some commenters opposed the proposed requirement for institutions to provide the disclosures as a stand-alone document that student must sign, while others supported requiring the disclosures to be made clearly and directly, with text specified by the Department. Other commenters recommended exploring other means of distributing the disclosures to students, such as providing a video to each school to use or posting a video on YouTube that describes the disclosure information. Some commenters stressed the need to consult with the CFPB, FTC, and State attorneys general to determine whether prospective students should be asked to sign a document confirming that they received a copy of the disclosure template and, if so, what it should say and when and how it should be conveyed to maximize the effectiveness of the disclosures.

We also received several comments about the requirement that institutions provide the disclosures before a prospective student signs an enrollment agreement, completes registration, or makes a financial commitment to the institution. Some commenters recommended allowing institutions to obtain written confirmation from a student that they received a copy of the disclosure template at the same time as when the student signs an enrollment agreement, provided that new students are not penalized if they fail to attend any course sessions after the first seven days of the beginning of the term. Other commenters, in contrast, argued that the Department should specify a minimum length of time before students can enroll or make a financial commitment to the institution after receiving the disclosure template. Some of these commenters recommended instituting a minimum waiting period of three days after providing a prospective student with the disclosures before enrolling the student in order to provide students with sufficient time to review and understand the intricacies of their enrollment contracts.

Several commenters recommended allowing institutions to use a variety of means to confirm that the disclosures were provided to prospective students, including email messages, telephone calls, or other means that can be documented. The commenters argued that requiring written confirmation could complicate students’ planning and would pose significant compliance challenges for institutions. The commenters also noted that students often enroll at community colleges without selecting their course of study or program and that the regulations should reflect this reality.

Discussion: We agree with the commenters who stated that it is important that prospective students receive the information in the disclosure template directly and clearly prior to enrolling in a program. We recognize, however, that not all enrollment processes will take place in person, and that hand-delivering the disclosure template as a written, stand-alone document may not be feasible in all situations. In addition, in light of commenter confusion about how the student warning and disclosure template delivery requirements worked together in the proposed regulations, we believe that it would facilitate institutional compliance if the delivery requirements aligned, to the extent possible. Accordingly, we are revising §668.412(e) to provide that the same written delivery methods may be used to deliver the disclosure template as may be used to deliver student warnings. Specifically, the disclosure template may be provided to a prospective student or a third party acting on behalf of the prospective student by hand-delivering the disclosure template to the prospective student or third party individually or as part of a group presentation or sending the disclosure template as the only substantive content in an email to the primary email address used by the institution for communicating with the prospective student or third party about the program. As provided in the proposed regulations, the institution must obtain acknowledgement that the student or third party has received a copy of the disclosure template. If the disclosure template is delivered by hand, the acknowledgement must be in writing. If the disclosure template is sent by email to a prospective student or third party, an institution may satisfy the acknowledgement requirement through a variety of methods such as a pop-screen that asks the student to click “continue” or “I understand” before proceeding. Requiring these types of acknowledgements does not impose a significant burden on institutions or prospective students, yet provides adequate assurance that a prospective student has received important information about the program. Institutions must also maintain records of their efforts to provide the disclosure template.

We appreciate the commenters’ suggestions about additional and alternative methods for delivering the disclosure template to prospective students. Although we encourage institutions to consider innovative ways to deliver information about program outcomes to students, we believe that, to facilitate institutional compliance, it is preferable to have one, clear delivery requirement in the regulations. As discussed in the NPRM and elsewhere in this document, we will conduct consumer testing to test the manner of delivery of the disclosure template. In the course of consumer testing, we may also consult with one or more of the entities recommended by the commenters.

It is critical that prospective students receive the disclosure template before enrolling in a program so that the information on the template can inform their decision about whether to enroll in the program. Although we believe it is imperative that, for programs that are subject to the student warning requirement, prospective students have a cooling-off period between receiving the warning and enrolling in the program, it is not necessary for programs that are not at risk of losing eligibility based on their D/E rates for the next award year. In all cases, students will be able to access the information on the disclosure template through the program’s Web site and via its promotional materials prior to receiving the disclosure template directly from the institution.

Lastly, students must enroll in an eligible program in order to be eligible for title IV, HEA program funds. Any prospective student who has indicated that he or she intends to enroll in a GE program must be provided these disclosures.

Changes: We have revised §668.412(e) to specify that the disclosure template may be delivered to prospective students or a third party acting on behalf of the student by hand-delivering the disclosure template to the prospective student or third party individually or as part of a group presentation or sending the disclosure template to the primary email address used by the institution for communicating with the prospective student or third party about the program. We have also revised the regulations to require that, if the disclosure template is provided by email, the template must be the only substantive content in the email, the institution must receive written or other electronic acknowledgement of the prospective student’s or third party’s receipt of the disclosure template, and the institution must send the disclosure template using a different address or method of delivery if the institution receives a response that the email could not be delivered. We also have revised the regulations to require institutions to maintain records of their efforts to provide the disclosure template.

§668.413 Calculating, Issuing, and Challenging Completion Rates, Withdrawal Rates, Repayment Rates, Median Loan Debt, Median Earnings, and Program Cohort Default Rates

Completion and Withdrawal Rates

Comments: A commenter contended that students who are excluded from the D/E rates calculation under §668.404(e) should similarly be excluded from the completion and withdrawal rate calculations.

Discussion: In calculating the D/E rates and the repayment rate for a program, as provided in §668.404(e) and §668.413(b)(3)(vi) respectively, we exclude a student if he or she (1) has a loan that was in a military deferment status, (2) has a loan that may be discharged based on total and permanent disability, (3) was enrolled in another eligible program at the institution for which these rates are calculated or at another institution, or (4) died. We exclude these students because a student’s ability to work and have earnings or repay a loan could be diminished under any of the circumstances listed, which could adversely affect a program’s results, even though the circumstances are the result of student choices or unfortunate events that have nothing to do with program performance. Of these circumstances, only two are reasonably appropriate for holding an institution harmless for the purpose of determining completion and withdrawal rates--if the student died or became totally and permanently disabled while he or she was enrolled in the program. Therefore, as a general matter we agree to account for students in these two groups by excluding them from the completion and withdrawal rates.

However, our ability to identify these individuals is limited. For a student who borrowed, we may learn of a disability if the student has applied for or received a disability discharge of a loan. However, in instances where the individual seeks that discharge after the draft rates are calculated, or where we are not aware of a borrower’s death, the institution will have to provide relevant documentation during the challenge process described in §668.413(d) to support the exclusion. For a student who does not borrow, i.e., receives a Pell Grant only, we would not typically know if the student becomes disabled or dies while enrolled in the program. Again, the institution will have to identify and provide documentation to support the exclusion of these students during the challenge process described in §668.413(d).

For instances where an institution identifies a student who borrowed but has not applied for a disability discharge of a loan before the draft completion and withdrawal rates are calculated, or for a student who does not borrow, we will assess whether the student may be excluded from the calculation of the rates on the basis of a medical condition by applying the standard we use in §668.404(e)(2) to determine if the disability exclusion applies for the purpose of the D/E rates measure. Specifically, under 34 CFR 682.402(c)(5)-(6) and 34 CFR 685.213(b)(6)-(7), the Department reinstates a loan previously discharged on the basis of total and permanent disability if the borrower receives a loan after that previous loan was discharged. To be eligible for a loan, an individual must be enrolled to attend postsecondary school on at least a half-time basis. 34 CFR 685.200(a)(1). That is, the existing regulations infer that an individual who is able to attend school on at least a half-time basis is not totally and permanently disabled.

Accordingly, we are providing in §668.413(a)(2)(ii) that a student may be excluded from the calculation of the completion rates or withdrawal rates, as applicable, if the student became totally and permanently disabled while enrolled in the program and unable to continue enrollment on at least a half-time basis.

Changes: We have revised §668.413(b) to provide that a student who died while enrolled in the program is excluded from the enrollment cohort used for calculating completion and withdrawal rates. We have also provided in this section that a student who became totally and permanently disabled, while enrolled in the program, and who was unable to continue enrollment in school on at least a half-time basis, is excluded from the enrollment cohort used for calculating completion and withdrawal rates.

Comments: Some commenters expressed concern that disclosing four different completion rates would be excessive and potentially overwhelming for prospective students.

Discussion: Although we believe that the various completion rates would capture the experience of full-time and part-time students in a way that would be beneficial to both enrolled and prospective students, as well as institutions as they work to improve student outcomes, we agree that providing four completion rates on the disclosure template may be overwhelming for students and prospective students. Accordingly, as was the case in the NPRM, we have provided that we will use consumer testing to assess which of the disclosures, including the various options for completion and withdrawal rates, are most meaningful for students and prospective students. The disclosure template will include only those items identified by the Secretary as required disclosures for a particular year.

Changes: None.

Comments: Commenters asked the Department to clarify the methodology for calculating completion rates and withdrawal rates. Specifically, some commenters asked that we define the cohort of students for whom completion rates and withdrawal rates are calculated and address whether the cohort includes all students who received title IV, HEA program funds in a particular award year, or at any time in the past.

A number of commenters suggested that, for the purpose of calculating completion rates, we determine a student’s enrollment status at a fixed point after the start of a term, rather than on the first day of the student’s enrollment in the program, because many students may subsequently change their enrollment status. Another commenter suggested that institutions, and not the Department, calculate the completion and withdrawal rates that will be included in the disclosures.

Discussion: The Department will calculate the disclosure items indicated in §668.413 in order to ensure accuracy and consistency in the calculations.

With regard to the comments about the cohort used to calculate the completion and withdrawal rates, we clarify that the “enrollment cohort” is comprised of all the students who began enrollment in a GE program during a particular award year, where students are those individuals receiving title IV, HEA program funds. For example, all students who began enrollment in a GE program at any time during the 2011-2012 award year comprise the enrollment cohort for that award year. The Department will track the students in the enrollment cohort to calculate a completion rate at the end of the calendar date for each measurement period, i.e., at 100, 150, 200, and 300 percent of the length of the program. We will apply the same process for the next enrollment cohort for the program--the students who began enrollment during the 2012-2013 award year--and for every subsequent enrollment cohort for that program.

However, because students may enroll in a program at any time during an award year, we will determine on a student-by-student basis whether a student completed the program within the length of the program or the applicable multiple of the program. As an example, consider the calculation of the 100 percent of normal time completion rate associated with a two-year program for the students that enrolled in the program during the 2011-2012 award year, assuming that 100 students began enrollment in the program at various times during that award year. We will determine for each student individually whether he or she completed the program within two years by comparing for each student, the date the student began enrollment in the program to the date they completed the program. If, for example, 75 of those students completed the program within two years of when they began enrollment, the 100 percent of normal time completion rate for the 2011-2012 enrollment cohort would be 75 percent. Both completion and withdrawal rates under the regulations will be calculated using this methodology.

Changes: We have revised §668.413(b)(1) to clarify that the enrollment cohort for an award year represents the students who began the GE program at any time during that award year.

Repayment Rate

Comments: Some commenters asked whether there is a distinction made for the repayment rate calculation cohort period for medical or dental programs that require a residency.

Discussion: We see no reason to make a distinction in the cohort period for medical and dental programs that require a residency. For the D/E rates calculation, we adjust the cohort period because we would not expect students, while in a residency or other type of required training, to have earnings at a level that is reflective of the training they received. In comparison, we do expect borrowers to repay their loans while in residency or other training. Consequently, modifying the cohort period would not be appropriate.

Changes: None.

Comments: With respect to the repayment rate methodology in §668.413(b)(3), some commenters objected to the breadth of the exclusion for students enrolled in another eligible program at the institution or another institution, specifically noting the absence of any requirement that the institution provide documentation to validate the exclusion. On the other hand, some commenters supported the exclusion for borrowers currently enrolled in an eligible program regardless of whether it is the same program as that in which they originally enrolled, and for borrowers in military deferment. These commenters suggested expanding the exclusion to include other borrowers in deferment status, other than deferments for unemployment or economic hardship, including students working in the Peace Corps.

Discussion: The repayment rate disclosure will show consumers how effectively those who are expected to repay their loans are actually repaying them and, from that information, allow consumers to evaluate program performance. We exclude from the repayment rate calculation, as well as the D/E rates calculation, students who are in school or in military deferment because those statuses are reflective of individual choices that have little to do with the effectiveness of the program (see §668.412(a)(13) and §668.404(d)(3)). We decline to add an exclusion for borrowers in the Peace Corps because there is no longer a separate deferment in the title IV, HEA program regulations for such borrowers, and, therefore, there would be no way to easily identify these students from other students with an economic hardship deferment. As we do not expect the number of borrowers with an economic hardship deferment due to Peace Corps service to be significant, we believe the advantage to consumers of including all students in economic hardship status in the repayment rate calculation greatly outweighs any benefit from excluding all such students because they may include Peace Corps volunteers.

Changes: None.

Comments: Some commenters asserted that rehabilitated loans, which are defaulted loans subsequently paid in full or defaulted loans that returned to active repayment status, should not be treated as defaulted loans for the purpose of calculating loan repayment rates.

Discussion: We disagree that rehabilitated loans that were once in default should not be considered defaulted for the purpose of the repayment rate calculation. The repayment rate is intended to assess whether a program’s borrowers are able to manage their debt. A borrower’s default on a loan at some previous time, even if the loan is no longer in default status, indicates that the borrower was unable to manage his or her debt burden. This information should be reflected in a program’s repayment rate.

Changes: None.

Comments: One commenter contended that the determination of the outstanding balance for each of a borrower’s loans at the beginning and end of the award year is unduly complicated because of the need to prorate payments for the reporting of consolidated or multiple loans in a borrower’s loan profile. Further, the commenter suggested that measurement of active repayment of a borrower’s entire portfolio, possibly using a “weighted” method of calculating a student’s loan portfolio based on the amount of debt, would be more accurate and solve the potential problem of negative outcomes of simple proration for those earning higher degrees.

Discussion: We do not believe that the loan repayment rate calculations are overly complex. If a borrower has made a payment sufficient to reduce the outstanding balance of a consolidation loan during the measurement period, the borrower is included as a borrower in active repayment. A consolidation loan may have been used to pay off one or more original loans obtained for the program being measured, for that program and other programs offered by the same institution, or for that program and programs offered by other institutions. There is no practicable way to allocate payments made by a borrower among the components of the consolidation debt corresponding to the original loans, and the commenter proposed no reasonable basis to allocate payments made among a borrower’s original loan and other loans associated with other programs or other institutions. Regardless, the Department, and not the institution, calculates a program’s repayment rate using data already reported by the institution, so the burden of calculating the rates will fall on the Department, rather than the institution.

Changes: None.

Comments: Some commenters asserted that a borrower making full payments in an income-driven repayment plan, such as Income Based Repayment, Income Contingent Repayment, and Pay As You Earn, should count positively towards the program’s repayment rate by being included in the numerator of the calculation even if the borrower’s principal year-end balance is not reduced. These commenters argued that because the Department has made income-driven repayment plans available to borrowers to assist them in managing their debt, programs should not be penalized if a student takes advantage of such a plan as the institution does not have control over whether the plan will result in negative amortization.

Discussion: The loan repayment rate presents a simple measurement: the proportion of borrowers who are expected to be repaying their loans during a given year who are actually paying enough during that year to owe less at the end of the year than they owed at the start of the year (i.e. paid all interest and at least one dollar of principal). Income-driven plans are available to assist borrowers whose loan debt in relation to their income places them in a “partial financial hardship”; a program where many borrowers are forced to enroll in such plans is not leading to good outcomes. As a result, a repayment rate disclosure that treated such borrowers as in “active repayment” would not provide meaningful information to consumers about a program’s student outcomes and, worse, may give prospective students unrealistic expectations about the likely outcomes of their investment in such a program.

Changes: None.

Program Cohort Default Rate

Comments: None.

Discussion: As discussed in “§668.403 Gainful Employment Program Framework,” program cohort default rates will be used in the regulations as a potential disclosure under §668.412 only, rather than as a standard for determining program eligibility. To reflect that change, we are removing from §§668.407, 668.408, and 668.409 the provisions that established that the Secretary will use the methodology and procedures, including challenge procedures, in subpart R to calculate program cohort default rates; the provisions relating to the notice to institutions of their draft program cohort default rates; and the provisions relating to the issuance and publication of an official program cohort default rate.

Changes: We have revised §§668.413(b)-(f) to: establish that the Secretary will use the methodology and procedures, including challenge procedures, in subpart R to calculate program cohort default rates; and to incorporate provisions relating to the notice to institutions of their draft program cohort default rates and relating to the issuance and publication of an official program cohort default rate.

Comments: None.

Discussion: As discussed in “§668.403 Gainful Employment Program Framework,” program cohort default rates will be used in the regulations as a potential disclosure under §668.412 only, rather than as a standard for determining program eligibility, and we will use the procedures in subpart R to calculate the rate. However, certain sections of subpart R pertained to eligibility and are not necessary for these final regulations and we are removing those sections from the final regulations. Specifically, §668.506 of subpart R addressed the effect of a program cohort default rate on the continued eligibility of a program. Other provisions in subpart R governed challenges to the accuracy and completeness of the data used to calculate program cohort default rates and, additionally, appeals of results that might have led to loss of program eligibility. With respect to appeals, §668.513 would have permitted an institution to appeal a loss of eligibility based on academic success for disadvantaged students. Section 668.514 would have permitted an institution to appeal a loss of eligibility based on the number of students who borrowed title IV loans as a percentage of the total number of individuals enrolled in the program. Section 668.515 would have permitted an institution to appeal a loss of eligibility if at least two of the three program cohort default rates are calculated as average rates and would be less than 30 percent if calculated for the fiscal year alone. These provisions are being removed from the final regulations.

The provisions that remain serve the purpose of ensuring that the calculation process results in an accurate rate.

Changes: We have removed and reserved §§668.506, 668.513, 668.514, and 668.515 of subpart R.

Comments: One commenter objected to proposed §668.504(c)(1), which would allow an institution to submit a participation rate index challenge only to a draft program cohort default rate that could result in loss of eligibility of a program. The commenter believed that institutions should be allowed to assert a participation rate index challenge to any draft rate, because a successful assertion of a challenge, which would be relatively inexpensive and readily demonstrated, would eliminate the need to pursue more complicated, detailed, and costly challenges on other grounds to the draft and final program cohort default rates.

Discussion: As previously stated, in these regulations we are using program cohort default rates only as a disclosure. We therefore retain only those provisions of proposed subpart R that do not relate to loss of eligibility. Challenges based on a participation rate index would not have changed the calculation of the official rate, but would have only relieved the institution from loss of eligibility for the affected program. Because program cohort default rates will not affect eligibility, there is no reason to adopt a procedure that affected only whether the program would lose eligibility. The rate itself is useful information for consumers, and should be disclosed.

Changes: We have removed and reserved §668.504(c).

Comments: None.

Discussion: Proposed §668.502(a) provided that we would begin the program cohort default rate calculation process by counting whether at least 30 borrowers entered repayment in the fiscal year at issue; if fewer than 30 did so, we then counted whether at least 30 borrowers entered repayment in that year and the two preceding years. This approach conformed to institutional CDR requirements but is no longer applicable given that we are not adopting the program cohort default rate as an accountability metric. Because the rate will be used only as a disclosure, we will apply the minimum n-size of 10 that, as discussed in “§668.412 Disclosure Requirements for GE Programs,” we have established for all of the disclosure items.

This change requires a number of conforming changes to various provisions in subpart R. We are revising §§668.502, 668.504, and 668.516 to reflect the use of a minimum cohort size of 10 for the purposes of calculating, challenging, and appealing program cohort default rates.

Changes: We have revised §668.502(a) to provide for the Department to calculate a program cohort default rate for a program as long as that rate is based on a cohort of 10 or more borrowers. We also have revised §668.502(d) to reflect the use of cohorts with 10 or more borrowers in the calculation and §668.502(d)(2) describes how we will calculate the rate if there are fewer than 10 borrowers in a cohort for a fiscal year. We have made conforming changes in §668.504(a)(2) regarding draft program cohort default rates.

We have revised §668.516 to describe our determination of an official program cohort default rate more accurately and to provide that an institution may not disclose an official program cohort default rate under §668.412(a)(12) if the number of borrowers in the applicable cohorts is fewer than 10. As revised, §668.516 explains that we notify the institution if we determine that the applicable cohort has fallen to fewer than 10.

Comments: None.

Discussion: In considering the changes to subpart R previously described, we determined that as proposed, the regulations did not explicitly address how the Department, in the first two years that rates are calculated under the regulations, would calculate a program’s rate where the number of borrowers in the fiscal year was fewer than 10 and for which the Department would include in the calculation borrowers from the prior two fiscal years’ cohorts. In turn, the regulations did not explicitly address how an institution would challenge a program’s draft cohort default rate in these circumstances. Specifically, an institution would not have had an opportunity to challenge--at the draft rate stage--the data on borrowers from the prior two years, because the Department would not have calculated rates for those years. We are, therefore, revising §668.502(d)(2) and §668.504(a)(2) to clarify how this process will work to allow an opportunity to make that challenge.

Section 668.502(d)(2), as revised in these regulations, sets forth how the Department will calculate a program’s cohort default rate if there are fewer than 10 borrowers. Section 668.502(d)(2)(i) provides that, in the first two years that we calculate a program’s cohort default rate, we include in our calculation the number of borrowers in that cohort and in the two most recent prior cohorts for which we have relevant data. Under §668.502(d)(2)(ii), for other fiscal years, we include in our calculation the number of borrowers in the program cohort and in the two most recent program cohorts as previously calculated by the Department.

We are revising §668.504(a)(2) to provide that, except as set forth in §668.502(d)(2)(i), the draft cohort default rate of a program is always calculated using data for that fiscal year alone.

With these changes, we make it clear that the challenge process under §668.504(b) includes challenges with respect to rates with fewer than 10 borrowers in the first two years for which the Department uses data from the two most recent prior fiscal years.

Changes: We have revised §668.502(d)(2), and made conforming changes to §668.504(a)(2), to describe how the Department, in the first two years in which it calculates a program’s cohort default rates under these regulations, will calculate a rate for a program that has fewer than 10 borrowers in the fiscal year being measured and for which the Department uses data on borrowers from the prior two years to calculate the rate and to clarify that an institution may challenge that data once it receives its draft program cohort default rate or official program cohort default rate.

Comments: Some commenters objected to the adoption of institution level CDR rules in determining the cohort default rate of a program on the grounds that those rules measured only the percentage of borrowers who actually defaulted on their loans within the three-year period, without regard to the number who would likely have defaulted but were placed, often by reason of extensive efforts by the institution, in deferment or forbearance status so that default would likely be forestalled until after the close of the three-year period. These rules, they asserted, made CDR an inadequate measure of the repayment performance of the affected borrowers, and the commenters urged the Department to measure program cohort default rates using only the performance of borrowers who entered into repayment status and were not in deferment or forbearance status for a significant portion of the three-year period.

Discussion: As explained in the NPRM and in this preamble, we will calculate the program cohort default rate using the process and standards already used to calculate institutional cohort default rates, in part because institutions are already familiar with those procedures. We do not believe it would be appropriate to change the calculation method to exclude those in deferment or forbearance because it would lead to inconsistency between institutional CDR and program cohort default rates which could be confusing to consumers.

Changes: None.

Comments: One commenter asserted that in instances in which a cohort of borrowers entering repayment is very small, default by one or two borrowers may produce a failing program cohort default rate but that rate would not be meaningful information for consumers.

Discussion: As discussed, we agree that disclosures based on cohorts consisting of fewer than ten borrowers are not justified for privacy concerns, but we see no reason, and the commenter did not offer one, that a rate based on that number would not be useful to consumers. We note that each of the required disclosures must be made if the cohort on which the data are based includes 10 or more individuals, and that rate or data could always be affected by actions of a very small number. Nevertheless, we consider all that data useful to the consumer, and see no reason to designate some disclosures based on small numbers as useful, but others, such as default rate, as uninformative. An institution that considers a program cohort default rate to be misleading because the number of borrowers involved was small is free to provide that explanation to prospective students.

Changes: None.

Comments: One commenter noted that proposed §668.507 would give the Department discretion whether to include in the program cohort default rate calculation debt incurred for a GE program offered by another institution if the two institutions were under common ownership and control, but gave no indication of the conditions that would prompt the Department to do so. The commenter suggested that debt incurred at institutions under common ownership and control be included in the calculation for a program only if the institutions have the same accreditation and admission standards. The commenter contended that institutions with different accreditation and admission standards are so significantly independent that transfers from one to the other are not likely to be arranged in order to manipulate program cohort default rates, and that the regulations should not penalize an institution to which a borrower transfers in order to pursue a more advanced degree by attributing defaults at the institution from which the student is transferring to the institution to which the student is transferring.

Discussion: We believe that §668.503, which governs the determination of program cohort default rates for programs that have undergone a change in status such as a merger or acquisition, addresses situations in which debt will ordinarily be combined to calculate the rate. We also believe that, by using program cohort default rate as a disclosure only, rather than as an accountability metric, there is less incentive to attempt to manipulate this rate. We therefore do not believe further changes to the regulations are necessary.

Changes: None.

General

Comments: Some commenters expressed concern regarding the minimum size of a cohort for disclosure of repayment rates.

Discussion: With respect to the concerns raised by the commenters, for the 2011 Final Rules, the Department provided sub-regulatory guidance to institutions instructing them not to disclose various data for a program if fewer than 10 students completed the program in the most recently completed award year.  We believe this guidance continues to provide a useful bright line, and it remains in effect.  As discussed in “§668.412 Disclosure Requirements for GE Programs,” because of privacy concerns, an institution may not disclose data described in §668.413 if that data is derived from a cohort of fewer than 10 students, and, for those data calculated and issued by the Department, the Department does not issue or make public any data it calculates from such a cohort.

Changes: We have added paragraph (g) to §668.413 to provide that we do not publish determinations made by the Department under §668.413, and an institution may not disclose a rate or amount determined under that section, if the determination is based on a cohort of fewer than ten students.

§668.414 Certification Requirements for GE Programs

Comments: Several commenters supported the proposed program certification requirements because, they believed, the requirements are streamlined, clear, and feasible to implement.

Discussion: We appreciate the commenters’ support.

Changes: None.

Comments: A number of commenters objected to the program certification requirements.  They contended that States, accrediting agencies, and the Department serve different roles, and that requiring certifications would be inconsistent with that framework.  The commenters asserted it would be more appropriate for the Department to rely on States and accreditors to monitor whether institutions have obtained the necessary program approvals from them because independent monitoring by the Department would be derivative and duplicative of their efforts. The commenters also argued that program quality and outcomes are more appropriately evaluated by an institutional accreditor and, similarly, that determining whether a program meets a State’s standards should be the responsibility of the State. Finally, one commenter stated that the certification requirements would contravene the HEA’s recognition requirements with respect to program accreditors.

Discussion:  The Department agrees that accrediting agencies and States play important roles in approving institutions to operate and offer programs and providing ongoing oversight of whether institutions and programs meet those State and accrediting requirements. However, this may not always guarantee that a program meets all minimum educational standards for students to obtain employment in the occupation the institution identified as being associated with that training. For example, in some States, for some types of programs, institutions are allowed to offer a program even if it does not meet the requirements for licensure or certification in that State. In such instances, under the regulations, for a program to be eligible for title IV, HEA program funds, the program will be required to meet State licensure, certification, and accreditation standards for the occupations the institution identifies for the program where it would not have had to in the absence of the certification requirements.

Even where the certification requirements are partly duplicative of State and accreditor efforts, there is no conflict with the HEA to require an institution to verify that a program meets applicable State and accrediting standards in light of the Department’s responsibility to protect students and ensure that title IV, HEA program funds are used for proper purposes, in this case, to prepare students for gainful employment in a recognized occupation. We believe there is minimal burden associated with providing this information to the Department.

The certification requirements have the added benefit of creating an enforcement mechanism for the Department to take action if a required approval has been lost, or if a certification that was provided was false. Further, Federal and State law enforcement agencies may be able to prosecute any misrepresentations made by institutions in their own investigations and enforcement actions.

Changes: None.

Comments: One commenter, while noting support for the proposed provisions, suggested that institutions that do not satisfy all State or Federal program-level accrediting and licensing requirements should not be eligible to participate in the title IV, HEA programs.

Discussion: Institutions will be required to ensure that the programs they offer have the necessary Federal, State, and accrediting agency approvals to meet the requirements for the jobs associated with those programs. If a program does not meet these requirements, the institution will have to either obtain the necessary approvals or risk losing title IV, HEA program eligibility.

Changes: None.

Comments: A number of commenters asserted that the initial and continuing reporting requirements to update the certifications would be burdensome. They noted that for existing programs, institutions would be required to submit transitional certifications and reporting covering several years of data at the same time. The commenters were concerned that institutions would make unintentional errors for which they would be held liable. They were also concerned about how the implementation of the regulations would affect the timing of an institution’s PPA recertification.

Discussion: The Department estimates that there will be minimal additional administrative burden associated with the certification requirements. We believe that any burden is outweighed by the benefits of the requirements which, as described previously, will help ensure that programs meet minimum standards for students to obtain employment in the occupations for which they receive training. Furthermore, after the initial period where institutions will be required to submit transitional certifications for existing programs by December 31st of the year that the regulations take effect, the continuing certification procedure will be combined and synchronized with the existing PPA recertification process to minimize any increased institutional burden and facilitate compliance. This will have no bearing on the timing of an institution’s PPA recertification process. The only time an institution will need to update its existing program certification separately from the PPA recertification process will be when there is a change in the program or in its approvals that makes the existing program certification no longer accurate. Institutions will be required under 34 CFR 600.21 to update the program certification within 10 days of such a change. Regarding the commenter’s concern that the certification requirements will increase institutions’ possible liability and exposure to litigation, these requirements could affect complaints that are based upon violations of the new requirements but, in other cases, could also reduce complaints as students and prospective students receive better and more transparent information.

Changes: We have revised §668.14(b) to provide that an institution must update a program certification within 10 days of any change in the program or in its approvals that makes the existing certification no longer accurate. We have also made a conforming change to §600.21 to include program certifications in the list of items that an institution must update within 10 days.

Comments: One commenter suggested that the Department clarify in the regulations how the certifications would work together with the debt measures to establish that a program meets all of the gainful employment standards. Another commenter requested assurances that the existing certification requirements would continue to apply even after the D/E rates measure is implemented.

Discussion: The certification requirements are an independent pillar of the accountability framework of these regulations that complement the metrics-based standards. To determine whether a program provides training that prepares students for gainful employment as required by the HEA, these regulations provide procedures to establish a program’s eligibility and to measure its outcomes on a continuing basis. Accordingly, the certification requirements will continue to apply after the D/E rates measure becomes operational.

Changes: None.

Comments: Some commenters stated that providing certifications for GE programs would provide an important baseline for key information about a program, and suggested that the certification requirements should be expanded. In this regard, commenters argued that the Department should require institutions to affirm that programs lead to gainful employment for their graduates, add additional certification requirements for institutions with failing or zone programs, or require institutions that do not meet the certification requirements to pay monetary penalties.

Discussion: An expanded certification process as suggested by the commenters is unnecessary in light of the requirements already provided in the regulations. An important goal of the certification requirements is to ensure that institutions assess on an ongoing basis whether their programs meet all required Federal, State, and accrediting standards. Furthermore, we do not believe that additional certification requirements for institutions with failing or zone programs are needed, because the Department has existing procedures that consider an institution’s financial responsibility and administrative capability at least annually, and an institution with demonstrated problems, such as having failing or zone programs under the regulations, may be subject to additional restrictions and oversight. Consequently, an expanded certification process would add little to the existing requirements.

Changes: None.

Comments: One commenter recommended that we require institutions to provide separate certifications for programs by location.

Discussion: If a program does not meet the certification requirements in any State where an institution is located, then the program as a whole would be considered deficient and could not be certified. Consequently, we do not believe it is necessary to require separate certifications.

Changes: None.

Comments: Some commenters argued that institutions should be required to certify that their programs provide students with access to information about the licensure and certifications required by employers, or that meet industry standards nationwide, and provide an explanation to students of the certification options available in a particular field. The commenters suggested that the provision of such information by institutions would demonstrate that they are sufficiently aware of requirements for employment in the industries for which they are preparing students to work. Similarly, some commenters suggested that the regulations should require institutions to provide new data or information to students prior to enrolling to help them understand the certificates or licenses that are needed for a particular occupation so that the students can make better decisions. On the other hand, one commenter asserted that institutions should not be required to identify the licensure and certifications required by all employers.

Some commenters suggested more information about how a program provides training that prepares students for gainful employment should be included in the transitional certification along with an affirmation signed by the senior executive at the institution. Specifically, the commenters asserted that institutions should provide affirmations about job outcomes for programs subject to the transitional certification requirements because those programs are already participating in the title IV, HEA programs and information about their student outcomes is available.

Discussion: We appreciate the suggestion that more detailed information should be required as a part of the certifications, but believe that the regulations strike an appropriate balance between affirming that a program meets certain requirements while not creating ambiguity or increasing burden in providing more detailed statements about the program’s outcomes. Requiring institutions to certify that their programs provide the training necessary to obtain certifications expected by employers or industry organizations would be impractical as preferences will likely vary among employers and organizations. Without objective and reliable standards, such as those set by State or Federal agencies like the Federal Aviation Administration or the Department of Transportation, or by accrediting agencies, the Department would be unable to enforce such a requirement.

Further, we do not believe that requiring institutions to provide additional information in their certifications would further the objectives of these provisions as the certifications are limited in scope to whether a program meets certain objective minimum standards. Further, we believe that the D/E rates measure and required disclosures address the commenter’s suggestions.

Changes: None.

Comments: Some commenters made suggestions regarding the Department’s approval of institutions’ certifications. Specifically, one commenter asserted that the Department should give special consideration to whether programs that are significantly longer or require a higher credential than comparable programs should be approved.

Discussion: Because the Department does not review program content, it cannot make determinations about the appropriate credential level for a particular program. With respect to program length, additional requirements are not necessary because existing regulations at §668.14(b)(26) already provide that a program must demonstrate a reasonable relationship between the length of the program and entry-level requirements for employment in the occupation for which the program provides training. Under §668.14(b)(26), the relationship is considered to be reasonable if the number of clock hours of the program does not exceed by more than 50 percent the minimum number of clock hours required for training that has been established by the State in which the program is located. Also, where it is unclear whether a program’s length is excessive, the Department may check with the applicable State or accrediting agency to resolve the issue.

Changes: None.

Comments: Some commenters expressed concern that, for new programs, the proposed regulations would require an application only in those instances where the new program is the same, or substantially similar to, a failing or ineligible program offered by the same institution. These commenters noted that an institution could circumvent the certification process by misrepresenting a new program as not substantially similar to the failing, zone, or ineligible program.

Discussion: An institution that offered a program that lost eligibility, or that voluntarily discontinued a program when it was failing or in the zone under the D/E rates measure, may not offer a new program that is substantially similar to the ineligible, zone, or discontinued program for three years. We recognize the possibility that some institutions might make minor changes to a program and represent that the new program is not substantially similar to its predecessor. To address the commenter’s concern, we are removing the definition of “substantially similar” from the definition of CIP code, and establishing in §668.410 that two programs are substantially similar if they share a four-digit CIP code. We believe that precluding institutions from establishing new programs within the same four-digit CIP code will deter institutions from making small changes to a program solely for the purpose of representing that the new program is not substantially similar to the discontinued program, other than in instances where a program could be associated with a range of CIP codes, as suggested by the commenters. To address this concern that a similar program could be established using a different four-digit CIP code, we are revising §668.414(d)(4) to require an institution that is establishing a new program to explain in the program certification that is submitted to the Department how the new program is different from any program the institution offered that became ineligible or was voluntarily discontinued within the previous three years. The institution must also identify a CIP code for the new program. We will presume that a new program is not substantially similar to the ineligible or discontinued program if it does not share a four-digit CIP code with the other program. The certification and explanation reported by the institution may be reviewed on a case-by-case basis to determine if the two programs are not substantially similar. A program established in contravention of these provisions would be considered ineligible and the institution would be required to return the title IV, HEA program funds received for that program.

We believe that these changes will make it more difficult for an institution to continue to offer the same, or a similar program and claim that it is not substantially similar to an ineligible or discontinued program, while allowing an institution to establish new programs in different areas that may better serve their students.

Changes: We have revised the certification requirements to include a requirement in §668.414(d)(4) that an institution affirm in its certification that, and provide an explanation of how, a new program is not substantially similar to a program that became ineligible or was a zone or failing program that was voluntarily discontinued in the previous three years.

Comments: Several commenters urged the Department to create an approval process for all new programs before an institution could start enrolling students who receive title IV, HEA program funds to mitigate the risk of students incurring significant amounts of debt in programs unlikely to pass the D/E rates measure. One commenter suggested that limiting certifications to the PPA is not sufficient, and that applications for all new program approvals should require certification regarding licensing and certification.

While some commenters said approval requirements should apply to all new programs, other commenters suggested that an institution should be required to seek new program approval only if it had one or more failing programs at that time under the D/E rates measure, regardless of their similarity to the new program. Other commenters expressed the view that an institution that wished to build upon a successful existing program, such as by adding a graduate-level program, should be exempted from any new program approval process, or be subject to a streamlined approval process.

As a part of a new program approval requirement, some commenters proposed that institutions should have to certify that they conducted a reasoned analysis of the expected debt and earnings of graduates, as well as expected completion rates, and add that information to their PPA certification before starting any new program.

Discussion: The Department did not propose and is not including in the final regulations an approval process for new programs. As previously stated, we believe that the D/E rates measure is the best measure of whether a program prepares students for gainful employment. While we agree that it is important for institutions to conduct a reasoned analysis of expected program outcomes such as the expected debt and earnings of graduates, or expected completion rates, we will not require institutions to submit this information or certify that it was conducted because there is no basis upon which the Department could assess such information to determine whether the analysis was sufficient or that the analysis indicates that the program will indeed pass the D/E rates measure in the future. Without this ability, we do not believe adding such requirements would be useful.

Changes: None.

Comments: To increase transparency, some commenters suggested that an institution’s PPA, or the portions related to its GE programs, should be published on a public Web site to provide the public and policy makers the opportunity to assess the institution’s analysis, discussed in the previous comment, that the program would meet the D/E rates. They argued that this additional reporting should not be particularly burdensome for an institution because it should already be conducting such analysis. They also argued that the Department should strengthen its procedures to verify the accuracy and veracity of the information contained in a PPA, arguing that, otherwise, an institutional officer providing a false certification would have little risk of being identified and held accountable.

Discussion: As the Department is not requiring the analysis of potential debt and earnings outcomes as requested by the commenter, we are also declining to publish institutions’ PPAs. As discussed in “§668.412 Disclosure Requirements for GE Programs,” there also is little variation in the PPA and the disclosure and certification requirements already provide sufficient protections for students. Similarly, it would not be beneficial to modify procedures to verify the information contained in institutions’ PPAs. As with any representation made by an institution, the Department has the authority to investigate and take action against an institution that fraudulently misrepresents information in its PPA when those issues are identified during audits, program reviews, or when investigating complaints about an institution or program.

Changes:  None.

Comments:  Several commenters argued that six months is an insufficient amount of time for institutions to submit transitional certifications after the regulations become effective. They recommended increasing the time period or eliminating the transitional certifications altogether and require only that institutions provide the certifications as a part of their periodic PPA recertification.

Discussion:  The Department understands that there is some administrative burden associated with submitting the transitional program certifications.  However, programs should already be meeting the minimum requirements regarding accreditation, licensure, and certification, so the additional burden on institutions of providing this information should be minimal. This reporting burden is outweighed by the importance of promptly confirming after the regulations become effective that all programs meet the certification requirements. This will reduce the potential harm to students who become enrolled, or continued harm to students already enrolled, in programs that do not meet the minimum standards. If we were to wait until PPA recertification, a significant amount of time could pass before a program’s deficiencies would come to light during which students would continue to accumulate debt and exhaust title IV, HEA program eligibility in a program providing insufficient preparation.

Changes:  None.

Comments: Some commenters argued that institutions offering a program in multiple States might not meet the licensure, certification, and accreditation requirements in each State. They suggested that institutions should be prohibited from enrolling students in a State where these requirements are not met.  Other commenters recommended requiring institutions to disclose to students when a program does not meet the applicable certification requirements for the State where the student is located, but that the student should still be able to choose to enroll in that program.  Several commenters asserted that a student might still choose to enroll in such a program because the student intends to move to, and work in, a different State where the program would meet any applicable certification requirements.  

Some commenters criticized the requirements of the proposed regulations to obtain necessary programmatic accreditation and State-level approvals where the MSA within which they operate spans multiple States. Several commenters were concerned that it would be difficult for programs to meet the requirements of all of these States. The commenters stated that the State-MSA requirement could lead to confusion in a large MSA where an institution might not be aware of which governmental agencies have requirements and of differing requirements between States. One commenter suggested that the MSA requirement would be contrary to provisions in OMB Bulletin 13-01. Another commenter asserted that the State-MSA requirement would limit an institution’s ability to offer programs specialized to meet local labor market needs.

Some commenters argued that that the use of MSAs was not appropriate for online programs, because they are not bound by physical location. Other commenters asserted that the physical location of students should determine the relevant States whose requirements must be met rather than the physical location of institutions. They suggested that the certifications should apply to any State in which a sizeable number or plurality of students are enrolled.

Discussion:  We do not agree that it is too difficult for an institution to identify all of the governmental agencies that have licensure, certification, and accreditation requirements in the States that intersect with the MSA where a program is located. It is an institution’s responsibility to be aware of the requirements in the States where its students are likely to seek employment and ensure that their programs meet those requirements. However, we recognize that in some cases, State requirements may conflict in such a way that it would be impossible to concurrently meet the requirements of multiple States. For example, Ohio and Kentucky, which are a part of the Cincinnati, Ohio MSA, require nail technicians to receive a minimum of 200 and 600 clock hours of training, respectively, in order to obtain a license. However, the regulations at §668.14(b)(26) provide that the length of a program cannot exceed 150 percent of the minimum number of clock hours of training established by a State for the relevant occupation. In this case, a nail technician program in Cincinnati could not concurrently meet the requirements for both Ohio and Kentucky because a program length beyond 300 hours would violate §668.14(b)(26), jeopardizing the program’s title IV, HEA program eligibility. As a result, we are revising the regulations to remove the MSA certification requirement. However, institutions will still be required under §668.412(a)(14) to disclose whether a program meets applicable requirements in each State in the institution’s MSA.

We are addressing this potential conflict between different State requirements within an institution’s MSA by eliminating the proposal for program certifications to cover the States within an MSA, and requiring instead that the institution provide applicable program certifications in any State where the institution is otherwise required to obtain State approval under 34 CFR 600.9.

The current State authorization regulations apply to States where an institution has a physical location, and the program certification requirements also apply in those States so those two sets of requirements are aligned. If any changes are made in the future to extend the State authorization requirements in 34 CFR 600.9 to apply in other States, we intend the program certification requirements to remain aligned. Since institutions will have to ensure they maintain appropriate State approvals under the State authorization regulations, we anticipate that institutions will actively address any potential conflicts at that time. We believe that the requirements for the applicable program certifications should also be provided for those States. This will ensure a program and the institution that provides the program have the necessary State approvals for purposes of the Title IV, HEA programs. Linking the State certification requirements in §668.414(d)(2) with the State authorization regulations in §600.9 to identify States where institutions must obtain the applicable approvals benefits students and prospective students because the State authorization requirements include additional student protections for the students enrolled in the programs for which certifications would be required.

While institutions will not be prohibited from enrolling students in a program that does not meet the requirements of any particular State, a program that does not meet the applicable requirements in the State where it is located for the jobs for which it trains students will be ineligible to receive title IV, HEA program funds. As discussed in “§668.412 Disclosure Requirements for GE Programs,” institutions may be required to include on a program’s disclosure template whether the program meets the licensure, certification, and accreditation requirements of States, in addition to the States in the institution’s MSA, for which the institution has made a determination regarding those requirements so that students who intend to seek employment in those other States can consider this information before enrolling in the program.

Changes:  We have removed from §668.414 the requirement that an institution’s certification regarding programmatic accreditation and licensure and certification must be made with respect to each State that intersects with the program’s MSA. We have revised this section to require that the institution’s program certification is required in any State in which the institution is otherwise required to obtain State approval under 34 CFR 600.9.

§668.415 Severability

Comments: One commenter recommended that we omit the provisions of §668.415 regarding the severability of the provisions of subpart Q. Specifically, the commenter argued that the provisions of the regulations are too intertwined such that if a court found any part of the regulations invalid, it would not allow the remaining provisions to stand. In that event, the commenter argued, the remaining provisions would not serve the Department’s intent and the rulemaking process would be undermined.

Discussion: We believe that the provisions of subpart Q are severable. Each provision of subpart Q serves a distinct purpose within the accountability and transparency frameworks and provides value to students, prospective students, and their families and the public, taxpayers, and the Government that is separate from, and in addition to, the value provided by the other provisions. Although we recognize that severability is an issue to be decided by a court, §668.415 makes clear our intent that the provisions of subpart Q operate independently and the potential invalidity of one or more provisions should not affect the remainder of the provisions.¹⁸⁵

Changes: None.

Executive Orders 12866 and 13563

Regulatory Impact Analysis

Under Executive Order 12866, the Secretary must determine whether this regulatory action is “significant” and, therefore, subject to the requirements of the Executive order and subject to review by the Office of Management and Budget (OMB). Section 3(f) of Executive Order 12866 defines a “significant regulatory action” as an action likely to result in a rule that may--

(1) Have an annual effect on the economy of $100 million or more, or adversely affect a sector of the economy, productivity, competition, jobs, the environment, public health or safety, or State, local, or tribal governments or communities in a material way (also referred to as an “economically significant” rule);

(2) Create serious inconsistency or otherwise interfere with an action taken or planned by another agency;

(3) Materially alter the budgetary impacts of entitlement grants, user fees, or loan programs or the rights and obligations of recipients thereof; or

(4) Raise novel legal or policy issues arising out of legal mandates, the President's priorities, or the principles stated in the Executive order.

This final regulatory action will have an annual effect on the economy of more than $100 million because the estimated Federal student aid, institutional revenues, and instructional expenses associated with students that drop out of postsecondary education, transfer, or remain in programs that lose eligibility for title IV, HEA funds as a result of the regulations is over $100 million on an annualized basis. The estimated annualized costs and transfers associated with the regulations are provided in the “Accounting Statement” section of this Regulatory Impact Analysis (RIA). Therefore, this final action is “economically significant” and subject to review by OMB under section 3(f)(1) of Executive Order 12866. Notwithstanding this determination, we have assessed the potential costs and benefits, both quantitative and qualitative, of this final regulatory action and have determined that the benefits justify the costs.

We have also reviewed these regulations under Executive Order 13563, which supplements and explicitly reaffirms the principles, structures, and definitions governing regulatory review established in Executive Order 12866. To the extent permitted by law, Executive Order 13563 requires that an agency--

(1) Propose or adopt regulations only on a reasoned determination that their benefits justify their costs (recognizing that some benefits and costs are difficult to quantify);

(2) Tailor its regulations to impose the least burden on society, consistent with obtaining regulatory objectives and taking into account--among other things and to the extent practicable--the costs of cumulative regulations;

(3) In choosing among alternative regulatory approaches, select those approaches that maximize net benefits (including potential economic, environmental, public health and safety, and other advantages; distributive impacts; and equity);

(4) To the extent feasible, specify performance objectives, rather than the behavior or manner of compliance a regulated entity must adopt; and

(5) Identify and assess available alternatives to direct regulation, including economic incentives--such as user fees or marketable permits--to encourage the desired behavior, or provide information that enables the public to make choices.

Executive Order 13563 also requires an agency “to use the best available techniques to quantify anticipated present and future benefits and costs as accurately as possible.” The Office of Information and Regulatory Affairs of OMB has emphasized that these techniques may include “identifying changing future compliance costs that might result from technological innovation or anticipated behavioral changes.”

We are issuing these final regulations only on a reasoned determination that their benefits justify their costs. In choosing among alternative regulatory approaches, we selected those approaches that maximize net benefits. Based on the analysis that follows, the Department believes that these final regulations are consistent with the principles in Executive Order 13563.

We also have determined that this regulatory action does not unduly interfere with State, local, or tribal governments in the exercise of their governmental functions.

In this regulatory impact analysis we discuss the need for regulatory action, the potential costs and benefits, net budget impacts, assumptions, limitations, and data sources, as well as regulatory alternatives we considered.

Elsewhere in this section, under Paperwork Reduction Act of 1995, we identify and explain burdens specifically associated with information collection requirements.

A detailed analysis, including our Regulatory Flexibility Analysis, is found in Appendix A to this document.

Paperwork Reduction Act of 1995

The Paperwork Reduction Act of 1995 does not require you to respond to a collection of information unless it displays a valid OMB control number. We display the valid OMB control numbers assigned to the collections of information in these regulations at the end of the affected sections of the regulations.

Sections 668.405, 668.406, 668.410, 668.411, 668.412, 668.413, 668.414, 668.504, 668.509, 668.510, 668.511, and 668.512 contain information collection requirements. Under the Paperwork Reduction Act of 1995 (PRA) (44 U.S.C. 3507(d)), the Department has submitted a copy of these sections, related forms, and Information Collection Requests (ICRs) to the Office of Management and Budget (OMB) for its review.

The OMB Control numbers associated with the regulations and related forms are 1845-0123 (identified as 1845-NEW1 in the NPRM), 1845-0122 (identified as 1845-NEW2 in the NPRM), and 1845-0121 (identified as 1845-NEW3 in the NPRM). Due to the removal of the pCDR measure as an accountability metric, the number of GE programs and enrollments in those programs have been reduced throughout this section.

§668.405 Issuing and Challenging D/E Rates

Requirements: Under the regulations, the Secretary will create a list of students who completed a GE program during the applicable cohort period from data reported by the institution. The list will indicate whether the list is of students who completed the program in the two-year cohort period or in the four-year cohort period, and it will also indicate which of the students on the list will be excluded from the debt-to-earnings (D/E) rates calculations under §668.404(e), for one of the following reasons: a military deferment, a loan discharge for total and permanent disability, enrollment on at least a half-time basis, completing a higher undergraduate or graduate credentialed program, or death.

The institution will then have the opportunity, within 45 days of being provided the student list from the Secretary, to propose corrections to the list. After receiving the institution’s proposed corrections, the Secretary will notify the institution whether a proposed correction is accepted and will use any corrected information to create the final list.

Burden Calculation: We have estimated that the 2010-2011 and the 2011-2012 total number of students enrolled in GE programs is projected to be 6,436,806 (the 2010-2011 total of 3,341,856 GE students plus the 2011-2012 total of 3,094,950 GE students).

We estimate that 89 percent of the total enrollment in GE programs will be at for-profit institutions, 2 percent will be at private non-profit institutions, and 9 percent will be at public institutions. As indicated in connection with the 2011 Final Rules (75 FR 66933), we estimate that 16 percent of students enrolled in GE programs will complete their course of study. Therefore, we estimate that there will be 916,601 students who complete their programs at for-profit institutions (6,436,806 students times 89 percent of total enrollment at for-profit institutions times 16 percent, the percentage of students who complete programs) during the two-year cohort period.

On average, we estimate that it will take for-profit institutional staff 0.17 hours (10 minutes) per student to review the list to determine whether a student should be included or excluded under §668.404(e) and, if included, whether the student's identity information requires correction, and then to obtain the evidence to substantiate any inclusion, exclusion, or correction, increasing burden by 155,822 hours (916,601 students times .17 hours) under OMB 1845-0123.

We estimate that there will be 20,598 students who complete their programs at private non-profit institutions (6,436,806 students times 2 percent of total enrollment at private non-profit institutions times 16 percent, the percentage of students who complete programs) during the two-year cohort period.

On average, we estimate that it will take private non-profit institutional staff 0.17 hours (10 minutes) per student to review the list to determine whether a student should be included or excluded under §668.404(e) and, if included, whether the student's identity information requires correction, and then to obtain the evidence to substantiate any inclusion, exclusion, or correction, increasing burden by 3,502 hours (20,598 students times .17 hours) under OMB 1845-0123.

We estimate that there will be 92,690 students who complete their programs at public institutions (6,436,806 students times 9 percent of the total enrollment at public institutions times 16 percent, the percentage of students who complete programs) during the two-year cohort period.

On average, we estimate that it will take public institutional staff 0.17 hours (10 minutes) per student to review the list to determine whether a student should be included or excluded under §668.404(e) and, if included, whether the student's identity information requires correction, and then to obtain the evidence to substantiate any inclusion, exclusion, or correction, increasing burden by 15,757 hours (92,690 students times .17 hours) under OMB 1845-0123.

Collectively, the total number of students who complete their programs and who will be included on the lists that will be provided to institutions to review for accuracy is a projected 1,029,889 students, thus increasing burden by 175,081 hours under OMB Control Number 1845-0123.

Requirements: Under §668.405(d), after finalizing the list of students, the Secretary will obtain from SSA the mean and median earnings, in aggregate form, of those students on the list whom SSA has matched to its earnings data for the most recently completed calendar year for which SSA has validated earnings information. SSA will not provide to the Secretary individual data on these students; rather, SSA will advise the Secretary of the number of students it could not, for any reason, match against its records of earnings. In the D/E rates calculation, the Secretary will exclude from the loan debts of the students on the list the same number of loan debts as SSA non-matches, starting with the highest loan debt. The remaining debts will then be used to calculate the median debt for the program for the listed students. The Secretary will calculate draft D/E rates using the higher of the mean or median annual earnings reported by SSA under §668.405(e), notify the institution of the GE program's draft D/E rates, and provide the institution with the individual loan data on which the rates were calculated.

Under §668.405(f), the institution will have the opportunity, within 45 days of the Secretary's notice of the draft D/E rates, to challenge the accuracy of the rates, under procedures established by the Secretary. The Secretary will notify the institution whether a proposed challenge is accepted and use any corrected information from the challenge to recalculate the GE program's draft D/E rates.

Burden Calculation: There are 8,895 programs that will be evaluated under the regulations. Our analysis estimates that of those 8,895 programs, with respect to the D/E rates measure, 6,913 programs will be passing, 1,253 programs will be in the zone, and 729 programs will fail.

We estimate that the number of students at for-profit institutions who complete programs that are in the zone will be 77,693 (485,583 students enrolled in zone programs times 16 percent, the percentage of students who complete programs) and the number who complete failing programs at for-profit institutions will be 66,200 (413,747 students enrolled in failing programs times 16 percent, the percentage of students who complete programs), for a total of 143,893 students (77,693 students plus 66,200 students).

We estimate that it will take institutional staff an average of 0.25 hours (15 minutes) per student to examine the loan data and determine whether to select a record for challenge, resulting in a burden increase of 35,973 hours (143,893 students times .25 hours) in OMB Control Number 1845-0123.

We estimate that the number of students at private non-profit institutions who complete programs that are in the zone will be 760 (4,747 students enrolled in zone programs times 16 percent, the percentage of students who complete programs) and the number who complete failing programs at private non-profit institutions will be 272 (1,701 students enrolled in failing programs times 16 percent, the percentage of students who complete programs), for a total of 1,032 students (760 students plus 272 students).

We estimate that it will take institutional staff an average of 0.25 hours (15 minutes) per student to examine the loan data and determine whether to select a record for challenge, resulting in a burden increase of 258 hours (1,032 students times .25 hours) in OMB Control Number 1845-0123.

We estimate that the number of students at public institutions who complete programs that are in the zone will be 109 (684 students enrolled in zone programs times 16 percent, the percentage of students who complete programs) and the number who complete failing programs at public institutions will be 84 (523 students enrolled in failing programs times 16 percent, the percentage of students who complete programs), for a total of 193 students (109 students plus 84 students).

We estimate that it will take institutional staff an average of 0.25 hours (15 minutes) per student to examine the loan data and determine whether to select a record for challenge, resulting in a burden increase of 48 hours (193 students times .25 hours) in OMB Control Number 1845-0123.

Collectively, the burden for institutions to examine loan records and to determine whether to make a draft D/E rates challenge will increase burden by 36,279 hours under OMB Control Number 1845-0123.

The total increase in burden for §668.405 will be 211,360 hours under OMB Control Number 1845-0123.

§668.406 D/E Rates Alternate Earnings Appeals

Alternate Earnings Appeals

Requirements: The regulations will allow an institution to submit to the Secretary an alternate earnings appeal if, using data obtained from SSA, the Secretary determined that the program was failing or in the zone under the D/E rates measure. In submitting an alternate earnings appeal, the institution will seek to demonstrate that the earnings of students who completed the GE program in the applicable cohort period are sufficient to pass the D/E rates measure. The institution will base its appeal on alternate earnings evidence from either a survey conducted in accordance with standards included on an Earnings Survey Form developed by NCES or from State-sponsored data systems.

In either instance, the alternate earnings data will be from the same calendar year for which the Secretary obtained earnings data from SSA for use in the D/E rates calculations.

An institution with a GE program that is failing or in the zone that wishes to submit alternate earnings appeal information must notify the Secretary of its intent to do so no earlier than the date that the Secretary provides the institution with its draft D/E rates and no later than 14 business days after the date the Secretary issues the notice of determination of the program's D/E rates. No later than 60 days after the date the Secretary issues the notice of determination, the institution must submit its appeal information under procedures established by the Secretary. The appeal information must include all supporting documentation related to recalculating the D/E rates using alternate earnings data.

Survey: An institution that wishes to submit an appeal by providing survey data must include in its survey all the students who completed the program during the same cohort period that the Secretary used to calculate the final D/E rates under §668.404 or a comparable cohort period, provided that the institution may elect to exclude from the survey population all or some of the students excluded from the D/E rates calculation under §668.404(e).

The Secretary will publish in the Federal Register an Earnings Survey Form developed by NCES. The Earnings Survey Form will be a pilot-tested universe survey that may be used by an institution in accordance with the survey standards, such as a required response rate or subsequent non-response bias analysis that the institution must meet to guarantee the validity and reliability of the results. Although use of the pilot-tested universe survey will not be required and the Earnings Survey Form will be provided by NCES only as a service to institutions, an institution that chooses not to use the Earnings Survey Form will be required to conduct its survey in accordance with the published NCES standards, including presenting to the survey respondent, in the same order and in the same manner, the same survey items included in the NCES Earnings Survey Form.

Under the regulations, the institution will certify that the survey was conducted in accordance with the standards of the NCES Earnings Survey Form and submit an examination-level attestation engagement report prepared by an independent public accountant or independent governmental auditor, as appropriate. The attestation will be conducted in accordance with the attestation standards contained in the GAO's Government Auditing Standards promulgated by the Comptroller General of the United States and with procedures for attestations contained in guides developed by, and available from, the Department's Office of Inspector General.

Burden Calculation: We estimate that for-profit institutions will have 1,225 gainful employment programs in the zone and that 718 programs will be failing for a total of 1,943 programs. We expect that most institutions will determine that SSA data reflect accurately the earnings of students and will therefore not elect to conduct the survey. Accordingly, we estimate that for-profit institutions will submit alternate earnings appeals under the survey appeal option for 10 percent of those programs, which will equal 194 appeals annually. We estimate that conducting the survey, providing the institutional certification, and obtaining the examination-level attestation engagement report will total, on average, 100 hours of increased burden, therefore burden will increase 19,400 hours (194 survey appeals times 100 hours) under OMB Control Number 1845-0122.

We estimate that private-non-profit institutions will have 20 gainful employment programs in the zone and that 8 programs will be failing for a total of 28 programs. We expect that most institutions will determine that SSA data reflect accurately the earnings of students and will therefore not elect to conduct the survey.

Accordingly, we estimate that private non-profit institutions will submit alternate earnings appeals under the survey appeal option for 10 percent of those programs, which will equal 3 appeals annually. We estimate that conducting the survey, providing the institutional certification, and obtaining the examination-level attestation engagement report will total, on average, 100 hours of increased burden, therefore burden will increase 300 hours (3 survey appeals times 100 hours) under OMB Control Number 1845-0122.

We estimate that public institutions will have 8 gainful employment programs in the zone and that 3 programs will be failing for a total of 11 programs. We expect that most institutions will determine that SSA data reflect accurately the earnings of students and will therefore not elect to conduct the survey. Accordingly, we estimate that public institutions will submit alternate earnings appeals under the survey appeal option for 10 percent of those programs, which will equal 1 appeal annually. We estimate that conducting the survey, providing the institutional certification, and obtaining the examination-level attestation engagement report will total, on average, 100 hours of increased burden, therefore burden will increase 100 hours (1 survey appeals times 100 hours) under OMB Control Number 1845-0122.

Collectively, the projected burden associated with conducting an alternative earnings survey will increase burden by 19,800 hours under OMB Control Number 1845-0122.

State data systems

An institution that wishes to submit an appeal by providing State data will include in the list it submits to the State or States all the students who completed the program during the same cohort period that the Secretary used to calculate the final D/E rates under §668.404 or a comparable cohort period, provided that the institution may elect to exclude from the survey population all or some of the students excluded from the D/E rates calculated under §668.404(e). The earnings information obtained from the State or States must match 50 percent of the total number of students included on the institution's list, and the number matched must be 30 or more.

Burden Calculation: We estimate that there will be 718 failing GE programs at for-profit institutions and 1,225 programs in the zone, for a total of 1,943 programs. We expect that most institutions will determine that SSA data reflect accurately the earnings of students who completed a program and will therefore not elect to submit earnings data from a State-sponsored system. Accordingly, we estimate that in 10 percent of those cases, institutions will obtain earnings data from a State-sponsored system, resulting in approximately 194 appeals.

We estimate that, on average, each appeal will take 20 hours, including execution of an agreement for data sharing and privacy protection under the Family Educational Rights and Privacy Act (20 U.S.C. 1232g) (FERPA) between the institution and a State agency (when the State agency is located in a State other than the State in which the institution resides), preparing the list(s), submitting the list(s) to the appropriate State agency, reviewing the results, calculating the revised D/E rates, and submitting those results to the Secretary. Therefore, burden will increase by 3,880 hours (194 State system appeals times 20 hours) under OMB Control Number 1845-0122.

We estimate that there will be 8 failing GE programs at private non-profit institutions and 20 programs in the zone, for a total of 28 programs. We expect that most institutions will determine that SSA data reflect accurately the earnings of students who completed a program and will therefore not elect to submit earnings data from a State-sponsored system. Accordingly, we estimate that in 10 percent of those cases, institutions will obtain earnings data from a State-sponsored system, resulting in 3 appeals.

We estimate that, on average, each appeal will take 20 hours, including execution of an agreement for data sharing and privacy protection under FERPA between the institution and a State agency (when the State agency is located in a State other than the State in which the institution resides), preparing the list(s), submitting the list(s) to the appropriate State agency, reviewing the results, calculating the revised D/E rates, and submitting those results to the Secretary. Therefore burden will increase by 60 hours (3 State system appeals times 20 hours) under OMB Control Number 1845-0122.

We estimate that there will be 3 failing GE programs at public institutions and 8 programs in the zone, for a total of 11 programs. We expect that most institutions will determine that SSA data reflect accurately the earnings of students who completed a program and will therefore not elect to submit earnings data from a State-sponsored system. Accordingly, we estimate that in 10 percent of those cases institutions will obtain earnings data from a State-sponsored system, resulting in approximately 1 appeal. We estimate that, on average, each appeal will take 20 hours, including execution of an agreement for data sharing and privacy protection under FERPA between the institution and a State agency (when the State agency is located in a State other than the State in which the institution resides), preparing the list(s), submitting the list(s) to the appropriate State agency, reviewing the results, calculating the revised D/E rates, and submitting those results to the Secretary. Therefore, burden will increase by 20 hours (1 State system appeal times 20 hours) under OMB Control Number 1845-0122.

Collectively, the projected burden associated with conducting an alternative earnings based on State data systems will increase burden by 3,960 hours under OMB Control Number 1845-0122.

Requirements: Under the regulations, to pursue an alternate earnings appeal, the institution must notify the Secretary of its intent to submit an appeal. This notification must be made no earlier than the date the Secretary provides the institution with draft D/E rates and no later than 14 business days after the Secretary issues the final D/E rates.

Burden Calculation: We estimated above that for-profit institutions will have 194 alternate earnings survey appeals and 194 State-sponsored data system appeals, for a total of 388 appeals per year. We estimate that completing and submitting a notice of intent to submit an appeal will take, on average, 0.25 hours per submission or 97 hours (388 submissions times 0.25 hours) under OMB Control 1845-0122.

We estimated above that private non-profit institutions will have 3 alternate earnings survey appeals and 3 State-sponsored data system appeals, for a total of 6 appeals per year. We estimate that completing and submitting a notice of intent to submit an appeal will take, on average, 0.25 hours per submission or 2 hours (6 submissions times 0.25 hours) under OMB Control 1845-0122.

We estimated above that public institutions will have 1 alternate earnings survey appeal and 1 State-sponsored data system appeal, for a total of 2 appeals per year. We estimate that completing and submitting a notice of intent to submit an appeal will take, on average, 0.25 hours per submission or 1 hour (2 submissions times 0.25 hours) under OMB Control 1845-0122.

Collectively, the projected burden associated with completing and submitting a notice of intent will increase burden by 100 hours under OMB Control Number 1845-0122.

The total increase in burden for §668.406 will be 23,860 hours under OMB Control Number 1845-0122.

§668.410 Consequences of the D/E Rates Measure

Requirements: Under §668.410(a), we require institutions to provide warnings to students and prospective students in any year for which the Secretary notifies an institution that the program could become ineligible based on its final D/E rates measure for the next award year. Within 30 days after the date of the Secretary’s notice of determination under §668.409, the institution must provide a written warning directly to each student enrolled in the program. To the extent practicable, an institution must provide this warning in other languages for enrolled students for whom English is not their first language.

In the warning, an institution must describe the options available to the student to continue his or her education in the event that the program loses its eligibility for title IV, HEA program funds. Specifically, the warning will inform the student of academic and financial options available to continue his or her education at the institution; whether the institution will allow the student to transfer to another program at the institution; continue to provide instruction in the program to allow the student to complete the program; whether the student’s earned credits could be transferred to another institution; or refund the tuition, fees, and other required charges paid by, or on behalf of, the student to enroll in the program.

Under §668.410(a)(5), an affected institution must provide a written warning by hand-delivering it individually or as part of a group presentation, or via email.

Burden Calculation: We estimate that the written warnings will be hand-delivered to 10 percent of the affected students, delivered through a group presentation to another 10 percent of the affected students, and delivered through the student's primary email address used by the institution to the remaining 80 percent. Based upon 2009-2010 reported data, 2,703,851 students were enrolled at for-profit institutions. Of that number, we estimate that 327,468 students were enrolled in zone programs and 844,488 students were enrolled in failing programs at for-profit institutions. Thus, the warnings will have to be provided to 1,171,956 students (327,468 students plus 844,488 students) enrolled in GE programs at for-profit institutions.

Of the 1,171,956 projected number of warnings to be provided to enrolled students at for-profit institutions, we estimate that 117,196 students (1,171,956 students times 10 percent) will receive the warning individually and that it will take on average 0.17 hours (10 minutes) per warning to print the warning, locate the student, and deliver the warning to each affected student. This will increase burden by 19,923 hours (117,196 students times 0.17 hours) under OMB Control Number 1845-0123.

Of the 1,171,956 projected warnings to be provided to enrolled students at for-profit institutions, we estimate that 117,196 students (1,171,956 students times 10 percent) will receive the warning at a group presentation and that it will take on average 0.33 hours (20 minutes) per warning to print the warning, conduct the presentation, and answer questions about the warning to each affected student. This will increase burden by 38,675 hours (117,196 times 0.33 hours) under OMB Control Number 1845-0123.

Of the 1,171,956 projected warnings to be provided to enrolled students at for-profit institutions, we estimate that 937,564 students (1,171,956 students times 80 percent) will receive the warning via email and that it will take on average 0.017 hours (1 minute) per warning to send the warning to each affected student. This will increase burden by 15,939 hours (937,565 students times 0.017 hours) under OMB Control Number 1845-0123.

Based upon 2009-2010 reported data, 57,700 students were enrolled at private non-profit institutions. Of that number of students, we estimate that 2,308 students will be enrolled in zone programs and 5,423 students will be enrolled in failing programs at private non-profit institutions. Thus, the warnings will have to be provided to 7,731 students (2,308 students plus 5,423 students) enrolled in GE programs at private non-profit institutions.

Of the 7,731 projected number of warnings to be provided to enrolled students at non-profit institutions, we estimate that 773 students (7,731 students times 10 percent) will receive the warning individually and that it will take on average 0.17 hours (10 minutes) per warning to print the warning, locate the student, and deliver the warning to each affected student. This will increase burden by 131 hours (773 students times 0.17 hours) under OMB Control Number 1845-0123.

Of the 7,731 projected warnings to be provided to enrolled students at non-profit institutions, we estimate that 773 students (7,731 students times 10 percent) will receive the warning at a group presentation and that it will take on average 0.33 hours (20 minutes) per warning to print the warning, conduct the presentation, and answer questions about the warning to each affected student. This will increase burden by 255 hours (773 times 0.33 hours) under OMB Control Number 1845-0123.

Of the 7,731 projected warnings to be provided to enrolled students at non-profit institutions, we estimate that 6,185 students (7,731 students times 80 percent) will receive the warning via email and that it will take on average 0.017 hours (1 minute) per warning to send the warning to each affected student. This will increase burden by 105 hours (6,185 students times 0.017 hours) under OMB Control Number 1845-0123.

Based upon 2009-2010 reported data, 276,234 students were enrolled at public institutions. Of that number of students, we estimate that 628 students will be enrolled in zone programs and 13,178 students will be enrolled in failing programs at public institutions. Thus, the warnings will have to be provided to 13,806 students (628 students plus 13,178 students) enrolled in GE programs at public institutions.

Of the 13,806 projected number of warnings to be provided to enrolled students at public institutions, we estimate that 1,381 students (13,806 students times 10 percent) will receive the warning individually and that it will take on average 0.17 hours (10 minutes) per warning to print the warning, locate the student, and deliver the warning to each affected student. This will increase burden by 235 hours (1,381 students times 0.17 hours) under OMB Control Number 1845-0123.

Of the 13,806 projected warnings to be provided to enrolled students at public institutions, we estimate that 1,381 students (13,806 students times 10 percent) will receive the warning at a group presentation and that it will take on average 0.33 hours (20 minutes) per warning to print the warning, conduct the presentation, and answer questions about the warning to each affected student. This will increase burden by 456 hours (1,381 times 0.33 hours) under OMB Control Number 1845-0123.

Of the 13,806 projected warnings to be provided to enrolled students at public institutions, we estimate that 11,044 students (13,806 students times 80 percent) will receive the warning via email and that it will take on average 0.017 hours (1 minute) per warning to send the warning to each affected student. This will increase burden by 188 hours (11,044 students times 0.017 hours) under OMB Control Number 1845-0123.

Collectively, providing the warnings will increase burden by 75,907 hours under OMB Control Number 1845-0123.

Students will also be affected by the warnings. On average, given the alternatives available to institutions, we estimate that it will take each student 0.17 hours (10 minutes) to read the warning and ask any questions.

Burden will increase by 199,233 hours (1,171,956 students times 0.17 hours) for the students who will receive warnings from for-profit institutions under one of the three delivery options, under OMB Control Number 1845-0123.

Burden will increase by 1,314 hours (7,731 students times 0.17 hours) for the students who will receive warnings from private non-profit institutions under one of the three delivery options, under OMB Control Number 1845-0123.

Burden will increase by 2,347 hours (13,806 students times 0.17 hours) for the students who will receive warnings from public institutions under one of the three delivery options, under OMB Control Number 1845-0123.

Collectively, students reading the warning will increase burden by 202,894 hours under OMB Control Number 1845-0123.

Requirements: Under §668.410(a)(6)(ii), institutions must provide a warning about a possible loss of eligibility for title IV, HEA program funds directly to prospective students prior to their signing an enrollment agreement, registering, or making any financial commitment to the institution. The warning may be hand-delivered as a separate warning, or as part of a group presentation, or sent via email to the primary email address used by the institution for communicating with prospective students. To the extent practicable, an institution will have to provide this warning in other languages for those students and prospective students for whom English is not their first language.

Burden Calculation: Most institutions will have to contact, or be contacted by, a larger number of prospective students to yield institutions’ desired net enrollments. The magnitude of this activity will be different depending on the type and control of the institution, as detailed below.

We estimate that the number of prospective students that must contact or be contacted by for-profit institutions will be 6 times the number of expected enrollments. As noted above, we estimate that 1,171,956 students (327,468 students enrolled in zone programs plus 844,488 students enrolled in failing programs) will be enrolled in programs at for-profit institutions that require a warning to students and prospective students. Therefore, for-profit institutions will be required to provide 7,031,736 warnings (1,171,956 times 6), with an estimated per student time of 0.10 hours (6 minutes) to deliver, increasing burden by 703,174 hours (7,031,736 prospective students times 0.10 hours) under OMB Control Number 1845-0123.

We estimate that the number of prospective students that must contact or be contacted by private non-profit institutions will be 1.8 times the number of expected enrollments. As noted above, we estimate that 7,731 students (2,308 students enrolled in zone programs plus 5,423 students enrolled in failing programs) will be enrolled in programs at private non-profit institutions that require a warning to students and prospective students. Therefore, private non-profit institutions will be required to provide 13,916 warnings (7,731 students times 1.8), with an estimated per student time of 0.10 hours (6 minutes) to deliver, increasing burden by 1,392 hours (13,916 prospective students times 0.10 hours) under OMB Control Number 1845-0123.

We estimate that the number of prospective students that must contact or be contacted by public institutions will be 1.5 times the number of expected enrollments. As noted above, we estimate that 13,806 students (628 students enrolled in zone programs plus 13,178 students enrolled in failing programs) will be enrolled in programs at public institutions that require a warning to students and prospective students. Therefore, public institutions will be required to provide 20,709 warnings (13,806 students times 1.5), with an estimated per student time of 0.10 hours (6 minutes) to deliver, increasing burden by 2,071 hours (20,709 prospective students times 0.10 hours) under OMB Control Number 1845-0123.

Collectively, burden will increase by 706,637 hours under OMB Control Number 1845-0123.

The prospective students will also be affected by the warnings. On average, given the alternatives available to institutions, we estimate that it will take each student 0.08 hours (5 minutes) to read the warning and ask any questions.

Burden will increase by 562,539 hours (7,031,736 times 0.08 hours) for the prospective students who will receive warnings from for-profit institutions, under OMB Control Number 1845-0123.

Burden will increase by 1,113 hours (13,916 times 0.08 hours) for the prospective students who will receive warnings from private non-profit institutions, under OMB Control Number 1845-0123.

Burden will increase by 1,657 hours (20,709 times 0.08 hours) for the prospective students who will receive warnings from public institutions, under OMB Control Number 1845-0123.

Collectively, prospective students reading the warning will increase burden by 565,309 hours under OMB Control Number 1845-0123.

Requirements: Under §668.410(a)(6)(ii)(B)(2), if more than 30 days have passed from the date the initial warning is provided, the prospective student must be provided an additional written warning and may not enroll until three business days later.

Burden Calculation: We estimate that 50 percent of students enrolling in a failing program will do so more than 30 days after receiving the initial prospective student warning. Burden for institutions will increase by 281,269 hours for the 3,515,868 students (7,031,736 prospective students times 50 percent times .08 hours) for whom for-profit institutions must provide subsequent warnings.

Burden will increase by 557 hours for the 6,958 students (13,916 prospective students times 50 percent times .08 hours) for whom private non-profit institutions will provide subsequent warnings.

Burden will increase by 828 hours for the 10,355 students (20,709 prospective students times 50 percent times .08 hours) for whom public institutions will provide subsequent warnings.

Collectively, subsequent warning notices will increase burden by 282,654 hours under OMB Control Number 1845-0123.

Similarly, it will take the recipients of subsequent warnings time to read the second warning. Burden for students will increase by 281,269 hours for the 3,515,868 students (7,031,736 prospective students times 50 percent times .08 hours) to read the subsequent warnings from for-profit institutions, OMB Control Number 1845-0123.

Burden will increase by 557 hours for the 6,958 students (13,916 prospective students times 50 percent times .08 hours) to read the subsequent warnings from private non-profit institutions.

Burden will increase by 828 hours for the 10,355 students (20,709 prospective students times 50 percent times .08 hours) to read the subsequent warnings from public institutions.

Collectively, burden to students to read the subsequent warnings will increase by 282,654 hours under OMB Control Number 1845-0123.

The total increase in burden for §668.410 will be 2,116,055 hours under OMB Control Number 1845-0123.

§668.411 Reporting Requirements for GE Programs

Requirements: Under §668.411, institutions will report, for each student enrolled in a GE program during an award year who received title IV, HEA program funds for enrolling in that program: (1) Information needed to identify the student and the institution the student attended; (2) the name, CIP code, credential level, and length of the GE program; (3) whether the GE program is a medical or dental program whose students are required to complete an internship or residency; (4) the date the student initially enrolled in the GE program; (5) the student's attendance dates and attendance status in the GE program during the award year; and (6) the student's enrollment status as of the first day of the student's enrollment in the GE program.

Further, if the student completed or withdrew from the GE program during the award year, the institution will report: (1) The date the student completed or withdrew; (2) the total amount the student received from private education loans for enrollment in the GE program that the institution is, or should reasonably be, aware of; (3) the total amount of institutional debt the student owes any party after completing or withdrawing from the GE program; (4) the total amount for tuition and fees assessed the student for the student’s entire enrollment in the program; and (5) the total amount of allowances for books, supplies, and equipment included in the student's title IV, Cost of Attendance for each award year in which the student was enrolled in the program, or a higher amount if assessed by the institution to the student.

By July 31 of the year the regulations take effect, institutions will be required to report this information for the second through seventh award years prior to that date. For medical and dental programs that require an internship or residency, institutions will need to include the eighth award year no later than July 31. For all subsequent award years, institutions will report not later than October 1, following the end of the award year, unless the Secretary establishes a different date in a notice published in the Federal Register. The regulations give the Secretary the flexibility to identify additional reporting items, or to specify a reporting deadline different than October 1, in a notice published in the Federal Register.

Finally, the regulations will require institutions to provide the Secretary with an explanation of why any missing information is not available.

Burden Calculation: There are 2,526 for-profit institutions that offer one or more GE programs. We estimate that, on average, it will take 6 hours for each of those institutions to modify or develop manual or automated systems for reporting under §668.411. Therefore burden will increase for these institutions by 15,156 hours (2,526 institutions times 6 hours).

There are 318 private non-profit institutions that offer one or more GE programs. We estimate that, on average, it will take 6 hours for each of those institutions to modify or develop manual or automated systems for reporting under §668.411. Therefore burden will increase for these institutions by 1,908 hours (318 institutions times 6 hours).

There are 1,117 public institutions that offer one or more GE programs. We estimate that, on average, it will take 6 hours for each of those institutions to modify or develop manual or automated systems for reporting under §668.411. Therefore burden will increase for these institutions by 6,702 hours (1,117 institutions times 6 hours).

Collectively, burden to develop systems for reporting will increase by 23,766 hours (under OMB Control Number 1845-0123.

Requirements: Under §668.411(a)(3), if an institution is required by its accrediting agency or State to calculate a placement rate for either the institution or the program, or both, the institution is required to report to the Department the required placement rate, using the required methodology, and to report the name of the accrediting agency or State.

Burden Calculation: The Department will be developing a database to collect this data. Therefore, under the Paperwork Reduction Act, the Department will construct an information collection (IC) closer to the time of system development which the public will have an opportunity to provide comment prior to the IC’s submission to OMB for approval.

Requirements: Section 668.411(b) requires that, by no later than July 31 of the year the regulations take effect, institutions report the information required by §668.411(a) for the second through seventh award years prior to that date. For medical and dental programs that require an internship or residency, institutions will need to include the eighth completed award year prior to July 31.

Burden Calculation: According to our analysis of previously reported GE program enrollment data, there were 2,703,851 students enrolled in GE programs offered by for-profit institutions during the 2009-2010 award year. Based on budget baseline estimates as provided in the general background information, we estimate that enrollment in GE programs at for-profit institutions for 2008-2009 was 2,219,280. Going forward, we estimate that enrollment in GE programs at for-profit institutions for 2010-2011 was 2,951,154, for 2011-2012 enrollment was 2,669,084, for 2012-2013 enrollment was 2,426,249, and for 2013-2014 enrollment will be 2,227,230. This results in a total of 15,196,848 enrollments.

We estimate that, on average, the reporting of GE program information by for-profit institutions will take 0.03 hours (2 minutes) per student as we anticipate that, for most for-profit institutions, reporting will be an automated process. Therefore, GE reporting by for-profit institutions will increase burden by 455,905 hours (15,196,848 students times .03 hours) in OMB Control Number 1845-0123.

According to our analysis of previously reported GE program enrollment data, there were 57,700 students enrolled in GE programs offered by private non-profit institutions during the 2009-2010 award year. Based on budget baseline estimates as provided in the general background information, we estimate that enrollment in GE programs at private non-profit institutions for 2008-2009 was 49,316. Going forward, we estimate that enrollment in GE programs at private non-profit institutions for 2010-2011 was 67,509, for 2011-2012 enrollment was 73,585, for 2012-2013 enrollment was 70,641, and for 2013-2014 enrollment will be 65,697. This results in a total of 384,448 enrollments.

We estimate that, on average, the reporting of GE program information by private non-profit institutions will take 0.03 hours (2 minutes) per student as we anticipate that, for most private non-profit institutions, reporting will be an automated process. Therefore, GE reporting by private non-profit institutions will increase burden by 11,533 hours (384,448 students times .03 hours) in OMB Control Number 1845-0123.

According to our analysis of previously reported GE program enrollment data, there were 276,234 students enrolled in GE programs offered by public institutions during the 2009-2010 award year. Based on budget baseline estimates as provided in the general background information, we estimate that enrollment in GE programs at public institutions for 2008-2009 was 236,097. Going forward, we estimate that enrollment in GE programs at public institutions for 2010-2011 was 323,194, for 2011-2012 enrollment was 352,281, for 2012-2013 enrollment was 338,190, and for 2013-2014 enrollment will be 314,517. This results in a total of 1,840,513 enrollments.

We estimate that, on average, the reporting of GE program information by public institutions will take 0.03 hours (2 minutes) per student as we anticipate that, for most public institutions, reporting will be an automated process. Therefore, GE reporting by public institutions will increase burden by 55,215 hours (1,840,513 students times .03 hours) in OMB Control Number 1845-0123.

Collectively, we estimate that burden upon institutions to meet the initial reporting requirements under §668.411 will increase burden by 522,653 hours in OMB Control Number 1845-0123.

The total increase in burden for §668.411 will be 546,419 hours under OMB Control Number 1845-0123.

§668.412 Disclosure Requirements for GE Programs

Requirements: Section 668.412 requires institutions to disclose items, using the disclosure template provided by the Secretary. Under §668.412, the Department has flexibility to tailor the disclosure in a way that will be most useful to students and minimize burden to institutions.

These disclosure items could include:

(1) The primary occupations (by name and SOC code) that the GE program prepares students to enter, along with links to the corresponding occupational profiles on O*Net or its successor site;

(2) The GE program's completion rates for full-time and less-than-full-time students and GE program’s withdrawal rates;

(3) The length of the program in calendar time;

(4) The number of clock or credit hours or equivalent, as applicable, in the program;

(5) The total number of individuals enrolled in the program during the most recently completed award year;

(6) The loan repayment rate for any one or all of the following groups of students who entered repayment on title IV loan during the two-year cohort period: all students who enrolled in the program, students who completed the program, and students who withdrew from the program;

(7) The total cost of tuition and fees, books, supplies, and equipment that students will incur for completing the program within the length of the program;

(8) The placement rate for the program, if the institution is required by its accrediting agency or State to calculate a placement rate either for the program or the institution or both, using the required methodology of that accrediting agency or State;

(9) Of the individuals enrolled in the program during the most recently completed award year, the percentage who received a title IV loan or a private loan for enrollment in the program;

(10) As calculated by the Secretary, the median loan debt incurred by any or all of the following groups: students who completed the program during the most recently completed award year, students who withdrew from the program during the most recently completed award year, or both those groups of students;

(11) The mean or median earnings of any one or all of the following groups: students who completed the program during the applicable cohort period used to calculate the most recent D/E rates for the program, students who were in withdrawn status at the end of the applicable cohort period used to calculate the most recent D/E rates for the program, or both those groups of students;

(12) The most recent program cohort default rate;

(13) The most recent annual earnings rate as calculated by the Secretary under §668.404;

(14)

(i) Whether the program does or does not satisfy--

(A) The applicable educational prerequisites for professional licensure or certification in each State within the institution’s MSA; and

(B) The applicable educational prerequisites for professional licensure or certification in any other State for which the institution has made a determination regarding such requirements; and

(ii) For any States not described in paragraph (14)(i) of this section, a statement that the institution has not made a determination with respect to the licensure or certification requirements of those States.

(15) Whether the program is programmatically accredited and the name of the accrediting agency.

(16) A link to the U.S. Department of Education’s College Navigator Web site, or its successor site, or other similar Federal resource.

The Secretary will conduct consumer testing to determine how to make the disclosures as meaningful as possible. After we have the results of the consumer testing, each year the Secretary will identify which of these items institutions must include in their disclosures, along with any other information that must be included, and publish those requirements in a notice in the Federal Register.

Institutions must update their GE program disclosure information annually. They must make it prominently available in their promotional materials and make it prominent, readily accessible, clear, conspicuous, and directly available on any Web page containing academic, cost, financial aid, or admissions information about a GE program.

An institution that offers a GE program in more than one program length must publish a separate disclosure template for each length of the program.

Burden Calculation: We estimate that of the 37,589 GE programs that reported enrollments in the past, 12,250 programs will be offered by for-profit institutions. We estimate that, annually, the amount of time it will take to collect the data from institutional records, from information provided by the Secretary, and from the institution's accreditor or State, and the amount of time it will take to ensure that promotional materials either include the disclosure information or provide a Web address or direct link to the information will be, on average, 4 hours per program. Additionally, we estimate that revising the institution's Web pages used to disseminate academic, cost, financial aid, or admissions information to also contain the disclosure information about the program will, on average, increase burden by an additional 1 hour per program. Therefore, burden will increase by 5 hours per program for a total of 61,250 hours of increased burden (12,250 programs times 5 hours per program) under OMB Control Number 1845-0123.

We estimate that of the 37,589 GE programs that reported enrollments in the past, 2,343 programs will be offered by private non-profit institutions. We estimate that, annually, the amount of time it will take to collect the data from institutional records, from information provided by the Secretary, and from the institution's accreditor or State, and the amount of time it will take to ensure that promotional materials either include the disclosure information or provide a Web address or direct link to the information will be, on average, 4 hours per program. Additionally, we estimate that revising the institution's Web pages used to disseminate academic, cost, financial aid, or admissions information about the program to also contain the disclosure information will, on average, increase burden by an additional 1 hour per program. Therefore, burden will increase by 5 hours per program for a total of 11,715 hours of increased burden (2,343 programs times 5 hours per program) under OMB Control Number 1845-0123.

We estimate that of the 37,589 GE programs that reported enrollments in the past, 22,996 programs will be offered by public institutions. We estimate that the amount of time it will take to collect the data from institutional records, from information provided by the Secretary, and from the institution's accreditor or State, and the amount of time it will take to ensure that promotional materials either include the disclosure information or provide a Web address or direct link to the information will be, on average, 4 hours per program. Additionally, we estimate that revising the institution's Web pages used to disseminate academic, cost, financial aid, and admissions information about the program to also contain the disclosure information will, on average, increase burden by an additional 1 hour per program. Therefore, on average, burden will increase by 5 hours per program for a total of 114,980 hours of increased burden (22,996 programs times 5 hours per program) under OMB Control Number 1845-0123.

Collectively, we estimate that burden will increase by 187,945 hours in OMB Control Number 1845-0123.

Under §668.412(e), an institution must provide, as a separate document, a copy of the disclosure information to a prospective student. Before a prospective student signs an enrollment agreement, completes registration at, or makes a financial commitment to the institution, the institution must obtain written acknowledgement from the prospective student that he or she received the copy of the disclosure information.

We estimate that the enrollment in the 12,250 GE programs offered by for-profit institutions for 2013-2014 included 2,227,230 prospective students. As noted earlier, most institutions will have to contact, or be contacted by, a larger number of prospective students to yield institutions' desired net enrollments.

We estimate that the number of prospective students that must contact or be contacted by for-profit institutions will be 6 times the number of expected enrollment. As noted above, we estimate that 13,363,380 (2,227,230 students for 2013-2014 times 6) students will be enrolled in GE programs at for-profit institutions. Therefore, for-profit institutions will be required to provide 13,363,380 disclosures to prospective students. On average, we estimate that it will take institutional staff 0.03 hours (2 minutes) per prospective student to provide a copy of the disclosure information which can be hand-delivered, delivered as part of a group presentation, or by sending the disclosure template via the institution’s primary email address (used to communicate with students and prospective students). We also estimate that, on average, it will take institutional staff 0.10 hours (6 minutes) to obtain written acknowledgement and answer any questions from each prospective student. Therefore, we estimate that the total burden associated with providing the disclosure information and obtaining written acknowledgement by for-profit institutions will be 0.13 hours (8 minutes) per prospective student. Burden will increase by 1,737,239 hours for for-profit institutions (13,363,380 prospective students times 0.13 hours) under OMB Control Number 1845-0123.

We estimate that the burden on each prospective student will be 0.08 hours (5 minutes) to read the disclosure information and provide written acknowledgement of receipt. Burden will increase by 1,069,070 hours for prospective students at for-profit institutions (13,363,380 prospective students times 0.08 hours) under OMB Control Number 1845-0123.

We estimate that the enrollment in the 2,343 GE programs offered by private non-profit institutions for 2013-2014 included 65,697 prospective students. As noted earlier, most institutions will have to contact, or be contacted by, a larger number of prospective students to yield their enrollments.

We estimate that the number of prospective students that must contact or be contacted by private non-profit institutions will be 1.8 times the number of expected enrollment. As noted above we estimate that 65,697 students will be enrolled in GE programs at private non-profit institutions. Therefore, private non-profit institutions will be required to provide 118,255 disclosures (65,697 times 1.8) to prospective students. On average, we estimate that it will take institutional staff 0.03 hours (2 minutes) per prospective student to provide a copy of the disclosure information which can be hand-delivered, delivered as a part of a group presentation, or by sending the disclosure template via the institution’s primary email address (used to communicate with students and prospective students). We also estimate that, on average, it will take institutional staff 0.10 hours (6 minutes) to obtain written acknowledgement and answer any questions from each prospective student. Therefore, we estimate that the total burden associated with providing the disclosure information and obtaining written acknowledgement by private-non-profit institutions will be 0.13 hours (8 minutes) per prospective student. Burden will increase by 15,373 hours for private non-profit institutions (118,255 prospective students times 0.13 hours) under OMB Control Number 1845-0123.

We estimate that the burden on each prospective student will be 0.08 hours (5 minutes) to read the disclosure information and provide written acknowledgement of receipt. Burden will increase by 9,460 hours for prospective students at private non-profit institutions (118,255 prospective students times 0.08 hours) under OMB Control Number 1845-0123.

We estimate that the enrollment in the 22,996 GE programs offered by public institutions for 2013-2014 included 314,517 prospective students. As noted earlier, most institutions will have to contact, or be contacted by, a larger number of prospective students to yield their enrollments.

We estimate that the number of prospective students that must contact or be contacted by public institutions will be 1.5 times the number of expected enrollment. As noted above, we estimate that 314,517 students will be enrolled in GE programs at public institutions. Therefore, public institutions will be required to provide 471,776 disclosures (314,517 times 1.5) to prospective students. On average, we estimate that it will take institutional staff 0.03 hours (2 minutes) per prospective student to provide a copy of the disclosure information which can be hand-delivered, delivered as part of a group presentation, or by sending the disclosure template via the institution’s primary email address (used to communicate to students and prospective students). We also estimate that, on average, it will take institutional staff 0.10 hours (6 minutes) to obtain written acknowledgement and answer any questions from each prospective student. Therefore, we estimate that the total burden associated with providing the disclosure information and obtaining written acknowledgement by public institutions will be 0.13 hours (8 minutes) per prospective student. Burden will increase by 61,331 hours for public institutions (471,776 prospective students times 0.13 hours) under OMB Control Number 1845-0123.

We estimate that the burden on each prospective student will be 0.08 hours (5 minutes) to read the disclosure information and provide written acknowledgement of receipt. Burden will increase by 37,742 hours for prospective students at public institutions (471,776 prospective students times 0.08 hours) under OMB Control Number 1845-0123.

Collectively, burden will increase by 2,930,215 hours under OMB Control Number 1845-0123.

The total increase in burden for §668.412 will be 3,118,160 hours under OMB Control Number 1845-0123.

§668.413 Calculating, Issuing, and Challenging Completion Rates, Withdrawal Rates, Repayment Rates, Median Loan Debt, Median Earnings, and Program Cohort Default Rate

Requirements: As discussed in connection with §668.412, an institution will be required to disclose, among other information, completion and withdrawal rates, repayment rates, and median loan debt and median earnings for a GE program. Using the procedures in §668.413 and based partially on the information that an institution will report under §668.411, the Secretary will calculate and make available to the institution for disclosure: Completion rates, withdrawal rates, repayment rates, median loan debt, and median earnings for a GE program.

An institution will have an opportunity to correct the list of students who withdrew from a GE program and the list of students who completed or withdrew from a GE program prior to the Secretary sending the lists to SSA for earnings information.

For the median earnings calculation under §§668.413(b)(9) and (b)(10), after the Secretary provides a list of the relevant students to the institution, the institution may provide evidence showing that a student should be included on the list or removed from the list as a result of meeting the definitions of an exclusion under §668.413(b)(11). The institution may also correct or update a student's identity information or attendance information on the list.

Burden Calculation: For the 12,250 GE programs at for-profit institutions, we estimate, on average, that it will take institutional staff 2 hours to review each of the two lists to determine whether a student should be included or excluded under §668.413(b)(11) and, if included, whether the student's identity information or attendance information requires correction, and then to obtain the evidence to substantiate any inclusion, exclusion, or correction. Burden will increase by 49,000 hours (12,250 programs times 2 lists times 2 hours) under OMB Control Number 1845-0123.

For the 2,343 GE programs at private non-profit institutions, we estimate, on average, that it will take institutional staff 2 hours to review each of the two lists to determine whether a student should be included or excluded and, if included, whether the student's identity information or attendance information requires correction, and then to obtain the evidence to substantiate any inclusion, exclusion, or correction. Burden will increase by 9,372 hours (2,343 programs times 2 lists times 2 hours) under OMB Control Number 1845-0123.

For the 22,996 GE programs at public institutions, we estimate, on average, that it will take institutional staff 2 hours to review each of the two lists to determine whether a student should be included or excluded and, if included, whether the student's identity information or attendance information requires correction, and then to obtain the evidence to substantiate any inclusion, exclusion, or correction. Burden will increase by 91,984 hours (22,996 programs times 2 lists times 2 hours) under OMB Control Number 1845-0123.

Collectively, burden will increase by 150,356 hours under OMB Control Number 1845-0123.

Under §668.413(d)(1), an institution may challenge the Secretary's calculation of the draft completion rates, withdrawal rates, repayment rates, and median loan debt.

The Secretary will develop the completion rates, withdrawal rates, repayment rates, and median loan debt calculations for each of the estimated 12,250 GE programs at for-profit institutions. For the purpose of challenging the completion, withdrawal, and repayment rates and median

loan debt we estimate that, on average, it will take institutional staff 20 hours per program to review the calculations, compare the data to institutional records, and determine whether challenges need to be made to the calculations. Therefore, burden will increase by 245,000 hours (12,250 programs times 20 hours) under OMB Control Number 1845-0123.

The Secretary will develop the completion rates, withdrawal rates, repayment rates, and median loan debt calculations for each of the estimated 2,343 GE programs at private non-profit institutions. For the purpose of challenging the completion, withdrawal, and repayment rates and median loan debt we estimate that, on average, it will take institutional staff 20 hours per program to review the calculations, compare the data to institutional records, and determine whether challenges need to be made to the calculations. Therefore, burden will increase by 46,860 hours (2,343 programs times 20 hours) under OMB Control Number 1845-0123.

The Secretary will develop the completion rates, withdrawal rates, repayment rates, and median loan debt calculations for each of the estimated 22,996 GE programs at public institutions. For the purpose of challenging the completion, withdrawal, and repayment rates and median loan debt we estimate that, on average, it will take institutional staff 20 hours per program to review the calculations, compare the data to institutional records, and determine whether challenges need to be made to the calculations. Therefore, burden will increase by 459,920 hours (22,996 times 20 hours) under OMB Control Number 1845-0123.

Collectively, burden will increase by 751,780 under OMB Control Number 1845-0123.

The total increase in burden for §668.413 will be 902,136 under OMB Control Number 1845-0123.

§668.414 Certification Requirements for GE Programs

Requirements: Under §668.414(a) each institution participating in the title IV, HEA programs will be required to provide a ”transitional certification” to supplement its current program participation agreement (PPA). The transitional certification will be submitted no later than December 31 of the year in which the regulations take effect. The transitional certification will be signed by the institution's most senior executive officer that each of its currently eligible GE programs included on its Eligibility and Certification Approval Report meets the GE program eligibility certification requirements of this section and will update within 10 days if there are any changes in the approvals for a program, or other changes that make an existing certification inaccurate. Under §668.414(d), the certification will provide that each GE program meets certain requirements (PPA certification requirements), specifically that each GE program is:

1. Approved by a recognized accrediting agency, is included in the institution's accreditation, or is approved by a recognized State agency for the approval of public postsecondary vocational education in lieu of accreditation;

2. Programmatically accredited, if required by a Federal governmental entity or required by a governmental entity in the State in which the institution is located or in which the institution is otherwise required to obtain State approval under 34 CFR 600.9; and

3. Satisfies licensure or certification requirements in the State where the institution is located or in which the institution is otherwise required to obtain State approval, each eligible program it offers satisfies the applicable educational prerequisites for professional licensure or certification requirements in that State so that the student who completes the program and seeks employment in that State qualifies to take any licensure or certification exam that is needed for the student to practice or find employment in an occupation that the program prepares students to enter.

A program is substantially similar to another program if the two programs share the same four-digit CIP code. The Secretary presumes a program is not substantially similar to another program if the two programs have different four-digit CIP codes, but the institution must provide an explanation of how the new program is not substantially similar to an ineligible or voluntarily discontinued program with its certification under §668.414.

Burden Calculation: We estimate that it will take the 2,526 for-profit institutions that offer GE programs 0.5 hours to draft a certification statement and obtain the signature of the institution's senior executive for submission to the Department and, when applicable, provide an explanation of how a new program is not substantially similar to an ineligible or voluntarily discontinued program. This will increase burden by 1,263 hours (2,526 institutions times 0.5 hours) under OMB Control Number 1845-0123.

We estimate that it will take the 318 private non-profit institutions that offer GE programs 0.5 hours to draft a certification statement and obtain the signature of the institution's senior executive for submission to the Department and, when applicable, provide an explanation of how a new program is not substantially similar to an ineligible or voluntarily discontinued program. This will increase burden by 159 hours (318 institutions times 0.5 hours) under OMB Control Number 1845-0123.

We estimate that it will take the 1,117 public institutions that offer GE programs 0.5 hours to draft a certification statement and obtain the signature of the institution's senior executive for submission to the Department and, when applicable, provide an explanation of how a new program is not substantially similar to an ineligible or voluntarily discontinued program. This will increase burden by 559 hours (1,117 institutions times 0.5 hours) under OMB Control Number 1845-0123.

The total increase in burden for §668.414 will be 1,981 hours under OMB Control Number 1845-0123.

Subpart R--Program Cohort Default Rates

Requirements: Under subpart R, the Secretary will calculate a GE program's cohort default rate using a structure that will generally mirror the structure of the iCDR regulations in subpart N of part 668 of the regulations. Thus, depending on the pCDR of a program, an institution will have the opportunity to submit a challenge, request an adjustment, or appeal the pCDR. Detailed information about each of these opportunities and our burden assessments follow. For all requests for challenges, adjustments, or appeals, institutions will receive a loan record detail report (LRDR) provided by the Department.

Burden Calculation: The pCDR regulations in subpart R, although specific to programs, generally mirror the structure of the institutional cohort default rate (iCDR) regulations in subpart N of part 668 of the regulations. However, because pCDR is used as a potential disclosure, and not as a standard for assessing eligibility (as with iCDR), the available appeals are limited to factual corrections and challenges and the burden assessments that follow recognize that, although institutions will have the option of submitting challenges, requests for adjustments, and certain appeals for all of their GE programs in every year for which we calculate a pCDR, institutions will in all likelihood exercise those rights only in those instances in which we calculate a pCDR rate of 20 percent or higher.

Of the 6,815 GE programs that we estimate will be evaluated for pCDR, we estimate that 943 programs will have rates of 30 percent or more and therefore have the highest likelihood of having pCDR challenges, adjustments, or appeals. In addition, we estimate that half of the 1,840 GE programs with a pCDR rate of 20 percent to 29.9 percent will also make challenges, request adjustments, or submit appeals, adding another 920 programs to the 943 that had rates of 30 percent or more for a total of 1,863 programs. We estimate that 92 percent of the 1,863 will be GE programs at for-profit institutions, 3 percent will be GE programs at private non-profit institutions, and 5 percent will be GE programs at public institutions.

We used an analysis of the FY 2011 iCDR data to estimate the percentage of the possible 1,863 programs where a challenge, adjustment request, or appeal may be submitted. Those percentages varied by the type of challenge, adjustment, or appeal, as indicated in each of the regulatory sections that follow and are used to project the distribution of pCDR challenges, adjustments, and appeals.

§668.504 Draft Cohort Program Default Rates and Your Ability To Challenge Before Official Program Cohort Default Rates Are Issued

Requirements: Incorrect Data Challenges: Under §668.504(b), the institution may challenge the accuracy of the data included on the LRDR by sending an incorrect data challenge to the relevant data manager(s) within 45 days of receipt of the LRDR from the Department. The challenge will include a description of the information in the LRDR that the institution believes is incorrect along with supporting documentation.

Burden Calculation: Based upon FY 2011 submissions, there were 353 iCDR challenges for incorrect data of a total of 510 challenges, requests for adjustments, and appeals, a 69 percent submission rate. Therefore 69 percent of the projected 1,863 challenges, adjustments, and appeals, or 1,285, are projected to be challenges for incorrect data.

We estimate that out of the likely 1,285 submissions, 1,182 (92 percent) will be from for-profit institutions. We estimate that the average institutional staff time needed to review a GE program's LRDR for each of these 1,182 programs and to gather and prepare incorrect data challenges will be 4 hours (1.5 hours for list review and 2.5 hours for documentation submission). This will increase burden by 4,728 hours (1,182 programs times 4 hours) under OMB Control Number 1845-0121.

We estimate that out of the likely 1,285 submissions, 39 (3 percent) will be from private non-profit institutions. We estimate that the average institutional staff time needed to review a GE program's LRDR for each of these 39 programs and to gather and prepare the challenges will be 4 hours (1.5 hours for list review and 2.5 hours for documentation submission). This will increase burden by 156 hours (39 programs times 4 hours) under OMB Control Number 1845-0121.

We estimate that, out of the likely 1,285 submissions, 64 (5 percent) will be from public institutions. We estimate that the average institutional staff time needed to review a GE program's LRDR for each of these 64 programs and to gather and prepare the challenges will be 4 hours (1.5 hours for list review and 2.5 hours for documentation submission). This will increase burden by 256 hours (64 programs times 4 hours) under OMB Control Number 1845-0121.

The total increase in burden for §668.504 will be 5,140 hours under OMB Control Number 1845-0121.

§668.509 Uncorrected Data Adjustments

Requirements: An institution may request an uncorrected data adjustment for the most recent cohort of borrowers used to calculate a GE program's most recent official pCDR, if in response to the institution's incorrect data challenge, a data manager agreed correctly to change data but the changes were not reflected in the official pCDR.

Burden Calculation: Based upon FY 2011 submissions, there were 116 uncorrected data adjustments of the total 510 challenges, requests for adjustments, and appeals. Therefore, 23 percent of the projected 943 challenges, adjustments, and appeals or 217 are projected to be uncorrected data adjustments.

We estimate that the average institutional staff time needed is 1 hour for list review and 0.5 hours for documentation submission, for a total of 1.5 hours.

We estimate that 200 (92 percent) of the 217 projected uncorrected data adjustments will be from for-profit institutions. Therefore, burden will increase at for-profit institutions by 300 hours (200 adjustments times 1.5 hours) under OMB Control Number 1845-0121.

We estimate that 6 (3 percent) of the 217 projected uncorrected data adjustments will be from private non-profit institutions. Therefore, burden will increase at private non-profit institutions by 9 hours (6 adjustments times 1.5 hours) under OMB Control Number 1845-0121.

We estimate that 11 (5 percent) of the 217 projected uncorrected data adjustments will be from public institutions. Therefore, burden will increase at public institutions by 17 hours (11 adjustments times 1.5 hours) under OMB Control Number 1845-0121.

The total increase in burden for §668.509 will be 326 hours under OMB Control Number 1845-0121.

§668.510 New Data Adjustments

Requirements: An institution could request a new data adjustment for the most recent cohort of borrowers used to calculate the most recent official pCDR for a GE program, if a comparison of the LRDR for the draft rates and the LRDR for the official rates shows that data have been newly included, excluded, or otherwise changed and the errors are confirmed by the data manager.

Burden Calculation: Based upon FY 2011 submissions, there were 12 new data adjustments of the total 510 challenges, requests for adjustments, and appeals. Therefore, 2 percent of the projected 943 challenges, adjustments, and appeals or 19 are projected to be new data adjustments. We estimate that the average institutional staff time needed is 3 hours for list review and 1 hour for documentation submission, for a total of 4 hours.

We estimate that 17 (92 percent) of the 19 projected new data adjustments will be from for-profit institutions. Therefore, burden will increase at for-profit institutions by 68 hours (17 adjustments times 4 hours) under OMB Control Number 1845-0121.

We estimate that 1 (3 percent) of the 19 projected new data adjustments will be from private non-profit institutions. Therefore, burden will increase at private non-profit institutions by 4 hours (1 adjustment times 4 hours) under OMB Control Number 1845-0121.

We estimate that 1 (5 percent) of the 19 projected new data adjustments will be from public institutions. Therefore, burden will increase at public institutions by 4 hours (1 adjustment times 4 hours) under OMB Control Number 1845-0121.

The total increase in burden for §668.510 will be 76 hours under OMB Control Number 1845-0121.

§668.511 Erroneous Data Appeals

Requirements: An institution could appeal the calculation of a pCDR if it disputes the accuracy of data that was previously challenged under §668.504(b) (challenge for incorrect data) or if a comparison of the LRDR that we provided for the draft rate and the official rate shows that data have been newly included, excluded, or otherwise changed, and the accuracy of the data has been disputed. The institution must send a request for verification of data to the applicable data manager(s) within 15 days of receipt of the notice of the official pCDR, and it must include a description of the incorrect information and all supporting documentation to demonstrate the error.

Burden Calculation: Based upon the fact that in FY 2011 there were no iCDR erroneous data appeals, we have no basis to establish erroneous data appeals burden for pCDRs.

§668.512 Loan Servicing Appeals

Requirements: An institution could appeal the calculation of a pCDR on the basis of improper loan servicing or collection.

Burden Calculation: Based upon FY 2011 submissions, there were 19 loan servicing appeals of the total 510 challenges, requests for adjustments, and appeals. Therefore, 4 percent or 38 of the projected 943 challenges, adjustments, and appeals are projected to be loan servicing appeals. We estimate that, on average, to gather, analyze, and submit the necessary documentation, each appeal will take 3 hours.

We estimate that 35 (92 percent) of the 38 projected loan servicing appeals will be from for-profit institutions. Therefore, burden will increase at for-profit institutions by 105 hours (35 servicing appeals times 3 hours) under OMB Control Number 1845-0121.

We estimate that 1 (3 percent) of the 38 projected loan servicing appeals will be from private non-profit institutions. Therefore, burden will increase at private non-profit institutions by 3 hours (1 servicing appeal times 3 hours) under OMB Control Number 1845-0121.

We estimate that 2 (5 percent) of the 38 projected loan servicing appeals will be from public institutions. Therefore, burden will increase at public institutions by 6 hours (2 servicing appeals times 3 hours) under OMB Control Number 1845-0121.

The total increase in burden for §668.512 will be 114 hours under OMB Control Number 1845-0121.

Consistent with the discussion above, the following chart describes the sections of the regulations involving information collections, the information being collected, the collections that the Department will submit to OMB for approval and public comment under the PRA, and the estimated costs associated with the information collections. The monetized net costs of the increased burden on institutions and borrowers, using wage data developed using BLS data, available at www.bls.gov/ncs/ect/sp/ecsuphst.pdf, is $209,247,305, as shown in the chart below. This cost was based on an hourly rate of $36.55 for institutions and $16.30 for students.

Collection of Information

Regulatory section	Information collection	OMB Control No. and estimated burden	Estimated costs
668.405-Issuing and challenging D/E rates	The regulations provide institutions an opportunity to correct information about students who have completed their programs and who are on the list provided by the Department to the institution.	OMB 1845-0123 This will be a new collection. We estimate that the burden will increase by 211,360 hours.	$7,725,208
668.406-D/E rates alternate earnings appeals	The regulations will allow institutions to make an alternate earnings appeal to the D/E rates, when the final D/E rates are failing or in the zone under the D/E rates measure.	OMB 184-0122 This will be a new collection. We estimate that the burden will increase by 23,860 hours.	$872,083
668.410-Consequences of the D/E rates measure.	The regulations provide that for any year the Secretary notifies the institution that a GE program could become ineligible based on its D/E rates for the next award year the institution must provide student warnings.	OMB 1845-0123 This will be a new collection. We estimate that the burden for institutions will increase by 1,065,198 hours. We estimate that burden will increase for individuals by 1,050,857 hours.	$56,061,956
668.411-Reporting requirements for GE programs.	The regulations will require institutions to report to the Department information about students in GE programs.	OMB 1845-0123 This will be a new collection. We estimate that the burden will increase by 546,419 hours.	$19,971,614
668.412-Disclosure requirements for GE programs.	The regulations will require certain information about GE programs to be disclosed by institutions to enrolled and prospective students.	OMB 1845-0123 This will be a new collection. We estimate that the burden for institutions will increase by 2,001,888 hours. We estimate that the burden for individuals will increase by 1,116,272 hours.	$91,364,240
668.413-Calculating, issuing, and challenging completion rates, withdrawal rates, repayment rates, median loan debt, and median earnings, and program cohort default rates.	The regulations allow institutions to challenge the rates and median earnings calculated by the Department.	OMB 1845-0123 This will be a new collection. We estimate that the burden will increase by 902,136 hours.	$32,973,071
668.414-Certification requirements for GE programs.	The regulations will add a requirement that an institution certify that GE programs it offers are approved or accredited by an accrediting agency or the State. The regulations also add a requirement that the institution must provide an explanation of how a new GE program is not substantially similar to an ineligible or voluntarily discontinued program.	OMB 1845-0123 This will be a new collection. We estimate that the burden will increase by 1,981 hours.	$72,406
668.504-Draft program cohort default rates and challenges.	The regulations will allow an institution to challenge the draft program cohort default rates.	OMB 1845-0121 This will be a new collection. We estimate that the burden will increase 5,140 hours.	$187,867
668.509-Uncorrected data adjustments.	The regulations will allow institutions to request a data adjustment when agreed-upon data changes were not reflected in the official program cohort default rate.	OMB 1845-0121 This will be a new collection. We estimate that the burden will increase by 326 hours.	$11,915
668.510-New data adjustments.	The regulations will allow institutions to request a new data adjustment if a comparison of the draft and final LRDR show that data have been included, excluded, or otherwise changed and the errors are confirmed by the data manager.	OMB 1845-0121 This will be a new collection. We estimate that the burden will increase by 76 hours.	$2,778
668.511-Erroneous data appeals.	The regulations will allow an institution to appeal the program cohort default rate calculation when the accuracy was previously challenged on the basis of incorrect data.	OMB 1845-0121 This will be a new collection. We estimate that the burden will increase by 0 hours.	$0
668.512-Loan Servicing Appeal.	The regulations will allow an institution to appeal on the basis of improper loan servicing or collection where the institution can prove that the servicer failed to perform required servicing or collections activities.	OMB 1845-0121 This will be a new collection. We estimate that the burden will increase by 114 hours.	$4,167

The total burden hours and change in burden hours associated with each OMB Control number affected by the regulations follows:

-----------------------------------------------------------

Total current Change

Control No. burden hours in burden hours

-----------------------------------------------------------------------------------------------------------------------------

1845-0123………………. 0 + 6,896,111

1845-0122………………. 0 23,860

1845-0121………………. 0 5,656

------------------------------------------------------

Total………………………. 0 6,925,627

Assessment of Educational Impact

In the NPRM we requested comments on whether the proposed regulations would require transmission of information that any other agency or authority of the United States gathers or makes available.

Based on the response to the NPRM and on our review, we have determined that these final regulations do not require transmission of information that any other agency or authority of the United States gathers or makes available.

Accessible Format: Individuals with disabilities can obtain this document in an accessible format (e.g., braille, large print, audiotape, or compact disc) on request to the program contact person listed under FOR FURTHER INFORMATION CONTACT.

Electronic Access to This Document: The official version of this document is the document published in the Federal Register. Free Internet access to the official edition of the Federal Register and the Code of Federal Regulations is available via the Federal Digital System at: www.gpo.gov/fdsys. At this site you can view this document, as well as all other documents of this Department published in the Federal Register, in text or Adobe Portable Document Format (PDF). To use PDF you must have Adobe Acrobat Reader, which is available free at the site.

You may also access documents of the Department published in the Federal Register by using the article search feature at: www.federalregister.gov. Specifically, through the advanced search feature at this site, you can limit your search to documents published by the Department.

(Catalog of Federal Domestic Assistance Numbers: 84.007 FSEOG; 84.032 Federal Family Education Loan Program; 84.033

Federal Work-Study Program; 84.038 Federal Perkins Loan Program; 84.063 Federal Pell Grant Program; 84.069A LEAP; 84.268 William D. Ford Federal Direct Loan Program; 84.376 ACG/Smart; 84.379 TEACH Grant Program; 84.069B Grants for Access and Persistence Program)

List of Subjects

34 CFR Part 600

Colleges and universities, Foreign relations, Grant programs-education, Loan programs-education, Reporting and recordkeeping requirements, Student aid, Vocational education

34 CFR Part 668

Administrative practice and procedure, Aliens, Colleges and universities, Consumer Protection, Grant programs-education, Loan programs—education, Reporting and recordkeeping requirements, Selective Service System, Student aid, Vocational education

Dated:

_____________________

Arne Duncan, Secretary of Education.

For the reasons discussed in the preamble, the Secretary of Education proposes to amend parts 600 and 668 of title 34 of the Code of Federal Regulations as follows:

PART 600--INSTITUTIONAL ELIGIBILITY UNDER THE HIGHER EDUCATION ACT OF 1965, AS AMENDED

1. The authority citation for part 600 continues to read as follows:

AUTHORITY: 20 U.S.C. 1001, 1002, 1003, 1088, 1091, 1094, 1099b, and 1099c, unless otherwise noted.

2. Section 600.2 is amended by:

A. Revising the definition of “Recognized occupation.”

B. Revising the authority citation at the end of the section.

The revisions read as follows:

§600.2  Definitions.

* * * * *

Recognized occupation: An occupation that is--

(1) Identified by a Standard Occupational Classification (SOC) code established by the Office of Management and Budget (OMB) or an Occupational Information Network O*Net-SOC code established by the Department of Labor, which is available at www.onetonline.org or its successor site; or

(2) Determined by the Secretary in consultation with the Secretary of Labor to be a recognized occupation.

* * * * *

(Authority: 20 U.S.C. 1001, 1002, 1071, et seq., 1078-2, 1088, 1091, 1094, 1099b, 1099c, 1141; 26 U.S.C. 501(c))

3. Section 600.10 is amended by:

A. Revising paragraphs (c)(1), (c)(2), and (c)(3)(i).

B. Revising the authority citation at the end of the section.

The revisions read as follows:

§600.10  Date, extent, duration, and consequence of eligibility.

* * * * *

(c) Educational programs. (1) An eligible institution that seeks to establish the eligibility of an educational program must--

(i) For a gainful employment program under 34 CFR part 668, subpart Q of this chapter, update its application under §600.21, and meet any time restrictions that prohibit the institution from establishing or reestablishing the eligibility of the program as may be required under 34 CFR 668.414;

(ii) Pursuant to a requirement regarding additional programs included in the institution’s program participation agreement under 34 CFR 668.14, obtain the Secretary’s approval; and

(iii) For a direct assessment program under 34 CFR 668.10, and for a comprehensive transition and postsecondary program under 34 CFR 668.232, obtain the Secretary’s approval.

(2) Except as provided under §600.20(c), an eligible institution does not have to obtain the Secretary’s approval to establish the eligibility of any program that is not described in paragraph (c)(1)(i), (ii), or (iii) of this section.

(3) * * *

(i) Fails to comply with the requirements in paragraph (c)(1) of this section; or

* * * * *

(Authority: 20 U.S.C. 1001, 1002, 1088, 1094, and 1141)

4. Section 600.20 is amended by:

A. Revising the introductory text of paragraph (c)(1).

B. Revising the authority citation at the end of the section.

The revisions read as follows:

§600.20  Notice and application procedures for establishing, reestablishing, maintaining, or expanding institutional eligibility and certification.

* * * * *

(c) * * *

(1) Add an educational program or a location at which the institution offers or will offer 50 percent or more of an educational program if one of the following conditions applies, otherwise it must report to the Secretary under §600.21:

* * * * *

(Authority: 20 U.S.C. 1001, 1002, 1088, 1094, and 1099c)

5. Section 600.21 is amended by:

A. Adding paragraph (a)(11).

B. Revising the authority citation at the end of the section.

The revisions read as follows:

§600.21  Updating application information.

(a) * * *

(11) For any gainful employment program under 34 CFR part 668, subpart Q--

(i) Establishing the eligibility or reestablishing the eligibility of the program;

(ii) Discontinuing the program’s eligibility under 34 CFR 668.410;

(iii) Ceasing to provide the program for at least 12 consecutive months;

(iv) Losing program eligibility under §600.40;

(v) Changing the program’s name, CIP code, as defined in 34 CFR 668.402, or credential level;

or

(vi) Updating the certification pursuant to §668.414(b).

* * * * *

(Authority: 20 U.S.C. 1094, 1099b)

PART 668--STUDENT ASSISTANCE GENERAL PROVISIONS

6. The authority citation for part 668 continues to read as follows:

AUTHORITY: 20 U.S.C. 1001, 1002, 1003, 1088, 1091, 1094, 1099b, and 1099c, unless otherwise noted.

7. Section 668.6 is amended by:

A. Removing and reserving paragraph (a).

B. Adding a new paragraph (d).

C. Revising the authority citation.

The addition and revision read as follows:

§668.6 Reporting and disclosure requirements for programs that prepare students for gainful employment in a recognized occupation.

* * * * *

(d) Sunset provisions. Institutions must comply with the requirements of this section through December 31, 2016.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.7 [Removed and Reserved]

8. Remove and reserve section 668.7.

§668.8 [Amended]

9. Section 668.8 is amended by:

A. In paragraph (d)(2)(iii), removing the reference to “§668.6” and adding, in its place, a reference to “subpart Q of this part”.

B. In paragraph (d)(3)(iii), removing the reference to “§668.6” and adding, in its place, a reference to “subpart Q of this part”.

10. Section 668.14 is amended by revising paragraph (a)(26) to read as follows:

§668.14  Program participation agreement.

(a) * * *

(26) If an educational program offered by the institution is required to prepare a student for gainful employment in a recognized occupation, the institution must--

(i) Demonstrate a reasonable relationship between the length of the program and entry level requirements for the recognized occupation for which the program prepares the student. The Secretary considers the relationship to be reasonable if the number of clock hours provided in the program does not exceed by more than 50 percent the minimum number of clock hours required for training in the recognized occupation for which the program prepares the student, as established by the State in which the institution is located, if the State has established such a requirement, or as established by any Federal agency;

(ii) Establish the need for the training for the student to obtain employment in the recognized occupation for which the program prepares the student; and

(iii) Provide for that program the certification required in §668.414.

* * * * *

11. Add subpart Q to read as follows:

Subpart Q—Gainful Employment (GE) Programs

Sec.

668.401 Scope and purpose.

668.402 Definitions.

668.403 Gainful employment framework.

668.404 Calculating D/E rates.

668.405 Issuing and challenging D/E rates.

668.406 D/E rates alternate earnings appeals.

668.407 [Reserved].

668.408 [Reserved].

668.409 Final determination of the D/E rates measure.

668.410 Consequences of the D/E rates measure.

668.411 Reporting requirements for GE programs.

668.412 Disclosure requirements for GE programs.

668.413 Calculating, issuing, and challenging completion rates, withdrawal rates, repayment rates, median loan debt, median earnings, and program cohort default rate.

668.414 Certification requirements for GE programs.

668.415 Severability.

Subpart Q—Gainful Employment (GE) Programs

§668.401 Scope and purpose.

This subpart applies to an educational program offered by an eligible institution that prepares students for gainful employment in a recognized occupation, and establishes the rules and procedures under which--

(a) The Secretary determines that the program is eligible for title IV, HEA program funds;

(b) An institution reports information about the program to the Secretary; and

(c) An institution discloses information about the program to students and prospective students.

(Authority: 20 U.S.C. 1001, 1002, 1088, 1231a)

§668.402 Definitions.

The following definitions apply to this subpart.

Annual earnings rate. The percentage of a GE program’s annual loan payment compared to the annual earnings of the students who completed the program, as calculated under §668.404.

Classification of instructional program (CIP) code. A taxonomy of instructional program classifications and descriptions developed by the U.S. Department of Education’s National Center for Education Statistics (NCES). The CIP code for a program is six digits.

Cohort period. The two-year cohort period or the four-year cohort period, as applicable, during which those students who complete a program are identified in order to assess their loan debt and earnings. The Secretary uses the two-year cohort period when the number of students completing the program is 30 or more. The Secretary uses the four-year cohort period when the number of students completing the program in the two-year cohort period is less than 30 and when the number of students completing the program in the four-year cohort period is 30 or more.

Credential level. The level of the academic credential awarded by an institution to students who complete the program. For the purposes of this subpart, the undergraduate credential levels are: undergraduate certificate or diploma, associate degree, bachelor’s degree, and post-baccalaureate certificate; and the graduate credential levels are graduate certificate (including a postgraduate certificate), master’s degree, doctoral degree, and first-professional degree (e.g., MD, DDS, JD).

Debt-to-earnings rates (D/E rates). The discretionary income rate and annual earnings rate as calculated under §668.404.

Discretionary income rate. The percentage of a GE program’s annual loan payment compared to the discretionary income of the students who completed the program, as calculated under §668.404.

Four-year cohort period. The cohort period covering four consecutive award years that are--

(1) The third, fourth, fifth, and sixth award years prior to the award year for which the D/E rates are calculated pursuant to §668.404. For example, if D/E rates are calculated for award year 2014-2015, the four-year cohort period is award years 2008-2009, 2009-2010, 2010-2011, and 2011-2012; or

(2) For a program whose students are required to complete a medical or dental internship or residency, the sixth, seventh, eighth, and ninth award years prior to the award year for which the D/E rates are calculated. For example, if D/E rates are calculated for award year 2014-2015, the four-year cohort period is award years 2005-2006, 2006-2007, 2007-2008, and 2008-2009. For this purpose, a required medical or dental internship or residency is a supervised training program that--

(i) Requires the student to hold a degree as a doctor of medicine or osteopathy, or a doctor of dental science;

(ii) Leads to a degree or certificate awarded by an institution of higher education, a hospital, or a health care facility that offers post-graduate training; and

(iii) Must be completed before the student may be licensed by a State and board certified for professional practice or service.

Gainful employment program (GE program). An educational program offered by an institution under §668.8(c)(3) or (d) and identified by a combination of the institution’s six-digit Office of Postsecondary Education ID (OPEID) number, the program’s six-digit CIP code as assigned by the institution or determined by the Secretary, and the program’s credential level.

Length of the program. The amount of time in weeks, months, or years that is specified in the institution’s catalog, marketing materials, or other official publications for a student to complete the requirements needed to obtain the degree or credential offered by the program.

Metropolitan Statistical Area (MSA). The Metropolitan Statistical Area as published by the U.S. Office of Management and Budget and available at www.census.gov/population/metro/ or its successor site.

Poverty Guideline. The Poverty Guideline for a single person in the continental United States as published by the U.S. Department of Health and Human Services and available at http://aspe.hhs.gov/poverty or its successor site.

Prospective student. An individual who has contacted an eligible institution for the purpose of requesting information about enrolling in a GE program or who has been contacted directly by the institution or by a third party on behalf of the institution about enrolling in a GE program.

Student. An individual who received title IV, HEA program funds for enrolling in the GE program.

Title IV loan. A loan authorized under the Federal Perkins Loan Program (Perkins Loan), the Federal Family Education Loan Program (FFEL Loan), or the William D. Ford Direct Loan Program (Direct Loan).

Two-year cohort period. The cohort period covering two consecutive award years that are--

(1) The third and fourth award years prior to the award year for which the D/E rates are calculated pursuant to §668.404. For example, if D/E rates are calculated for award year 2014-2015, the two-year cohort period is award years 2010-2011 and 2011-2012; or

(2) For a program whose students are required to complete a medical or dental internship or residency, the sixth and seventh award years prior to the award year for which the D/E rates are calculated. For example, if D/E rates are calculated for award year 2014-2015, the two-year cohort period is award years 2007-2008 and 2008-2009. For this purpose, a required medical or dental internship or residency is a supervised training program that--

(i) Requires the student to hold a degree as a doctor of medicine or osteopathy, or as a doctor of dental science;

(ii) Leads to a degree or certificate awarded by an institution of higher education, a hospital, or a health care facility that offers post-graduate training; and

(iii) Must be completed before the student may be licensed by a State and board certified for professional practice or service.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.403 Gainful employment program framework.

(a) General. A program provides training that prepares students for gainful employment in a recognized occupation if the program--

(1) Satisfies the applicable certification requirements in §668.414; and

(2) Is not an ineligible program under the D/E rates measure.

(b) Debt-to-earnings rates (D/E rates). For each award year and for each eligible GE program offered by an institution, the Secretary calculates two D/E rates, the discretionary income rate and the annual earnings rate, using the procedures in §§668.404 through 668.406.

(c) Outcomes of the D/E rates measure. (1) A GE program is “passing” the D/E rates measure if--

(i) Its discretionary income rate is less than or equal to 20 percent; or

(ii) Its annual earnings rate is less than or equal to eight percent.

(2) A GE program is “failing” the D/E rates measure if--

(i) Its discretionary income rate is greater than 30 percent or the income for the denominator of the rate (discretionary earnings) is negative or zero; and

(ii) Its annual earnings rate is greater than 12 percent or the denominator of the rate (annual earnings) is zero.

(3) A GE program is “in the zone” for the purpose of the D/E rates measure if it is not a passing GE program and its--

(i) Discretionary income rate is greater than 20 percent but less than or equal to 30 percent; or

(ii) Annual earnings rate is greater than eight percent but less than or equal to 12 percent.

(4) For the purpose of the D/E rates measure, subject to paragraph (c)(5) of this section, a GE program becomes ineligible if the program either--

(i) Is failing the D/E rates measure in two out of any three consecutive award years for which the program’s D/E rates are calculated; or

(ii) Has a combination of zone and failing D/E rates for four consecutive award years for which the program’s D/E rates are calculated.

(5) If the Secretary does not calculate or issue D/E rates for a program for an award year, the program receives no result under the D/E rates measure for that award year and remains in the same status under the D/E rates measure as the previous award year; provided that if the Secretary does not calculate D/E rates for the program for four or more consecutive award years, the Secretary disregards the program’s D/E rates for any award year prior to the four-year period in determining the program’s eligibility.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.404 Calculating D/E rates.

(a) General. Except as provided in paragraph (f) of this section, for each award year, the Secretary calculates D/E rates for a GE program as follows:

(1) Discretionary income rate = annual loan payment / (the higher of the mean or median annual earnings – (1.5 x Poverty Guideline)). For the purposes of this paragraph, the Secretary applies the Poverty Guideline for the calendar year immediately following the calendar year for which annual earnings are obtained under paragraph (c) of this section.

(2) Annual earnings rate = annual loan payment / the higher of the mean or median annual earnings.

(b) Annual loan payment. The Secretary calculates the annual loan payment for a GE program by--

(1)(i) Determining the median loan debt of the students who completed the program during the cohort period, based on the lesser of the loan debt incurred by each student as determined under paragraph (d)(1) of this section and the total amount for tuition and fees and books, equipment, and supplies for each student as determined under paragraph (d)(2) of this section;

(ii) Removing, if applicable, the appropriate number of highest loan debts as described in §668.405(e)(2); and

(iii) Calculating the median of the remaining amounts.

(2) Amortizing the median loan debt--

(i)(A) Over a 10-year repayment period for a program that leads to an undergraduate certificate, a post-baccalaureate certificate, an associate degree, or a graduate certificate;

(B) Over a 15-year repayment period for a program that leads to a bachelor's degree or a master's degree; or

(C) Over a 20-year repayment period for a program that leads to a doctoral or first-professional degree; and

(ii) Using an annual interest rate that is the average of the annual statutory interest rates on Federal Direct Unsubsidized Loans that were in effect during--

(A) The three-year period prior to the end of the cohort period, for undergraduate certificate programs, post-baccalaureate certificate programs, and associate degree programs. For these programs, the Secretary uses the Federal Direct Unsubsidized Loan interest rate applicable to undergraduate students;

(B) The three-year period prior to the end of the cohort period, for graduate certificate programs and master’s degree programs. For these programs, the Secretary uses the Federal Direct Unsubsidized Loan interest rate applicable to graduate students;

(C) The six-year period prior to the end of the cohort period, for bachelor’s degree programs. For these programs, the Secretary uses the Federal Direct Unsubsidized Loan interest rate applicable to undergraduate students; and

(D) The six-year period prior to the end of the cohort period, for doctoral programs and first professional degree programs. For these programs, the Secretary uses the Federal Direct Unsubsidized Loan interest rate applicable to graduate students.

Note to paragraph (b)(2)(ii): For example, for an undergraduate certificate program, if the two-year cohort period is award years 2010-2011 and 2011-2012, the interest rate would be the average of the interest rates for the years from 2009-2010 through 2011-2012.

(c) Annual earnings. (1) The Secretary obtains from the Social Security Administration (SSA), under §668.405, the most currently available mean and median annual earnings of the students who completed the GE program during the cohort period and who are not excluded under paragraph (e) of this section; and

(2) The Secretary uses the higher of the mean or median annual earnings to calculate the D/E rates.

(d) Loan debt and assessed charges. (1) In determining the loan debt for a student, the Secretary includes--

(i) The amount of title IV loans that the student borrowed (total amount disbursed less any cancellations or adjustments) for enrollment in the GE program (Federal PLUS Loans made to parents of dependent students, Direct PLUS Loans made to parents of dependent students, and Direct Unsubsidized Loans that were converted from TEACH Grants are not included);

(ii) Any private education loans as defined in 34 CFR 601.2, including private education loans made by the institution, that the student borrowed for enrollment in the program and that were required to be reported by the institution under §668.411; and

(iii) The amount outstanding, as of the date the student completes the program, on any other credit (including any unpaid charges) extended by or on behalf of the institution for enrollment in any GE program attended at the institution that the student is obligated to repay after completing the GE program, including extensions of credit described in clauses (1) and (2) of the definition of, and excluded from, the term “private education loan” in 34 CFR 601.2;

(2) The Secretary attributes all of the loan debt incurred by the student, and attributes the amount reported for the student under §668.411(a)(2)(iv) and (v), for enrollment in any--

(i) Undergraduate GE program at the institution to the highest credentialed undergraduate GE program subsequently completed by the student at the institution as of the end of the most recently completed award year prior to the calculation of the draft D/E rates under this section; and

(ii) Graduate GE program at the institution to the highest credentialed graduate GE program completed by the student at the institution as of the end of the most recently completed award year prior to the calculation of the draft D/E rates under this section; and

(3) The Secretary excludes any loan debt incurred by the student for enrollment in programs at other institutions. However, the Secretary may include loan debt incurred by the student for enrollment in GE programs at other institutions if the institution and the other institutions are under common ownership or control, as determined by the Secretary in accordance with 34 CFR 600.31.

(e) Exclusions. The Secretary excludes a student from both the numerator and the denominator of the D/E rates calculation if the Secretary determines that--

(1) One or more of the student’s title IV loans were in a military-related deferment status at any time during the calendar year for which the Secretary obtains earnings information under paragraph (c) of this section;

(2) One or more of the student’s title IV loans are under consideration by the Secretary, or have been approved, for a discharge on the basis of the student’s total and permanent disability, under 34 CFR 674.61, 682.402, or 685.212;

(3) The student was enrolled in any other eligible program at the institution or at another institution during the calendar year for which the Secretary obtains earnings information under paragraph (c) of this section;

(4) For undergraduate GE programs, the student completed a higher credentialed undergraduate GE program at the institution subsequent to completing the program as of the end of the most recently completed award year prior to the calculation of the draft D/E rates under this section;

(5) For graduate GE programs, the student completed a higher credentialed graduate GE program at the institution subsequent to completing the program as of the end of the most recently completed award year prior to the calculation of the draft D/E rates under this section; or

(6) The student died.

(f) D/E rates not issued. The Secretary does not issue draft or final D/E rates for a GE program under §668.405 if--

(1) After applying the exclusions in paragraph (e) of this section, fewer than 30 students completed the program during the two-year cohort period and fewer than 30 students completed the program during the four-year cohort period; or

(2) SSA does not provide the mean and median earnings for the program as provided under paragraph (c) of this section.

(g) Transition period. (1) The transition period is determined by the length of the GE program for which the Secretary calculates D/E rates under this subpart. The transition period is--

(i) The first five award years for which the Secretary calculates D/E rates under this subpart if the length of the program is one year or less;

(ii) The first six award years for which the Secretary calculates D/E rates under this subpart if the length of the program is between one and two years; and

(iii) The first seven award years for which the Secretary calculates D/E rates if the length of the program is more than two years.

(2) If a GE program is failing or in the zone based on its draft D/E rates for any award year during the transition period, the Secretary calculates transitional draft D/E rates for that award year by using--

(i) The median loan debt of the students who completed the program during the most recently completed award year; and

(ii) The earnings used to calculate the draft D/E rates under paragraph (c) of this section.

(3) For any award year for which the Secretary calculates transitional draft D/E rates for a program, the Secretary determines the final D/E rates for the program based on the lower of the draft or transitional draft D/E rates.

(4) An institution may challenge or appeal the draft or transitional draft D/E rates, or both, under the procedures in §668.405 and §668.406, respectively.

(Authority: 20 U.S.C. 1001, 1002, 1088, 1094)

§668.405 Issuing and challenging D/E rates.

(a) Overview. For each award year, the Secretary determines the D/E rates for a GE program at an institution by--

(1) Creating a list of the students who completed the program during the cohort period and providing the list to the institution, as provided in paragraph (b) of this section;

(2) Allowing the institution to correct the information about the students on the list, as provided in paragraph (c) of this section;

(3) Obtaining from SSA the mean and median annual earnings of the students on the list, as provided in paragraph (d) of this section;

(4) Calculating draft D/E rates and providing them to the institution, as provided in paragraph (e) of this section;

(5) Allowing the institution to challenge the median loan debt used to calculate the draft D/E rates, as provided in paragraph (f) of this section;

(6) Calculating final D/E rates and providing them to the institution, as provided in paragraph (g) of this section; and

(7) Allowing the institution to appeal the final D/E rates as provided in §668.406.

(b) Creating the list of students. (1) The Secretary selects the students to be included on the list by--

(i) Identifying the students who completed the program during the cohort period from the data provided by the institution under §668.411; and

(ii) Indicating which students would be removed from the list under §668.404(e) and the specific reason for the exclusion.

(2) The Secretary provides the list to the institution and states which cohort period was used to select the students.

(c) Institutional corrections to the list. (1) The Secretary presumes that the list of students and the identity information for those students are correct unless, as set forth in procedures established by the Secretary, the institution provides evidence to the contrary satisfactory to the Secretary. The institution bears the burden of proof that the list is incorrect.

(2) No later than 45 days after the date the Secretary provides the list to the institution, the institution may--

(i) Provide evidence showing that a student should be included on or removed from the list pursuant to §668.404(e); or

(ii) Correct or update a student’s identity information and the student’s program attendance information.

(3) After the 45-day period expires, the institution may no longer seek to correct the list of students or revise the identity or program information of those students included on the list.

(4) The Secretary considers the evidence provided by the institution and either accepts the correction or notifies the institution of the reasons for not accepting the correction. If the Secretary accepts the correction, the Secretary uses the corrected information to create the final list. The Secretary provides the institution with the final list and indicates the cohort period or cohort periods used to create the final list.

(d) Obtaining earnings data. The Secretary submits the final list to SSA. For the purposes of this section, SSA returns to the Secretary--

(1) The mean and median annual earnings of the students on the list whom SSA has matched to SSA earnings data, in aggregate and not in individual form; and

(2) The number, but not the identities, of students on the list that SSA could not match.

(e) Calculating draft D/E rates. (1)(i) If the SSA earnings data includes reports from records of earnings on at least 30 students, the Secretary uses the higher of the mean or median annual earnings provided by SSA to calculate draft D/E rates for a GE program, as provided in §668.404.

(ii) If the SSA earnings data includes reports from records of earnings on fewer than 30 but at least 10 students, the Secretary uses the earnings provided by SSA only for the purpose of disclosure under §668.412(a)(13).

(2) If SSA reports that it was unable to match one or more of the students on the final list, the Secretary does not include in the calculation of the median loan debt the same number of students with the highest loan debts as the number of students whose earnings SSA did not match. For example, if SSA is unable to match three students out of 100 students, the Secretary orders by amount the debts of the 100 listed students and excludes from the D/E rates calculation the three largest loan debts.

(3)(i) The Secretary notifies the institution of the draft D/E rates for the program and provides the mean and median annual earnings obtained from SSA and the individual student loan information used to calculate the rates, including the loan debt that was used in the calculation for each student.

(ii) The draft D/E rates and the data described in paragraphs (b) through (e) of this section are not considered public information.

(f) Institutional challenges to draft D/E rates. (1) The Secretary presumes that the loan debt information used to calculate the median loan debt for the program under §668.404 is correct unless the institution provides evidence satisfactory to the Secretary, as provided in paragraph (f)(2) of this section, that the information is incorrect. The institution bears the burden of proof to show that the loan debt information is incorrect and to show how it should be corrected.

(2) No later than 45 days after the Secretary notifies an institution of the draft D/E rates for a program, the institution may challenge the accuracy of the loan debt information that the Secretary used to calculate the median loan debt for the program under §668.404 by submitting evidence, in a format and through a process determined by the Secretary, that demonstrates that the median loan debt calculated by the Secretary is incorrect.

(3) In a challenge under this section, the Secretary does not consider--

(i) Any objection to the mean or median annual earnings that SSA provided to the Secretary;

(ii) More than one challenge to the student-specific data on which draft D/E rates are based for a program for an award year; or

(iii) Any challenge that is not timely submitted.

(4) The Secretary considers the evidence provided by an institution challenging the median loan debt and notifies the institution of whether the challenge is accepted or the reasons why the challenge is not accepted.

(5) If the information from an accepted challenge changes the median loan debt of the program, the Secretary recalculates the program’s draft D/E rates.

(6) Except as provided under §668.406, an institution that does not timely challenge the draft D/E rates for a program waives any objection to those rates.

(g) Final D/E rates. (1) After expiration of the 45-day period and subject to resolution of any challenge under paragraph (f) of this section, a program’s draft D/E rates constitute its final D/E rates.

(2) The Secretary informs the institution of the final D/E rates for each of its GE programs by issuing the notice of determination described in §668.409(a).

(3) After the Secretary provides the notice of determination to the institution, the Secretary may publish the final D/E rates for the program.

(h) Conditions for corrections and challenges. An institution must ensure that any material that it submits to make any correction or challenge under this section is complete, timely, accurate, and in a format acceptable to the Secretary and consistent with any instructions provided to the institution with the notice of its draft D/E rates and the notice of determination.

(Authority: 20 U.S.C. 1001, 1002, 1088, 1094)

§668.406 D/E rates alternate earnings appeals.

(a) General. If a GE program is failing or in the zone under the D/E rates measure, an institution may file an alternate earnings appeal to request recalculation of the program’s most recent final D/E rates issued by the Secretary. The alternate earnings must be from the same calendar year for which the Secretary obtained earnings data from SSA to calculate the final D/E rates under §668.404.

(b) Basis for appeals. (1) The institution may use alternate earnings from an institutional survey conducted under paragraph (c) of this section, or from a State-sponsored data system under paragraph (d) of this section, to recalculate the program’s final D/E rates and file an appeal if by using the alternate earnings--

(i) For a program that was failing the D/E rates measure, the program is passing or in the zone with respect to the D/E rates measure; or

(ii) For a program that was in the zone for the purpose of the D/E rates measure, the program is passing the D/E rates measure.

(2) When submitting its appeal of the final D/E rates, the institution must--

(i) Use the annual loan payment used in the calculation of the final D/E rates; and

(ii) Use the higher of the mean or median alternate earnings.

(3) The institution must include in its appeal the alternate earnings of all the students who completed the program during the same cohort period that the Secretary used to calculate the final D/E rates under §668.404 or a comparable cohort period, provided that the institution may elect--

(i) If conducting an alternate earnings survey, to exclude from the survey, in accordance with the standards established by NCES, all or some of the students excluded from the D/E rates calculation under §668.404(e); or

(ii) If obtaining annual earnings data from one or more State-sponsored data systems, and in accordance with paragraph (d)(2) of this section, to exclude from the list of students submitted to the administrator of the State-administered data system all or some of the students excluded from the D/E rates calculation under §668.404(e).

(c) Survey requirements for appeals. An institution must--

(1) In accordance with the standards included on an Earnings Survey Form developed by NCES, conduct a survey to obtain annual earnings information of the students described in paragraph (b)(3) of this section. The Secretary will publish in the Federal Register the Earnings Survey Form that will include a pilot-tested universe survey as well as the survey standards. An institution is not required to use the Earnings Survey Form but, in conducting a survey under this section, must adhere to the survey standards and present to the survey respondent in the same order and same manner the same survey items, included in the Earnings Survey Form; and

(2) Submit to the Secretary as part of its appeal--

(i) A certification signed by the institution’s chief executive officer attesting that the survey was conducted in accordance with the survey standards in the Earnings Survey Form, and that the mean or median earnings used to recalculate the D/E rates was accurately determined from the survey results;

(ii) An examination–level attestation engagement report prepared by an independent public accountant or independent governmental auditor, as appropriate, that the survey was conducted in accordance with the requirements set forth in the NCES Earnings Survey Form. The attestation must be conducted in accordance with the attestation standards contained in the Government Accountability Office’s Government Auditing Standards promulgated by the Comptroller General of the United States (available at www.gao.gov/yellowbook/overview or its successor site), and with procedures for attestations contained in guides developed by and available from the Department of Education’s Office of Inspector General; and

(iii) Supporting documentation requested by the Secretary.

(d) State-sponsored data system requirements for appeals. An institution must--

(1) Obtain annual earnings data from one or more State-sponsored data systems by submitting a list of the students described in paragraph (b)(3) of this section to the administrator of each State-sponsored data system used for the appeal;

(2) Demonstrate that annual earnings data were obtained for more than 50 percent of the number of students in the cohort period not excluded pursuant to paragraph (b)(3) of this section, and that number of students must be 30 or more; and

(3) Submit as part of its appeal--

(i) A certification signed by the institution’s chief executive officer attesting that it accurately used the State-provided earnings data to recalculate the D/E rates; and

(ii) Supporting documentation requested by the Secretary.

(e) Appeals procedure. (1) For any appeal under this section, in accordance with procedures established by the Secretary and provided in the notice of draft D/E rates under §668.405 and the notice of determination under §668.409, the institution must--

(i) Notify the Secretary of its intent to submit an appeal no earlier than the date that the Secretary provides the institution the draft D/E rates under §668.405(e)(3), but no later than 14 days after the date the Secretary issues the notice of determination under §668.409(a) informing the institution of the final D/E rates under §668.405(g); and

(ii) Submit the recalculated D/E rates, all certifications, and specified supporting documentation related to the appeal no later than 60 days after the date the Secretary issues the notice of determination.

(2) An institution that timely submits an appeal that meets the requirements of this section is not subject to any consequences under §668.410 based on the D/E rates under appeal while the Secretary considers the appeal. If the Secretary has published final D/E rates under §668.405(g), the program’s final D/E rates will be annotated to indicate that they are under appeal.

(3) An institution that does not submit a timely appeal waives its right to appeal the GE program’s failing or zone D/E rates for the relevant award year.

(f) Appeals determinations. (1) Appeals denied. If the Secretary denies an appeal, the Secretary notifies the institution of the reasons for denying the appeal, and the program’s final D/E rates previously issued in the notice of determination under §668.409(a) remain the final D/E rates for the program for the award year.

(2) Appeals granted. If the Secretary grants the appeal, the Secretary notifies the institution that the appeal is granted, that the recalculated D/E rates are the new final D/E rates for the program for the award year, and of any consequences of the recalculated rates under §668.410. If the Secretary has published final D/E rates under §668.405(g), the program’s published rates will be updated to reflect the new final D/E rates.

(g) Conditions for alternate earnings appeals. An institution must ensure that any material that it submits to make an appeal under this section is complete, timely, accurate, and in a format acceptable to the Secretary and consistent with any instructions provided to the institution with the notice of determination.

(Authority: 20 U.S.C. 1001, 1002, 1088, 1094)

§668.407 [Reserved].

§668.408 [Reserved].

§668.409 Final determination of the D/E rates measure.

(a) Notice of determination. For each award year for which the Secretary calculates a D/E rates measure for a GE program, the Secretary issues a notice of determination informing the institution of the following:

(1) The final D/E rates for the program as determined under §668.404, §668.405, and, if applicable, §668.406;

(2) The final determination by the Secretary of whether the program is passing, failing, in the zone, or ineligible, as described in §668.403, and the consequences of that determination;

(3) Whether the program could become ineligible based on its final D/E rates for the next award year for which D/E rates are calculated for the program;

(4) Whether the institution is required to provide the student warning under §668.410(a); and

(5) If the program’s final D/E rates are failing or in the zone, instructions on how it may make an alternate earnings appeal pursuant to §668.406.

(b) Effective date of Secretary’s final determination. The Secretary’s determination as to the D/E rates measure is effective on the date that is specified in the notice of determination. The determination, including, as applicable, the determination with respect to an appeal under §668.406, constitutes the final decision of the Secretary with respect to the D/E rates measure and the Secretary provides for no further appeal of that determination.

(Authority: 20 U.S.C. 1001, 1002, 1088, 1094)

§668.410 Consequences of the D/E rates measure.

(a) Student warning. (1) Events requiring a warning to students and prospective students. The institution must provide a warning with respect to a GE program to students and to prospective students for any year for which the Secretary notifies an institution that the program could become ineligible based on its final D/E rates measure for the next award year.

(2) Content of warning. Unless otherwise specified by the Secretary in a notice published in the Federal Register, the warning must--

(i) State that: “This program has not passed standards established by the U.S. Department of Education. The Department based these standards on the amounts students borrow for enrollment in this program and their reported earnings.

If in the future the program does not pass the standards, students who are then enrolled may not be able to use federal student grants or loans to pay for the program, and may have to find other ways, such as private loans, to pay for the program.”; and

(ii) Refer students and prospective students to (and include a link for) College Navigator, its successor site, or another similar Federal resource, for information about other similar programs.

(iii) For warnings provided to enrolled students--

(A) Describe the academic and financial options available to students to continue their education in another program at the institution, including whether the students could transfer credits earned in the program to another program at the institution and which course credits would transfer, in the event that the program loses eligibility for title IV, HEA program funds;

(B) Indicate whether or not the institution will--

(1) Continue to provide instruction in the program to allow students to complete the program; and

(2) Refund the tuition, fees, and other required charges paid to the institution by, or on behalf of, students for enrollment in the program; and

(C) Explain whether the students could transfer credits earned in the program to another institution.

(3) Consumer testing. The Secretary will conduct consumer testing to determine how to make the student warning as meaningful as possible.

(4) Alternative languages. To the extent practicable, the institution must provide alternatives to the English-language student warning for those students and prospective students for whom English is not their first language.

(5) Delivery to students. (i) An institution must provide the warning required under this section in writing to each student enrolled in the program no later than 30 days after the date of the Secretary’s notice of determination under §668.409 by--

(A) Hand-delivering the warning as a separate document to the student individually or as part of a group presentation; or

(B) Sending the warning to the primary email address used by the institution for communicating with the student about the program. 

(ii) If the institution sends the warning by email, the institution must--

(A) Ensure that the warning is the only substantive content in the email;

(B) Receive electronic or other written acknowledgement from the student that the student has received the email;

(C) Send the warning using a different address or method of delivery if the institution receives a response that the email could not be delivered; and

(D) Maintain records of its efforts to provide the warnings required by this section.

(6) Delivery to prospective students. (i) General. An institution must provide any warning required under this section to each prospective student or to each third party acting on behalf of the prospective student at the first contact about the program between the institution and the student or the third party acting on behalf of the student by--

(A) Hand-delivering the warning as a separate document to the prospective student or third party individually, or as part of a group presentation;

(B) Sending the warning to the primary email address used by the institution for communicating with the prospective student or third party about the program;

(C) Providing the prospective student or third party a copy of the disclosure template as required by §668.412(e) that includes the student warning required by this section; or

(D) Providing the warning orally to the student or third party if the contact is by telephone. 

(ii) Special warning requirements before enrolling a prospective student. (A) Before an institution enrolls, registers, or enters into a financial commitment with a prospective student with respect to the program, the institution must provide any warning required under this section to the prospective student in the manner prescribed in paragraph (a)(6)(i)(A) through (C) of this section.

(B) An institution may not enroll, register, or enter into a financial commitment with the prospective student with respect to the program earlier than--

(1) Three business days after the institution first provides the student warning to the prospective student; or

(2) If more than 30 days have passed from the date the institution first provided the student warning to the prospective student, three business days after the institution provides another warning as required by this paragraph.

(iii) Email delivery and acknowledgement. If the institution sends the warning to the prospective student or the third party by email, including by providing the prospective student or third party an electronic copy of the disclosure template, the institution must--

(A) Ensure that the warning is the only substantive content in the email;

(B) Receive electronic or other written acknowledgement from the prospective student or third party that the student or third party has received the email;

(C) Send the warning using a different address or method of delivery if the institution receives a response that the email could not be delivered; and

(D) Maintain records of its efforts to provide the warning required under this section.

(7) Disclosure template. Within 30 days of receiving notice from the Secretary that the institution must provide a student warning for the program, the institution must update the disclosure template described in §668.412 to include the warning in paragraph (a)(2) of this section or such other warning specified by the Secretary in a notice published in the Federal Register.

(b) Restrictions. (1) Ineligible program. Except as provided in §668.26(d), an institution may not disburse title IV, HEA program funds to students enrolled in an ineligible program.

(2) Period of ineligibility. (i) An institution may not seek to reestablish the eligibility of a failing or zone program that it discontinued voluntarily, reestablish the eligibility of a program that is ineligible under the D/E rates measure, or establish the eligibility of a program that is substantially similar to the discontinued or ineligible program, until three years following the date specified in the notice of determination informing the institution of the program’s ineligibility or the date the institution discontinued the failing or zone program.

(ii) An institution may not seek to reestablish the eligibility of a program that it discontinued voluntarily after receiving draft D/E rates that are failing or in the zone, or establish the eligibility of a program that is substantially similar to the discontinued program, until--

(A) Final D/E rates that are passing are issued for the program for that award year; or

(B) If the final D/E rates for the program for that award year are failing or in the zone, three years following the date the institution discontinued the program.

(iii) For the purposes of this section, an institution voluntarily discontinues a program on the date the institution provides written notice to the Secretary that it relinquishes the title IV, HEA program eligibility of that program.

(iv) For the purposes of this subpart, a program is substantially similar to another program if the two programs share the same four-digit CIP code. The Secretary presumes a program is not substantially similar to another program if the two programs have different four-digit CIP codes but the institution must provide an explanation of how the new program is not substantially similar to the ineligible or voluntarily discontinued program with its certification under §668.414.

(3) Restoring eligibility. An ineligible program, or a failing or zone program that an institution voluntarily discontinues, remains ineligible until the institution establishes the eligibility of that program under §668.414(c).

(Authority: 20 U.S.C. 1001, 1002, 1088, 1094, 1099c)

§668.411 Reporting requirements for GE programs.

(a) In accordance with procedures established by the Secretary, an institution must report--

(1) For each student enrolled in a GE program during an award year who received title IV, HEA program funds for enrolling in that program--

(i) Information needed to identify the student and the institution;

(ii) The name, CIP code, credential level, and length of the program;

(iii) Whether the program is a medical or dental program whose students are required to complete an internship or residency, as described in §668.402;

(iv) The date the student initially enrolled in the program;

(v) The student’s attendance dates and attendance status (e.g., enrolled, withdrawn, or completed) in the program during the award year; and

(vi) The student’s enrollment status (e.g., full-time, three-quarter time, half-time, less than half-time) as of the first day of the student’s enrollment in the program;

(2) If the student completed or withdrew from the GE program during the award year--

(i) The date the student completed or withdrew from the program;

(ii) The total amount the student received from private education loans, as described in §668.404(d)(1)(ii), for enrollment in the program that the institution is, or should reasonably be, aware of;

(iii) The total amount of institutional debt, as described in §668.404(d)(1)(iii), the student owes any party after completing or withdrawing from the program;

(iv) The total amount of tuition and fees assessed the student for the student’s entire enrollment in the program; and

(v) The total amount of the allowances for books, supplies, and equipment included in the student’s title IV Cost of Attendance (COA) for each award year in which the student was enrolled in the program, or a higher amount if assessed the student by the institution;

(3) If the institution is required by its accrediting agency or State to calculate a placement rate for either the institution or the program, or both, the placement rate for the program, calculated using the methodology required by that accrediting agency or State, and the name of that accrediting agency or State; and

(4) As described in a notice published by the Secretary in the Federal Register, any other information the Secretary requires the institution to report.

(b)(1) An institution must report the information required under paragraphs (a)(1) and (a)(2) of this section no later than--

(i) July 31, following the date these regulations take effect, for the second through seventh award years prior to that date;

(ii) For medical and dental programs that require an internship or residency, July 31, following the date these regulations take effect for the second through eighth award years prior to that date; and

(iii) For subsequent award years, October 1, following the end of the award year, unless the Secretary establishes different dates in a notice published in the Federal Register.

(2) An institution must report the information required under paragraph (a)(3) of this section on the date and in the manner prescribed by the Secretary in a notice published in the Federal Register.

(3) For any award year, if an institution fails to provide all or some of the information in paragraph (a) of this section to the extent required, the institution must provide to the Secretary an explanation, acceptable to the Secretary, of why the institution failed to comply with any of the reporting requirements.

(Authority: 20 U.S.C. 1001, 1002, 1088, 1231a)

§668.412 Disclosure requirements for GE programs.

(a) Disclosure template. An institution must use the disclosure template provided by the Secretary to disclose information about each of its GE programs to enrolled and prospective students. The Secretary will conduct consumer testing to determine how to make the disclosure template as meaningful as possible. The Secretary identifies the information that must be included in the template in a notice published in the Federal Register. That information may include, but is not limited to:

(1) The primary occupations (by name and SOC code) that the program prepares students to enter, along with links to occupational profiles on O*NET (www.onetonline.org) or its successor site.

(2) As calculated by the Secretary under §668.413, the program’s completion rates for full-time and less-than-full-time students and the program’s withdrawal rates.

(3) The length of the program in calendar time (i.e., weeks, months, years).

(4) The number of clock or credit hours or equivalent, as applicable, in the program.

(5) The total number of individuals enrolled in the program during the most recently completed award year.

(6) As calculated by the Secretary under §668.413, the loan repayment rate for any one or all of the following groups of students who entered repayment on title IV loans during the two-year cohort period:

(i) All students who enrolled in the program.

(ii) Students who completed the program.

(iii) Students who withdrew from the program.

(7) The total cost of tuition and fees, and the total cost of books, supplies, and equipment, that a student would incur for completing the program within the length of the program.

(8) The placement rate for the program, if the institution is required by its accrediting agency or State to calculate a placement rate either for the program or the institution, or both, using the required methodology of that accrediting agency or State.

(9) Of the individuals enrolled in the program during the most recently completed award year, the percentage who received a title IV loan or a private loan for enrollment in the program.

(10) As calculated by the Secretary, the median loan debt as determined under §668.413 of any one or all of the following groups:

(i) Those students who completed the program during the most recently completed award year.

(ii) Those students who withdrew from the program during the most recently completed award year.

(iii) All of the students described in paragraphs (a)(10)(i) and (ii) of this section.

(11) As provided by the Secretary, the mean or median earnings of any one or all of the following groups of students:

(i) Students who completed the program during the cohort period used by the Secretary to calculate the most recent D/E rates for the program under this subpart.

(ii) Students who were in withdrawn status at the end of the cohort period used by the Secretary to calculate the most recent D/E rates for the program under this subpart.

(iii) All of the students described in paragraph (a)(11)(i) and (ii) of this section.

(12) As calculated by the Secretary under §668.413, the most recent program cohort default rate.

(13) As calculated by the Secretary under §668.404, the most recent annual earnings rate.

(14) (i) Whether the program does or does not satisfy--

(A) The applicable educational prerequisites for professional licensure or certification in each State within the institution’s MSA; and

(B) The applicable educational prerequisites for professional licensure or certification in any other State for which the institution has made a determination regarding such requirements.

(ii) For any States not described in paragraph (a)(14)(i) of this section, a statement that the institution has not made a determination with respect to the licensure or certification requirements of those States.

(15) Whether the program is programmatically accredited and the name of the accrediting agency.

(16) A link to the U.S. Department of Education’s College Navigator Web sitehttp://nces.ed.gov/collegenavigator/, or its successor site, or other similar Federal resource.

(b) Disclosure updates. (1) In accordance with procedures and timelines established by the Secretary, the institution must update at least annually the information contained in the disclosure template with the most recent data available for each of its GE programs.

(2) The institution must update the disclosure template to include any student warning as required under §668.410(a)(7).

(c) Program Web pages. (1) On any Web page containing academic, cost, financial aid, or admissions information about a GE program maintained by or on behalf of an institution, the institution must provide the disclosure template for that program or a prominent, readily accessible, clear, conspicuous, and direct link to the disclosure template for that program.

(2) The Secretary may require the institution to modify a Web page if it provides a link to the disclosure template and the link is not prominent, readily accessible, clear, conspicuous, and direct.

(d) Promotional materials. (1) All promotional materials made available by or on behalf of an institution to prospective students that identify a GE program by name or otherwise promote the program must include--

(i) The disclosure template in a prominent manner; or

(ii) Where space or airtime constraints would preclude the inclusion of the disclosure template, the Web address (URL) of, or the direct link to, the disclosure template, provided that the URL or link is prominent, readily accessible, clear, conspicuous, and direct and the institution identifies the URL or link as “Important Information about the educational debt, earnings, and completion rates of students who attended this program” or as otherwise specified by the Secretary in a notice published in the Federal Register.

(2) Promotional materials include, but are not limited to, an institution’s catalogs, invitations, flyers, billboards, and advertising on or through radio, television, print media, the Internet, and social media.

(3) The institution must ensure that all promotional materials, including printed materials, about a GE program are accurate and current at the time they are published, approved by a State agency, or broadcast.

(e) Direct distribution to prospective students.

(1) Before a prospective student signs an enrollment agreement, completes registration, or makes a financial commitment to the institution, the institution must provide the prospective student or a third party acting on behalf of the prospective student, as a separate document, a copy of the disclosure template.

(2) The disclosure template may be provided to the prospective student or third party by--

(i) Hand-delivering the disclosure template to the prospective student or third party individually or as part of a group presentation; or

(ii) Sending the disclosure template to the primary email address used by the institution for communicating with the prospective student or third party about the program.

(3) If the institution hand-delivers the disclosure template to the prospective student or third party, it must obtain written confirmation from the prospective student or third party that the prospective student or third party received a copy of the disclosure template.

(4) If the institution sends the disclosure template to the prospective student or third party by email, the institution must--

(i) Ensure that the disclosure template is the only substantive content in the email;

(ii) Receive electronic or other written acknowledgement from the prospective student or third party that the prospective student or third party received the email;

(iii) Send the disclosure template using a different address or method of delivery if the institution receives a response that the email could not be delivered; and

(iv) Maintain records of its efforts to provide the disclosure template required under this section.

(f) Disclosure templates by program length, location, or format. (1) An institution that offers a GE program in more than one program length must publish a separate disclosure template for each length of the program. The institution must ensure that each disclosure template clearly identifies the applicable length of the program.

(2) An institution that offers a GE program in more than one location or format (e.g., full-time, part-time, accelerated) may publish a separate disclosure template for each location or format if doing so would result in clearer disclosures under paragraph (a) of this section. An institution that chooses to publish separate disclosure templates for each location or format must ensure that each disclosure template clearly identifies the applicable location or format.

(3) If an institution publishes a separate disclosure template for each length, or for each location or format, of the program, the institution must disaggregate, by length of the program, location, or format, those disclosures set forth in paragraphs (a)(4)-(a)(5), (a)(7)-(a)(9), and (a)(14) and as otherwise provided by the Secretary in a notice published in the Federal Register.

(g) Privacy considerations. An institution may not include on the disclosure template any of the disclosures described in paragraphs (a)(2), (a)(5), and (a)(6) or paragraphs (a)(8) through (a)(13) of this section if they are based on fewer than 10 students.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.413 Calculating, issuing, and challenging completion rates, withdrawal rates, repayment rates, median loan debt, median earnings, and program cohort default rate.

(a)(1) General. Under the procedures in this section, the Secretary determines the completion rates, withdrawal rates, repayment rates, median loan debt, median earnings, and program cohort default rate an institution must disclose under §668.412 for each of its GE programs, notifies the institution of that information, and provides the institution an opportunity to challenge the calculations.

(2) Enrollment cohort. (i) Subject to paragraph (a)(2)(ii) of this section, for the purpose of calculating the completion and withdrawal rates under paragraph (b) of this section, the enrollment cohort is comprised of all the students who began enrollment in a GE program during an award year. For example, the students who began enrollment in a GE program during the 2014-2015 award year constitute the enrollment cohort for that award year.

(ii) A student is excluded from the enrollment cohort for the purpose of calculating the completion and withdrawal rates under paragraph (b) of this section if, while enrolled in the program, the student died or became totally and permanently disabled and was unable to continue enrollment on at least a half-time basis, as determined under the standards in 34 CFR 685.213.

(b) Calculating completion rates, withdrawal rates, repayment rates, median loan debt, median earnings, and program cohort default rate.

(1) Completion rates. For each enrollment cohort, the Secretary calculates the completion rates of a GE program as follows:

(i) For students whose enrollment status is full-time on the first day of the student’s enrollment in the program:

Number of full-time students in the enrollment cohort who completed the program within 100% of the length of the program

Number of full-time students in the enrollment cohort

and

Number of full-time students in the enrollment cohort who completed the program within 150% of the length of the program

Number of full-time students in the enrollment cohort

(ii) For students whose enrollment status is less than full-time on the first day of the student’s enrollment in the program:

Number of less-than-full-time students in the enrollment cohort who completed the program within 200% of the length of the program

Number of less-than-full-time students in the enrollment cohort

and

Number of less-than-full-time students in the enrollment cohort who completed the program within 300% of the length of the program

Number of less-than-full-time students in the enrollment cohort

(2) Withdrawal rate. For each enrollment cohort, the Secretary calculates two withdrawal rates for a GE program as follows:

(i) The percentage of students in the enrollment cohort who withdrew from the program within 100 percent of the length of the program;

(ii) The percentage of students in the enrollment cohort who withdrew from the program within 150 percent of the length of the program.

(3) Loan repayment rate. For an award year, the Secretary calculates a loan repayment rate for borrowers not excluded under paragraph (b)(3)(vi) of this section who enrolled in a GE program as follows:

Number of borrowers paid in full plus number of borrowers in active repayment

Number of borrowers entering repayment

(i) Number of borrowers entering repayment. The total number of borrowers who entered repayment during the two-year cohort period on FFEL or Direct Loans received for enrollment in the program.

(ii) Number of borrowers paid in full. Of the number of borrowers entering repayment, the number who have fully repaid all FFEL or Direct Loans received for enrollment in the program.

(iii) Number of borrowers in active repayment. Of the number of borrowers entering repayment, the number who, during the most recently completed award year, made loan payments sufficient to reduce by at least one dollar the outstanding balance of each of the borrower’s FFEL or Direct Loans received for enrollment in the program, including consolidation loans that include a FFEL or Direct Loan received for enrollment in the program, by comparing the outstanding balance of each loan at the beginning and end of the award year.

(iv) Loan defaults. A borrower who defaulted on a FFEL or Direct Loan is not included in the numerator of the loan repayment rate formula even if that loan has been paid in full or meets the definition of being in active repayment.

(v) Repayment rates for borrowers who completed or withdrew. The Secretary may modify the formula in this paragraph to calculate repayment rates for only those borrowers who completed the program or for only those borrowers who withdrew from the program.

(vi) Exclusions. For the award year the Secretary calculates the loan repayment rate for a program, the Secretary excludes a borrower from the repayment rate calculation if the Secretary determines that--

(A) One or more of the borrower’s FFEL or Direct loans were in a military-related deferment status at any time during the most recently completed award year;

(B) One or more of the borrower’s FFEL or Direct loans are either under consideration by the Secretary, or have been approved, for a discharge on the basis of the borrower’s total and permanent disability, under 34 CFR 682.402 or 685.212;

(C) The borrower was enrolled in any other eligible program at the institution or at another institution during the most recently completed award year; or

(D) The borrower died.

(4) Median loan debt for students who completed the GE program. For the most recently completed award year, the Secretary calculates a median loan debt for the students described in §668.412(a)(10)(i) who completed the GE program during the award year. The median is calculated on debt described in §668.404(d)(1).

(5) Median loan debt for students who withdrew from the GE program. For the most recently completed award year, the Secretary calculates a median loan debt for the students described in §668.412(a)(10)(ii) who withdrew from the program during the award year. The median is calculated on debt described in §668.404(d)(1).

(6) Median loan debt for students who completed and withdrew from the GE program. For the most recently completed award year, the Secretary calculates a median loan debt for the students described in §668.412(a)(10)(iii) who completed the GE program during the award year and those students who withdrew from the GE program during the award year. The median is calculated on debt described in §668.404(d)(1).

(7) Median earnings. The Secretary calculates the median earnings of a GE program as described in paragraphs (b)(8) through (b)(12) of this section.

(8) Median earnings for students who completed the GE program. (i) The Secretary determines the median earnings for the students who completed the GE program during the cohort period by--

(A) Creating a list of the students who completed the program during the cohort period and providing it to the institution, as provided in paragraph (b)(8)(ii) of this section;

(B) Allowing the institution to correct the information about the students on the list, as provided in paragraph (b)(8)(iii) of this section;

(C) Obtaining from SSA the median annual earnings of the students on the list, as provided in paragraph (b)(8)(iv) of this section; and

(D) Notifying the institution of the median annual earnings for the students on the list.

(ii) Creating the list of students. (A) The Secretary selects the students to be included on the list by--

(1) Identifying the students who were enrolled in the program and completed the program during the cohort period from the data provided by the institution under §668.411; and

(2) Indicating which students would be removed from the list under paragraph (b)(11) of this section and the specific reason for the exclusion.

(B) The Secretary provides the list to the institution and states which cohort period was used to select the students.

(iii) Institutional corrections to the list. (A) The Secretary presumes that the list of students and the identity information for those students are correct unless the institution provides evidence to the contrary that is satisfactory to the Secretary. The institution bears the burden of proof that the list is incorrect.

(B) No later than 45 days after the date the Secretary provides the list to the institution, the institution may--

(1) Provide evidence showing that a student should be included on or removed from the list pursuant to paragraph (b)(11) of this section or otherwise; or

(2) Correct or update a student’s identity information and the student’s program attendance information.

(C) After the 45-day period expires, the institution may no longer seek to correct the list of students or revise the identity or program information of those students included on the list.

(D) The Secretary considers the evidence provided by the institution and either accepts the correction or notifies the institution of the reasons for not accepting the correction. If the Secretary accepts the correction, the Secretary uses the corrected information to create the final list. The Secretary notifies the institution which students are included on the final list and the cohort period used to create the list.

(iv) Obtaining earnings data. If the final list includes 10 or more students, the Secretary submits the final list to SSA. For the purposes of this section, SSA returns to the Secretary--

(A) The median earnings of the students on the list whom SSA has matched to SSA earnings data, in aggregate and not in individual form; and

(B) The number, but not the identities, of students on the list that SSA could not match.

(9) Median earnings for students who withdrew from the program. (i) The Secretary determines the median earnings for the students who withdrew from the program during the cohort period by--

(A) Creating a list of the students who were enrolled in the program but withdrew from the program during the cohort period and providing it to the institution, as provided in paragraph (b)(9)(ii) of this section;

(B) Allowing the institution to correct the information about the students on the list, as provided in paragraph (b)(9)(iii) of this section;

(C) Obtaining from SSA the median annual earnings of the students on the list, as provided in paragraph (b)(9)(iv) of this section; and

(D) Notifying the institution of the median annual earnings for the students on the list.

(ii) Creating the list of students. (A) The Secretary selects the students to be included on the list by--

(1) Identifying the students who were enrolled in the program but withdrew from the program during the cohort period from the data provided by the institution under §668.411; and

(2) Indicating which students would be removed from the list under paragraph (b)(11) of this section and the specific reason for the exclusion.

(B) The Secretary provides the list to the institution and states which cohort period was used to select the students.

(iii) Institutional corrections to the list. (A) The Secretary presumes that the list of students and the identity information for those students are correct unless the institution provides evidence to the contrary that is satisfactory to the Secretary, in a format and process determined by the Secretary. The institution bears the burden of proof that the list is incorrect.

(B) No later than 45 days after the date the Secretary provides the list to the institution, the institution may--

(1) Provide evidence showing that a student should be included on or removed from the list pursuant to paragraph (b)(11) of this section or otherwise; or

(2) Correct or update a student’s identity information and the student’s program attendance information.

(C) After the 45-day period expires, the institution may no longer seek to correct the list of students or revise the identity or program information of those students included on the list.

(D) The Secretary considers the evidence provided by the institution and either accepts the correction or notifies the institution of the reasons for not accepting the correction. If the Secretary accepts the correction, the Secretary uses the corrected information to create the final list. The Secretary notifies the institution which students are included on the final list and the cohort period used to create the list.

(iv) Obtaining earnings data. If the final list includes 10 or more students, the Secretary submits the final list to SSA. For the purposes of this section SSA returns to the Secretary--

(A) The median earnings of the students on the list whom SSA has matched to SSA earnings data, in aggregate and not in individual form; and

(B) The number, but not the identities, of students on the list that SSA could not match.

(10) Median earnings for students who completed and withdrew from the program. The Secretary calculates the median earnings for both the students who completed the program during the cohort period and students who withdrew from the program during the cohort period in accordance with paragraphs (b)(8) and (b)(9) of this section.

(11) Exclusions from median earnings calculations. The Secretary excludes a student from the calculation of the median earnings of a GE program if the Secretary determines that--

(i) One or more of the student’s title IV loans were in a military-related deferment status at any time during the calendar year for which the Secretary obtains earnings information under this section;

(ii) One or more of the student’s title IV loans are under consideration by the Secretary, or have been approved, for a discharge on the basis of the student’s total and permanent disability, under 34 CFR 674.61, 682.402 or 685.212;

(iii) The student was enrolled in any other eligible program at the institution or at another institution during the calendar year for which the Secretary obtains earnings information under this section; or

(iv) The student died.

(12) Median earnings not calculated. The Secretary does not calculate the median earnings for a GE program if SSA does not provide the median earnings for the program.

(13) Program cohort default rate. The Secretary calculates the program cohort default rate using the methodology and procedures set forth in subpart R of this part.

(c) Notification to institutions. The Secretary notifies the institution of the--

(1) Draft completion, withdrawal, and repayment rates calculated under paragraph (b)(1) through (b)(3) of this section and the information the Secretary used to calculate those rates.

(2) Median loan debt of the students who completed the program, as described in paragraph (b)(4) of this section, the students who withdrew from the program, as described in paragraph (b)(5) of this section, and both the students who completed and withdrew from the program, as described in paragraph (b)(6) of this section, in each case during the cohort period.

(3) Median earnings of the students who completed the program, as described in paragraph (b)(8) of this section, the students who withdrew from the program, as described in paragraph (b)(9) of this section, or both the students who completed the program and the students who withdrew from the program, as described in paragraph (b)(10) of this section, in each case during the cohort period.

(4) Draft program cohort default rate, as described in paragraph (b)(13) of this section.

(d) Challenges to completion rates, withdrawal rates, repayment rates, median loan debt, median earnings, and program cohort default rate. (1) Completion rates, withdrawal rates, repayment rates, and median loan debt. (i) No later than 45 days after the Secretary notifies an institution of a GE program’s draft completion rate, withdrawal rate, repayment rate, and median loan debt, the institution may challenge the accuracy of the information that the Secretary used to calculate the draft rates and the draft median loan debt by submitting, in a form prescribed by the Secretary, evidence satisfactory to the Secretary demonstrating that the information was incorrect.

(ii) The Secretary considers any evidence provided by the institution challenging the accuracy of the information the Secretary used to calculate the rates and the median loan debt and notifies the institution whether the challenge is accepted or the reasons the challenge is not accepted. If the Secretary accepts the challenge, the Secretary uses the corrected data to calculate the rates or median loan debt.

(iii) An institution may challenge the Secretary’s calculation of the completion rates, withdrawal rates, repayment rates, and median loan debt only once for an award year. An institution that does not timely challenge the rates or median loan debt waives any objection to the rates or median loan debt as stated in the notice.

(2) Median earnings. The Secretary does not consider any challenges to the median earnings calculated under this section.

(3) Program cohort default rate. The Secretary considers any challenges to the program cohort default rate under the procedures for challenges set forth in subpart R of this part.

(e) Final calculations. (1) Completion rates, withdrawal rates, repayment rates, and median loan debt. (i) After expiration of the 45-day period, and subject to resolution of any challenge under paragraph (d)(1) of this section, a program’s draft completion rate, withdrawal rate, repayment rate, and median loan debt constitute the final rates and median loan debt for that program.

(ii) The Secretary informs the institution of the final completion rate, withdrawal rate, repayment rate, and median loan debt for each of its GE programs by issuing a notice of determination.

(iii) Unless paragraph (g) of this section applies, after the Secretary provides the notice of determination, the Secretary may publish the final completion rate, withdrawal rate, repayment rate, and median loan debt.

(2) Median earnings. The median earnings of a program calculated by the Secretary under this section constitute the final median earnings for that program. After the Secretary provides the institution with the notice in paragraph (c) of this section, the Secretary may publish the final median earnings for the program.

(3) Program cohort default rate. Subject to resolution of any challenge under subpart R of this part, a program’s program cohort default rate calculated by the Secretary under subpart R constitutes the official program cohort default rate for that program. After the Secretary provides the notice of determination, the Secretary may publish the official program cohort default rate.

(f) Conditions for challenges. An institution must ensure that any material that it submits to make any corrections or challenge under this section is--

(1) Complete, timely, accurate, and in a format acceptable to the Secretary as described in this subpart and, with respect to program cohort default rate, in subpart R of this part; and

(2) Consistent with any instructions provided to the institution with the notice of its draft completion, withdrawal, and repayment rates, median loan debt, or program cohort default rate.

(g) Privacy considerations. The Secretary does not publish a determination described in paragraphs (b)(1)-(b)(6), (b)(8) through (b)(10), and(b)(13) of this section, and an institution may not disclose a determination made by the Secretary or make any disclosures under those paragraphs, if the determination is based on fewer than 10 students.

(Authority: 20 U.S.C. 1001, 1002, 1088, 1094)

§668.414 Certification requirements for GE programs.

(a) Transitional certification for existing programs. (1) Except as provided in paragraph (a)(2) of this section, an institution must provide to the Secretary no later than December 31 of the year in which this regulation takes effect, in accordance with procedures established by the Secretary, a certification signed by its most senior executive officer that each of its currently eligible GE programs included on its Eligibility and Certification Approval Report meets the requirements of paragraph (d) of this section. The Secretary accepts the certification as an addendum to the institution’s program participation agreement with the Secretary under §668.14.

(2) If an institution makes the certification in its program participation agreement pursuant to paragraph (b) of this section between July 1 and December 31 of the year in which this regulation takes effect, it is not required to provide the transitional certification under this paragraph.

(b) Program participation agreement certification. As a condition of its continued participation in the title IV, HEA programs, an institution must certify in its program participation agreement with the Secretary under §668.14 that each of its currently eligible GE programs included on its Eligibility and Certification Approval Report meets the requirements of paragraph (d) of this section. An institution must update the certification within 10 days if there are any changes in the approvals for a program, or other changes for a program that make an existing certification no longer accurate.

(c) Establishing eligibility and disbursing funds. (1) An institution establishes the eligibility for title IV, HEA program funds of a GE program by updating the list of the institution’s eligible programs maintained by the Department to include that program, as provided under 34 CFR 600.21(a)(11)(i). By updating the list of the institution’s eligible programs, the institution affirms that the program satisfies the certification requirements in paragraph (d) of this section. Except as provided in paragraph (c)(2) of this section, after the institution updates its list of eligible programs, the institution may disburse title IV, HEA program funds to students enrolled in that program.

(2) An institution may not update its list of eligible programs to include a GE program, or a GE program that is substantially similar to a failing or zone program that the institution voluntarily discontinued or became ineligible as described in §668.410(b)(2), that was subject to the three-year loss of eligibility under §668.410(b)(2), until that three-year period expires.

(d) GE program eligibility certifications. An institution certifies for each eligible program included on its Eligibility and Certification Approval Report, at the time and in the form specified in this section, that--

(1) Each eligible GE program it offers is approved by a recognized accrediting agency or is otherwise included in the institution’s accreditation by its recognized accrediting agency, or, if the institution is a public postsecondary vocational institution, the program is approved by a recognized State agency for the approval of public postsecondary vocational education in lieu of accreditation;

(2) Each eligible GE program it offers is programmatically accredited, if such accreditation is required by a Federal governmental entity or by a governmental entity in the State in which the institution is located or in which the institution is otherwise required to obtain State approval under 34 CFR 600.9;

(3) For the State in which the institution is located or in which the institution is otherwise required to obtain State approval under 34 CFR 600.9, each eligible program it offers satisfies the applicable educational prerequisites for professional licensure or certification requirements in that State so that a student who completes the program and seeks employment in that State qualifies to take any licensure or certification exam that is needed for the student to practice or find employment in an occupation that the program prepares students to enter; and

(4) For a program for which the institution seeks to establish eligibility for title IV, HEA program funds, the program is not substantially similar to a program offered by the institution that, in the prior three years, became ineligible for title IV, HEA program funds under the D/E rates measure or was failing, or in the zone with respect to, the D/E rates measure and was voluntarily discontinued by the institution. The institution must include with its certification an explanation of how the new program is not substantially similar to any such ineligible or discontinued program.

(Authority: 20 U.S.C. 1001, 1002, 1088, 1094, 1099c)

§668.415 Severability.

If any provision of this subpart or its application to any person, act, or practice is held invalid, the remainder of the subpart or the application of its provisions to any person, act, or practice shall not be affected thereby.

(Authority: 20 U.S.C. 1001, 1002, 1088)

12. Add subpart R to read as follows:

Subpart R—Program Cohort Default Rate

Sec.

668.500 Purpose of this subpart

668.501 Definitions of terms used in this subpart

668.502 Calculating and applying program cohort default rates

668.503 Determining program cohort default rates for GE programs at institutions that have undergone a change in status

668.504 Draft program cohort default rates and your ability to challenge before official program cohort default rates are issued

668.505 Notice of the official program cohort default rate of a GE program

668.506 [Reserved]

668.507 Preventing evasion of program cohort default rates

668.508 General requirements for adjusting and appealing official program cohort default rates

668.509 Uncorrected data adjustments

668.510 New data adjustments

668.511 Erroneous data appeals

668.512 Loan servicing appeals

668.513 [Reserved]

668.514 [Reserved]

668.515 [Reserved]

668.516 Fewer-than-ten-borrowers determinations.

Subpart R—Program Cohort Default Rate

§668.500 Purpose of this subpart.

General. The program cohort default rate is a measure of a GE program offered by the institution. This subpart describes how program cohort default rates are calculated, and how you may request changes to your program cohort default rates or appeal the rate. Under this subpart, you submit a “challenge” after you receive your draft program cohort default rate, and you request an “adjustment” or “appeal” after your official program cohort default rate is published.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.501 Definitions of terms used in this subpart. We use the following definitions in this subpart:

(a) Cohort. Your cohort is a group of borrowers used to determine your program cohort default rate. The method for identifying the borrowers in a cohort is provided in §668.502(b).

(b) Data manager.

(1) For FFELP loans held by a guaranty agency or lender, the guaranty agency is the data manager.

(2) For FFELP loans that we hold, we are the data manager.

(3) For Direct Loan Program loans, the Secretary’s servicer is the data manager.

(c) Days. In this subpart, “days” means calendar days.

(d) Default. A borrower is considered to be in default for program cohort default rate purposes under the rules in §668.502(c).

(e) Draft program cohort default rate. Your draft program cohort default rate is a rate we issue, for your review, before we issue your official program cohort default rate. A draft program cohort default rate is used only for the purposes described in §668.504.

(f) Entering repayment.

(1) Except as provided in paragraphs (f)(2) and (f)(3), loans are considered to enter repayment on the dates described in 34 CFR 682.200 (under the definition of “repayment period”) and in 34 CFR 685.207, as applicable.

(2) A Federal SLS Loan is considered to enter repayment--

(i) At the same time the borrower’s Federal Stafford Loan enters repayment, if the borrower received the Federal SLS Loan and the Federal Stafford Loan during the same period of continuous enrollment; or 

(ii) In all other cases, on the day after the student ceases to be enrolled at an institution on at least a half-time basis in an educational program leading to a degree, certificate, or other recognized educational credential.

(3) For the purposes of this subpart, a loan is considered to enter repayment on the date that a borrower repays it in full, if the loan is paid in full before the loan enters repayment under paragraphs (f)(1) or (f)(2).

(g) Fiscal year. A fiscal year begins on October 1 and ends on the following September 30. A fiscal year is identified by the calendar year in which it ends.

(h) GE program. An educational program offered by an institution under §668.8(c)(3) or (d) and identified by a combination of the institution’s six-digit Office of Postsecondary Education ID (OPEID) number, the program’s six-digit CIP code as assigned by the institution or determined by the Secretary, and the program’s credential level, as defined in §668.402.

(i) Loan record detail report. The loan record detail report is a report that we produce. It contains the data used to calculate your draft or official program cohort default rate.

(j) Official program cohort default rate. Your official program cohort default rate is the program cohort default rate that we publish for you under §668.505.

(k) We. We are the Department, the Secretary, or the Secretary’s designee.

(l) You. You are an institution. We consider each reference to “you” to apply separately to the institution with respect to each of its GE programs.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.502 Calculating program cohort default rates.

(a) General. This section describes the four steps that we follow to calculate your program cohort default rate for a fiscal year:

(1) First, under paragraph (b), we identify the borrowers in your GE program’s cohort for the fiscal year. If the total number of borrowers in that cohort is fewer than 10, we also include the borrowers in your cohorts for the two most recent prior fiscal years for which we have data that identifies those borrowers who entered repayment during those fiscal years.

(2) Second, under paragraph (c), we identify the borrowers in the cohort (or cohorts) who are considered to be in default by the end of the second fiscal year following the fiscal year those borrowers entered repayment. If more than one cohort will be used to calculate your program cohort default rate, we identify defaulted borrowers separately for each cohort.

(3) Third, under paragraph (d), we calculate your program cohort default rate.

(4) Fourth, we apply your program cohort default rate to your program at all of your locations-- 

(i) As you exist on the date you receive the notice of your official program cohort default rate; and

(ii) From the date on which you receive the notice of your official program cohort default rate until you receive our notice that the program cohort default rate no longer applies.

(b) Identify the borrowers in a cohort.

(1) Except as provided in paragraph (b)(3), your cohort for a fiscal year consists of all of your current and former students who, during that fiscal year, entered repayment on any Federal Stafford Loan, Federal SLS Loan, Direct Subsidized Loan, or Direct Unsubsidized Loan that they received to attend the GE program, or on the portion of a loan made under the Federal Consolidation Loan Program or the Federal Direct Consolidation Loan Program that is used to repay those loans.

(2) A borrower may be included in more than one of your cohorts and may be included in the cohorts of more than one institution in the same fiscal year.

(3) A TEACH Grant that has been converted to a Federal Direct Unsubsidized Loan is not considered for the purpose of calculating and applying program cohort default rates.

(c) Identify the borrowers in a cohort who are in default.

(1) Except as provided in paragraph (c)(2), a borrower in a cohort for a fiscal year is considered to be in default if, before the end of the second fiscal year following the fiscal year the borrower entered repayment--

(i) The borrower defaults on any FFELP loan that was used to include the borrower in the cohort or on any Federal Consolidation Loan Program loan that repaid a loan that was used to include the borrower in the cohort (however, a borrower is not considered to be in default on a FFELP loan unless a claim for insurance has been paid on the loan by a guaranty agency or by us);

(ii) The borrower fails to make an installment payment, when due, on any Direct Loan Program loan that was used to include the borrower in the cohort or on any Federal Direct Consolidation Loan Program loan that repaid a loan that was used to include the borrower in the cohort, and the borrower’s failure persists for 360 days;

(iii) You or your owner, agent, contractor, employee, or any other affiliated entity or individual make a payment to prevent a borrower’s default on a loan that is used to include the borrower in that cohort; or

(iv) The borrower fails to make an installment payment, when due, on a Federal Stafford Loan that is held by the Secretary or a Federal Consolidation Loan that is held by the Secretary and that was used to repay a Federal Stafford Loan, if such Federal Stafford Loan or Federal Consolidation Loan was used to include the borrower in the cohort, and the borrower’s failure persists for 360 days.

(2) A borrower is not considered to be in default based on a loan that is, before the end of the second fiscal year following the fiscal year in which it entered repayment--

(i) Rehabilitated under 34 CFR 682.405 or 34 CFR 685.211(e); or

(ii) Repurchased by a lender because the claim for insurance was submitted or paid in error.

(d) Calculate the program cohort default rate. Except as provided in §668.503, if there are--

(1)(i) Ten or more borrowers in your cohort for a fiscal year, your program cohort default rate is the percentage that is calculated by--

(ii) Dividing the number of borrowers in the cohort who are in default, as determined under paragraph (c), by the number of borrowers in the cohort, as determined under paragraph (b).

(2) Fewer than 10 borrowers in your cohort for a fiscal year, your program cohort default rate is the percentage that is calculated by--

(i) For the first two years we attempt to calculate program cohort default rates under this part for a program, dividing the total number of borrowers in that program’s cohort and in the two most recent prior cohorts for which we have data to identify the individuals comprising the cohort who are in default, as determined for each program’s cohort under paragraph (c), by the total number of borrowers in that program cohort and the two most recent prior cohorts for which we have data to identify the individuals comprising the cohort, as determined for each program cohort under paragraph (b).

(ii) For other fiscal years, by dividing the total number of borrowers in that program cohort and in the two most recent prior program cohorts who are in default, as determined for each program cohort under paragraph (c), by the total number of borrowers in that program cohort and the two most recent prior program cohorts as determined for each program cohort under paragraph (b).

(iii) If we identify a total of fewer than ten borrowers under paragraph (d)(2), we do not calculate a draft program cohort default rate for that fiscal year.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.503 Determining program cohort default rates for GE programs at institutions that have undergone a change in status.

(a) General.

(1) If you undergo a change in status identified in this section, the program cohort default rate of a GE program you offer is determined under this section.

(2) In determining program cohort default rates under this section, the date of a merger, acquisition, or other change in status is the date the change occurs.

(3) [Reserved]

(4) If the program cohort default rate of a program offered by another institution is applicable to you under this section with respect to a program you offer, you may challenge, request an adjustment, or submit an appeal for the program cohort default rate under the same requirements that would be applicable to the other institution under §§668.504 and 668.508.

(b) Acquisition or merger of institutions. If you offer a GE program and your institution acquires, or was created by the merger of, one or more institutions that participated independently in the title IV, HEA programs immediately before the acquisition or merger and that offered the same GE program, as identified by its 6-digit CIP code and credential level--

(1) Those program cohort default rates published for a GE program offered by any of these institutions before the date of the acquisition or merger are attributed to the GE program after the merger or acquisition; and

(2) Beginning with the first program cohort default rate published after the date of the acquisition or merger, the program cohort default rates for that GE program are determined by including in the calculation under §668.502 the borrowers who were enrolled in that GE program from each institution that offered that program and that was involved in the acquisition or merger.

(c) [Reserved]

(d) Branches or locations becoming institutions. If you are a branch or location of an institution that is participating in the title IV, HEA programs, and you become a separate, new institution for the purposes of participating in those programs--

(1) The program cohort default rates published for a GE program before the date of the change for your former parent institution are also applicable to you when you offer that program;

(2) Beginning with the first program cohort default rate published after the date of the change, the program cohort default rates for a GE program for the next three fiscal years are determined by including the applicable borrowers who were enrolled in the GE program from your institution and from your former parent institution (including all of its locations) in the calculation under §668.502.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.504 Draft program cohort default rates and your ability to challenge before official program cohort default rates are issued.

(a) General.

(1) We notify you of the draft program cohort default rate of a GE program before the official program cohort default rate of the GE program is calculated. Our notice includes the loan record detail report for the draft program cohort default rate.

(2) Except as provided in §668.502(d)(2)(i), regardless of the number of borrowers included in the program cohort, the draft program cohort default rate of a GE program is always calculated using data for that fiscal year alone, using the method described in §668.502(d)(1).

(3) The draft program cohort default rate of a GE program and the loan record detail report are not considered public information and may not be otherwise voluntarily released to the public by a data manager.

(4) Any challenge you submit under this section and any response provided by a data manager must be in a format acceptable to us. This acceptable format is described in materials that we provide to you. If your challenge does not comply with these requirements, we may deny your challenge.

(b) Incorrect data challenges.

(1) You may challenge the accuracy of the data included on the loan record detail report by sending a challenge to the relevant data manager, or data managers, within 45 days after you receive the data. Your challenge must include--

(i) A description of the information in the loan record detail report that you believe is incorrect; and 

(ii) Documentation that supports your contention that the data are incorrect.

(2) Within 30 days after receiving your challenge, the data manager must send you and us a response that--

(i) Addresses each of your allegations of error; and

(ii) Includes the documentation that supports the data manager’s position.

(3) If your data manager concludes that draft data in the loan record detail report are incorrect, and we agree, we use the corrected data to calculate your program cohort default rate.

(4) If you fail to challenge the accuracy of data under this section, you cannot contest the accuracy of those data in an uncorrected data adjustment under §668.509, or in an erroneous data appeal, under §668.511.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.505 Notice of the official program cohort default rate of a GE program.

(a) We notify you of the official program cohort default rate of a GE program after we calculate it. After we send our notice to you, we publish a list of GE program cohort default rates for all institutions.

(b) If one or more borrowers who were enrolled in a GE program entered repayment in the fiscal year for which the rate is calculated, you will receive a loan record detail report as part of your notification package for that program.

(c) You have five business days, from the date of our notification, as posted on the Department’s Web site, to report any problem with receipt of the notification package.

(d) Except as provided in paragraph (e), timelines for submitting, adjustments, and appeals begin on the sixth business day following the date of the notification package that is posted on the Department’s Web site.

(e) If you timely report a problem with receipt of your notification package under paragraph (c) and the Department agrees that the problem was not caused by you, the Department will extend the challenge, appeal, and adjustment deadlines and timeframes to account for a re-notification package.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.506 [Reserved]

§668.507 Preventing evasion of program cohort default rates.

In calculating the program cohort default rate of a GE program, the Secretary may include loan debt incurred by the borrower for enrolling in GE programs at other institutions if the institution and the other institutions are under common ownership or control, as determined by the Secretary in accordance with 34 CFR 600.31.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.508 General requirements for adjusting and appealing official program cohort default rates.

(a) [Reserved]

(b) Limitations on your ability to dispute a program cohort default rate.

(1) You may not dispute the calculation of a program cohort default rate except as described in this subpart.

(2) You may not request an adjustment, or appeal a program cohort default rate, under §668.509, §668.510, §668.511, or §668.512, more than once.

(c) Content and format of requests for adjustments and appeals. We may deny your request for adjustment or appeal if it does not meet the following requirements:

(1) All appeals, notices, requests, independent auditor’s opinions, management’s written assertions, and other correspondence that you are required to send under this subpart must be complete, timely, accurate, and in a format acceptable to us. This acceptable format is described in materials that we provide to you.

(2) Your completed request for adjustment or appeal must include-- 

(i) All of the information necessary to substantiate your request for adjustment or appeal; and 

(ii) A certification by your chief executive officer, under penalty of perjury, that all the information you provide is true and correct.

(d) Our copies of your correspondence. Whenever you are required by this subpart to correspond with a party other than us, you must send us a copy of your correspondence within the same time deadlines. However, you are not required to send us copies of documents that you received from us originally.

(e) Requirements for data managers’ responses.

(1) Except as otherwise provided in this subpart, if this subpart requires a data manager to correspond with any party other than us, the data manager must send us a copy of the correspondence within the same time deadlines.

(2) If a data manager sends us correspondence under this subpart that is not in a format acceptable to us, we may require the data manager to revise that correspondence’s format, and we may prescribe a format for that data manager’s subsequent correspondence with us.

(f) Our decision on your request for adjustment or appeal.

(1) We determine whether your request for an adjustment or appeal is in compliance with this subpart.

(2) In making our decision for an adjustment, under §668.509 or §668.510, or an appeal, under §668.511 or §668.512--

(i) We presume that the information provided to you by a data manager is correct unless you provide substantial evidence that shows the information is not correct; and

(ii) If we determine that a data manager did not provide the necessary clarifying information or legible records in meeting the requirements of this subpart, we presume that the evidence that you provide to us is correct unless it is contradicted or otherwise proven to be incorrect by information we maintain.

(3) Our decision is based on the materials you submit under this subpart. We do not provide an oral hearing.

(4) We notify you of our decision before we notify you of your next official program cohort default rate.

(5) You may not seek judicial review of our determination of a program cohort default rate until we issue our decision on all pending requests for adjustments or appeals for that program cohort default rate.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.509 Uncorrected data adjustments.

(a) Eligibility. You may request an uncorrected data adjustment for a GE program’s most recent cohort of borrowers used to calculate the most recent official program cohort default rate if, in response to your challenge under §668.504(b), a data manager agreed correctly to change the data, but the changes are not reflected in your official program cohort default rate.

(b) Deadlines for requesting an uncorrected data adjustment. You must send us a request for an uncorrected data adjustment, including all supporting documentation, within 30 days after you receive your loan record detail report from us.

(c) Determination. We recalculate your program cohort default rate, based on the corrected data, and correct the rate that is publicly released, if we determine that--

(1) In response to your challenge under §668.504(b), a data manager agreed to change the data; 

(2) The changes described in paragraph (c)(1) are not reflected in your official program cohort default rate; and 

(3) We agree that the data are incorrect.

(Authority: 20 U.S.C. 1001, 1002, 1088) 

§668.510 New data adjustments.

(a) Eligibility. You may request a new data adjustment for the most recent program cohort of borrowers, used to calculate the most recent official program cohort default rate for a GE program, if--

(1) A comparison of the loan record detail reports that we provide to you for the draft and official program cohort default rates shows that the data have been newly included, excluded, or otherwise changed; and

(2) You identify errors in the data described in paragraph (a)(1) that are confirmed by the data manager.

(b) Deadlines for requesting a new data adjustment.

(1) You must send to the relevant data manager, or data managers, and us a request for a new data adjustment, including all supporting documentation, within 15 days after you receive your loan record detail report from us.

(2) Within 20 days after receiving your request for a new data adjustment, the data manager must send you and us a response that--

(i) Addresses each of your allegations of error; and 

(ii) Includes the documentation used to support the data manager’s position.

(3) Within 15 days after receiving a guaranty agency’s notice that we hold an FFELP loan about which you are inquiring, you must send us your request for a new data adjustment for that loan. We respond to your request as set forth under paragraph (b)(2).

(4) Within 15 days after receiving incomplete or illegible records or data from a data manager, you must send a request for replacement records or clarification of data to the data manager and us.

(5) Within 20 days after receiving your request for replacement records or clarification of data, the data manager must--

(i) Replace the missing or illegible records; 

(ii) Provide clarifying information; or 

(iii) Notify you and us that no clarifying information or additional or improved records are available.

(6) You must send us your completed request for a new data adjustment, including all supporting documentation--

(i) Within 30 days after you receive the final data manager’s response to your request or requests; or 

(ii) If you are also filing an erroneous data appeal or a loan servicing appeal, by the latest of the filing dates required in paragraph (b)(6)(i) or in §668.511(b)(6)(i) or §668.512(c)(10)(i).

(c) Determination. If we determine that incorrect data were used to calculate your program cohort default rate, we recalculate your program cohort default rate based on the correct data and make corrections to the rate that is publicly released.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.511 Erroneous data appeals.

(a) Eligibility. Except as provided in §668.508(b), you may appeal the calculation of a program cohort default rate if--

(1) You dispute the accuracy of data that you previously challenged on the basis of incorrect data under §668.504(b); or

(2) A comparison of the loan record detail reports that we provide to you for the draft and official program cohort default rates shows that the data have been newly included, excluded, or otherwise changed, and you dispute the accuracy of that data.

(b) Deadlines for submitting an appeal.

(1) You must send a request for verification of data errors to the relevant data manager, or data managers, and to us within 15 days after you receive the notice of your official program cohort default rate. Your request must include a description of the information in the program cohort default rate data that you believe is incorrect and all supporting documentation that demonstrates the error.

(2) Within 20 days after receiving your request for verification of data errors, the data manager must send you and us a response that--

(i) Addresses each of your allegations of error; and 

(ii) Includes the documentation used to support the data manager’s position.

(3) Within 15 days after receiving a guaranty agency’s notice that we hold an FFELP loan about which you are inquiring, you must send us your request for verification of that loan’s data errors. Your request must include a description of the information in the program cohort default rate data that you believe is incorrect and all supporting documentation that demonstrates the error. We respond to your request as set forth under paragraph (b)(2).

(4) Within 15 days after receiving incomplete or illegible records or data, you must send a request for replacement records or clarification of data to the data manager and us.

(5) Within 20 days after receiving your request for replacement records or clarification of data, the data manager must--

(i) Replace the missing or illegible records; 

(ii) Provide clarifying information; or 

(iii) Notify you and us that no clarifying information or additional or improved records are available.

(6) You must send your completed appeal to us, including all supporting documentation-- 

(i) Within 30 days after you receive the final data manager’s response to your request; or 

(ii) If you are also requesting a new data adjustment or filing a loan servicing appeal, by the latest of the filing dates required in paragraph (b)(6)(i) or in §668.510(b)(6)(i) or §668.512(c)(10)(i).

(c) Determination. If we determine that incorrect data were used to calculate your program cohort default rate, we recalculate your program cohort default rate based on the correct data and correct the rate that is publicly released.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.512 Loan servicing appeals.

(a) Eligibility. Except as provided in §668.508(b), you may appeal, on the basis of improper loan servicing or collection, the calculation of the most recent program cohort default rate for a GE program.

(b) Improper loan servicing. For the purposes of this section, a default is considered to have been due to improper loan servicing or collection only if the borrower did not make a payment on the loan and you prove that the responsible party failed to perform one or more of the following activities, if that activity applies to the loan:

(1) Send at least one letter (other than the final demand letter) urging the borrower to make payments on the loan.

(2) Attempt at least one phone call to the borrower.

(3) Send a final demand letter to the borrower.

(4) For a FFELP loan held by us or for a Direct Loan Program loan, document that skip tracing was performed if the applicable servicer determined that it did not have the borrower’s current address.

(5) For an FFELP loan only-- 

(i) Submit a request for preclaims or default aversion assistance to the guaranty agency; and 

(ii) Submit a certification or other documentation that skip tracing was performed to the guaranty agency.

(c) Deadlines for submitting an appeal.

(1) If the loan record detail report was not included with your official program cohort default rate notice, you must request it within 15 days after you receive the notice of your official program cohort default rate.

(2) You must send a request for loan servicing records to the relevant data manager, or data managers, and to us within 15 days after you receive your loan record detail report from us. If the data manager is a guaranty agency, your request must include a copy of the loan record detail report.

(3) Within 20 days after receiving your request for loan servicing records, the data manager must-- 

(i) Send you and us a list of the borrowers in your representative sample, as described in paragraph (d) (the list must be in Social Security number order, and it must include the number of defaulted loans included in the program cohort for each listed borrower); 

(ii) Send you and us a description of how your representative sample was chosen; and 

(iii) Either send you copies of the loan servicing records for the borrowers in your representative sample and send us a copy of its cover letter indicating that the records were sent, or send you and us a notice of the amount of its fee for providing copies of the loan servicing records.

(4) The data manager may charge you a reasonable fee for providing copies of loan servicing records, but it may not charge more than $10 per borrower file. If a data manager charges a fee, it is not required to send the documents to you until it receives your payment of the fee.

(5) If the data manager charges a fee for providing copies of loan servicing records, you must send payment in full to the data manager within 15 days after you receive the notice of the fee.

(6) If the data manager charges a fee for providing copies of loan servicing records, and--

(i) You pay the fee in full and on time, the data manager must send you, within 20 days after it receives your payment, a copy of all loan servicing records for each loan in your representative sample (the copies are provided to you in hard copy format unless the data manager and you agree that another format may be used), and it must send us a copy of its cover letter indicating that the records were sent; or

(ii) You do not pay the fee in full and on time, the data manager must notify you and us of your failure to pay the fee and that you have waived your right to challenge the calculation of your program cohort default rate based on the data manager’s records. We accept that determination unless you prove that it is incorrect.

(7) Within 15 days after receiving a guaranty agency’s notice that we hold an FFELP loan about which you are inquiring, you must send us your request for the loan servicing records for that loan. We respond to your request under paragraph (c)(3).

(8) Within 15 days after receiving incomplete or illegible records, you must send a request for replacement records to the data manager and us.

(9) Within 20 days after receiving your request for replacement records, the data manager must either-- 

(i) Replace the missing or illegible records; or 

(ii) Notify you and us that no additional or improved copies are available.

(10) You must send your appeal to us, including all supporting documentation-- 

(i) Within 30 days after you receive the final data manager’s response to your request for loan servicing records; or 

(ii) If you are also requesting a new data adjustment or filing an erroneous data appeal, by the latest of the filing dates required in paragraph (c)(10)(i) or in §668.510(b)(6)(i) or §668.511(b)(6)(i).

(d) Representative sample of records.

(1) To select a representative sample of records, the data manager first identifies all of the borrowers for whom it is responsible and who had loans that were considered to be in default in the calculation of the program cohort default rate you are appealing.

(2) From the group of borrowers identified under paragraph (d)(1), the data manager identifies a sample that is large enough to derive an estimate, acceptable at a 95 percent confidence level with a plus or minus 5 percent confidence interval, for use in determining the number of borrowers who should be excluded from the calculation of the program cohort default rate due to improper loan servicing or collection.

(e) Loan servicing records. Loan servicing records are the collection and payment history records--

(1) Provided to the guaranty agency by the lender and used by the guaranty agency in determining whether to pay a claim on a defaulted loan; or 

(2) Maintained by our servicer that are used in determining your program cohort default rate.

(f) Determination.

(1) We determine the number of loans, based on the loans included in your representative sample of loan servicing records, that defaulted due to improper loan servicing or collection, as described in paragraph (b).

(2) Based on our determination, we use a statistically valid methodology to exclude the corresponding percentage of borrowers from both the numerator and denominator of the calculation of the program cohort default rate for the GE program, and correct the rate that is publicly released.

(Authority: 20 U.S.C. 1001, 1002, 1088)

§668.513 [Reserved]

§668.514 [Reserved]

§668.515 [Reserved]

§668.516 Fewer-than-ten-borrowers determinations.

We calculate an official program cohort default rate regardless of the number of borrowers included in the applicable cohort or cohorts. However, an institution may not disclose an official program cohort default rate under §668.412(a)(12) or otherwise, if the number of borrowers in the applicable cohorts is fewer than ten.

(Authority: 20 U.S.C. 1001, 1002, 1088)

Appendix A -- Regulatory Impact Analysis

This regulatory impact analysis (RIA) is divided into the following sections:

1. Need for Regulatory Action

In “Background” and “Outcomes and Practices” we discuss how high debt and relatively poor earnings affect students who enroll in gainful employment programs (“GE programs”). In “Basis of Regulatory Approach,” we consider the legislative history of the statutory provisions pursuant to which the Department is promulgating these regulations. “Regulatory Framework” provides an overview of the Department’s efforts, through these regulations, to establish an institutional accountability system for GE programs and to increase transparency of student outcomes in GE programs for the benefit of students, prospective students, and their families, the public, taxpayers, the Government, and institutions of higher education.

2. Analysis of the Regulations

Using data reported by institutions pursuant to the 2011 Prior Rule, we estimate how existing GE programs would have fared under these regulations and how students would have been impacted.

3. Costs, Benefits, and Transfers

The impact estimates provided in “Analysis of the Regulations” are used to consider the costs and benefits of the regulations to students, institutions, the Federal Government, and State and local governments. In “Net Budget Impacts” we estimate the budget impact of the regulations. We also provide a “Sensitivity Analysis” to demonstrate how alternative student and program impact assumptions would change our budget estimates.

4. Regulatory Alternatives Considered

In this section, we describe the other approaches the Department considered for key features of the regulations, including components of the D/E rates measures and possible alternative metrics.

5. Regulatory Flexibility Analysis

The RIA concludes with an analysis of the potential impact of the regulations on small businesses and non-profit institutions.

1. Need for Regulatory Action

Background

These regulations are intended to address growing concerns about educational programs that, as a condition of eligibility for title IV, HEA program funds, are required by statute to provide training that prepares students for gainful employment in a recognized occupation, but instead are leaving students with unaffordable levels of loan debt in relation to their income.

Through this regulatory action, the Department establishes: (1) an accountability framework for GE programs that defines what it means to prepare students for gainful employment in a recognized occupation by establishing measures by which the Department will evaluate whether a GE program remains eligible for title IV, HEA program funds, and (2) a transparency framework that will increase the quality and availability of information about the outcomes of students enrolled in GE programs.

The accountability framework defines what it means to prepare students for gainful employment by establishing measures that will assess whether programs provide quality education and training that allow students to pay back their student loan debt.

The transparency framework establishes reporting and disclosure requirements that will increase the transparency of student outcomes of GE programs so that information is disseminated to students, prospective students, and their families that is accurate and comparable to help them make better informed decisions about where to invest their time and money in pursuit of a postsecondary degree or credential. Further, this information will provide the public, taxpayers, and the Government with relevant information to understand the outcomes of the Federal investment in these programs. Finally, the transparency framework will provide institutions with meaningful information that they can use to improve the outcomes of students that attend their programs.

Outcomes and Practices

GE programs include non-degree programs, including diploma and certificate programs, at public and private non-profit institutions such as community colleges and nearly all educational programs at for-profit institutions of higher education regardless of program length or credential level. Common GE programs provide training for occupations in fields such as cosmetology, business administration, medical assisting, dental assisting, nursing, and massage therapy.

For fiscal year (FY) 2010, 37,589 GE programs with an enrollment of 3,985,329 students receiving title IV, HEA program funds reported program information to the Department.¹⁸⁶ About 61 percent of these programs are at public institutions, 6 percent at private non-profit institutions, and 33 percent at for-profit institutions. The Federal investment in students attending these programs is significant. In FY 2010, students attending GE programs received approximately $9.7 billion in Federal student aid grants and approximately $26 billion in Federal student aid loans.

Table 1.1 provides, by two-digit Classification of Instructional Program (CIP) code, the number of GE programs for which institutions reported program information to the Department in FY 2010. Table 1.2 provides the enrollment of students receiving title IV, HEA program funds in GE programs, by two-digit CIP code, for which institutions reported program information to the Department.

Table 1.1: FY 2010 GE Program Count

2-Digit CIP Code	2-Digit CIP Name	Public		Private		Proprietary							Total for All Sectors
		Ugrad cert	Post bacc cert	Ugrad cert	Post bacc cert	Ugrad cert	Associates	Bachelor's	Post bacc cert	Master's	Doctoral	First prof
51	Health Professions and Related Sciences	4,735	291	404	274	2,493	1,078	155	16	87	18	11	9,562
52	Business Management and Administrative Services	3,401	117	127	166	474	649	376	30	119	23	1	5,483
12	Personal and Miscellaneous Services	1,059	1	47	3	2,354	127	28	0	3	0	17	3,639
47	Mechanics and Repairs	2,254	2	54	0	266	84	0	0	0	0	0	2,660
11	Computer and Information Sciences	1,613	51	52	38	292	342	219	7	39	5	0	2,658
15	Engineering Related Technologies	1,689	11	42	6	143	145	23	1	1	0	0	2,061
50	Visual and Performing Arts	583	28	53	72	107	238	275	0	38	1	0	1,395
13	Education	389	298	29	389	52	19	57	22	78	30	1	1,364
43	Protective Services	869	11	15	21	55	189	112	6	23	3	0	1,304
48	Precision Production Trades	1,047	0	22	0	41	13	0	0	0	0	0	1,123
46	Construction Trades	956	0	24	0	98	26	2	0	0	0	0	1,106
22	Law and Legal Services	312	5	40	19	118	197	40	5	2	1	10	749
19	Home Economics	667	15	12	8	15	11	13	2	2	1	0	746
1	Agricultural Business and Production	502	2	5	0	7	1	1	0	0	0	0	518
10	Telecommunications Technologies	378	0	4	1	31	42	55	0	3	0	0	514
44	Public Administration and Services	146	41	7	21	0	8	11	2	16	6	0	258
9	Communications	131	15	10	22	19	15	37	0	5	0	0	254
49	Transportation and Material Moving Workers	170	0	5	2	28	7	6	1	2	0	0	221
31	Parks, Recreation, Leisure, and Fitness Studies	106	5	7	2	36	21	15	2	2	0	0	196
24	Liberal Arts and Sciences, General Studies and Humanities	130	1	4	4	2	22	17	1	4	1	0	186
30	Multi-interdisciplinary Studies	60	52	12	30	5	2	15	2	3	0	0	181
45	Social Sciences and History	79	48	4	22	1	4	18	0	3	0	0	179
42	Psychology	9	29	4	55	0	3	16	6	27	21	0	170
14	Engineering	39	44	1	14	4	6	15	1	8	0	0	132
16	Foreign Languages and Literature	105	11	2	8	1	0	5	0	0	0	0	132
23	English Language and Literature/Letters	53	24	10	7	7	2	10	0	3	0	0	116
39	Theological Studies and Religious Vocations	1	0	45	43	0	2	9	0	5	2	0	107
26	Biological and Biomedical Sciences	35	30	1	13	1	2	10	0	0	0	0	92
3	Conservation and Renewable Natural Resources	62	4	2	4	1	0	8	1	2	0	0	84
41	Science Technologies	70	1	0	0	2	5	0	0	0	0	0	78
4	Architecture and Related Programs	39	6	1	6	1	0	3	0	2	0	1	59
5	Area, Cultural, Ethnic, and Gender Studies	20	24	3	7	0	0	1	0	0	0	0	55
25	Library Studies	22	11	0	7	0	0	1	0	0	0	0	41
40	Physical Sciences	12	11	0	5	1	0	2	0	0	0	0	31
54	History	2	6	0	2	0	2	6	3	4	0	0	25
27	Mathematics and Statistics	4	14	3	1	0	1	1	0	0	0	0	24
38	Philosophy and Religious Studies	0	3	7	4	0	0	4	0	2	1	0	21
32	Basic Skills	10	1	1	0	3	0	0	0	0	0	0	15
34	Health-related Knowledge and Skills	6	0	2	1	4	0	0	0	0	0	0	13
36	Leisure and Recreational Activities	5	1	3	0	0	0	2	0	1	0	0	12
28	Reserve Officer Training Corps	1	0	0	0	2	1	1	1	0	0	0	6
60	Residency Programs	0	5	0	1	0	0	0	0	0	0	0	6
21	Technology/Education Industrial Arts	0	1	0	1	0	1	1	0	0	0	0	4
29	Military Technologies	0	0	0	0	1	2	1	0	0	0	0	4
33	Citizenship Activities	2	1	0	0	0	0	0	0	0	0	0	3
37	Personal Awareness and Self Improvement	1	0	0	0	0	0	0	0	0	0	0	1
53	High School/Secondary Diplomas and Certificates	1	0	0	0	0	0	0	0	0	0	0	1
Total		21,775	1,221	1,064	1,279	6,665	3,267	1,571	109	484	113	41	37,589

Table 1.2: FY 2010 Title IV Enrollment in GE Programs

2-Digit CIP Code	2-Digit CIP Name	Public			Private			Proprietary
		Ugrad cert	Post bacc cert	Ugrad cert		Post bacc cert	Ugrad cert		Associates	Bachelor's	Post bacc cert	Master's	Doctoral	First prof	Total for All Sectors
51	Health Professions and Related Sciences	277,010	2,475	35,356		3,130	445,923		306,061	94,512	735	41,885	5,035	9,116	1,221,238
52	Business Management and Administrative Services	129,593	1,690	3,904		2,180	16,174		231,033	308,843	2,184	109,180	15,357	0	820,138
12	Personal and Miscellaneous Services	44,669	0	3,169		6	198,590		34,860	5,857	0	15	0	568	287,734
43	Protective Services	57,765	152	841		171	3,209		115,239	85,657	90	8,098	1,014	0	272,236
11	Computer and Information Sciences	36,207	385	1,252		436	14,659		100,225	88,824	222	6,089	771	0	249,070
47	Mechanics and Repairs	67,155	6	3,878		0	79,074		15,040	0	0	0	0	0	165,153
13	Education	13,697	6,376	1,124		6,932	1,838		21,473	29,290	1,616	58,768	21,659	4	162,777
50	Visual and Performing Arts	14,935	153	1,104		548	6,573		36,354	66,897	0	3,166	13	0	129,743
15	Engineering Related Technologies	25,641	36	1,479		17	21,879		48,954	11,964	14	695	0	0	110,679
42	Psychology	1,021	711	10		1,071	0		463	36,866	218	18,666	12,990	0	72,016
22	Law and Legal Services	10,629	235	768		875	5,047		31,550	7,948	213	724	591	5,742	64,322
30	Multi-interdisciplinary Studies	1,448	507	57		209	74		32,287	23,772	117	2,076	0	0	60,547
19	Home Economics	50,594	133	946		78	785		999	2,846	85	1,442	446	0	58,354
44	Public Administration and Services	5,624	458	147		233	0		18,642	18,865	35	10,339	3,955	0	58,298
46	Construction Trades	21,776	0	1,988		0	13,271		2,529	51	0	0	0	0	39,615
48	Precision Production Trades	29,078	0	1,356		0	6,566		972	0	0	0	0	0	37,972
10	Telecommunications Technologies	9,587	0	105		2	3,730		4,841	12,737	0	490	0	0	31,492
24	Liberal Arts and Sciences, General Studies and Humanities	14,539	1	10		435	14		9,178	1,318	97	138	174	0	25,904
45	Social Sciences and History	741	381	76		391	89		61	14,869	0	740	0	0	17,348
23	English Language and Literature/Letters	8,436	156	1,142		21	2,059		3,668	1,476	0	119	0	0	17,077
9	Communications	3,684	85	63		112	2,046		873	8,424	0	277	0	0	15,564
49	Transportation and Material Moving Workers	4,109	0	725		22	7,518		436	430	3	146	0	0	13,389
31	Parks, Recreation, Leisure, and Fitness Studies	2,445	824	165		3	2,073		3,271	3,263	19	645	0	0	12,708
14	Engineering	980	385	7		289	46		149	5,241	1	174	0	0	7,272
1	Agricultural Business and Production	6,562	12	116		0	236		2	42	0	0	0	0	6,970
54	History	9	28	0		2	0		140	2,473	44	1,629	0	0	4,325
4	Architecture and Related Programs	2,718	114	1		89	2		0	114	0	97	0	532	3,667
3	Conservation and Renewable Natural Resources	1,253	5	5		52	7		0	2,075	6	258	0	0	3,661
16	Foreign Languages and Literature	2,574	48	4		47	27		0	30	0	0	0	0	2,730
38	Philosophy and Religious Studies	0	6	64		5	0		0	2,146	0	411	2	0	2,634
41	Science Technologies	1,602	3	0		0	169		422	0	0	0	0	0	2,196
26	Biological and Biomedical Sciences	482	282	1		45	71		107	719	0	0	0	0	1,707
39	Theological Studies and Religious Vocations	1	0	780		361	0		54	341	0	73	3	0	1,613
34	Health-related Knowledge and Skills	103	0	27		1	1,320		0	0	0	0	0	0	1,451
21	Technology/Education Industrial Arts	0	4	0		2	0		761	305	0	0	0	0	1,072
25	Library Studies	575	130	0		177	0		0	1	0	0	0	0	883
32	Basic Skills	176	1	10		0	366		0	0	0	0	0	0	553
5	Area, Cultural, Ethnic, and Gender Studies	133	140	14		17	0		0	1	0	0	0	0	305
36	Leisure and Recreational Activities	171	1	15		0	0		0	114	0	4	0	0	305
28	Reserve Officer Training Corps	5	0	0		0	11		17	139	10	0	0	0	182
40	Physical Sciences	70	34	0		36	0		0	17	0	0	0	0	157
27	Mathematics and Statistics	32	77	5		2	0		28	12	0	0	0	0	156
29	Military Technologies	0	0	0		0	12		62	4	0	0	0	0	78
60	Residency Programs	0	14	0		9	0		0	0	0	0	0	0	23
33	Citizenship Activities	6	1	0		0	0		0	0	0	0	0	0	7
37	Personal Awareness and Self Improvement	7	0	0		0	0		0	0	0	0	0	0	7
53	High School/Secondary Diplomas and Certificates	1	0	0		0	0		0	0	0	0	0	0	1
Total		847,843	16,049	60,714		18,006	833,458		1,020,751	838,483	5,709	266,344	62,010	15,962	3,985,329

Table 1.3 provides the percentage of students receiving title IV, HEA program funds in GE programs who fall within the following demographic categories: Pell grant recipients; received zero estimated family contribution (EFC) as indicated by their Free Application for Federal Student Aid (FAFSA); married; over the age of 24; veteran; and female.

Table 1.3: Characteristics of Students Enrolled in GE Programs (FY 2010)¹⁸⁷

Sector	Institution type	Credential level	Percent Pell Recipient	Percent zero estimated family contribution	Percent married	Percent above 24 in age	Percent of veteran	Percent female
Public	All		70.50%	41.50%	30.10%	66.20%	3.70%	70.10%
	< 2 year	Certificate	67.50%	37.30%	39.30%	72.00%	3.60%	83.70%
	2-3 year	Certificate	71.10%	43.20%	28.90%	65.20%	3.70%	69.60%
	4+ year	Certificate	63.60%	33.20%	30.30%	63.60%	4.30%	67.50%
		Post-Bacc Certificate	n/a	15.40%	47.00%	94.30%	4.00%	65.00%
Private	All		67.80%	40.80%	31.20%	63.60%	3.40%	67.00%
	< 2 year	Certificate	81.40%	52.10%	31.90%	63.30%	3.00%	53.90%
		Post-Bacc Certificate	n/a	33.30%	66.70%	100.00%	0.00%	66.70%
	2-3 year	Certificate	56.80%	38.60%	31.50%	64.20%	3.90%	71.00%
		Post-Bacc Certificate	n/a	26.70%	6.70%	93.30%	0.00%	86.70%
	4+ year	Certificate	69.10%	47.60%	28.60%	53.60%	2.60%	68.40%
		Post-Bacc Certificate	n/a	17.40%	37.30%	89.10%	5.10%	68.30%
For-Profit	All		63.70%	34.10%	36.60%	68.80%	10.50%	64.10%
	< 2 year	Certificate	75.60%	47.00%	27.10%	55.50%	2.90%	74.10%
		Associate's	96.00%	80.60%	34.30%	50.30%	2.30%	57.50%
		1st Professional Degree	n/a	51.30%	31.70%	56.20%	0.00%	94.70%
	2-3 year	Certificate	74.90%	43.40%	27.80%	53.90%	4.70%	65.40%
		Associate's	74.20%	44.40%	24.20%	54.00%	5.00%	62.90%
		Post-Bacc Certificate	n/a	16.80%	44.40%	86.00%	2.80%	79.20%
	4+ year	Certificate	72.10%	45.30%	33.60%	61.30%	4.60%	76.50%
		Associate's	60.00%	35.60%	38.90%	66.70%	11.80%	63.20%
		Bachelor's	55.30%	27.00%	39.40%	75.20%	14.70%	59.50%
		Post-Bacc Certificate	n/a	15.50%	43.70%	97.90%	8.00%	75.50%
		Master's	n/a	19.00%	48.30%	94.50%	14.00%	66.00%
		Doctoral	n/a	16.50%	48.90%	97.90%	14.60%	66.90%
		1st Professional Degree	n/a	27.10%	32.70%	80.90%	10.90%	52.40%
All	All		64.90%	34.70%	36.10%	68.50%	10.00%	64.50%

Research has demonstrated the significant benefits of postsecondary education. Among them are private pecuniary benefits¹⁸⁸ such as higher wages and social benefits such as a better educated and flexible workforce and greater civic participation.^{189 190 191 192} Even though the costs of postsecondary education have risen, there is evidence that the average financial returns to graduates have also increased.¹⁹³

Our analysis, provided in more detail in “Analysis of the Regulations,” reveals that low earnings and high rates of student loan default are common in many GE programs. For example, 27 percent of the 5,539 GE programs that the Department estimates would be assessed under the accountability metrics of the final regulations produced graduates with mean and median annual earnings below those of a full-time worker earning no more than the Federal minimum wage ($15,080).^{194 195} Approximately 22 percent of borrowers who attended programs that the Department estimates would be assessed under the accountability metrics of the final regulations defaulted on their Federal student loans within the first three years of entering repayment.¹⁹⁶

In light of the low earnings and high rates of default of graduates and borrowers at some GE programs, the Department is concerned that all students at these programs may not be making optimal educational and borrowing decisions. While many students appear to borrow less than might be optimal, either because they are risk averse or lack access to credit,¹⁹⁷ the outcomes previously described indicate that overborrowing may be a significant problem for at least some students.

Over the past three decades, student loan debt has grown rapidly as increases in college costs have outstripped increases in family income,¹⁹⁸ State and local postsecondary education funding has flattened,¹⁹⁹ and relatively expensive for-profit institutions have proliferated.²⁰⁰ Roughly only one-quarter of the increase in student debt in the past twenty-five years can be directly attributed to Americans obtaining more education.²⁰¹ Student loan debt now stands at over $1,096.5 billion nationally and rose by 80 percent, or $463.2 billion, between FY2008 and FY2013,²⁰² a period when other forms of consumer debt were flat or declining.²⁰³ Since 2003, the percentage of 25-year-olds with student debt has nearly doubled, increasing from 25 percent to 43 percent.²⁰⁴ Young people with student debt also owe more; the average student loan balance among 25-year-olds with debt has increased from $10,649 in 2003 to $20,326 in 2012.²⁰⁵

The increases in the percentage of young people with student debt and in average student debt loan balances have coincided with sluggish growth in State tax appropriations for higher education.²⁰⁶ While State funding for public institutions has stagnated, Federal student aid has increased dramatically. Overall Federal Pell Grant expenditures have grown from $7.96 billion in award year 2000-01 to approximately $32 billion in award year 2012-13, and Stafford Loan volumes have increased from $29.5 billion to $78 billion between award year 2000-01 and 2013-14.²⁰⁷ Much of the growth in overall Pell Grant expenditure is driven by an increase in recipients from approximately 4 million in award year 2000-01 to 8.8 million in 2013-14 and because the maximum Pell Grant grew by 10 percent after adjusting for inflation between 2003-2004 and 2013-2014.²⁰⁸

Other evidence suggests that student borrowing may not be too high for all students and at all institutions but rather, overborrowing results from specific and limited conditions.²⁰⁹ Although students may have access to information on average rates of return, they may not understand how their own abilities, choice of major, or choice of institution may affect their job outcomes or the expected value of the investment they make in their education.²¹⁰ Further, overborrowing may result because students do not understand the true cost of loans, because they overestimate their chance of graduating, or because they overestimate the earnings associated with the completion of their program of study.²¹¹

Inefficiently high borrowing can cause substantial harm to borrowers. There is some evidence suggesting that high levels of student debt decrease the long-term probability of marriage.²¹² For those who do not complete a degree, greater amounts of student debt may raise the probability of bankruptcy.²¹³ There is also evidence that it increases the probability of being credit constrained, particularly if students underestimate the probability of dropping out.²¹⁴ Since the Great Recession, student debt has been found to be associated with reduced home ownership rates.²¹⁵ And, high student debt may make it more difficult for borrowers to meet new mortgage underwriting standards, tightened in response to the recent recession and financial crisis.²¹⁶

Further, when borrowers default on their loans, everyday activities like signing up for utilities, obtaining insurance, and renting an apartment can become a challenge.²¹⁷ Such borrowers become subject to losing Federal payments and tax refunds and wage garnishment.²¹⁸ Borrowers who default might also be denied a job due to poor credit, struggle to pay fees necessary to maintain professional licenses, or be unable to open a new checking account.²¹⁹

There is ample evidence that students are having difficulty repaying their loans. The national two-year cohort default rate on Stafford loans has increased from 5.2 percent in 2006 to 10 percent in 2011.²²⁰ As of 2012, approximately 6 million borrowers were in default on Federal loans, owing $76 billion.²²¹

The determinants of default, which include both student and institutional characteristics, have been examined by many researchers. A substantial body of research suggests that “completing a postsecondary program is the strongest single predictor of not defaulting regardless of institution type.”²²² In a study of outcomes 10 years after graduation for students receiving BS/BA degrees in 1993, Lochner and Monge-Naranjo found that both student debt and post-school income levels are significant predictors of repayment and nonpayment, although the estimated effects were modest.²²³ In another study, Belfield examined the determinants of Federal loan repayment status of a more recent cohort of borrowers and found that loan balances had only a trivial influence on default rates.²²⁴ However, Belfield found substantial differences between students who attended for-profit institutions and those who attended public institutions. Even when controlling for student characteristics, measures of college quality, and college practices, students at for-profit institutions, especially two-year colleges, borrow more and have lower repayment rates than students at public institutions.²²⁵ Two recent studies also found that students who attend for-profit colleges have higher rates of default than comparable students who attend public colleges.^{226 227}

The causes of excessive debt, high default rates, and low earnings of students at GE programs include aggressive or deceptive marketing practices, a lack of transparency regarding program outcomes, excessive costs, low completion rates, deficient quality, and a failure to satisfy requirements such as licensing, work experience, and programmatic accreditation requirements needed for students to obtain higher paying jobs in a field. The outcomes of students who attend GE programs at for-profit educational institutions are of particular concern.

The for-profit sector has experienced tremendous growth in recent years,²²⁸ fueled in large part by Federal student aid funding and the increased demand for postsecondary education during the recent recession.²²⁹ The share of Federal student financial aid going to students at for-profit institutions has grown from approximately 13 percent of all title IV, HEA program funds in award year 2000-2001 to 19 percent in award year 2013-2014.²³⁰

The for-profit sector plays an important role in serving traditionally underrepresented populations of students. For-profit institutions are typically open-enrollment institutions that are more likely to enroll students who are older, women, Black, or Hispanic, or with low incomes.²³¹ Single parents, students with a certificate of high school equivalency, and students with lower family incomes are also more commonly found at for-profit institutions than community colleges.²³²

For-profit institutions develop curriculum and teaching practices that can be replicated at multiple locations and at convenient times, and offer highly structured programs to help ensure timely completion.²³³ For-profit institutions “are attuned to the marketplace and are quick to open new schools, hire faculty, and add programs in growing fields and localities.”²³⁴

At least some research suggests that for-profit institutions respond to demand that public institutions are unable to handle. Recent evidence from California suggests that for-profit institutions absorb students where public institutions are unable to respond to demand due to budget constraints.^{235 236} Additional research has found that “[c]hange[s] in for-profit college enrollments are more positively correlated with changes in State college-age populations than are changes in public-sector college enrollments.”²³⁷

Other evidence, however, suggests that for-profits are facing increasing competition from community colleges and traditional universities, as these institutions have started to expand their programs in online education. According to one annual report recently filed by a large, publically traded for-profit institution, “a substantial proportion of traditional colleges and universities and community colleges now offer some form of . . . online education programs, including programs geared towards the needs of working learners. As a result, we continue to face increasing competition, including from colleges with well-established brand names. As the online . . . learning segment of the postsecondary education market matures, we believe that the intensity of the competition we face will continue to increase.”²³⁸

On balance, we believe, and research confirms, that the for-profit sector has many positive features. There is also, however, growing evidence of troubling outcomes and practices at some for-profit institutions.

For-profit institutions typically charge higher tuitions than public postsecondary institutions. Among first-time full-time degree- or certificate-seeking undergraduates at title IV institutions operating on an academic calendar system and excluding students in graduate programs, average tuition and required fees at less-than-two-year for-profit institutions are more than double the average cost at less-than-two-year public institutions and average tuition and required fees at two-year for-profit institutions are about four times the average cost at two-year public institutions.^{239 240}

While for-profit institutions may need to charge more than public institutions because they do not have the State and local appropriation dollars and must pass the educational cost onto the student, there is some indication that even when controlling for government subsidies, for-profit institutions charge more than their public counterparts. To assess the role of government subsidies in driving this cost differential, Cellini conducted a sensitivity analysis comparing the costs of for-profit and community college programs. Her research found the primary costs to students at for-profit institutions, including foregone earnings, tuition, and loan interest, amounted to $51,600 per year on average, as compared with $32,200 for the same primary costs at community colleges. Further, Cellini’s analysis estimated taxpayer contributions, such as government grants, of $7,600 per year for for-profit institutions and $11,400 for community colleges.²⁴¹

Because aid received from grants has not kept pace with rising tuition in the for-profit sector, in contrast to other sectors, the net cost to students has increased sharply in recent years.²⁴² Not surprisingly, “student borrowing in the for-profit sector has risen dramatically to meet the rising net prices.”²⁴³ Students at for-profit institutions are more likely to receive Federal student financial aid and have higher average student debt than students in public and non-profit non-selective institutions.²⁴⁴

In 2011-2012, 60 percent of certificate-seeking students who were enrolled at for-profit institutions took out title IV, HEA student loans during that year compared to 10 percent at public two-year institutions.²⁴⁵ Of those who borrowed, the median loan amount borrowed by students enrolled in certificate programs at two-year for-profit institutions was $6,629 as opposed to $4,000 at public two-year institutions.²⁴⁶ In 2011-12, 66 percent students enrolled at for-profit institutions took out student loans, while only 20 percent of students enrolled at public two-year institutions took out student loans.²⁴⁷ Of those who borrowed in 2011-12, students enrolled in associate degree programs at two-year for-profit institutions had a median loan amount borrowed during 2011-12 of $7,583 in comparison to $4,467 for students at public two-year institutions.²⁴⁸

Although student loan default rates have increased in all sectors in recent years, they have consistently been highest among students attending for-profit institutions.^{249 250} Approximately 19 percent of borrowers who attended for-profit institutions default on their Federal student loans within the first three years of entering repayment as compared to about 13 percent of borrowers who attended public institutions.²⁵¹ Estimates of “cumulative lifetime default rates,” based on the number of loans, rather than borrowers, yield average default rates of 24, 23, and 31 percent, respectively, for public, private, and for-profit two-year institutions in the 2007-2011 cohort years. Based on estimates using dollars in those same cohort years (rather than loans or borrowers to estimate defaults) the average lifetime default rate is 50 percent for students who attended two-year for-profit institutions in comparison to 35 percent for students who attended two-year public and private institutions.²⁵²

There is growing evidence that many for-profit programs may not be preparing students for careers as well as comparable programs at public institutions. A 2011 GAO report reviewed results of licensing exams for 10 occupations that are among the largest fields of study, by enrollment, and found that, for nine out of 10 licensing exams, graduates of for-profit institutions had lower rates of passing than graduates of public institutions.²⁵³

Many for-profit institutions devote greater resources to recruiting and marketing than they do to instruction or to student support services.²⁵⁴ An investigation by the U.S. Senate Committee on Health, Education, Labor & Pensions (Senate HELP Committee) of 30 prominent for-profit institutions found that almost 23 percent of revenues were spent on marketing and recruiting but only 17 percent on instruction.²⁵⁵ A review of data provided by some of those institutions showed that they employed 35,202 recruiters compared with 3,512 career services staff and 12,452 support services staff.²⁵⁶

Lower rates of completion at many for-profit institutions are also a cause for concern. The six-year degree/certificate attainment rate of first-time undergraduate students who began at a four-year degree-granting institution in 2003-2004 was 34 percent at for-profit institutions in comparison to 67 percent at public institutions.²⁵⁷ However, it is important to note that, among first-time undergraduate students who began at a two-year degree-granting institution in 2003-2004, the six-year degree/certification attainment rate was 40 percent at for-profit institutions compared to 35 percent at public institutions.²⁵⁸

The slightly lower degree/certification attainment rates of two-year public institutions may at least be partially attributable to higher rates of transfer from two-year public institutions to other institutions.²⁵⁹ Based on available data, it appears that relatively few students transfer from for-profit institutions to other institutions. Survey data indicate about 5 percent of all student transfers originate from for-profit institutions, while students transferring from public institutions represent 64 percent of all transfers occurring at any institution (public two-year institutions to public four-year institutions being the most common type of transfer).²⁶⁰

Additionally, students who transfer from for-profit institutions are substantially less likely to be able to successfully transfer credits to other institutions than students who transfer from public institutions. According to a recent NCES study, an estimated 83 percent of first-time beginning undergraduate students who transferred from a for-profit institution to an institution in another sector were unable to successfully transfer credits to their new institution. In comparison, 38 percent of first-time beginning undergraduate students who transferred between two public institutions were unable to transfer credits to their new institution.²⁶¹

The higher costs of for-profit institutions and resulting greater amounts of debt incurred by their students, together with generally lower rates of completion, continue to raise concerns about whether some for-profit programs lead to earnings that justify the investment made by students, and additionally, taxpayers through the title IV, HEA programs.

In general, we believe that most programs operated by for-profit institutions produce positive educational and career outcomes for students. One study estimated moderately positive earnings gains, finding that “[a]mong associate’s degree students, estimates of returns to for-profit attendance are generally in the range of 2 to 8 percent per year of education.”²⁶² However, recent evidence suggests “students attending for-profit institutions generate earnings gains that are lower than those of students in other sectors.”²⁶³ The same study that found gains resulting from for-profit attendance in the range of 2 to 8 percent per year of education also found that gains for students attending public institution are “upwards of 9 percent.”²⁶⁴ But, other studies fail to find significant differences between the returns to students on educational programs at for-profit institutions and other sectors.²⁶⁵

Analysis of data collected on the outcomes of 2003-2004 first-time beginning postsecondary students as a part of the Beginning Postsecondary Students Longitudinal Study shows that students who attend for-profit institutions are more likely to be idle--not working or in school--six years after starting their programs of study in comparison to students who attend other types of institutions.²⁶⁶ Additionally, students who attend for-profit institutions and are no longer enrolled in school six years after beginning postsecondary education have lower earnings at the six-year mark than students who attend other types of institutions.²⁶⁷

These outcomes are troubling for two reasons. First, some students will not have sufficient earnings to repay the debt they incurred to enroll in these programs. Second, because the HEA limits the amounts of Federal grants and loans students may receive, their options to transfer to higher-quality and affordable programs may be constrained as they may no longer have access to sufficient student aid.²⁶⁸ These limitations make it even more critical that students’ initial choices in GE programs prepare them for employment that provides adequate earnings and do not result in excessive debt.

We also remain concerned that some for-profit institutions have taken advantage of the lack of access to reliable information about GE programs to mislead students. In 2010, the GAO released the results of undercover testing at 15 for-profit colleges across several States.²⁶⁹ Thirteen of the colleges tested gave undercover student applicants “deceptive or otherwise questionable information” about graduation rates, job placement, or expected earnings.²⁷⁰ The Senate HELP Committee investigation of the for-profit education sector also found evidence that many of the most prominent for-profit institutions engage in aggressive sales practices and provide misleading information to prospective students.²⁷¹ Recruiters described “boiler room”-like sales and marketing tactics and internal institutional documents showed that recruiters are taught to identify and manipulate emotional vulnerabilities and target non-traditional students.²⁷²

There has been growth in the number of qui tam lawsuits brought by private parties alleging wrongdoing at for-profit institutions, such as overstating job placement rates.²⁷³ Moreover, a growing number of State and other Federal law enforcement authorities have launched investigations into whether for-profit institutions are using aggressive or even deceptive marketing and recruiting practices.

Several State Attorneys General have sued for-profit institutions to stop fraudulent marketing practices, including manipulation of job placement rates. In 2013, the New York State Attorney General announced a $10.25 million settlement with Career Education Corporation (CEC), a private for-profit education company, after its investigation revealed that CEC significantly inflated its graduates’ job placement rates in disclosures made to students, accreditors, and the State.²⁷⁴ The State of Illinois sued Westwood College for misrepresentations and false promises made to students enrolling in the company’s criminal justice program.²⁷⁵ The Commonwealth of Kentucky has filed lawsuits against several private for-profit institutions, including National College of Kentucky, Inc., for misrepresenting job placement rates, and Daymar College, Inc., for misleading students about financial aid and overcharging for textbooks.²⁷⁶ And most recently, a group of 13 State Attorneys General issued Civil Investigatory Demands to Corinthian Colleges, Inc., Education Management Co., ITT Educational Services, Inc., and CEC, seeking information about job placement rate data and marketing and recruitment practices.²⁷⁷ The States participating include Arizona, Arkansas, Connecticut, Idaho, Iowa, Kentucky, Missouri, Nebraska, North Carolina, Oregon, Pennsylvania, Tennessee, and Washington.

Federal agencies have also begun investigations into the practices of some for-profit institutions. For example, the Consumer Financial Protection Bureau issued Civil Investigatory Demands to Corinthian Colleges, Inc. and ITT Educational Services, Inc. in 2013, demanding information about their marketing, advertising, and lending policies.²⁷⁸ The Securities and Exchange Commission also subpoenaed records from Corinthian Colleges, Inc. in 2013, seeking student information in the areas of recruitment, attendance, completion, placement, and loan defaults.²⁷⁹ And, the Department is also gathering and reviewing extensive amounts of data from Corinthian Colleges, Inc. regarding, in particular, the reliability of its disclosures of placement rates.²⁸⁰

The 2012 Senate HELP Committee report also found extensive evidence of aggressive and deceptive recruiting practices, excessive tuition, and regulatory evasion and manipulation by for-profit colleges in their efforts to enroll service members, veterans, and their families. The report described veterans being viewed as “dollar signs in uniform.”²⁸¹ The Los Angeles Times reported that recruiters from for-profit colleges have been known to recruit at Wounded Warriors centers and at veterans hospitals, where injured soldiers are pressured into enrolling through promises of free education and more.²⁸² There is evidence that some for-profit colleges take advantage of service members and veterans returning home without jobs through a number of improper practices, including by offering post-9/11 GI Bill benefits that are intended for living expenses as “free money.”²⁸³ Many veterans enroll in online courses simply to gain access to the monthly GI Bill benefits even if they have no intention of completing the coursework.²⁸⁴ In addition, some institutions have recruited veterans with serious brain injuries and emotional vulnerabilities without providing adequate support and counseling, engaged in misleading recruiting practices onsite at military installations, and failed to accurately disclose information regarding the graduation rates of veterans.²⁸⁵ In 2012, an investigation by 20 States, led by the Commonwealth of Kentucky’s Attorney General, resulted in a $2.5 million settlement with QuinStreet, Inc. and the closure of GIBill.com, a Web site that appeared as if it was an official site of the U.S. Department of Veterans Affairs, but was in reality a for-profit portal that steered veterans to 15 colleges, almost all for-profit institutions, including Kaplan University, the University of Phoenix, Strayer University, DeVry University, and Westwood College.²⁸⁶

Basis of Regulatory Approach

The components of the accountability framework that a program must satisfy to meet the gainful employment requirement are rooted in the legislative history of the predecessors to the statutory provisions of sections 101(b)(1), 102(b), 102(c), and 481(b) of the HEA that require institutions to establish the title IV, HEA program eligibility of GE programs. 20 U.S.C. 1001(b)(1), 1002(b)(1)(A)(i), (c)(1)(A), 1088(b).

The legislative history of the statute preceding the HEA that first permitted students to obtain federally financed loans to enroll in programs that prepared them for gainful employment in recognized occupations demonstrates the conviction that the training offered by these programs should equip students to earn enough to repay their loans. APSCU v. Duncan, 870 F.Supp.2d at 139; see also 76 FR 34392. Allowing these students to borrow was expected to neither unduly burden the students nor pose “a poor financial risk” to taxpayers. 76 FR 34392. Specifically, the Senate Report accompanying the initial legislation (the National Vocational Student Loan Insurance Act (NVSLIA), Pub. L. 89-287) quotes extensively from testimony provided by University of Iowa professor Dr. Kenneth B. Hoyt, who testified on behalf of the American Personnel and Guidance Association. On this point, the Senate Report sets out Dr. Hoyt’s questions and conclusions:

Would these students be in a position to repay loans following their training? . . .

If loans were made to these kinds of students, is it likely that they could repay them following training? Would loan funds pay dividends in terms of benefits accruing from the training students received? It would seem that any discussion concerning this bill must address itself to these questions. . . . .

We are currently completing a second-year followup of these students and expect these reported earnings to be even higher this year. It seems evident that, in terms of this sample of students, sufficient numbers were working for sufficient wages so as to make the concept of student loans to be [repaid] following graduation a reasonable approach to take. . . . I have found no reason to believe that such funds are not needed, that their availability would be unjustified in terms of benefits accruing to both these students and to society in general, nor that they would represent a poor financial risk.

Sen. Rep. No. 758 (1965) at 3745, 3748-49 (emphasis added).

Notably, both debt burden to the borrower and financial risk to taxpayers and the Government were clearly considered in authorizing federally backed student lending. Under the loan insurance program enacted in the NVSLIA, the specific potential loss to taxpayers of concern was the need to pay default claims to banks and other lenders if the borrowers defaulted on the loans. After its passage, the NVSLIA was merged into the HEA, which in title IV, part B, has both a direct Federal loan insurance component and a Federal reinsurance component, under which the Federal Government reimburses State and private non-profit loan guaranty agencies upon their payment of default claims. 20 U.S.C. 1071(a)(1). Under either HEA component, taxpayers and the Government assume the direct financial risk of default. 20 U.S.C. 1078(c) (Federal reinsurance for default claim payments), 20 U.S.C. 1080 (Federal insurance for default claims).

Not only did Congress consider expert assurances that vocational training would enable graduates to earn wages that would not pose a “poor financial risk” of default, but an expert observed that this conclusion rested on evidence that “included both those who completed and those who failed to complete the training.” APSCU v. Duncan, 870 F.Supp.2d at 139, citing H.R. Rep. No. 89-308, at 4 (1965), and S. Rep. No. 89-308, at 7, 1965 U.S.C.C.A.N. 3742, 3748.

The concerns regarding excessive student debt reflected in the legislative history of the gainful employment eligibility provisions of the HEA are as relevant now as they were then. Excessive student debt affects students and the country in three significant ways: payment burdens on the borrower; the cost of the loan subsidies to taxpayers; and the negative consequences of default (which affect borrowers and taxpayers).

The first consideration is payment burdens on the borrower. As we said in the NPRM, loan payments that outweigh the benefits of the education and training for GE programs that purport to lead to jobs and good wages are an inefficient use of a borrower's resources.

The second consideration is taxpayer subsidies. Borrowers who have low incomes but high debt may reduce their payments through income-driven repayment plans. These plans can either be at little or no cost to taxpayers or, through loan cancellation, can cost taxpayers as much as the full amount of the loan with interest. Deferments and repayment options are important protections for borrowers because, although postsecondary education generally brings higher earnings, there is no guarantee for the individual. Policies that assist those with high debt burdens are a critical form of insurance. However, these repayment options should not mean that institutions should increase the level of risk to the individual student or taxpayers through high-cost, low-value programs nor should institutions be the only parties without risk.

The third consideration is default. The Federal Government covers the cost of defaults on Federal student loans. These costs can be significant to taxpayers. Loan defaults also harm students and their families. They have to pay collection costs, their tax refunds and wages can be garnished, their credit rating is damaged, undermining their ability to rent a house, get a mortgage, or purchase a car, and, to the extent they can still get credit, they pay much higher interest. Increasingly, employers consider credit records in their hiring decisions. And, former students who default on Federal loans cannot receive additional title IV, HEA program funds for postsecondary education. See section 484(a)(3) of the HEA, 20 U.S.C. 1091(a)(3).

In accordance with the legislative intent behind the gainful employment eligibility provisions now found in sections 101, 102, and 481 of the HEA and the significant policy concerns they reflect, these regulations introduce certification requirements to establish a program’s eligibility and, to assess continuing eligibility, institute metrics-based standards that measure whether students will be able to pay back the educational debt they incur to enroll in the occupational training programs that are the subject of this rulemaking. 20 U.S.C. 1001(b)(1), 1002(b)(1)(A)(i), (c)(1)(A), 1088(b).

Regulatory Framework

As stated previously, the Department’s goals in the regulations are twofold: to establish an accountability framework and to increase the transparency of student outcomes of GE programs.

As part of the accountability framework, to determine whether a program provides training that prepares students for gainful employment as required by the HEA, the regulations set forth procedures to establish a program’s eligibility and to measure its outcomes on a continuing basis. To establish a program’s eligibility, an institution will be required to certify, among other things, that each of its GE programs meets all applicable accreditation and licensure requirements necessary for a student to obtain employment in the occupation for which the program provides training. This certification will be incorporated into the institution’s program participation agreement.

A GE program’s continuing eligibility will be assessed under the D/E rates measure, which compares the debt incurred by students who completed the program against their earnings. The regulations set minimum thresholds for the D/E rates measure. Programs with outcomes that meet the standards established by the thresholds will be considered to be passing the D/E rates measure and remain eligible to receive title IV, HEA program funds. Additionally, programs that do not meet the minimum requirements to be assessed under the D/E rates measure will also remain eligible to receive title IV, HEA program funds. Programs that are consistently unable to meet the standards of the D/E rates measure will eventually become ineligible to participate in the title IV, HEA programs.

An extensive body of research exists on the appropriate thresholds by which to measure the appropriateness of student debt levels relative to earnings. A 2006 study by Baum and Schwartz for the College Board defined “reliable benchmarks” to inform appropriate “levels of debt that will not unduly constrain the life choices facing former students.” The study determined that “the payment-to-income ratio should never exceed 18 to 20 percent” of discretionary income.²⁸⁷ A 2001 study by King and Frishberg found that students tend to overestimate the percentage of income they will be able to dedicate to student loan repayment, and asserted that based on lender recommendations, “8 percent of income is the most students should be paying on student loan repayment... assuming that most borrowers will be making major purchases, such as a home, in the 10 years after graduation.”²⁸⁸ Other studies have acknowledged or used the 8 percent standard as the basis for their work. In 2004, Harrast analyzed undergraduates’ ability to repay loans and cited the 8 percent standard to define excess debt as the difference between debt at graduation and lender-recommended levels for educational loan payments, finding that in all but a few cases, graduates in the upper debt quartile exceed the recommended level by a “significant margin.”²⁸⁹ Additionally, King and Bannon issued a report in 2002 acknowledging the 8 percent standard, and used it as the basis to estimate that 39 percent of all student borrowers graduate with unmanageable student loan debt.²⁹⁰

Several studies have proposed alternate measures and ranges for benchmarking debt burden, yet still acknowledge the 8 percent threshold as standard practice. In studying the repercussions from increasing student loan limits for Illinois’ students, the Illinois Student Assistance Commission noted in 2001 that other studies capture a range from 5 percent to 15 percent of gross income, but still indicated “it is generally agreed that when this ratio exceeds 8 percent, real debt burden may occur.”²⁹¹ The Commission also credited the National Association of Student Financial Aid Administrators (NASFAA) with adopting the 8 percent standard in 1986, after which it picked up wide support in the field.²⁹² A 2003 study by Baum and O’Malley analyzing how borrowers perceive their own levels of debt, recognized 8 percent standard for student loan debt but noted that “many loan administrators, lenders, and observers anecdotally suggest that a range of 8 to 12 percent may be considered acceptable.”²⁹³ This study also suggested that graduates devoting 7 percent or more of their income to student loan payments are much more likely to report repayment difficulty than those devoting smaller percentages of their incomes to loan payments. This is based on borrowers’ perceptions that repayment will rarely be problematic when payments are between 7 and 17 percent. In a 2012 study analyzing whether students were borrowing with the appropriate frequency and volume, Avery and Turner noted that 8 percent was both the most commonly referenced standard and a “manageable” one, but referenced a 2003 GAO study that set the benchmark at 10 percent.²⁹⁴

In addition to the accountability framework, the regulations include institutional reporting and disclosure requirements designed to increase the transparency of student outcomes for GE programs. Institutions will be required to report information that is necessary to implement aspects of the regulations that support the Department’s two goals of accountability and transparency. This includes information needed to calculate the D/E rates, as well as some of the specific required disclosures. The disclosure requirements will operate independently of the eligibility requirements and ensure that relevant information regarding GE programs is made available to students, prospective students, and their families, the public, taxpayers, and the Government, and institutions. The disclosure requirements will provide for transparency throughout the admissions and enrollment process so that students, prospective students, and their families can make informed decisions. Specifically, institutions will be required to make information regarding program costs and student completion, debt, earnings, and loan repayment available in a meaningful and easily accessible format.

Together, the certification requirements, accountability metrics, and disclosure requirements are designed to make improved and standardized market information about GE programs available to students, prospective students, and their families, the public, taxpayers, and the Government, and institutions; lead to a more competitive marketplace that encourages institutions to improve the quality of their programs and promotes institutions with high-performing programs; reduce costs and student debt; strengthen graduates’ employment prospects and earnings; eliminate poor performing programs; and improve the return on educational investment for students, families, taxpayers, and the Government.

The D/E Rates Measure

As previously stated, as part of the accountability framework, the D/E rates measure will be used to determine whether a GE program remains eligible for title IV, HEA program funds. The debt-to-earnings measures under both the 2011 Prior Rule and these regulations assess the debt burden incurred by students who completed a GE program in relation to their earnings.

The D/E rates measure will evaluate the amount of debt students who completed a GE program incurred to enroll in that program in comparison to those same students’ discretionary and annual earnings after completing the program. The regulations establish the standards by which the program will be assessed to determine, for each year rates are calculated, whether it passes or fails the D/E rates measure or is “in the zone.” Under the regulations, to pass the D/E rates measure, the GE program must have a discretionary income rate less than or equal to 20 percent or an annual earnings rate less than or equal to 8 percent. GE programs that have a discretionary income rate between 20 percent and 30 percent or an annual earnings rate between 8 percent and 12 percent will be considered to be in the zone. GE programs with a discretionary income rate over 30 percent and an annual earnings rate over 12 percent will fail the D/E rates measure. Under the regulations, a GE program will become ineligible for title IV, HEA program funds if it fails the D/E rates measure for two out of three consecutive years, or has a combination of D/E rates that are in the zone or failing for four consecutive years. The D/E rates measure and the thresholds are intended to assess whether students who complete a GE program face excessive debt burden relative to their income.

To allow institutions an opportunity to improve, the regulations include a transition period for the first several years after the regulations become effective. During these years, the transition period and zone together will allow institutions to make improvements to their programs in order to become passing.

The D/E rates measure assesses program outcomes that, consistent with legislative intent, indicate whether a program is preparing students for gainful employment. It is designed to reflect and account for two of the three primary reasons that a program may fail to prepare students for gainful employment, with former students unable to earn wages adequate to manage their educational debt: (1) a program does not train students in the skills they need to obtain and maintain jobs in the occupation for which the program purports to train students and (2) a program provides training for an occupation for which low wages do not justify program costs. The third primary reason that a program may fail to prepare students for gainful employment is that it is experiencing a high number of withdrawals or “churn” because relatively large numbers of students enroll but few, or none, complete the program, which can often lead to default.

The D/E rates measure assesses the outcomes of only those students who complete the program. The calculation includes former students who received title IV, HEA program funds--both loans and grants. For those students who have debt, the D/E rates take into account private loans and institutional financing in addition to title IV, HEA program loans.

The D/E rates measure primarily assesses whether the loan funds obtained by students “pay dividends in terms of benefits accruing from the training students received,” and whether such training has indeed equipped students to earn enough to repay their loans such that they are not unduly burdened. H.R. Rep. No. 89-308, at 4 (1965); S. Rep. No. 89-758, at 7 (1965). In addition to addressing Congress’ concern of ensuring that students’ earnings would be adequate to manage their debt, research also indicates that debt-to-earnings is an effective indicator of unmanageable debt burden. An analysis of a 2002 survey of student loan borrowers combined borrowers’ responses to questions about perceived loan burden, hardship, and regret to create a “debt burden index” that was significantly positively associated with borrowers’ actual debt-to-income ratios. In other words, borrowers with higher debt-to-income ratios tended to feel higher levels of burden, hardship, and regret.²⁹⁵

Accordingly, the D/E rates measure identifies programs that fail to adequately provide students with the occupational skills needed to obtain employment or that train students for occupations with low wages or high unemployment. The D/E rates also provide evidence of the experience of borrowers and, specifically, where borrowers may be struggling with their debt burden.

2. Analysis of the Regulations

Data and Methodology

Data

After the effective date of reporting and disclosure requirements under the 2011 Prior Rule on July 1, 2011, the Department received, pursuant to the reporting requirements, information from institutions on their GE programs for award years 2006-2007 through 2010-2011 (the “GE Data”). The GE Data is stored in the National Student Loan Database System (NSLDS), maintained by the Department’s Office of Federal Student Aid (FSA). The GE Data originally included information on students who received title IV, HEA program funds, as well as students who did not.  After the decisions in APSCU v. Duncan, the Department removed from NSLDS and destroyed the data on students²⁹⁶ who did not receive title IV, HEA program funds.

Using the remaining GE Data, student loan information also stored in NSLDS, and earnings information obtained from SSA, the Department calculated two debt-to-earnings (D/E) ratios, or rates, for GE programs. These D/E rates are the annual earnings rate and the discretionary income rate. The methodology that was used to calculate both rates is described in further detail below. We refer to the D/E rates data as the “2012 GE informational D/E rates.” The 2012 GE informational D/E rates are stored in a data file maintained by the Department that is accessible on its Web site.²⁹⁷ In addition to the D/E rates, we also calculated informational program level cohort default rates (pCDR) and repayment rates (RR).

A GE program is defined by a unique combination of the first six digits of its institution’s Office of Postsecondary Education Identification (“OPEID”) code, also referred to as the six-digit OPEID, the CIP code, and the program’s credential level. The terms OPEID code, CIP code, and credential level are defined below.

The 2012 GE informational D/E rates were calculated for programs in the GE Data based on the debt and earnings of the cohort of students receiving title IV, HEA program funds who completed GE programs during an “applicable two-year cohort period,” between October 1, 2007 and September 30, 2009 (the “08/09 D/E rates cohort”).²⁹⁸ The annual loan payment component of the debt-to-earnings formulas for the 2012 GE informational D/E rates was calculated for each program using student loan information from the GE Data and from NSLDS. The earnings components of the D/E rates formulas were calculated for each program using information obtained from SSA for the 2011 calendar year.

Unless otherwise specified, in accordance with the regulations, the Department analyzed the 2012 GE informational D/E rates only for those programs with 30 or more students who completed the program during the applicable two-year cohort period--that is, those programs that met the minimum “n-size” requirements.^{299 300} Of the 37,589 GE programs for which institutions reported program information for FY 2010 to the Department, 5,539 met the minimum n-size of 30 for the 2012 GE informational D/E rates calculations.

We estimated the number of programs that would, under the provisions in the regulations for the D/E rates measure, “pass,” “fail,” or fall in the “zone” based on their 2012 GE informational D/E rates results.

• Pass: Programs with an annual earnings rate less than or equal to 8 percent OR a discretionary income rate less than or equal to 20 percent.

• Zone: Programs that are not passing and have an annual earnings rate greater than 8 percent and less than or equal to 12 percent OR a discretionary income rate greater than 20 percent and less than or equal to 30 percent.

• Fail: Programs with an annual earnings rate over 12 percent AND a discretionary income rate over 30 percent.

Under the regulations, a program becomes ineligible for title IV, HEA program funds if it fails the D/E rates measure for two out of three consecutive years, or has a combination of D/E rates that are in the zone or failing for four consecutive years. The regulations establish a transition period for the first several years after the regulations become effective on July 1, 2015, to allow institutions an opportunity to improve their D/E rates by reducing the cost of their programs or the loan debt of their students.

The Department also analyzed the estimated impact of the regulations on GE programs using the following criteria:

• Enrollment: Number of students receiving title IV, HEA program funds for enrollment in a program. In order to estimate enrollment, we used the unduplicated count of students receiving title IV, HEA program funds in FY 2010.^{301 302}

• OPEID: Identification number issued by the Department that identifies each postsecondary educational institution (institution) that participates in the Federal student financial assistance programs authorized under title IV of the HEA.

• CIP code: Six-digit code that identifies instructional program specialties within educational institutions. These codes are derived from the Department’s National Center for Education Statistics’ (NCES) Classification of Instructional Programs, which is a taxonomy of instructional program classifications and descriptions.

• Sector: The sector designation for a program’s institution--public non-profit, private non-profit, or for-profit--using NSLDS sector data as of November 2013.³⁰³

• Institution type: The type designation for a program’s institution--less than 2 years, 2-3 years, and 4 years or more--using NSLDS data as of November 2013.

• Credential level: A program’s credential level--certificate, associate degree, bachelor’s degree, post-baccalaureate certificate, master’s degree, doctoral degree, and first professional degree.³⁰⁴

Methodology for D/E rates calculations

The methodology used by the Department to calculate the 2012 GE informational D/E rates departs slightly in some cases from the provisions in the regulations. We have identified those departures in footnotes.

As stated previously, the D/E rates measure is comprised of two debt-to-earnings ratios, or rates. The first, the discretionary income rate, is based on discretionary income, and the second, the annual earnings rate, is based on annual earnings. The formulas for the two D/E rates are:

discretionary income rate = annual loan payment_

discretionary income

annual earnings rate = annual loan payment

annual earnings

For the 2012 GE informational D/E rates, the annual earnings rates and discretionary income rates calculations are truncated two digits after the decimal place.

Although the Department calculated D/E rates for programs with an n-size of 10 or more, for the “Student demographics analysis of the final regulations” and “Impact Analysis of Final Regulations” sections of the RIA, the Department analyzed only those programs in the 2012 GE informational D/E rates data set with an n-size of 30 or more students who completed programs during the applicable two-year cohort period (FYs 2008-2009). It is important to note that under the regulations, if less than 30 students completed a program during the two-year cohort period, a four-year cohort period will be applied. If 30 or more students completed the program during the four-year cohort period, D/E rates will be calculated for that program. The 2012 GE informational D/E rates data set does not apply the four-year cohort period “look back” provisions.

A program’s annual loan payment is the median annual loan payment of the 08/09 D/E rates cohort and is calculated based on the cohort’s median total loan debt.³⁰⁵

• Each student’s total loan debt includes both FFEL and Direct Loans (except PLUS Loans made to parents or Direct Unsubsidized loans that were converted from TEACH Grants), private loans, and institutional loans that the student received for enrollment in the program.³⁰⁶

• In cases where a student completed multiple GE programs at the same institution, all loan debt is attributed to the highest credentialed program that the student completed and the student is not included in the calculation of D/E rates for the lower credentialed programs that the student completed.

• The total loan debt associated with each student is capped at an amount equivalent to the program’s tuition and fees³⁰⁷ if: (1) tuition and fees information for the student was provided by the institution, and (2) the amount of tuition and fees was less than the student’s total loan debt. This tuition and fees cap was applied to approximately 15 percent of student records for the 08/09 2012 D/E rates cohort.

• Excluded from the calculations are students whose loans were in military deferment or who were enrolled at an institution of higher education for any amount of time in the earnings calendar year, as defined below, or whose loans were discharged because of disability or death.

The median annual loan payment for each program was derived from the median total loan debt by assuming an amortization period and annual interest rate based on the credential level of the program.

• Amortization period:

  • 10 years for undergraduate certificate, associate degree, and post-baccalaureate certificate programs;³⁰⁸

  • 15 years for bachelor’s and master’s degree programs;

  • 20 years for doctoral and first professional degree programs.³⁰⁹

• Interest rate:

  • 6.8 percent for undergraduate certificate and associate degree programs;

  • 6.8 percent for post-baccalaureate certificate and master’s degree programs;

  • 5.42 percent for bachelor’s degree programs;

  • 5.42 percent for doctoral and first professional programs.

For undergraduate certificate, associate degree, post-baccalaureate certificate, and master’s degree programs, the rate is the average interest rate on Federal Direct Unsubsidized loans over the three years prior to the end of the applicable cohort period, in this case, the average rate over 2007-2009. For bachelor’s degree, doctoral, and first professional programs, the rate is the average interest rate on Federal Direct Unsubsidized loans over the six years prior to the end of the applicable cohort period, in this case, the average rate over 2004-2009. For undergraduate programs (certificate, associate degree, bachelor’s degree), the undergraduate Unsubsidized rate was applied, and for graduate programs (post-baccalaureate certificate, master’s, doctoral, first professional) the graduate rate was applied.³¹⁰

The annual earnings for the annual earnings rate calculation is either the mean or median annual earnings, whichever is higher, of the 08/09 D/E rates cohort for the calendar year immediately prior to the fiscal year for which the D/E rates are calculated. In this case, the D/E rates were calculated for the 2012 fiscal year. Accordingly, annual earnings were obtained from the SSA for the 2011 calendar year. Annual earnings include wages, salaries, tips, and self-employment income.

For calculating the discretionary income rate, discretionary income is the amount of the program’s mean or median, whichever is applicable, annual earnings above 150 percent of the Federal Poverty Guideline for a single person in the continental United States (FPL) for the annual earnings calendar year, in this case 2011, as published by the U.S. Department of Health and Human Services. The FPL for 2011 was $10,890.^{311 312}

Methodology for pCDR calculations

Program cohort default rates (“pCDR”) measure the proportion of a program’s borrowers who enter repayment on their loans in a given fiscal year that default within the first three years of repayment. The formula for pCDR is:

pCDR = borrowers whose loans are in default

borrowers whose loans entered repayment

The pCDR calculations are truncated to two-digits after the decimal point.

Generally, we analyzed pCDR only for those programs with a minimum n-size of 30 or more borrowers whose FFEL and Direct Loans for enrollment in the program entered repayment between October 1, 2008 and September 30, 2009 (FY 2009). However, if fewer than 30 students entered repayment during that fiscal year, we also included borrowers who entered repayment over the previous two fiscal years, October 1, 2006 to September 30, 2008 (FYs 2007 and 2008). If a program still did not reach 30 borrowers entering repayment, then a pCDR was not calculated. Of the 5,539 programs in the 2012 informational D/E rates data, 4,420 met the pCDR n-size requirements.³¹³ We refer to the pCDR data as the “2012 GE informational pCDRs.”

For the 2012 GE informational pCDRs, the denominator of the calculation is the number of borrowers whose loans entered repayment on their FFEL or Direct Loans in FY 2009, or if applicable, in FYs 2007-2009. The numerator of the calculation is the number of those borrowers who defaulted on FFEL or Direct Loans on or before September 30, 2011 (or if applicable, on or before September 30, 2009 and September 30, 2010 for borrowers entering repayment in FYs 2007 and 2008 respectively).

Methodology for repayment rate calculations

Repayment rates measure the proportion of a program’s borrowers who enter repayment on their loans in a given fiscal year that paid one dollar of principal in their third year of repayment. We refer to the repayment rate data as the 2011 GE informational repayment rates. The formula for repayment rate is:

RR= loans paid in full + loans in active repayment

original outstanding principal balance

Repayment rates were calculated by program for students who entered repayment on FFEL or Direct Loans received for enrollment in the program between October 1, 2006 and September 30, 2008 (FYs 2007 and 2008). We refer to these data as the “2011 GE informational repayment rates.”

For the 2011 GE informational repayment rates, the denominator of the calculation is the total original outstanding principal balance of FFEL and Direct Loans for borrowers who entered repayment in FYs 2007 and 2008. The numerator of the calculation is the total original outstanding principal balance of FFEL and Direct Loans for borrowers who entered repayment in FYs 2007 and 2008 on loans that have never been in default and that are fully paid plus the total original outstanding principal balance of FFEL and Direct Loans for borrowers who entered repayment in FYs 2007 and 2008 on loans that have never been in default and, for the period between October 1, 2010 and September 30, 2011 (FY 2011), whose balance was lower by at least one dollar at the end of the period than at the beginning. To account for negative amortization loans where borrowers could have been making full payments but still not paying down a dollar of principal, 3 percent of the original outstanding principal balance in the denominator was added to the numerator.

Student Demographics Analysis

In the 2014 NPRM, the Department provided the results of several regression analyses examining the relationship between demographic factors and program results under the D/E rates and pCDR measures. Several commenters cited to analysis by Charles River Associates and the Parthenon Group arguing that the Department provided insufficient detail regarding the methodology, data sources and data cleaning process, and types of regression models and variables it used for the regression analysis. These commenters also asserted that the Department should have reported more results than the R-squared statistics. Specifically, they contended that the Department should have provided the point-estimates and T-statistics. Although we believe that we sufficiently explained our analysis in the NPRM, we restate our analysis in greater detail here. We then provide the results of the Department’s student demographic analysis of the final regulations.

Explanation of terms

A regression analysis is a statistical method that can be used to measure relationships between variables. The demographic variables we analyze, provided below, are referred to as “independent” variables because they represent the potential inputs or causes of outcomes. The annual earnings rate and pCDR measures are referred to as “dependent” variables because they are the variables on which the effect of the independent variables are examined.

The output of a regression analysis contains several relevant points of information. The “coefficient,” also known as the point estimate, for each independent variable is the average amount that a dependent variable, in this case the annual earnings rate and pCDR, is expected to change with a one unit change in the associated independent variable, holding all other independent variables constant. The “T-statistic” is the ratio of the coefficient to its standard error. The T-statistic is commonly used to determine whether the relationship between the independent and dependent variables is “statistically significant.” The “R-squared” is the fraction of the variance of the dependent variable that is explained by the independent variables.

Student demographics analysis of 2014 NPRM

Methodology for student demographics analysis of 2014 NPRM

In the 2014 NPRM, the Department examined the association between demographic factors (independent variables) and the annual earnings rate and, separately, the pCDR measure (dependent variables). The Department did not conduct a regression analysis for the discretionary income rate because the discretionary income rate is simply a linear transformation of the annual earnings rate. As a result, the relationships that demographic factors have with the annual earnings rate will be broadly similar to those with the discretionary income rate.

For the NPRM, we used an ordinary least squares regression (robust standard errors), a common methodology that is used to model the relationship between a dependent variable and one or more independent variables by fitting a linear equation to observed data. One commenter argued that a Tobit regression would be more appropriate but, based on the commenter’s own analysis, acknowledged that both approaches lead to similar results. Because the ordinary least squares regression model is widely used, easily understood, and would not yield substantially different results, we have not changed our methodology for the student demographics analysis of the final regulations.

The first set of analysis we conducted examined the association of socioeconomic background and race and ethnicity with program outcomes. In performing these analyses, the Department used 2012 GE informational rate data, NSLDS data, and data reported by institutions to the Integrated Post-Secondary Education Data System (IPEDS).

The Department chose to use the proportion of title IV students enrolled in programs who were Pell Grant recipients (percent Pell) as a proxy for the average socioeconomic background of the students in GE programs because household income is the primary determinant of whether students qualify for Pell Grants. For both the annual earnings rate analysis and pCDR analysis, the proportion of Pell Grant recipients in each program was drawn from NSLDS. The percent Pell variable was determined by calculating the percentage of programs’ students who entered repayment on title IV, HEA program loans between October 1, 2007 and September 30, 2009, who also received a Pell Grant for attendance at the programs’ respective institutions between July 1, 2004 and July 30, 2009. The Department chose this five-year timeframe so that students who may have received a Pell Grant for a prior course of study but were no longer in economic hardship when they enrolled in the program being analyzed would not be assigned low socioeconomic status. We determined percent Pell for 4,938 of the 5,539 programs in the 2012 GE informational D/E rates data. We were unable to determine the percent Pell for all programs in the annual earnings rate regression analysis because some programs with a sufficient number of students who completed the program (30) between October 1, 2007 and September 30, 2009, to calculate D/E rates did not have any students entering repayment on title IV, HEA program loans during that period. For the pCDR regression analysis, we determined percent Pell for all programs in the 2012 GE informational pCDR data.

Because the Department does not collect race or ethnicity information from individual students receiving title IV, HEA program funds, we used data from IPEDs to estimate the proportion of minority students in programs (percent minority). The estimates for percentage of minority students in programs were derived differently for the annual earnings rate analysis and the pCDR analysis.

For the annual earnings rate analysis, we used the proportion of minority individuals who completed GE programs as reported in IPEDS 2008. For the purpose of this analysis, the term “minority” refers to individuals from American Indian or Alaska Native (Indian), Black or African American (Black), Hispanic or Latino/Hispanic (Hispanic), backgrounds, race and ethnicity groups that have historically been and continue to be underrepresented in higher education. For the annual earnings rate regression analysis, we determined percent minority for 3,886 of the 5,539 programs in the 2012 GE informational D/E rates data set. The remaining programs were excluded in the annual earnings rate regression. Many programs could not be matched primarily because IPEDS and NSLDS use different reporting mechanisms. For example, IPEDS and NSLDS use different unit identifiers for institutions. In addition, in reporting to the two systems, different CIPs are sometimes used. As a result, using IPEDS data for percent minority restricts the data set and provides at best an approximation of the racial and ethnic makeup of each program.

One commenter provided their own analysis using IPEDS data and argued that IPEDS data requires cleaning and manipulation. This commenter adjusted the IPEDS data for instances where the race and ethnicity categories do not add up to 100 percent, removed Puerto Rican programs from the sample, converted 2000 CIP codes to 2010 CIP codes, and aggregated branch programs reported in IPEDS to the GE program level. In the NPRM analysis, the Department converted IPEDS credential levels to GE credential levels and IPEDS OPEIDs³¹⁴ to six-digit OPEIDs and then aggregated the number of individuals who completed to the GE program level defined by unique combinations of six-digit OPEID, CIP code, and credential level in order to match IPEDS data to GE data. We did not adjust CIP codes or remove specific programs. Since then, the Department re-ran the analysis with all CIP codes converted to 2010 CIP codes, but results were not materially different. One commenter asserted that the proportion of individuals across categories of race and ethnicity may not add up to 100 percent for every program as a result of reporting errors to IPEDS.³¹⁵ However, the Department confirmed that the proportion of students in all race and ethnicity categories totaled to 100 percent of the total completions for each program in IPEDS. We do not agree that certain programs, such as Puerto Rican programs, should be removed as all programs under the regulation are relevant for the student demographics analysis.

As noted above, the sample size was limited for the percent Pell and minority variables. We determined percent minority for 3,886 and percent Pell for 4,938 of the 5,539 programs in the 2012 GE informational D/E rates data set . The resulting sample size of programs for which we determined both variables was 3,455. This may have biased the sample because the average annual earnings rate was 6.2 percent (standard deviation=4.7 percent) compared to an average annual earnings rate of 4.2 percent (standard deviation=4.6 percent) for the sample that did not have corresponding demographic data.

For the pCDR measure analysis, we used institution-level fall 2007 IPEDs data as a proxy for program-level percentages of minority students. Since the pCDR measure includes both students who do and who do not complete a program, there was no direct way in the data the Department had available to fully measure the race or ethnicity of students in the pCDR cohorts. The Department elected not to use the IPEDS program-level race or ethnicity data for individuals who completed a program because the race or ethnicity of students who completed a given GE program might differ substantially from the race or ethnicity of students who did not complete.

One commenter asserted that the use of institution-level data was not an appropriate methodology for this type of analysis. We acknowledge that institution-level data does not perfectly measure program-level demographic characteristics; however, there was no better source of data to approximate, at the program level, the percentage of minority students who both complete and do not complete a program.

While the first set of regression models in the NPRM analyzed the simple relationships between socioeconomic status and race or ethnicity and outcomes, the second set of regression models in the NPRM examined the effects of a broader range of characteristics on outcomes by controlling for the following additional independent variables:

• Institution Sector and Type: Public < 2 years, Public 2-3 years, Public 4+ years, Private < 2 years, Private 2-3 years, Private 4+ years, For-Profit < 2 years, For-Profit 2-3 years, For-Profit 4+ years.

• Credential Level: (01) Undergraduate certificate, (02) Associate degree, (03) Bachelor’s degree, (04) Post-Baccalaureate certificate, (05) Master’s degree, (06) Doctoral degree, (07) First Professional degree.

• Percentage of students that were:

  • Female.

  • Over the age of 24. We considered age over 24 as an indicator of nontraditional students because most traditional students begin their academic careers at an earlier age.

  • Had a zero estimated family contribution (EFC). We consider zero EFC status as an indicator of socioeconomic status because EFC is calculated based on household income.

The percent female, above age 24, and zero EFC for each program was determined using 2008 demographic profile data in NSLDS on students who entered repayment on title IV, HEA loans between October 1, 2007 and September 30, 2009. Some students who entered repayment in this time period did not have a 2008 demographic profile, so not all programs in the 2012 GE informational D/E rates and pCDR data sets had corresponding demographic data. Further, we were unable to determine the percent female, above age 24, and zero EFC for all programs in the annual earnings rate regression analysis because some programs with a sufficient number of students who completed the program (30) between October 1, 2007 and September 30, 2009, to calculate D/E rates did not have any students entering repayment on title IV, HEA program loans during that period. For the annual earnings rate regression analysis, we determined percent female, above age 24, and zero EFC for 4,687 of the 5,539 programs in the 2012 GE informational D/E rates data set. The resulting sample size of programs for which we determined all of the variables was 3,282. This may have biased the sample because the average annual earnings rate of these programs was 6.6 percent (standard deviation=4.7 percent) compared to an average annual earnings rate of 3.9 percent (standard deviation=4.6 percent) for the sample that did not have corresponding demographic data.

One commenter asserted that more variables should have been used in the regression, specifically enrollment status, average amount of title IV, HEA program funds received, and credential level. The commenter asserted that average amount of title IV, HEA program funds received is a better proxy of income than percent Pell because it provides detail on income level. . Although credential level was not identified as a variable in the description of the NPRM regression analysis, it was among the variables included in the second set of regression models in the NPRM. We did not include amount of title IV, HEA program funds received as a variable, however, because it is sensitive to cost of attendance and other factors. Finally, we did not include enrollment status because we were more accurately able to determine at the program level age above 24, which, like enrollment status, is also a proxy for nontraditional students.

One commenter argued that the sample of programs for the student demographics analysis was not large enough because it was limited to only programs in the 2012 GE informational D/E rates and pCDR data sets. As evidence of this, the commenter asserted that the top four program categories (health, business, computer/information science, and personal and culinary services) comprise 50 percent of the overall universe but 70 percent of the sample. We believe it is appropriate to analyze only those programs that our data estimates will be assessed under the regulations. Further, we do not believe the sample size is too small as there is significant variation within the sample of programs we analyzed. For example, percent Pell of the programs analyzed ranges from zero to 100 percent with a standard deviation of 25 percent (mean = 65 percent). The percent minority of the programs analyzed also ranges from zero to 100 percent with a standard deviation of 31 percent (mean = 36 percent).

Results of student demographics analysis of 2014 NPRM

The results of the Department’s student demographics regression analyses of the 2014 NPRM using annual earnings rates as the dependent variable are restated in greater detail below. We do not provide the same for the analysis using pCDR as the dependent variable as pCDR is not an accountability metric in the final regulations.

Table 2.1: Annual earnings rate regression- percent Pell & minority status (2014 NPRM version)

	Coefficient	Robust standard error	T-statistic	P-value
Minority	-0.01	0.003	-1.89	0.059
Pell	0.021	0.004	5.25	0

In order to investigate the criticism that the annual earnings rate measures primarily the socioeconomic status and racial/ethnic composition of the student body, the Department regressed program annual earnings rates on percent Pell and percent minority. As Table 2.1 shows, the Department found that programs with higher proportions of students who received Pell Grants tended to have slightly higher annual earnings rates, when controlling for percent minority. This relationship is statistically significant, but is small in magnitude. The results suggest that a one percent increase in a program's percentage of Pell students yields a 0.02 percent (coefficient) increase in the annual earnings rate. The T-statistic for minority status indicates the relationship between the percent minority variable and the annual earnings rate is not statistically significant when controlling for percent Pell.

Further, percent Pell and percent minority explained approximately one percent (R-squared) of the variance in annual earnings rate results. This suggests that a program’s annual earnings rate is influenced by much more than the socioeconomic and minority status of its students.

Table 2.2: Annual earnings rate regression- all independent variables (2014 NPRM version)

Number of obs	3282
R-squared	0.3583
	Coefficient	Robust standard error	T-statistic	P-value
Minority	-0.012	0.003	-3.870	0.000
Pell	-0.006	0.005	-1.130	0.260
Zero EFC	0.040	0.007	5.940	0.000
Female	0.001	0.002	0.510	0.608
Above age 24	-0.047	0.004	-11.800	0.000
Private4	-1.043	0.510	-2.050	0.041
Private<2	0.262	0.491	0.530	0.593
For-Profit2-3	1.330	0.303	4.390	0.000
For-Profit4	3.029	0.387	7.830	0.000
For-Profit<2	0.980	0.312	3.140	0.002
Public2-3	-3.006	0.303	-9.920	0.000
Public4	-2.330	0.517	-4.510	0.000
Public<2	-1.594	0.343	-4.650	0.000
Level_2	2.420	0.240	10.080	0.000
Level_3	3.080	0.438	7.040	0.000
Level_4	0.464	0.568	0.820	0.415
Level_5	-1.514	0.600	-2.520	0.012
Level_6	-2.974	0.737	-4.040	0.000
Level_7	0.635	4.629	0.140	0.891
Constant	6.495	0.493	13.170	0.000
Private 2-3 years used as reference group for sector and credential level 01 (undergraduate certificate) used as reference group for credential levels

To investigate whether other demographic or non-demographic factors could explain more of the variation in program annual earnings rates, the Department conducted a second regression with additional independent variables. The second regression used percent zero EFC, female, and above age 24 as independent variables in addition to percent Pell and percent minority. We controlled for the sector and type of a program’s institution and the credential level of the program. Holding constant other demographic, program, and institutional characteristics, the relationship between percent Pell and the annual earnings rate was no longer statistically significant. Another indicator of socioeconomic status, percent zero EFC, was positively associated with program annual earnings rate. However, interpretations of the percent Pell and percent zero EFC coefficients should be taken with caution because percent Pell and percent zero EFC are highly correlated (correlation coefficient = 0.72). These correlations are taken into account in the student demographics analysis of the final regulations provided below. In addition, percent above 24 was negatively associated with program annual earnings rate. Almost 36 percent (R-squared) of the variance in annual earnings rate results can be explained by the variables used in this analysis.

Several commenters referenced reports by Charles River Associates and the Parthenon Group which attempted to replicate the Department’s regression analysis in the NPRM using publicly available data and included additional analysis of the relationship between student characteristics and debt-to-earnings ratios using student-level data from a sample of for-profit institutions. The Parthenon Group analyzed Health-related programs, and engaged in a process to clean IPEDS data, which resulted in a sample set of 1,095 programs. The Parthenon Group asserted that the results of their regression analysis with annual earnings rate as the dependent variable and minority status, gender, age, Pell eligibility, average aid, enrollment status, and degree level as independent variables indicated that student characteristics explained 47 percent of the variance in annual earnings rates. The Parthenon Group’s analysis with pCDR as the dependent variable concluded that 63 percent of the variation resulted from student characteristics. Charles River Associates’ analysis used annual earnings rate and the pCDR from the 2012 GE informational D/E rates and pCDRs as dependent variables and IPEDS institutional Pell Grant data and program-level race and ethnicity data on the percentage of students who are Black, Indian, or Hispanic as independent variables. The R-squared value of the Charles River Associates model was 0.025 compared to less than 0.02 in the Department’s analysis. From its analysis, Charles River Associates concluded that Pell Grant status had a positive and significant relationship with both annual earnings rate and pCDR and minority status was positively correlated with pCDR but there was no statistically significant relationship between minority status and annual earnings rate.

Student demographics analysis of final regulations

In response to the NPRM, commenters asserted that the proposed regulations primarily measure student characteristics instead of program quality and that the regulations would deny postsecondary opportunities to low-income, minority, and female students by restricting access to postsecondary education. Some commenters conducted their own analyses with both publicly available data from IPEDS and non-publicly available data from several for-profit institutions. These commenters argued their analysis shows that the Department underestimated the explanatory power of student demographics on program results and that student demographics play an important part in explaining postsecondary outcomes.

Specifically, Charles River Associates conducted an analysis using student-level data for 10 different for-profit institutions combining NSLDS data with demographic information provided by institutions. These data were used in logistic regressions with three dummy dependent variables representing whether students completed, ever borrowed, or defaulted. The results were a series of odds ratios for propensity to graduate, borrow, and default that indicated that minority and Pell status matter for student outcomes. Among the findings were that African American students were less likely to borrow than white students (.92 percent compared to a reference group of white students), but 13 percent more likely to default. Hispanic students were not statistically different from white students with respect to the likelihood of graduation, but were 13 percent more likely to borrow and 36 percent more likely to default. Students who received Pell Grants were two times more likely to graduate and five times more likely to borrow, and, among students who borrow, 14 percent more likely to default. When limited to students who complete a program, Pell Grant recipients were 3.8 times more likely to borrow and 20 percent more likely to default than students who do not receive a Pell Grant. Regression with the another dependent variable, cumulative amount borrowed, indicated that the strongest predictors of amount borrowed are credential level and completion status, with students who do not complete borrowing approximately $6,700 less than students who do complete after accounting for the factors in the model.

To respond to these comments and to further examine the relationship between student demographics and program results under the annual earnings rate, the Department conducted additional analysis for the final regulations.

Methodology for student demographic analysis of final regulations

Similar to the NPRM methodology, the Department used ordinary least squares regressions to examine the relationship between student demographics and the program results under the final regulations. In addition, the Department conducted descriptive analyses of the 2012 GE informational D/E rates programs. Specifically, we examined the demographic composition of programs, comparing the composition of passing, zone, and failing programs.

We conducted regression analysis using only annual earnings rate as the dependent variable because pCDR is not an accountability metric in the final regulations. For this analysis, percent white, Black, Hispanic, Asian, Indian, two or more races, female, zero EFC, independent, and mother completed college, institutional sector and type, and program credential level were used as independent variables.³¹⁶

For the race and ethnicity variables, we used the proportion of individuals in each race and ethnicity category reported in the IPEDS 2008 data set. To match the IPEDS data to the 2012 GE informational D/E rates data set, the Department converted IPEDS credential levels to GE credential levels, converted IPEDS OPEIDs to six-digit OPEIDs, and converted all CIP codes to 2010 CIP codes.³¹⁷ We aggregated the number of completions reported for each program in IPEDS to the corresponding GE program definition of six-digit OPEID, CIP code, and credential level. While D/E rates measure only the outcomes of students who completed a program and received title IV, HEA program funds, IPEDS completions data include both title IV graduates and non-title IV graduates. We believe the IPEDS data provides a reasonable approximation of the proportion, by race and ethnicity, of title IV graduates completing GE programs. Unlike the NPRM analysis, we did not group multiple race and ethnicity categories into a single minority status variable because definitions of minority status may vary.³¹⁸ For the annual earnings rate regression analysis, we determined percent of each race and ethnicity category for 4,173 of the 5,539 programs in the 2012 GE informational D/E rates data set. Many programs could not be matched primarily because, as stated above, IPEDS and NSLDS use different reporting mechanisms, and the two reporting systems may not be consistent in matching data at the GE program-level. Because this resulted in a limited data set, the regression analysis was conducted both with and without the percent race and ethnicity variables.³¹⁹

Percent Pell for this analysis is the percentage of title IV students who completed a GE program between October 1, 2007 and September 30, 2009, who received a Pell Grant at any time in their academic career. Unlike the determination of percent Pell in the NPRM, which was based on all borrowers, we determined percent Pell based on all students who completed a program because those are the students whose outcomes are assessed by the annual earnings rate. Further, because Pell status is being used as a proxy for socioeconomic background, we counted students if they had received a Pell Grant at any time in their academic career, even if they did not receive it for enrollment in the program.

The following variables that were used in the NPRM analysis were also used in the analysis for the final regulations:

• Institution Sector. Public, Private, or For-Profit

• Credential Level. (01) Undergraduate certificate, (02) Associate degree, (03) Bachelor’s degree, (04) Post-Baccalaureate certificate, (05) Master’s degree, (06) Doctoral degree, (07) First Professional degree.

• Percentage of students:

  • Female.

  • Zero EFC. We consider zero EFC status as an indicator of socioeconomic status because EFC is calculated based on household income.

The percentage of students with the following characteristics were used as additional variables in the analysis for the final regulations but were not used in the NPRM analysis:

• Independent. Independent status is determined by a number of factors, including age, marital status, veteran status, and whether a student is claimed as a dependent by anyone for purposes of a tax filing.³²⁰ We consider independent students as an indicator that the student is non-traditional because most traditional students begin their studies as dependents.

• Married. Students who were married at the beginning of their academic careers. We consider married status to indicate the student is non-traditional because most traditional students are unmarried at the start of their academic careers.

• Mother of Students with College Education. Students whose mothers completed college. Children of mothers who completed college are more likely to attend and complete college.³²¹

The percent female, zero EFC, independent, married, and with mothers who completed college for each program were determined from the earliest demographic record (post-1995) in NSLDS for any title IV student who completed a GE program between October 1, 2007 and September 30, 2009. Unlike the determination of percentages of these variables in the NPRM, which was based on all borrowers, we determined the percentage of each of these variables based on all students who completed a program because those are the students whose outcomes are assessed by the annual earnings rate. Also, we determined these characteristics from each student’s earliest NSLDS record rather than just their status while in the program since these characteristics are being used as a proxy for socioeconomic background or to indicate that the student is non-traditional. With respect to these variables, we determined the composition of over 99 percent of the programs in the 2012 GE informational D/E rates data set.

Table 2.3 provides the program level descriptive statistics for the demographic variables.

Table 2.3: Descriptive Statistics of the Demographic Variables

Variable	Observations	Median	Mean	Standard deviation	Minimum	Maximum
Percent Asian	4,173	1.1	4.5	10.8	0.0	100.0
Percent Black	4,173	10.3	18.4	21.7	0.0	100.0
Percent Hispanic	4,173	7.1	19.3	27.5	0.0	100.0
Percent Indian	4,173	0.0	1.0	3.6	0.0	85.0
Percent Two or More Races	4,173	0.0	0.2	1.1	0.0	23.8
Percent White	4,173	63.2	56.6	31.9	0.0	100.0
Percent Zero EFC	5,537	40.2	42.0	21.3	0.0	100.0
Percent Independent	5,537	56.4	53.7	19.0	0.0	100.0
Percent Female	5,537	81.5	67.0	31.9	0.0	100.0
Percent Mothers College	5,537	25.0	26.3	11.4	0.0	100.0
Percent Pell	5,537	78.6	75.3	18.2	0.0	100.0

Student demographics descriptive analysis of final regulations

Table 2.4: Percent White, Black, Hispanic, Pell, Zero EFC, Female, Independent, Married, and Mothers Completing College in Passing, Zone, and Failing Programs

Table 2.4 shows that passing, zone, and failing programs have very similar proportions of low-income, non-traditional, female, white, Black, and Hispanic students.³²²

Table 2.5: Number of Passing, Zone, and Failing Programs Disaggregated by Quartiles* of Percent White

*The first quartile represents the programs with the lowest proportion of white students and the fourth quartile represents programs with the highest proportion.

Table 2.5 shows that the passing rates across all quartiles of percent white are similar, except the fourth quartile has a slightly higher passing rate.

Table 2.6: Number of Passing, Zone, and Failing Programs Disaggregated by Quartiles* of Percent Black

*The first quartile represents the programs with the lowest proportion of Black students and the fourth quartile represents programs with the highest proportion.

Table 2.6 shows that the passing rates across all quartiles of percent Black are similar, except the first quartile has a slightly higher passing rate.

Table 2.7: Number of Passing, Zone, and Failing Programs Disaggregated by Quartiles* of Percent Hispanic

*The first quartile represents the programs with the lowest proportion of Hispanic students and the fourth quartile represents programs with the highest proportion.

Table 2.7 shows that the passing rates across all quartiles of percent Hispanic are similar.

Table 2.8: Number of Passing, Zone, and Failing Programs Disaggregated by Quartiles* of Percent Pell

*The first quartile represents the programs with the lowest proportion of Pell students and the fourth quartile represents programs with the highest proportion.

Table 2.8 shows that the passing rates across all quartiles of percent Pell are similar.

Table 2.9: Number of Passing, Zone, and Failing Programs Disaggregated by Quartiles* of Percent Zero EFC

*The first quartile represents the programs with the lowest proportion of zero EFC students and the fourth quartile represents programs with the highest proportion.

Table 2.9 shows that the passing rates across all quartiles of percent zero EFC are almost the same.

Table 2.10: Number of Passing, Zone, and Failing Programs Disaggregated by Quartiles* of Percent Female

*The first quartile represents the programs with the lowest proportion of female students and the fourth quartile represents programs with the highest proportion.

Table 2.10 shows that the passing rates across all quartiles of percent female are similar.

Table 2.11: Number of Passing, Zone, and Failing Programs Disaggregated by Quartiles of Percent Independent

*The first quartile represents the programs with the lowest proportion of independent students and the fourth quartile represents programs with the highest proportion.

Table 2.11 shows that the passing rates across all quartiles of percent independent are similar, except the first quartile has a slightly lower passing rate.

***

These results suggest that the regulations do not primarily measure student demographics because indicators of many student characteristics have similar passing rates across quartiles.

Student demographics regression analysis of final regulations

As described in “Methodology for student demographics analysis of final regulations,” to further examine the relationship between student demographics and program results under the final regulations, we analyzed the degree to which individual demographic characteristics might be associated with a program’s annual earnings rate while holding other characteristics constant. This method allowed us to investigate whether there are any particular demographic characteristics that may place programs at a substantial disadvantage under the D/E rates measure.

For this analysis, the Department created a regression model with annual earnings rate as the dependent variable and multiple independent variables that are indicators of student, program, and institutional characteristics. The independent variables in the regression analysis are zero EFC, independent, female, mothers completing college, and the following race and ethnicity categories: American Indian or Alaska Native (Indian), Asian/Native Hawaiian/Other Pacific Islander (Asian), Black or African American (Black), Hispanic or Latino/Hispanic (Hispanic), White/White non-Hispanic (White), and Two or More Races.³²³ In addition, we included program credential level and institutional sector to control for non-demographic characteristics of programs. As stated previously, we ran the regression models both with and without the race and ethnicity variables.

Table 2.12: Annual earnings rate regression- with race/ethnicity variables

Number of obs	4171
R-squared	0.44

Variable	Coefficient	Robust standard error	T-statistic	P-value
Asian	-0.022	0.005	-4.09	0.000
Black	0.019	0.003	6.07	0.000
Hispanic	-0.015	0.003	-4.62	0.000
Indian	-0.002	0.012	-0.13	0.896
race2or_more	-0.110	0.038	-2.92	0.004
zefc	-0.015	0.005	-3.00	0.003
independent	-0.017	0.004	-4.20	0.000
female	0.011	0.002	5.54	0.000
mother	0.031	0.008	4.07	0.000
Private	0.220	0.397	0.55	0.579
Public	1.231	0.646	1.90	0.057
Level_2	3.400	0.325	10.46	0.000
Level_3	4.775	0.412	11.60	0.000
Level_4	2.360	0.331	7.14	0.000
Level_5	-2.833	0.310	-9.14	0.000
Level_6	-2.192	0.397	-5.51	0.000
Level_7	-1.251	0.343	-3.64	0.000
Constant	2.168	0.251	8.65	0.000
Percent white used as reference group for race, for-profit used as reference group for sector, and credential level 01 (undergraduate certificate) used as reference group for credential levels
Demographic and dependent variable units in percentages

Table 2.13: Annual earnings rate regression- without race/ethnicity variables

Number of obs	5537
R-squared	0.3868

Variable	Coefficient	Robust standard error	T-statistic	P-value
zefc	-0.017	0.004	-4.48	0.000
independent	-0.005	0.004	-1.32	0.185
female	0.012	0.002	6.74	0.000
mother	0.044	0.007	6.09	0.000
Private	-2.855	0.210	-13.56	0.000
Public	-5.410	0.094	-57.72	0.000
Level_2	3.268	0.173	18.89	0.000
Level_3	2.978	0.318	9.37	0.000
Level_4	-2.131	0.322	-6.62	0.000
Level_5	-2.519	0.454	-5.55	0.000
Level_6	-2.947	0.743	-3.97	0.000
Level_7	4.614	3.210	1.44	0.151
Constant	5.357	0.376	14.25	0.000
For-profit used as reference group for sector and credential level 01 (undergraduate certificate) used as reference group for credential levels
Demographic and dependent variable units in percentages

The results of both regressions indicate that programs with greater proportions of zero EFC graduates have slightly lower annual earnings rates; programs with greater proportions of graduates mothers who completed college have slightly higher annual earnings rates; programs with greater proportions of Black graduates have slightly higher annual earnings rates; programs with greater proportions of Hispanic graduates have slightly lower annual earnings rates; programs with greater proportion of Asian graduates have slightly lower annual earnings rates; and programs with higher proportions of female graduates have slightly higher annual earnings rates. The percent American Indian variable does not have a statistically significant relationship with annual earnings rate. When controlling for race and ethnicity, programs with higher proportions of independent graduates have slightly lower annual earnings rates. Without controlling for race and ethnicity categories, the percent independent variable is not statistically significant. While many of the demographic variables are statistically significant, the magnitude of the coefficients is sufficiently small indicating that these factors have little impact on annual earnings rates and that it would be unlikely for a program to move from passing to failing solely by virtue of enrolling more students with these characteristics.

In response to the NPRM, commenters argued that the Department should further explore the results of the regression analysis where they contradict our own prior research on the relationship between student characteristics and education outcomes. For example, one commenter asserted that a recent study commissioned by the Department demonstrated that race, gender, and income were all significant in predicting student success in the form of degree attainment. We do not believe that the regression results described in this section contradict the Department’s prior research because we have not conducted similar research on D/E rates as calculated in the regulations.

To better understand the results of the regression analysis, we provide a correlation matrix of the variables that were used.

Table 2.14: Correlation Matrix

	D/E	earnings	loan_amount	asian	black	hispanic	indian	race2or_more	white	zefc	independent	female	mother
D/E	1.000
earnings	-0.217	1.000
loan_amount	0.805	0.181	1.000
asian	-0.014	0.069	0.058	1.000
black	0.108	-0.079	0.030	-0.110	1.000
hispanic	-0.073	-0.294	-0.155	-0.033	-0.210	1.000
indian	-0.020	0.001	-0.019	-0.008	-0.070	-0.059	1.000
race2or_more	0.018	-0.013	0.024	0.064	0.032	-0.011	0.006	1.000
white	-0.004	0.285	0.096	-0.235	-0.455	-0.703	-0.012	-0.068	1.000
zefc	-0.120	-0.593	-0.372	0.013	0.244	0.578	0.006	0.021	-0.672	1.000
independent	-0.137	-0.015	-0.162	0.043	0.252	-0.030	0.055	0.041	-0.168	0.346	1.000
female	0.010	-0.268	-0.124	-0.003	0.041	-0.038	0.008	-0.004	0.005	0.244	0.289	1.000
mother	0.140	0.327	0.329	-0.005	-0.177	-0.303	0.009	-0.016	0.383	-0.593	-0.442	-0.186	1.000
*Dummy variables for sector and credential level not included, those variables were not highly correlated with demographic variables

The correlation matrix demonstrates that there is some collinearity between zero EFC and percent mothers completing college, percent white, and percent Hispanic. To determine if the collinearity between these variables impacts the results of our analysis, we ran the regressions described above but without the race and ethnicity variables and without percent mothers completing college. These regressions show results similar to those in the original regressions, suggesting the results are robust to alternative specifications.³²⁴

The correlation matrix also shows the correlation between the demographic variables and annual earnings rate and its components, annual loan payment and annual earnings. To better understand the results of the correlation matrix, particularly those that appear counterintuitive, we further examined the relationship between low-income status, using the percent zero EFC variable, and annual earnings rate. The correlation matrix shows that percent zero EFC is negatively correlated with annual earnings rate and also with both of its components, annual loan payment and annual earnings. In other words, higher percent zero EFC is correlated with lower annual loan payment, lower annual earnings, and lower annual earnings rate. These correlations suggest that zero EFC students borrow less than other students and as a result, with respect to the relationship between percent zero EFC and annual earnings rate, the annual loan payment is more influential than annual earnings since lower annual earnings rate could only be the result of lower annual loan payments and not lower annual earnings.

To further examine this explanation, we used NPSAS:2012 data to determine the average cumulative amount borrowed by undergraduate students who are Pell Grant recipients and have zero EFC status. We limited the sample to students who received title IV, HEA program funds and completed a program because those are the students whose outcomes will be assessed under the D/E rates measure. We also limited our analysis to students who attended for-profit institutions and certificate students at private and public institutions to capture students in GE programs.

Table 2.15: Cumulative amount borrowed (NPSAS:12)- Pell & zero EFC

	Cumulative amount borrowed for undergrad (AVG)	Cumulative federal loan amount for undergrad (AVG)
Total	18,912.98	16,213.40
Expected Family Contribution in 2011-12
Zero 2011-12 EFC	17,315.38	14,859.19
2011-12 EFC > 1	21,064.62	18,037.23
Cumulative Pell amount
Never had Pell	21,570.68	17,672.99
Had Pell sometime in UG career	18,488.11	15,980.06

Table 2.15 confirms that zero EFC students and Pell Grant recipients in GE program tend to borrow less. These results could mean either that low-income students borrow less than other students enrolled in the same program, or low-income students tend to enroll in programs that lead to lower debt. Programs can lead to lower debt because they are either less expensive per credit or because they are shorter in time. To test these explanations, we conducted an ordinary least squares regression using student-level data for the programs in the 2012 GE informational D/E rates data set. Because we used the 2012 informational D/E rates data, the analysis was restricted to students who received title IV, HEA program funds who completed a GE program. To control for program cost, we used program-level fixed effects. The cumulative amount that a student borrowed to attend the program was used as the dependent variable and Pell status (received or not received) at any time in the student’s academic career was used as the independent variable.

Table 2.16: Cumulative amount borrowed regression- Pell status (2012 GE informational D/E rates)

Variable	Coefficient	Standard error	T-statistic	P-value
Intercept	0.00	25.16	0.00	1.00
Pell grant recipient	1016.06	62.86	16.16	<.0001

Note: Individual coefficients for each program not shown.

The results of this regression shows that when controlling for program effects, low-income students borrow more than other students. This finding suggests that the reason programs with a higher proportion of low-income students have better annual earnings rates is because low-income students tend to choose programs that typically lead to lower debt burdens.

Conclusions of student demographic analysis of final regulations

The Department acknowledges that student characteristics can play a role in postsecondary outcomes. However, based on the regression and descriptive analyses described above, the Department cannot conclude that the D/E rates measure is unfair towards programs that graduate high percentages of students who are minorities, low-income, female, or nontraditional or that demographic characteristics are largely determinative of results. If this were the case, we would expect to observe consistent results across all types of analyses indicating positive associations between the annual earnings rate and the demographic variables and dramatic differences in the demographic profiles of passing, zone, and failing programs. Instead, we find a negative association between the proportion of low-income students and the annual earnings rate when controlling for other demographic and non-demographic factors, similar passing rates across all quartiles of low-income variables, and similar demographic profiles in passing, zone, and failing programs for almost all of the variables examined. These and other results of our analyses suggest that the regulation is not primarily measuring student demographics.

Impact Analysis of Final Regulations

This impact analysis is based on the sample of 2012 GE informational rates generated from NSLDS as described in the “Data and Methodology for Analysis of the Regulations” above. For purposes of this impact analysis, the sample of programs only includes those that meet the minimum n-size threshold of 30. Of the 37,589³²⁵ GE programs in the FY 2010 reporting with total enrollment of 3,985,329 students receiving title IV, HEA program funds, 5,539 programs, representing 2,521,283 students receiving title IV, HEA program funds, had a minimum n-size of 30 and were evaluated in the 2012 GE informational D/E rates.

Table 2.17: 2012 GE Informational D/E Rates Program Count

2-Digit CIP Code	2-Digit CIP Name	Public	Private	For-Profit	Total
51	Health Professions and Related Sciences	648	121	1,677	2,446
12	Personal and Miscellaneous Services	77	9	868	954
52	Business Management and Administrative Services	54	20	381	455
47	Mechanics and Repairs	49	11	198	258
50	Visual and Performing Arts	2	5	237	244
11	Computer and Information Sciences	6	6	206	218
43	Protective Services	83	2	115	200
15	Engineering Related Technologies	8	8	115	131
13	Education	27	26	60	113
46	Construction Trades	39	10	59	108
22	Law and Legal Services	11	8	69	88
48	Precision Production Trades	23	5	28	56
10	Telecommunications Technologies	1	1	46	48
49	Transportation and Material Moving Workers	25	2	19	46
19	Home Economics	18	5	9	32
42	Psychology	1	1	24	26
9	Communications	0	0	21	21
24	Liberal Arts and Sciences, General Studies and Humanities	9	0	5	14
44	Public Administration and Services	1	1	10	12
31	Parks, Recreation, Leisure, and Fitness Studies	1	0	11	12
23	English Language and Literature/Letters	0	5	5	10
30	Multi-interdisciplinary Studies	2	0	6	8
1	Agricultural Business and Production	2	1	3	6
45	Social Sciences and History	1	1	4	6
14	Engineering	1	1	2	4
41	Science Technologies	2	0	2	4
34	Health-related Knowledge and Skills	0	0	4	4
39	Theological Studies and Religious Vocations	0	2	0	2
54	History	0	0	2	2
4	Architecture and Related Programs	0	1	1	2
21	Technology/Education Industrial Arts	0	0	2	2
32	Basic Skills	0	0	2	2
3	Conservation and Renewable Natural Resources	0	0	1	1
26	Biological and Biomedical Sciences	0	0	1	1
40	Physical Sciences	1	0	0	1
16	Foreign Languages and Literature	1	0	0	1
25	Library Studies	0	1	0	1
Total		1,093	253	4,193	5,539

Table 2.17 illustrates the type of programs, by sector, in the 2012 GE informational D/E rates. The most common types of programs offered were Health Professions and Related Sciences programs, Personal and Miscellaneous Services programs, and Business Management and Administrative Services programs. A substantial majority (over 75 percent) of these programs are offered by for-profit institutions. This table includes all programs in the sample at all credential levels.

Table 2.18: 2012 GE Informational D/E Rates Programs As a Percentage of All Programs in FY 2010 Reporting by Two-Digit CIP Code

2-Digit CIP Code	2-Digit CIP Name	Public	Private	For-Profit	Total
51	Health Professions and Related Sciences	12.9%	17.8%	43.5%	25.6%
52	Business Management and Administrative Services	1.5%	6.8%	22.8%	8.3%
12	Personal and Miscellaneous Services	7.3%	18.0%	34.3%	26.2%
47	Mechanics and Repairs	2.2%	20.4%	56.6%	9.7%
11	Computer and Information Sciences	0.4%	6.7%	22.8%	8.2%
15	Engineering Related Technologies	0.5%	16.7%	36.7%	6.4%
50	Visual and Performing Arts	0.3%	4.0%	36.0%	17.5%
13	Education	3.9%	6.2%	23.2%	8.3%
43	Protective Services	9.4%	5.6%	29.6%	15.3%
48	Precision Production Trades	2.2%	22.7%	51.9%	5.0%
46	Construction Trades	4.1%	41.7%	46.8%	9.8%
22	Law and Legal Services	3.5%	13.6%	18.5%	11.7%
19	Home Economics	2.6%	25.0%	20.5%	4.3%
1	Agricultural Business and Production	0.4%	20.0%	33.3%	1.2%
10	Telecommunications Technologies	0.3%	20.0%	35.1%	9.3%
44	Public Administration and Services	0.5%	3.6%	23.3%	4.7%
9	Communications	0.0%	0.0%	27.6%	8.3%
49	Transportation and Material Moving Workers	14.7%	28.6%	43.2%	20.8%
31	Parks, Recreation, Leisure, and Fitness Studies	0.9%	0.0%	14.5%	6.1%
24	Liberal Arts and Sciences, General Studies and Humanities	6.9%	0.0%	10.6%	7.5%
30	Multi-interdisciplinary Studies	1.8%	0.0%	22.2%	4.4%
45	Social Sciences and History	0.8%	3.8%	15.4%	3.4%
42	Psychology	2.6%	1.7%	32.9%	15.3%
14	Engineering	1.2%	6.7%	5.9%	3.0%
16	Foreign Languages and Literature	0.9%	0.0%	0.0%	0.8%
23	English Language and Literature/Letters	0.0%	29.4%	22.7%	8.6%
39	Theological Studies and Religious Vocations	0.0%	2.3%	0.0%	1.9%
26	Biological and Biomedical Sciences	0.0%	0.0%	7.7%	1.1%
3	Conservation and Renewable Natural Resources	0.0%	0.0%	8.3%	1.2%
41	Science Technologies	2.8%	0.0%	28.6%	5.1%
4	Architecture and Related Programs	0.0%	14.3%	14.3%	3.4%
5	Area, Cultural, Ethnic, and Gender Studies	0.0%	0.0%	0.0%	0.0%
25	Library Studies	0.0%	14.3%	0.0%	2.4%
40	Physical Sciences	4.3%	0.0%	0.0%	3.2%
54	History	0.0%	0.0%	13.3%	8.0%
27	Mathematics and Statistics	0.0%	0.0%	0.0%	0.0%
38	Philosophy and Religious Studies	0.0%	0.0%	0.0%	0.0%
32	Basic Skills	0.0%	0.0%	66.7%	13.3%
34	Health-related Knowledge and Skills	0.0%	0.0%	100.0%	30.8%
36	Leisure and Recreational Activities	0.0%	0.0%	0.0%	0.0%
60	Residency Programs	0.0%	0.0%	0.0%	0.0%
28	Reserve Officer Training Corps	0.0%	0.0%	0.0%	0.0%
21	Technology/Education Industrial Arts	0.0%	0.0%	100.0%	50.0%
29	Military Technologies	0.0%	0.0%	0.0%	0.0%
33	Citizenship Activities	0.0%	0.0%	0.0%	0.0%
53	High School/Secondary Diplomas and Certificates	0.0%	0.0%	0.0%	0.0%
37	Personal Awareness and Self-Improvement	0.0%	0.0%	0.0%	0.0%

Table 2.18 illustrates the percentage of programs in the 2012 GE informational D/E rates sample out of the universe of all GE programs³²⁶ for each two-digit CIP code ordered by the frequency of programs in the universe of GE programs. The first row shows that 12.9 percent of public health professions and related science programs (out of all public health professionals and related sciences programs) are in the sample. Also in the sample are 17.8 percent of private health professional and related science programs (out of all private health professionals and related sciences programs); and 43.5 percent of the for-profit health professional and related sciences programs (out of all for-profit health professionals and related sciences programs). In addition, 25.6 percent of health professionals and related sciences programs in all sectors are in the sample (out of all health professionals and related sciences programs in all sectors).

Table 2.19: 2012 GE Informational D/E Rates FY 2010 Enrollment Count

2-Digit CIP Code	2-Digit CIP Name	Public	Private	For-profit	Total
51	Health Professions and Related Sciences	82,308	26,749	689,375	798,432
52	Business Management and Administrative Services	6,339	3,082	564,141	573,562
12	Personal and Miscellaneous Services	8,396	1,597	183,441	193,434
43	Protective Services	11,248	336	163,685	175,269
11	Computer and Information Sciences	1,291	628	140,709	142,628
13	Education	3,325	3,338	96,037	102,700
47	Mechanics and Repairs	3,747	2,154	84,164	90,065
50	Visual and Performing Arts	148	299	86,178	86,625
15	Engineering Related Technologies	656	876	74,762	76,294
30	Multi-interdisciplinary Studies	151	0	55,203	55,354
42	Psychology	275	56	46,252	46,583
44	Public Administration and Services	54	64	39,432	39,550
22	Law and Legal Services	1,682	799	26,354	28,835
46	Construction Trades	2,686	1,778	11,833	16,297
24	Liberal Arts and Sciences, General Studies and Humanities	8,342	0	7,594	15,936
10	Telecommunications Technologies	435	52	13,570	14,057
45	Social Sciences and History	0	101	10,331	10,432
19	Home Economics	7,111	699	1,684	9,494
49	Transportation and Material Moving Workers	1,312	271	7,459	9,042
48	Precision Production Trades	1,642	1,165	5,887	8,694
23	English Language and Literature/Letters	0	1,101	5,659	6,760
9	Communications	0	0	6,034	6,034
14	Engineering	45	164	4,738	4,947
31	Parks, Recreation, Leisure, and Fitness Studies	815	0	3,377	4,192
34	Health-related Knowledge and Skills	0	0	1,320	1,320
54	History	0	0	1,293	1,293
21	Technology/Education Industrial Arts	0	0	1,066	1,066
4	Architecture and Related Programs	0	37	532	569
41	Science Technologies	192	0	253	445
3	Conservation and Renewable Natural Resources	0	0	420	420
1	Agricultural Business and Production	101	94	202	397
39	Theological Studies and Religious Vocations	0	167	0	167
32	Basic Skills	0	0	131	131
25	Library Studies	0	89	0	89
26	Biological and Biomedical Sciences	0	0	71	71
16	Foreign Languages and Literature	71	0	0	71
40	Physical Sciences	28	0	0	28
Total		142,400	45,696	2,333,187	2,521,283

Table 2.19 illustrates the enrollment count by sector for the 2012 GE informational D/E rates program sample. The types of programs with the highest number of FY 2010 enrollees were Health Professions and Related Sciences programs, Business Management and Ministry of Services programs, and Personal and Miscellaneous Services programs. Over ninety percent of enrollees attended programs offered by for-profit institutions and only two percent of enrollees attended programs offered by private nonprofit institutions.

Table 2.20: 2012 GE Informational D/E Rates Enrollment as a Percentage of All Enrollment in FY 2010 Reporting by Two-Digit CIP Code

2-Digit CIP Code	2-Digit CIP Name	Public	Private	For-Profit	Total
51	Health Professions and Related Sciences	29.4%	69.5%	76.3%	65.4%
52	Business Management and Administrative Services	4.8%	50.7%	82.6%	69.9%
12	Personal and Miscellaneous Services	18.8%	50.3%	76.5%	67.2%
43	Protective Services	19.4%	33.2%	76.7%	64.4%
11	Computer and Information Sciences	3.5%	37.2%	66.8%	57.3%
13	Education	16.6%	41.4%	71.3%	63.1%
47	Mechanics and Repairs	5.6%	55.5%	89.4%	54.5%
50	Visual and Performing Arts	1.0%	18.1%	76.3%	66.8%
15	Engineering Related Technologies	2.6%	58.6%	89.5%	68.9%
30	Multi-interdisciplinary Studies	7.7%	0.0%	94.6%	91.4%
42	Psychology	15.9%	5.2%	66.8%	64.7%
44	Public Administration and Services	0.9%	16.8%	76.1%	67.8%
22	Law and Legal Services	15.5%	48.6%	50.9%	44.8%
46	Construction Trades	12.3%	89.4%	74.7%	41.1%
24	Liberal Arts and Sciences, General Studies and Humanities	57.4%	0.0%	69.5%	61.5%
10	Telecommunications Technologies	4.5%	48.6%	62.3%	44.6%
45	Social Sciences and History	0.0%	21.6%	65.6%	60.1%
19	Home Economics	14.0%	68.3%	25.5%	16.3%
49	Transportation and Material Moving Workers	31.9%	36.3%	87.4%	67.5%
48	Precision Production Trades	5.6%	85.9%	78.1%	22.9%
23	English Language and Literature/Letters	0.0%	94.7%	77.3%	39.6%
9	Communications	0.0%	0.0%	51.9%	38.8%
14	Engineering	3.3%	55.4%	84.4%	68.0%
31	Parks, Recreation, Leisure, and Fitness Studies	24.9%	0.0%	36.4%	33.0%
34	Health-related Knowledge and Skills	0.0%	0.0%	100.0%	91.0%
54	History	0.0%	0.0%	30.2%	29.9%
21	Technology/Education Industrial Arts	0.0%	0.0%	100.0%	99.4%
4	Architecture and Related Programs	0.0%	41.1%	71.4%	15.5%
41	Science Technologies	12.0%	0.0%	42.8%	20.3%
3	Conservation and Renewable Natural Resources	0.0%	0.0%	17.9%	11.5%
1	Agricultural Business and Production	1.5%	81.0%	72.1%	5.7%
39	Theological Studies and Religious Vocations	0.0%	14.6%	0.0%	10.4%
32	Basic Skills	0.0%	0.0%	35.8%	23.7%
25	Library Studies	0.0%	50.3%	0.0%	10.1%
16	Foreign Languages and Literature	2.7%	0.0%	0.0%	2.6%
26	Biological and Biomedical Sciences	0.0%	0.0%	7.9%	4.2%
40	Physical Sciences	26.9%	0.0%	0.0%	17.8%
38	Philosophy and Religious Studies	0.0%	0.0%	0.0%	0.0%
36	Leisure and Recreational Activities	0.0%	0.0%	0.0%	0.0%
5	Area, Cultural, Ethnic, and Gender Studies	0.0%	0.0%	0.0%	0.0%
28	Reserve Officer Training Corps	0.0%	0.0%	0.0%	0.0%
27	Mathematics and Statistics	0.0%	0.0%	0.0%	0.0%
29	Military Technologies	0.0%	0.0%	0.0%	0.0%
60	Residency Programs	0.0%	0.0%	0.0%	0.0%
33	Citizenship Activities	0.0%	0.0%	0.0%	0.0%
37	Personal Awareness and Self-Improvement	0.0%	0.0%	0.0%	0.0%
53	High School/Secondary Diplomas and Certificates	0.0%	0.0%	0.0%	0.0%

Table 2.20 illustrates the percentage of FY 2010 enrollees in the 2012 GE informational D/E rates sample out of the universe of all FY 2010 GE reported enrollment for each two-digit CIP code ordered by the frequency of enrollees in the universe of GE programs. The first row shows that 29.4 percent of enrollees in public health professions and related science programs (out of all enrollees in public health professionals and related sciences programs) are in the sample. Also in the sample are 69.5 percent of enrollees in private health professional and related science programs (out of all enrollees in private health professionals and related sciences programs); 76.3 percent of enrollees in for-profit health professional and related sciences programs (out of enrollees in all for-profit health professionals and related sciences programs); and 65.4 percent of enrollees in health professionals and related sciences programs in all sectors (out of all enrollees in health professionals and related sciences programs in all sectors).

Table 2.21: 2012 GE Informational D/E Rates Program Results

Sector	IHE Type	Credential Level	Programs	Passing Programs	Zone Programs	Failing Programs	Enrollment	Enrollment in Passing Programs	Enrollment in Zone Programs	Enrollment in Failing Programs
Overall Total			5,539	4,094	928	517	2,521,283	1,679,616	453,904	387,763
Public	Total		1,093	1,090	2	1	142,400	142,077	277	46
	< 2 year	Certificate	157	157	0	0	11,439	11,439	0	0
	2-3 year	Certificate	824	823	1	0	119,615	119,559	56	0
	4-year	Certificate	86	84	1	1	8,102	7,835	221	46
		Post-Bacc Certificate	26	26	0	0	3,244	3,244	0	0
Private	Total		253	242	8	3	45,696	40,695	3,886	1,115
	< 2 year	Certificate	49	47	2	0	9,609	9,147	462	0
	2-3 year	Certificate	73	70	3	0	10,307	8,875	1,432	0
		Post-Bacc Certificate	1	1	0	0	17	17	0	0
	4-year	Certificate	91	86	3	2	20,666	17,679	1,992	995
		Post-Bacc Certificate	39	38	0	1	5,097	4,977	0	120
For-Profit	Total		4,193	2,762	918	513	2,333,187	1,496,844	449,741	386,602
	< 2 year	Certificate	1,100	877	185	38	216,363	154,749	51,207	10,407
		Associate's	5	4	1	0	195	195	0	0
		1st Professional Degree	4	4	0	0	312	312	0	0
	2-3 year	Certificate	1,223	903	264	56	365,500	255,040	97,385	13,075
		Associate's	452	215	160	77	105,750	41,914	34,921	28,915
		Post-Bacc Certificate	2	2	0	0	156	156	0	0
	4-year	Certificate	267	169	70	28	84,610	47,102	30,205	7,303
		Associate's	514	183	167	164	669,030	240,135	174,977	253,918
		Bachelor's	407	208	62	137	618,330	493,257	55,897	69,176
		Post-Bacc Certificate	8	8	0	0	1,950	1,950	0	0
		Master's	171	157	4	10	226,106	222,173	1,511	2,422
		Doctoral	30	28	2	0	37,676	36,754	922	0
		1st Professional Degree	10	4	3	3	7,209	3,107	2,716	1,386
Overall Total			5,539	4,094	928	517	2,521,283	1,679,616	453,904	387,763

Table 2.21 illustrates the 2012 GE informational D/E rates program results. This analysis shows that:

• 4,094 programs (74 percent³²⁷ of programs and comprising 67 percent (1,679,616) of the total enrollees) would pass the D/E rates measure.

• 928 programs (17 percent of programs with 453,904 enrollees (18 percent)) would fall into the zone.

• 517 of programs (9 percent of programs with 387,763 enrollees (15 percent)) would fail.

Almost all programs that would fail or fall in the zone were at for-profit institutions.

Table 2.22: Average Annual Loan Payment, Earnings, and Repayment Rate, Default Rate of 2012 GE Informational Rates Sample by Sector

	Annual Loan Payment	Annual Earnings	Repayment Rate	Default Rate
Public	351	32,488	55.6	11.7
Private	768	25,398	42.9	14.5
For-Profit	1,704	27,290	41.2	22.2

Table 2.22 provides the average program annual loan payment (weighted by the number of students completing a program), the average program earnings (weighted by the number of students completing a program), the average default rate (weighted by the number of applicable borrowers), and the average repayment rate (weighted by the number of applicable borrowers) for each sector.

Table 2.23: Average Annual Loan Payment, Earnings, and Repayment Rate, Default Rate of 2012 GE Informational Rates Sample by Status

	Annual Loan Payment	Annual Earnings	Repayment Rate	Default Rate
Pass	1,284	30,714	45.0	19.5
Zone	1,723	17,498	32.9	25.2
Fail	3,622	21,016	32.3	27.6

Table 2.23 provides the average program annual loan payment (weighted by the number of students completing a program), the average program earnings (weighted by the number of students completed a program), the average default rate (weighted by the number of applicable borrowers), and the average repayment rate (weighted by the number of applicable borrowers) for passing, zone, and failing programs.

Table 2.24: 2012 GE Informational D/E Rates Passing Programs Disaggregated by Annual Earnings Rate and Discretionary Income Rate

Table 2.24 shows that 60 percent of programs that passed overall passed both the annual earnings rate and the discretionary income rate. Thirty-three percent of programs that passed the D/E rates measure overall failed the discretionary income rate and passed the annual earnings rate whereas no programs that failed the annual earnings rate passed the discretionary income rate.

Table 2.25: 2012 GE Informational D/E Rates Zone Programs Disaggregated by Annual Earnings Rate and Discretionary Income Rate

Table 2.25 shows that eighty-three percent of programs in the zone failed the discretionary income rate but were in the zone for the annual earnings rate. Only 3 percent of zone programs failed the annual earnings rate but were in the zone for the discretionary income rate.

Table 2.26: Most Frequent Types of Programs in the 2012 Informational D/E Rates

CIP	Credential level	Programs	% of all programs	t4 students	% of all t4 students
MEDICAL/CLINICAL ASSISTANT.	Certificate	347	6.3%	176,086	7.0%
BUSINESS ADMINISTRATION AND MANAGEMENT, GENERAL.	Bachelor's	41	0.7%	168,054	6.7%
COSMETOLOGY/COSMETOLOGIST, GENERAL.	Certificate	590	10.7%	115,579	4.6%
LICENSED PRACTICAL/VOCATIONAL NURSE TRAINING*	Certificate	491	8.9%	79,923	3.2%
BUSINESS ADMINISTRATION AND MANAGEMENT, GENERAL.	Master's	28	0.5%	77,571	3.1%
MEDICAL/CLINICAL ASSISTANT.	Associate's	96	1.7%	64,012	2.5%
OFFICE MANAGEMENT AND SUPERVISION.	Associate's	2	0.0%	58,489	2.3%
BUSINESS ADMINISTRATION AND MANAGEMENT, GENERAL.	Associate's	42	0.8%	51,756	2.1%
AUTOMOBILE/AUTOMOTIVE MECHANICS TECHNOLOGY/TECHNICIAN.	Certificate	47	0.8%	33,349	1.3%
MASSAGE THERAPY/THERAPEUTIC MASSAGE.	Certificate	222	4.0%	33,267	1.3%
MEDICAL OFFICE ASSISTANT/SPECIALIST.	Associate's	10	0.2%	32,951	1.3%
CRIMINAL JUSTICE/LAW ENFORCEMENT ADMINISTRATION.	Bachelor's	8	0.1%	32,284	1.3%
COMPUTER SYSTEMS NETWORKING AND TELECOMMUNICATIONS.	Associate's	16	0.3%	32,283	1.3%
HEALTH INFORMATION/MEDICAL RECORDS TECHNOLOGY/TECHNICIAN.	Associate's	6	0.1%	29,372	1.2%
CORRECTIONS AND CRIMINAL JUSTICE, OTHER.	Associate's	2	0.0%	27,743	1.1%
BEHAVIORAL SCIENCES.	Associate's	1	0.0%	27,090	1.1%
BUSINESS ADMINISTRATION, MANAGEMENT AND OPERATIONS, OTHER.	Bachelor's	5	0.1%	25,469	1.0%
CULINARY ARTS/CHEF TRAINING.	Associate's	33	0.6%	24,997	1.0%
PHARMACY TECHNICIAN/ASSISTANT.	Certificate	116	2.1%	24,945	1.0%
ALL OTHER	ALL OTHER	3,436	62.0%	1,406,063	55.8%

Table 2.26 illustrates the most frequent types of programs (by enrollment count) in the 2012 informational D/E rates sample. The most frequent types of programs are cosmetology certificate programs, nursing certificate programs, medical/clinical assistant certificate programs, and massage therapy certificates.

Table 2.27: Average Program Annual Loan Payment, Earnings, Default Rate, and Repayment Rate for Most Frequent Type of Programs (by Enrollment Count)

CIP	Credential level	Annual loan payment	Earnings	Repayment rate	Default rate
MEDICAL/CLINICAL ASSISTANT.	Certificate	$1,074	$15,309	25.1%	24.8%
BUSINESS ADMINISTRATION AND MANAGEMENT, GENERAL.	Bachelor's	$2,495	$50,012	45.0%	19.6%
COSMETOLOGY/COSMETOLOGIST, GENERAL.	Certificate	$856	$12,272	42.5%	17.2%
LICENSED PRACTICAL/VOCATIONAL NURSE TRAINING*	Certificate	$983	$33,852	44.1%	12.9%
BUSINESS ADMINISTRATION AND MANAGEMENT, GENERAL.	Master's	$2,182	$63,822	45.8%	7.0%
MEDICAL/CLINICAL ASSISTANT.	Associate's	$1,942	$19,223	23.5%	22.5%
OFFICE MANAGEMENT AND SUPERVISION.	Associate's	$2,041	$38,570	37.5%	33.9%
BUSINESS ADMINISTRATION AND MANAGEMENT, GENERAL.	Associate's	$1,811	$27,367	33.5%	27.9%
AUTOMOBILE/AUTOMOTIVE MECHANICS TECHNOLOGY/TECHNICIAN.	Certificate	$1,322	$23,603	52.0%	21.5%
MASSAGE THERAPY/THERAPEUTIC MASSAGE.	Certificate	$1,002	$16,118	41.2%	21.7%
MEDICAL OFFICE ASSISTANT/SPECIALIST.	Associate's	$2,086	$22,343	25.7%	34.6%
CRIMINAL JUSTICE/LAW ENFORCEMENT ADMINISTRATION.	Bachelor's	$3,105	$38,541	35.1%	24.8%
COMPUTER SYSTEMS NETWORKING AND TELECOMMUNICATIONS.	Associate's	$4,098	$28,872	33.8%	31.4%
HEALTH INFORMATION/MEDICAL RECORDS TECHNOLOGY/TECHNICIAN.	Associate's	$2,639	$24,392	31.0%	35.8%
CORRECTIONS AND CRIMINAL JUSTICE, OTHER.	Associate's	$2,211	$30,857	25.9%	43.9%
BEHAVIORAL SCIENCES.	Associate's	$2,485	$18,781	23.3%	38.0%
BUSINESS ADMINISTRATION, MANAGEMENT AND OPERATIONS, OTHER.	Bachelor's	$1,989	$49,629	46.1%	12.7%
CULINARY ARTS/CHEF TRAINING.	Associate's	$4,387	$22,378	38.6%	26.1%
PHARMACY TECHNICIAN/ASSISTANT.	Certificate	$983	$16,994	29.9%	21.4%
ALL OTHER	ALL OTHER	$1,651	$29,219	42.9%	20.9%

Table 2.27 provides the average program annual loan payment (weighted by the number of students completing a program), the average program earnings (weighted by the number of students completing a program), the average default rate (weighted by the number of applicable borrowers), and the average repayment rate (weighted by the number of applicable borrowers).

Table 2.28: Most Frequent Types of Zone or Failing Programs in the 2012 GE Informational D/E Rates Sample (by Enrollment Count)

CIP	Credential level	Zone/Fail Programs	% Zone/Fail programs in CIP-cred	Zone/Fail t4 students	% Zone/Fail t4 students in CIP-cred
MEDICAL/CLINICAL ASSISTANT.	Certificate	92	26.5%	59,955	34.0%
COSMETOLOGY/COSMETOLOGIST, GENERAL.	Certificate	182	30.8%	56,771	49.1%
MEDICAL/CLINICAL ASSISTANT.	Associate's	69	71.9%	56,262	87.9%
MEDICAL OFFICE ASSISTANT/SPECIALIST.	Associate's	5	50.0%	32,129	97.5%
HEALTH INFORMATION/MEDICAL RECORDS TECHNOLOGY/TECHNICIAN.	Associate's	5	83.3%	28,615	97.4%
BEHAVIORAL SCIENCES.	Associate's	1	100.0%	27,090	100.0%
COMPUTER SYSTEMS NETWORKING AND TELECOMMUNICATIONS.	Associate's	6	37.5%	26,248	81.3%
CULINARY ARTS/CHEF TRAINING.	Associate's	31	93.9%	24,446	97.8%
CRIMINAL JUSTICE/LAW ENFORCEMENT ADMINISTRATION.	Associate's	10	83.3%	24,033	98.0%
ELECTRICAL, ELECTRONIC AND COMMUNICATIONS ENGINEERING TECHNOLOGY/TECHNICIAN.	Associate's	4	44.4%	23,509	96.6%
BEHAVIORAL SCIENCES.	Bachelor's	1	50.0%	18,853	97.5%
TEACHER ASSISTANT/AIDE.	Associate's	1	100.0%	16,025	100.0%
HUMAN SERVICES, GENERAL.	Associate's	1	100.0%	15,064	100.0%
CRIMINAL JUSTICE/SAFETY STUDIES.	Associate's	18	69.2%	14,616	68.9%
CAD/CADD DRAFTING AND/OR DESIGN TECHNOLOGY/TECHNICIAN.	Associate's	5	100.0%	14,321	100.0%
SECURITIES SERVICES ADMINISTRATION/MANAGEMENT.	Associate's	13	100.0%	13,473	100.0%
MEDICAL INSURANCE CODING SPECIALIST/CODER.	Associate's	9	81.8%	13,290	97.7%
BUSINESS/COMMERCE, GENERAL.	Associate's	3	75.0%	12,550	75.2%
GRAPHIC DESIGN.	Bachelor's	19	82.6%	11,983	91.0%
ALL OTHER	ALL OTHER	970	22.4%	352,434	19.3%

Table 2.28 shows that the most frequent types of zone and failing programs in the 2012 GE informational D/E rates sample (by enrollment count) were medical/clinical assistant certificate programs, cosmetology certificate programs, and medical/clinical assistant associate degree programs.

Table 2.29: Most Frequent Types of Programs That Are Failing or in the Zone in 2012 Informational Rates (by Enrollment Count)

CIP	Credential Level	Annual loan payment			Earnings			Repayment rate			Default rate
		Pass	Zone	Fail	Pass	Zone	Fail	Pass	Zone	Fail	Pass	Zone	Fail
COSMETOLOGY/COSMETOLOGIST, GENERAL.	Certificate	975	1,247	1,697	16,189	13,467	12,900	28	20	20	23	28	24
MEDICAL/CLINICAL ASSISTANT.	Certificate	557	1,220	1,501	12,306	12,663	10,970	50	40	30	16	17	21
MEDICAL/CLINICAL ASSISTANT.	Associate's	1,193	1,925	2,603	20,805	19,689	16,961	32	25	18	19	20	28
MASSAGE THERAPY/THERAPEUTIC MASSAGE.	Certificate	1,351	2,233	2,823	20,105	22,943	17,357	31	23	21	13	36	17
BUSINESS ADMINISTRATION AND MANAGEMENT, GENERAL.	Associate's	1,187	2,366	3,531	16,845	25,703	26,664	24	37	29	29	41	27
PHARMACY TECHNICIAN/ASSISTANT.	Certificate			2,485			18,781			23			38
CULINARY ARTS/CHEF TRAINING.	Associate's	2,531	2,876	4,435	34,103	28,680	27,969	37	40	33	26	22	33
GRAPHIC DESIGN.	Associate's	1,716	2,338	4,655	25,156	22,980	22,275	20	41	39	11	22	27
CRIMINAL JUSTICE/SAFETY STUDIES.	Associate's	1,293	2,031	3,581	23,735	21,227	19,939	21	21	16	31	27	35
LEGAL ASSISTANT/PARALEGAL.	Associate's	1,652	3,810	4,496	32,229	33,746	30,320	52	35	32	18	31	40
INTERIOR DESIGN.	Bachelor's	2,128	2,657		43,331	29,449		38	30		7	25
MEDICAL ADMINISTRATIVE/EXECUTIVE ASSISTANT AND MEDICAL SECRETARY.	Certificate			2,310			14,637			26			40
BARBERING/BARBER.	Certificate		2,393			22,588			23			42
AUTOMOBILE/AUTOMOTIVE MECHANICS TECHNOLOGY/TECHNICIAN.	Certificate	1,543	2,174	2,341	25,756	22,888	17,299	27	29	15	33	27	33
MEDICAL OFFICE ASSISTANT/SPECIALIST.	Certificate		3,269	4,546		26,175	28,290		39	32		27	38
DENTAL ASSISTING/ASSISTANT.	Certificate		2,623	3,027		22,517	20,743		21	15		30	32
GRAPHIC DESIGN.	Bachelor's	1,114	2,035	3,086	21,325	23,722	19,191	16	28	15	12	19	25
HEATING, AIR CONDITIONING, VENTILATION AND REFRIGERATION MAINTENANCE TECHNOLOGY/TECHNICIAN (HAC, HACR, HVAC, HVACR).	Certificate	2,154	2,478		41,023	25,676		27	28		20	28
MEDICAL INSURANCE CODING SPECIALIST/CODER.	Certificate	2,659	3,033	3,898	34,788	27,684	26,297	56	45	42	8	19	21
All others		1,373	1,911	3,742	34,034	19,083	20,926	46	36	34	19	24	24

Table 2.29 provides the average program annual loan payment (weighted by the number of students completing a program), the average program earnings (weighted by the number of students completing a program), the average default rate (weighted by the number of applicable borrowers), and the average repayment rate (weighted by the number of applicable borrowers) for the most frequent types of programs that were failing or in the zone (by enrollment count).

Table 2.30: Program Results by Institution³²⁸ in the 2012 GE Informational Rates Sample

Table 2.30 illustrates that a large majority of institutions in the 2012 GE informational D/E rates sample have all passing programs.

Table 2.31: Concentration of Zone and Failing Programs by Institution³²⁹ in 2012 GE Informational D/E Rates Sample

Table 2.31 illustrates that most of the zone and failing programs in the 2012 GE informational D/E rates sample are concentrated in a small number of institutions.

Table 2.32: Concentration of Enrollment in Zone and Failing Programs by Institution³³⁰ in 2012 GE Informational D/E Rates Sample

Table 2.32 illustrates that most of the enrollment in zone and failing programs in the 2012 GE informational D/E rates sample are concentrated in a small number of institutions.

In response to the NPRM, analysis submitted by a commenter used data from the 2012 IPEDS files to construct a data set of 13,426 certificate programs, 9,993 associate degree programs, and 5,402 bachelor’s degree programs at for-profit institutions and identified physical locations with alternatives within the same credential level and similar CIP codes.³³¹ Programs were defined by six-digit CIP code and program length and the IPEDS unit identifier to represent a campus location. Programs that were online only were excluded from the analysis. Substitute programs were defined in a variety of ways: (1) programs at the same for-profit institution within the same credential level and a similar CIP code (four-digit and two-digit CIP codes analyzed); (2) programs at for-profit institutions within the same credential level, similar CIP code, and same five-digit zip code or three-digit zip code prefix; and (3) nearby programs in a similar CIP code at public or private not-for-profit institutions. This analysis found that 26.26 percent of students enrolled in for-profit institutions have an alternative within the same 6-digit CIP code and 5-digit zip code and, under the most expansive parameters of the analysis, that 95.78 percent of students attending for-profit institutions have at least one alternative within the same 2-digit CIP code and three-digit zip code prefix. The report provided that these results did not account for factors that might inhibit students from pursuing alternative programs including unwillingness to make even minor changes in locations or areas of study, a lack of qualifications or prerequisites to enter an alternative program, a lack of capacity in potential alternative programs, a lack of new programs to absorb students, and the possibility that accepting students with high debt amounts and high default potential would cause the receiving programs to fail the accountability metrics of the regulations. The report concluded that the Department’s estimates of students affected by the regulations who would be able to find alternative programs is overstated and, as a result, the Department underestimated the number of students who will lose access to postsecondary education as a result of the regulations.

We believe that the commenter’s analysis does not provide a useful assessment of transfer options because it evaluates transfer options for students in all programs rather than for those in zone and failing programs who will be most likely to seek alternatives as a result of their program’s performance under the regulations. Further, the commenter’s analysis did not consider as transfer options programs offered via distance education, which includes many online programs.

The Department conducted its own analysis to estimate the short-term transfer options that may be available to students in zone and failing programs (the Department assumes that in the long term, education markets will adjust and transfer options will change as student and employer demand will increase supply). Since 2012 GE informational D/E rates data are aggregated to each unique combination of the six-digit OPEID, six-digit CIP code, and credential level we do not have precise data on geographic location. For example, a GE program can have multiple branch locations in different cities and States. At some of these locations, the program could be offered as an online program. And at other locations, the program could be offered as an in-person program. But each of these locations would present as a single program in our data set without detail regarding precise location or format. To address this, the Department matched the 2012 GE informational D/E rates data with IPEDS data, which has more precise information regarding program location. As noted above, NSLDS and IPEDS have different reporting mechanisms and as a result, matching data from the two systems provides at best an approximation of the location of programs.

In order to identify geographical regions where potential transfer options may exist, we used the Core Based Statistical Area (CBSA) (or five-digit ZIP code instead if the CBSA is not applicable). For each combination of CBSA, CIP code, and credential level, we determined the number of programs available and the number of programs that would pass, fail, or fall in the zone under the D/E rates measure. For the programs not offered by distance education identified in IPEDS corresponding to the programs in the 2012 GE informational D/E rates that would not pass the D/E rates measure, we determined whether there were other programs in the same CBSA that had the same CIP and credential level and that would pass the D/E rates measure, would not be evaluated under the D/E rates measure (do not meet the n-size requirement), or is a non-GE program with an open admissions policies. We separately considered the availability of distance education programs as transfer options for students in in-person failing and zone programs in addition to in-person options. Finally, we also analyzed whether students in distance education programs that would fail or fall in the zone under the D/E rates measure would have available other distance education programs as transfer options.

Table 2.33: Transfer options for students in zone and failing programs

CIP code	Percent of enrollees in zone and failing programs not offered via distance education without transfer options (options include distance education options)	Percent of enrollees in zone and failing programs not offered via distance education without transfer options (options exclude distance education options)	Percent of enrollees in zone and failing programs offered via distance education without transfer options (options include ONLY distance education options)
Transfer options to the same 6-digit CIP	6.2%	32.3%	5.9%
Transfer options to the same 4-digit CIP	1.0%	9.8%	1.4%
Transfer options to the same 2-digit CIP	0.0%	2.1%	0.0%

Our analysis indicates that, under a static scenario assuming no reaction to the regulations, about 32 percent of students in in-person zone and failing programs will not have nearby transfer options to an in-person program with the same six-digit CIP code and credential level. This decreases to about 10 percent when in-person programs in the same four-digit CIP code are included. When online options in the same six-digit CIP code and credential level are considered, the percentage decreases from 32 percent to about 6 percent.

We recognize that there are some communities, particularly in rural areas, in which alternative programs in the same field may not be available. We also agree that students served by GE programs may have ties to a particular location that could limit their ability to pursue opportunities at physical campuses far from their home. However, we continue to believe that the substantial majority of students will find alternatives. The increased availability of online or distance programs, the chance that students will change their field or level of study in light of the data available under the regulations, and the possibility of new entrants and expanded capacity remained options for absorbing students affected by the regulations.

3. Costs, Benefits, and Transfers

Assumptions and Methodology

The budget model

To calculate the net budget impacts estimate, as in the NPRM, the Department developed a model based on assumptions regarding enrollment, program performance, student response to program performance, and average amount of title IV, HEA program funds per student to estimate the budget impact of these regulations. As discussed in more detail below, as a result of comments and, additionally, internal reconsideration, we revised the model used to create the budget estimate for the NPRM. The revised model: (1) takes into account a program’s past results under the D/E rates measure to predict future results, and (2) tracks a program’s cumulative results across multiple cycles of results under the D/E rates measure.

Budget model assumptions

We made assumptions in three areas in order to estimate the budget impact of the final regulations:

1. Program performance under the regulations;

2. Student behavior in response to program performance; and,

3. Enrollment of students in GE programs.

Program transition assumptions

Some commenters were critical of the model used by the Department to estimate the budget impact for the NPRM because it made no assumption regarding the probability that a program would transition from passing or in the zone to a second failure or ineligibility. As stated previously and described in detail below, the Department’s revised budget model accounts for this by tracking a program’s results across multiple cycles. With this capability, the revised model uses cumulative past results to predict future results.

Some commenters criticized the NPRM’s budget model on the basis that the assumptions for the probability that a program is failing did not distinguish whether the program fails due to its D/E rates or because of its pCDR. We do not address this comment here as the revised budget model for the final regulations makes no assumptions regarding pCDR results because the measure is not included as an accountability metric in the final regulations.

As in the NPRM, given a program’s status under the D/E rates measure in any year--passing, in the zone, failing, ineligible, or not evaluated because the program did not meet the minimum n-size requirements--we developed assumptions for the likelihood that the program’s performance would place it in each of the same five categories in the subsequent year:

1. Passing;

2. In the zone;

3. Failing;

4. Ineligible (a program could become ineligible in one of two ways: (1) by failing the D/E rates measure for two out of three consecutive years, or (2) by not achieving a passing status in four consecutive years); or,

5. Not evaluated because the program failed to meet the minimum n-size requirements for the D/E rates measure.

The budget model applies assumptions for three transitions between program results (year 0 to 1 to 2 to 3). It assumes that after year 3, which marks the beginning of the fourth transition in results, the rates of program transition will reach a steady state.

The program assumptions track results through each cycle of the model. Stated differently, results do not reset after each cycle. Rather, past results impact future results. For example, a program that falls in the zone in year 0 and passes in year 1 would not simply be considered a passing program. Its zone result in year 0 would continue to influence the probabilities of its year 2 results. If a program’s performance reaches ineligible status (2 fails in 3 years or no passes in 4 years), the program becomes, and remains, ineligible for all future years. The model assigns probabilities for all potential combinations of results for each transition.

Year 0 to year 1 program transition assumptions

The assumptions for the year 0 to year 1 transition in program results (ex: the probability that a program is in the zone in year 0 and passing in year 1) is the observed comparison of actual D/E informational rates results for two consecutive cohorts of students in the GE Data. As in the NPRM, the initial assignment of performance categories in year 0 is based on the 2012 GE informational D/E rates data for students who completed GE programs in fiscal years 2008 and 2009. The program transition assumption for year 0 to year 1 are based on the outcomes of students who completed GE programs in fiscal years 2007 and 2008, and the outcomes of students who completed GE programs in fiscal years 2008 and 2009. For the observed results that are the basis for the year 0 to year 1 program transition assumption, we applied a minimum n-size of 10, instead of 30 as is required under the final regulations and used in the “Analysis of the Regulations” section of this RIA, for the D/E rates calculations to maximize the number of observations in the two-year comparative analysis used to create the program transition assumptions. Program results under the D/E rates measure for the 2007/2008 cohort of students who completed the program were calculated using the same methodology used to calculate the 2012 GE informational D/E rates except that, as with the 2008/2009 cohort, a minimum n-size of 10 was applied. It is important to note that the results in the “Analysis of the Regulations” section in this RIA are based on a minimum n-size of 30 for the D/E rates measure as is required under the regulations but the budget model for the “Discussion of Costs, Benefits and Transfers” and the “Net Budget Impact” sections used a minimum n-size of 15 for the D/E rates measure. This was done to simulate the effect of the four-year cohort period “look back” provisions of the regulations so that the net budget impact would not be underestimated as a result of treating programs that will likely be evaluated under the regulations as not having a result in the budget model. Only the results of programs with students who completed the programs in FY 2008 were compared because these programs would have results for both cohorts.

The observed year 0 to year 1 results also serve as the baseline for each subsequent transition of results (year 1 to year 2, etc.). As described below, the model applies additional assumptions from that baseline for each transition beginning with year 1 to year 2.

Table 3.1: Year 0 to Year 1 (Observed) Program Transition Assumptions

Year 1 Result	Year 2 Result	% w/ Year 2 Result
Pass	Pass	81.7%
	Zone	4.0%
	Fail	0.6%
	Not Evaluated	13.7%
Zone	Pass	22.0%
	Zone	39.8%
	Fail	24.9%
	Not Evaluated	13.4%
Fail	Pass	3.9%
	Zone	20.3%
	Fail	64.7%
	Not Evaluated	11.2%
Not Evaluated	Pass	6.3%
	Zone	0.8%
	Fail	0.6%
	Not Evaluated	92.3%

Because the year 0 and year 1 assumptions are the actual observed results of programs based on a cohort of students that completed programs prior to the Department’s GE rulemaking efforts, the year 0 and year 1 assumptions do not account for changes that institutions have made to their programs in response to the Department’s regulatory actions or will make after the final regulations are published.

Year 1 to year 2 program transition assumptions

After the year 0 to year 1 transition, the model assumes that institutions will take at least some steps to improve program performance during the transition period by, beginning with the year 1 to year 2 transition, increasing the baseline observed probability for all combinations with a passing result in year 2 by five percentage points. Because the total probabilities for each subsequent year result for any single prior year result cannot exceed 100 percent, the 5 percentage point year 2 “improvement increase” in the probability of passing is offset by a three percentage point zone probability decrease and two percentage point fail probability decrease.

We also assumed that programs with recent passing results would have a greater chance of future passing results, and programs with recent failing results would likewise be more likely to fail in the future. A zone result in year 0 or 1 was considered to have a neutral effect on future results. For each passing result a program had in years 0 and 1, we increased the proportion of passing programs in year 2 for all combinations of year 0-year 1 results by five percentage points. Each 5 percentage point year 2 “momentum increase” in the probability of passing is offset by a three percentage point zone probability decrease and two percentage point fail probability decrease. Similarly, for each failing result a program had in years 0 and 1, we decreased the proportion of passing programs in year 2 for all combinations of year 0-year 1 results by five percentage points. Each 5 percentage point year 2 “momentum decrease” in the probability of passing is offset by a two percentage point zone probability increase and three percentage point fail probability increase.

To demonstrate the effect of the year 1 to year 2 transition assumptions, we provide as an example the probability of each of a program’s possible results in year 2 if it was in the zone in year 0 and passing in year 1. For the year 1 to year 2 pass-pass transition probability, a 5 percent improvement increase and a 5 percent momentum increase due to the year 1 pass result are added to the baseline observed 81.5 percent pass-pass probability, resulting in an assumed probability of 91.5 percent that a program is passing in year 2 after it was in the zone in year 0 and passing in year 1. In most cases, the 10 percentage point year 2 pass probability increase would be offset in the model by a 6 percentage point year 2 zone probability decrease (3 percentage points for each 5 percentage point increase) and a 4 percentage point year 2 fail probability decrease (2 percentage points for each 5 percentage point increase) from the baseline observed pass-zone and pass-fail probabilities respectively. In this case, the baseline observed probabilities are decreased from 4 percent to 0 percent for pass-zone and 1 percent to 0 percent for pass-fail. Because the baseline observed pass-ineligible probability is already 0 percent, the remaining 5 percent offset amount is taken from the baseline observed pass-not evaluated probability, reducing it from 13.5 percent to 8.5 percent. To summarize, for a program that is in the zone in year 0 and passing in year 1, the probabilities of the program’s year 2 results are as follows: pass, 91.5 percent (81.5 + 5 + 5); zone, 0 percent (4 – 4); fail 0 percent (1 – 1); not evaluated, 8.5 percent (13.5 – 5); ineligible, 0 percent.

Table 3.2: Year 1 to Year 2 Program Transition Assumptions

Prior Years Results				Pass	Zone	Fail	NE	Ineligible
YR0 Pass
	Yr1 Pass			96.5	0	0	3.5	0
	Y1 Zone			32	34	21	13	0
	YR1 Fail			9	17	0	11	63
	YR1 Not Evaluated			16	0	0	84	0
	YR1 Ineligible			0	0	0	0	100
YR0 Zone
		Yr1 Pass		91.5	0	0	8.5	0
		Y1 Zone		27	37	23	13	0
		YR1 Fail		4	20	0	11	65
		YR1 Not Evaluated		11	0	0	89	0
		YR1 Ineligible		0	0	0	0	100
YR0 Fail
		Yr1 Pass		86.5	1	0	12.5	0
		Y1 Zone		22	40	3	13	22
		YR1 Fail		0	0	0	0	100
		YR1 Not Evaluated		6	1	0	92	1
		YR1 Ineligible		0	0	0	0	100
YR0 Not Evaluated
			Yr1 Pass	91.5	0	0	8.5	0
			Y1 Zone	27	37	23	13	0
			YR1 Fail	4	20	0	11	65
			YR1 Not Evaluated	11	0	0	89	0
			YR1 Ineligible	0	0	0	0	100

Program transition assumptions for year 3 and after

Beginning with year 3, the budget model assumes a program falls into one of six categories based upon the program’s past performance and then, for each of these categories, assumes a probability for each possible result the program could have in the subsequent year (pass, zone, fail, not evaluated, or ineligible). The six performance categories are as follows:

• High Performing: Programs that have zero probability of failure in the following year. These programs have no recent zone or failing results.

• Improving: Programs with a most recent result that is better than the prior year’s result.

• Declining: Programs with multiple zone results in previous years or programs with a most recent result that is worse than the prior year’s result.

• Facing Ineligibility: Programs that could become ineligible the following year. Any program with a failing result in the most recent year is in this category, along with any program that has only zone or failing results in the previous three years.

• Ineligible: Programs that have already become ineligible.

• Not Evaluated: Programs with an n-size under 15.

As with the year 0 to year 2 assumptions, for each performance category, the probability of a program’s result in the following year is based on the baseline observed results provided in Table 3.1. Also like the year 0 to year 2 assumptions, the model assumes ongoing improvement by increasing the baseline observed probability for all combinations with a passing result in the following year by five percentage points.

The probability that a high performing program will pass the following year is the baseline observed probability of pass-pass increased by 10 percentage points and additionally by the 5 percentage point improvement increase. The probability that an improving program will pass the following year is the baseline observed probability of zone-pass increased by 10 percentage points and additionally by the 5 percentage point improvement increase. The probability that a declining program will pass the following year is the baseline observed probability of zone-pass decreased by 10 percentage points and offset by the 5 percentage point improvement increase. The probability that a program facing ineligibility will pass the following year is the baseline observed probability of fail-pass decreased by 10 percentage points and offset by the 5 percentage point improvement increase. The probability that an ineligible program will pass in the following year is of course zero. The probability that a not evaluated program will pass the following year was only adjusted for the 5 percentage point improvement increase. Where a program’s subsequent year’s pass probability was increased or decreased, the model offsets the adjustment by increasing or decreasing the corresponding zone and fail probabilities from the baseline observed probabilities in the same amounts applied to the year 1 to year 2 transition probabilities.

To demonstrate the effect of the year 3 and after transition assumptions, we provide as an example the probability of each of a high performing program’s possible results for the following year. For the probability that a high performing program will pass the following year, a 5 percent improvement increase and a 10 percent momentum increase are added to the baseline observed 81.5 percent pass-pass probability, resulting in an assumed probability of 96.5 percent. The probability that this program would fall in the zone, fail, not be evaluated, or become ineligible the following year is determined by apportioning the 15 percentage point pass offset to the baseline observed probabilities that the program would fall in the zone, fail, or not be evaluated after passing the previous year. The zone probability is reduced from 4 percent to 0 percent, the fail probability from 1 percent to 0 percent, and the not evaluated probability from 13.5 percent to 3.5 percent.

Table 3.3: Post Year 3 Program Transition Assumptions

	Pass	zone	Fail	Ne	Ineligible
Prior Year Group:
Good	96.5	0	0	3.5	0
Improving	37	31	19	13	0
Poor/Declining	17	42	28	13	0
Fail Next	0	21	0	11	68
Ineligible	0	0	0	0	100
NE	11	0	0	89	0

Student response assumptions

In the NPRM, the Department provided two primary budget impact estimates, one based on a “low” student response to program performance and the other based on a “high” student response to program performance. For clarity, we provide for the final regulations a single primary budget impact estimate based on a single set of student response assumptions and have reserved all alternate impact scenarios for the “Sensitivity Analysis” section of this RIA.

As in the NPRM, the budget model applies assumptions for the probability that a student will transfer, remain in a program, or drop out of a program in reaction to the program’s performance--passing, in the zone, failing, ineligible, or not evaluated. The model assumes that student response will increase as a program gets closer to ineligibility. The budget model assumptions regarding student responses to program results are provided in Table 3.4. These assumptions are based on our best judgment and consideration of comments. Coupled with the scenarios presented in the “Sensitivity Analysis,” these assumptions are intended to provide a reasonable estimation of the range of impact that the regulations could have on the budget.

Table 3.4: Student response assumptions

	Student Response
Program Result	Stay	Transfer	Drop
First Zone	0.80	0.15	0.05
Second Zone	0.60	0.30	0.10
Third Zone	0.40	0.45	0.15
Fourth Zone	0.20	0.60	0.20
First Failure	0.40	0.45	0.15
Second Failure	0.20	0.60	0.20
Ineligibility	0.20	0.60	0.20
Passing or Not Evaluated	1.00	0.00	0.00

In comparison to the NPRM, the budget model for the final regulations assumes different levels of student response for each number of years that a program is in the zone. This adjustment is consistent with the modifications to the program performance assumptions to account for cumulative past program results. We made other adjustments to the student response assumptions for greater simplicity and clarity, such as increasing or decreasing in equal amounts the proportion of students that are assumed to stay, transfer, and drop out for each result that brings a program closer to ineligibility. We continue to assume that a high proportion of students in poorly performing programs will transfer as a large majority of programs will meet the standards of the regulations and students will have access to information that will help them identify programs that lead to good outcomes, and, as our analysis shows, most students will have transfer options within geographic proximity or will be able to enroll in online programs. Further, as stated previously, we believe that institutions with programs that perform well under the regulations will grow existing programs and offer new ones.

In the revised model, the assumptions for student responses are always applied to the estimated enrollment in each program determined by the enrollment growth assumptions. While we expect that the disclosure of poor program performance to students, along with institutional reactions to a program’s performance under the D/E rates measure, could result in reduced enrollment in poor-performing programs, we are applying the student response assumptions to the baseline enrollment to demonstrate the maximum impact of the regulations for the scenario presented.

Enrollment growth rate assumptions

For FYs 2016 to 2024, the budget model assumes a yearly rate of growth or decline in enrollment of students receiving title IV, HEA program funds in GE programs. The loan volume projections in the Department’s FY 2015 President’s Budget (PB) are used as a proxy for the rate of change in enrollment.

To estimate the rate of change in enrollment for programs at public and private non-profit institutions, we used the projected growth rates in loan volumes for 2-year or less than 2-year public and non-profit institutions because almost all GE programs in these sectors are offered by such institutions. With respect to programs at for-profit institutions, we applied the projected loan volume growth rates for 2-year or less than 2-year for-profit institutions and 4-year private for-profit institutions, depending on the credential level of the program.

The Department used actual loan volume data through September 2013 for the growth rate estimates for FYs 2011 through 2013. The growth rate estimates for FY 2014 and subsequent years are the projected loan volume growth rates from the FY 2015 PB. For subsequent years, we assumed a reversion to long-run historical trends in loan growth for our enrollment assumption.

Table 3.5: Enrollment Growth Rate Assumptions

Sector	2010-16	2017	2018	2019	2020	2021	2022	2023
Public and Private Nonprofit	-6.00%	3%	3%	3%	3%	3%	3%	3%
For-Profit 4-year	-33%	2%	2%	2%	2%	2%	2%	2%
For-Profit 2-year or less	-32%	3%	3%	3%	3%	3%	3%	3%

Some commenters argued that the budget model in the NPRM underestimated the enrollment growth rate for the for-profit sector. In their analysis, these commenters used the average annual growth rate of enrollment at for-profit institutions over the past twenty years to estimate future enrollment. One commenter presented three student response scenarios using this enrollment growth rate assumption.³³² In the first scenario, the commenter assumed that 100 percent of students in a program that is made ineligible would not continue their education at an eligible program; in the second, 50 percent of students would continue; and, in the third, 25 percent of students would continue.  In the 50 percent scenario, the analysis estimated between one and two million fewer students would access postsecondary education by 2020 and four million over a decade. The commenters’ analysis of the 50 percent scenario estimated that by 2020, 736,000 to 1.25 million fewer female students, 268,000 to 430,000 fewer African-American students, and 199,000 to 360,000 fewer Hispanic students would continue their postsecondary education.  In the 25 percent and 100 percent scenarios, the analysis estimated that three million to 5.7 million and 3.9 million to 7.5 million fewer students, respectively, would access postsecondary education by 2024. 

We do not agree with the assertion that future enrollment patterns at for-profit institutions will be similar to enrollment over the past twenty years.  Total fall enrollment in for-profit institutions participating in the title IV, HEA programs increased from 546,053 students in 1995 to 2,175,031 students in 2012, down from a peak of approximately 2.43 million in 2010.³³³ Between 1995 and 2012, the average rate of enrollment growth at for-profit institutions that participate in the title IV, HEA programs was approximately 8.84 percent.³³⁴ There is no evidence to suggest that enrollment at for-profit institutions will continue to grow at this rate, particularly in light of the recent decline in enrollment. The Department’s estimate takes this more recent data into account and predicts a significant decline in loan volume, and accordingly enrollment, between FYs 2010 and 2016. After FY 2016, the Department predicts a 3 percent growth in loan volume, and enrollment, for all types of institutions in all sectors except four-year for-profit institutions, which we estimate to grow at a rate of 2 percent annually. We continue to believe that the PB loan volume projections used in the NPRM are reasonable and we have again adopted them for the purpose of estimating enrollment in this analysis.

Methodology for net budget impact

The budget model estimates a yearly enrollment of students in GE programs for FYs 2016 to 2024 and the distribution of those students in programs by result (pass, zone, fail, not evaluated, ineligible). The net budget impact for each year is calculated by applying assumptions regarding the average amount of title IV, HEA program funds received to this distribution of students and programs.

To establish initial program performance results (passing, zone, failing, ineligible, and not evaluated) for FY 2016, we calculated program results under the D/E rates measure using the same methodology used to calculate the 2012 GE informational D/E rates except that a minimum n-size of 15 was applied to simulate the impact of the applicable four-year cohort period “look back” provisions of the regulations. Because the final regulations apply a four-year applicable cohort period for programs that do not have 30 or more students who completed the program over a two-year cohort period, the budget estimate is based on a minimum n-size of 15 because we assume programs with 15 students who completed the program over two years would have 30 students who completed the program over four years, making them subject to the regulations.

The yearly enrollment for each GE program is determined by using the actual enrollment of students in GE programs in FY 2010, as reported by institutions in the GE Data, as a starting point. Each subsequent year’s enrollment in these programs, including for FYs 2016 to 2024, is estimated by applying the yearly enrollment growth rate assumptions provided in Table 3.5 to each program’s FY 2010 enrollment.

Table 3.6 provides the estimated initial 2016 distribution of programs and enrollment by program result prior to any program transition or student response.

Table 3.6: Estimated Initial 2016 Distribution of Programs and Enrollment by Program Result

Performance Category	Programs	Enrollment
Passing	6,710	1,520,101
Zone	1,205	380,946
Failing	713	321,269
Ineligible	-	-
Not Evaluated	28,475	1,015,654
Total	37,103	3,237,970
Note: Model limited to programs that had enrollment in FY2010 GE reporting

To this initial distribution of programs and students, the budget model applies the student response assumptions in Table 3.4 to estimate the number of students who will transfer to another program, drop-out, or remain in their program in reaction to the initial program results. The model then applies the program transition assumptions to the initial program results to create a new distribution of programs by result. The model repeats this process for each fiscal year through 2024.

This process produces a yearly estimate for the number of students receiving title IV, HEA program funds who will choose to (1) enroll in a better-performing program; (2) remain in a zone, failing, or ineligible program; or (3) drop out of postsecondary education altogether after their program receives a zone or failing result or becomes ineligible. An estimated net savings for the title IV, HEA programs results from students who drop out of postsecondary education in the year after their program receives D/E rates that are in the zone or failing or who remain at a program that becomes ineligible for title IV, HEA program funds. We assume no budget impact on the title IV, HEA programs from students who transfer from programs that are failing or in the zone to better-performing programs as the students’ eligibility for title IV, HEA program funds carries with them across programs.

To estimate the yearly Pell Grant and loan volume that would be removed from the system based on the primary budget assumptions, we multiply the number of students who leave postsecondary education or who remain in ineligible programs by the average Pell grant amount and average loan amount for each type of title IV, HEA program loan per student by sector and credential level as reported in NPSAS:2012. Consistent with the requirements of the Credit Reform Act of 1990, budget cost estimates for the title IV, HEA programs also reflect the estimated net present value of all future non-administrative Federal costs associated with a cohort of loans. To determine the estimated impact from reduced loan volume, the yearly loan volumes are multiplied by the PB 2015 subsidy rates for the relevant loan type.

Methodology for costs, benefits, and transfers

The estimated number of students who transfer, dropout, or stay in ineligible programs based on the student response assumption is used to quantify the costs and transfers resulting from the final regulations for each year from 2017 to 2024. We quantify a transfer of title IV, HEA program funds from programs that lose students to programs that gain students. We also quantify the transfer of instructional expenses as students shift programs as well as the cost associated with additional instructional expenses to educate students who transfer.

In this analysis, student transfers could result from students who enrolled in one set of programs and switch to other programs or prospective students who choose to enroll in a program other than the one they would have chosen in the absence of the regulations.

To calculate the amounts of student aid that could transfer with students each year, we multiply the estimated number of students receiving title IV, HEA program funds transferring from ineligible, failing, or zone programs each year by the average Pell Grant, Stafford subsidized loan, unsubsidized loan, PLUS loan, and GRAD PLUS loan per student as reported in NPSAS:2012. To annualize the amount of title IV, HEA program fund transfers from 2016 to 2024, we calculate the net present value (NPV) of the yearly transfers using a discount rate of 3 percent and a discount rate of 7 percent.³³⁵

To calculate the transfer of instructional expenses, we apply the $4,529 average 2-year for-profit instructional expense per enrollee for award year 2010-2011 from IPEDS to the estimated number of annual student transfers for 2017 to 2024. To determine the additional cost of educating transferring students, we used the instructional expense per enrollee data from IPEDS to calculate the average instructional expense per enrollee of passing, zone, and failing programs in the 2012 GE informational D/E rates. As determined by this calculation, we apply a difference of $1,405 for students who transfer from failing to passing programs and $1,287 for those who transfer from zone to passing programs to the estimated number of students who will transfer between FYs 2017 and 2024.

Discussion of Costs, Benefits, and Transfers

We have considered the primary costs, benefits, and transfers of the transparency framework and accountability framework for the following groups or entities that will be affected by the final regulations:

• Students

• Institutions and State and local government

• Federal government

We discuss first the anticipated benefits of the regulations, including improved market information. We then assess the expected costs and transfers for students, institutions, the Federal government, and State and local governments.

Benefits

We expect the potential primary benefits of the regulations to be: (1) improved and standardized market information about GE programs that will increase the transparency of student outcomes for better decision making by students, prospective students, and their families, the public, taxpayers, and the Government, and institutions, leading to a more competitive marketplace that encourages improvement; (2) improvement in the quality of programs, reduction in costs and student debt, and increased earnings; (3) elimination of poor performing programs; (4) better return on educational investment for students, prospective students, and their families, as well as for taxpayers and the Federal Government; (5) greater availability of programs that provide training in occupational fields with many well-paying jobs; and (6) for institutions with high-performing programs, potential growth in enrollments and revenues resulting from the additional market information that will permit those institutions to demonstrate to consumers the value of their GE programs.

Improved Market Information

The regulations will provide a standardized process and format for students, prospective students, and their families to obtain information about the outcomes of students who enroll in GE programs such as cost, debt, earnings, completion, and repayment outcomes. This information will result in more educated decisions based on reliable information about a program’s outcomes. Students, prospective students, and their families will have extensive, comparable, and reliable information to assist them in choosing programs where they believe they are most likely to complete their education and achieve the earnings they desire, while having debt that is manageable.

The improved information that will be available as a result of the regulations will also benefit institutions. Information about student outcomes will provide a clear indication to institutions about whether their students are achieving positive results. This information will help institutions determine whether it would be prudent to expand programs or whether certain programs should be improved, by increasing quality and reducing costs, or eliminated. Institutions may also use this information to offer new programs in fields where students are experiencing positive outcomes, including higher earnings and steady employment. Additionally, institutions will be able to identify and learn from programs that produce exceptional results for students.

The taxpayers and the Government will also benefit from improved information about GE programs. As the funders and stewards of the title IV, HEA programs, these parties have an interest in knowing whether title IV, HEA program funds are benefiting students. The information provided in the disclosures will allow for more effective monitoring of the Federal investment in GE programs.

The Department received many comments about the utility and scope of the disclosures, as well as about the burden associated with the disclosure and related reporting obligations. These comments are addressed in §§ 668.411 and 668.412 of the preamble and in the PRA.

Benefits to Students

Students will benefit from lower costs, and as a result, lower debt, and better program quality as institutions improve programs that fail or fall in the zone under the D/E rates measure. Efforts to improve programs by offering better student services, working with employers to ensure graduates have needed skills, increasing academic quality, and helping students with career planning will lead to better outcomes and higher earnings over time. Students will also benefit by transferring to passing programs, increasing the availability of successful programs providing high-quality training at lower costs, and from the availability of new programs in fields where there are more jobs and greater earnings. Students who graduate with manageable debts and adequate earnings will be more likely to pay back their loans, marry, form families, purchase a car, buy a home, start or invest in a business, and save for retirement.

Benefits to Institutions and State and local governments

For institutions, the impact of the regulations will likely be mixed. Institutions with programs that do not pass the D/E rates measure, including programs that lose eligibility, are likely to see lower revenues and possibly reduced profit margins. On the other hand, institutions with high-performing programs are likely to see growing enrollment and revenue and to benefit from additional market information that permits institutions to demonstrate the value of their programs.

Although low-performing programs may experience a drop in enrollment and revenues, we believe disclosures will increase enrollment and revenues in well-performing programs. Improved information from disclosures will increase market demand for programs performing well in areas such as completion, debt, earnings after completion, and repayment rates. We also believe these increases in revenue will offset any additional costs incurred and revenues lost by institutions as they improve the quality of their programs and lower their tuition prices in response to the regulations in order to ensure the long-term viability of their programs. While the increases or decreases in revenues for institutions are costs or benefits from the institutional perspective, they are transfers from a social perspective. The additional demand for education due to program quality improvement may be considered a social benefit.

State and local governments will benefit from improved oversight of their investments in postsecondary education. Additionally, State and local postsecondary education funding will be allocated more efficiently to higher-performing programs

Benefits to the Federal Government

A primary benefit of the regulations will be improved oversight and administration of the title IV, HEA programs. Additionally, Federal taxpayer funds will be allocated more efficiently to higher-performing programs, where students are more likely to graduate with manageable amounts of debt and gain stable employment in a well-paying field, increasing the positive benefits of Federal investment in title IV, HEA programs. Students will also be more likely to repay their loans, which will lower the cost of loans subsidized by the Federal Government.

Costs

Costs to students

Students may incur some costs as a result of the regulations. We expect that over the long term, all students will have increased access to programs that lead to successful outcomes. In the short term, although we believe that many students in failing and zone programs will be able to transfer to passing programs, new programs, or non-GE programs that provide equivalent training, at least some students may be temporarily left without transfer options. We expect that many of these students will re-enter postsecondary education later, but understand that some students may not continue.

Costs to Institutions and State and local governments

As the regulations are implemented, institutions will incur costs as they make changes needed to comply with the regulations, including costs associated with the reporting and disclosure requirements. These costs could include: (1) training of staff for additional duties, (2) potential hiring of new employees, (3) purchase of new software or equipment, and (4) procurement of external services. This additional burden is discussed in more detail under Paperwork Reduction Act of 1995.

Institutions that make efforts to improve the outcomes of failing and zone programs will face additional costs. For example, institutions that reduce the tuition and fees of programs will see decreased revenue. An institution could also choose to spend more on curriculum development to for example, link a program’s content to the needs of in-demand and well-paying jobs in the workforce, or allocate more funds toward other functions, such as hiring better faculty; providing training to existing faculty; offering tutoring or other support services to assist struggling students; providing career counseling to help students find jobs; or other areas where increased investment could yield improved performance on the D/E rates measure.

The costs of program changes in response to the regulations are difficult to quantify generally as they would vary significantly by institution and ultimately depend on institutional behavior. For example, institutions with all passing programs could elect to commit only minimal resources toward improving outcomes. On the other hand, they could instead make substantial investments to expand passing programs and meet increased demand from prospective students, which could result in an attendant increase in enrollment costs. Institutions with failing or zone programs could decide to devote significant resources towards improving performance, depending on their capacity, or could instead elect to discontinue one or more of the programs.

Many commenters argued that the types of investments and activities described by the Department here and in the NPRM that would improve program outcomes are not likely to affect program performance in the near term, so institutions would have to incur such costs in the expectation that program improvement would be reflected in future D/E rates. These comments are addressed in “§668.404 Calculating D/E rates” of the preamble.

State and local governments may experience increased costs as enrollment in public institutions increases as a result of some students transferring from programs at for-profit institutions. Several commenters argued that it costs taxpayers more to educate students at public institutions. These commenters relied on analysis³³⁶ that examined direct costs and calculated that at for-profit 2-year institutions produce graduates at a cost to taxpayers that is $25,546 lower on a per-student basis than the public 2-year institutions. ³³⁷ Another study estimated that public institutions receive $19.38 per student in direct tax support and private non-profit institutions receive $8.69 per student for every $1 dollar received by for-profit institutions,³³⁸ while another found that taxpayer costs of 4-year public institutions averaged $9,709 per student compared to $99 per student at for-profit institutions.³³⁹ Focusing on State and local support only, updated data from the Digest of Education Statistics indicates that State and local government grants, contracts, and appropriations per full-time equivalent student in 2011-12 to 2-year public institutions (constant 2012-13 dollars) totaled $6,280 compared to $91 to 2-year for-profit institutions.³⁴⁰

Another study cited by commenters found that if the number of graduates from nine for-profit institutions in four states, California, New York, Ohio, and Texas, in the five-year period from AYs 2007-08 to 2011-12 transferred to public 2-year or 4-year institutions, it would have cost those States an additional $6.4 billion for bachelor’s graduates and $4.6 billion for associate graduates (constant 2013 dollars).³⁴¹ The analysis submitted by commenters does not reflect the expected effect of the regulations as the majority of programs, even at for-profit institutions, are expected to pass the D/E rates measure and many students who switch programs are expected to do so within the for-profit sector, substantially reducing the impact on State and Local governments estimated in the studies cited by commenters. The Department recognizes that a shift in students to public institutions could result in higher State and Local government costs, but the extent of this is dependent on student transfer patterns and State and local government choices.

Further, if States choose to expand the enrollment capacity of passing programs at public institutions, it is not necessarily the case that they will face marginal costs that are similar to their average cost or that they will only choose to expand through traditional brick-and-mortar institutions. The Department continues to find that many States across the country are experimenting with innovative models that use different methods of instruction and content delivery, including online offerings, that allow students to complete courses faster and at lower cost. Forecasting the extent to which future growth would occur in traditional settings versus online education or some other model is outside the scope of this analysis.

Transfers

As students drop out of postsecondary education or remain in programs that lose eligibility for title IV, HEA Federal student aid, there will be a transfer of Federal student aid from those students to the Federal Government. Under the primary budget scenario, the annualized amount of this transfer of title IV, HEA programs funds over the FY 2014 to FY 2024 budget window is $423 million.

Additionally, as students change programs based on program performance and disclosures, revenues and expenses associated with students will transfer between postsecondary institutions. We estimate that approximately $2.55 billion (7 percent discount rate) or $2.52 billion (3 percent discount rate) in title IV, HEA Pell Grant and loan volume will transfer from zone, failing, and ineligible programs to passing programs on an annualized basis. These amounts reflect the anticipated high level of initial transfers as institutions adapt to the proposed regulations and failing and zone programs eventually lose eligibility for title IV, HEA program funds. We expect the title IV, HEA program funds associated with student transfers related to the final regulations to decline in future years. Additionally, we estimate that $1.24 billion (7 percent discount rate) or $1.22 billion (3 percent discount rate) in instructional expenses will transfer among postsecondary institutions.

Net Budget Impacts

As previously discussed, the Department made several assumptions about program transition, student response to program performance and enrollment growth in order to estimate the net budget impact of the regulations. The vast majority of students are assumed to resume their education at the same or another program in the event the program they are attending voluntarily closes, fails or falls in the zone under the D/E rates measure, or loses eligibility to participate in the title IV, HEA programs and the Department estimates no significant net budget impact from those students who continue their education. The student response scenarios presented in this RIA also assume that some students will not pursue, or continue to pursue, postsecondary education if warned about poor program performance or if their program loses eligibility, while other students will remain in an ineligible program that remains operational even though they will be unable to receive title IV, HEA program funds. The estimated potential net impact on the Federal budget results from Federal loans and Pell Grants not taken by these students.

As provide in Table 3.7, we estimate, under the primary student and program response scenario, that the regulations will result in reduced costs of $4.3 billion due to Pell Grants not taken between fiscal years 2014 and 2024. The estimated reductions in Pell Grant costs will be slightly offset by approximately $695 million in reduced net returns associated with lower Federal Direct Unsubsidized and PLUS loan volume. Accordingly, we estimate the net budget impact of the regulations will be $4.2 billion over the FY 2014 to FY 2024 budget window.

Table 3.7: Primary Budget Estimate

			2016	2017	2018	2019	2020	2021	2022	2023	2024
Enrollment			3,237,970	3,318,628	3,401,376	3,486,270	3,573,367	3,662,728	3,754,413	3,848,485	3,945,008
	Passing		1,520,101	1,557,072	1,465,515	1,683,584	1,697,608	1,749,853	1,814,372	1,879,106	1,945,703
	Zone		380,946	390,131	301,382	140,204	108,548	61,283	34,637	19,013	9,187
	Failing		321,269	328,179	126,553	55,957	47,978	25,867	15,457	6,574	3,294
	Ineligible		-	-	217,907	321,986	444,273	538,048	590,099	632,859	663,475
	Not Evaluated		1,015,654	1,043,246	1,290,019	1,284,538	1,274,960	1,287,677	1,299,848	1,310,933	1,323,349
Transfers or Dropouts from Zone, Failing, or Ineligible Programs
	Transfers			206,200	276,987	277,829	339,454	365,875	379,837	393,080	404,614
	Dropouts			68,733	92,329	92,610	113,151	121,958	126,612	131,027	134,871
	Remaining			443,376	276,525	147,709	148,195	137,365	133,743	134,340	136,470
Title IV Aid Associated with Students who Drop or Remain in Ineligible Programs
	Pell Grants			192,242,071	376,559,358	434,059,521	558,900,388	633,566,568	674,304,716	709,840,137	737,247,652
	Subsidized Loans			186,263,125	368,492,350	423,861,995	544,426,799	618,459,691	658,925,156	693,594,389	720,125,092
	Unsubsidized Loans			236,400,514	467,439,419	536,448,198	687,496,110	780,004,345	830,590,192	873,923,217	907,054,310
	PLUS Loans			34,018,998	67,706,866	78,698,969	102,172,579	116,899,155	124,958,062	131,839,083	137,124,154
Estimated Net Budget Impact using PB 2015 Subsidy Rates
	Pell Grants			192,242,071	376,559,358	434,059,521	558,900,388	633,566,568	674,304,716	709,840,137	737,247,652
	Subsidized loans			17,974,392	43,076,756	53,533,770	72,517,650	86,646,203	97,784,493	104,316,596	108,882,914
	Unsubsidized loans			(32,055,910)	(54,269,717)	(58,204,629)	(69,918,354)	(74,880,417)	(72,510,524)	(75,244,789)	(78,188,082)
	PLUS Loans			(9,307,598)	(16,913,175)	(18,462,778)	(23,642,735)	(26,384,139)	(27,178,378)	(28,622,265)	(29,125,170)
	Total			168,852,955	348,453,222	410,925,883	537,856,948	618,948,215	672,400,307	710,289,679	738,817,314

In the NPRM, the Department estimated that the net budget impact of the proposed regulations would be $666 million in the “low reaction” scenario or $973 million in the “high reaction” scenario. The increased estimate in these regulations is due to the modified methodology for the budget model described in “Methodology for net budget impacts” that applies the student response assumption to the baseline estimated enrollment and not the decreased enrollment as a result of student transfers in prior years. We believe this revised approach captures the title IV, HEA program aid that students would have continued to receive in the absence of the regulations, not only for the first year after they drop out or remain in an eligible program, but also for subsequent years as they continued their educations. While Table 3.8 presents the approximate effect on the estimated initial 37,103 programs with student enrollment in FY2010 that would first be evaluated under the regulations, it does not take into account new programs that may have been established since that time.

Table 3.8: Estimated Effect of the Regulations on Gainful Employment Programs

Programs	2016	2017	2018	2019	2020	2021	2022	2023	2024
Passing	6,710	6,710	7,471	10,510	10,413	10,695	11,180	11,604	12,009
Zone	1,205	1,205	1,178	562	415	231	123	65	31
Failing	713	713	653	237	197	91	53	23	11
Ineligible	-	-	463	919	1,282	1,614	1,769	1,868	1,922
Not Evaluated	28,475	28,475	27,338	24,875	24,795	24,471	23,978	23,543	23,130

The Department’s calculations of the net budget impacts represent our best estimate of the effect of the regulations on the Federal student aid programs. However, these estimates will be heavily influenced by actual program performance, student response to program performance, and potential increases in enrollment and retention rates as a result of the regulations. For example, if students, including prospective students, react more strongly to the consumer disclosures or potential ineligibility of programs than anticipated and, if many of these students leave postsecondary education, the impact on Pell Grants and loans could increase substantially. Similarly, if institutions react to the regulations by modifying their program offerings, enrollment strategies, or pricing, the assumed enrollment and aid amounts could be overstated.

Over the last several years, we believe that institutions in the for-profit sector have made changes to improve program performance, particularly by reducing cost and eliminating some poorly performing offerings. Because the data available to analyze the regulations are based on older cohorts of students, the budget estimates may not reflect these changes. In addition, we are unable to predict the extent to which institutions will take advantage of the transition period provisions of the regulations to reduce costs to students in failing and zone programs. Although these factors are not explicitly accounted for in the estimates, we expect that they will operate to reduce the number of failing and zone programs and affected students, and in turn, lower the net budget impact estimate.

As previously stated, we do not estimate any significant budget impact stemming from students who transfer to another institution when a program they are attending or planned to attend voluntarily closes, fails or falls in the zone under the D/E rates measure, or loses eligibility to participate in the title IV, HEA programs. Although it is true that programs have varied costs across sector, CIP code, credential level, location, and other factors, the students’ eligibility for title IV, HEA program funds carries with them across programs. It is possible that passing programs that students choose to transfer to could have lower prices than zone, failing or ineligible programs, and the amount of title IV, HEA program funds to GE programs may be reduced as a result of those transfers. However, students or counselors may also use the disclosures and earnings information to choose a different field of study or credential level which could result in increased aid volume. In general, we anticipate that overall aid to students who transfer among GE programs or to non-GE programs will not change significantly, so no net budget impact was estimated for these students.

The effects previously described represent the estimated effects of the regulations during the initial period of time after the regulations take effect. We expect that the budget effects of the regulations will decline over time as programs that are unable to pass will be eliminated and using data about program outcomes, including D/E rates, institutions will be better able to ensure that their programs consistently meet the standards of the regulations.

This gradual decline in impact of the regulations may be similar to the pattern observed when institutional cohort default rates (CDR) were introduced in 1989 with an initial elimination of the worst-performing institutions followed by an equilibrium where institutions overwhelmingly meet the CDR standards. We do not expect the impact of the regulations to drop off as sharply as occurred with the introduction of institutional CDR because of the four year zone and due to the transition period provisions which could potentially extend eligibility for programs that might otherwise become ineligible.

Accounting Statement

As required by OMB Circular A-4 (available at http://www.whitehouse.gov/sites/default/files/omb/assets/omb/circulars/a004/a-4.pdf), the accounting statement in Table 3.9 provides the classification of the expenditures associated with the regulations. The accounting statement represents our best estimate of the impact of the regulations on the Federal student aid programs.

Expenditures are classified as transfers from the Federal Government to students receiving title IV, HEA program funds and from low-performing programs to higher-performing programs. Transfers are neither costs nor benefits, but rather the reallocation of resources from one party to another.

Table 3.9: Accounting Statement

Category	Benefits
Improved market information and development of measures linking programs to labor market outcomes	Not Quantified
Better return on money spent on education	Not Quantified
Category	Costs
Discount Rate	3%	7%
Additional expense of educating transfer students at passing programs	$373	$379
Cost of Compliance with Paperwork Burden	$50	$54
Category	Transfers
Discount Rate	3%	7%
Transfer of Federal student aid money from failing programs to the Federal government when students drop out of programs	$423	$423
Estimated Transfer of revenues from non-passing programs to passing or zone programs as students transfer	$2,515	$2,554
Estimated Transfer of instructional expenses from non-passing programs to passing or zone programs as students transfer	$1,216	$1,235

Costs and Transfers Sensitivity Analysis

We also provide alternative accounting statements using varied program transition and student response assumptions to demonstrate the sensitivity of the net budget impacts to these factors. These scenarios illuminate how different student and program responses could affect the title IV, HEA programs and institutions offering GE programs. We offer extreme scenarios in order to bound the estimates of effects, although we believe these extreme scenarios are unlikely to occur.

Alternative program transition assumptions

In addition to the primary program transition assumptions provided in Tables 3.1-3.3, we assumed two additional program transition scenarios, zero program transition and positive program transition. For the zero program transition, an extreme worst case scenario, we assume institutions will have no success in improving programs. Accordingly, for this scenario, the year 0 program results, calculated based on the outcomes of students who completed GE programs in FYs 2008 and 2009 as described in “Program transition assumptions,” are held constant for each cycle of the budget model. For the positive program transition, we assumed institutions would be highly successful in improving programs. This scenario simulates the effects of 25 percent greater improvement over the primary program transition scenario described in “Program transition assumptions.” Tables 3.10 and 3.11 provide the program transition assumptions for these alternative scenarios.

Table 3.10: Zero Program Transition Assumptions

	Performance Category in Year 1
Prior Year Result	Pass	Zone	Fail	NE	Ineligible
Pass	1	0	0	0	0
Zone	0	1	0	0	0
Fail	0	0	0	0	1
Not Evaluated	0	0	0	1	0

		Performance Category in Year 2
Prior Years Results			Pass	Zone	Fail	NE	Ineligible
YR0 Pass
	Yr1 Pass		100	0	0	0	0
	Y1 Zone		0	100	0	0	0
	YR1 Fail		0	0	0	0	100
	YR1 Not Evaluated		0	0	0	100	0
	YR1 Ineligible		0	0	0	0	100
YR0 Zone
	Yr1 Pass		100	0	0	0	0
	Y1 Zone		0	100	0	0	0
	YR1 Fail		0	0	0	0	100
	YR1 Not Evaluated		0	0	0	100	0
	YR1 Ineligible		0	0	0	0	100
YR0 Fail
	Yr1 Pass		100	0	0	0	0
	Y1 Zone		0	100	0	0	0
	YR1 Fail		0	0	0	0	100
	YR1 Not Evaluated		0	0	0	100	0
	YR1 Ineligible		0	0	0	0	100
YR0 Not Evaluated
	Yr1 Pass		100	0	0	0	0
	Y1 Zone		0	100	0	0	0
	YR1 Fail		0	0	0	0	100
	YR1 Not Evaluated		0	0	0	100	0
	YR1 Ineligible		0	0	0	0	100

	Performance Category in Subsequent Year
		Pass	Zone	Fail	Ne	Ineligible
Prior Year Group:
Good		100	0	0	0	0
Improving		0	100	0	0	0
Poor/Declining		0	0	100	0
Fail Next		0	0	0	0	100
Ineligible		0	0	0	0	100
Not Evaluated		0	0	0	100	0

Table 3.11: Positive (+ 25 percent) Program Transition Assumptions

	Performance Category in Year 1
	Pass	Zone	Fail	NE	Ineligible
Pass	82.75	3	0.75	13.5	0
Zone	38.25	30	18.75	13	0
Fail	25.25	15	0	11	48.75
Not Evaluated	6.5	0.75	0.75	92	0

			Performance Category in Year 2
Prior Years Results		Pass		Zone	Fail	NE	Ineligible
YR0 Pass
	Yr1 Pass	97.75		0	0	2.25	0
	Y1 Zone	48.25		24	14.75	13	0
	YR1 Fail	30.25		12	0	11	46.75
	YR1 Not Evaluated	16.5		0	0	83.5	0
	YR1 Ineligible	0		0	0	0	100
YR0 Zone
	Yr1 Pass	92.75		0	0	7.25	0
	Y1 Zone	43.25		27	16.75	13	0
	YR1 Fail	25.25		15	0	11	48.75
	YR1 Not Evaluated	11.5		0	0	88.5	0
	YR1 Ineligible	0		0	0	0	100
YR0 Fail
	Yr1 Pass	87.75		0	0	12.25	0
	Y1 Zone	38.25		30	18.75	13	0
	YR1 Fail	0		0	0	0	100
	YR1 Not Evaluated	6.5		0.75	0	92	0.75
	YR1 Ineligible	0		0	0	0	100
				`
YR0 Not Evaluated
	Yr1 Pass	92.75		0	0	7.25	0
	Y1 Zone	43.25		27	16.75	13	0
	YR1 Fail	25.25		15	0	11	48.75
	YR1 Not Evaluated	11.5		0	0	88.5	0
	YR1 Ineligible	0		0	0	0	100

	Pass	Zone	Fail	Ne	Ineligible
Prior Year Group:
Good	97.75	0	0	2.25	0
Improving	53.25	21	12.75	13	0
Poor/Declining	33.25	32	21.75	13	0
Fail Next	20.25	17	0	11	51.75
Ineligible	0	0	0	0	100
NE	11.5	0	0	88.5	0

Alternative student response assumptions

We also assumed two additional student response scenarios, zero student response and strong student response. For the zero program response, an extreme worst case scenario, we assumed students in zone and failing programs would not react to warnings and disclosures and instead, would remain in their programs until they are made ineligible. For the strong student response, we assumed students would be highly responsive to program performance. This scenario simulates the effects of 25 percent greater student reaction over the primary student response scenario described in “Student response assumptions.” Tables 3.12 and 3.13 provide the student response assumptions for these alternative scenarios.

Table 3.12: Zero Student Response Assumptions

	Student Response
Program Result	Stay	Transfer	Drop
First Zone	1.00	0.00	0.00
Second Zone	1.00	0.00	0.00
Third Zone	1.00	0.00	0.00
Fourth Zone	1.00	0.00	0.00
First Failure	1.00	0.00	0.00
Second Failure	1.00	0.00	0.00
Ineligibility	1.00	0.00	0.00
Passing or Not Evaluated	1.00	0.00	0.00

Table 3.13: Strong (+ 25 percent) Student Response Assumption

	Student Response
Program Result	Stay	Transfer	Drop
First Zone	0.75	0.19	0.06
Second Zone	0.50	0.38	0.13
Third Zone	0.25	0.56	0.19
Fourth Zone	0.00	0.75	0.25
First Failure	0.25	0.56	0.19
Second Failure	0.00	0.75	0.25
Ineligibility	0.00	0.75	0.25
Passing or Not Evaluated	1.00	0.00	0.00

The costs and transfers associated with the combinations of primary and alternative program and student response scenarios are provided in Tables 3.14 – 3.16.

Table 3.14: Costs and Transfers Associated with Zero Student Response Assumptions

Estimates	Low Program, Low Student		Main Program, Low Student		High Program, Low Student
Average Annual Student Transfers over 2017-2024	-		-		-
Average Annual Student Dropouts over 2017-2024	-		-		-
	3%	7%	3%	7%	3%	7%
Additional expense of educating transfer students at passing programs	$0	$0	$0	$0	$0	$0
Transfer of Federal student aid money from failing programs to the Federal government when students drop out of programs or remain in ineligible programs	$1,291	$1,275	$918	$905	$574	$567
Estimated Transfer of revenues from non-passing programs to passing or zone programs as students transfer	$0	$0	$0	$0	$0	$0
Estimated Transfer of instructional expenses from non-passing programs to passing or zone programs as students transfer	$0	$0	$0	$0	$0	$0

Table 3.15: Costs and Transfers Associated with Primary Student Response Assumptions

Estimates	Low Program, Main Student Assumptions		Main Program, Main Student Assumptions		High Program, Main Student Assumptions
Average Annual Student Transfers over 2017-2024	416,538		330,484		223,719
Average Annual Student Dropouts over 2017-2024	138,846		110,161		74,573
	3%	7%	3%	7%	3%	7%
Additional expense of educating transfer students at passing programs	$470	$476	$373	$379	$254	$260
Transfer of Federal student aid money from failing programs to the Federal government when students drop out of programs or remain in ineligible programs	$565	$565	$423	$423	$277	$280
Estimated Transfer of revenues from non-passing programs to passing or zone programs as students transfer	$3,170	$3,212	$2,515	$2,554	$1,719	$1,763
Estimated Transfer of instructional expenses from non-passing programs to passing or zone programs as students transfer	$1,530	$1,550	$1,216	$1,235	$829	$851

Table 3.16: Costs and Transfers Associated with Strong (+ 25 percent) Student Response Assumptions

Estimates	Low Prog, High Stu		Main Program, High Student Assumptions		High Program, High Student Assumptions
Average Annual Student Transfers over 2017-2024	520,813		404,069		279,640
Average Annual Student Dropouts over 2017-2024	173,916		134,964		93,404
	3%	7%	3%	7%	3%	7%
Additional expense of educating transfer students at passing programs	$588	$595	$466	$473	$317	$325
Transfer of Federal student aid money from failing programs to the Federal government when students drop out of programs or remain in ineligible programs	$384	$388	$299	$303	$214	$218
Estimated Transfer of revenues from non-passing programs to passing or zone programs as students transfer	$3,964	$4,016	$3,143	$3,192	$2,025	$2,077
Estimated Transfer of instructional expenses from non-passing programs to passing or zone programs as students transfer	$1,913	$1,938	$1,520	$1,543	$1,036	$1,063

4. Regulatory Alternatives Considered

As part of the development of these regulations, the Department engaged in a negotiated rulemaking process in which we received comments and proposals from non-Federal negotiators representing institutions, consumer advocates, students, financial aid administrators, accreditors, and State Attorneys General. The non-Federal negotiators submitted a variety of proposals relating to placement rates, protections for students in failing programs, exemptions for programs with low borrowing or default rates, rigorous approval requirements for existing and new programs, as well as other issues. Information about these proposals is available on the GE Web site at http://www2.ed.gov/policy/highered/reg/hearulemaking/2012/gainfulemployment.html. The Department also published proposed regulations in a notice of proposed rulemaking and invited public comment. We received comments, including proposals, on a wide range of issues related to the regulations. We have responded to these comments in the preamble of the final regulations.

In addition to the proposals from the non-Federal negotiators and the public, the Department considered alternatives to the regulations based on its own analysis, including alternative provisions for the D/E rates measure, as well as alternative metrics. Important alternatives that were considered are discussed below.

Alternative Components of the D/E Rates Measure

N-Size

For the purpose of calculating the D/E rates measure, we considered reducing the n-size for program evaluation to 10 students who completed a program in a two-year cohort period. At an n-size of 10, about 50 percent of GE programs would be subject to evaluation under the D/E rates measure. However, these additional programs account for a relatively small proportion of students receiving title IV, HEA program funds for enrollment in GE programs. Although we believe an n-size of 10 would be reasonable for the D/E rates measure, we elected to retain the n-size of 30 and to include those who completed over a four-year period if needed to achieve a 30-student cohort for a given program. Our data show that, using the two-year cohort period, 5,539 programs have enough students who completed the program to satisfy an n-size of 30. These 5,539 programs represent approximately 60 percent of students who received title IV, HEA program funds for enrolling in a GE program. Further, we estimate that, using the four-year cohort period, 3,356 additional programs would meet an n-size of 30.

Table 4.1: Effect of N-Size on Programs Evaluated under the D/E Rates Measure

Result	N>=10		N>=30
	Programs	Enrollment	Programs	Enrollment
Pass	8,799	2,021,644	4,094	1,679,616
Zone	1,394	500,055	928	453,904
Fail	857	427,982	517	387,763
Total	11,050	2,949,681	5,539	2,521,283

Interest Rates

As demonstrated by Table 4.2, the interest rate used in the D/E rates calculations has a substantial effect on a program’s performance under the D/E rates measure.

Table 4.2: Interest Rate Impact on D/E Rates Results (Total 5,539 programs)

Although the calculation of the D/E rates measure is based on a group of students who completed a program over a particular two- or four-year period, the dates on which each of these students may have taken out a loan, and the interest rates on those loans, vary. The Department considered several options for the interest rate to apply to the D/E rates measure calculation. For the NPRM, we used the average interest rate over the six years prior to the end of the applicable cohort period on Federal Direct Unsubsidized loans. This proposal was designed to approximate the interest rate that a large percentage of the students in the calculation received, even those students who attended four-year programs, and to mitigate any year-to-year fluctuations in the interest rates that could lead to volatility in the results of programs under the D/E rates measure. Some commenters suggested using the actual interest rates on an individual borrower level, but we believe that would be unnecessarily complicated. Other commenters suggested that we adopt a sliding scale, with interest rates averaged over a number of years that corresponds to program length. As discussed in “§668.404 Calculating D/E Rates” in Analysis of Comments and Changes, we adopted this proposal for the final regulations. For certificate, associate, and master’s degree programs, the average interest rate over the three years prior to the end of the applicable cohort period on Federal Direct Unsubsidized loans will be used to calculate the D/E rates measure. For bachelor’s, doctoral, and first professional degree programs, the average interest rate over the six years prior to the end of the applicable cohort period on Federal Direct Unsubsidized loans will be used. The undergraduate interest rate on these loans will be applied to undergraduate programs, and the graduate interest rate will be applied to graduate programs.

Table 4.3: Options for Determining Interest Rate³⁴² for D/E Rates Calculation

		Four-year average			Three-year average			Two-year average
2YP	2YPMED	UG	GRAD	MED	UG	GRAD	MED	UG	GRAD	MED
08-09	05-06	6.43%	6.43%	4.04%	6.80%	6.80%	4.03%	6.80%	6.80%	4.34%
11-12	08-09	6.80%	6.80%	6.43%	6.80%	6.80%	6.80%	6.80%	6.80%	6.80%
12-13	09-10	6.80%	6.80%	6.80%	6.80%	6.80%	6.80%	6.80%	6.80%	6.80%
13-14	10-11	6.07%	6.45%	6.80%	5.82%	6.34%	6.80%	5.33%	6.11%	6.80%
14-15	11-12	5.61%	6.38%	6.80%	5.21%	6.24%	6.80%	4.42%	5.97%	6.80%
15-16	12-13	5.26%	6.42%	6.80%	4.75%	6.29%	6.80%	5.19%	6.73%	6.80%
16-17	13-14	4.99%	6.53%	6.45%	5.37%	6.90%	6.34%	5.57%	7.09%	6.11%

Amortization Period

The regulations apply the same 10-, 15-, 20-year amortization periods by credential level as under the 2011 Prior Rule. In calculating the annual loan payment for the purpose of the D/E rates measure, a 10-year amortization period would be used for certificate and associate degree programs, 15 years for bachelor’s and master’s degree programs, and 20 years for doctoral and first professional degree programs. We presented at the negotiations, as an alternative, a 10-year amortization period for all programs, which we believe is a reasonable assumption. In the NPRM, we invited comment on a 10-year schedule for all programs and also on a 20-year schedule for all programs.

As discussed in the NPRM, we analyzed available data on the repayment plans that existing borrowers have selected and the repayment patterns of older loan cohorts and considered the repayment schedule options available under consolidation loan repayment rules. Although the prevalence of the standard 10-year repayment plan and data related to older cohorts could support a 10-year amortization period for all credential levels, the Department has retained the split amortization approach in the regulation. Growth in loan balances, the introduction of plans with longer repayment periods than were available when those older cohorts were in repayment, and some differentiation in repayment periods by credential level in more recent cohorts contributed to this decision.

As provided in Tables 4.4 and 4.5, extending the amortization periods for lower credentials would reduce the number of programs that fail or fall in the zone under the D/E rates measure, and shortening the amortization period for higher credentials would increase the number of failing and zone programs. The greatest effect would be on graduate-level programs.

Table 4.4: D/E Rates Results by Sector and Credential (N-Size of 30, 10-Year Amortization for all Credential Levels)

Sector	IHE Type	Credential Level	Programs	Passing Programs	Zone Programs	Failing Programs	Enrollment	Enrollment in Passing Programs	Enrollment in Zone Programs	Enrollment in Failing Programs
Public	Total		1,093	1,090	2	1	142,400	142,077	277	46
	< 2 year	Certificate	157	157	0	0	11,439	11,439	0	0
	2-3 year	Certificate	824	823	1	0	119,615	119,559	56	0
	4-year	Certificate	86	84	1	1	8,102	7,835	221	46
		Post-Bacc Certificate	26	26	0	0	3,244	3,244	0	0
Private	Total		253	242	8	3	45,696	40,695	3,886	1,115
	< 2 year	Certificate	49	47	2	0	9,609	9,147	462	0
	2-3 year	Certificate	73	70	3	0	10,307	8,875	1,432	0
		Post-Bacc Certificate	1	1	0	0	17	17	0	0
	4-year	Certificate	91	86	3	2	20,666	17,679	1,992	995
		Post-Bacc Certificate	39	38	0	1	5,097	4,977	0	120
For-Profit	Total		4,193	2,723	908	562	2,333,187	1,440,196	474,526	418,465
	< 2 year	Certificate	1,100	877	185	38	216,363	154,749	51,207	10,407
		Associate's	5	4	1	0	195	195	0	0
		1st Professional Degree	4	4	0	0	312	312	0	0
	2-3 year	Certificate	1,223	903	264	56	365,500	255,040	97,385	13,075
		Associate's	452	215	160	77	105,750	41,914	34,921	28,915
		Post-Bacc Certificate	2	2	0	0	156	156	0	0
	4-year	Certificate	267	169	70	28	84,610	47,102	30,205	7,303
		Associate's	514	183	167	164	669,030	240,135	174,977	253,918
		Bachelor's	407	176	52	179	618,330	447,758	74,024	96,548
		Post-Bacc Certificate	8	8	0	0	1,950	1,950	0	0
		Master's	171	153	6	12	226,106	214,922	7,909	3,275
		Doctoral	30	27	1	2	37,676	34,085	2,669	922
		1st Professional Degree	10	2	2	6	7,209	1,878	1,229	4,102
Overall Total			5,539	4,055	918	566	2,521,283	1,622,968	478,689	419,626

Table 4.5: D/E Rates Results by Sector and Credential (N-Size of 30, 20-Year Amortization for all Credential Levels)

Sector	IHE Type	Credential Level	Programs	Passing Programs	Zone Programs	Failing Programs	Enrollment	Enrollment in Passing Programs	Enrollment in Zone Programs	Enrollment in Failing Programs
Public	Total		1,093	1,092	1	0	142,400	142,354	46	0
	< 2 year	Certificate	157	157	0	0	11,439	11,439	0	0
	2-3 year	Certificate	824	824	0	0	119,615	119,615	0	0
	4-year	Certificate	86	85	1	0	8,102	8,056	46	0
		Post-Bacc Certificate	26	26	0	0	3,244	3,244	0	0
Private	Total		253	250	2	1	45,696	44,581	998	117
	< 2 year	Certificate	49	49	0	0	9,609	9,609	0	0
	2-3 year	Certificate	73	73	0	0	10,307	10,307	0	0
		Post-Bacc Certificate	1	1	0	0	17	17	0	0
	4-year	Certificate	91	89	1	1	20,666	19,671	878	117
		Post-Bacc Certificate	39	38	1	0	5,097	4,977	120	0
For-Profit	Total		4,193	3,643	364	186	2,333,187	1,921,377	302,473	109,337
	< 2 year	Certificate	1,100	1,063	34	3	216,363	206,008	9,731	624
		Associate's	5	5	0	0	195	195	0	0
		1st Professional Degree	4	4	0	0	312	312	0	0
	2-3 year	Certificate	1,223	1,169	49	5	365,500	352,788	12,189	523
		Associate's	452	379	57	16	105,750	77,226	16,125	12,399
		Post-Bacc Certificate	2	2	0	0	156	156	0	0
	4-year	Certificate	267	239	24	4	84,610	77,307	7,002	301
		Associate's	514	350	118	46	669,030	415,112	206,900	47,018
		Bachelor's	407	233	73	101	618,330	527,631	44,833	45,866
		Post-Bacc Certificate	8	8	0	0	1,950	1,950	0	0
		Master's	171	159	4	8	226,106	222,831	2,055	1,220
		Doctoral	30	28	2	0	37,676	36,754	922	0
		1st Professional Degree	10	4	3	3	7,209	3,107	2,716	1,386
Overall Total			5,539	4,985	367	187	2,521,283	2,108,312	303,517	109,454

D/E Rates Thresholds and the Zone

We also considered the related issues of the appropriate thresholds for the D/E rates measure and whether there should be a zone. The regulations establish stricter passing thresholds than the thresholds in the 2011 Prior Rule. The passing threshold for the discretionary income rate is 20 percent instead of 30 percent, and the threshold for the annual earnings rate is 8 percent instead of 12 percent. Additionally, the regulations add a zone category for programs with a discretionary income rate greater than 20 percent but less than or equal to 30 percent or an annual earnings rate greater than 8 percent but less than or equal to 12 percent.

The passing thresholds for the discretionary income rate and the annual earnings rate are based upon mortgage industry practices and expert recommendations. The justification for these thresholds is included in the Preamble.

Table 4.6: D/E Rates Measure Results for Alternative Thresholds

Estimated Effects of the D/E Rates Alternatives

In order to consider the alternatives for calculation of the D/E rates, we estimated the budget impact of the alternatives on program results under the D/E rate measure. The results are summarized in Table 4.7. To evaluate the alternatives, we used the same data, methods, and assumptions as the estimates described in “Methodology for Costs, Benefits, and Transfers” and the “Net Budget Impacts” sections of this RIA. The alternatives considered would result in different estimated distributions of enrollment in passing, zone, and failing programs under the regulations, leading to the results in Table 4.7.

Table 4.7: Estimated Effects of D/E Rates Alternatives

Estimates	N10, 10-15-20 Amortization
Average Annual Student Transfers over 2017-2024	329,914
Average Annual Student Dropouts over 2017-2024	109,971
	3%	7%
Additional expense of educating transfer students at passing programs	$382	$388
Transfer of Federal student aid money from failing programs to the Federal government when students drop out of programs or remain in ineligible programs	$433	$433
Estimated Transfer of revenues from non-passing programs to passing or zone programs as students transfer	$2,576	$2,616
Estimated Transfer of instructional expenses from non-passing programs to passing or zone programs as students transfer	$1,246	$1,266

Estimates	N30, 10 year Amortization for all		N30, 20-year Amortization for all
Average Annual Student Transfers over 2017-2024	321,663		162,551
Average Annual Student Dropouts over 2017-2024	107,221		54,184
	3%	7%	3%	7%
Additional expense of educating transfer students at passing programs	$368	$374	$179	$180
Transfer of Federal student aid money from failing programs to the Federal government when students drop out of programs or remain in ineligible programs	$412	$412	$192	$191
Estimated Transfer of revenues from non-passing programs to passing or zone programs as students transfer	$2,517	$2,558	$1,229	$1,241
Estimated Transfer of instructional expenses from non-passing programs to passing or zone programs as students transfer	$1,200	$1,219	$584	$589

Discretionary Income Rate

Instead of two debt-to-earnings ratios, the annual earnings rate and the discretionary income rate, we considered a simpler approach where only the discretionary income rate would be used as a metric. However, this would have led to any program with earnings below the discretionary income level failing the measure. Removing the annual earnings rate altogether would make ineligible programs that, based on expert analysis, leave students with manageable levels of debt. In some cases, programs may leave graduates with low earnings, but these students may also have minimal debt that is manageable at those earnings levels.

For these programs, rather than establish a minimum earnings threshold through a single discretionary earnings rate measure, we believe that students, using the information about program outcomes that will be available as a result of the disclosures, should be able to make their own assessment of whether the potential earnings will meet their goals and expectations.

Pre- and Post-Program Earnings Comparison

The Department also considered an approach that would compare pre-program and post-program earnings to capture the near-term effect of the program. This approach had been suggested by commenters responding to the 2011 Prior Rule and to the NPRM, especially for short-term programs, and has some merit conceptually. While it is important that programs lead to earnings gains, we believe that the D/E rates measure better achieves the objectives of these regulations by assessing earnings in the context of whether they are at a level that would allow borrowers to manage their debt and avoid default.

pCDR

pCDR Measure

In the NPRM, the Department proposed that programs must pass a program-level cohort default rate (pCDR) measure, in addition to the D/E rates measure. Unlike the D/E rates measure, the pCDR measure would assess the outcomes of both students who complete GE programs and those who do not. The pCDR measure adopted almost all of the statutory and regulatory requirements of the institutional cohort default rate (iCDR) measure that is used to measure default rates at the institutional level for all title IV eligible institutions. As proposed, GE programs would fail the measure if more than 30 percent of borrowers defaulted on their FFEL or Direct Loans within the first three years of entering repayment. Programs that failed the pCDR measure for three consecutive years would become ineligible.

The Department strongly believes in the importance of holding GE programs accountable for the outcomes of students who do not complete a program and ensuring that institutions make meaningful efforts to increase completion rates. However, given the wealth of feedback we received, we believe further study is necessary before we adopt pCDR or another accountability metric that would take into account the outcomes of students who do not complete a program. Therefore, we are not adopting pCDR as an accountability metric. Using the information we receive from institutions through reporting, we will work to develop a robust measure of outcomes for students who do not complete their programs.

We continue to believe that default rates are important for students to consider as they decide where to pursue, or continue, their postsecondary education and whether or not to borrow to attend a particular program. Accordingly, we are retaining pCDR as one of the disclosures that institutions may be required to make under §668.412. We believe that requiring this disclosure, along with other potential disclosures such as completion, withdrawal, and repayment rates, will bring a level of accountability and transparency to GE programs with high rates of non-completion.

Table 4.8: Estimated Results under pCDR measure

Sector	IHE Type	Credential Level	Programs	Passing Programs	Failing Programs	Enrollment	Enrollment in Passing Programs	Enrollment in Failing Programs
Public	Total		902	850	52	121,650	108,995	12,655
	< 2 year	Certificate	119	115	4	9,489	9,293	196
	2-3 year	Certificate	701	655	46	104,399	92,090	12,309
	4-year	Certificate	60	58	2	5,055	4,905	150
		Post-Bacc Certificate	22	22	0	2,707	2,707	0
Private	Total		262	236	26	40,039	36,317	3,722
	< 2 year	Certificate	33	25	8	5,655	4,427	1,228
	2-3 year	Certificate	66	63	3	8,877	8,603	274
		Post-Bacc Certificate	1	1	0	17	17	0
	4-year	Certificate	94	79	15	19,263	17,043	2,220
		Post-Bacc Certificate	68	68	0	6,227	6,227	0
For-Profit	Total		5,651	4,786	865	2,583,388	1,921,468	661,920
	< 2 year	Certificate	1,027	869	158	196,484	157,098	39,386
		Associate's	4	3	1	87	34	53
		1st Professional Degree	3	2	1	262	262	0
	2-3 year	Certificate	1,386	1,128	258	349,369	270,025	79,344
		Associate's	832	700	132	135,988	109,139	26,849
		Post-Bacc Certificate	2	2	0	156	156	0
	4-year	Certificate	398	337	61	90,875	83,496	7,379
		Associate's	958	746	212	774,875	302,358	472,517
		Bachelor's	721	679	42	737,414	701,022	36,392
		Post-Bacc Certificate	26	26	0	3,960	3,960	0
		Master's	218	218	0	235,113	235,113	0
		Doctoral	67	67	0	51,931	51,931	0
		1st Professional Degree	9	9	0	6,874	6,874	0
Overall Total			6,815	5,872	943	2,745,077	2,066,780	678,297

Table 4.9: Estimated Results under D/E rates measure + pCDR measure

Sector	IHE Type	Credential Level	Programs	Passing Programs	Zone Programs	Failing Programs	Enrollment	Enrollment in Passing Programs	Enrollment in Zone Programs	Enrollment in Failing Programs
Public	Total		1,507	1,453	1	53	195,087	182,165	221	12,701
	< 2 year	Certificate	179	175	0	4	12,203	12,007	0	196
	2-3 year	Certificate	1,178	1,132	0	46	169,275	156,966	0	12,309
	4-year	Certificate	115	111	1	3	9,955	9,538	221	196
		Post-Bacc Certificate	35	35	0	0	3,654	3,654	0	0
Private	Total		345	310	6	29	52,305	43,776	3,692	4,837
	< 2 year	Certificate	54	45	1	8	9,796	8,172	396	1,228
	2-3 year	Certificate	86	81	2	3	10,952	9,374	1,304	274
		Post-Bacc Certificate	1	1	0	0	17	17	0	0
	4-year	Certificate	127	107	3	17	24,706	19,499	1,992	3,215
		Post-Bacc Certificate	77	76	0	1	6,834	6,714	0	120
For-Profit	Total		6,082	4,071	748	1,263	2,666,984	1,517,809	301,309	847,866
	< 2 year	Certificate	1,275	938	151	186	224,500	138,444	38,452	47,604
		Associate's	5	3	1	1	195	142	0	53
		1st Professional Degree	4	3	0	1	312	312	0	0
	2-3 year	Certificate	1,505	1,010	195	300	379,498	220,076	71,970	87,452
		Associate's	839	513	137	189	139,033	63,153	30,337	45,543
		Post-Bacc Certificate	2	2	0	0	156	156	0	0
	4-year	Certificate	412	274	57	81	93,097	52,045	27,557	13,495
		Associate's	971	510	140	321	781,846	148,293	81,531	552,022
		Bachelor's	738	509	58	171	746,345	602,143	46,313	97,889
		Post-Bacc Certificate	27	27	0	0	3,999	3,999	0	0
		Master's	227	213	4	10	238,863	234,930	1,511	2,422
		Doctoral	67	65	2	0	51,931	51,009	922	0
		1st Professional Degree	10	4	3	3	7,209	3,107	2,716	1,386
Overall Total			7,934	5,834	755	1,345	2,914,376	1,743,750	305,222	865,404

pCDR Thresholds

As described above, we modeled the proposed pCDR measure on the iCDR measure that is currently used to determine institutional eligibility to participate in title IV, HEA programs. In addition to adopting the iCDR threshold under which an institution loses eligibility if it has three consecutive fiscal years of an iCDR of 30 percent or greater, we considered adopting the second iCDR threshold, pursuant to which an institution loses eligibility if it has one year of an iCDR of 40 percent or greater. Of the 6,815 programs in the 2012 GE informational rates sample with pCDR data, 233 have a default rate of 40 percent or more.

Negative Amortization

The Department also considered in its design of the NPRM a variation on a repayment metric that would compare the total amounts that borrowers, both students who completed a program and students who did not, owed on their FFEL and Direct Loans at the beginning and end of their third year of repayment to determine if borrower payments reduced the balance on their loans over the course of that year. Different variations of this measure were considered, including a comparison of total balances and a comparison of principal balances. We considered using this metric in addition to the D/E rates measure to measure the performance of students who did not complete the program as well as those that did. Ultimately, the Department decided not to propose negative amortization as an eligibility metric in the proposed regulations because we were unable to draw clear conclusions at this time from the data available.

Programs with Low Rates of Borrowing

Several negotiators and, as discussed in the preamble, many commenters argued that programs for which a majority of students do not borrow should not be subject to the D/E rates measure or should be considered to be passing the measure because results would not accurately reflect the level of borrowing by individuals enrolled in the program and the low cost of the program. They contended that low rates of borrowing indicate that a program is low cost and, therefore, of low financial risk to students, prospective students, and taxpayers.

In the NPRM, institutions would have been permitted to demonstrate that a program with D/E rates that are failing or in the zone should instead be deemed to be passing the D/E rates measure because less than 50 percent of all individuals who completed the program, both those who received title IV, HEA program funds, and those who did not, had to assume any debt to enroll in the program.

As discussed in detail in “668.401 Scope and Purpose,” we have not retained these provisions for the final regulations. We do not believe the commenters presented an adequate justification for us to depart from the purpose of the regulations--to evaluate the outcomes of students receiving title IV, HEA program funds and a program’s continuing eligibility to receive title IV, HEA program funds based solely on those outcomes--even for the limited purpose of demonstrating that a program is “low risk.” Further, we agree with the commenters who suggested that a program for which fewer than 50 percent of individuals borrow is not necessarily low risk to students and taxpayers. Because the proposed showing of mitigating circumstances would be available to large programs with many students, and therefore there may be significant title IV, HEA program funds borrowed for a program, it is not clear that the program poses less risk simply because those students, when considered together with individuals who do not receive title IV, HEA program funds, comprise no more than 49 percent of all students. We also note that, if a program is indeed “low cost” or does not have a significant number of borrowers, it is very likely that the program will pass the D/E rates measure.

Borrower Protections

During the negotiated rulemaking sessions, members of the negotiated rulemaking committee offered various proposals to provide relief to students in programs that become ineligible, for example, requiring institutions to make arrangements to reduce student debt. Although we developed a debt reduction proposal for consideration by the rulemaking committee, we did not include any borrower relief provisions in the NPRM and have not done so in the final regulations.

We developed our debt reduction proposal in response to suggestions from negotiators representing consumer advocates and students. We presented regulatory provisions that would have required an institution with a program that could lose eligibility the following year to make sufficient funds available to enable the Department, if the program became ineligible, to reduce the debt burden of students who attended the program during that year. The amount of funds would have been approximately the amount needed to reduce the debt burden of students to the level necessary for the program to pass the D/E rates measure and pCDR measure. If the program were to lose eligibility, the Department would use the funds provided by the institution to pay down the loans of students who were enrolled at that time or who attended the program during the following year. We also included provisions that, during the transition period, would have alternatively allowed an institution to offer to every enrolled student for the duration of their program, and every student who subsequently enrolled while the program’s eligibility remained in jeopardy, institutional grants in the amounts necessary to reduce loan debt to a level that would result in the program passing the D/E rates and pCDR measures. If an institution took advantage of this option, a program that would otherwise lose eligibility would avoid that consequence during the transition period.

We acknowledge the desire to ease the debt burden of students attending programs that become ineligible and to shift the risk to the institutions that are enrolling students in these programs. We also recognize that the loan reduction plan proposal would give institutions with the means to institute such a program more control over their performance under the D/E rates measure. However, the discussions among the negotiators made it clear that the issues remain extremely complex, as negotiators raised concerns about the extent to which relief would be provided, what cohort of students would receive relief, and whether the proposals made by negotiators would be sufficient. The Department is not prepared to address these concerns in these regulations at this time, but we will continue to explore options to address these concerns. However, we note that under these regulations, the student warnings and disclosure template will provide students with resources to compare programs where they may continue their training and potentially apply academic credits they have earned toward completion of another program.

5. Final Regulatory Flexibility Analysis

This Final Regulatory Flexibility Analysis presents an estimate of the effect on small entities of the regulations. The U.S. Small Business Administration Size Standards define “for-profit institutions” as “small businesses” if they are independently owned and operated and not dominant in their field of operation with total annual revenue below $7,000,000, and defines “non-profit institutions” as small organizations if they are independently owned and operated and not dominant in their field of operation, or as small entities if they are institutions controlled by governmental entities with populations below 50,000. In the NPRM, the Secretary invited comments from small entities as to whether they believe the proposed changes would have a significant economic impact on them and requested evidence to support that belief. This final analysis responds to and addresses comments that were received.

Description of the Reasons that Action by the Agency Is Being Considered

The Secretary is creating through these final regulations a definition of “gainful employment in a recognized occupation” by establishing what we consider, for purposes of meeting the requirements of section 102 of the HEA, to be a reasonable relationship between the loan debt incurred by students in a training program and income earned from employment after the student completes the training.

As described in this RIA, the trends in graduates’ earnings, student loan debt, defaults, and repayment underscore the need for the Department to act. The gainful employment accountability framework takes into consideration the relationship between total student loan debt and earnings after completion of a postsecondary program.

Succinct Statement of the Objectives of, and Legal Basis for, the Regulations

As discussed in the NPRM, these final regulations are intended to address growing concerns about high levels of loan debt for students enrolled in postsecondary education programs that presumptively provide training that leads to gainful employment in a recognized occupation. The HEA applies different criteria for determining the eligibility of these programs to participate in the title IV, HEA programs. In the case of shorter programs and programs of any length at for-profit institutions, eligibility is restricted to programs that “prepare students for gainful employment in a recognized occupation.” Generally, the HEA does not require degree programs greater than one year in length at public and non-profit institutions to meet this gainful employment requirement in order to be eligible for title IV, HEA program funds. This difference in eligibility is longstanding and has been retained through many amendments to the HEA. As recently as August 14, 2008, when the HEOA was enacted, Congress again adopted the distinct treatment of for-profit institutions while adding an exception for certain liberal arts baccalaureate programs at some for-profit institutions.

Description of and, Where Feasible, an Estimate of the Number of Small Entities to which the Regulations Will Apply

The regulations will apply to programs that, as discussed above, must prepare students for gainful employment in a recognized occupation to be eligible for title IV, HEA program funds. The Department estimates that significant number of programs offered by small entities will be subject to the regulations. As stated in connection with the 2011 Prior Rule, given private non-profit institutions are considered small entities regardless of revenues, a wide range of institutions will be covered by the regulations. These entities may include institutions with multiple programs, a few of which are covered by the regulations, as well as single-program institutions with well-established ties to a local employer base. Many of the programs that will be subject to the regulations are offered by for-profit institutions and public and private non-profit institutions with programs less than two years in length. We expect that small entities with a high percentage of programs that are failing or in the zone under the D/E rates measure will be more likely to discontinue operations than will large entities.

The structure of the regulations and the n-size provisions reduce the effect of the regulations on small entities but complicate the analysis. The regulations provide for the evaluation of individual GE programs offered by postsecondary institutions, but these programs are administered by the institution, either at the branch level or on a system-wide basis, so the status as a small entity is determined at the institutional level. Table 5.1 presents the distribution of programs and enrollment at small entities by performance on the 2012 informational rates.

Table 5.1: Performance on D/E Rates Measure by Programs at Small Entities

Sector	IHE_type	Program count				Enrollment count
		Total	Pass	Zone	Fail	Total	Pass	Zone	Fail
Private	< 2 year	49	47	2	0	9,609	9,147	462	462
	2-3 year	74	71	3	0	10,324	8,892	1,432	1,432
	4+ year	130	124	3	3	25,763	22,656	1,992	1,992
For-Profit	< 2 year	767	613	125	29	107,304	77,498	24,710	24,710
	2-3 year	435	313	85	37	48,142	32,473	11,148	11,148
	4+ year	51	30	10	11	6,532	4,128	846	846

One factor that could contribute to the effect of the regulations on a small entity is the number of programs it offers that are covered by the regulations and how those programs perform. If an institution only has a limited number of programs, the effect on the institution could be greater. Table 5.2 provides an estimate of the number of small entities that offer a limited number of GE programs and the number of these small entities where 50 percent or more of their programs could fail or fall in the zone under the D/E rates measure.

Table 5.2: Distribution and D/E Rates Measure Performance of Small Entities by Number of Programs

Number of Programs per Small Entity	Number of Small Entities	Number of Small Entities with More Than 50% Failing	Number of small entities with more than 50% of zone or failing
1	709	58	186
2	179	1	9
3	47	1	5
4	32	0	4
5	15	0	1
6	6	0	1
7	1	0	0
8	2	0	0
9	1	0	0
12	1	0	0
15	1	0	0

While private non-profit institutions are classified as small entities, our estimates indicate that very few programs at those institutions are likely to fail the D/E rates measure, with an even smaller number likely to be found ineligible. The governmental entities controlling public sector institutions are not expected to fall below the 50,000 population threshold for small status under the Small Business Administration’s Size Standards, but, even if they do, programs at public sector institutions are highly unlikely to fail the D/E rates measure. Accordingly, our analysis of the effects on small entities focuses on the for-profit sector.

Description of the Projected Reporting, Recordkeeping, and Other Compliance Requirements of the Regulations, Including an Estimate of the Classes of Small Entities that Will Be Subject to the Requirements and the Type of Professional Skills Necessary for Preparation of the Report or Record

Table 5.3 relates the estimated burden of each information collection requirement to the hours and costs estimated in Paperwork Reduction Act of 1995. This additional workload is discussed in more detail under Paperwork Reduction Act of 1995.  Additional workload would normally be expected to result in estimated costs associated with either the hiring of additional employees or opportunity costs related to the reassignment of existing staff from other activities.  In total, these regulations are estimated to increase burden on small entities participating in the title IV, HEA programs by 1,947,273 hours in the initial year of reporting.  The monetized cost of this additional burden on institutions, using wage data developed using BLS data available at www.bls.gov/ncs/ect/sp/ecsuphst.pdf, is $71,172,816. In subsequent years, this burden would be reduced as institutions would only be reporting for a single year and we would expect the annual cost to be approximately $18 million.  This cost was based on an hourly rate of $36.55.

Table 5.3: Paperwork Reduction Act

Provision	Reg Section	OMB Control Number	Hours	Costs
Issuing and Challenging D/E Rates	668.405	OMB 1845-0123	80,670	2,948,481
Alternate Earnings Appeal	668.406	OMB 1845-0122	9,736	355,857
Consequences of GE Measures	668.410	OMB 1845-0123	427,091	15,610,175
Reporting Requirements of GE Programs	668.411	OMB 1845-0123	475,424	17,376,731
Disclosure Requirements for GE Programs	668.412	OMB 1845-0123	748,282	27,349,710
Calculating, Issuing, and Challenging Completion, Withdrawal, and Repayment Rates	668.413	OMB 1845-0123	203,147	7,425,023
Certification and Application Requirement for GE Programs	668.414	OMB 1845-0123	665	24,323
Draft Program Cohort Default Rates and Challenges	668.504	OMB 1845-0121	2,052	74,998
Program CDR - Uncorrected Data Adjustments	668.509	OMB 1845-0121	129	4,726
Program CDR - New Data Adjustments	668.510	OMB 1845-0121	31	1,143
Program CDR -Erroneous Data Appeals	668.511	OMB 1845-0121	-	-
Program CDR -Loan Servicing Appeals	668.512	OMB 1845-0121	45	1,649
Total			1,947,273	71,172,816

Identification, to the Extent Practicable, of All Relevant Federal Regulations that May Duplicate, Overlap, or Conflict with the Regulations

The regulations are unlikely to conflict with or duplicate existing Federal regulations. Under existing law and regulations, institutions are required to disclose data in a number of areas related to the regulations.

Alternatives Considered

As previously described, we evaluated several alternative provisions for the regulations and their effect on different types of institutions, including small entities. As discussed in “Regulatory Alternatives Considered,” several different approaches were analyzed, including, regarding the D/E rates measure, the use of different interest rates, amortization periods, and minimum n-size for programs to be evaluated, and additional or alternative metrics such as pCDR, placement rates, pre- and post-program earnings comparison, and a negative amortization test. These alternatives are not specifically targeted at small entities, but the n-size alternative of 10 students completing a program may have had a larger effect on programs at small entities.

1 Please see “§668.404 Calculating D/E Rates” for details about the calculation of the D/E rates.

2 Please see §668.404(a)(1) for the definition of the discretionary income rate.

3 Please see §668.404(a)(2) for the definition of the annual earnings rate.

4 Please see the “Analysis of the Regulations: Methodology for pCDR Calculations” in the Regulatory Impact Analysis.

5 NPSAS:2012

6 Schneider, M. (2014). American Enterprise Institute. Are Graduates from Public Universities Gainfully Employed? Analyzing Student Loan Debt and Gainful Employment.

7 Kantrowitz, M. (2014). Edvisors Network Inc., Student Aid Policy Analysis. U.S. Department of Education Proposes Stricter Gainful Employment Rule.

8 National Postsecondary Student Aid Study (NPSAS) 2012. Unpublished analysis of restricted-use data.

9 Postsecondary Education: Student Outcomes Vary at For-Profit, Nonprofit, and Public Schools (GAO-12-143), GAO, December 7, 2011.

10 “For Profit Higher Education: The Failure to Safeguard the Federal Investment and Ensure Student Success,” Senate HELP Committee, July 30, 2012.

11 National Center for Education Statistics (NCES) (2014). Digest of Education Statistics (Table 222). Available at: http://nces.ed.gov/programs/digest/d12/tables/dt12_222.asp. This table provides evidence of the growth in fall enrollment. For evidence of the growth in the number of institutions, please see the Digest of Education Statistics (Table 306) available at http://nces.ed.gov/programs/digest/d12/tables/dt12_306.asp.

12 Deming, D., Goldin, C., and Katz, L. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

13 U.S. Department of Education, Federal Student Aid, Title IV Program Volume Reports, available at https://studentaid.ed.gov/about/data-center/student/title-iv. The Department calculated the percentage of Federal Grants and FFEL and Direct student loans (excluding Parent PLUS) originated at for-profit institutions (including foreign) for award year 2000-2001 and award year 2013-2014.

14 Deming, D., Goldin, C., and Katz, L. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

15 Id.

16 Id.

17 Id.

18 Keller, J. (2011, January 13). Facing new cuts, California's colleges are shrinking their enrollments. Chronicle of Higher Education. Retrieved from http://chronicle.com/article/Facing-New-Cuts-

Californias/125945/.

19 Cellini, S. R. (2009). Crowded Colleges and College Crowd-Out: The Impact of Public Subsidies on the Two-Year College Market. American Economic Journal: Economic Policy, 1(2): 1-30.

20 Deming, D.J., Goldin, C., and Katz, L.F. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

21 Apollo Group, Inc. (2013). Form 10-K for the fiscal year ended August 31, 2013. Available at www.sec.gov/Archives/edgar/data/929887/000092988713000150/apol-aug312013x10k.htm.

22

23 Id.

24 NCES, Digest of Education Statistics 2013 (Table 318.40) available at http://nces.ed.gov/programs/digest/d13/tables/dt13_318.40.asp. Indicates that in 2011-12, of 855,562 degrees and certificates awarded at for-profit institutions, approximately 75% (637,565)_were certificates or associate degrees. At public institutions in 2011-12, approximately 45%, or 1,280,470 of 2,846,394 degrees and certificates awarded, were certificates or associate degrees.

25 Id.

26 Cellini, S. R., and Darolia, R. (2013). College Costs and Financial Constraints: Student Borrowing at For-Profit Institutions. Unpublished manuscript. Available at www.upjohn.org/stuloanconf/Cellini_Darolia.pdf.

27 Id.

28 Deming, D.J., Goldin, C., and Katz, L.F. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

29 National Postsecondary Student Aid Study (NPSAS) 2012. Unpublished analysis of restricted-use data using the NCES PowerStats tool available at http://nces.ed.gov/datalab/postsecondary/index.aspx.

30 Id.

31

32

33 Darolia, R. (2013). Student Loan Repayment and College Accountability. Federal Reserve Bank of Philadelphia.

34 Deming, D.J., Goldin, C., and Katz, L.F. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

35 Based on the Department’s analysis of the three-year cohort default rates for fiscal year 2011, U.S. Department of Education, available at http://www2.ed.gov/offices/OSFAP/defaultmanagement/schooltyperates.pdf.

36 Federal Student Aid, Default Rates for Cohort Years 2007-2011, www.ifap.ed.gov/eannouncements/attachments/060614DefaultRatesforCohortYears20072011.pdf.

37 Postsecondary Education: Student Outcomes Vary at For-Profit, Nonprofit, and Public Schools (GAO-12-143), GAO, December 7, 2011.

38 For Profit Higher Education: The Failure to Safeguard the Federal Investment and Ensure Student Success, Senate HELP Committee, July 30, 2012.

39 Id.

40 Id.

41 The commenter suggests that the fact that the report was not “voted on” by the committee renders the report suspect. The commenter cites no rule that requires reports issued “by the committee” or even by committee staff to be voted on. The report states that it is “Prepared by the Committee on Health, Education, Labor, and Pensions, United States Senate.” S. Prt. No. 112-37. Because no bill accompanied the report, it is not clear why any vote would be in order.

42 We cite findings in the HELP report in three paragraphs on two pages of the preamble of the NPRM. 79 FR 16434, 16435 (virtually identical language is repeated in the Regulatory Impact Analysis at 79 FR 16937, 16938). Two of those paragraphs also cite to the GAO report. We note that the same commenter asserts that Congress has already “addressed” these abuses by banning incentive compensation for recruiters, proscriptions that an industry trade group has vigorously opposed in litigation. APSCU v. Duncan, 681 F.3d 427 (D.C. Cir. 2012).

43 Id.

44 “Students Attending For-Profit Postsecondary Institutions: Demographics, Enrollment Characteristics, and 6-Year Outcomes” (NCES 2012-173). Available at: http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2012173

45 Id.

46

47

48 NCES, “Transferability of Postsecondary Credit Following Student Transfer or Coenrollment,” NCES 2014-163, table 8.

49 Cellini, S. R., and Darolia, R. (2013). College Costs and Financial Constraints: Student Borrowing at For-Profit Institutions. Unpublished manuscript. Available at www.upjohn.org/stuloanconf/Cellini_Darolia.pdf.

50 Darolia, R. (2013). Student Loan Repayment and College Accountability. Federal Reserve Bank of Philadelphia.

51 Cellini, S. R., and Darolia, R. (2013). College Costs and Financial Constraints: Student Borrowing at For-Profit Institutions. Unpublished manuscript. Available at www.upjohn.org/stuloanconf/Cellini_Darolia.pdf.

52 Lang, K., and Weinstein, R. (2013). “The Wage Effects of Not-for-Profit and For-Profit Certifications: Better Data, Somewhat Different Results.” NBER Working Paper.

53 Deming, D., Goldin, C., and Katz, L. The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, vol. 26, no. 1, Winter 2012.

54

55 For-Profit Colleges: Undercover Testing Finds Colleges Encouraged Fraud and Engaged in Deceptive and Questionable Marketing Practices (GAO-10-948T), GAO, August 4, 2010 (reissued November 30, 2010).

56 Id.

57 For Profit Higher Education: The Failure to Safeguard the Federal Investment and Ensure Student Success, Senate HELP Committee, July 30, 2012.

58 Id.

59 “U.S. to Join Suit Against For-Profit College Chain,” The New York Times, May 2, 2011. Available at: http://www.nytimes.com/2011/05/03/education/03edmc.html?_r=0.

60 “A.G. Schneiderman Announces Groundbreaking $10.25 Million Dollar Settlement with For-Profit Education Company That Inflated Job Placement Rates to Attract Students,” press release, Aug. 19, 2013. Available at: www.ag.ny.gov/press-release/ag-schneiderman-announces-groundbreaking-1025-million-dollar-settlement-profit.

61 “Attorneys General Take Aim at For-Profit Colleges’ Institutional Loan Programs,” The Chronicle of Higher Education, March 20, 2012. Available at: http://chronicle.com/article/Attorneys-General-Take-Aim-at/131254/.

62 “Kentucky Showdown,” Inside Higher Ed, Nov. 3, 2011. Available at: www.insidehighered.com/news/2011/11/03/ky-attorney-general-jack-conway-battles-profits.

63 “For Profit Colleges Face New Wave of State Investigations,” Bloomberg, Jan. 29, 2014. Available at: www.bloomberg.com/news/2014-01-29/for-profit-colleges-face-new-wave-of-coordinated-state-probes.html.

64 Id.

65 “Corinthian Colleges Crumbles 14% on SEC probe,” Fox Business, June 11, 2013. Available at: www.foxbusiness.com/government/2013/06/11/corinthian-colleges-crumbles-14-on-sec-probe/.

66 U.S. Department of Education, Press Release, “Education Department Names Seasoned Team to Monitor Corinthian Colleges,” July 18, 2014. Available at: www.ed.gov/news/press-releases/education-department-names-seasoned-team-monitor-corinthian-colleges.

67 Cellini, S. R. (2012). For Profit Higher Education: An Assessment of Costs and Benefits. National Tax Journal, 65 (1): 153-180.

68 Guryan, J., and Thompson, M. Charles River Associates. (2014). Report on the Proposed Gainful Employment Regulation.

69 Guryan, J., and Thompson, M. Charles River Associates. (2014). Report on the Proposed Gainful Employment Regulation.

70 Please note that race and ethnicity status was derived from data reported by institutions to IPEDS. See the Regulatory Impact Analysis for additional details on methodology.

71 Bowen, W, Chingos, M., and McPherson, M. Crossing the Finish Line: Completing College at America's Public Universities. Princeton, NJ: Princeton UP, 2009.

72 Bound, J., Lovenheim, M., and Turner, S. 2007. “Understanding the Decrease in College Completion Rates and the Increased Time to the Baccalaureate Degree.” PSC Research Report No. 07-626. November 2007.

73 U.S. Senate, Committee on Health, Education, Labor and Pensions, For Profit Higher Education: The Failure to Safeguard the Federal Investment and Ensure Student Success, Washington: Government Printing Office, July 30, 2012.

74 Gicheva, D. “In Debt and Alone? Examining the Causal Link between Student Loans and Marriage.” Working Paper (2012).

75 Gicheva, D., and U. N. C. Greensboro. “The Effects of Student Loans on Long-Term Household Financial Stability.” Working Paper (2013).

76 Id.

77 Shand, J. M. (2007). “The Impact of Early-Life Debt on the Homeownership Rates of Young Households: An Empirical Investigation.” Federal Deposit Insurance Corporation Center for Financial Research.

78 Brown, M., and Sydnee, C. (2013). Young Student Loan Borrowers Retreat from Housing and Auto Markets. Liberty Street Economics, retrieved from: http://libertystreeteconomics.newyorkfed.org/2013/04/young-student-loan-borrowers-retreat-from-housing-and-auto-markets.html.

79 https://studentaid.ed.gov/repay-loans/default.

80 www.asa.org/for-students/student-loans/managing-default/

81 Avery, C., and Turner, S. (2013). Student Loans: Do College Students Borrow Too Much-Or Not Enough? Journal of Economic Perspectives, 26(1), 165-192.

82 Moretti, E. (2004). Estimating the Social Return to Higher Education: Evidence from Longitudinal and Repeated Cross-Sectional Data. Journal of Econometrics, 121(1), 175-212.

83 Kane, Thomas J., and Rouse, C. E. (1995). Labor Market Returns to Two- and Four-Year College. The American Economic Review, 85 (3), 600-614.

84 Cellini, Stephanie R. and Chaudhary, L. (2012). “The Labor Market Returns to For-Profit College Education.” Working paper.

85 Baum, S., Ma, J., and Payea, K. (2013) “Education Pays 2013: The Benefits of Education to Individuals and Society” College Board. Available at http://trends.collegeboard.org/.

86 Avery, C., and Turner, S. (2013). Student Loans: Do College Students Borrow Too Much-Or Not Enough? Journal of Economic Perspectives, 26(1), 165-192.

87

Baum, S. & O’Malley, M. (2003). College on credit: How borrowers perceive their education debt. Results of the 2002 National Loan Survey (Final Report). Braintree, MA: Nellie Mae.

88 NCES, “Degrees of Debt,” NCES 2014-11.

89 Avery, C., and Turner, S. (2013). Student Loans: Do College Students Borrow Too Much-Or Not Enough? Journal of Economic Perspectives, 26(1), 165-192.

90 Kantrowitz, M. (2010). Finaid.com. What is Gainful Employment? What is Affordable Debt?, available at www.finaid.org/educators/20100301gainfulemployment.pdf.

91 Baum, S., and Schwartz, S. (2006). How Much Debt is Too Much? Defining Benchmarks for Managing Student Debt.

92 The GAO report was not undertaken to determine acceptable debt burdens, but rather, as stated in the report, “to determine how often students who were federal financial aid recipients received aid that was greater than their federally defined financial need.” GAO-03-508 at 19. The report contains neither an analysis of debt burden nor reference to the 10 percent debt burden rate as a “generally-agreed upon” standard; the GAO report merely cites, without comment, the 10 percent figure as a Department performance indicator.

93 The Department used the 10 percent debt/income indicator without elaboration. The stated purpose of the indicator was for the Department to assess its own progress in meeting certain standards, including the debt-to-earnings ratios of students. See page 165, available at http://www2.ed.gov/about/reports/annual/2002report/index.html.

94 Healthcare and Education Reconciliation Act of 2010, PL 111-152, §2213, March 30, 2010, 124 Stat 1029, 1081.

95 Indeed, in the notice of proposed rulemaking for the 2011 Prior Rule, the Department proposed counting the full amount of loan debt for calculating the debt-to-earnings ratios. 75 FR 43639. In response to comments, in the 2011 Prior Rule, the Department capped the loan debt at the lesser of tuition and fees or the total amount borrowed. 76 FR 34450.

96 See, e.g., Kantrowitz, M. (2010). Finaid.com. What is Gainful Employment? What is Affordable Debt?, available at www.finaid.org/educators/20100301gainfulemployment.pdf. The article addresses the proposed standard included in the notice of proposed rulemaking for the 2011 Prior Rule, which included all debt, and states “The most common standards promoted by personal finance experts are 10% and 15% of [gross] income.” At 10.

97 Baum, S., and Schwartz, S. (2006). How Much Debt is Too Much? Defining Benchmarks for Managing Student Debt. See also S. Baum, “Gainful Employment,” posting to The Chronicle of Higher Education, http://chronicle.com/blogs/innovations/gainful-employment/26770, in which Baum described the 2006 study: 

This paper traced the history of the long-time rule of thumb that students who had to pay more than 8% of their incomes for student loans might face difficulties and looked for better guidelines. It concluded that manageable payment-to-income ratios increase with incomes, but that no former student should have to pay more than 20% of their discretionary income for all student loans from all sources.

98 Id.

99 Id. at 2-3.

100 Greiner, K. (1996). How Much Student Loan Debt Is Too Much? Journal of Student Financial Aid, 26(1), 7–19.

101 Scherschel, P. (1998). Student Indebtedness: Are Borrowers Pushing the Limits? USA Group Foundation.

102 Harrast, S.A. (2004). Undergraduate Borrowing: A Study of Debtor Students and their Ability to Retire Undergraduate Loans. NASFAA Journal of Student Financial Aid, 34(1), 21–37.

103 King, T., & Frishberg, I. (2001). Big Loans, Bigger Problems: A Report on the Sticker Shock of Student Loans. Washington, DC: The State PIRG’s Higher Education Project. Available at www.pirg.org/highered/highered.asp?id2=7973.

104 Illinois Student Assistance Commission (2001). Increasing College Access . . . or Just Increasing Debt? A Discussion about Raising Student Loan Limits and the Impact on Illinois Students.

105 Baum, S., and O’Malley, M. (2002, February 6). College on Credit: How Borrowers Perceive their Education Debt: Results of the 2002 National Student Loan Survey. Final Report. Braintree, MA: Nellie Mae Corporation.

106 Baum, S., and Schwartz, S. (2006). How Much Debt is Too Much? Defining Benchmarks for Managing Student Debt.

107 Id., at 3.

108 Id., at 12, Table 10

109 FHA, Risk Management Initiatives: New Manual Underwriting Requirements, 78 FR 75238, 75239 (December 11, 2013).

110 Vornovytskyy, M., Gottschalck, A., and Smith, A., Household Debt in the U.S.: 2000 to 2011, U.S. Census Bureau, Survey of Income and Program Participation Panels. Available at www.census.gov/people/wealth/files/Debt%20Highlights%202011.pdf. Table A-2 shows that median credit card debt of households under 35 years of age as of 2011 was $3,000, and median other unsecured debt for that same cohort, including student loans and other unsecured debt, was $13,000. The “other” debt accounts for 81 percent of unsecured household debt. Assuming that the lending standards described here allocate 12 percent to non-housing debt, and 81 percent of that allocation is 9.75 percent allocable to non-credit card debt, which includes student loan debt, the 8 percent annual earnings rate appears to fall within this range.

111 Bricker, J., Kennickell, A., Moore, K., and Sabelhaus, J. (2012). “Changes in U.S. Family Finances from 2007 to 2010: Evidence from the Survey of Consumer Finances,” Federal Reserve Bulletin, 98(2). Available at www.federalreserve.gov/pubs/bulletin/2012/pdf/scf12.pdf.

112 Vornovytskyy, M., Gottschalck, A., and Smith, A., Household Debt in the U.S.: 2000 to 2011, U.S. Census Bureau, Survey of Income and Program Participation Panels. Available at www.census.gov/people/wealth/files/Debt%20Highlights%202011.pdf. Table A-2 shows that median credit card debt of households under 35 years of age as of 2011 was $3,000, and median other unsecured debt for that same cohort, including student loans and other unsecured debt, was $13,000. The “other” debt accounts for 81 percent of unsecured household debt. Assuming that the lending standards described here allocate 12 percent to non-housing debt, and 81 percent of that allocation is 9.75 percent allocable to non-credit card debt, which includes student loan debt, the 8 percent annual earnings rate appears to fall within this range.

113 2012 GE informational D/E rates.

114 Id.

115 Id.

116 Id.

117 National Bureau of Economic Research (2014), US Business Cycle Expansions and Contractions, available at www.nber.org/cycles.html.

118 Baum, S., and Schwartz, S. (2006). How Much Debt Is Too Much? Defining Benchmarks for Managing Student Debt.

119 Greiner, K. (1996). How Much Student Loan Debt Is Too Much? Journal of Student Financial Aid, 26(1), 7–19.

120 Scherschel, P. (1998). Student Indebtedness: Are Borrowers Pushing the Limits? USA Group Foundation.

121 Harrast, S.A. (2004). Undergraduate Borrowing: A Study of Debtor Students and their Ability to Retire Undergraduate Loans. NASFAA Journal of Student Financial Aid, 34(1), 21–37.

122 King, T., & Frishberg, I. (2001). Big Loans, Bigger Problems: A Report on the Sticker Shock of Student Loans. Washington, DC: The State PIRG’s Higher Education Project. Available at www.pirg.org/highered/highered.asp?id2=7973.

123 Kantrowitz, M. (2010). Finaid.com. What is Gainful Employment? What is Affordable Debt?, available at www.finaid.org/educators/20100301gainfulemployment.pdf.

124 NCES, “Degrees of Debt,” NCES 2014-11.

125 NCES, Degrees of Debt (2014). See Figure 1 for percent of bachelor’s degree recipients who did not borrow and Figure 7 for the ratio of monthly loan payments to monthly income. The analysis uses data from U.S. Department of Education, National Center for Education Statistics, 1993/94, 2000/01, and 2008/09 Baccalaureate and Beyond Longitudinal Studies (B&B:93/94, B&B:2000/01, and B&B:08/09).

126 Cellini, S., and Chaudhary, L. (2012). “The Labor Market Returns to For-Profit College Education.” Working paper.

127 Kane, T., and Rouse, C. E. (1995). Labor Market Returns to Two- and Four-Year College. The American Economic Review, 85(3), 600-614.

128 Avery, C., and Turner, S. (2013). Student Loans: Do College Students Borrow Too Much-Or Not Enough? Journal of Economic Perspectives, 26(1), 165-192.

129 Deming, D., Goldin, C., and Katz, L. (2013). For Profit Colleges. Future of Children, 23(1), 137-164.

130 Schneider, M. (2014). American Enterprise Institute. Are Graduates from public Universities Gainfully Employed? Analyzing Student Loan Debt and Gainful Employment.

131 National Bureau of Economic Research (2014). US Business Cycle Expansions and Contractions, available at www.nber.org/cycles.html.

132 National Bureau of Economic Research (2014). U.S. Business Cycle Expansions and Contractions, available at www.nber.org/cycles.html.

133 The two-digit CIP code, 13, is the classification for the education programs including Early Childhood Education and Training, Elementary Education and Teaching, and many other types of programs related to education.

134 Available at http://aspe.hhs.gov/poverty/faq.cfm.

135 We note that, because the D/E rates are calculated based on a 100 percent sample of the students in the cohort, the median of debt is the value at the 50^th percentile (i.e., the midpoint of the distribution of debt) and the values on either side of the median do not influence the value of the median.

136 Available at www.ifap.ed.gov/fsahandbook/attachments/1415FSAHbkVol3Ch2.pdf.

137 Department of Education analysis of NSLDS data.

138 The best private student loans will have interest rates of LIBOR + 2.0% or PRIME - 0.50% with no fees. Such loans will be competitive with the Federal PLUS Loan. Unfortunately, these rates often will be available only to borrowers with good credit who also have a creditworthy cosigner. It is unclear how many borrowers qualify for the best rates, although the top credit tier typically encompasses about 20 percent of borrowers. See Private Student Loans, Finaid.Org, available at www.finaid.org/loans/privatestudentloans.phtml.

139 Id.

140 Private Student Loans, Finaid.Org, available at www.finaid.org/loans/privatestudentloans.phtml.

141 We are unable to provide more precise probabilities for the scenario of a program that fails the D/E rates measure in two out of three years. Because some students are common to consecutive two-year cohort periods for the D/E rates calculations, we cannot rely on the assumption that each year's D/E rates are statistically independent from the previous and subsequent year’s D/E rates. Without the assumption of independence between years, there is no widely accepted method for calculating the probability of a program failing the D/E rates measure in two out of three years.

142 United States Government Accountability Office, “For Profit Schools: Large Schools and Schools that Specialize in Healthcare Are More Likely to Rely Heavily on Federal Student Aid,” October 2010, available at www.gao.gov/new.items/d114.pdf.

143 Available at http://federalstudentaid.ed.gov/datacenter/proprietary.html.

144 Introduction To State And Local Coverage And Section 218, available at www.ssa.gov/section218training/basic_course_4.htm#8.

145 Office of Data Exchange and Policy Publications, SSA; see 2014 General Instructions for Forms W-2 and W-3, Department of Treasury, Internal Revenue Service, December 17, 2013, available at www.irs.gov/pub/irs-pdf/iw2w3.pdf.

146 Office of Data Exchange and Policy Publications, SSA

147 Internal Revenue Service, Wage Compensation for S Corporation Officers, FS-2008-25, August 2008, available at www.irs.gov/uac/Wage-Compensation-for-S-Corporation-Officers.

148 “Approximately 90 percent of the wage reports received by SSA each year are posted to the MEF without difficulty. After the computerized routines are applied, approximately 96 percent of wage items are successfully posted to the MEF (GAO 2005).” Anya Olsen and Russell Hudson. “Social Security Administration’s Master Earnings File: Background Information.” Social Security Bulletin, Vol. 69, No. 3, 2009, www.ssa.gov/policy/docs/ssb/v69n3/v69n3p29.html.

Social Security Administration's Master Earnings File: Background Information, available at www.ssa.gov/policy/docs/ssb/v69n3/v69n3p29.html.

149 “In previous reports, SSA acknowledged that unauthorized noncitizens’ intentional misuse of SSNs has been a major contributor to the ESF’s growth.” Employers Who Report Wages with Significant Errors in the Employee Name and SSN (A-08-12-13036), Office of Inspector General, Department of Health and Human Services, at 4.

150 Source: internal programming statistics, SSA, Office of Deputy Commissioner for Systems; see also Johnson, M., Growth of the Social Security Earnings Suspense File Points to the Rising Potential Cost of Unauthorized Work To Social Security, The Senior Citizens League, Feb. 2013, table 2, available at http://seniorsleague.org/2013/growth-of-the-social-security-earnings-suspense-file-points-to-the-rising-potential-cost-of-unauthorized-work-to-social-security-2/.

151 Household Employer’s Tax Guide, IRS Publication 926, available at www.irs.gov/publications/p926/ar02.html#en_US_2014_publink100086732.

152 Source: ED records from response files received from SSA as refined based on additional SSA explanations of its exclusion from verified individuals of those verified individuals whose records show an indication that the wage earner died. Where an exchange consisted of multiple component data sets, each has been listed separately and then totaled. Data on all but the first of these exchanges was provided to the commenter pursuant to a FOIA request.

153 BLS, Databases, Tables & Calculators by Subject, available at http://data.bls.gov/timeseries/LNU04000000?years_option=all_years&periods_option=specific_periods&periods=Annual+Data.

154 BLS, Labor Force Statistics from the Current Population Survey, Frequently Asked Questions, available at www.bls.gov/cps/faq.htm#Ques5.

155 NCES, Unemployment rates of persons 16 to 64 years old, by age group and educational attainment: Selected years, 1975 through 2013 (derived from BLS, Office of Employment and Unemployment Statistics, unpublished annual average data from the Current Population Survey (CPS), selected years, 1975 through 2013), available at http://nces.ed.gov/programs/digest/d13/tables/dt13_501.80.asp.

For the purposes of this report:

The unemployment rate is the percentage of persons in the civilian labor force who are not working and who made specific efforts to find employment sometime during the prior 4 weeks. The civilian labor force consists of all civilians who are employed or seeking employment.

156 Mark Kantrowitz, Student Aid Policy Analysis - Analysis of FY2011 Gainful Employment Data, July 13, 2012, available at www.finaid.org/educators/20120713gainfulemploymentdata.pdf.

157 The duration of unemployment for those unemployed during 2010 and 2011 grew as well: 15.3 percent of those unemployed who found work during 2010, and 13.8 percent of the unemployed who found work during 2011, had been unemployed for 27 to 52 weeks [; in addition, of those unemployed who found work during 2010, 11 percent had been unemployed for a year or more, and of those reemployed during 2011, 12.9 percent had been unemployed for a year or more. Ilg, Randy E., and Theodossiou, Eleni, Job search of the unemployed by duration of unemployment, Monthly Labor Review, March 2012, available at www.bls.gov/opub/mlr/2012/03/art3full.pdf.

158 IRS Guidance, Reporting Tip Income-Restaurant Tax Tips, available at www.irs.gov/Businesses/Small-Businesses-&-Self-Employed/Reporting-Tip-Income-Restaurant-Tax-Tips (“Tips your employees receive from customers are generally subject to withholding. Employees are required to claim all tip income received. This includes tips you paid over to the employee for charge customers and tips the employee received directly from customers. . . . Employees must report tip income on Form 4070, Employee's Report of Tips to Employer, (PDF) or on a similar statement. This report is due on the 10th day of the month after the month the tips are received. . . . No report is required from an employee for months when tips are less than $20.”).

159 See Baystate Medical Center v. Leavitt, 545 F.Supp.2d 20 (D.D.C. 2008), on which the commenter chiefly relies, describes the “repeated recognition in case law that the agency must use ‘the most reliable data available’ to produce figures that can be considered sufficiently ‘accurate.’” Baystate, 545 F.Supp.2d at 41 (citation omitted). The accuracy of the determination “cannot be weighed in a vacuum, but instead must be evaluated by reference to the data that was available to the agency at the relevant time.” Id. An agency that used the most reliable data available in making a determination need not “recalculate” based on “subsequently corrected data” or where, for instance, “the data failed to account for part-time workers.” Id. (internal citations omitted).

160 See: SSA, Annual benefits paid from the OASI Trust Fund, by type of benefit, calendar years 1937-2013, available at www.ssa.gov/oact/STATS/table4a5.html; The Board of Trustees, Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, 2014 Annual Report, available at

www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/ReportsTrustFunds/downloads/tr2014.pdf.

161 The commenters do not challenge the regulations by contending that they could be read to bar a challenge based on actual return information were the institution able to secure such information by, for example, obtaining copies of IRS earnings records with the consent of each of the students in the cohort. This option would be highly impractical, however, and therefore we did not consider it to be viable for purposes of these regulations.  We also are unaware of any comments that suggested that we adopt such an option. 

162 www.census.gov/cps/methodology/unreported.html

163 www.bls.gov/news.release/ocwage.tn.htm

164 http://nces.ed.gov/statprog/2002/glossary.asp#cross-sectional

165 For example, BLS uses these data to produce the occupational earnings analysis that the Department does not now consider to be a sufficiently precise measure to justify its continued use as a source of earnings for the purpose of calculating D/E rates, for the reasons already explained.

166“The establishment of orders for child support enforcement cases . . . occurs through either judicial or administrative processes. . . . In 30 States, imputation is practiced if the non-custodial parent fails to provide relevant information or is currently unemployed or underemployed. Five States impute income only if the non-custodial parent fails to provide relevant information such as pay stubs, income tax returns or financial affidavits. Thirteen States impute income only if the non-custodial parent is unemployed or underemployed.

Most of the 48 States that impute income consider a combination of factors in determining the amount of income to be imputed to the non-custodial parent. Thirty-five States base imputed awards on the premise that the non-custodial parent should be able to work a minimum wage job for 40 hours per week. Fifteen of the States consider the area wage rate and 10 of the States look at the area employment rate to determine imputed income. Seventeen States consider the non-custodial parent’s level of education while 14 account for disabilities hindering full employment. Thirty-five States evaluate the non-custodial parent’s skills and experience and thirty-one base imputations on most recent employment, where information is available.”

Office of Inspector General, Department of Health and Human Services (2000), State policies used to establish child support orders for low-income non-custodial parents, at 5, 15. Available at https://oig.hhs.gov/oei/reports/oei-05-99-00391.pdfhttps://oig.hhs.gov/oei/reports/oei-05-99-00391.pdf.

167 National Conference of State Legislatures, Child Support Digest (Volume 1, Number 3) www.ncsl.org/research/military-and-veterans-affairs/child-support-digest-volume-1-number-3.aspx.

168 In order to impute income to a parent who has demonstrated an inability to pay the specified amount, courts must determine that the party is voluntarily unemployed or underemployed. States allow for exceptions to the general rule regarding voluntary income decreases if the party can demonstrate that the decrease was based on a “good faith reason” (e.g., taking a lower paying job that has greater long-term job security and potential for future earnings).

National Conference of State Legislatures, Child Support 101.2: Establishing and modifying support orders , available at www.NCSL.Org/research/human-services/enforcement-establishing-and-modifying-orders.aspx.

169 Disclosures may be “appropriately required ... in order to dissipate the possibility of consumer confusion or deception.” Zauderer v. Office of Disciplinary Counsel of Supreme Court of Ohio, 471 U.S. 626, 651 (1985). If a requirement is “directed at misleading commercial speech and imposes only a disclosure requirement rather than an affirmative limitation on speech, the less exacting scrutiny set out in Zauderer v. Office of Disciplinary Counsel of Supreme Court of Ohio, 471 U.S. 626, 105 S.Ct. 2265, 85 L.Ed.2d 652, governs.” Milavetz, Gallop & Milavetz, P.A. v. United States, 559 U.S. 229, 230 2010).

170 Section 485 of the HEA was enacted in 1980 and has been repeatedly amended, most recently in 2013. Section 485 requires an institution to disclose to employees and current and prospective students myriad details regarding campus security policies and statistics on crimes committed on or near campuses, 20 U.S.C. 1092(f); statistics regarding the number and costs of, and revenue from, its athletic programs, 20 U.S.C. 1092(g); and some 23 categories of information about the educational programs and student outcomes, including disclosures of some of the very kinds of information--for the institution as a whole--as required for GE programs in these regulations, including completion rate, placement rate, and retention rate. 20 U.S.C. 1092(a)(1)(L), (R), (U). Not only does the HEA require these disclosures, but the HEA also specifies the manner in which the rate is to be calculated. See, e.g., 20 U.S.C. 1092(a)(3) (completion rate). These disclosures must be made through various media, including “electronic media.” See 20 U.S.C. 1092(a)(1). In addition, section 487(a)(8) of the HEA requires an institution that advertises job placement rates as a means of attracting students to enroll to make available to prospective students “the most recent data concerning employment statistics, graduation statistics, and any other information necessary to substantiate the truthfulness of the advertisements.” 20 U.S.C. 1094(a)(8).

171 The warning is required only after the Department has issued a notice of determination informing the institution of its final D/E rates and that the institution is subject to the student warning requirements. That determination is the outcome of an administrative appeal process and, as final agency action, is subject to review by a Federal court under the Administrative Procedure Act. By the time the warning is required, therefore, the institution’s opportunity to controvert the determination is over.

172 The congressional findings state that “education is fundamental to the development of individual citizens and the progress of the Nation as a whole” and that student consumers and their parents must be able to obtain information to make an “informed judgment about the educational benefits available at a given institution.” Pub. L. 101–542, §102, November 8, 1990, 104 Stat 2381.

173 Congress demonstrated this most recently in Pub. L. 110-315, §110, August 8, 2008, enacting section 132 of the Higher Education Act, which in subsection (h) requires institutions to disclose on their own Web sites a “net price calculator” regarding their programs, while subsection (a) requires the Department to implement a “College Navigator” Web site displaying a wide range of data, including some similar data. 20 U.S.C. 1015a(a), (h). In that same law, Congress also amended section 485 of the HEA to add at least seven new disclosures to those already required of the institution itself. 20 U.S.C. 1085(a), as amended by Pub. L. 110-315, §488(a), 122 Stat. 3293.

174 The regulations also require even more detailed counseling by the institution for students exiting the institution. 34 CFR 685.304(b).

175 See, e.g., 16 CFR 14.9, Requirements concerning clear and conspicuous disclosures in foreign language advertising and sales materials: Where “clear and conspicuous disclosures are required,” the disclosure shall appear in the “predominant language of the publication in which the advertisement or sales material appears.” See also FTC Final Rule, Free Annual File Disclosures, 75 FR 9726, 9733 (Mar. 3, 2010) (noting “the Commission’s belief that a disclosure in a language different from that which is principally used in an advertisement would be deceptive”).

176 See For-Profit Colleges: Undercover Testing Finds Colleges Encouraged Fraud and Engaged in Deceptive and Questionable Marketing Practices (GAO-10-948T), GAO, August 4, 2010 (reissued November 30, 2010); For Profit Higher Education: The Failure to Safeguard the Federal Investment and Ensure Student Success, Senate HELP Committee, July 30, 2012.

177 As noted in the NPRM, the Department has previously expressly interpreted section 437(c) of the HEA in controlling regulations to provide no relief for a claim that the loan was arranged for enrollment in an institution that was ineligible, or that the institution arranged the loan for enrollment in an “ineligible program.” 34 CFR 682.402(e); 59 FR 22462, 22470 (April 29, 1994), 59 FR 2486, 2490 (Jan. 14, 1994).

178 Loans and grants are treated similarly, but slightly differently, under §668.26(d). With respect to Direct Loans, the loss of eligibility will be expected to occur during a “period of enrollment”--a term defined as a period that--

must coincide with one or more bona fide academic terms established by the school for which institutional charges are generally assessed (e.g., a semester, trimester, or quarter in weeks of instructional time; an academic year; or the length of the program of study in weeks of instructional time).

34 CFR 685.102. The period of enrollment is referred to as the “loan period.” The maximum period for which a Direct Loan may be made is an academic year, 34 CFR §685.203, and therefore the “loan period” for a loan cannot exceed an academic year even if the program of study is longer than an academic year. Section 668.26(d)(3) limits the disbursements that may be made after loss of eligibility to those made on “a loan,” if all of the following conditions are met: the borrower must be enrolled on the date on which eligibility is lost; the loss of eligibility must take place during a loan period; a first disbursement on the loan has already been made before the date on which eligibility is lost; and the institution must continue to provide training in that GE program at least through the scheduled completion date of the academic year for which the loan was scheduled, or the length of the program, whichever falls earlier. With respect to Pell Grants, the institution may disburse Pell Grant funds under similar conditions: the student is enrolled on the date program eligibility ceases, the institution has already received a valid output record for the student, the requested Pell Grant is intended to be disbursed for the “payment period” [academic term or portion of a term, see: 34 CFR 668.4] during which the loss of eligibility occurs, or a prior payment period, and the institution continues to provide training in the program until at least the completion of the payment period. 34 CFR 668.26(d)(1).

179[NSLDS’s] “overall purpose” has never included the collection of information on students who do not receive and have not applied for either federal grants or federal loans. To expand it in that way would make the database no longer “a system (or a successor system) that ... was in use by the Secretary, directly or through a contractor, as of the day before August 14, 2008.” 20 U.S.C. 1015c(b)(2). The Department could not create a student unit record system of information on all students in gainful employment programs; nor can it graft such a system onto a pre-existing database of students who have applied for or received Title IV assistance. For that reason--and not, as the court previously held, because the added information is unnecessary for the operation of any Title IV program--the expansion is barred by the statutory prohibition on new databases of personally identifiable student information.

APSCU v. Duncan, 930 F.Supp.2d at 221 (emphasis added).

180 The NSLDS is currently renumbered as 18-11-06.

181 (4) Disclaimer. Estimates of an individual net price determined using a net price calculator required under paragraph (3) shall be accompanied by a clear and conspicuous notice-- (A) stating that the estimate--
(i) does not represent a final determination, or actual award, of financial assistance;
(ii) shall not be binding on the Secretary, the institution of higher education, or the State; and
(iii) may change;

20 U.S.C. 1015a(h). Institutions must also make available, directly or indirectly through Department sites, not only tuition and fees for the three most recent academic years for which data are available, but a statement of the percentage changes in those costs over that period. 20 U.S.C. 1015a(i)(5).

182 See n. 242, supra..

183 The disclosures required by §668.412(a) consist of either statistical data elements--for example, dollar amounts, ratios, time periods--and simple facts, such as whether the program meets any educational prerequisites for obtaining a license in a given State. The disclosures are required under §668.412 only after the institution itself has calculated the data, or the Department has calculated the data and given the institution an opportunity to challenge each such determination under §668.413.

184 FTC, .com Disclosures, March 2013. The FTC advises a party, when using a hyperlink to lead to a disclosure, to--

- Make the link obvious;
- Label the hyperlink appropriately to convey the importance, nature, and relevance of the information it leads to;
- Use hyperlink styles consistently, so consumers know when a link is available;
- Place the hyperlink as close as possible to the relevant information it qualifies and make it noticeable;
- Take consumers directly to the disclosure on the click-through page;
- Assess the effectiveness of the hyperlink by monitoring click-through rates and other information about consumer use and make changes accordingly.

Id. at ii. Available at www.ftc.gov/sites/default/files/attachments/press-releases/ftc-staff-revises-online-advertising-disclosure-guidelines/130312dotcomdisclosures.pdf

185 “Whether an administrative agency’s order or regulation is severable, permitting a court to affirm it in part and reverse it in part, depends on the issuing agency’s intent.”  Davis Cty. Solid Waste Mgmt. v. EPA, 108 F.3d 1454, 1459 (D.C. Cir. 1997) (quoting North Carolina v. FERC, 730 F.2d 790, 795-96 (D.C. Cir. 1984).  “Severance and affirmance of a portion of an administrative regulation is improper if there is ‘substantial doubt’ that the agency would have adopted the severed portion on its own.”  Davis, 108 F.3d at 1459.  Additionally, a court looks to whether a rule can function as designed if a portion is severed.  “Whether the offending portion of a regulation is severable depends upon the intent of the agency and upon whether the remainder of the regulation could function sensibly without the stricken provision.”  MD/DC/DE Broadcasters Ass’n v. FCC, 236 F.3d 13, 22 (D.C. Cir. 2001) (citations omitted). 

186 NSLDS.

187 Pell grant recipient percentages are based on students at undergraduate GE programs who entered repayment on title IV, HEA program loans between October 1, 2007 and September 30, 2009 and received a Pell grant for attendance at the institution between July 1, 2004 to June 30, 2009. Graduate programs are not included in calculation of Pell recipient percentages. Other percentages are based on students at GE programs who entered repayment on title IV, HEA program loans between October 1, 2007 and September 30, 2009 and had a demographic record in NSLDS in 2008. Sector and credential averages are generated by weighting program results by FY 2010 enrollment.

188 Avery, C., and Turner, S. (2013). Student Loans: Do College Students Borrow Too Much-Or Not Enough? Journal of Economic Perspectives, 26(1), 165-192.

189 Moretti, E. (2004). Estimating the Social Return to Higher Education: Evidence from Longitudinal and Repeated Cross-Sectional Data. Journal of Econometrics, 121(1), 175-212.

190 Kane, T. J., and Rouse, C. E. (1995). Labor Market Returns to Two- and Four-Year College. The American Economic Review, 85 (3), 600-614.

191 Cellini, S., and Chaudhary, L. (2012). “The Labor Market Returns to For-Profit College Education.” Working paper.

192 Baum, S., Ma, J., and Payea, K. (2013) “Education Pays 2013: The Benefits of Education to Individuals and Society” College Board. Available at http://trends.collegeboard.org/.

193 Avery, C., and Turner, S. (2013). Student Loans: Do College Students Borrow Too Much-Or Not Enough? Journal of Economic Perspectives, 26(1), 165-192.

194 At the Federal minimum wage of $7.25 per hour (www.dol.gov/whd/minimumwage.htm), an individual working 40 hours per week for 52 weeks per year would have annual earnings of $15,080.

195 2012 GE informational D/E rates.

196 2012 GE informational D/E rates. The percent of borrowers who default is calculated based on pCDR data

197 Dunlop, E. “What Do Student Loans Actually Buy You? The Effect of Stafford Loan Access on Community College Students,” Working Paper (2013).

198 Martin, A., and Andrew L., “A Generation Hobbled by the Soaring Cost of College,” New York Times, May 12, 2012.

199 Deming, D., Goldin, C., and Katz, L. (2013). For Profit Colleges. Future of Children, 23(1), 137-164.

200 Id.

201 Akers, B., and Chingos, M. (2014). Is a Student Loan Crisis on the Horizon. Brookings Institution.

202 U.S. Department of Education, Federal Student Aid Portfolio Summary, National Student Loan Data System available at https:/studentaid.ed.gov/about/data-center/student/portfolio.

203 Federal Reserve Bank of New York (2012, November). Quarterly Report on Household Debt and Credit. Retrieved from www.newyorkfed.org/research/national economy/household credit/DistrictReport_Q32012.pdf.

204 Brown, M., and Sydnee, C. (2013). Young Student Loan Borrowers Retreat from Housing and Auto Markets. Liberty Street Economics, retrieved from: http://libertystreeteconomics.newyorkfed.org/2013/04/young-student-loan-borrowers-retreat-from-housing-and-auto-markets.html.

205 Id.

206 Deming, D., Goldin, C., and Katz, L. (2013). For Profit Colleges. Future of Children, 23(1), 137-164.

207 U.S. Department of Education, Federal Student Aid, Title IV Program Volume Reports, available at https://studentaid.ed.gov/about/data-center/student/title-iv.. Stafford Loan comparison based on FFEL and Direct Loan student volume excluding Graduate PLUS loans that did not exist in 2000-01.

208 Baum, S and Payea, K. (2013). Trends in Student Aid, College Board.

209 Avery, C., and Turner, S. Student Loans: Do College Students Borrow Too Much Or Not Enough? The Journal of Economic Perspectives 26, no. 1 (2012): 189.

210 Id. at 165-192.

211 Id.

212 Gicheva, D. “Student Loans or Marriage? A Look at the Highly Educated,” Working paper(2014).

213 Gicheva, D., and U. N. C. Greensboro. “The Effects of Student Loans on Long-Term Household Financial Stability.” Working Paper (2013).

214 Id.

215 Shand, J. M. (2007). “The Impact of Early-Life Debt on the Homeownership Rates of Young Households: An Empirical Investigation.” Federal Deposit Insurance Corporation Center for Financial Research.

216 Brown, M., and Sydnee, C. (2013). Young Student Loan Borrowers Retreat from Housing and Auto Markets. Liberty Street Economics, available at: http://libertystreeteconomics.newyorkfed.org/2013/04/young-student-loan-borrowers-retreat-from-housing-and-auto-markets.html.

217 https://studentaid.ed.gov/repay-loans/default.

218 https://studentaid.ed.gov/repay-loans/default.

219 www.asa.org/in-default/consequences/.

220 U.S. Department of Education (2014). 2-year official national student loan default rates. Federal Student Aid. Retrieved from http://www2.ed.gov/offices/OSFAP/defaultmanagement/defaultrates.html.

221 Martin, A., “Debt Collectors Cashing In on Student Loans,” New York Times, September 8, 2012.

222 Gross, J. P., Cekic, O., Hossler, D., & Hillman, N. (2009). What Matters in Student Loan Default: A Review of the Research Literature. Journal of Student Financial Aid, 39(1), 19–29.

223 Lochner, L., and Monge-Naranjo, A. (2014). “Default and Repayment Among Baccalaureate Degree Earners.” NBER Working Paper No. w19882.

224 Belfield, C. R. (2013). “Student Loans and Repayment Rates: The Role of For-Profit Colleges.” Research in Higher Education, 54(1): 1-29.

225 Id.

226 Deming, D., Goldin, C., and Katz, L. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

227 Hillman, N. W. “College on Credit: A Multilevel Analysis of Student Loan Default.” The Review of Higher Education 37.2 (2014): 169-195. Project MUSE. Web. 12 Mar. 2014.

228 NCES. (2014). Digest of Education Statistics (Table 222). Available at: http://nces.ed.gov/programs/digest/d12/tables/dt12_222.asp. This table provides evidence of the growth in fall enrollment. For evidence of the growth in the number of institutions, please see the Digest of Education Statistics (Table 306) available at http://nces.ed.gov/programs/digest/d12/tables/dt12_306.asp.

229 Deming, D., Goldin, C., and Katz, L. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

230 U.S. Department of Education, Federal Student Aid, Title IV Program Volume Reports, available at https://studentaid.ed.gov/about/data-center/student/title-iv. The Department calculated the percentage of Federal Grants and FFEL and Direct student loans (excluding Parent PLUS) originated at for-profit institutions (including foreign) for award year 2000-2001 and award year 2013-2014.

231 Deming, D., Goldin, C., and Katz, L. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

232 Id.

233 Id.

234 Deming, D., Goldin, C., and Katz, L. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

235 Keller, J. (2011, January 13). Facing New Cuts, California’s Colleges Are Shrinking Their Enrollments. Chronicle of Higher Education. Available at http://chronicle.com/article/Facing-New-Cuts-Californias/125945/.

236 Cellini, S. R.. (2009). Crowded Colleges and College Crowd-Out: The Impact of Public Subsidies on the Two-Year College Market. American Economic Journal: Economic Policy, 1(2): 1-30.

237 Deming, D.J., Goldin, C., and Katz, L.F. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

238 Apollo Group, Inc. (2013). Form 10-K for the fiscal year ended August 31, 2013. Available at www.sec.gov/Archives/edgar/data/929887/000092988713000150/apol-aug312013x10k.htm.

239 IPEDS First-Look (July 2013), table 2. Average costs (in constant 2012-13 dollars) associated with attendance for full-time, first-time degree/certificate-seeking undergraduates at Title IV institutions operating on an academic year calendar system, and percentage change, by level of institution, type of cost, and other selected characteristics: United States, academic years 2010-11 and 2012-13.

240 Id.

241 Cellini, S. R. (2012). For Profit Higher Education: An Assessment of Costs and Benefits. National Tax Journal, 65 (1): 153-180.

242 Cellini, S. R., and Darolia, R. (2013). College Costs and Financial Constraints: Student Borrowing at For-Profit Institutions. Unpublished manuscript.

243 Id.

244 Deming, D.J., Goldin, C., and Katz, L.F. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

245 National Postsecondary Student Aid Study 2012.

246 Id.

247 Id.

248 Id.

249 Darolia, R. (2013). Student Loan Repayment and College Accountability. Federal Reserve Bank of Philadelphia.

250 Deming, D.J., Goldin, C., and Katz, L.F. (2012). The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, 26(1), 139-164.

251 Based on the Department’s analysis of the three-year cohort default rates for fiscal year 2011, U.S. Department of Education, available at http://www2.ed.gov/offices/OSFAP/defaultmanagement/schooltyperates.pdf.

252 Federal Student Aid, Default Rates for Cohort Years 2007-2011, available at www.ifap.ed.gov/eannouncements/060614DefaultRatesforCohortYears20072011.html.

253 Postsecondary Education: Student Outcomes Vary at For-Profit, Nonprofit, and Public Schools (GAO-12-143), GAO, December 7, 2011.

254 For Profit Higher Education: The Failure to Safeguard the Federal Investment and Ensure Student Success, Senate HELP Committee, July 30, 2012.

255 Id.

256 Id.

257 “Students Attending For-Profit Postsecondary Institutions: Demographics, Enrollment Characteristics, and 6-Year Outcomes” (NCES 2012-173). Available at: http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2012173

258 Id.

259

260

261 NCES, Transferability of Postsecondary Credit Following Student Transfer or Coenrollment, NCES 2014-163, table 8.

262 Cellini, S. R., and Darolia, R. (2013). College Costs and Financial Constraints: Student Borrowing at For-Profit Institutions. Unpublished manuscript. Available at www.upjohn.org/stuloanconf/Cellini_Darolia.pdf.

263 Darolia, R. (2013). Student Loan Repayment and College Accountability. Federal Reserve Bank of Philadelphia.

264 Cellini, S. R., and Darolia, R. (2013). College Costs and Financial Constraints: Student Borrowing at For-Profit Institutions. Unpublished manuscript. www.upjohn.org/stuloanconf/Cellini_Darolia.pdf.

265 Lang, K., and Weinstein, R. (2013). “The Wage Effects of Not-for-Profit and For-Profit Certifications: Better Data, Somewhat Different Results.” NBER Working Paper.

266 Deming, D., Goldin, C., and Katz, L. The For-Profit Postsecondary School Sector: Nimble Critters or Agile Predators? Journal of Economic Perspectives, vol. 26, no. 1, Winter 2012.

267 Id.

268 See section 401(c)(5) of the HEA, 20 U.S.C. 1070a(c)(5), for Pell Grant limitation; see section 455(q) of the HEA, 20 U.S.C. 1087e(q), for the 150 percent limitation. Specifically, Federal law sets lifetime limits on the amount of grant and subsidized loan assistance students may receive: Federal Pell Grants may be received only for the equivalent of 12 semesters of full-time attendance, and Federal subsidized loans may be received for no longer than 150 percent of the published program length.

269 For-Profit Colleges: Undercover Testing Finds Colleges Encouraged Fraud and Engaged in Deceptive and Questionable Marketing Practices (GAO-10-948T), GAO, August 4, 2010 (reissued November 30, 2010).

270 Id.

271 For Profit Higher Education: The Failure to Safeguard the Federal Investment and Ensure Student Success, Senate HELP Committee, July 30, 2012.

272 Id.

273 “U.S. to Join Suit Against For-Profit College Chain,” The New York Times, May 2, 2011. Available at: http://www.nytimes.com/2011/05/03/education/03edmc.html?_r=0.

274 “A.G. Schneiderman Announces Groundbreaking $10.25 Million Dollar Settlement with For-Profit Education Company That Inflated Job Placement Rates to Attract Students,” press release, Aug. 19, 2013. Available at: www.ag.ny.gov/press-release/ag-schneiderman-announces-groundbreaking-1025-million-dollar-settlement-profit.

275 “Attorneys General Take Aim at For-Profit Colleges’ Institutional Loan Programs,” The Chronicle of Higher Education, March 20, 2012. Available at: http://chronicle.com/article/Attorneys-General-Take-Aim-at/131254/.

276 “Kentucky Showdown,” Inside Higher Ed, Nov. 3, 2011. Available at: www.insidehighered.com/news/2011/11/03/ky-attorney-general-jack-conway-battles-profits.

277 “For Profit Colleges Face New Wave of State Investigations,” Bloomberg, Jan. 29, 2014. Available at: www.bloomberg.com/news/2014-01-29/for-profit-colleges-face-new-wave-of-coordinated-state-probes.html.

278 “For Profit Colleges Face New Wave of State Investigations, Bloomberg, Jan. 29, 2014. Available at: www.bloomberg.com/news/2014-01-29/for-profit-colleges-face-new-wave-of-coordinated-state-probes.html.

279 “Corinthian Colleges Crumbles 14% on SEC probe,” Fox Business, June 11, 2013. Available at: www.foxbusiness.com/government/2013/06/11/corinthian-colleges-crumbles-14-on-sec-probe/.

280 U.S. Department of Education, Press Release, “Education Department Names Seasoned Team to Monitor Corinthian Colleges.” Available at: www.ed.gov/news/press-releases/education-department-names-seasoned-team-monitor-corinthian-colleges.

281 “Dollar Signs In Uniform,” Los Angeles Times, Nov. 12, 2012. Available at: http://articles.latimes.com/2012/nov/12/opinion/la-oe-shakely-veterans-college-profit-20121112; citing “Harkin Report,” S. Prt. 112-37, For Profit Higher Education: The Failure to Safeguard the Federal Investment and Ensure Student Success, July 30, 2012.

282 Id.

283 Id.

284 Id.

285 “We Can’t Wait: President Obama Takes Action to Stop Deceptive and Misleading Practices by Educational Institutions that Target Veterans, Service Members and their Families,” White House Press Release, April 26, 2012. Available at: www.whitehouse.gov/the-press-office/2012/04/26/we-can-t-wait-president-obama-takes-action-stop-deceptive-and-misleading.

286 “$2.5M Settlement over ‘GIBill.com’,” Inside Higher Ed, June 28, 2012. Available at: www.insidehighered.com/news/2012/06/28/attorneys-general-announce-settlement-profit-college-marketer.

287 Baum, S., & Schwartz, S. (2006). How Much is Debt is Too Much? Defining Benchmarks for Manageable Student Debt. New York: The College Board.

288 King, T., & Frishberg, I. (2001). Big loans, bigger problems: A report on the sticker shock of student loans. Washington, DC: The State PIRG’s Higher Education Project.

289 Harrast, S.A. (2004). Undergraduate borrowing: A study of debtor students and their ability to retire undergraduate loans. NASFAA Journal of Student Financial Aid, 34(1), 21–37.

290 King, T., & Bannon, E. (2002). The burden of borrowing: A report on rising rates of student loan debt. Washington, DC: The State PIRG’s Higher Education Project.

291 Illinois Student Assistance Commission (2001). Increasing college access…or just increasing debt? A discussion about raising student loan limits and the impact on Illinois students. Available at: http://www.collegezone.com/media/research_access_web.pdf

292 Id.

293 Baum, S., & O’Malley, M. (2003). College on credit: How borrowers perceive their education. The 2002 National Student Loan Survey. Boston: Nellie Mae Corporation.

294 Avery, C. & Turner, S., (2012). Student Loans: Do College Students Borrow Too Much—Or Not Enough? Journal of Economic Perspectives Vol 26(1).

295 Baum, S. & O’Malley, M. (2003). College on credit: How borrowers perceive their education debt. Results of the 2002 National Loan Survey (Final Report). Braintree, MA: Nellie Mae.

296 In the “Analysis of the Regulation” the term “students” for the most part, refers to individuals who receive title IV, HEA program funds for a GE program as defined in “§668.402 Definitions” The Department’s analysis of the effect of the rule is based on the defined term, but the references to commenters’ analysis and some background information may refer to students more generally.

297 http://www2.ed.gov/policy/highered/reg/hearulemaking/2012/gainfulemployment.html.

298 This cohort uses fiscal years, whereas the regulations use award years for the computation of the D/E rates. Since the earnings data available are tied to cohorts defined in terms of fiscal years, the 2012 GE informational D/E rates are based on a fiscal year calendar.

299 In comparison, for programs that do not meet this minimum n-size, programs with 30 or more students who completed the program during a four-year cohort period will also be evaluated under the regulations.

300 The 2012 GE informational D/E rates files on the Department’s Web site also include debt-to-earnings rates for variations on n-size for comparative purposes.

301 FY 2010 enrollment is the most recent NSLDS data available to the Department regarding enrollment in GE programs. It is important to note that this data may not account reflect the overall decline in postsecondary enrollment since FY 2010.

302 A small number of programs in the 2012 GE informational D/E rates data set did not have FY 2010 enrollment data.

303 November 2013 NSLDS data was the closest existing data capture of sector and type to the approximate time for which rates would have been calculated for all measures evaluated in this Regulatory Impact Analysis.

304 In the final regulations the definition of “credential level” has been revised to clarify that postgraduate certificates are included in the post-baccalaureate certificate credential level.

305 We used fiscal years for the computation of the 2012 GE informational D/E rates, whereas the regulations use award years.

306 In comparison, under the regulations, Perkins loans will also be included in total loan debt. As such, informational rates analysis results should be considered an approximation of the implementation of the GE regulation.

307 Under the regulations, loan debt is capped for each student at the amount charged for tuition and fees, books, supplies, and equipment.

308 The regulations clarify that postgraduate certificates would be included in the post-baccalaureate certificate credential level.

309 The 2012 GE informational rates files also include debt-to-earnings rates calculated using variations of the amortization schedule for comparative purposes.

310 For the 2012 informational D/E rates cohort, the applicable average interest rates are the same for undergraduate and graduate programs. In comparison, undergraduate and graduate interest rates differ from each other in future cohort periods.

311 The Poverty Guideline is the Federal poverty guideline for an individual person in the continental United States as issued by the U.S. Department of Health and Human Services. The Department used the 2011 Poverty Guideline of $10,890 to conduct our analysis

312 Informational rates published in the past may have used a different year’s Poverty Guideline.

313 The pCDR n-size requirements apply to borrowers while the D/E rates n-size requirements apply to students who complete the program.

314 IPEDS 2011 OPEIDs used because that would be close to the time of calculation of rates for the cohort.

315 The denominator of percent minority includes all race categories including American Indian, Asian, Black, Hispanic, White, Two or More Races, Race Unknown, Nonresident Alien

316 The annual earnings rate for this analysis differs slightly from the annual earnings rate used in the NPRM in that it reflects interest rate changes made to the regulations since the NPRM.

317 IPEDS 2011 OPEIDs used because that would be close to the time of calculation of rates for the cohort.

318 The proportion of students who completed programs in the race unknown and nonresident alien categories were not considered in the Department’s analysis

319 Unmatched programs may bias results that include race/ethnicity variables. The sample with matched programs had a mean annual earnings rate of 5.6 (standard deviation=5) in comparison to the sample that did not match which had a mean annual earnings rate of 6.4 (standard deviation=5)

320 Details on determining dependence/independence are available at https://studentaid.ed.gov/fafsa/filling-out/dependency#dependent-or-independent.

321 Goldrick-Rab, S., and Sorensen, K. (2010, Fall). Unmarried Parents in College, Future of Children, Journal Issue: Fragile Families (20).

322 Average percent Asian was similar across passing, zone, and failing programs (all categories between four and five percent), average percent American Indian was also similar across the categories (roughly one percent in all categories)

323 For purposes of this analysis, nonresident aliens and race unknown categories were excluded in the denominator in the calculation of percentages

324 Detailed results are not provided here.

325 A small number of informational rate programs did not have FY 2010 enrollment data.

326 This program count includes either GE programs that reported FY 2010 title IV enrollment and/or reported 2012 informational D/E rates (n>10)and/or had Department-calculated 2012 informational pCDR rates.

327 Percentages not provided in table

328 Defined as a unique six-digit OPEID.

329 Defined as a unique six-digit OPEID.

330 Defined as a unique six-digit OPEID.

331 Jonathan Guryan and Matthew Thompson, Charles River Associates, Report on the Proposed Gainful Employment Regulation, 76-85.

332 Jonathan Guryan and Matthew Thompson, Charles River Associates, Report on the Proposed Gainful Employment Regulation, 67-69.

333 U.S. Department of Education, Digest of Education Statistics 2013, Table 303.20, “Total fall enrollment in all postsecondary institutions participating in Title IV programs and annual percentage change,” available at http://nces.ed.gov/programs/digest/d13/tables/dt13_303.20.asp; Data from IPEDS, “Fall Enrollment Survey” (IPEDS-EF:95-99); and IPEDS Spring 2001 through Spring 2013, Enrollment component (prepared October 2013).

334 Id.

335 Office of Management and Budget, Circular A4: Regulatory Analysis (September 2003), available at www.whitehouse.gov/sites/default/files/omb/assets/omb/circulars/a004/a-4.pdf.

336 Charles River Associates (2011)

337 Bradford Cornell & Simon M. Cheng, Charles River Assoc. for the Coalition for Educ. Success, An Analysis of Taxpayer Funding Provided for Post-Secondary Education: For-profit and Not-for-profit Institutions 2 (Sept. 8, 2010) 16

338 Shapiro & Pham, The Public Costs of Higher Education: A Comparison of Public, Private Not-For-Profit, and Private For-Profit Institutions, (Sonoco 2010) 5.

339 Klor de Alva, Nexus, For‐Profit Colleges and Universities: America’s Least Costly and Most Efficient System of Higher Education, August 2010.

340 U.S. Department of Education, Digest of Education Statistics 2013, Table 333.10 and Table 333.55

341 Jorge Klor de Alva & Mark Schneider, Do Proprietary Institutions of Higher Education Generate Savings for States? The Case of California, New York, Ohio and Texas available at http://nexusresearch.org/reports/StateSaving/How%20Much%20Does%20Prop%20Ed%20Save%20States%20v9.pdf

342 Projected interest rates from Budget Service used in calculations requiring interest rates for future award years.

48