Monthly Survey of Rates and Terms on Conventional 1-Family Nonfarm Mortgage Loans

FEDERAL HOUSING FINANCE BOARD

[97-N-1]

Monthly Survey of Rates and Terms on Conventional 1-Family

Nonfarm Mortgage Loans

AGENCY: Federal Housing Finance Board.

ACTION: Request for comments.

-----------------------------------------------------------------------

SUMMARY: The Federal Housing Finance Board (Finance Board) is seeking

comments on several aspects of its Monthly Survey of Rates and Terms on

Conventional 1-Family Nonfarm Mortgage Loans. The Finance Board seeks

comments on whether it should continue to publish mortgage information

by lender type. If not, then the Finance Board seeks comments on

whether the sampling and weighting design for this survey should draw

lenders without regard to lender type. If so, the Finance Board seeks

suggestions for alternative sampling and weighting methodologies. The

Finance Board also seeks comments on the designation of successor

adjustable-rate mortgage indexes if it decides to stop publishing data

by lender type.

DATES: Comments must be received by April 18, 1997.

ADDRESSES: Mail comments to Elaine L. Baker, Executive Secretary,

Federal Housing Finance Board, 1777 F Street, N.W., Washington, D.C.

20006. Comments will be available for inspection at this address.

FOR FURTHER INFORMATION CONTACT: Joseph A. McKenzie (202) 408-2845,

Associate Director, Office of Policy, Federal Housing Finance Board,

1777 F Street, N.W., Washington, D.C. 20006.

SUPPLEMENTARY INFORMATION:

A. Background

The Finance Board is responsible for conducting the Monthly Survey

of Rates and Terms on Conventional 1-Family Nonfarm Mortgage Loans.

This survey, usually called the ``Monthly Interest Rate Survey'' or

``MIRS,'' asks a sample of approximately 350 mortgage lenders to report

the terms and conditions on all conventional mortgage loans for the

purchase of single-family, nonfarm homes that they close during the

last five working days of the month. The sample of lenders includes

savings associations, mortgage companies, commercial banks, and savings

banks that have volunteered to participate in the survey. MIRS provides

national and regional data on mortgage interest rates, mortgage terms,

and house prices. The Finance Board's regulations describe MIRS more

thoroughly. See 12 CFR 902.3.

From 1963 to September 1989, the former Federal Home Loan Bank

Board conducted MIRS. Law requires the Finance Board to conduct this

survey. The statutory mandate to conduct MIRS appears in identical

provisions in the Federal National Mortgage Association (Fannie Mae)

Charter Act, 12 U.S.C. 1717(b)(2), and the Federal Home Loan Mortgage

Corporation (Freddie Mac) Act, 12 U.S.C. 1454(a)(2). These provisions

allow the two agencies annually to adjust the maximum size of mortgage

loans that they can purchase or guarantee by the October-over-October

percentage price change in house prices as reported in MIRS.

More recently, the 1994 Department of Housing and Urban Development

(HUD) appropriation act tied the high-cost area limits for Federal

Housing Administration (FHA)-insured mortgages to the purchase-price

limitations of Fannie Mae and Freddie

[[Page 9768]]

Mac, thus linking the FHA limits indirectly to MIRS. See Department of

Veterans Affairs and Housing and Urban Development, and Independent

Agencies Appropriations Act, Pub. L. No. 103-327, 108 Stat. 2298

(1994). In addition, the Internal Revenue Service uses the data from

MIRS to set the safe-harbor purchase-price limits for mortgages

purchased with the proceeds of mortgage revenue bond issues. See 26 CFR

6a.103A-2(f)(5).

Beyond its use for indexing the conforming loan limit, MIRS

provides information for general statistical purposes and program

evaluation. Economic policy makers use the data to determine interest

rates, down payments, terms to maturity, terms on adjustable-rate

mortgages (ARMs), initial fees and charges on mortgage loans, and other

trends in mortgage markets. Information from MIRS regularly appears in

the popular and trade press.

On or about the 26th of each month the Finance Board publishes a

MIRS press release with mortgage rate and term information by property

type (all, newly built, and previously occupied; Table I), by loan type

(adjustable-rate and fixed-rate; Table II), and by lender type (savings

association, mortgage company, commercial bank, savings bank; Table

III), and a table providing data on 15- and 30-year conforming fixed-

rate loans (Table V). In addition, it publishes quarterly tables with

rate and term information for metropolitan areas (Table IV) and for

Federal Home Loan Bank districts (Table VI).

An ARM index derived from MIRS--the National Average Contract

Mortgage Rate for the Purchase of Previously Occupied Homes--was the

only ARM index that Federally chartered savings institutions could use

for a period in the early 1980's. A very small proportion of existing

ARMs may use another interest-rate series from MIRS as an index.

B. Sampling and Weighting the Data

The Finance Board samples all savings associations, mortgage

companies, commercial bank, and savings banks for MIRS because it

publishes monthly aggregate data by lender type. In addition, the

Finance Board samples lenders representing all regions because it

publishes quarterly data for 32 selected large metropolitan areas,

quarterly data for the 12 Federal Home Loan Bank districts, and annual

data for all 50 states and for 60 metropolitan statistical areas

(MSAs).

MIRS presents a ``clustered sampling'' problem. The item of

interest is individual loans, but the Finance Board must sample lenders

to get the individual loan data. The loans must come from all regions

and must represent all lender types. Several recent developments have

improved the geographical dispersion of MIRS loans. First, some large

national mortgage companies participate in MIRS. This means that one

lender may report loans from 20 or more states. Second, the continuing

trend toward the consolidation of depository institutions has resulted

in large institutions that originate loans in many states.

As with most survey data, the tabulated MIRS data reflects the

weighting of the individual responses. The current weighting draws

depository institutions with equal probabilities of selection from

``lender-type geo strata'' (for example, commercial banks in Nebraska,

savings associations from the Cincinnati MSA, or savings banks from the

Boston CMSA.) Since the sample of loans reported in a given month may

differ from true lending experience (for example, over -or under-

represent certain regions), the MIRS data is weighted to comport with

information on lending patterns derived from independent sources:

(1) The data is adjusted so that the distribution of loans by

lender type matches the lender-type distribution in the latest release

of HUD's Survey of Mortgage Lending Activity, and

(2) The data is adjusted so that the distribution of loans by

Federal Home Loan Bank district matches the state pattern of mortgage

originations annually reported by HUD.

The weighting process builds up the national data from four

separate subsamples based on lender type, where the shares of loans by

lender type come from the HUD data. On balance, this weighting process

significantly increases the importance of loans reported by commercial

banks and reduces the importance of loans reported by savings

associations because commercial bank loans are under-represented in the

sample. Regional adjustment of the data does not have a significant

effect on the results because the geographic pattern of responses

approximates aggregate lending patterns.

C. Sampling by Lender Type

The Finance Board publishes data by lender type principally because

the former Federal Home Loan Bank Board published the data that way

when it conducted MIRS. Accordingly, the Finance Board draws four

separate subsamples corresponding to savings associations, mortgage

companies, commercial banks and, savings banks. As the financial

services sector evolves, the distinctions between commercial banks and

thrifts continue to erode. If the institutional distinctions between

commercial bank and thrift are blurred, then published data by lender

type may no longer be useful or meaningful.

While the overall samples of savings associations, savings banks,

and mortgage companies are adequate, the Finance Board has had

persistent trouble in recruiting commercial banks for the sample. Over

the past several years, the Finance Board has contacted more than 2,000

commercial banks, all with at least 10 percent of their assets in

residential mortgage loans, and asked them to participate in MIRS. Most

of the banks contacted never responded to the solicitation. Many banks

that did respond said that either they make no mortgages or that a

subsidiary mortgage company originates all the loans that they hold.

Many banks that responded positively never submitted any loan data.

Despite the Finance Board's recruitment efforts, only 118

commercial banks reported a total of 5,437 loans in 1996. This

represents only 4 percent of the total number of loans reported in

1996. However, HUD's Survey of Mortgage Lending Activity reports that

commercial banks originate about one-quarter of all single-family

mortgage loans. As a result, the MIRS weighting process weighs up each

commercial bank loan by a factor of about six.

While the MIRS sample has few large commercial banks, the overall

sample contains many loans originated by the mortgage banking

subsidiaries of large commercial banks that have large mortgage

investments.

The Finance Board specifically requests comments on the following:

--Should it continue to report MIRS data by lender type?

--Should it continue to sample MIRS lenders by lender type?

--Do institutional changes render the data by lender type meaningless?

--Are there alternative ways to increase commercial bank participation

in the sample?

D. Home Mortgage Disclosure Act Data

The HUD data on mortgage originations by lender type is crucial to

the MIRS weighting process. However, some observers believe the HUD

data may overstate the commercial bank share of mortgage originations.

Very few large commercial banks originate mortgage loans. Most of the

large commercial banks with significant portfolio concentrations of

residential mortgages have purchased these loans from subsidiary

mortgage companies

[[Page 9769]]

that have significant origination volumes.

Home Mortgage Disclosure Act (HMDA) data may provide an alternative

data source for the lender type shares for MIRS. HMDA requires lenders

to submit information on single-family mortgage applications. The data

includes a disposition code, so it is possible to use HMDA information

on loans closed. The scope of the HMDA data includes information on all

nonmetropolitan mortgage originations but from the smallest lenders.

The more important of these omissions is loans in nonmetropolitan

areas. Approximately one-fifth of the nation's population lives outside

metropolitan areas. Secondly, very small lenders are not subject to

HMDA reporting. The Finance Board specifically requests comments on

whether it could or should use the HMDA data as the basis for

developing the lender-type adjustment in the MIRS weighting process.

The Finance Board also requests comments on whether another data source

is available that it could use in developing shares of aggregate

lending by lender type.

Beyond the use of the HMDA data to develop the lender-type

adjustment, the Finance Board requests comments on whether it could

develop a size-stratified weighting scheme based on individual lender

origination volumes reported in the HMDA data. A HMDA-based weighting

scheme would group lenders by origination volume and sample lenders,

without regard to charter type, with decreasing frequency (and

increasing weight) as origination volume declines. The implicit

assumption is that loans originated by one type of lender (for example,

commercial banks) are no different from loans originated by another

type of lender.

The Finance Board requests comments on whether it should change its

MIRS weighting methodology. Should it adopt a size-stratified weighting

methodology using HMDA data? If so, how should it surmount the omission

in the HMDA data of nonmetropolitan lending data and loans from small

lenders? (The MIRS data now contains loans from nonmetropolitan lenders

as well as loans made by metropolitan lenders in nonmetropolitan

areas.) Is there another weighting methodology that is more appropriate

than either the current methodology or the one suggested that uses the

HMDA data?

E. Data Edit Limits

Most statistical surveys incorporate certain validity checks that

the data must pass. MIRS contains validity checks or edits on allowable

interest-rate ranges, loan sizes, purchase prices, loan fee amounts,

and consistency of ZIP code with state of the property. The Finance

Board established the current maximum allowable value of $500,000 for

loan size and $750,000 for property price in November 1991. These edits

would reject loans where the responding lender omitted a decimal point

from dollar values, which would have the effect of reporting a loan

amount or purchase price 100 times larger than the actual amount. The

edits also exclude certain typographical errors, especially when the

purchase price contains an extra zero. For example, a reported $50,000

loan on a $900,000 property is more likely to be a $50,000 loan on a

$90,000 property. The current edits would reject this transaction.

While the edits screen out incorrect transactions, they also may

exclude some valid transactions. Since the Finance Board established

the current price and loan-size limits in November 1991, housing prices

have increased modestly. The Finance Board seeks comments on an

appropriate methodology to adjust the house size and loan amount edit

limits to allow for housing price appreciation. The Finance Board does

not plan to change the lower loan size and property price limit of

$10,000.

While it is not possible precisely to quantify the effect that the

changes in the edit limits will have on the reported average house

prices, the Finance Board believes the effect will be small because the

proportion of loans between the old and any higher new edit limits is

likely to be small. MIRS now has few transactions in bands just below

the current edit limits. In 1996, only 0.7 percent of MIRS loans had

balances between $400,000 and $500,000, and only 1.2 percent of MIRS

loans financed homes with prices between $500,000 and $750,000.

Transactions in these bands are skewed toward the lower end of the

bands. Therefore, the Finance Board expects that only a small fraction

of 1 percent of the survey's loans will fall between the old and any

higher new edit limits.

F. Adjustable-Rate Mortgage Index

A very small number of ARMs may use as an index a MIRS interest

rate series by lender type. This information appears on Table III of

the regular monthly MIRS release. If the Finance Board were to adopt a

changed MIRS sampling methodology that no longer separately sampled

lenders by lender type, then it probably would stop the publication of

Table III in the monthly MIRS release.

Section 402(e)(4) of the Financial Institutions Reform, Recovery

and Enforcement Act of 1989 ``FIRREA,'' Public Law No. 101-73, 103

Stat. 183 (August 9, 1989), requires the Chairperson of the Finance

Board to designate a ``substantially similar'' successor index if the

Finance Board no longer makes available any index from MIRS. If the

Finance Board were to stop Table III, then it proposes to designate

that the National Average Contract Mortgage Rate for the Purchase of

All Homes by Combined Lenders be the successor index for any ARM index

that uses a contract rate from Table III. It also proposes to designate

the National Average Effective Mortgage Rate for the Purchase of All

Homes by Combined Lenders be the successor index for any ARM index that

uses an effective rate from Table III. The Finance Board publishes both

of the proposed successor index rates in the top panel of Table I in

the monthly MIRS release, and the current value of both interest rates

is available on a recording maintained by the Finance Board.

The Finance Board is proposing these successor index rates because

the loans reported in Table III by lender type include loans on both

newly built and previously occupied homes. The proposed successor index

rates also include loans on both newly built and previously occupied

homes. The only difference is that the data in Table I combines loans

from all types of lenders whereas Table III reports mortgage data by

type of lender.

The Finance Board seeks comments on these proposed successor index

rates.

G. Effective Date and Transition Provisions

The Finance Board would adopt any changes to the MIRS sampling and

weighting methodology effective at the beginning of 1998. Before

implementing any changes, the Finance Board would consult with the

technical staff of other Federal agencies and instrumentalities to

obtain their views and suggestions about the MIRS sampling and

weighting methodology.

The Finance Board also would make available special tabulations so

that Fannie Mae and Freddie Mac would have data calculated on the same

basis for their determination of the conforming loan limit for 1999.

This calculation would occur in November 1998.

By the Federal Housing Finance Board.

[[Page 9770]]

Dated: February 26, 1997.

Rita I. Fair,

Managing Director.

[FR Doc. 97-5266 Filed 3-3-97; 8:45 am]

BILLING CODE 6725-01-P