NMDB-NSMB--PRA Supporting Statement (revised 2013-12-12)

“NATIONAL SURVEY OF MORTGAGE BORROWERS"
OMB NUMBER 2590-00XX
REVISED SUPPORTING STATEMENT

BACKGROUND
The Federal Housing Finance Agency (FHFA) is seeking OMB clearance under the Paperwork
Reduction Act (PRA) for a new collection of information called the “National Survey of Mortgage
Borrowers” (NSMB). The NSMB will be a quarterly survey of individuals who have recently
obtained a loan secured by a first mortgage on single-family residential property. The survey will
be carried out primarily by a contractor under the direction of FHFA. The survey questionnaire
will consist of approximately 80-85 questions designed to obtain information about individual
residential mortgages and borrowers that is not available elsewhere. This includes specific
information on: the terms of the mortgage obtained; the mortgaged property; the borrower’s
experience with the loan application and closing processes; and the borrower’s financial resources
and financial knowledge. A copy of the draft survey questionnaire, revised as of November 22,
2013, is attached to this supporting statement.
The NSMB is one component of a larger project, known as the “National Mortgage Database.”
The National Mortgage Database is a joint effort of FHFA and the Consumer Financial Protection
Bureau (CFPB) and is a uniquely comprehensive database of information on the residential
mortgage origination market. Although the overall National Mortgage Database project is being
co-sponsored by FHFA and CFPB, FHFA alone is responsible for administering the NSMB. In
order to provide the context for this collection of information, the background of the National
Mortgage Database project is summarized below.
The need for the National Mortgage Database grew out of the subprime mortgage crisis that began
in the United States in 2007. The subprime crisis, and the financial crisis that followed, revealed
the extent to which regulators lacked sufficient data to foresee and adequately respond to a crisis in
the mortgage markets.
Section 1324 of the Housing and Economic Recovery Act of 2008 (HERA) addressed one
important mortgage finance-related data deficiency by requiring that FHFA conduct a monthly
survey to collect data on the characteristics of individual prime and subprime mortgages, and on
the borrowers and properties associated with those mortgages. Specifically, FHFA is required to
collect data on: the sales price of the mortgaged property; the loan-to-value ratio of the mortgage;
the terms of the mortgage; the creditworthiness of the borrowers; whether borrowers on subprime
mortgages would have qualified for prime lending; and whether the mortgage was purchased by
1

Fannie Mae or Freddie Mac (both of which are regulated by FHFA). 1 The stated purposes of the
monthly mortgage survey required under HERA are: (1) to enable FHFA to prepare a detailed
annual report on the mortgage market activities of Fannie Mae and Freddie Mac relative to the rest
of the market for the Committee on Banking, Housing, and Urban Affairs of the Senate and the
Committee on Financial Services of the House of Representatives; 2 and (2) to compile a database
of timely and otherwise unavailable residential mortgage market information to be made available
to the public. 3
In order to fulfill those and other statutory mandates, as well as to support policymaking and
research efforts, FHFA, along with CFPB, committed in July 2012 to fund, build and manage the
National Mortgage Database. A key purpose of the National Mortgage Database is to make
accessible accurate, comprehensive information for monitoring the residential mortgage market.
Some key uses are monitoring the relative health of mortgage markets and consumers; gaining new
insights on consumer decision making; monitoring new and emerging products in the mortgage
market; examining both first and second lien mortgages for a given borrower; and understanding
the impact of consumers’ overall debt burden.
FHFA purchases the core data for the National Mortgage Database from Experian—one of the
three national credit repositories—which maintains a common database of credit information on
individual consumers with the two other repositories (Equifax and Trans Union). The credit
repositories maintain the largest (close to the universe) and most representative set of mortgage
data currently available. Because the credit repositories archive their data, FHFA has been able to
create the National Mortgage Database retrospectively, building a random 1-in-20 sample of all
first-lien residential mortgages in Experian’s database outstanding at any time between January
1998 and June 2012. Since June 2012, Experian has provided, and going forward will continue to
provide, an additional random 1-in-20 sample of newly-reported mortgages drawn each quarter.
This credit repository data complements information from the Home Mortgage Disclosure Act
(HMDA) database that is maintained by the Federal Financial Institutions Examination Council,4
property valuation models, and data files maintained by Fannie Mae and Freddie Mac. As
explained more fully below, the purpose of the NSMB is to complete the National Mortgage
1

See 12 U.S.C. § 4544(c).
See 12 U.S.C. §§ 4544(a), (b).
3
See 12 U.S.C. § 4544(c)(3).
4
HMDA was enacted by Congress in 1975 and was implemented by the Federal Reserve Board's Regulation C.
Effective July 21, 2011, the rule-writing authority over Regulation C was transferred to the CFPB. Regulation C
requires that certain financial institutions, including banks, savings associations, credit unions, and other mortgage
lending institutions periodically report specified information about each mortgage loan that is originated during the
reporting period. Using the loan data submitted by these financial institutions, the Federal Financial Institutions
Examination Council creates aggregate tables for each metropolitan statistical area or metropolitan division, and
individual institution disclosure reports. This data is made available to the public in order to assist: in determining
whether financial institutions are serving the housing needs of their communities; public officials in distributing publicsector investments so as to attract private investment to areas where it is needed; and in identifying possible
discriminatory lending patterns. The HMDA Loan/Application Register, by which the HMDA database information is
collected, is currently assigned OMB control number 7100-0247, with an expiration date of September 30, 2013.
2

2

Database, by obtaining critical information that is not available from those existing sources.
Together, this information will allow for more accurate analysis of the current stock of outstanding
mortgages, new originations and terminations, as well as historic performance going back to 1998.

A. JUSTIFICATION
1. Circumstances necessitating the collection of information
FHFA is required to conduct a monthly mortgage market survey under section 1324 of HERA. As
part of that mandate, FHFA must collect information on the characteristics of individual subprime
and nontraditional mortgages, as well as on the characteristics of borrowers on such mortgages,
including information on the creditworthiness of those borrowers and information sufficient to
determine whether those borrowers would have qualified for prime lending. 5 The NSMB
questionnaire is designed to elicit this information directly from borrowers, who are likely to be the
most reliable and accessible—and, in some cases, the only—source for this information. In
addition, the questionnaire is designed to elicit more complete information on mortgage terms,
mortgaged properties, and borrowers’ household demographics than can be obtained from the
existing sources.
In the judgment of FHFA, the information obtained from the NSMB, in combination with that
obtained from the existing sources described above, will make the National Mortgage Database a
high quality, and timely resource for information on developments in the residential mortgage
finance market. The NSMB will be especially critical in ensuring that the National Mortgage
Database contains uniquely comprehensive information on the range of nontraditional and
subprime mortgage products being offered, the methods by which these mortgages are being
marketed and the characteristics, and particularly creditworthiness, of borrowers for these types of
loans. These data will enable FHFA to fulfill its regulator mandate set out in HERA and support
general rulemaking. .

2. Use of data
The information in the National Mortgage Database, including that obtained through the NSMB,
will be used for three primary purposes: (1) to prepare the report to Congress on the mortgage
market activities of Fannie Mae and Freddie Mac that FHFA is required to submit under section
1324 of HERA; (2) for research and analysis by FHFA and other federal agencies that have
regulatory and supervisory responsibilities/mandates related to mortgage markets; and (3) to
provide a resource for research and analysis by academics and other interested parties outside of
the government. Generally, the National Mortgage Database will allow Congress, regulators and
other interested parties to track emerging trends in the mortgage origination process throughout the
United States and will allow them to determine more quickly and accurately when the mortgage
origination process is changing in a way that may adversely affect financial markets, borrowers,
and consumers. FHFA intends that the availability of this information, as well as the research and
5

See 12 U.S.C. § 4544(c)(2).

3

analyses derived from it, will provide sufficient warning to allow it and other regulators to take
steps to avoid, or at least to mitigate, major mortgage market crises in the future.

3. Use of information technology
The NSMB will primarily use machine-readable paper questionnaires. Completed questionnaires
will be scanned and the responses will be automatically uploaded into the electronic National
Mortgage Database. However, the survey materials will also give recipients the option of
completing the survey online in either English or Spanish.

4. Efforts to identify duplication
As explained above, the vast majority of data included in the National Mortgage Database has been
and will continue to be drawn from existing sources—i.e., one of the consumer credit databases
maintained by the national credit repositories; the HMDA database maintained by the Federal
Financial Institutions Examination Council; and other information in the possession of FHFA, its
regulated entities or CFPB. The NSMB is designed to obtain needed information that is not
available from these other sources.

5. Impact on small entities
The information collection will not have a significant economic impact on a substantial number of
small entities. The survey recipients will be individuals only.

6. Consequences of less frequent collection and obstacles to burden reduction
Section 1324 of HERA requires that FHFA undertake a survey of mortgage markets on a monthly
basis. 6 While the performance data on existing mortgages in the National Mortgage Database is
pulled from the credit repository database on a monthly basis, newly-originated mortgages are
added to the National Mortgage Database on a quarterly basis. The NSMB questionnaires will be
sent to borrowers that originated their mortgage in the year and quarter that corresponds to the
quarterly draws from the credit depository database for newly-originated mortgages. One
important purpose of the survey is to monitor loan origination trends. Ideally, monthly housing
surveys would provide the best feedback; however, quarterly surveys are sufficient.
The NSMB questionnaire will be sent out quarterly to a simple, random sample of 7,000 borrowers
of newly originated mortgages. The sampling frame for these mortgages corresponds to loans in the

6

See 12 U.S.C. § 4544(c).

4

National Mortgage Database. 7 This NSMB sample size assumes a 40 percent overall response
rate. 8 It allows for representative samples of targeted subpopulations that are of particular interest
and that make up about 10 percent of the sample, such as those mortgages where borrowers
received counseling, mortgages for first-time homebuyers, and mortgages on homes purchased
from builders.

7. Circumstances requiring special information collection
There are no special circumstances that require FHFA to conduct the information collection in a
manner inconsistent with the guidelines provided in Item 7.

8. Solicitation of comments on information collection
In accordance with the requirements of 5 CFR 1320.8(d), FHFA published a request for public
comments regarding this information collection in the Federal Register on April 25, 2013. See 78
FR 24420 (Apr. 25, 2013). The 60-day comment period closed on June 24, 2013. FHFA received
no public comments.

9. Provision of payments or gifts to respondents
Survey recipients will receive a cash payment, currently expected to be five dollars, as an
inducement to complete and return the NSMB questionnaire. Recipients who fail to respond to the
first two survey solicitations may receive an additional cash inducement of a similar amount.

10. Assurance of confidentiality
Section 1324 of HERA authorizes FHFA to modify the mortgage data released to the public as
necessary to ensure that it contains no “representation of information that permits the identity of a
borrower to which the information relates to be reasonably inferred by either direct or indirect
means.” 9 The National Mortgage Database, including data obtained pursuant to the NSMB, will
not contain any personally identifiable information and the agencies will take appropriate
precautions to ensure that individual consumers cannot be identified through the database or
through any datasets that may be made available to researchers or the public. The NSMB
questionnaire will contain a statement to this effect and, in addition, will make clear that
respondents should not include their name, address or any other personally identifiable information
on the completed questionnaire.
7

The credit repository database contains data on mortgages originated by borrowers. On average, about 2.5 million
mortgages are originated quarterly. The National Mortgage Database is a random sample of 1-in-20 (i.e., about
125,000 mortgage tradelines) of the mortgages reported to Experian quarterly.
8
The 40 percent overall response rate assumes the NSMB will yield a minimum of 2,800 surveys responses.
9
See 12 U.S.C. § 4544(c)(3), (4).

5

FHFA published a System of Records (SORN) notice for the National Mortgage Database in the
Federal Register on December 10, 2012. 10 However, as the National Mortgage Database project
has evolved, the benefits of additional data collection have become apparent. Therefore, FHFA is
currently developing a revised SORN covering a broader range of record categories. FHFA will
provide OMB with a copy of the revised SORN as soon as it is finalized.

11. Questions of a sensitive nature
The information collection does not contain any questions of a sensitive nature.

12. Estimates of the hour burden of the information collection
FHFA has analyzed the cost and hour burden on members of the public associated with this
information collection.
The estimated total annual hour burden on the public is 14,000 hours. FHFA estimates that it will
take one survey respondent approximately 30 minutes to gather the necessary information and to
complete the survey questionnaire. This estimate is based on the reported experience of
respondents to a series of similar surveys that were conducted by Freddie Mac (these are described
in detail in part B).
•
•
•

Completion time: 0.5 hours per survey.
Total surveys: 28,000 per year (4 x 7,000 per quarter)
Total hours: 14,000 hours per year.

13. Estimated total annualized cost burden to respondents
There will be no costs imposed upon the respondents by this information collection. Postage costs
to return the survey questionnaire will be pre-paid by the survey contractor.

14. Estimated cost to the federal government
The estimated annual burden to the federal government is $452,800 and 40 hours, calculated as
follows:
FHFA analyst embeds NSMB data into query-based electronic database:

10

See 77 Fed. Reg. 73464 (Dec. 10, 2012). The currently-effective SORN can be viewed at:
http://www.fhfa.gov/webfiles/25097/FHFA-21%20SORN.pdf.

6

•
•
•
•
•

Processing time: 10 hours per quarterly survey
Total surveys: 4 surveys per year
Total hours: 40
Hourly rate: $70 (includes salary, benefits, and overhead)
Total cost: $2,800

Approximately $450,000 will be paid annually to the contractor hired to conduct the surveys. Of
this, approximately $215,000 will be attributable to the cash incentive payments to survey
recipients; approximately $60,000 will be for printing costs; and approximately $50,000 will be for
postage costs.

15. Reasons for change in burden
This is a new information collection, so there is no change in burden from any previous estimates.

16. Plans for tabulation, statistical analysis and publication
There are no specific plans for tabulation, statistical analysis or publication at this time, other than
those items discussed in the answer to question #2 above. In the future, FHFA may produce
periodic reports for the public based on aggregate data or make high-level data available to the
public in electronic form.
The project schedule is attached to this revised supporting statement. The project schedule
assumes a December 23, 2013 start date. However, FHFA now anticipates that week 1 of the
“Prepare for Mailing” phase, as identified in the project schedule, will occur in January 2014.

17. If seeking approval to not display the expiration date for OMB approval of the
information collection, explain the reasons why display would be inappropriate
FHFA plans to display the expiration date for OMB approval.

18. Explain each exception to the topics of the certification statement identified in
“certification for paperwork reduction act submission.”
There are no exceptions to the topics of the certification statement identified in the “Certification
for Paperwork Reduction Act Submission.”

7

B. COLLECTIONS OF INFORMATION INVOLVING STATISTICAL METHODS
Background—Freddie Mac Surveys
The methodology for National Mortgage Database, including the NSMB, is based upon
methodology that was developed and tested during a series of three surveys funded and carried out
by Freddie Mac between 2010 and 2012. Freddie Mac (which purchases, but does not originate,
mortgage loans) undertook those surveys with the intent of obtaining a better understanding of the
gaps and potential problems borrowers may face when obtaining home mortgages, and of ways to
improve the mortgage origination process from the borrower’s perspective. The Freddie Mac
survey questionnaires requested information on borrowers’ knowledge of and experience with
shopping for a mortgage, the mortgage loan received, the property associated with the mortgage,
characteristics of household members, and borrowers’ expectations.
In considering how to fulfill the statutory mandate for a monthly mortgage market survey imposed
by HERA, FHFA looked closely at the Freddie Mac surveys and decided upon a larger and more
complete version of those surveys as the appropriate means of fulfilling that mandate. In
developing the NSMB, FHFA has worked closely with those who developed and implemented the
Freddie Mac surveys in order to ensure that the lessons learned in conducting those surveys are
carried over to the NSMB. Freddie Mac’s experience with its surveys are discussed in the answers
below to explain the reasons for FHFA’s intended approach to the NSMB and to support FHFA’s
estimates as to the response rates.

Question 1. Describe (including a numerical estimate) the potential respondent
universe and any sampling or other respondent selection methods to be used. Data on
the number of entities (e.g., establishments, State and local government units,
households, or persons) in the universe covered by the collection and in the
corresponding sample are to be provided in tabular form for the universe as a whole
and for each of the strata in the proposed sample. Indicate expected response rates
for the collection as a whole. If the collection had been conducted previously, include
the actual response rate achieved during the last collection.
The NSMB will represent a universe of 8 to 10 million mortgages that are newly originated in the
U.S. annually. As mentioned above, a 1-in-20 simple random sample, or 400,000 to 500,000, of
those newly originated mortgages is added to the National Mortgage Database annually.
The sampling frame for the NSMB is a simple random sample of borrowers. Based on the second
and third of the Freddie Mac surveys conducted between 2010 and 2012, FHFA expects the
response rate for useable NSMB questionnaires to be between 40 and 57 percent. The sample size
allows for representative samples of targeted subpopulations that are of particular interest and that
make up about 10 percent of the sample, such as those mortgages where borrowers received
counseling, mortgages for first-time homebuyers, and mortgages on homes purchased from
builders. Assuming a universe of 10 million mortgages annually (or a random simple sample of
8

125,000 mortgages quarterly), a measurement error of 3 percent, 40 percent overall response rate,
the survey sample will be sent quarterly to approximately 7,000 borrowers.
The first Freddie Mac survey was conducted in 2010. The sampling frame for that survey was tied
to a prototype database supplied by Experian that was one-twenty-fifth the size of the National
Mortgage Database. A random sample of 1,500 loans originated between January and December
2009 and consisting of a mix of purchases and refinances of primary residences, second homes, and
investment properties was selected from the Experian credit bureau data. Simmons Research, a
contractor, was retained to conduct a mail survey, which was fielded from May 25 to July 22, 2010.
The 2010 Freddie Mac survey validated the use of the credit repository database from which to
draw the sampling frame, pretested and fielded the survey components of the project, and allowed
for evaluation of the quality and feasibility of matching external data elements to the repository
data. Relatively few problems were uncovered with the survey questions, but the response rate was
low—only 16 percent. In order to improve the response rate for its second survey, Freddie Mac
engaged Dr. Don A. Dillman, a leading national expert in mail surveys, to help in revising the
presentation of the survey questionnaire.
The second Freddie Mac survey was conducted in 2011. For the 2011 survey, a revised random
sample of 1,000 borrowers on mortgages originated between October and December 2010 and sold
to Freddie Mac was drawn from the Experian database. The sample represented a mix of purchases
and refinances of primary residences, second homes, and investment properties. Freddie Mac
fielded a mail survey from February 28 to April 29, 2011. Based upon Dr. Dillman’s advice, the
survey questions remained the same as those used in the 2010 survey, but a number of changes in
the appearance of the questionnaire, survey communications, and packaging were made. As a
result, Freddie Mac’s 2011 survey had a vastly improved overall response rate of 60 percent.
The third Freddie Mac survey was conducted in 2012. Dr. Dillman remained a consultant on the
2012 survey and reviewed all proposed changes prior to survey execution. The execution and
communication strategies for the 2012 mail survey were the same as those for the 2011 survey,
except for one minor change to the questionnaire to clarify skip directions relating to whether the
borrower had a refinanced loan or a home purchase loan. For the 2012 survey, a representative
sample of 5,000 borrowers on newly originated mortgages was drawn from the credit repository
database. Simmons Research was again retained to conduct the survey, which was fielded between
September 17 and November 19, 2012. Although unauthorized changes to the communications
package made by Simmons Research lowered the response rate for the 2012 survey, the response
rate was still over 40 percent, which confirmed the effectiveness of the repository sampling frame.
The changes made by Simmons and their effect on the response rate are discussed in more detail in
response to Question B.2.
Table 1 provides further detail of responses to the mail questionnaire for the 2011 and 2012
surveys. In 2011, 1,000 mail surveys were mailed to borrowers at the mortgage property
address. The net sample consisted of 928 borrowers after 72 packages were returned as
undeliverable. A total of 556 questionnaires, or 60 percent, were returned completed: 436 in the
first mailing and 120 in the second mailing. Of these only 527 questionnaires had usable
9

responses, thus lowering the effective response rate to 57 percent. The difference was screened out
in question 1, where 25 respondents indicated they did not have a mortgage in 2011. Purchase
loans made up 46 percent of responses and refinances made up the remaining 54 percent of
responses.
The 2012 Freddie Mac survey consisted of a sample of 5,000 borrowers. 54 survey questionnaires
were returned as undeliverable, for a net sample of 4,946. The returned questionnaires totaled
2,167, or 44 percent. In the first mailing, 1,739 were returned, accounting for 80 percent of the
returned surveys. 186 of the 2,167 returned questionnaires were not usable because the
respondents indicated that they had no mortgage in 2012. Accordingly, 1,981 of the returned
questionnaires, or 40 percent of the net sample, had usable responses. New loans accounted for 39
percent and refinances accounted for 61 percent of the useable returns in the 2012 survey.
Table 2 breaks down non-response rates by question for the 2011 and 2012 Freddie Mac surveys.
Improvements in response rates between the two surveys are attributable to the clarification of skip
directions relating to whether the borrower had a refinanced loan or a home purchase loan. For the
majority of the questions the non-response rate was less than 3 percent of respondents. Generally
the non-response rate improved from 2011 to 2012 with one exception (question 74).

Question 2. Describe the procedures for the collection of information, including
• Statistical methodology for stratification and sample selection,
• Estimation procedure,
• Degree of accuracy needed for the purpose described in the justification,
• Unusual problems requiring specialized sampling procedures, and
• Any use of periodic (less frequently than annual) data collection cycles to reduce
burden.
Data for the NSMB will be collected through a single-blind mail survey format. Questionnaires
will be mailed to the borrower(s) on the mortgage at the property address associated with the
mortgage. Borrowers are not identified on the questionnaire and, therefore, respondents will
remain anonymous. A key is maintained to track respondents while maintaining their anonymity.
A mail survey reduces costs but can substantially lower response rates. The methodology for
maximizing response rate is discussed earlier in Question 1. The response rate for the usable
survey is expected to be 40 to 57 percent.
The survey’s sample selection is discussed earlier in Question 1. The NSMB is a simple, random
sample of mortgage originations and is not stratified. Alternatives designed for stratifying,
clustering, or cut-off samples were not considered. While subpopulations are of interest, the key
purpose of the survey is to gather more information on the origination characteristics of the
sampled borrowers. Not enough information exists on the relationship between origination and
subpopulation to demonstrate that cut-off samples would be more precise or bias free. As more
information is available as an outgrowth of this survey, this issue could be revisited. No estimate
10

procedures are needed in interpreting the survey responses. Similarly, no hypotheses are being
tested and no unusual problems exist that require specialized sampling procedures.

Question 3. Describe methods to maximize response rates and to deal with issues of nonresponse. The accuracy and reliability of information collected must be shown to be
adequate for intended uses. For collections based on sampling, a special justification
must be provided for any collection that will not yield "reliable" data that can be
generalized to the universe studied.
As mentioned in the response to Question 1, Freddie Mac hired Dr. Don Dillman, a leading expert
in modern survey methods, to guide its 2011 survey after the survey response rate for its 2010
survey had been only 16 percent. Dr. Dillman’s recommendations resulted in important
improvements in three critical areas of the survey: the execution/implementation; the
communications package; and the questionnaire content and format. These changes greatly
improved the response rate, which improved to 60 percent for the 2011 survey.
One important recommendation that was implemented for the 2011 and 2012 Freddie Mac surveys
was to have four planned mail contacts with borrowers. The first contact with potential
respondents consisted of the questionnaire, an upfront five dollar incentive, and a cover letter. The
second contact, a reminder letter sent to all respondents, occurred in the second week of
implementation. The third contact, in the fifth week, was sent only to non-responders and included
a second reminder letter, another copy of the questionnaire, and a second five dollar incentive. The
final contact, a third reminder letter, was sent in the seventh week to non-responders only. A due
date for returning the survey questionnaire was included in the last mailing, which closed the
communication loop with potential respondents.
Another adopted recommendation resulted in a change to two features of the communications
package. First, the tone of the communications was made friendlier to reflect a personal and
sincere request for help. All correspondence was signed by an officer of the survey sponsor
(Freddie Mac) and included contact information for authenticity. Second, the questionnaires were
mailed in a plain white envelope so as not be mistaken for “junk mail”. The envelope had only the
return address for Freddie Mac visible to provide legitimacy to the survey. These features were
incorporated successfully into the materials for the 2011 Freddie Mac survey, leading to a dramatic
increase in the response rate.
Freddie Mac contracted with Simmons Research to conduct the survey in 2012. Simmons
Research failed to follow Dr. Dillman’s protocols for the communications package, which likely
gave recipients the impression that the envelopes contained “junk mail” and resulted in a lowerthan-expected response rate of 44 percent for the 2012 survey. Specifically, Simmons Research:
used address labels instead of hand writing the addresses on the envelop; did not use the plain
white envelope that was specified; and printed Experian’s logo instead of Freddie Mac’s logo on
the return envelope, which created an inconsistency with the information in the cover letter.

11

A critical lesson from the 2012 Freddie Mac survey is that Dr. Dillman’s recommendations on the
communications package need to be followed exactly to avoid: (1) the survey mailing being
perceived as junk mail by the borrower; (2) undermining the legitimacy of the survey as a FHFA
effort by using Experian’s address on the return envelope; and (3) inconsistency and confusion
about who is conducting the survey—i.e., FHFA and not Experian. The NSMB will incorporate all
of Dr. Dillon’s recommendations regarding the multiple mail contacts and the content, tone and
appearance of the communications package.
To further increase response rates, survey recipients will also be given the option of completing the
survey online in either English or Spanish. The first mailing will contain an insert, in both English
and Spanish, that will inform recipients of those options and provide the appropriate web addresses
to access those electronic versions of the survey. The insert will also provide Spanish-speaking
recipients with a telephone number through which they may request a paper copy of the survey in
Spanish. In addition, the survey package will include instructions on how to log onto the FHFA
website and view a video in which the Director of FHFA will explain the importance of the survey
and thank recipients for taking the time to complete the questionnaire.
In the NSMB, two types of non-response may occur. The first type is unit nonresponse. It occurs
when data is not obtained for the sample unit, i.e., a respondent chooses not to participate in
survey. The second type is item non-response. It occurs when a respondent fails to answer one or
more questions on the survey. To correct for these biases, non-response adjustment weights will be
created and applied to the survey responses.
National Mortgage Database staff will document unit and item non-response at the conclusion of
each survey. Survey response will be evaluated to determine if any biases between respondents
and non-respondents exist. Other administrative data will be used to compare respondents and
non-respondents characteristics to determine if response rates between the two groups vary. Staff
will use proven methodology to create non-response adjustments. Weights will be employed to
estimated probability of response and actual response rates for mutually exclusive subgroups of the
sample. Based on its findings, National Mortgage Database staff will determine future steps in
addressing any biases found.
Response rates will be tracked and reported based on the week the survey response was received.
All responses will be totaled two weeks after the third reminder letter (sent in Week 7) and divided
by all survey units that were identified as being eligible. The numerator of the response rate will
include all surveys that included all the required items for the report period.

12

Question 4. Describe any tests of procedures or methods to be undertaken. Testing is
encouraged as an effective means of refining collections of information to minimize
burden and improve utility. Tests must be approved if they call for answers to
identical questions from 10 or more respondents. A proposed test or set of tests may
be submitted for approval separately or in combination with the main collection of
information.
As explained in detail above, the three surveys conducted by Freddie Mac on its own initiative
between 2010 and 2012 allowed for extensive testing of the procedures and methods that have now
been adopted for the National Mortgage Database generally and for the NSMB specifically. FHFA
has worked closely with those that planned and executed the Freddie Mac surveys to ensure that
the lessons learned in carrying out those surveys are reflected in the procedures and methods for
the NSMB. Therefore, FHFA has no plans to undertake any further testing of these methods or
procedures, or of the survey questions, prior to fielding the first survey in the first quarter of 2014.

Question 5. Provide the name and telephone number of individuals consulted on
statistical aspects of the design and the name of the agency unit, contractor(s),
grantee(s), or other person(s) who will actually collect and/or analyze the information
for the agency.
All stakeholders including those who will analyze the data are provided in the list in Table 3. The
name and telephone number of individuals consulted on statistical aspects of the design include
staff from CFPB, NMDB and Freddie Mac, and Dr. Don Dillman.

13

Table 1
Summary of Survey 2 and Survey 3 Responses
Bounce Net
1st
2nd
Sample Backs Sample Mailing Mailing
Survey 2
1,000
72
928
436
120
Percent of Surveys
78% 22%
Response Rate
Survey 3
5,000
54
4,946 1739 428
Percent of Surveys
80% 20%
Response Rate

Surveys
No
Useable
Useable Responses
Returne Mortgage Response Purchase Refinance Unknown
556
25
527
241
286
0
46%
54%
0%
60%
57%
2,167
186
1,981
778
1,198
5
39%
61%
0%
44%
40%

14

Table 2
Non-response Comparison
Survey 2 and Survey 3
NO ANSWER
LOAN APPLICATION PROCESS
S3%
S2%
Q3 Co-signers
**
**
Q4 Mortgage literacy – familiar [+]
**
3.8
Q5 Useful services NEW [+]
**
na
Q6 Mortgage application with
*
**
Q7 First contact
*
**
Q8 Choosing a lender – important [+]
9.7
13.1
Q9 Idea of mortgage you wanted [+]
**
**
Q10 Information lender provided NEW [+]
6.1
na
Q11 Lender ask you about NEW [+]
9.2
na
Q12 Actions taken during application [+]
3.6
10.9
Q13 Choosing a mortgage – important NEW[+]
4.5
na
Q14 Satisfaction with mortgage NEW [+]
**
na
Q15 Satisfaction with lender [+]
**
**
Q16 Get Counseling
*
**
Q31 Unpleasant surprises
**
**
MORTGAGE
Q20 Mortgage amount
**
5.9
Q21 Monthly payment
**
5.1
Q22 Taxes/Insurance included
*
**
Q23 Interest rate
3.5
5.7
Q24 ARM loan
*
**
Q25 Discount points paid
**
3.2
Q26 Origination points paid
**
4.4
Q27 Rate lock
**
**
Q28 Prepayment penalty
**
**
Q28 Credit life insurance
**
3.6
Q28 Escrow account
**
**
Q28 Non-amort payment [NEW]
4.8
na
Q29 Take out another loan (=No)
*
18.2
Q30 Amount of 2nd loan
16
*
Q45 Acquisition cost
5.1
6.7
Q49 Property value today
3.1
3.2
NEW LOAN
Q29 Take out another loan (=Yes)
**
11.4
Q33 Pre-approval
**
*
Q34 Proceeds from another sale
*
**
Q35 How much money used
7.5
12.3
Q36 Down payment percent
**
**
Q37 Down payment sources [+]
30.5
35.3
Q38 Closing costs paid [+]
37.4
44.7
* < 1%
** < 3%
[+] Multiple response question – worst case provided

15

Table 2 (continued)
Non-response Comparison
Survey 2 and Survey 3
NO ANSWER
REFINANCE
S3 %
S2%
Q39 Why refinance-important [+]
14.2
22.7
Q40 Refinanced amount
9.1
24.5
Q41 Size of new loan
3.6
14.3
Q42 Use of new money [+]
27
19.5
MORTGAGED PROPERTY
Q43 When acquired property
2.9
7.1
Q44 Investment or good place to live
**
**
Q46 Acquire from
**
**
Q47 Type of dwelling
*
*
Q50 Rent property
**
**
Q52 Property primary residence
**
4.4
Q53 When moved into property
10.7
15.8
HOUSEHOLD DEMOGRAPHICS
Q55 Marital status
**
**
Q57 Age
4
4.8
Q58 Gender
**
**
Q59 Owned other properties
**
3.2
Q60 Education
**
**
Q61 Hispanic/Latino
**
**
Q62 Race
**
4
Q63 Work status
**
**
Q67 Others in household
**
3.4
Q70 Help pay expenses
4.1
*
Q71 Income
6.3
8.4
  Q72 Change in income NEW
3.4
na 
LIFE EVENTS/EXPECTATIONS
Q73 Life events [+]
7.8
8.8
Q74 Financial events [+]
8.4
6.5
Q75 Financial crisis [+]
7.3
10.3
Q76 Financial knowledge NEW [+]
3.8
na
Q77 Financial risk
2.7
4.4
Q78 Agree/disagree NEW [+]
6.4
na
Q79 Income vs cost of living
3.3
5.3
Q80 Future home prices
**
4
Q81 Likely to move/refinance [+]
5.2
9.1
Q82 Likelihood of a personal crisis [+]
6.3
7
* < 1%
** < 3%
[+] Multiple response question – worst case provided

16

17