E8-16501. Home Mortgage Disclosure

AGENCY:

Board of Governors of the Federal Reserve System.

ACTION:

Proposed rule; proposed staff interpretation.

SUMMARY:

The Board is proposing to amend Regulation C (Home Mortgage Disclosure) to revise the rules for reporting price information on higher-priced loans. The rules would be conformed to the definition of “higher-priced mortgage loan” adopted by the Board under Regulation Z (Truth in Lending) contemporaneously with this proposal. Regulation C currently requires lenders to report the spread between the annual percentage rate (APR) on a loan and the yield on Treasury securities of comparable maturity if the spread meets or exceeds 3.0 percentage points for a first-lien loan (or 5.0 percentage points for a subordinate-lien loan). Under the proposal, a lender would report the spread between the loan's APR and a survey-based estimate of rates currently offered on prime mortgage loans of a comparable type if the spread meets or exceeds 1.5 percentage points for a first-lien loan (or 3.5 percentage points for a subordinate-lien loan).

DATES:

Comments must be received by August 29, 2008.

ADDRESSES:

You may submit comments, identified by Docket No. R-1321, by any of the following methods:

• Agency Web Site: http://www.federalreserve.gov. Follow the instructions for submitting comments at http://www.federalreserve.gov/​generalinfo/​foia/​ProposedRegs.cfm.
• Federal eRulemaking Portal: http://www.regulations.gov. Follow the instructions for submitting comments.
• E-mail: [e-mail]. Include the docket number in the subject line of the message.
• Fax: [phone] or [phone].
• Mail: Jennifer J. Johnson, Secretary, Board of Governors of the Federal Reserve System, 20th Street and Constitution Avenue, NW., Washington, DC 20551.

All public comments are available from the Board's Web site at: http://www.federalreserve.gov/​generalinfo/​foia/​ProposedRegs.cfm as submitted, unless modified for technical reasons. Accordingly, your comments will not be edited to remove any identifying or contact information. Public comments may also be viewed electronically or in paper in Room MP-500 of the Board's Martin Building (20th and C Streets, NW.) between 9 a.m. and 5 p.m. on weekdays.

FOR FURTHER INFORMATION CONTACT:

John C. Wood, Counsel, or Paul Mondor, Senior Attorney, Division of Consumer and Community Affairs, Board of Governors of the Federal Reserve System, Washington, DC 20551, at [phone] or [phone]. For users of Telecommunications Device for the Deaf (TDD) only, contact [phone].

SUPPLEMENTARY INFORMATION:

I. Background on HMDA and Regulation C

The Home Mortgage Disclosure Act (HMDA) requires depository and certain for-profit, nondepository institutions to collect, report to regulators, and disclose to the public data about originations and purchases of home mortgage loans (home purchase and refinancing) and home improvement loans, as well as loan applications that do not result in originations (for example, applications that are denied or withdrawn).

HMDA data can be used to help determine whether institutions are serving the housing needs of their communities. The data help public officials target public investment to attract private investment where it is needed. HMDA data also assist in identifying possible discriminatory lending patterns and in enforcing antidiscrimination statutes.

The Board's Regulation C implements HMDA. The data reported under Regulation C include, among other items, application date; loan type, purpose, and amount; the property location and type; the race, ethnicity, sex, and annual income of the loan applicant; the action taken on the loan application (approved, denied, withdrawn, etc.), and the date of that action; whether a loan is covered by the Home Ownership and Equity Protection Act (HOEPA); lien status (first lien, subordinate lien, or unsecured); and loan pricing (rate spread).^[1]

HMDA and Regulation C were adopted in 1975, and have been amended numerous times over the years. The loan price reporting requirement was added in the most recent amendments and took effect beginning with the collection of data for calendar year 2004. (67 FR 7222, February 15, 2002; 67 FR 30771, May 8, 2002; and 67 FR 43218, June 27, 2002.) Institutions must report the difference between a loan's APR and the yield on Treasury securities of comparable maturity if that difference is 3.0 percentage points or more for a first-lien loan, or 5.0 percentage points or more for a subordinate-lien loan. If the rate spread for a loan is less than the 3.0 or 5.0 percentage point threshold, it is not reported. The Treasury yield used is as of the 15th day of a month most closely preceding the date the loan's interest rate was set by the institution for the final time before closing (rate lock date). The Board provides Treasury yields for various maturities, via the Federal Financial Institutions Examination Council (FFIEC) Web site, to assist institutions in calculating the rate spread.

II. Summary of Proposal

The Board is proposing a method for determining when price information is reported that is similar in concept to Regulation C's current method but different in the particulars. The proposed rule, like the current rule, would set a threshold above a market rate to trigger reporting. But the market rate the Board is proposing is different, and therefore so is the threshold. Instead of yields on Treasury securities of comparable maturity, the proposed rule would use a survey-based estimate of market rates for the lowest-risk prime mortgages, referred to as the “average prime offer rate,” for comparable types of transactions.

The survey the Board would rely on for the foreseeable future is the Primary Mortgage Market Survey® (PMMS) conducted by Freddie Mac. The Board would conduct its own survey if it became appropriate or necessary to do so. The reporting threshold would be set at 1.5 percentage points above the average prime offer rate for first-lien loans, and 3.5 points for subordinate-lien loans. The lender would report the difference between the transaction's APR and the average prime offer rate on a comparable type of transaction if the difference met or exceeded the threshold.

The proposed amendments are intended to facilitate regulatory compliance by conforming the test for rate spread reporting under Regulation C to the definition of higher-priced mortgage loans under Regulation Z. The proposed amendments will also provide better and more useful pricing data on higher-priced loans reported under Regulation C.

III. Reasons for Improving HMDA Rate Spread Reporting

Since the Board adopted Regulation C's reporting benchmark of yields on Treasury securities of comparable maturity, HMDA reporters and others have on various occasions identified shortcomings of this benchmark. Commenters to the January 2008 proposal under Regulation Z (73 FR 1672, January 9, 2008), under which the Board proposed to use Treasury yields as the benchmark to identify higher-priced loans warranting stricter regulations, again identified these shortcomings. Many of these commenters urged the Board to use a benchmark that more closely tracks mortgage rates. They also urged the Board to use the same test for these two purposes under Regulations C and Z, respectively. The Board considered these comments, conducted its own analysis, and concluded that both regulations should rely on a benchmark index that more closely tracks mortgage rates. Accordingly, this proposal would implement essentially the same rule the Board is adopting under Regulation Z.

A. Drawbacks of Using Treasury Security Yields

There are significant advantages to using Treasury yields to set the threshold for reporting price information. Treasuries are traded in a highly liquid market; Treasury yield data are published for many different maturities and can easily be calculated for other maturities; and the integrity of published yields is not subject to question. For these reasons, Treasuries are also commonly used in federal statutes, such as HOEPA, for benchmarking purposes.

As recent events have highlighted, however, using Treasury yields to set the APR threshold for HMDA rate spread reporting has two major disadvantages. The most significant disadvantage is that the spread between Treasuries and mortgage rates changes in the short term and in the long term. Moreover, the comparable Treasury security for a given mortgage loan is quite difficult to determine accurately.

The Treasury-mortgage spread can change for at least three different reasons. First, credit risk may change on mortgages, even for the highest-quality borrowers. For example, credit risk increases when house prices fall. Second, competition for prime borrowers can increase, tightening spreads, or decrease, allowing lenders to charge wider spreads. Third, movements in financial markets can affect Treasury yields but have no effect on lenders' cost of funds or, therefore, on mortgage rates. For example, Treasury yields fall disproportionately more than mortgage rates during a “flight to quality.”

Recent events illustrate how much the Treasury-mortgage spread can swing. The spread averaged about 170 basis points in 2007 but increased to an average of about 220 basis points in the first half of 2008. In addition, the spread was highly volatile in this period, swinging as much as 25 basis points in a week. Thus, the spread may vary significantly from time to time, and long-term predictions of future spreads are highly uncertain.

Changes in the Treasury-mortgage spread can undermine key objectives of the regulation. These changes mean that rate spreads for loans with identical credit risk are reported in some periods but not in others, contrary to the objective of consistent and predictable coverage over time. Moreover, lenders' uncertainty as to when such changes will occur can cause them to set an internal threshold below the regulatory threshold. This may reduce credit availability directly (if a lender's policy is not to make higher-priced loans, to avoid having to report loan pricing for them) or indirectly, by increasing regulatory burden. The recent volatility might lead lenders to set relatively conservative cushions.

Adverse consequences of volatility in the spread between mortgages rates and Treasuries could be reduced simply by setting the regulatory threshold at a high enough level to ensure exclusion of all prime loans. But a threshold high enough to accomplish this objective would likely fail to meet another, equally important objective of covering essentially all of the subprime market. Instead, the Board is proposing to use a benchmark index that more closely follows mortgage market rates, which would make any changes in the spread between mortgage rates and Treasuries largely academic.

The second major disadvantage of using Treasury yields to set the threshold is that the comparable Treasury security for a given mortgage loan is quite difficult to determine accurately. Regulation C determines the comparable Treasury security on the basis of maturity: a loan is matched to a Treasury with the same contract term to maturity. For example, the regulation matches a 30-year mortgage loan to a 30-year Treasury security. This method does not, however, account for the fact that very few loans reach their full maturity, and it causes significant distortions when the yield curve changes shape.^[2] These distortions can bias coverage, sometimes in unpredictable ways, and consequently might influence the preferences of lenders to offer certain loan products in certain environments.

B. Reasons for Following the Regulation Z Final Rule

As noted above, the Board's objective in setting the rate spread reporting threshold has been to cover subprime mortgages and avoid covering prime mortgages. The same purpose underlies the definition of “higher-priced mortgage loan” the Board has just adopted under Regulation Z. For the reasons discussed in the Regulation Z final rule, the Board believes the definition under Regulation Z, if applied to Regulation C, would better achieve this purpose and ensure more consistent and more useful data. Moreover, using the same definition in both Regulation Z and Regulation C will relieve compliance burdens.

IV. The Board's Proposal

A. Rates From the Prime Mortgage Market

To address the principal drawbacks of Treasury security yields, discussed above, the Board is proposing a rule that relies instead on a rate that more closely tracks rates in the prime mortgage market. Proposed § 203.4(a)(12)(ii) would define an “average prime offer rate” as an annual percentage rate derived from average interest rates, points, and other pricing terms offered by a representative sample of creditors for mortgage transactions that have low-risk pricing characteristics. Comparing a transaction's annual percentage rate to this average offered annual percentage rate, rather than to an average offered contract interest rate, should make reporting more accurate and consistent. If a loan's APR exceeds the average prime offer rate for a comparable transaction by 1.5 or more percentage points for a first-lien loan, or 3.5 or more percentage points for a subordinate-lien loan, the creditor would report the difference. (The basis for selecting these thresholds is explained further in part IV.B. below.) The lender would use the most recently available average prime offer rate as of the date on which the lender sets the rate for the final time before consummation.

To facilitate compliance, the proposed rule and commentary would provide that the Board will derive average prime offer rates from survey data according to a methodology it will make publicly available, and publish these rates in a table on the Internet on at least a weekly basis. This table would indicate how to identify a comparable transaction.

As noted above, the survey the Board intends to use for the foreseeable future is Freddie Mac's PMMS, which contains weekly average rates and points offered by a representative sample of creditors to prime borrowers seeking a first-lien, conventional, conforming mortgage and who would have at least 20 percent equity. The PMMS contains pricing data for four types of transactions: “1-year ARM,” “5/1-year ARM,” “30-year fixed,” and “15-year fixed.” For the two types of ARMs, PMMS pricing data are based on ARMs that adjust according to the yield on one-year Treasury securities; the pricing data include the margin and the initial rate (if it differs from the sum of the index and margin). These data are updated every week and are published on Freddie Mac's Web site (see http://www.freddiemac.com/​dlink/​html/​PMMS/​display/​PMMSOutputYr.jsp).

The Freddie Mac PMMS is the most viable option for obtaining average prime offer rates. This is the only publicly available data source that has rates for more than one kind of fixed-rate mortgage (the 15-year and the 30-year) and more than one kind of variable-rate mortgage (the 1-year ARM and the 5/1-year ARM). Having rates on at least two fixed-rate products and at least two variable-rate products supplies a firmer basis for estimating rates for other fixed-rate and variable-rate products (such as a 20-year fixed or a 3/1 ARM).

Other publicly available surveys the Board considered are less suitable for the purposes of this proposal. Only one ARM rate is collected by the Mortgage Bankers Association's Weekly Mortgage Applications Survey and the Federal Housing Finance Board's Monthly Survey of Interest Rates and Terms on Conventional Single-Family Non-Farm Mortgage Loans. Moreover, the FHFB Survey has a substantial lag because it is monthly and reports rates on closed loans. The Board also evaluated two non-survey options involving Fannie Mae and Freddie Mac. One is the Required Net Yield, the prices these institutions will pay to purchase loans directly. The other is the yield on mortgage-backed securities issued by Fannie Mae and Freddie Mac. With either option, data for ARM yields would be difficult to obtain.

These other data sources, however, provide useful benchmarks to evaluate the accuracy of the PMMS. The PMMS has closely tracked these other indices, according to a Board staff analysis. The Board would continue to use them periodically to help it determine whether the PMMS remains an appropriate source of data for average prime offer rates. If the PMMS ceased to be available, or if circumstances arose that rendered it unsuitable for this rule, the Board would consider other alternatives including conducting its own survey.

The Board would use the pricing terms from the PMMS, such as interest rate and points, to calculate an annual percentage rate (consistent with Regulation Z, 12 CFR 226.22) for each of the four types of transactions that the PMMS reports. These annual percentage rates would be the average prime offer rates for transactions of those types. The Board would derive annual percentage rates for other types of transactions from the loan pricing terms available in the survey. The method of derivation the Board would use is being published as part of this proposal (see Attachment I to this Federal Register notice). When finalized, the method would be published on the Internet along with the table of annual percentage rates.

B. Threshold for Rate Spread Reporting

The Board is proposing a threshold of 1.5 percentage points above the average prime offer rate for a comparable transaction for first-lien loans and 3.5 percentage points for second-lien loans. These thresholds are the same as adopted under Regulation Z's definition of “higher-priced mortgage loan.”

As discussed above, the rate spread reporting requirement was intended to cover the subprime market and generally exclude the prime market; and in the face of uncertainty it is appropriate to err on the side of covering somewhat more than the subprime market. Based on available data, it appears that the existing thresholds capture all of the subprime market and a portion of the alt-A market.^[3] Based also on available data, the Board believes that the thresholds it is proposing would cover all, or virtually all, of the subprime market and a portion of the alt-A market. The Board considered loan-level origination data for the period 2004 to 2007 for subprime and alt-A securitized pools. The proprietary source of these data is FirstAmerican Loan Performance.^[4] The Board also ascertained from a proprietary database of mostly government-backed and prime loans (McDash Analytics) that coverage of the prime market during the first three quarters of 2007 at these thresholds would have been very limited. The Board recognizes that the recent mortgage market disruption began at the end of this period, but it is the latest period the Board has been able to study in this database.

The Board is proposing a threshold for subordinate-lien loans of 3.5 percentage points. This is consistent with the existing rule under Regulation C, which sets the threshold over Treasury yields for these loans two percentage points above the threshold for first-lien loans. See 12 CFR 203.4(a)(12). The Board recognizes that it would be preferable to set a threshold for second-lien loans above a measure of market rates for second-lien loans, but it does not appear that a suitable measure of this kind exists. Although data are very limited, the Board believes it is appropriate to apply the same difference of two percentage points to the thresholds above market mortgage rates. As noted in the Regulation Z final rule, with rare exceptions, commenters explicitly endorsed, or at least did not raise any objection to, this approach in connection with that rulemaking; the Board is proposing to maintain consistency between the two rules.

The Board recognizes that there are limitations to making judgments about the future scope of this proposed rule based on past data. For example, once a final rule takes effect, the risk premiums for alt-A loans compared to the prime loans reported in the PMMS may be higher than the risk premiums for the period 2004-2007. In that case, coverage of alt-A loans would be higher than an estimate for that period would indicate.

Another important example is prime “jumbo” loans, or loans extended to borrowers with low-risk mortgage pricing characteristics, but in amounts that exceed the threshold for loans eligible for purchase by Freddie Mac or Fannie Mae. The PMMS collects pricing data only on loans eligible for purchase by one of these entities (“conforming loans”). Prime jumbo loans have always had somewhat higher rates than prime conforming loans, but the spread has widened significantly and become much more volatile since August 2007. If this spread remains wider and more volatile when this proposal takes effect in final form, the rule would cover a significant share of transactions that would be prime jumbo loans. While covering prime jumbo loans is not the Board's objective, the Board does not believe that it should set the threshold at a higher level to avoid what may be only temporary coverage of these loans relative to the long time horizon for this rule.

Credit risk and liquidity risk can vary by many factors, including geography, property type, and type of loan. This may suggest to some that different thresholds should be applied to different classes of transactions. This approach would make the regulation inordinately complicated and subject it to frequent revision, which would not be in the interest of creditors, investors, or consumers. Although the simpler approach the Board is proposing—just two thresholds, one for first-lien loans and another for subordinate-lien loans—has its disadvantages, the Board believes they would be outweighed by its benefits of simplicity and stability.

C. Timing of Determining the Reporting Threshold

Regulation C currently determines the threshold as of the 15th of the month before the rate is locked. This proposal would determine the threshold for a transaction on a more current basis. The proposal would require a creditor to use the most recent average prime offer rate available as of the rate lock date. As the PMMS is updated weekly, the Board will also update average prime offer rates weekly. The Board anticipates that using a more current benchmark will improve reporting accuracy without increasing regulatory burden.

V. Effective Date

Under the final rule published simultaneously with this proposal, the Regulation Z amendments concerning higher-priced mortgage loans take effect on October 1, 2009. The Board contemplates that any final amendments to Regulation C under this rulemaking would take effect for data collection beginning January 1, 2009. Switching rules for HMDA rate spread reporting in the middle of a calendar year would make the data more difficult to use and interpret. If the Board were to make it effective January 1, 2010, lenders would be required to report HMDA data in 2009 using the old (current) rule based on Treasury security yields while, in October through December of 2009, determining applicability of the Regulation Z higher-priced mortgage loan provisions using the new rule based on average prime offer rates. An effective date of January 1, 2009 would ensure that lenders would not need to maintain two separate systems for determining higher-priced mortgage loans during the final quarter of 2009.

If a loan were consummated on or after January 1, 2009, the lender would be required to determine whether the loan is higher-priced (and, if so, report the rate spread) using the new rule, while if the loan were consummated before January 1, 2009 the lender would continue to use the old (current) rule. The Board recognizes that some loans that close in 2009 will have had their rates locked sometime in 2008 (or earlier). Thus, some loans that close in 2009 (and accordingly would be reported on a lending institution's HMDA report for calendar year 2009) would require a creditor to use pre-2009 average prime offer rates to determine their rate spreads. To address this issue, the Board would publish average prime offer rates on the Internet dating from the beginning of October 2008, which lenders could use for loans that are locked in on or after October 1, 2008 but originated in 2009. Lenders that locked in a rate prior to October 1, 2008 but originated the loan in 2009 (or later) would determine whether and how to report price information for such loans using the old (current) rule. To help data users identify these loans, the Board contemplates adding a notation to each such loan in the publicly available data report for 2009 (based on application date, as the closest available proxy for rate-lock date). The Board expects such loans to comprise a very small percentage (one percent or less) of the 2009 HMDA data, based on staff analysis of past years' data.

VI. Requests for Comment

The Board requests comments on (1) the proposal to change the reporting benchmark from Treasury yields to average prime offer rates; (2) the Board's plan to use the Freddie Mac PMMS to estimate average prime offer rates, including comment on whether there are more appropriate sources of data; (3) the method the Board proposes to use to derive average prime offer rates from the PMMS data, which is being published as Attachment I to this proposal; (4) the proposed 1.5 and 3.5 percentage point thresholds; (5) the proposed timing for rate spread determination (rate-lock date, with weekly updating of the average prime offer rate benchmarks); (6) the proposed effective date of these amendments; and (7) the costs and benefits of the proposal generally.

VII. Paperwork Reduction Act

In accordance with section 3506 of the Paperwork Reduction Act of 1995 (44 U.S.C. Ch. 35; 5 CFR part 1320 appendix A.1), the Board has reviewed the proposed rule under the authority delegated to the Board by Office of Management and Budget (OMB). The Board may not conduct or sponsor, and an organization is not required to respond to, this information collection unless it displays a currently valid OMB number. The OMB control number is 7100-0247.

The information collection requirements that would be revised by this rulemaking appear in 12 CFR part 203. The information collection is mandatory under 12 U.S.C. 2801-2810. It generates data used to help determine whether financial institutions are serving the housing needs of their communities, to help target investment to promote private investment where it is needed, and to provide data to assist in identifying possibly discriminatory lending patterns and in enforcing antidiscrimination statutes.

The respondents are all types of financial institutions that meet the tests for coverage under the regulation. Under the Paperwork Reduction Act (PRA), however, the Board accounts for the burden of the paperwork associated with the regulation only for state member banks, their subsidiaries, subsidiaries of bank holding companies, U.S. branches and agencies of foreign banks (other than federal branches, federal agencies, and insured state branches of foreign banks), commercial lending companies owned or controlled by foreign banks, and organizations operating under section 25 or 25A of the Federal Reserve Act (12 U.S.C. 601-604a; 611-631). Other federal agencies account for the paperwork burden for the institutions they supervise. Respondents must maintain their loan/application registers and modified registers for three years, and their disclosure statements for five years.

The Board has determined that the data collection and reporting are required by law; completion of the loan/application register, submission to the Board, and disclosure to the public upon request are mandatory. The data, as modified according to the regulation, are made publicly available and are not considered confidential. Information that might identify an individual borrower or applicant is given confidential treatment under exemption 6 of the Freedom of Information Act (5 U.S.C. 552(b)(6)).

The current total annual burden to comply with the provisions of Regulation C is estimated to be 156,910 hours for 680 Board-regulated institutions that are deemed to be respondents for the purposes of the PRA. The reporting, recordkeeping, and disclosure burden for this information collection is estimated to vary from 12 to 12,000 hours per respondent per year, with an average of 242 hours for state member banks and an average of 192 hours for mortgage banking subsidiaries and other respondents. This estimated burden includes time to: Gather and maintain the data needed, review the instructions, and complete the register. The Board estimates that respondents regulated by the Board would take, on average, 16 hours (two business days) to revise and update their systems to comply with the proposed threshold for rate spread reporting. This one-time revision would increase the burden by 10,880 hours to 167,790.

Comments are invited on: (1) Whether the proposed collection of information is necessary for the proper performance of the Board's functions; including whether the information has practical utility; (2) the accuracy of the Board's estimate of the burden of the proposed information collection, including the cost of compliance; (3) ways to enhance the quality, utility, and clarity of the information to be collected; and (4) ways to minimize the burden of information collection on respondents, including through the use of automated collection techniques or other forms of information technology. Comments on the collection of information should be sent to Michelle Shore, Federal Reserve Board Clearance Officer, Division of Research and Statistics, Mail Stop 151-A, Board of Governors of the Federal Reserve System, Washington, DC 20551, with copies of such comments sent to the Office of Management and Budget, Paperwork Reduction Project (7100-0247), Washington, DC 20503.

VIII. Initial Regulatory Flexibility Analysis

The Regulatory Flexibility Act (5 U.S.C. 601 et seq.) (RFA) generally requires an agency to perform an assessment of the impact a rule is expected to have on small entities. However, under section 605(b) of the RFA, 5 U.S.C. 605(b), the regulatory flexibility analysis otherwise required under section 604 of the RFA is not required if an agency certifies, along with a statement providing the factual basis for such certification, that the rule will not have a significant economic impact on a substantial number of small entities. Based on its analysis and for the reasons stated below, the Board believes that this proposed rule will not have a significant economic impact on a substantial number of small entities. A final regulatory flexibility analysis will be conducted after consideration of comments received during the public comment period.

A. Statement of the Objectives of and Legal Basis for the Proposal

The Board is proposing amendments to Regulation C to make the rules for reporting higher-priced loans in the annual Home Mortgage Disclosure Act (HMDA) data consistent with the definition of higher-priced loan in the amendments to Regulation Z (Truth in Lending) that the Board is adopting in final form. The amendments are intended to reduce regulatory burden by allowing mortgage lenders to use a single definition of higher-priced loan, rather than different definitions under the two regulations. The amendments are also intended to result in more useful HMDA data because the new definition of higher-priced loan uses a survey-based estimate of market mortgage rates as the benchmark for reporting.

The purpose of HMDA is to provide to public officials, and to the public, information to enable them to determine whether lending institutions are fulfilling their obligations to serve the housing needs of their communities. The purpose of the law is also to assist public officials in determining the distribution of public sector investments in a manner designed to improve the private investment environment. HMDA data also assist in identifying possibly discriminatory lending patterns and in enforcing antidiscrimination statutes. 12 U.S.C. 2801(b). HMDA authorizes the Board to prescribe regulations to carry out the purposes of the statute. 12 U.S.C. 2804(a).

The act expressly states that the Board's regulations may contain “such classifications, differentiations, or other provisions * * * as in the judgment of the Board are necessary and proper to effectuate the purposes of [HMDA], and prevent circumvention or evasion thereof, or to facilitate compliance therewith.” 12 U.S.C. 2804(a). The Board believes that the amendments to Regulation C discussed above are within Congress's broad grant of authority to the Board to adopt provisions that carry out the purposes of the statute.

B. Small Entities Affected by the Proposal

The proposed rule would apply to all institutions that are required to report under HMDA. The Board does not have complete data on the asset sizes of all HMDA reporting institutions. Through data from Reports of Condition and Income (“call reports”) of depository institutions and certain subsidiaries of banks and bank holding companies, however, the Board can determine numbers of small entities among those categories. For the majority of HMDA respondents that are non-depository institutions exact asset size information is not available. The Board has somewhat reliable estimates based in large measure on self-reporting from approximately five percent of the non-depository respondents. Based on the best information available for each category of respondent, the Board makes the following estimate of small entities that would be affected by this proposal: Of all HMDA respondents in 2008 (for 2007 activities), which number approximately 8,625, approximately 4,520 had total domestic assets of $165 million or less and thus would be considered small entities for purposes of the Regulatory Flexibility Act.

C. Other Federal Rules

The Board believes no federal rules duplicate, overlap, or conflict with the proposed revisions to Regulation C. However, the Board solicits comment on this matter.

D. Significant Alternatives to the Proposed Revisions

The Board solicits comment on any significant alternatives that may provide additional ways to reduce regulatory burden associated with this proposed rule.

IX. Solicitation of Comments Regarding the Use of “Plain Language”

Section 722 of the Gramm-Leach-Bliley Act of 1999 requires the Board to use “plain language” in all proposed and final rules published after January 1, 2000. The Board invites comments on whether the proposed rules are clearly stated and effectively organized, and how the Board might make the proposed text easier to understand.

List of Subjects in 12 CFR Part 203

• Banks
• Banking
• Federal Reserve System
• Mortgages
• Reporting and recordkeeping requirements

Text of Proposed Revisions

Certain conventions have been used to highlight the proposed revisions to the text of Regulation C, Appendix A, and the Official Staff Commentary. New language is shown inside bold arrows, while language that would be deleted is set off in brackets.

Authority and Issuance

For the reasons set forth in the preamble, the Board proposes to amend 12 CFR part 203 as follows:

PART 203—HOME MORTGAGE DISCLOSURE (REGULATION C)

1. The authority citation for part 203 continues to read as follows:

Authority: 12 U.S.C. 2801-2810.

2. Section 203.4 is amended by revising paragraph (a)(12) to read as follows:

§ 203.4

Compilation of loan data.

(a) * * *

(12) ▸(i)◂ For originated loans subject to Regulation Z, 12 CFR part 226, the difference between the loan's annual percentage rate (APR) and the [yield on Treasury securities having comparable periods of maturity] ▸average prime offer rate for a comparable transaction as of the date the interest rate is set◂, if that difference is equal to or greater than [3] ▸1.5◂ percentage points for loans secured by a first lien on a dwelling, or equal to or greater than [5] ▸3.5◂ percentage points for loans secured by a subordinate lien on a dwelling. [The lender shall use the yield on Treasury securities as of the 15th day of the preceding month if the rate is set between the 1st and the 14th day of the month and as of the 15th day of the current month if the rate is set on or after the 15th day, as prescribed in appendix A to this part.]

▸(ii) “Average prime offer rate” means an annual percentage rate that is derived from average interest rates, points, and other loan pricing terms currently offered to consumers by a representative sample of creditors for mortgage loans that have low-risk pricing characteristics. The Board publishes average prime offer rates for a broad range of types of mortgage in a table updated at least weekly as well as the methodology the Board uses to derive these rates.◂

* * * * *

3. In appendix A to part 203, under I. Instructions for Completion of Loan/Application Register, paragraphs I.G.1. and I.G.2. are revised to read as follows:

Appendix A to Part 203—Form and Instructions for Completion of HMDA Loan/Application Register

* * * * *

I. Instructions for Completion of Loan/Application Register

* * * * *

G. Pricing-Related Data

1. Rate Spread

a. For a home-purchase loan, a refinancing, or a dwelling-secured home improvement loan that you originated, report the spread between the annual percentage rate (APR) and the ▸average prime offer rate for a comparable transaction◂ [applicable Treasury yield] if the spread is equal to or greater than ▸1.5◂ [3] percentage points for first-lien loans or ▸3.5◂ [5] percentage points for subordinate-lien loans. To determine whether the rate spread meets this threshold, use the ▸average prime offer rate for the type of transaction, pursuant to § 203.4(a)(12) and staff commentary thereunder, as of the date◂ [Treasury yield for securities of a comparable period of maturity as of the 15th day of a given month, depending on when] the interest rate was set, and use the APR for the loan, as calculated and disclosed to the consumer under § 226.6 or 226.18 of Regulation Z (12 CFR part 226). Use the ▸most recently available average prime offer rate.◂ [15th day of a given month for any loan on which the interest rate was set on or after that 15th day through the 14th day of the next month. (For example, if the rate is set on September 17, 2004, use the Treasury yield as of September 15, 2004; if the interest rate is set on September 3, 2004, use the Treasury yield as of August 15, 2004). To determine the applicable Treasury-security yield, the financial institution must use] ▸Current and historic average prime offer rates are set forth in◂ the table published on the FFIEC's Web site (http://www.ffiec.gov/​hmda) entitled ▸“Average Prime Offer Rates.”◂ [“Treasury Securities of Comparable Maturity under Regulation C.”]

* * * * *

d. Enter the rate spread to two decimal places, and use a leading zero. For example, enter 03.29. If the difference between the APR and the ▸average prime offer rate◂ [Treasury yield] is a figure with more than two decimal places, round the figure or truncate the digits beyond two decimal places.

e. If the difference between the APR and the ▸average prime offer rate◂ [Treasury yield] is less than ▸1.5◂ [3] percentage points for a first-lien loan and less than ▸3.5◂ [5] percentage points for a subordinate-lien loan, enter “NA.”

2. Date the interest rate was set. The relevant date to use to determine the ▸average prime offer rate for a comparable transaction◂ [Treasury yield] is the date on which the loan's interest rate was set by the financial institution for the final time before closing. If an interest rate is set pursuant to a “lock-in” agreement between the lender and the borrower, then the date on which the agreement fixes the interest rate is the date the rate was set. If a rate is re-set after a lock-in agreement is executed (for example, because the borrower exercises a float-down option or the agreement expires), then the relevant date is the date the rate is re-set for the final time before closing. If no lock-in agreement is executed, then the relevant date is the date on which the institution sets the rate for the final time before closing.

* * * * *

4. In Supplement I to Part 203, under Section 203.4—Compilation of Loan Data, 4(a) Data Format and Itemization, Paragraph 4(a)(12) Rate spread information, paragraph 4(a)(12)-1 is removed, new heading Paragraph 4(a)(12)(ii) is added, and new paragraphs 4(a)(12)(ii)-1, -2, and -3 are added, to read as follows:

Supplement I to Part 203—Staff Commentary

* * * * *

Section 203.4—Compilation of Loan Data

4(a) Data Format and Itemization

* * * * *

Paragraph 4(a)(12) Rate spread information.

[1] Treasury securities of comparable maturity. To determine the yield on a Treasury security, lenders must use the table entitled “Treasury Securities of Comparable Maturity under Regulation C,” which will be published on the FFIEC's Web site (http://www.ffiec.gov/​hmda) and made available in paper form upon request. This table will provide, for the 15th day of each month, Treasury security yields for every available loan maturity. The applicable Treasury yield date will depend on the date on which the financial institution set the interest rate on the loan for the final time before closing. See appendix A, Paragraphs I.G.1. and 2.]

▸Paragraph 4(a)(12)(ii)

1. Average prime offer rate. Average prime offer rates are annual percentage rates derived from average interest rates, points, and other loan pricing terms offered to borrowers by a representative sample of lenders for mortgage loans that have low-risk pricing characteristics. Other pricing terms include commonly used indices, margins, and initial fixed-rate periods for variable-rate transactions. Relevant pricing characteristics include a consumer's credit history and transaction characteristics such as the loan-to-value ratio, owner-occupant status, and purpose of the transaction. To obtain average prime offer rates, the Board uses a survey of lenders that both meets the criteria of § 203.4(a)(12)(ii) and provides pricing terms for at least two types of variable-rate transactions and at least two types of non-variable-rate transactions. An example of such a survey is the Freddie Mac Primary Mortgage Market Survey®.

2. Comparable transaction. The rate spread reporting requirement applies to a consumer credit transaction that is secured by the consumer's principal dwelling with an annual percentage rate that exceeds by the specified margin the average prime offer rate for a comparable transaction as of the date the interest rate is set. The table of market mortgage rates published by the Board indicates how to identify the comparable transaction.

3. Board table. The Board publishes on the Internet, in table form, average prime offer rates for a wide variety of transaction types. The Board calculates an annual percentage rate, consistent with Regulation Z (see 12 CFR 226.22 and part 226, appendix J), for each transaction type for which pricing terms are available from a survey. The Board estimates annual percentage rates for other types of transactions for which direct survey data are not available based on the loan pricing terms available in the survey and other information. The Board publishes on the Internet the methodology it uses to arrive at these estimates.◂

* * * * *

By order of the Board of Governors of the Federal Reserve System, July 15, 2008.

Jennifer J. Johnson,

Secretary of the Board.

Attachment I—Methodology for Determining Average Prime Offer Rate

The calculation of the Average Prime Offer Rate (APOR) is based on the Freddie Mac Primary Mortgage Market Survey® (PMMS). The survey collects data for a hypothetical “best quality” 80% LTV 1st lien for four mortgage products: (1) 30-year fixed-rate; (2) 15-year fixed-rate; (3) one-year variable-rate; and (4) five-year variable-rate. Each of the variable-rate products is assumed to adjust to an index based on the 1-year Treasury rate plus a margin and to adjust annually after the initial fixed-rate period.

The PMMS collects nationwide average offer prices during the Monday through Wednesday period each week and releases the averages on Thursday. For each loan type the average commitment loan rate and fees and points are reported, each expressed as percentages of the initial loan balance. For the fixed-rate products the commitment rate is the contract rate on the loan; for the variable-rate products it is the initial loan rate. For the variable-rate products, the average index margin is also reported (also expressed in percentage points).

The information provided by the PMMS survey is sufficient to compute an annual percentage rate (APR) for the 30- and 15-year fixed-rate products. However, additional information is needed for the two variable-rate products. Specifically, an estimate of the fully indexed rate (the sum of the index and margin, without regard for any temporary discount or premium) is needed. For the two variable-rate products, the fully indexed rate is calculated as the margin (collected in the survey) plus the future one-year Treasury rate, which is estimated by the current one-year Treasury rate.

The Board uses the rates prevailing during the three-day period in which the PMMS is conducted. Specifically, the average of the close-of-business one-year Treasury rates for Monday, Tuesday, and Wednesday of the survey week is used as the estimate of the “current” Treasury rate used for the fully-indexed component of the variable-rate APR calculations. (If data are available for fewer than three days, then only yields for the available days are used for the average.)

Survey data on the initial interest rate, fees and points, and the calculated fully indexed rate, are sufficient to compute an APR for the one- and five-year variable-rate mortgage products in the PMMS. In computing the APR a fully amortizing loan is assumed, with monthly compounding (similar assumptions are made for the fixed-rate products) and with a two-percentage-point cap in the annual interest rate adjustment.

The PMMS data provide information for only a subset of mortgage products. Specifically, the survey does not cover fixed-rate loans with terms of less than 15 years nor does it cover variable-rate rate mortgages with adjustment periods of other than one or five years. The Board uses interpolation techniques to estimate APRs for an additional range of products. The interpolation techniques rely on the relative yields of different Treasury products.

Currently, yields are tracked for Treasury securities with terms of: one, two, three, five, seven, and ten years. The Board uses these data to estimate APRs for two-, three-, seven-, and ten-year variable-rate rate mortgages which are identical to the one- and five-year variable-rate products surveyed in PMMS in all respects except the length of the initial interest rate period. The specific estimation technique is as follows.

The margin and fees and points for each interpolated variable-rate product are estimated as weighted averages of the margins and fees and points of the one-year and five-year variable-rate products reported in the PMMS. For the two-year variable-rate loan the weights are 3/4 for the one-year variable-rate and 1/4 for the five-year. For the three-year variable-rate product, the weights are 1/2 for both. For the seven- and ten-year variable-rate products, only the margin and fees and points of the five-year variable-rate are used.

The initial interest rate for each of the interpolated variable-rate products is estimated by a two-step process. First, a Treasury spread is computed as the weighted average of the spread between the initial interest on the one-year and five-year PMMS variable-rate products and the one- and five-year Treasury yields respectively. The weights used are the same as those used in the margin and fees and points calculations. The Treasury rates are taken from the Monday-Wednesday close-of-business averages cited above.

The second step is to add the Treasury spreads calculated from the PMMS data to the Treasury yield for the appropriate term. Thus, for example, for the two-year variable-rate product, the estimated spread is added to the two-year Treasury rate, while the ten-year Treasury rate is used for the ten-year variable-rate estimate.

Thus estimated, the initial rates, margins, points and fees are used to calculate a fully indexed rate and ultimately an APR for the two-, three-, seven- and ten-year variable-rate products.

To calculate APRs for fixed-rate loans with terms of ten years or less, the Board uses the initial interest rates (and fees and points) of the one-, two-, three-, five-, seven-, and ten-year variable-rate loan products calculated above to estimate APRs for fixed-rate loans with a term of one, two, three, five, seven, and ten years respectively.

Altogether the Board estimates APRs for ten additional products (two-, three-, seven-, and ten-year 30-year term variable-rates and one-, two-, three-, five-, seven-, and ten-year fixed-rate term loans) to use along with the four products directly surveyed in the PMMS. If survey data become available for any of the ten interpolated products, survey-based inputs will be used instead of the estimates. These 14 products cover most mortgages in current use. Assignment rules allow coverage of all other products.

For example, a four-year variable-rate loan will be matched to the five-year variable-rate product threshold APR; a six-year to the seven-year and any variable-rate loan with a repricing interval of more than seven years will be matched to the ten-year variable-rate product threshold APR. Similar assignments will be used for fixed-rate loans, with any fixed-rate loan with a term of more than 15 years matched to the 30-year fixed-rate product threshold APR and loans with terms between ten and 15 years matched to the 15-year fixed-rate loan threshold APR.

All of the information needed for the above calculations is publicly available on Thursday morning of each week. APRs for each of the 14 products are posted on the FFIEC Web site by Thursday night. All loans locking from Friday through the following Thursday use these APRs as the basis of their spread calculations.

Example:

The week of May 15, 2008 is used to illustrate the threshold APR methodology. On Thursday, May 15th, Freddie Mac released the following PMMS information reflecting national mortgage rate averages for the three day period May 12 to May 14 (each variable is expressed in percentage points):

30-year fixed-rate:
Contract rate	6.01
Fees & Points	0.6
15-year fixed-rate:
Contract rate	5.60
Fees & Points	0.5
Five-year variable-rate:
Initial rate	5.57
Fees & Points	0.6
Margin	2.75
One-year variable-rate:
Initial rate	5.18
Fees & Points	0.7
Margin	2.75

The Freddie Mac survey contract rate and points and fees for the 30-year and 15-year fixed-rate mortgages are sufficient to compute an APR for these two products. The APR is calculated assuming full amortization with one-month compounding. The calculated APRs are:

30-year fixed-rate	6.07
15-year fixed-rate	5.68

Additional information on the assumed fully-indexed rate is needed in order to calculate APRs for the one-year and five-year variable-rate products. Average close-of-business Treasury yields for the three days in which the survey was conducted are used for these calculations:

May 12th:
One-year Treasury	2.01
Two-year Treasury	2.30
Three-year Treasury	2.54
Five-year Treasury	3.00
Seven-year Treasury	3.34
Ten-year Treasury	3.78
May 13th:
One-year Treasury	2.08
Two-year Treasury	2.47
Three-year Treasury	2.70
Five-year Treasury	3.17
Seven-year Treasury	3.49
Ten-year Treasury	3.90
May 14th:
One-year Treasury	2.11
Two-year Treasury	2.53
Three-year Treasury	2.78
Five-year Treasury	3.22
Seven-year Treasury	3.50
Ten-year Treasury	3.92
Averaging these figures for the three days implies Treasury yields of:
One-year Treasury	2.07
Two-year Treasury	2.43
Three-year Treasury	2.67
Five-year Treasury	3.13
Seven-year Treasury	3.44
Ten-year Treasury	3.87

The fully-indexed rate (the estimated interest rate after one-year) for the one-year variable-rate mortgage is calculated as the appropriate Treasury yield plus the margin: 2.07 + 2.75 = 4.82. Similarly, the fully-indexed rate (the estimated interest rate after five-years) for the five-year variable-rate mortgage is calculated as: 3.13 + 2.75 = 5.88.

The initial rate, fees and points, and fully-indexed rate are sufficient to compute APRs for the one-year and five-year variable-rate products. Full amortization, monthly compounding, and a two-percentage-point cap in the annual change in rates are assumed. The calculated APRs are:

One-year variable-rate rate	4.91
Five-year variable-rate rate	5.82

Data for the interpolated two-year and three-year variable-rate mortgages are calculated as weighted averages of the figures for the one- and five-year variable-rates which is used in conjunction with the yields on the two- and three-year Treasuries as follows:

Two-year variable-rate:
Initial rate	[3×(5.18−2.07) + 1×(5.57−3.13)]/4 + 2.43 = 5.37
Fees & Points	[3×.7 + 1×.6]/4 = .7
Margin	[3×2.75 + 1×2.75]/4 = 2.75
Fully-indexed rate	2.75 + 2.43 = 5.18
Three-year variable-rate:
Initial rate	[2×(5.18−2.07) + 2×(5.57−3.13)]/4 + 2.67 = 5.45
Fees & Points	[2×.7 + 2×.6]/4 = .7
Margin	[2×2.75 + 2×2.75]/4 = 2.75
Fully-indexed rate	2.75 + 2.67 = 5.42

Full amortization, monthly compounding, and a two-percentage-point cap in the annual change in rates yields calculated APRs of:

Two-year variable-rate rate	5.27
Three-year variable-rate rate	5.49

APRs for seven-year and ten-year variable-rate mortgages are estimated using the survey data for the five-year variable-rate and yields on the seven- and ten-year Treasuries:

Seven-year variable-rate:
Initial rate	(5.57−3.13) + 3.44 = 5.88
Fees & Points	= .6
Margin	= 2.75
Fully-indexed rate	2.75 + 3.44 = 6.19
Ten-year variable-rate:
Initial rate	(5.57 − 3.13) + 3.87 = 6.31
Fees & Points	= .6
Margin	= 2.75
Fully-indexed rate	2.75 + 3.87=6.62

Full amortization, monthly compounding, and a two-percentage-point cap in the annual change in rates yields calculated APRs of:

Seven-year variable-rate rate	6.09
Ten-year variable-rate rate	6.47

The initial rate and fees and points of the variable-rate mortgages calculated above are used to estimate threshold APRs for fixed-rate products with terms of ten years or less. The estimates are as follows:

One-year fixed:
Initial rate	5.18
Fees & Points	.7
APR	5.96
Two-year fixed:
Initial rate	5.37
Fees & Points	.7
APR	6.06
Three-year fixed:
Initial rate	5.45
Fees & Points	.7
APR	5.92
Five-year fixed:
Initial rate	5.57
Fees & Points	.6
APR	5.82
Seven-year fixed:
Initial rate	5.88
Fees & Points	.6
APR	6.06
Ten-year fixed:
Initial rate	6.31
Fees & Points	.6
APR	6.44

Footnotes

1.  Institutions report these data to their supervisory agencies on an application-by-application basis using a register format. Institutions must make their loan/application registers available to the public, with certain fields redacted to preserve applicants' privacy. The Federal Financial Institutions Examination Council (FFIEC), on behalf of the supervisory agencies, compiles the reported data and prepares an individual disclosure statement for each institution, aggregate reports for all covered institutions in each metropolitan area, and other reports. These disclosure statements and reports are also available to the public.

2.  Robert B. Avery, Kenneth P. Brevoort, and Glenn B. Canner (2006), “Higher-Priced Home Lending and the 2005 HMDA Data,” Federal Reserve Bulletin, vol. 92 (September 8), pp. A123-66.

3.  The percentage of the first-lien mortgage market on which Regulation C has required rate spread reporting using a threshold of three percentage points has been greater than the percentage of the total market originations that one industry source has estimated to be subprime (25 percent vs. 20 percent in 2005; 28 percent vs. 20 percent in 2006). For industry estimates see Inside Mortgage Finance Publications, Inc., The 2007 Mortgage Market Statistical Annual vol. 1, at 4. Regulation C's coverage of higher-priced loans is not thought, however, to have reached the prime market in those years. Rather, in both 2005 and 2006 it reached into the alt-A market, which the same source estimated to be 12 percent in 2005 and 13 percent in 2006. In 2004, Regulation C captured a significantly smaller part of the market than an industry estimate of the subprime market (11 percent vs. 19 percent), but that year's HMDA data were somewhat anomalous because of a steep yield curve.

4.  Annual percentage rates were estimated from the contract rates in these data using formulas derived from a separate proprietary database of subprime loans that collects contract rates, points, and annual percentage rates. This separate database, which contains data on the loan originations of eight subprime mortgage lenders, is maintained by the Financial Services Research Program at George Washington University.