MIRS Supporting Statement 2007

"MONTHLY SURVEY OF RATES AND TERMS ON CONVENTIONAL 1FAMILY NONFARM MORTGAGE LOANS (MIRS)"
OMB NUMBER 3069-0001
SUPPORTING STATEMENT
A.

Justification

1. The Federal Housing Finance Board's (Finance Board) predecessor, the former Federal Home
Loan Bank Board (FHLBB), first provided data concerning a survey of mortgage interest rates in
1963. No statutory or regulatory provision explicitly required the FHLBB to conduct this survey,
commonly known as the Monthly Interest Rate Survey or MIRS, although references to the MIRS
did appear in several federal and state statutes. Responsibility for conducting the MIRS was
transferred to the Finance Board upon dissolution of the FHLBB in 1989. See Financial
Institutions Reform, Recovery, and Enforcement Act of 1989 (FIRREA), Pub. L. No. 101-73, tit.
IV, § 402(e)(3)-(4), 103 Stat. 183, codified at 12 U.S.C. § 1437 note, and tit. VII, § 731(f)(1) and
(f)(2)(B), 103 Stat. 433 (Aug. 9, 1989). In 1993, the Finance Board promulgated a final rule
describing the method by which it conducts the MIRS. See 58 Fed. Reg. 19195 (Apr. 13, 1993),
codified at 12 C.F.R. § 906.3. Since its inception, the MIRS has provided the only consistent
source of information on mortgage interest rates and terms and house prices for areas smaller
than the entire country.
Statutory references to the MIRS include the following:
•

Pursuant to their respective organic statutes, Fannie Mae and Freddie Mac use the MIRS
results as the basis for required annual adjustments to the maximum dollar limits for their
purchase of conventional mortgages. See 12 U.S.C. §§ 1454(a)(2) and 1717(b)(2). The
Fannie Mae and Freddie Mac limits were first tied to the MIRS by the Housing and
Community Development Act of 1980. See Pub. L. No. 96-399, tit. III, § 313(a)-(b), 94 Stat.
1644-1645 (Oct. 8, 1980). At that time, the nearly identical statutes required Fannie Mae and
Freddie Mac to base the dollar limit adjustments on "the national average one-family house
price in the monthly survey of all major lenders conducted by the [FHLBB]." See 12 U.S.C.
§§ 1454(a)(2) and 1717(b)(2) (1989). When Congress abolished the FHLBB in 1989, it
replaced the reference to the FHLBB in the Fannie Mae and Freddie Mac statutes with a
reference to the Finance Board. See FIRREA, tit. VII, § 731(f)(1) and (f)(2)(B), 103 Stat.
433.

•

Also in 1989, Congress required the Chairperson of the Finance Board to take necessary
actions to ensure that indices used to calculate the interest rate on adjustable-rate mortgages
(ARMs) remain available. See FIRREA, tit. IV, § 402(e)(3)-(4), 103 Stat. 183, codified at 12
U.S.C. § 1437 note. At least one ARM index, known as the National Average Contract
Mortgage Rate for the Purchase of Previously Occupied Homes by Combined Lenders, is
derived from the MIRS data. The statute permits the Finance Board to substitute a different
ARM index after notice and comment only if the new ARM index is based upon data

substantially similar to that of the original ARM index and substitution of the new ARM
index will result in an interest rate substantially similar to the rate in effect at the time the
new ARM index replaces the existing ARM index. See 12 U.S.C. § 1437 note.
•

Congress indirectly connected the high cost area limits for mortgages insured by the Federal
Housing Administration (FHA) of the Department of Housing and Urban Development to the
MIRS in 1994 when it statutorily linked these FHA insurance limits to the purchase price
limitations for Fannie Mae. See Pub. L. No. 103-327, 108 Stat. 2314 (Sept. 28, 1994),
codified at 12 U.S.C. § 1709(b)(2)(A)(ii).

•

The Internal Revenue Service uses the MIRS data in establishing “safe-harbor” limitations
for mortgages purchased with the proceeds of mortgage revenue bond issues. See 26 C.F.R.
§ 6a.103A-2(f)(5).

•

Statutes in several states and U.S. territories, including California, Michigan, Minnesota,
New Jersey, Wisconsin and the Virgin Islands, refer to, or rely upon, the MIRS. See, e.g.,
Cal. Civ. Code 1916.7 and 1916.8 (mortgage rates); Mich. Comp. Laws 445.1621(d)
(mortgage index rates); Minn. Stat. 92.06 (payments for state land sales); N.J. Rev. Stat.
31:1-1 (interest rates); Wis. Stat. 138.056 (variable loan rates); V.I. Code Ann. tit. 11, sec.
951 (legal rate of interest).

2. The information collection is used for the purposes described in the answer to Question 1. In
addition, the Finance Board uses the information collection to produce the MIRS and for general
statistical purposes and program evaluation. Economic policy makers, such as the Board of
Governors of the Federal Reserve System and the Council of Economic Advisors, use the MIRS
data to determine trends in the mortgage markets, including interest rates, down payments, terms
to maturity, terms on ARMs, and initial fees and charges on mortgage loans. Other federal
banking agencies use the MIRS results for research purposes. Information concerning the MIRS
is published regularly in the popular and trade press, in Finance Board press releases, including a
monthly 1-page ARM index release, a monthly 8- or 12-page release with mortgage rate and
term data, and an annual summary all available via the Finance Board’s web site, and in
publications of other federal agencies, including The Economic Report of the President and
Statistical Abstract of the United States. The Finance Board also responds to approximately
1,000 telephone calls annually and 1,400 to 1,900 inquiries monthly to its automated telephone
answering system for the ARM index and other MIRS information.
3. The Finance Board collects approximately 95 percent of the responses to the information
collection electronically. The Finance Board has strongly encouraged respondents to report
electronically since first instituting electronic reporting in November 1991.
4. The information collection avoids duplication by collecting all necessary information on a
single form known as Form FHFB 10-91 or through the MIRS software. The information
collected generally is not available from any other source and provides the only national survey of
home prices covering both new and existing homes.
2

5. The information collection does not have a significant economic impact on a substantial number
of small entities.
6. Pursuant to the statutory requirements described in the answer to Question 1, the Finance Board
conducts the information collection on a monthly basis. Since data on mortgage rates and terms and
home prices is vital economic information, collecting the information less frequently would affect
the ability of policy makers to make informed judgments about economic conditions. In addition,
most ARM contracts linked to the MIRS rely upon the data provided monthly.
7. There are no special circumstances that require the Finance Board to conduct the information
collection in a manner inconsistent with the guidelines in Question 7.
8. In accordance with the requirements of 5 C.F.R. § 1320.8(d), the Finance Board published a
request for public comments regarding this information collection in the Federal Register on
April 11, 2007. See 72 Fed. Reg. 18246 (Apr. 11, 2007). The 60-day comment period closed on
June 11, 2007. The Finance Board received no comments
9. The Finance Board does not provide a payment or gift to any respondent.
10. The Finance Board does not provide any assurance of confidentiality to respondents. However,
when the Finance Board responds to requests for individual loan data, which is similar to
information contained in public land records, it does not disclose the identity of the lender or any
loan identification number. The MIRS report contains no borrower identifier more specific than a
ZIP code.
11. The information collection does not include questions of a sensitive nature.
12. The estimated annualized cost to respondents of the information collection is $21,000. The
estimated annualized hour burden is 600 hours. The Finance Board based these estimates on the
following calculations:
Clerical staff completes and submits Form FHFB 10-91 or provides information through the MIRS
software.
ƒ
ƒ
ƒ
ƒ
ƒ

Review/processing time: 0.5 hours
Total forms: 1,200 (200 respondents x 6 monthly reports)
Total hours: 600
Hourly rate: $35 (includes salary, benefits, and overhead)
Total cost: $21,000

13. The Finance Board has not identified any costs to respondents other than the costs discussed
in detail in the answer to Question 12.

3

14. The estimated annualized cost to the Finance Board of the information collection is $80,125.
The estimated annualized hour burden is 1,144 hours. The Finance Board based these estimates
on the following:
ƒ

$80,125 for Finance Board staff salaries and benefits, which includes 40 percent of the time
of the principal analyst, 10 percent of the time of a senior information technology specialist,
5 percent of the time of the program supervisor, and a 25 percent allowance for benefits.

15. The reduction reported by the Finance Board in Item 13 of OMB Form 83-I is the result of
consolidation in the mortgage lending industry.
16. The Finance Board issues several types of releases based on the MIRS data in tabular form
including a monthly 1-page ARM index release, a monthly 8- or 12-page release with mortgage
rate and term data, and an annual summary all available via the Finance Board’s web site. The
Finance Board does not use any complex analytical techniques.
17. The Finance Board plans to display the expiration date for OMB approval of the information
collection.
18. There are no exceptions to the certification statement identified in Item 19 of OMB Form 83-I.
B. Collections of Information Employing Statistical Methods
1. The potential respondent universe includes the following:
•

A target sample of 105 of approximately 1,074 savings associations. In 2006, 69 savings
associations responded to the information collection.

•

A target sample of 45 of approximately 3,480 commercial banks and savings banks with (a) total
assets greater than $1 billion with mortgage-related assets of 10 percent, or (b) total assets less
than $1 billion with mortgage-related assets of 20 percent. In 2006, 39 commercial banks
responded to the information collection.

•

A sample of mortgage companies. Since mortgage companies are not subject to uniform
regulation or registration and are not subject to geographic restrictions on their operation, the
total universe and target sample are difficult to quantify. In 2006, 13 mortgage companies
responded to the information collection, including several very large mortgage companies that
operate in multiple states.

Approximately 50 percent of the lenders in the sample respond either every month or almost every
month, 25 percent respond infrequently, and the remaining 25 percent respond about every other
month. Lenders that do not respond in a 12-month period generally are dropped from the sample. In
2006, the Finance Board collected information on over 217,000 loans.

4

2. The survey is stratified on the basis of lender type and geographical location. The Finance Board
publishes monthly aggregate data by lender type, quarterly data for 32 selected large metropolitan
areas and the 12 Federal Home Loan Bank (Bank) districts, and annual data for all states. The
sampling method draws savings banks, commercial banks and savings associations with equal
probabilities of selection from a lender type/geo stratum.
The MIRS data is post-stratified or balanced on the basis of two considerations. First, the shares of
weighted loans are adjusted so the frequency distribution of weighted loans by lender type matches
the lender-type distribution in the latest Home Mortgage Disclosure Act (HMDA) release of the
Board of Governors of the Federal Reserve System. Second, the resulting weights are adjusted so
the frequency distribution of weighted loans by Bank district matches the geographic pattern of
mortgage originations annually reported in the HMDA file. Finally, individual loan weights are
capped at three standard deviations over the mean weight for the particular lender type. As a
practical matter, the balancing has the effect of weighting up loans reported by commercial banks
and weighting down loans reported by savings associations.
The Finance Board believes the national MIRS data is reliable and accurate for its intended
purposes. For some of the less populous states and smaller metropolitan areas, the inferences drawn
must be qualified by the sample size. The Finance Board tested the accuracy of the mortgage rate
data by comparing the average rate on 30-year conforming fixed-rate loans with similar data
reported in Freddie Mac's commitment-rate survey. After adjusting for a one month lag between
mortgage commitment and mortgage closing, the two series differed by an average of 0.04 percent.
It is not possible to test directly the reported home price data against other home-price series. The
data reported by the National Association of Realtors is median price data on purchase transactions
and includes some homes financed by FHA-insured mortgages. The MIRS specifically excludes
FHA mortgages and reports average, not median, prices. The new home price series of the Census
Bureau also includes some homes financed with FHA mortgages. The repeat-sales price index
jointly published by Fannie Mae and Freddie Mac excludes jumbo and other types of
nonconforming loans that are within the scope of the MIRS, and the repeat-sales index is likely to
include fewer ARMs than the MIRS.
3. The Finance Board prepares an internal monthly report indicating whether a lender responded to
the information collection and the number of loans it reported. Follow-up procedures generally
include the following:
•

Asking any of the 10 largest respondents that failed to report to do so before the monthly data is
tabulated.

•

Contacting a regular respondent by telephone if the respondent has failed to report for two or
three months. In most cases, the break in reporting is due to personnel changes at the
respondent.

•

Sending out address update forms each May to all respondents and again in July to respondents
that do not return the form.
5

•

Notifying respondents each October of the urgency of reporting October data for the purposes of
determining the Fannie Mae and Freddie Mac conforming loan limit.

•

Enlisting the survivor when a respondent merges or is acquired. These efforts are met with
limited success.

The Finance Board also addresses the issue of non-response by having the individual weights be the
number of lenders in the lender type/geo stratum divided by the number that actually report from
that lender type/geo stratum, including those lenders that report closing no loans. The reporting
system allows a no loans closed response, and lenders are periodically reminded of the importance
of this feature in the weighting scheme.
4. Form FHFB10-91 and the MIRS software were tested when last revised in June 1999. No testing
with individual respondents has occurred in conjunction with this request.
5. The sampling and weighting framework are attributable to Dr. Martin R. Frankel. The
framework, established in late 1991, has not been changed in any material respect. At that time, Dr.
Frankel was on the faculty of Baruch College, City University of New York. Dr. Frankel has not
been contacted in conjunction with this request.
The supervisory program official at the Finance Board is Dr. Joseph A. McKenzie, Deputy Chief
Economist, Office of Supervision, 202-408-2845 or [email protected]. The analyst responsible
for the day-to-day operation of the MIRS is David Roderer, 202-408-2540 or [email protected].

6