Justification Part A

JUSTIFICATION PART A

SUPPORTING STATEMENT FOR REQUEST FOR OMB APPROVAL

UNDER THE PAPERWORK REDUCTION ACT

PART A. JUSTIFICATION.

This is a justification for the Department's request for an extension of the Unemployment Insurance (UI) Benefit Accuracy Measurement (BAM) program. From its inception in 1987 until 1996, it was called Benefits Quality Control (BQC). It is authorized by regulation at 20 CFR 602. Attachment A contains the relevant sections of the Social Security Act on which the regulation is based. The Department's BAM information collection authority, under OMB number 1205-0245, is currently scheduled to expire on August 31, 2009.

This request does not seek to make any changes in the program.

A-1. Reasons for Data Collection

The BAM program consists of two comprehensive reviews: Paid Claims Accuracy (PCA) and Denied Claims Accuracy (DCA). States conduct intensive audits of statewide probability samples of UI payments and denials to determine their accuracy. The Department provides State Workforce Agency (SWAs) with software to edit the sampling frame files and select the weekly PCA and DCA samples.

BAM seeks to reduce waste, fraud, and abuse in the UI system. By investigating small representative weekly samples of paid and denied UI claims, it enables each State Workforce Agency (SWA) to estimate reliably the number of proper and improper payments (i.e., overpayments and underpayments) and denials, their rates of occurrence, and their types, causes, and responsibilities. For paid claims, BAM also estimates the dollar value and rate of improper payments. BAM PCA and DCA audits also provide information that can be used for program improvement, including the type of payment error, error cause, responsible party, point of detection within the system, and the actions of the claimant, employer, and agency prior to the BAM investigation.

In order to ensure uniformly thorough findings, all jurisdictions are required to apply the same BAM investigative methodology and coding instructions and to compile and report their data through the same automated data system. These are specified in Employment and Training (ET) Handbook 395, Benefits Accuracy Measurement State Operations Handbook.

The BAM program has undergone many changes over the years. The original BQC program was the outgrowth of experience with the Random Audit (RA) program which operated from 1981 to 1985. The BQC program was implemented voluntarily in 1986 and became mandatory for 52 SWAs-all but the Virgin Islands--on October 5, 1987.

Originally, all case investigations were required to be conducted in person. In July 1993, SWAs were allowed to conduct certain portions of a case investigation by mail, telephone or FAX. Beginning in 1997, they could exercise complete discretion in how case investigations were conducted. Also in 1997, to give states increased ability to conduct program improvement activities, the Department reduced minimum allocated paid claims sample sizes to 360 in the 10 states with the smallest UI workload (in terms of UI weeks paid) and 480 cases in other states. (States can choose to select and investigate larger samples, and some do so.) In 1996, the Department removed its requirement that SWAs release their BAM data to the public independently of the Department's public release of the data, based on the recommendation of the Federal-State workgroup that produced a framework for the Department’s performance management system, UI Performs. Beginning in August 2001, BAM was modified to include the sampling and investigation of UI claims denied for monetary, separation, or nonseparation issues. Effective January 2008, states were required to match BAM paid claims with the National Directory of New Hires (NDNH).

Paid Claims Accuracy

BAM paid claims accuracy is the means by which the UI system assesses the accuracy of UI benefit payments. Each week a random sample is selected of both intrastate and interstate original payments (including combined wage claims) made for a week of unemployment under the State UI, Unemployment Compensation for Federal Employees (UCFE), and Unemployment Compensation for Ex-servicemembers (UCX) programs. A sample of 360 cases per year is pulled in the ten states with the smallest UI program workloads (in terms of average annual UI weeks paid for the most recent five-year period) and 480 cases per year in the other states. State BAM staff audit each selected claim, examining all aspects of a claimant's eligibility to receive unemployment compensation during the sampled week. In their investigation, staff verify wages used to establish monetary entitlements, the claimant's reason for being unemployed, efforts to find work during the week and any other factors which would have affected the claimant’s entitlement to a benefit during the sampled week or the amount of the benefit paid.

Denied Claims Accuracy

States investigate BAM denied claims cases to determine claimant eligibility for UI payments in three broad areas: monetary determinations, separation determinations, and nonseparation determinations.

• A monetary determination is made when a claim is initially filed (or when a claim is made to establish a new benefit year) to verify that the claimant has sufficient wage credits in the base period and has satisfied other monetary requirements to demonstrate attachment to the labor force.

• A separation determination is made when the claim is initially filed or when an additional claim is filed in the claimant’s benefit year after a period of intervening employment. Separation determinations evaluate whether the claimant’s unemployment is involuntary and through no fault of the claimant.

• A nonseparation determination verifies that the claimant is meeting the eligibility requirements of state law for a specific week of unemployment: the claimant was able to work, available for work, actively sought work, and received sufficiently little income in the week to be considered “unemployed”.

Investigation of BAM denied claims follows the paid claims case investigation methodology. It evaluates denials accuracy by investigating random samples of each of the three types of denials. The Department has supplied each SWA with software that performs quality assurance edits of the sampling frames and randomly selects the BAM denied claims samples. The information on which the decision to deny benefits was made is verified by reviewing agency records, interviewing the claimant and contacting employers and other involved parties. All states sample a minimum of 150 cases of each type of denial in each calendar year. State BAM staff review agency records and contact claimants, employers, and all other relevant parties to verify information in agency records or obtain additional information pertinent to the determination that denies eligibility. Unlike the investigation of paid claims, in which all prior determinations affecting claimant eligibility for the compensated week selected for the sample are evaluated, the investigation of denied claims is limited to the issue upon which the denial determination is based.

States code the findings of their PCA and DCA investigations in a database that is maintained on a computer located in each SWA. The Department stores copies of these state databases (minus personally identifiable information) in a database maintained by the ETA Office of Workforce Security. The Department publishes annual performance results on the ETA Web site, http://workforcesecurity.doleta.gov/unemploy/.

A-2. Users, Purposes, and Consequences of Failure to Collect the Information

SWAs use both paid claims and denied claims data to evaluate the quality of their existing UC claims processes. It enables a SWA to meet its primary objective of strengthening the controls that prevent errors and/or fraud and abuse in the payment and denial of UI benefits.

The data collected in accordance with prescribed BAM methodology provides national and SWA administrators with accurate measurements of the rate of proper and improper payments and denials, the reasons for improper payments and denials, and who is responsible for them. Identification of specific types, causes, and responsibility for errors provides information about the effectiveness of state programs and the quality of their underlying policies, thereby serving as a basis to improve and strengthen program operations. BAM data can lead state and national program managers to make significant program improvements resulting in dollar savings, and continuing benefit payment integrity.

The Department uses BAM data to measure state performance with respect to UI payment integrity and to meet the Department’s reporting requirements of the Government Performance and Results Act (GPRA) and the Improper Payments Information Act (IPIA). Both National and Regional Office UI staff use the BAM data to provide technical assistance to state UI programs. The data are also used as part of the Department’s policy analysis and policy formulation functions, and are an essential component of UI Performs, the Department’s performance management system.

UI Performs promotes continuous improvement in UI performance, through the establishment of core performance measures and acceptable levels of performance (ALPs). One of these core measures, Overpayment Detection, includes BAM data. Under UI Performs, state and Federal staff work cooperatively to identify areas of UI programs that need improvement and develop appropriate plans through the annual State Quality Service Plan (SQSP). The Department believes that the SQSP mechanism is the most effective method for drawing attention to all performance deficiencies and providing opportunities to plan for improvements.

A-3. Technology and Obstacles Affecting Reporting Burden

In order to comply with the Government Paperwork Elimination Act, the BAM program uses an automated system for data collection, transmission, and retrieval that utilizes state-of-the-art information processing technology. This system was designed to maximize the use of data elements that are already collected by the SWA for processing UC claims thus minimizing the amount of additional effort required to collect this information.

The Department has provided each SWA with a Sun computer and is in the process of upgrading state systems with a Sun T2000 computer. The Department also provides states with an Informix relational database and applications software to enter, store, transmit, and retrieve BAM paid claims and denied claims data. Personal identifiers such as Social Security numbers (SSNs) are stored in the SWAs’ databases but are not transmitted to the Department.

The Department knows of no technical obstacles to operating the BAM program.

A-4. Duplication.

The BAM program does not duplicate any other UI reporting system. No other program involves the intensive case investigation of information for a sample of paid and denied claims through contacts with claimants, employers and third parties. The Department is not aware of any alternative to selecting samples of payments using a standard sample selection program, and validating information for measuring payment accuracy.

A-5. Burden on Small Business or Other Small Entities

There is a minor impact on small businesses. Although the formal respondents are SWAs, many of the employers contacted in the course of BAM case investigations are small businesses. Most contacts require less than half an hour of an employer's time (the only exception is an occasional base period wage investigation, which may take up to 45 minutes). Because the number of both paid and denied claims cases investigated average from 810 cases per year for the ten states with the smallest claims loads to 930 cases per year for the remaining 42 states, the likelihood that any small employer will be contacted more than once in a year is very small.

A-6. Consequences of Less Frequent Data Collection

BAM samples are drawn weekly and investigated on an on-going basis. The data are entered into the database as case investigations are completed. The Department runs a program each night to pick up any changes in the SWAs' databases. The current frequency of the data collection is necessary to assure the quality and integrity of the data for several reasons.

First, because sampling frames (populations) are assembled and samples are drawn weekly, sample and population characteristics can be compared to determine the representativeness of the samples and the integrity of the sampling frames. The Department has developed software which the Department and the SWAs use as a quality assurance tool. If flaws in the sampling procedure or population files are discovered, action can be taken immediately. If data were collected less frequently, any problems related to the collection process could compromise the integrity of the data. Useable information for the period affected would be lost, and important program management information would not be available until the next data collection period.

Second, experience in the BAM program has demonstrated that the review of completed cases is more accurate and efficient the sooner it occurs. State staff recall on questioned points is better, as is the quality of data derived from field investigations. Confidence in the use of BAM data depends in part on knowing the data have been reviewed thoroughly.

As noted above, BAM paid claims and denied claims are important parts of the UI Performs management system, which promotes continuous improvement in UI operational performance. The value of quality assessment information is directly related to its timeliness. Quality assessment systems must provide immediate feedback about problems that have been detected in order for system administrators to respond with corrective actions. Continuous data collection also enables program managers to evaluate the effectiveness of continuous improvement initiatives. The Department believes that because the UC system functions continuously, management information systems such as BAM, which monitor system quality, must reflect the characteristics of the operations they are evaluating. Sporadic or periodic collection of quality data will not meet these program management requirements.

A-7. Special Circumstances Involved in Collection of BAM Data

• In addition to the circumstances noted in A-6 above, it should be noted that weekly sampling and continuous data collection impose no added reporting burden on the states, since the states’ ADP systems are accessed overnight by the ETA OWS computer through automated data pick-up procedures.

• There is no requirement in the program to prepare written responses to a collection of information in fewer than 30 days; submit multiple copies of documents; or retain records for more than three years.

• The program is intended to produce statistical results that can be generalized (projected) to the population.

• All data classifications have been approved by OMB.

• Individuals and businesses contacted through the BAM program are not required to submit proprietary information or trade secrets.

A-8. Preclearance Notice and Responses.

The Department's preclearance notice was published in the Federal Register on March 31, 2009, for 60-day comment. (Vol. 74, No. 60). No comments were received.

A-9. Payments to Respondents.

Because the BAM program is a mandated data collection program, the Department provides funding to the participating states, which are the primary respondents. Persons contacted in the course of the case investigation, which are secondary respondents, do not receive payments or gifts.

A-10. Confidentiality

All states’ laws must conform to 20 CFR 603 (Federal-State Unemployment Compensation (UC) Program; Confidentiality and Disclosure of State UC Information). The BAM State Operations Handbook (ET Handbook No. 395, chapter VI) requires that, “States' written laws and policies are the bases for all determinations. Written policy is that policy that is distributed SWA-wide and, upon request, may be made available to the public.”

Respondents to the BAM data collection are informed that their responses are subject to state confidentiality statutes and that the Department will publish or disseminate data at a level of aggregation that will preclude the identification of individual respondents.

The Department maintains strict controls over the data gathered through the BAM program. The Department cannot identify an individual claimant from the BAM case in its database; the Federal BAM case record does not contains either the claimant’s name or SSN. Although the state data record contains the SSN, this field is not included when the case is uploaded to the Federal database. Users outside the state thus cannot identify individual claimants.

BAM data are published at the state and national level of aggregation. Statistics for population subgroups and characteristics are published for broad categories of UI program characteristics, such as the types of erroneous payments and improper denials, and the causes, responsible parties, and detection points of erroneous payments and denied claims.

A-11. Questions of a Sensitive Nature

The data collection instrument includes no questions of a sensitive nature.

A-12. Respondents' Burden and Cost of Collecting Information

The total burden comprises activities related to collecting and reporting the BAM paid and denied claims data. It is calculated as the sum of time spent by state staff to prepare for each case, review pertinent records, set up and conduct interviews and other fact finding efforts, and review and record the data. Although there is no payment made to these individuals, it also includes the time the claimants, employers and third parties spend providing information to the state investigators. Each SWA will select and investigate an average sample of 457 paid claims and 150 each of denied monetary, separation and nonseparation claims.

The investigation of BAM paid and denied claims involves 3,634 respondents and a total of 8,267 burden hours, for each state, as shown below.

BAM PCA / DCA Data Collection Burden Per State

 

	Paid Claims	Monetary Denied Claims	Separation Denied Claims	Nonseparation Denied Claims	Total
Cases	457*	150	150	150	907
Respondents/Case	4.90	3.20	3.00	3.10	4.01
Hours/Case	10.5	7.925	7.525	7.675	9.11
Total Respondents	2239.3	480	450	465	3634.30
Total Hours	4798.5	1188.75	1128.75	1151.25	8267.25

* Average for all 52 states. The 10 smallest states in terms of UI weeks paid sample at the rate of 360 cases per year; the other 42 states sample at the rate of 480 cases per year.

The total response burden for both paid and denied claims and the 52 SWAs is:

52 SWAs X 3,634.3 respondents = 188,983.6 respondents

52 SWAs X 8,267.25 hours = 429,897 hours

These estimates are based on the following data:

Respondents/ Case

For paid claims, based on FY 2008 BAM paid claims data, each case involves one state investigator, one claimant, 1.7 base period employers, 1.1 work search contact, and occasionally a third party such as a school or labor union (average contacts: 0.1).

All denied claims investigations involve one state investigator and one claimant. Based on FY 2008 BAM paid claims data, each monetary denial investigation involves approximately 1.2 base period employers; there is usually one separating employer; and nonseparation eligibility issues usually involve one employer and occasionally a third party such as a school or labor union (average contacts: 1.1).

Hours/Case

For paid claims, SWA investigators spend 5.1 hours, on average, to complete a BAM paid claims investigation, with an additional 3.17 hours for coding and entering data into a computerized database, reviewing completed cases, and transmitting the data to the Department, for a total of 8.27 hours per investigation.

For denied claims, the average time of completion and transmitting data by type of denial are: monetary denials - 6.825 hours; separation denials - 6.525 hours; and nonseparation denials - 6.625 hours.

Claimant, employer, and third party responses are approximately 0.5 hours per response per paid and denied case.

The Department estimates that the total cost of data collection and reporting for BAM paid and denied claims is $295,196 per state. It is computed as follows:

SWA staff

For paid claims, the Department estimates that the net burden investigating, collecting and transmitting the BAM information requested in this justification will be approximately 8.27 hours per case per year based on an allocated sample 23,764 cases, or an average of 457 cases per SWA. At the average rate of $39.17 per hour (based on the $67,016 and 1,711 hours per year the Department uses to compute UI state agency staff costs in FY 2009), the annual per state cost of the BAM paid claims burden for investigating, record keeping and reporting costs is:

8.27 hrs. X $39.17 X 457 = $148,039.

For denied claims, the Department estimates that the net burden will be approximately 6.66 hours per case for 450 cases at the average rate of $39.17 per hour the Department uses to compute UI staff costs (based on $67,016 and 1,711 hours per year):

6.66 hrs. X $39.17 X 450 = $117,392.

The cost for SWA staff for BAM paid and denied claims is $265,431 per state. Total cost for 52 SWAs is $13,802,412. Federal budget allocations cover these costs, as they cover the costs of other UI operations.

Claimants

For paid claims, the per claimant cost is estimated to be $3.28 based on the minimum wage of $6.55 and an estimated 0.5 hours to complete the claimant interview. The estimated claimant cost is:

0.5 hrs. X $6.55 X 457 cases = $1,497 (per state) x 52 states = $77,844.

Based on the 1987 and 1997-98 DCA pilots, it is assumed that half of the claimants denied UC would be employed by the time they are contacted by the DCA investigator; an hourly rate of $14.39 was estimated from the FY 2008 BAM data. The minimum wage of $6.55 was used for the other half of the claimants, who were assumed to be still unemployed at the time of the DCA interview:

0.5 hrs. X $14.39 X 225 = $1,619

0.5 hrs. X $6.55 X 225 = $737

The cost for denied claimant response is: 52 X ($1,619 + $737) = $122,512. The total cost for claimants responding for both paid and denied claims is: $200,356.

Employers and Third Parties

For paid and denied claims, an average hourly rate for employer and third party respondents is estimated to be $28.47, adjusted for inflation, based on the BLS National Occupational Employment and Wage Estimates (May 2007) for Human Resources and Labor Relations occupations. On average, 1.7 employers, 1.1 work search contacts, and 0.1 third parties respond per BAM paid claims case (estimated response time of 0.5 hours):

0.85 hrs. X $28.47 X 457 = $11,059 (per state) x 52 = $575,068

0.55 hrs. X $28.47 X 457 = $ 7,156 (per state) x 52 = $372,112

0.05 hrs. X $28.47 X 457 = $ 651 (per state) x 52 = $ 33,852

On average, 1.2 employers respond per BAM denied monetary claims cases; one employer per BAM denied separation cases, and 1.1 employers and third parties respond per BAM nonseparation cases:

0.6 hrs. X $28.47 X 150 = $2,562 (per state) X 52 = $133,224

0.5 hrs. X $28.47 X 150 = $2,135 (per state) X 52 = $111,020

0.55 hrs. X $28.47 X 150 = $2,349 (per state) X 52 = $122,148

Total cost for employers and third parties is $1,347,424.

Costs are summarized in the following table.

Cost Summary	Paid Claims	Denied Claims	Cost Per State	Cost - All States
SWA Staff	$148,039	$117,392	$265,431	$13,802,412
Claimants	$1,497	$2,356	$3,853	$200,356
Employers + 3rd Parties	$18,866	$7,046	$25,912	$1,347,424
Total All Costs	$168,402	$126,794	$295,196	$15,350,192

A-13. Annual Cost to Respondents

No major equipment purchases or similar start-up costs are required for respondents, as federal allocations underwrite respondents’ costs.

A-14. Annualized Federal Cost

Federal costs include the staff required to manage the BAM data collection and analyze the data and the maintenance the National Office computer system and database. The staff costs are summarized in the following table:

	Job Title	# of Staff	Staff Years	GS Level	Salary*	Cost
National Office
	Mathematical Statistician	1	0.9	13	$98,518	$88,666.20
	UI Program Specialist	1	0.8	13	$98,518	$78,814.40
	UI Program Specialist	1	1.0	12	$82,845	$82,845.00
						$250,325.60
Regional Offices
Region 1	UI Program Specialist	1	0.8	12	$83,437	$66,749.60
		1	0.3	13	$99,222	$24,805.50
Region 2	UI Program Specialist	1	0.8	12	$81,600	$65,280.00
Region 3	UI Program Specialist	1	0.5	12	$79,783	$39,891.50
Region 4	UI Program Specialist	2	0.8	12	$80,725	$129,160.00
Region 5	UI Program Specialist	1	0.9	13	$83,767	$75,390.30
	UI Program Specialist	1	0.5	12	$99,615	$49,807.50
Region 6	UI Program Specialist	1	0.5	13	$90,416	$45,208.00
		1	0.3	12	$107,522	$32,256.60
						$528,549.00
Total						$778,874.60

* Based on the GS grade step 5 rate as of January 2009.

Federal ADP costs to maintain the National Office computer and database are estimated to be approximately $130,000 annually.

Federal allocations to the SWAs also cover the costs in A-12 and A-13. There are no costs to the Federal government for claimant/employer/third parties time during the case investigations.

A-15. Reasons for Change in Burden

PCA hours per case, previously stated at 10.496, has been rounded to 10.5 hours. This technical adjustment results in an increase in the total burden hours of 92 hours. Also previously costs were erroneously assigned to the respondent SWAs. This has been corrected to indicate that those costs are, in fact, born by, federal allocations.

A-16. Publication Information

The Department publishes an annual summary of results and analyses of BAM paid and denied claims findings for each state and nationally. This publication is available to the public on the ETA Web site, http://workforcesecurity.doleta.gov/unemploy/. The most recent BAM data available are for calendar year 2007 at http://workforcesecurity.doleta.gov/unemploy/bam/2007/bam-cy2007.pdf.

A-17. Reasons for Not Displaying Date OMB Approval Expires.

The Department intends to display approval information.

A-18. Exceptions to Certification.

There are no exceptions to the certification statement in 83-I.

B-10