New information reported under HMDA and its application in fair lending enforcement.

Most lending institutions with offices in metropolitan statistical areas are required by the Home Mortgage Disclosure Act of 1975 (HMDA) to disclose information to the public about applications for home loans and the home loans that they originate or purchase during each calendar year. The law's requirements arose from concerns that, in some cases, lenders were contributing to the decline of certain neighborhoods by failing to provide adequate home financing to qualified applicants on reasonable terms and conditions. The disclosure of lending activity is intended to help determine whether lenders are adequately serving their communities' housing finance needs, to facilitate enforcement of the nation's fair lending laws, and to guide investment activities in both the public and the private sectors. HMDA is implemented by the Federal Reserve Board's Regulation C.

Underlying HMDA's disclosure requirements is a presumption that more publicly available information will improve market performance and help prevent market failures. The data reported under HMDA are certainly extensive: Taken together, the 8,853 lenders covered by the law as of the end of 2004 are estimated to have accounted for about 80 percent of home loans extended that year.

The Congress has amended HMDA on several occasions to extend the reach of the law to more lenders and to expand the types of information that must be disclosed. Amendments passed in 1989 have been the most sweeping to date. They require that lenders disclose the disposition of each application they process for home loans and the income, race, ethnicity, and sex of the individuals applying for the loans. As this new information became available, it revealed wide differences in rates of approval of loan applications across racial and ethnic lines and thereby heightened concerns about whether lending decisions complied with the nation's fair lending laws. The disclosures triggered a continuing debate about the proper interpretation of the data and the significance of the differences in lending decisions. Many lending institutions have responded to the concerns raised in the debate by adopting new loan-underwriting procedures to help ensure fair treatment of all applicants and by initiating a wide variety of community outreach and affordable lending programs intended to benefit minority borrowers and lower-income individuals and neighborhoods.

In 2002, in its most recent review of Regulation C, the Federal Reserve Board made a number of important changes to the disclosure requirements that substantially increase the types and amount of information made available through HMDA. (1) The revisions are intended to better advance the purposes of the law by keeping the regulation in step with recent developments in home-loan markets and by incorporating the revised standards of classification for the collection of information on race and ethnicity as established by the Office of Management and Budget (OMB).

Most of the recent changes in the information that is required to be reported under HMDA apply to data relating to loans extended in 2004. Individual lenders covered by HMDA were required to make their 2004 data available to the public beginning on March 31, 2005. However, only the September 2005 release of the data will have been comprehensively checked by the supervisory agencies for the errors and omissions that are detectable from a review of the data.

Perhaps the most important change to Regulation C is the requirement that lenders now disclose pricing (interest rates and fees) for loans with prices above designated thresholds. Loans with prices above the thresholds are referred to here as "higher-priced loans." Other important new information being reported under the revised regulation is whether the loan is a first lien, a junior lien, or unsecured (characteristics referred to here as a loan's lien status), whether it is secured by a manufactured home, and whether it is subject to the protections of the Home Ownership and Equity Protection Act of 1994 (HOEPA). These new pieces of information allow for a better understanding of lending activity in the higher-priced segment of the home-loan market, a segment that was virtually nonexistent a decade or so ago and is now a substantial part of the market. The growth of this market segment, while affording some consumers greater access to credit, has been accompanied by concerns about abusive lending practices, often referred to as "predatory lending." (2) These concerns lend importance to a better understanding of the higher-priced segment of the market and a greater ability to monitor the activities of the individual lenders involved in it.

This article presents a first look at the greatly expanded 2004 HMDA data and considers some of their implications for the continuing concerns about fair lending. (3) The analysis highlights some key relationships revealed in an initial review of the types of data that are new for 2004. Some parts of the analysis focus on nationwide statistics, and others examine patterns across groups of lenders, loan products, and various groupings of applicants, borrowers, and neighborhoods. The authors explore, in particular and in some depth, the strengths and limitations of the information on loan pricing.

We also describe how the Federal Reserve uses the HMDA data as part of a screening tool to facilitate the enforcement of the fair lending laws. In this regard, we discuss the way the expanded HMDA data, particularly the information on loan pricing, enhance the utility of the screening tool. At the same time, we emphasize that, although these data present valuable new opportunities for researchers and others to learn more about the home-loan market and for the regulatory agencies to improve the enforcement of fair lending laws, the data are not sufficient by themselves for drawing conclusions about the fairness of the lending process or the activities of any individual lender. For example, credit history scores and other factors not included in the HMDA data can be critical in determining loan prices. With regard to this issue, we collaborated with researchers at the Credit Research Center of Georgetown University, which has data on credit history scores and other loan-level factors relevant to loan pricing. The loan-level data were supplied to the Credit Research Center by a small group of lenders that are covered by HMDA and are active originators of loans in the higher-priced segment of the home-loan market.

Our examination of the 2004 data also focuses on the newly reported information about loans on manufactured homes. The disposition of applications for loans to buy, refinance, or improve such units has an important influence on the pattern of denial rates of all loans reported under HMDA. We also discuss the new information on HOEPA-related lending and certain requests for pre-approvals of home-purchase loans and assess their overall significance in the market. Finally, in the article's summary and conclusions, we review our key findings and emphasize that users of the data should exercise particular caution in drawing conclusions about lending patterns from HMDA data alone.

THE REQUIREMENTS OF REGULATION C

The Federal Reserve Board's Regulation C, which implements HMDA, applies to most depository institutions (commercial banks, savings institutions, and credit unions--hereafter, "banks") with a home or branch office in a metropolitan area. Banks that are exempt from Regulation C are small (currently those with assets of less than $34 million), or are not in the home-lending business, or have offices exclusively in rural (nonmetropolitan) areas. (4) Regulation C also extends to mortgage and consumer finance companies--hereafter, "mortgage companies"--whether such companies are independent or are subsidiaries of banks or affiliates of bank holding companies. Coverage of mortgage companies applies mainly to those that are active in the home-loan market--that is, those that extend 100 or more home-purchase or home-refinancing loans per year and operate in at least one metropolitan area. (5)

The HMDA data include information about applicants and borrowers, the home-loan products they seek, the disposition of their requests for credit, and details about the location of the property that relate to the application. For information about the channels through which the HMDA data are released and for a description of the data that were required of lenders before the 2002 revisions, see box "Distribution of HMDA Data and Pre-2004 Requirements of Regulation C."

Distribution of HMDA Data and Pre-2004 Requirements
of Regulation C

Under the Home Mortgage Disclosure Act (HMDA), lenders
use a "loan/application register" (HMDA/LAR) to
report information annually to their federal supervisory
agencies for each application and loan acted on during the
calendar year. Lenders must make their HMDA/LARs available
to the public by March 31 following the year to which
the data relate, and they must remove the two date-related
fields to help preserve applicants' privacy. (1)

The Federal Financial Institutions Examination Council
(FFIEC), acting on behalf of the federal supervisory agencies,
compiles the reported information and prepares an
individual disclosure statement for each institution--one for
each metropolitan statistical area (MSA) and metropolitan
division (MD) in which it has offices--as well as aggregate
reports for all covered lenders in each MSA and other
reports. (2) The disclosure and aggregate reports are detailed
tables of data on individual loans and applications.

The disclosure statements and reports are available to the
public at the FFIEC website for HMDA (www.ffiec.gov/
hmda), from the covered lenders themselves, and from
depositories (such as public libraries and other government
offices) in each MSA. (3) In addition, a copy of the HMDA/
LAR for each institution is available to the public on
CD-ROM for a nominal charge. The FFIEC also makes
available a copy of the file of population characteristics of
each census tract covered by the tables on individual institutions
and by the aggregate tables. The 2004 census tract file
is derived from the 2000 decennial census. MSA and MD
identifiers included on that file are based on the designations
of MSAs issued by the Office of Management and
Budget. (4)

Before the most recent revisions, in 2002, the Federal
Reserve Board's Regulation C required lenders to report the
following information on home-purchase and home-improvement
loans and on the refinancing of such loans:

For each application or loan

* application date and the date an action was taken on the
application

* action taken on the application

--approved and originated

--approved but not accepted by the applicant

--denied (with the reasons for denial--voluntary for
some lenders)

--withdrawn by the applicant

--file closed for incompleteness

* loan amount

* income relied on in loan underwriting

* loan type

--conventional

--insured by the Federal Housing Administration

--guaranteed by the Veterans Administration

--backed by the Farm Service Agency or Rural Housing

Service

* loan purpose

--home purchase

--refinance

--home improvement

* type of purchaser (if the lender subsequently sold the
loan)

For each applicant or co-applicant

* race or ethnicity

* sex

For each property

* location, by state, county, and census tract

* type (one- to four-family dwelling or dwelling with five
or more units)

* occupancy status (owner-occupied or nonowner-occupied)

Information is also reported on home loans purchased by an
institution during a calendar year. Under the 2002 revisions
to Regulation C, additional items became required beginning
in 2004.

(1.) Lenders must make their date-modified register available to the
public for a period of three years.

(2.) MSAs that have a single core with a population of 2.5 million or
more and meet certain other requirements contain MDs. Of the 370 MSAs
in the United States, 11 have a population of at least 2.5 million
and have a total of 29 MDs. Starting with the release of the 2004
HMDA data, disclosure reports will follow the guidance provided by
the Office of Management and Budget (OMB) regarding the appropriate
use of statistical-area definitions; see Office of Management and
Budget (2003), OMB Bulletin, no. 03-04 (June),
www.whitehouse.gov/omb/bulletins/b03-04.html. In conformance with OMB
guidance, the FFIEC will prepare disclosure reports for each MSA and
MD.

(3.) The FFIEC maintains the most recent three years of HMDA data. Data
for earlier years can be obtained from the National Technical
Information Center, Springfield, Virginia, www.ntis.gov.

(4.) See Office of Management and Budget, OMB Bulletin.

The New Reporting Requirements

The 2002 revisions to Regulation C are intended to improve the quality, consistency, and utility of the data reported under HMDA; they are also intended to ease regulatory burden, primarily by clarifying and simplifying parts of the regulation. The new requirements

* expand coverage to more nondepository lenders

* streamline the definitions of refinancing and home-improvement loan

* revise the definition of application to include certain requests for pre-approvals (however, in this article, applications are defined as being for a loan on a specific property; they are thus distinct in our analyses from requests for pre-approval, which are unrelated to a specific property)

* mandate for the first time the collection of lien status; property code (to distinguish between one- to four-family dwellings that are site-built and those that are manufactured homes); loan pricing; and HOEPA status

* incorporate changes to the rules on collecting and reporting information on race and ethnicity to conform to guidance issued by the OMB

* require lenders to request the race, ethnicity, and sex of prospective borrowers who apply by mail, Internet, or telephone

* revise the categories that identify the type of institution to which loans are sold

The disclosure of additional data and the revised definitions for some currently reported items serve several purposes. For example, the revised definition of refinancing is intended to reduce inconsistency in the data and to simplify reporting. (6) Some of the new data items--such as lien status and identification of loans for manufactured homes--allow more-precise differentiation among loan products and consequently reduce the possible analytical biases that arise when dissimilar loan products are grouped together. To ensure that nondepository institutions that are active home lenders are subject to the same reporting regime as are other lenders, coverage rules were changed by adding an annual dollar-volume threshold of $25 million of home-loan originations to the current criterion of 10 percent of total loan originations measured in dollars.

Pricing information increases the scope of analysis of HMDA data in support of fair lending enforcement and makes possible an assessment of pricing patterns in the higher-priced segment of the home-loan market. In addition, designation of HOEPA status can be used to identify lenders involved in that type of lending, to measure its incidence in the market, and to aid in fair lending evaluations and HOEPA compliance. Finally, the new information on requests for pre-approvals provides more data on consumers' experiences in the early stages of shopping for a loan to buy a home, a phase of the loan process that has heretofore gone largely unreported, and thus should also facilitate fair lending enforcement.

Transition Rules

To minimize reporting burden and to help ensure the quality and usefulness of the expanded data, the Federal Reserve established transition rules for HMDA compliance that generally did not require lenders to collect some of the new information for requests for pre-approval and applications submitted before January 1, 2004. Among the new information items affected by the transition rules were the data on pricing; the information on whether an application or loan involved a request for pre-approval and on whether the dwelling involved was a manufactured home; and the classifications of race and ethnicity.

Of all applications involving one- to four-family units in the 2004 HMDA data, about 2 million, or 7 percent, were filed before 2004, and thus the data reported on those applications (pertaining to about 1 million loans) might not reflect the new reporting rules. Users of the 2004 data should be aware of this limitation.

To help users of the HMDA data better distinguish loans subject to the transition rules, the Federal Financial Institutions Examination Council (FFIEC) has added a data item to the 2004 CD-ROM that contains a copy of the HMDA/LAR for each institution that indicates whether or not an application was filed before January 1, 2004 (see box "Distribution of HMDA Data and Pre-2004 Requirements of Regulation C"). Users of the 2004 data can make assumptions or restrict their analysis in various ways to address problems created by the transition rules. For example, in preparing the institution and aggregate MSA disclosure reports for 2004, the FFIEC excluded applications filed before January 1, 2004, from all tables reporting pricing (but not other) information.

The transition rules should have little effect on the data in future HMDA filings. However, because some applications have application filing dates that precede a decision on the application by more than a year, a few applications subject to the transition rules may be included when the 2005 HMDA data are reported in 2006.

Lien Status

Information on lien status differentiates home loans secured by a first lien, those secured by a junior (second or third) lien, and those not secured. (The last category arises only among home-improvement loans, for which a security interest in a property may or may not be taken.) Knowledge of lien status is basic to credit underwriting because loans secured by first liens have a lower incidence of default than loans secured by junior liens or unsecured loans; consequently, loans secured by a first lien are generally offered at the lowest rates of interest.

The information on lien status serves a number of public policy interests. First, the information improves the measurement of the overall size of the home-loan market and particular segments within that market, such as home-purchase lending. Although HMDA data have always included information about the purpose of a loan, recent market developments have made that information less useful for measuring lending. Today, many home purchases involve both first- and junior-lien loans. The junior-lien loan in such transactions is often used to avoid requirements to purchase private mortgage insurance (PMI) or to avoid exceeding the loan-size limits used by some secondary-market purchasers, especially Fannie Mae and Freddie Mac (see the appendix for more information about PMI and the availability of data on loans backed by PMI). In the past, a loan backed by a junior lien could not be distinguished directly in the home-purchase loan data from one backed by a first lien and was therefore often assumed to represent a separate home-purchase loan rather than to be one of two used to purchase a single property. (7) The expanded HMDA data allow such distinctions to be made and consequently help avoid the double counting of loans in the home-purchase market.

Second, lien status is essential for interpreting loan-pricing information and in conducting fair lending investigations. Regarding fair lending reviews, the historical lack of information on lien status in the HMDA data has hampered analyses focusing on potential differences in the pattern of the disposition of applications because distinguishing properly among loan products using only HMDA data has been difficult or impossible. (8) Because the use of various loan products and patterns of application disposition can vary across racial and ethnic groups, an inability to distinguish products can lead to spurious correlations and potentially inappropriate conclusions about the fairness of the application of credit-underwriting policies.

Manufactured-Home Status

Available evidence indicates that the credit profiles of individuals seeking loans backed by manufactured homes differ from those of individuals borrowing for site-built homes. (9) On the whole, loans to purchase manufactured homes involve relatively high credit risk, in part because the buyers of such homes tend to have weaker financial profiles than do those purchasing other single-family properties. This evidence has important implications for denial rates and pricing.

Analysis of past HMDA data implied that lenders denied about 60 percent of all applications for conventional home-purchase loans for manufactured homes, whereas they denied only about 12 percent of applications for other conventional home-purchase loans. (10) Until now, the general inability to accurately distinguish manufactured-home loans from loans related to site-built homes complicated the determination of whether differences in denial rates across groups of applicants arose from differences in underwriting practices across the groups or simply from different mixes of loan products sought by the groups. Identification of applications and loans involving manufactured homes in the expanded HMDA data allows for more-refined analysis of the sources of different denial rate patterns and for greater understanding of financing activities in this important market segment.

Loan Pricing

The home-loan market has evolved in a number of important respects over the past decade or so. Traditionally, lenders offered consumers a relatively limited array of products at prices that varied according to the characteristics of the loan and property but not according to the creditworthiness of the borrower. Effectively, borrowers either did or did not meet the underwriting criteria for a particular product, and those who met the criteria paid about the same price. This market characterization may explain why the congressional revisions to HMDA in 1989 focused on the disclosure of data on the disposition of applications rather than on loan prices.

Since then, improvements in information processing and the maturation of a robust secondary market for loans have spurred changes in the home-loan market. Prominent among these changes has been an evolution toward an explicitly risk-based pricing of credit. Now the creditworthiness of individual borrowers can lead to different prices for the same product. Less-creditworthy applicants, or those either unwilling or unable to document their creditworthiness or income, are increasingly less likely to be turned down for a loan; rather, they are offered credit at higher prices.

Borrowers in the higher-priced market generally fall into one of two market segments, "near prime" and "subprime," with individuals in the latter category paying the highest prices because they pose the greatest risk of default. In practice, the dividing line between these two "nonprime" markets is becoming increasingly amorphous, as is the line between the prime (lower-price) and nonprime markets.

Estimates of the annual volume of subprime lending vary, but all sources agree that this market has grown substantially in recent years. (11) One industry source estimates that over the period 1994-2004, the annual dollar volume of subprime home loans increased from about $35 billion to more than $530 billion. Consequently, subprime lending is no longer a minor segment of the market. Subprime loans are estimated to have accounted for about 19 percent of all home-loan originations in 2004, up from less than 5 percent in 1994. (12)

As significant pricing variability has emerged in the market, so have concerns about the fairness of creditor decisions in this regard. Little information has been available to assess the merits of these concerns, and only a few fair lending investigations focusing on pricing issues have been pursued by the federal banking agencies or the Department of Justice. In its review of Regulation C that led to the 2002 revisions, the Federal Reserve Board averred the importance of gathering information to facilitate assessments of the fairness of loan-pricing decisions, particularly for nondepository institutions, which are less likely to be subject to periodic fair lending examinations. Recognizing the costs incurred by lenders to comply with such a reporting and disclosure requirement, the Board limited the scope of the regulation to the disclosure of pricing on loan originations (not loans purchased from other entities or applications that did not result in a loan origination) in the higher-priced segment of the loan market and to focus within that segment only on dwelling-secured loans subject to Regulation Z (which does not cover "business purpose" loans--including some loans to individuals who do not intend to occupy the dwelling being financed). (13)

Specifically, the 2002 revisions to Regulation C require the reporting of the spread between the annual percentage rate (APR) on a loan and the rate on Treasury securities of comparable maturity for loans with spreads above designated thresholds. The APR was selected as the measure of the loan's pricing because it was regarded as the best single measure of the "true" cost of a loan. The thresholds for reporting differ by lien status: 3 percentage points for first liens and 5 percentage points for junior liens. To calculate the rate spread, the lender uses the yield on Treasury securities as of the fifteenth day of a given month depending on when the interest rate was set on the loan. (14)

In establishing this disclosure rule, the Federal Reserve sought to select thresholds that would exclude the vast majority of prime rate loans and include the vast majority of subprime loans. The selection of specific thresholds was based on loanprice data from several sources. (15) The analysis revealed that roughly 98 percent of prime first-lien loans have APRs that would likely fall below the threshold of 3 percentage points for reporting first liens. (16) The analysis also indicated that this threshold would require reporting for about 98 percent of the subprime loans backed by first liens and that the 5 percentage point threshold would capture about 95 percent of the subprime loans backed by junior liens. Overall, data from the Annual Housing Survey covering prime, near-prime, and subprime loans suggested that, in a typical year, the thresholds would fall somewhere in the near-prime range and would require the reporting of about 10 percent of all home loans backed by first liens and about 22 percent of all loans backed by junior liens.

In a given year, various factors may influence the proportion of loans that have prices placing them above or below the pricing thresholds. A change in interest rates can influence the volume and types of loans that exceed the pricing thresholds. With generally rising interest rates, for example, refinancing activity will be reduced, and consequently a larger proportion of loans reported above the thresholds will involve home purchases. Moreover, borrowers who refinance during a period of rising interest rates are likely to differ from those who borrow when rates are falling. When rates are rising, borrowers seeking to refinance their outstanding loans are likely to have more-urgent needs for additional funds that can be raised by a cash-out refinancing or are seeking to lower their total monthly payment obligations by lengthening the terms of their outstanding debt.

Changing interest rates also may affect the proportion of adjustable-rate versus fixed-rate loans originated over the course of a year and thus the mix of loans reported with rates above the threshold. To compensate borrowers for bearing the additional interest rate risk associated with adjustable-rate loans, such loans typically have lower initial APRs than do fixed-rate loans with the same term to maturity. If market participants expect interest rates to rise, these expectations tend to be built into the term structure of interest rates (the "yield curve") and to widen the difference between the initial rates on adjustable-rate and fixed-rate loans. This widening can increase the proportion of fixed-rate loans with APRs above the threshold because the APRs for longer-term adjustable-rate loans will not rise as much in such a market as will those for fixed-rate longer-term loans.

HOEPA Status

Long-standing concerns about predatory lending led the Congress to enact the Home Ownership and Equity Protection Act of 1994, the first federal statute to explicitly target such lending practices. HOEPA, which amends the Truth in Lending Act, applies to closed-end home loans (excluding home-purchase loans) bearing an APR or dollar-amount fees above specified thresholds. (17) The act imposes restrictions on certain loan features, including balloon payments and prepayment penalties, and requires improved disclosures for consumers. HOEPA, like most other federal consumer protection statutes, overrides weaker state laws but permits states to enact stricter rules.

The Federal Reserve Board, which has authority to modify some of the requirements of HOEPA (implemented by the Board's Regulation Z), made such a modification in 2001. The 2001 revisions to the regulation lowered the APR trigger for coverage of first-lien loans from 10 percentage points above the comparable-maturity Treasury security to 8 percentage points (the threshold for junior liens was left at 10 percentage points), adjusted the calculation of the dollar-amount trigger for fees to include amounts paid at closing for optional credit insurance products, prohibited or restricted certain practices, and required improved disclosures. (18) Although these amendments addressed some concerns, predatory lending continues to some degree. Since 1999, about thirty states and numerous local governments have enacted laws regarding predatory lending to address certain practices and contract terms. The Congress has also considered amendments to HOEPA to broaden its scope and to preempt state laws, but to date, no final action has been taken.

In its 2002 amendments to Regulation C (the HMDA regulation), the Board required lenders to report whether a loan is subject to HOEPA. In so doing, the Board recognized that obtaining information on the volume and pattern of lending covered under HOEPA would be useful for a better understanding of the size of the HOEPA-related segment of the market and would allow regulators to focus examinations on the loans and creditors posing the greatest concern. The HOEPA status of loans at banks could always be obtained through on-site examinations; but nondepository lenders are not subject to regular examinations, and thus the extent of their HOEPA-related lending has been largely unknown. Moreover, although banks are examined regularly, the collection of data on HOEPA status on the HMDA/LAR is a much more efficient way for the enforcement agencies to obtain the data and allows for some types of analysis to precede an on-site compliance examination.

Requests for Pre-Approvals

Prospective homebuyers are often asked by sellers to demonstrate that they are likely to qualify for financing. In recent years, many lending institutions have developed pre-approval programs to respond to that request. Such programs typically provide qualified prospective homebuyers with a binding written commitment to finance their purchase, subject to certain conditions related primarily to the property to be purchased and any changes in their financial circumstances. The request for a pre-approval does not generally identify a specific property so that, if granted, it can be used by the prospective buyer with more than one prospective seller. In the past, the HMDA records did not include data on requests for pre-approvals unless they ultimately resulted in an application related to a specific property. Under the expanded reporting requirements, lenders must also report requests for pre-approval that were denied. Disclosure of denials of pre-approval requests is intended to provide more-complete information on the availability of home financing and to facilitate fair lending enforcement. Lenders have the option of reporting pre-approvals that were granted but not acted on by the consumer. (19)

Changes in the Collection of Data Regarding Race and Ethnicity

The 2004 HMDA data incorporate the revised standards of classification for government collection of information on race and ethnicity as established by the OMB. (20) Perhaps the most important OMB revision allows individuals to select multiple racial and ethnic identifications, and HMDA reporting rules were modified to conform to these changes. For HMDA data collected before 2004, applicants for credit had no opportunity to designate both race and ethnicity but had to categorize themselves as being of Hispanic origin or as being in one of five racial categories (American Indian or Alaskan Native, Asian or Pacific Islander, black, white, or other). As of 2004, applicants may designate more than one racial category (American Indian or Alaska Native, Asian, black or African American, Native Hawaiian or other Pacific Islander, or white) and may designate one of two ethnicities (either "Hispanic or Latino" or "not Hispanic or Latino"). (Hereafter, for concision, we refer to the category "black or African American" as black and to the category "Hispanic or Latino" as Hispanic.)

The changes regarding race and ethnicity will make it difficult to align the HMDA data for 2004 with those for earlier years. Most important, applicants who in 2003 were classified as Hispanic were not also classified by their race. Consequently, a comparison of lending activity by race between 2004 and earlier years might lead some to conclude that lending to certain racial groups may have changed when, in fact, the only change was in the classification system.

Changes in the Data-Collection Requirements for Sales in the Secondary Market

The secondary market for home loans is the arena in which loans already originated are bought and sold. HMDA requires that, for a given year, covered institutions report the sales of loans that they originated in that year as well the sales of loans that they purchased in that year. For each sale, the institution must also report the type of purchaser.

HMDA data have long been one of the few sources of loan-level information describing secondary-market activities. The 2004 data are reported using codes that represent revised categories for identifying the secondary-market purchasers. For the first time, the HMDA data identify loans placed in private securitizations, which represent a growing segment of the secondary market. The revisions in the reporting categories are intended to improve the utility of the data.

SUMMARY OF RESULTS FROM THE 2004 HMDA DATA

For 2004, the FFIEC prepared disclosure statements for 8,853 HMDA-reporting lenders--3,946 commercial banks, 1,017 savings institutions, 2,030 credit unions, and 1,860 mortgage companies. Of the mortgage companies, most (1,464) were independent entities--that is, institutions that were neither subsidiaries of banks nor affiliates of bank holding companies (table 1). The disclosure statements consisted of 72,246 distinct reports, each covering the lending activity of a particular institution in each metropolitan statistical area (MSA) in which it had a home or branch office (table 2). The number of reporting institutions was up 9 percent from 2003, in part because OMB's revision of MSA boundaries added, on net, 242 previously rural counties to MSAs. (21)

The number of lenders covered by HMDA is large; however, most of these institutions, whether measured by number of reported applications or loans or by asset size, are small. For 2004, 60 percent of reporting institutions provided information on fewer than 250 loans or applications, accounting for 1.7 percent of the reported data (table 3). Sixty-three percent of the reporting banks had assets of less than $250 million, and they accounted for only 2.2 percent of the applications and loans in the 2004 HMDA data. (22)

At the other end of the spectrum, the twenty-five lenders reporting the largest number of applications accounted for about 42 percent of all the applications reported in the 2004 data (data not shown in table). If HMDA reporters are further aggregated to their highest level of corporate organization (such as a bank holding company), lending is even more concentrated. The twenty-five largest organizations reporting the largest number of applications accounted for 55 percent of the applications in the 2004 data (data not shown in table).

Volume of Applications and Loans

For 2004, lenders covered by HMDA reported on roughly 28.1 million home-loan applications (table 2)--9.8 million for purchasing one- to four-family homes, 16.1 million for refinancing existing home loans, 2.2 million for improving one- to four-family dwellings, and 62,000 related to multifamily dwellings (structures for five or more families). Lenders also reported on about 5.1 million loans they purchased from other institutions. In addition, lenders reported on roughly 330,000 requests for pre-approvals of home-purchase loans that were either turned down by the lender at the time the pre-approval was sought or granted but not acted on by the applicant (data not shown in table). In either case, those 330,000 requests for pre-approval did not reach the stage of an application for a loan for a specific property. (23) The 2004 volume of applications for refinancing fell about one-third from 2003, primarily because of a rise in interest rates.

Conventional and Government-Backed Loans

Among the reported applications for loans to purchase owner-occupied one- to four-family homes (both site-built and manufactured), about 90 percent were for conventional loans--that is, loans not involving a government backing for the lender--most of which involve first liens; the remainder were for government-backed forms of credit, mostly involving the Federal Housing Administration (FHA) (table 4). An even higher share of applications for refinancings (and home-improvement loans) were for conventional loans, an indication that borrowers with government-backed loans either tend to refinance into a conventional loan or tend not to refinance.

The share of HMDA-reported loans backed by the FHA has been declining over the past several years, from about 16 percent in 2000 to about 8 percent in 2004 (data not shown in tables). New, more flexibly underwritten conventional loan products are likely attracting borrowers that would otherwise seek FHA backing. Among these products are interest-only loans, adjustable-rate products that offer flexible payment options, and products that allow smaller down payments, a wider range of credit histories, and reduced documentation of incomes.

Lien Status

The 2004 data, which include for the first time information on the lien status of a loan, indicate that a significant minority of reported loans involve junior liens, particularly loans for home purchases. Among the loans to purchase owner-occupied homes, 13 percent involved junior (subordinate) liens (data not shown in tables). (24)

Homebuyers have various reasons for taking out subordinate loans when they purchase their homes. Some are seeking to raise funds to cover the down payments and closing costs of the first-lien loans used to buy their homes. In some cases, funds raised through the subordinate liens allow homebuyers to avoid the requirement to purchase PMI for first-lien loans with high loan-to-value ratios. In other cases, borrowers take out junior-lien loans to keep the amounts borrowed on their first-lien loans within the loan-size limits used by Fannie Mae and Freddie Mac (discussed below).

Owner Occupancy

Some commentators have attributed part of the strength in housing markets for the past several years to a growing number and share of home sales to investors or individuals purchasing second homes as distinct from those who intend to reside in the units being purchased. HMDA reports help document the role of investors in the housing market because the data indicate whether the property to which an application or loan relates is intended as the borrower's principal dwelling (that is, as an owner-occupied unit). (25) The HMDA data indicate that the share of reported lending for nonowner-occupied purposes remained steady from 1990 through 1995, primarily in the 4.5 percent to 5.5 percent range (whether measured in number of loans or dollar amount of loans), and then began rising. In 2004, the nonowner-occupied share of the home-purchase market in terms of number of loans was about 15 percent and in terms of dollar amount was roughly 13 percent (data not shown in tables).

Lender Specialization

Different types of lending institutions tend to specialize in different types of home loans. Mortgage companies, which extended 54 percent of all the home loans reported in 2004, accounted for roughly 67 percent of government-backed originations. Depository institutions extended 71 percent of reported home-improvement loans and about 89 percent of multifamily loans. Commercial banks and mortgage companies together accounted for about 90 percent of loans on manufactured homes in 2004.

Secondary-Market Activity

HMDA data document the importance of the secondary market for home loans. Of the 20.2 million home loans originated or purchased in 2004 by lenders covered by HMDA, 14.1 million, or roughly 70 percent, were sold in 2004 (data not shown in tables). (26)

Prominent in the secondary market are government-sponsored enterprises (GSEs)--in particular, Fannie Mae and Freddie Mac. (27) For the most part, the purchases of Fannie Mae and Freddie Mac in 2004 consisted of conventional loans originated to purchase homes or to refinance existing loans. These two institutions accounted for nearly 35 percent of the loans purchased by secondary-market institutions. Other types of purchasing institutions active in the secondary market include banks (8 percent of loans sold), private securitization pools (5 percent), and mortgage and insurance companies (9 percent). In some cases, the purchasing institution is affiliated with the originating lender--directly, as a subsidiary, or indirectly, as an affiliate of the holding company that owns the lender. Affiliated institutions accounted for 11 percent of loans sold in the secondary market.

Loans for Manufactured Homes

In the past, users of HMDA data had no certain way to identify which applications and loans involved manufactured homes. To help overcome this limitation, the Department of Housing and Urban Development (HUD) produced annually a list of reporting institutions (typically about twenty) that it believed were primarily in the business of extending such credit. (28) Users of the HMDA data often relied on the HUD list to identify, albeit imperfectly, loans and applications related to manufactured homes. This practice had its own limitations--it could not be used to identify applications and loans related to manufactured homes reported by lenders not on the HUD list, and users often assumed that all loans by the lenders on the list were for manufactured homes when some were not. The expanded HMDA data resolve this problem by explicitly including a code to identify applications and loans for manufactured homes. For background information on manufactured homes, see box "Manufactured Homes in the U.S. Housing Market."

Manufactured Homes in
the U.S. Housing Market

More than 23 million individuals, or roughly 8 percent of
the U.S. population, live in manufactured housing. Typically,
about 10 percent to 20 percent of all construction
starts for single-family housing each year are for manufactured
homes. Most manufactured homes are assembled
in factories, shipped to a home site, and never moved
once installed. Nearly 80 percent of all the manufactured
homes are owner occupied, a rate more than 10 percentage
points higher than that for site-built homes.

Manufactured housing is a significant source of affordable
housing. The average new unit cost about $55,000
in 2003, although prices varied, averaging about $32,000
for single-section homes and nearly $60,000 for multisection
("double wide") units (excluding land costs).
Because the price of a manufactured home is generally
lower per square foot than that of a site-built home, the
manufactured home is particularly attractive to households
with lower incomes. The average annual income of
households owning manufactured homes is less than half
that of those owning site-built homes and about the same
as that of households that rent their homes.

Loans for manufactured homes entail more credit risk than do most other forms of credit extended to consumers. (29) For example, the proportion of loans for manufactured homes that are thirty days or more past due is far higher than for most other consumer credit products and is about twice the rate for conventional loans secured by one- to four-family homes. (30) In part, the elevated credit risk arises from more uncertainty about whether the collateral backing the loan will retain its original value. Much of the credit risk arises from the poorer credit history profiles of the typical borrowers in the manufactured-home loan market compared with those in the site-built homeloan market.

An individual's credit history score (a statistical characterization of an individual's creditworthiness based exclusively on information in a credit record maintained by a credit-reporting agency) is a common metric of credit risk. (31) Among individuals who have manufactured-home loans (whether home loans or, as is the case for most manufactured homes, personal-property loans), the average credit history score as of June 30, 2003, was 666, a score nearly 70 points lower than the average among individuals with loans secured by one- to four-family site-built homes. Moreover, nearly 25 percent of the individuals with loans secured by manufactured homes had credit history scores below 600, a threshold that is often associated with high-risk lending, compared with only about 5 percent of the individuals with loans backed by site-built units.

Lenders recognize the elevated risks related to loans backed by manufactured homes and factor these risks into the interest rates they charge borrowers. (32) Lender caution is also reflected in the very high denial rates on applications for loans backed by manufactured homes.

The 2004 HMDA Data on Manufactured Housing

The 2004 HMDA data indicate that, on the basis of applications that lenders received after January 1, 2004, nearly 4,400 lenders extended more than 242,000 manufactured-home loans. (33) About 57 percent of these loans were for home purchases; most of the rest were for refinancing an earlier loan (data derived from table 5). Commercial banks, the largest source of loans on manufactured homes, extended 46 percent of the total number; mortgage companies extended 44 percent.

The data indicate further that manufactured-home lending is a relatively concentrated business. The ten lenders that extended the largest number of manufactured-home loans in 2004 accounted for one-third of all such loans that year, and the top twenty such lenders accounted for 42 percent (data not shown in tables). Likewise, 60 percent of the lenders that extended manufactured-home loans in 2004 extended ten or fewer such loans. The 2004 data indicate that thirty-five lenders could reasonably be considered to have specialized in manufactured-home lending that year (see box "The HUD List of Specialists in Manufactured-Home Lending and the 2004 HMDA Data").

The HUD List of Specialists
in Manufactured-Home Lending
and the 2004 HMDA Data

Before 2004, HMDA reporters were not required to identify
which of their applications and loans involved manufactured
homes, and identifying all of the lenders offering
such credit was impossible. For 2003, HUD's list of
manufactured-home loan specialists who are also HMDA
reporters identified 19 such lenders. Only 13 of the
19 lenders that reported 2003 HMDA data provided
2004 HMDA data under the same name and identification
number. These 13 lenders accounted for 15 percent
of all the manufactured-home loans reported in the 2004
data. Among the 6 manufactured-home lenders on the
2003 HUD list that did not report in 2004 under the same
name or identification number, 2 reported under different
names and identification numbers. These 2 lenders
reported information on only about 950 loans related
to manufactured homes (about 0.4 percent of the total).
Of the 15 manufactured-home loan specialists on the
HUD list that reported data for 2004, only 11 were in fact
primarily involved in extending loans on manufactured
homes.

For 35 lenders that supplied 2004 data (including the
11 on the 2003 HUD list), lending for manufactured
homes constituted at least 80 percent of their reported
lending activity, and so they may reasonably warrant
consideration as specialists in manufactured-home lending.
Among the rest of the approximately 4,400 reporting
lenders that had extended at least one manufactured-home
loan in 2004, about 500 indicated that the proportion
of their originations related to manufactured homes
was at least 20 percent but less than 80 percent, and the
remaining 3,900 indicated that the proportion was less
than 20 percent.

For a number of the largest lenders that extended manufactured-home loans (measured by the number of such loans), that business segment was only a very small portion of their lending activity, according to the 2004 data. In fact, among the twenty-five firms that extended the largest number of manufactured-home loans, only three could be characterized as focused primarily on that business segment. For virtually all the rest, manufactured-home lending amounted to 5 percent or less of their total lending activity.

Of those obtaining loans to purchase manufactured homes, 41 percent were of lower income, whereas of those borrowing to purchase site-built homes, about 24 percent had lower incomes (table 6). (34) On average, minority borrowers have lower incomes than do non-Hispanic white borrowers, but only about 18 percent of manufactured-home purchasers were members of a racial or ethnic minority group, whereas about 30 percent of purchasers of site-built homes were minorities (data derived from table 6). (35)

Requests for Pre-Approval

The 2004 data for the first time include information on certain types of requests for pre-approval of home-purchase loans. Since pre-approval programs pertain only to requests for loans to purchase a home, the HMDA data do not include pre-approval information for applications involving a refinancing or home-improvement loan. Although all requests for pre-approval that are turned down must be reported, lenders have the option of reporting requests for pre-approval that were approved but not acted on by the consumer. Because many lenders apparently chose not to report any optionally reportable requests for pre-approval, the new data do not account completely for pre-approval activity. Nonetheless, the new reporting scheme is sufficiently comprehensive to identify which individuals were denied at the pre-approval stage and which successful borrowers initiated the borrowing process through a pre-approval program.

Nearly half of all lenders reported some pre-approval activity, although the volume of such activity varied greatly across lenders. (36) The five lenders that reported the greatest number of requests for pre-approval accounted for one-third of all such requests. Some differences in the propensity to offer pre-approval programs were found by type of institution; more than half of the reporting credit unions, savings institutions, and mortgage companies reported requests for pre-approval, but only about one-third of the commercial banks reported such information.

Although requests for pre-approval are far fewer in number than home-purchase loan applications that do not begin through this channel, they are not rare events. The 2004 data include information about 1 million requests for pre-approval for first-lien loans to buy homes and about 100,000 for junior liens. Of those institutions with pre-approval programs, 82 percent did not report any pre-approval requests that were approved but not acted on by the consumer, an indication that these institutions chose not to report all their requests for pre-approval.

DENIALS AND PRICING IN THE 2004 DATA

A central element of the 1989 revisions to HMDA was the collection of loan-level data on the disposition of home-loan applications, and the 2002 revisions to Regulation C expanded this concept to include loan-level information on pricing. This section summarizes the aggregate outcomes on both points. Because the transition rules regarding the reporting of data create problems for assessing some of the 2004 data regarding loan pricing, as they do for manufactured homes and pre-approvals, the analysis that follows excludes "transition" applications--those submitted before January 1, 2004 (data on these applications are shown as memo items in tables 7 and 8). Otherwise, information is given on all applications reported under HMDA. For presentation, applications were grouped into twenty-five product categories based on loan and property type, purpose of the loan, and lien and owner-occupancy status. (37) For each product category, information is provided on the number of total and pre-approval applications, application denials, originated loans, loans with prices above the thresholds, loans covered by HOEPA, and the mean and median spreads for loans priced above certain thresholds.

Denial Rates

For the past fifteen years or so, the HMDA data have been the primary source of publicly available data on the disposition of applications for home loans. The expanded HMDA data for 2004 provide new opportunities to assess patterns in the disposition of applications at different stages of the lending process and across product lines and applicant groups.

Denial Rates across Products

The incidence of denials differs substantially across loan products. Lenders deny only about 15 percent of the applications for home-purchase loans on one- to four-family site-built homes, whether the loans are secured as a first lien or a junior lien and whether they are conventional or government backed (table 7). In contrast, about 30 percent to 36 percent of refinancings and home-improvement loan applications involving first liens are denied, as are about 50 percent of the applications for manufactured homes. The main exception to this pattern is applications for government-backed first-lien loans for refinancings, which have a denial rate similar to that of home-purchase loans.

Of particular importance are the disposition patterns for applications for manufactured homes. As noted, past HMDA data did not distinguish applications for manufactured homes from those for site-built properties. Analysis of the HUD list of manufactured-home loan specialists suggested that such lenders had very high denial rates, and that for lenders offering both manufactured-home loans and other home loans a distorted picture of their propensity to deny credit could easily be drawn. The 2004 data confirm the importance of distinguishing applications for manufactured homes from those for site-built properties. For example, adding the applications for conventional home-purchase first liens for manufactured homes to those for one- to four-family site-built homes would increase the number of total lender actions on applications only 7 percent but the number of denials more than 25 percent. The denial rate for the category "conventional home-purchase first liens" would increase from 14.9 percent to 17 percent (data derived from table). Although this change might not appear large in the aggregate, for some lenders it could create a major distortion.

Denials of Requests for Pre-Approval

Denial rates for applications that begin with requests for pre-approval can be computed in different ways, especially since lenders need not report approved requests for pre-approval not acted on by the borrowers. One way to assess the disposition of applications received through the pre-approval process is to compute denial rates for requests for pre-approval separately from the denial rates for subsequent applications related to a specific property. Another way of assessing denial rates is to combine the two stages (pre-approval requests and subsequent applications for a specific property) and to treat a denial at either stage as a denial.

The denial rates for pre-approval requests (column 3 of table 8) are similar to the denial rates for all applications for home loans on specific properties (column 4 of table 7). Not surprisingly, the denial rates on applications for a specific property that began as requests for pre-approval (derived from columns 5 and 6 of table 8) are lower than the denial rates on pre-approval requests and on applications for a specific property that did not come through the pre-approval process. But although they are relatively low, the denial rates for pre-approved borrowers are not zero: More than 8 percent of pre-approved applicants for conventional first-lien home-purchase loans are turned down when they apply for a loan on a specific property.

If we view requests for pre-approval and applications for loans to purchase a specific property as elements of a single process, the data suggest that the overall denial rates for applicants for home loans on specific properties who came through the pre-approval process are about the same as for applicants who did not first request a pre-approval. Seventeen percent of the applicants for conventional first-lien home-purchase loans who came through the pre-approval process were denied versus 15 percent of those who did not first request a pre-approval (data derived from tables 7 and 8).

However, origination rates for the two groups were very different. Only 49 percent of the applicants for conventional first-lien home-purchase loans who began the process with a request for a pre-approval ended up with a loan, compared with 67 percent of other applicants. This difference appears to stem not from lender actions but from markedly different rates of withdrawal from the process by applicants. Thirty-four percent of applications for conventional first-lien home-purchase loans that started at the pre-approval stage are withdrawn by the applicant at some point (or not acted on by the lender). The withdrawal rate for other applications is much lower (19 percent).

Reported Reasons for Denial

The HMDA data include information from lenders on why they turned down an application. Lenders generally provide the information voluntarily; however, two federal bank supervisory agencies, the Office of Thrift Supervision and the Office of the Comptroller of the Currency, require the institutions they supervise to report this information. (38)

Institutions are allowed to cite up to three reasons (from a list of nine) that an application was turned down. Overall, one or more reasons for denial were provided for about 81 percent of the denials across all loan products and for about 75 percent of the denials for home-purchase loans (data not shown in tables). Poor or no credit history was the most frequently cited reason for denying applications: Credit-related issues were cited in about 26 percent of the denials of applications for conventional first-lien loans to purchase one- to four-family site-built homes and in about 52 percent of the denials of applications for such loans to purchase manufactured homes. Other reasons often cited for credit denials involved excessive debt-to-income ratios, issues related to collateral, and unverifiable or incomplete information on applications; a catch-all category in the HMDA data labeled "other" was also frequently cited.

Loan Pricing

Because of the transition rules, some unknown proportion of higher-priced loans was reported in the same way as loans that did not meet the threshold requirements. (39) The inability to identify higher-priced loans that were originated in 2004 but had application dates preceding that year means that users of the data need to take special account of these applications when conducting analyses.

Reasons for Loan-Price Variation

The HMDA data on loans in the higher-priced segment of the home-loan market do not include much of the information that might explain variations in the prices of reported loans. Among the factors reflected in loan pricing are the cost of the funds to be lent, credit risk, prepayment risk, overhead expenses, loan-servicing costs, the negotiating abilities and inclinations of the creditor and borrower, the possibility of discriminatory pricing, and variations in the channels through which a loan application at a given lender may be processed (see box "Reasons for Loan Price Variation").

Reasons for Loan Price Variation

As in credit underwriting, loan pricing reflects a wide
variety of factors.

Cost of funds. The cost of funds is the largest element in the
overall cost of extending prime-market home loans and a
significant factor for loans in the higher-priced segment of
the market. Funding costs vary with the expected duration
of the debt and the creditworthiness of the borrower. Also,
many creditors originate loans for subsequent sale in
the secondary market; consequently, the prices offered by
secondary-market participants for home loans bear heavily
on the pricing decisions for such loans.

Credit risk. Credit risk is the probability that a loan will go
into default. Loans that involve greater credit risk carry
higher prices. On average, loans in the prime market entail
substantially lower credit risk than do those in the subprime
market. (1) Interest rates on loans increase with the rate of
serious delinquency, even for subprime loans, an indication
that loans that pose greater credit risk carry higher rates of
interest (chart A).

[GRAPHIC OMITTED]

Credit risk is a function of the creditworthiness of the
borrower, the equity in the home securing the loan, and the
likelihood that proceeds of a foreclosure sale of the home
will satisfy the obligation if default occurs. In general, the
creditworthiness of borrowers is related to their income and
employment prospects; available assets if financial problems
arise; claims on their income from servicing other
debts; and credit history, which, in part, reflects their willingness
and ability to repay credit. As noted, in underwriting
loans, credit history is often summarized and measured
by a credit history score. Equity in a home is measured at
the time of loan origination by a loan-to-value ratio (LTV).

The importance of credit history in loan pricing is illustrated
by the fact that interest rates are higher for loans with
lower credit history scores (chart B). For the most part,
borrowers in the prime segment of the market have credit
history scores that indicate they pose relatively little credit
risk. Borrowers in the higher-priced segment of the market
typically have weaker credit history profiles for one or
more of several reasons: previous failures to make loan
payments as scheduled, collection agency actions, bankruptcy
or adverse court judgments, or little or no previous
experience with credit.

[GRAPHIC OMITTED]

Prepayment risk. Prepayment risk measures the possibility
that a loan will be repaid before the end of the loan term.
Most early payoffs of home loans are attributable either to
the sale of the home or the refinancing of the loan, typically
when rates have fallen sufficiently from the rate on the
existing loan. Because a prepayment results in payment of
the principal ahead of schedule, the lender (or secondary-market
investor) must reinvest the funds at the new market
rate, which may be lower than the old rate, particularly in
the case of a refinancing.

Although the possibility of prepayment is well understood,
estimating when it will happen is quite difficult. For
this reason, lenders compensate for the risk either by including
prepayment penalties in their loan contracts or by pricing
the risk in their calculation of the interest rate on the
loan. The first of these options, the prepayment penalty, is
rare in the prime segment of the market but is more common
in the subprime segment.

Borrowers in the higher-priced segment of the home-loan
market have higher prepayment rates than others because
many of them improve their credit profiles over time as they
make regular payments, and this improvement in turn
allows them to qualify for a lower-rate loan. Our review of
depersonalized credit record information from one of the
three national credit-reporting agencies indicates that, as of
June 30, 2003, almost one-fourth of those with outstanding
home-loan debt and with credit history scores between 580
and 620 (a credit score range associated with individuals
with subprime credit quality) increased their credit scores
40 points or more over the ensuing eighteen-month period.
A change in credit score of this magnitude would typically
be sufficient to move their credit risk profiles into the
near-prime or prime segment of the market. (2)

The effect of even a small improvement in the credit
history score is much larger for borrowers in the higher-priced
segment of the home-loan market than for those
in the prime segment. For a higher-priced loan, a small
improvement in the borrower's credit history score may
translate into a substantial reduction in interest rates and
may encourage prepayment (chart B). (3) Because pricing in
the prime segment varies little by credit score, borrowers in
the prime market are less likely to obtain a lower-priced
loan if their credit scores improve.

Another factor that may result in elevated rates of prepayment
in the higher-priced portion of the market is the
practice referred to as "loan flipping." Flipping is inducing
a borrower to refinance a loan repeatedly--even though
the refinancing may not be in the borrower's interest--and
charging high fees with each refinancing.

Overhead expenses. Overhead expenses represent a fairly
small component of the cost of lending for most home
loans. However, borrowers who have experienced payment
problems in the past, or who have little or no credit history,
or who are unable or unwilling to document their employment
histories or income are likely to require more time to
underwrite. The higher cost of underwriting may be passed
on to such borrowers and can result in prices that place their
loans in the higher-priced segment of the market. Marketing
and other expenses incurred to identify market opportunities
and solicit customers may also differ across segments of
the home-loan market.

Servicing costs. Servicing costs are expenses incurred to
process and distribute loan payments, monitor accounts,
and deal with borrowers who fall behind in their payments.
Servicing costs can be particularly high if the loan involves
a foreclosure--that is, a forced sale. Because the higher-priced
segment of the market has high rates of serious
delinquency, servicing costs are higher than in the prime
market. And because higher-priced loans tend to be smaller
than prime loans, the costs of servicing and the costs of
extra underwriting efforts (noted earlier) must be spread
over a smaller dollar volume of loans. Borrowing a relatively
small amount increases the possibility that elevated
costs will lead to a higher-priced loan because any given
amount of fixed costs passed on to the borrower increases
the APR more on a smaller loan than on a larger loan.
Hence, these costs have a larger effect on loan prices in the
higher-priced segment of the market than in the prime
segment.

In general, the cost and risk-related factors noted earlier
may be measured in an objective way and are demonstrably
related to the costs, and hence the prices established, for
credit. Two additional, and related, pricing factors are not
necessarily objective and, moreover, are more likely than
others to raise fair lending concerns: discretionary pricing
by loan officers and price negotiations between creditor and
applicant.

Discretionary pricing. Many creditors provide their loan
officers and agents working on their behalf (for example,
mortgage brokers) with rate sheets that indicate the creditors'
minimum prices by product (for example, for conventional
loans of various types or with various types of
government backing), loan characteristics (for example,
term to maturity and LTV ratio), and borrower creditworthiness
(for example, credit history score and debt-to-income
ratio). In some cases, loan officers and brokers are allowed
to deviate from these prices as market conditions warrant
or allow. A loan officer may quote a prospective borrower
a price above the rate sheet (sometimes referred to as an
"overage"), and if the consumer accepts the price without
demanding cash back to offset loan fees or other closing
costs, the contract interest rate or loan fees on such "overaged
loans" will be higher than they might otherwise have
been.

Discretionary pricing can be a legitimate business practice
and can help ensure that markets allocate resources
in the most efficient way. However, when loan officers are
permitted latitude in establishing prices, the lender runs the
risk that differential treatment on a basis prohibited by law
may arise. Obtaining overages more often, or in higher
amounts, from minority borrowers or targeting only minorities
for overaging may constitute a fair lending violation
unless some legitimate, nondiscriminatory reason exists for
the result.

Price negotiations. Price variation can also arise because
less sophisticated or less knowledgeable borrowers are not
as likely to shop for credit or to realize that they may
negotiate with the lender over the interest rate and fees.
Moreover, lower-income borrowers may be disproportionately
represented in the category of less sophisticated borrowers.
Given that minority borrowers have disproportionately
lower incomes, there is some likelihood that they will
be overrepresented among borrowers with overaged loans.
Such results may be interpreted by some as demonstrating
unlawful discriminatory pricing by lenders.

Differences in the extent to which borrowers negotiate or
shop for the best deal may result in a pattern of overage
loans that is not illegal but that nonetheless may be difficult
for a lender to document and explain. Moreover, instances
of different negotiating strengths among borrowers can be
difficult to distinguish from illegal discriminatory treatment
in which loan officers quote loan rates or provide information
or assistance that varies according to the race, ethnicity,
sex, or other prohibited characteristic of the borrower.

Variations in loan-processing channels. The delivery channels
through which borrowers obtain loans vary widely
across lenders. On the one hand, underwriting and pricing
may be centrally controlled even though the application
may begin on the Internet or with a mailed solicitation or
at a bank branch. On the other hand, in complex financial
organizations with bank branches, multiple affiliates, decentralized
loan production offices, indirect brokerage operations,
and nonbank subsidiaries, each application may
be subject to a different underwriting and pricing regime
depending on its point of initiation. The 2004 HMDA
pricing data suggest that the delivery channel through
which a borrower obtains a loan may matter. As discussed
elsewhere in this article (see section "Incidence
of Higher-Priced Lending for Selected Subgroups"), the
incidence of higher-priced lending is higher for borrowers
who live outside the assessment areas of lenders covered by
the Community Reinvestment Act of 1977 (CRA) than for
those who live inside these areas. (4) This difference may be
due to a reliance on different delivery channels for loans
within and outside these lenders' assessment areas.

(1.) See Amy Crews Cutts and Robert Van Order (2004), "On the Economics
of Subprime Lending," Freddie Mac Working Paper 04-01 (Washington:
Freddie Mac, January), www.freddiemac.com/corporate/reports.

(2.) The three national credit-reporting agencies are Equifax
(www.equifax.com), Experian (www.experian.com), and Trans Union
Corporation (www.transunion.com).

(3.) Of course, prepayment penalties may deter prepayment among some
borrowers in the higher-priced segment of the market even when their
credit scores improve.

(4.) The assessment areas of lenders covered by the CRA include
principally the locales in which a lender has its main or branch
offices and its deposit-taking automated teller machines. For a more
complete definition of CRA assessment areas, see the Federal Reserve
Board's Regulation BB, section 228.41. See also Robert B. Avery,
Glenn B. Canner, Shannon C. Mok, and Dan S. Sokolov (2005), "Community
Banks and Rural Development: Research Relating to Proposals to Revise
the Regulations That Implement the Community Reinvestment Act,"
Federal Reserve Bulletin, vol. 91 (Spring), pp. 202-35.

Issues Raised by Expanding the Disclosure of Pricing-Related Items

Although disclosures that are more comprehensive could improve the understanding of loan pricing, it would impose new costs on lenders to collect and report the additional data, raise difficult reporting issues, and might pose privacy concerns for consumers and reveal otherwise nonpublic information about lenders' business strategies. Adding new data elements to mandated disclosures would require institutions to train staff, modify data collection and reporting software, and expand controls to ensure the reporting of correct data.

Further, the fact that lenders differ in the factors they consider in setting loan prices makes it difficult to select additional data elements that would allow a complete understanding of the determinants of a particular lender's pricing method. Also, some loan-pricing items that might be added to the HMDA data raise technical issues about what, precisely, to report. For example, if lenders were required to report credit scores, getting consistent data across lenders would be difficult because institutions rely on different types of credit scores in underwriting--for example, some lenders rely on generic FICO credit history scores (see text note 31), whereas others use proprietary credit scores developed from information on their own experience with lending.

The potential for compromising consumer privacy is also a consideration. More than 90 percent of the loan records in a given year's HMDA data are unique--that is, an individual lender reported only one loan in a given census tract for a specific loan amount. These unique loan records can be matched with other publicly available information, such as property deed records, to determine the identities of individual borrowers. With such a match, any data item in the HMDA database, such as loan pricing, becomes publicly known. During the Board's review of Regulation C, some commenters cited this circumstance as a reason not to require the reporting of price information. Expanding HMDA to include data items such as credit scores that may be considered highly personal would likely also raise privacy concerns.

Finally, requiring lenders to disclose additional information about their lending activities may result in the disclosure of otherwise nonpublic information about lenders' business strategies. HMDA now requires disclosure of information about lending patterns--for example, pricing patterns--that otherwise would not be public. In general, such disclosure is pro-competitive because it helps possible entrants to the market identify business opportunities and lowers the information advantage of market incumbents. An argument could be made that disclosing detailed information about lenders' business strategies through HMDA might discourage lenders from testing new products or entering new markets by creating a risk that, because of such disclosure, a lender would lose its competitive advantage before it had recouped the fixed costs of entry. The likelihood of such discouragement would depend critically on whether potential competitors could discern the essential elements of a lender's business strategy (a discernment that would depend, in part, on which data items had to be disclosed) and, further, distinguish successful business strategies from unsuccessful ones (a distinction that could not be made on the basis of HMDA data alone). Ultimately, any decision to add data items to the reporting requirements of HMDA should be based on a careful weighing of the costs and benefits of such additional reporting.

The Interest Rate Situation in 2004

The interest rates prevailing in a given year can significantly affect the proportion of loans that exceed the thresholds established by the Federal Reserve for determining whether a loan is "higher-priced." For 2004, the rate on Treasury securities used to calculate the spread for home loans with thirty-year terms varied from 4.67 percent to 5.54 percent. This variation implies that the threshold for reporting a first-lien loan as higher-priced ranged from 7.67 percent to 8.54 percent over the year. For junior liens, which typically have a shorter term to maturity than do first liens, the reporting threshold ranged from about 8.78 percent to 9.79 percent for a fifteen-year loan (different terms to maturity would yield somewhat different ranges).

Data derived by Freddie Mac from its Primary Mortgage Market Survey show that the spread for average interest rates for first-lien conforming mortgages extended in 2004 imply a typical gap between the thirty-year Treasury rate and an estimated APR for prime-rate loans of between 1 percent and 1.25 percent. (40) This gap implies that a thirty-year first-lien home loan would have to have been priced between 1.75 percentage points and 2 percentage points above a prime-rate home loan to exceed the HMDA price-reporting threshold. Such a price spread was around the upper end of the near-prime market, but it was probably still below the levels associated with most subprime loans. In future years, the thresholds may cover a greater proportion of the near-prime segment of the market.

Results: The Incidence of Higher-Priced Lending

Several patterns are revealed in the 2004 HMDA pricing data (table 7). First, in almost all cases, government-backed loan products show lower incidences of higher-priced lending than do comparable conventional loan products. For example, among first-lien home-purchase loans for site-built homes, 11.5 percent of conventional loans have APRs above the pricing threshold versus only 1.3 percent of government-backed loans. Second, with few exceptions, first-lien loans have a substantially lower incidence of higher-priced lending than do junior-lien loans for the same purposes. For example, nationally the incidence of higher-priced lending for conventional first-lien refinance loans was 15.5 percent whereas for comparable junior-lien loans it was 27.4 percent. Third, manufactured-home loans exhibit the greatest incidence of higher pricing across all loan products, a result consistent with the elevated credit risk associated with such lending. For example, 57.1 percent of conventional first-lien loans used to purchase manufactured homes were higher priced, in sharp contrast to the 11.5 percent rate for comparable loans for site-built homes. Finally, the lower incidence of higher-priced lending (shown in the memo item in table 7) for loans initiated in the transition period reinforces the decision to exclude such loans from the pricing analysis.

Rate spreads for higher-priced loans. Variation in mean and median spreads across products for loans with rates above the threshold is much smaller than variation in the incidence of higher-priced lending. Because the threshold for reporting is set higher for junior liens than for first liens, higher-priced junior-lien products have higher mean and median spreads. Once again, manufactured-home loans stand out in that they have the highest average spreads among all the loan products with comparable lien status.

Except for loans backed by manufactured homes, the vast majority of higher-priced loans have prices within 1 or 2 percentage points of the pricing thresholds. Only a very small proportion of higher-priced first-lien loans have spreads that exceed 7 percentage points. Similarly, only a small proportion of junior-lien loans have spreads of 9 percentage points or more: But home-improvement loans provide two exceptions--30.7 percent of conventional junior-lien home-improvement loans and 47.6 percent of government-backed junior-lien home-improvement loans have spreads of 9 percentage points or more. Reflecting these distributions, the mean and median spreads for most loan products fall within 2 percentage points of the reporting thresholds. The exceptions include loans backed by manufactured homes and junior-lien home-improvement loans, for which the distribution of prices is more even.

Lenders of higher-priced loans. Most lenders covered by HMDA reported extending few if any higher-priced loans for 2004 (data not shown in tables). Nearly 3,300 lenders reported making no such loans, and an additional 2,300 reported making between one and nine such loans. Nearly 500 lenders reported making more than 100 higher-priced loans; these more-active lenders accounted for 96 percent of all reported higher-priced lending of this type. Moreover, the 10 lenders with the largest volume extended 38 percent of all higher-priced loans.

Variation across metropolitan areas. The analysis of separate geographic markets shows that higher-priced lending varies considerably across MSAs. (41) For this exercise, the focus is on the incidence of higher-priced lending among conventional first-lien home-purchase loans for site-built, owner-occupied homes. The MSA with the lowest incidence of higher-priced lending for this product is the San Francisco --San Mateo-Redwood City area in California, at 2 percent; the MSA with the highest incidence is the McAllen-Edinburg-Pharr area in Texas, at 42 percent. (42) A review of the full list of MSAs indicates that most of the areas with the highest incidence are in the southern region of the country, whereas those with the lowest incidence are much more dispersed.

Although a comprehensive analysis of the reasons for such wide variation in the incidence of higher-priced lending is beyond the scope of this article, a review of data from one of the three national credit-reporting agencies finds a close association between the proportion of individuals in an MSA county with a low credit score and the incidence of higher-priced lending in that area.

Loans Covered by HOEPA

The 2004 HMDA data indicate whether a loan is subject to the Home Ownership and Equity Protection Act of 1994. Before 2004, little information was publicly available about the extent of such lending or the number or type of institutions involved in such activities. However, HMDA data do not capture all HOEPA-related lending. Some HOEPA loans are extended by institutions not covered by HMDA, and some HOEPA loans that are made by HMDA-covered institutions are not reported under the Federal Reserve Board's Regulation C, which implements HMDA. In particular, if the proceeds of a home-secured loan are not used to refinance an existing home loan or to finance home improvement, then the loan may be covered by HOEPA but is not reportable under Regulation C. (43)

Incidence of HOEPA-Related Lending

For 2004, 1,948 lenders reported extending 24,594 loans covered by HOEPA (table 7). The HOEPA loans accounted for only 0.003 percent of all the originations of home-secured refinance or home-improvement loans reported for 2004 (derived from the table). (44)

HOEPA lending is relatively concentrated: The ten lenders that reported the largest number of HOEPA originations accounted for 37 percent of all reported HOEPA loans (data not shown in tables). At the other extreme, 801 institutions reported only one HOEPA loan, and 327 reported only two such loans. Most HOEPA loans were extended by banks (50 percent of the total) or by bank subsidiaries or affiliates of bank holding companies (14 percent of the total); independent mortgage companies extended the rest.

Characteristics of HOEPA-Related Lending

As noted, HOEPA applies only to closed-end home loans (whether for refinancing or home improvement) and not to home-purchase loans or home equity lines of credit. The vast majority of HOEPA loans reported in the 2004 data involved conventional products: Only 7 percent of reported HOEPA loans were government backed (derived from table 7). About 50 percent of the reported HOEPA loans involved first-lien conventional loans (more than 80 percent of these were for refinancings, and the rest were for home improvement), and about 40 percent involved junior-lien conventional loans (more than half of these were for home improvement).

On average, reported HOEPA loans are not large (data not shown in tables). For example, for conventional refinancing loans covered by HOEPA, the average size of a first-lien loan was $98,650, and the average size of a junior-lien loan was $31,705. In contrast, the average sizes of such loans not covered by HOEPA were $173,125 for a first-lien loan and $54,581 for a junior-lien loan.

Reported HOEPA lending varies among borrowers sorted by borrower income, race, and ethnicity and among census tracts sorted by census tract income, population, and location. However, the data do not indicate that HMDA-reportable HOEPA lending is focused primarily on lower-income or minority individuals or on those residing in lower-income neighborhoods or neighborhoods with high concentrations of minority individuals. For example, although reported HOEPA loans were extended to borrowers in all income groups, about three-fourths were extended to middle- and higher-income borrowers (data not shown in tables). Similarly, most reported HOEPA loans were extended to non-Hispanic white borrowers. Most of the homes secured by reported HOEPA loans were in middle- or higher-income areas and in areas with a minority population that was less than 20 percent of the total population.

LENDING OUTCOMES BY RACE, ETHNICITY, AND SEX

One of the primary purposes of the HMDA data is to allow comparison of the outcomes of the lending process for applicants and borrowers grouped along many dimensions, including by race, ethnicity, and sex. Outcomes reported in the HMDA data include the disposition of applications (denial rates) and, as of 2004, the pricing of loans. Gross outcomes for different groups of borrowers can be compared, but HMDA data include information on a number of items whose presence or absence for borrowers can be made consistent ("controlled for") in conducting the comparisons. Clearly the HMDA data do not include all the factors that are involved in credit underwriting and pricing. However, by controlling for variations so as to make borrowers as similar as possible on the dimensions of the data that are available, one can account for some of the factors that may explain differences in the outcomes of the lending process among groups.

The HMDA data allow individuals to be matched by loan type and purpose, type of property securing the loan, lien status, owner-occupancy status, property location (for example, same MSA or even same census tract), income relied on for underwriting, loan amount, and time of year when the loan was made as well as by whether the loan involved a co-applicant. In comparing lending outcomes across racial and ethnic groups, one can match for the sex of the applicant and co-applicant, and in comparing outcomes by sex, one can match for race and ethnicity.

Comparisons in outcomes across groups can be conducted at the level of an individual institution, groups of institutions (for example, manufactured-home lending specialists), geographic market, or populations as a whole. Further, a variety of statistical methodologies can be used to control for the effects of the credit-related or other factors in HMDA noted above. The full range of these comparisons is beyond the scope of this article. However, to gain an understanding of the differences that are likely to be important, we analyzed the 2004 data using statistical matching criteria similar to those used in the Federal Reserve's statistical analysis program (described in the section "Using the Expanded HMDA Data as a Screening Tool for Fair Lending Enforcement").

We restrict the analysis to denial rates, the incidence of higher-priced lending, and the mean spreads paid by borrowers with higher-priced loans, and we compare these outcomes across eleven groups--nine racial or ethnic groups and the two sexes. We conduct the analysis for thirteen of the twenty-five loan products covered in table 7. (45)

We present the comparisons at three levels, one unadjusted and two adjusted. The first level for each group is the raw, or unadjusted, average outcome. The second level is the average outcome as adjusted for the borrower-related factors reported in the HMDA data--income, loan amount, location (MSA) of the property, presence of a co-applicant, and (in the comparisons by race and ethnicity) sex or (in the comparisons by sex) race and ethnicity; applying this adjustment is hereafter termed "adjusting (or controlling or accounting) for borrower-related factors." The third level is the average outcome as adjusted for all the items in the second level (the borrower-related factors) plus the lending institution--applying this adjustment is hereafter termed "adjusting (or controlling, or accounting) for borrower-related factors plus lender."

Applications subject to the transition rules were excluded from the pricing comparisons; however, they were included for the denial-rate comparisons. (46) Also excluded from the sample are applicants residing outside the fifty states and the District of Columbia, applications deemed to be business-related, and requests for pre-approval that were denied by the lender or that were granted by the lender but not acted upon by the borrower. Otherwise, the sample includes all 2004 HMDA applications acted upon by the lender in the thirteen product areas.

Unadjusted and adjusted comparisons for lending outcomes across groups are discussed in the sections below. For purposes of presentation, the adjusted outcomes shown in the tables are normalized so that, for the base comparison group (non-Hispanic whites in the case of comparison by race and ethnicity, and males in the case of comparison by sex), the adjusted mean at each adjustment level is the same as the unadjusted mean. Consequently, the adjusted outcomes for any other group represent the expected average outcome if the members of that group had the same distribution of control factors as that of the base comparison group.

Denial Rates across Groups

Unadjusted mean denial rates vary across loan categories for all groups of borrowers (table 9). For example, the mean unadjusted denial rate for Asians is lowest for government-backed first-lien home-purchase loans (12.4 percent) and highest for conventional junior liens for home improvements (46.1 percent).

For every loan category, American Indians, blacks, Hispanic whites, and the group for which race was missing have higher unadjusted mean denial rates than non-Hispanic whites, with the highest rates generally for blacks and the rates for Hispanic whites lying about halfway between those for blacks and those for non-Hispanic whites. The denial rates for each of the other minority groups vary in their relationship with the rates for non-Hispanic whites.

With few exceptions, controlling for borrower-related factors reduces the differences among racial and ethnic groups. (Although the effect of controlling for borrower-related factors can widen the racial and ethnic differences in denial rates.) Accounting for lender almost always reduces differences further, although statistically significant differences remain between non-Hispanic whites and most of the other racial and ethnic groups.

For example, for conventional first-lien home-purchase loans, the unadjusted mean denial rate for blacks was 24.7 percent and for non-Hispanic whites 10.9 percent, a difference of 13.8 percentage points. Accounting for income, loan amount, and other borrower-related factors in the HMDA data reduces the difference 2.4 percentage points. Controlling for borrower-related factors plus lender significantly reduces the gap further, to 7.3 percentage points.

The reduction for conventional first-lien refinancing is even more dramatic. The unadjusted difference between black and non-Hispanic white denial rates is 17.6 percentage points, a difference cut by more than half, to 8.1 percentage points, when adjusted for borrower-related factors plus lender.

Differences in denial rates exhibit no consistent pattern with regard to the sex of the applicant. For some products, males have higher denial rates, and for others, females do; but in general, the size of the difference by sex is small. Furthermore, controlling for borrower-related factors plus lender has an inconsistent (but generally small) effect. In fact, in some cases adjustment reverses the sign of the difference--that is, for example, some denial rates that were higher for females than for males before adjustment become higher for males than for females after adjustment.

Incidence of Higher-Priced Lending across Racial and Ethnic Groups

Although most borrowers do not have higher-priced loans, the incidence of higher-priced lending varies substantially across racial and ethnic groups (table 10). Moreover, both the overall incidence of higher-priced lending and the differences across groups varies substantially across loan product categories. For government-backed loan products, small proportions of borrowers have higher-priced loans, and no meaningful differences appear across racial and ethnic groups. At the other extreme, the majority of borrowers for manufactured homes have higher-priced loans; and for this product, significant differences appear across racial and ethnic groups (although these differences are smaller than for some other products).

Differences in the incidence of higher-priced lending across loan products make it difficult to identify consistent patterns across all types of lending. Conventional home-purchase lending and conventional refinance lending include most of the borrowers covered by the HMDA data, and consequently those loan types are the main focus of the discussion of higher-priced lending that follows. In general, unadjusted differences in the incidence of higher-priced lending between non-Hispanic whites, on the one hand, and blacks, on the other, are large, but these differences are substantially reduced after controlling for borrower-related factors plus lender.

Most of the reduction in the difference in the incidence of higher-priced lending across groups comes from adding the control for lender to the control for borrower-related factors. For conventional first-lien home-purchase loans, the mean unadjusted incidence of higher-priced lending was 32.4 percent for blacks and 8.7 percent for non-Hispanic whites, a difference of 23.7 percentage points. Borrower-related factors account for about one-fourth of the difference. Adding to this adjustment the control for lender reduces the remaining gap markedly, to 7 percentage points. For conventional first-lien refinancings, the unadjusted difference between blacks and non-Hispanic whites is 21.7 percentage points; this difference is reduced to 4.7 percentage points after controlling borrower-related factors plus lender, and about two-thirds of that reduction comes from the addition of the control for lender. Similar patterns for differences between black and non-Hispanic white borrowers are found for the junior-lien products.

The picture for Asians differs greatly from that for blacks. Asians have a lower unadjusted mean incidence of higher-priced lending for first-lien conventional home-purchase and refinance loans than do non-Hispanic whites. The gap is narrowed by controlling for borrower-related factors plus lender, although the incidence of higher-priced lending remains lower for Asians than for non-Hispanic whites. Hispanic whites show a pattern similar to that of blacks, but with smaller differences relative to non-Hispanic whites. For first-lien conventional loan products, the adjusted differences for Hispanic whites are less than one-half those for blacks; for conventional home-purchase junior liens, the differences for Hispanic whites are close to those for blacks.

The foregoing analysis indicates that the information in the HMDA data--that is, adjusting the HMDA data for borrower-related factors plus lender--is insufficient to account fully for racial or ethnic differences in the incidence of higher-priced lending; significant differences remain unexplained. Explaining the remaining differences is likely to require more details about such factors as the specific credit circumstances of each borrower, the specific loan products they seek, and the business practices of the institutions they approach for credit.

Understanding the patterns within the HMDA data is also likely to require more information. Two somewhat offsetting patterns are shown in the data. On the one hand, minority borrowers (except for Asians) tend to disproportionately use government-backed products, which show a much lower incidence of higher-priced loans. The implication is that the difference in the incidence of higher-priced lending between minorities and non-Hispanic whites is lower for the combined loan product, conventional and government-backed loans, than for each product separately. On the other hand, for a given loan type, whether government-backed or conventional, non-Asian minorities tend to disproportionately borrow from lenders that have higher incidences of higher-priced loans, a tendency on the part of minority borrowers that accounts for much of the aggregate difference in outcomes between them and non-Hispanic white borrowers.

The disproportionate borrowing by non-Asian minorities from higher-priced lenders could occur because of often benign factors such as a "segmented" marketplace in which different lenders offer different products and borrower groups self-select the product-lender combination that best matches their credit or other circumstances. Such a marketplace does not necessarily raise public-policy concerns regarding fair lending: For example, compared with non-Hispanic whites, minority groups on average may seek loans with higher loan-to-value ratios (perhaps because on average they may have less savings to meet down-payment and closing cost requirements), which are typically higher priced and which are the specialty of certain lenders. This explanation could account for differences in lender choice, but demonstrating it requires loan-specific information--such as loan-to-value ratios--as well as other information that is not in the HMDA data.

However, a situation that might suggest an inadequately functioning marketplace--and that could trigger fair lending concerns--would occur if minority borrowers are incurring prices on their loans that are higher than is warranted by their credit characteristics. Such a problem could arise in one or both of the following circumstances: (1) neighborhoods with high proportions of minority residents may be less well served by lenders offering prime products, a circumstance that would make obtaining lower-priced loans more difficult for well-qualified minorities, or (2) some minority borrowers may be steered to lenders who typically charge higher prices than the credit characteristics of these borrowers warrant. The data reported under HMDA are insufficient to tell us whether either explanation (or any other) is correct, nor do they tell us why minorities disproportionately use government-backed products.

Incidence of Higher-Priced Lending for Selected Subgroups

Although, for reasons discussed above, we cannot definitively explain the racial and ethnic pricing differences that remain after adjusting for borrower-related factors plus lender or the reasons that minorities may disproportionately obtain credit from higher-priced lenders, some additional insights may be gained by analyzing subgroups of lenders and borrowers (table 11). We present data only for blacks and Hispanic whites because they are the two minority groups whose outcomes differ most from those of non-Hispanic whites in the incidence of higher-priced lending. We also restrict the comparison to the two loan-products that account for the largest number of loans--conventional home-purchase loans and refinancings. We analyze these two minority groups and the two loan products with regard to five factors: (1) the borrower's relative income; (2) the relative income of the census tract in which the property related to the loan is located; (3) the location of the property relative to the lender's assessment area as defined under the Community Reinvestment Act of 1977 (CRA); (4) the disposition of the loan--that is, whether the loan was retained or sold and, if it was sold, to whom; and (5) the proportion of the lender's loans that are higher priced (less than 10 percent, 10-49 percent, or 50 percent or more--the last proportion being taken to indicate that the lender specializes in higher-priced lending). (47)

Results for borrowers grouped by their relative income and their census tract income classification offer little evidence that unexplained racial differences in pricing or lender choice vary by income. Although the incidence of higher-priced loans is higher for lower-income borrowers and those who live in lower-income census tracts, the same is true for all three racial groups. Thus, although unexplained price differences are somewhat lower for higher-income borrowers, price patterns are generally similar to those of the overall decompositions.

The decompositions for CRA assessment area and loan sales provide evidence on whether the channel through which a loan was obtained and the subsequent disposition of the loan affect racial or ethnic groups differently. Although the overall incidence of higher-priced loans reported in the 2004 HMDA data is much lower for loans sold to the GSEs, the data offer scant evidence that the disposition (sold or retained) is related to unexplained racial differences.

However, whether the loan was originated by an institution in its CRA assessment area does matter. Differences across groups for lending within an assessment area are about one-third of those for lending outside the assessment area. Moreover, for all racial and ethnic groups, lending within an assessment area exhibits a much lower incidence of higher-priced lending.

One possible explanation for the assessment-area effect may be that the channel through which loans are originated matters. Loans extended to borrowers outside an institution's assessment area may be more likely to have come through mortgage brokers, who may price differently or who operate in areas with different market conditions than do institutions that originate loans directly. Although this pattern may suggest that brokers charge higher prices, particularly to minorities, it is not necessarily evidence of unfair treatment or that the existence of the broker channel adversely affects minorities. It may indicate that brokers serve markets or individuals who are more costly to serve, or whose credit profiles are weaker, and price accordingly. If so, then were it not for broker activity, some of these borrowers might not be served at all or might pay higher prices. Determining whether brokers treat minority borrowers fairly is a complex undertaking and requires information unavailable in the HMDA data. The same can be said about lenders that originate loans through different channels.

Finally, although the aggregate lending patterns of specialists in higher-priced lending exhibit unexplained racial or ethnic differences, so do the aggregate patterns of other lenders to approximately the same degree. And for the higher-priced specialists with the highest incidence of higher-priced lending, differences across racial and ethnic groups are, in some instances, lower than for other lenders. Regarding the racial and ethnic differences in the lending patterns that do exist among higher-priced lending specialists, the analysis shows that income, loan amount, and other HMDA factors appear to explain little of those differences.

Differences in Mean Price Spreads across Racial and Ethnic Groups

Patterns across racial and ethnic groups for the mean spreads paid by those with higher-priced loans are quite different from patterns across such groups for the incidence of higher-priced lending (table 12). For the loan products with the largest numbers of borrowers, the unadjusted mean spreads are lower for all minority groups except blacks than they are for non-Hispanic whites. Typically, Asian borrowers have the lowest mean spreads. Furthermore, adjusting for differences in loan amounts, incomes, and other HMDA factors plus lender either does little to the average spreads or changes them in different directions for different groups. One consistent difference that persists after adjustment is the difference between blacks and non-Hispanic whites. After adjustment, the gap in the mean spreads between these groups for the most common loan products is between 0.10 and 0.15 percentage points.

Differences in Pricing by Sex of Borrower

As with denial rates, there is little evidence of systematic differences in pricing when borrowers are distinguished by sex (table 10). The differences in the unadjusted incidence of higher-priced lending are almost always small across loan products and generally narrow when HMDA factors plus lender are taken into account. Of the nineteen loan product comparisons with evidence of some pricing difference by sex, males have a higher incidence of higher-priced loans in nine cases, while females do in ten cases. Nearly identical patterns are exhibited for mean spreads, and there is no evidence that one sex consistently pays more than the other (table 12).

The Role of Factors Not Included in HMDA

An important limitation of the decompositions reported earlier is that controls were possible only for borrower-related factors included in the HMDA data plus lender when assessing differences in loan pricing among racial, ethnic, or other groups. As noted, many factors relevant to underwriting and pricing are not included in the HMDA data and thus cannot be accounted for in analyses that rely exclusively on the data. To provide some insight into how important controlling for these other factors might be in accounting for pricing differences across borrower groups, we collaborated with Georgetown University's Credit Research Center (CRC) to perform additional analyses using a CRC database. The CRC data consist of both HMDA data and a wide range of other loan-level factors not in the HMDA data that may relate to credit underwriting and pricing.

The CRC data, which were provided by eight lenders that specialize in subprime lending, are equivalent to the 2004 HMDA filings of those lenders (for the loans they originated) plus non-HMDA information on many other characteristics of the loans and borrowers. (48) The non-HMDA information consists of credit history scores (in this instance, FICO scores), loan-to-value ratios for first-lien loans, the appraised value of property, and information on whether the interest rate on the loan was adjustable or fixed, whether underwriting for the loan waived certain certifications by the borrower (that is, whether the loan was a "low documentation" product), whether the loan carried a prepayment penalty, and whether the loan was originated through a broker. Unlike the HMDA data, the CRC data are not disaggregated by individual lender; consequently, the analysis of the data could not control for lender. Together, the eight lenders accounted for about 22 percent of the higher-priced conventional first- or junior-lien home-purchase or refinance loans related to owner-occupied properties reported by HMDA filers for 2004.

The CRC analyzed pricing differences among the combined filings of the eight lenders in a manner identical to that used to prepare table 10--that is, using only data reported in HMDA--and then it extended the adjustment of differences by using the non-HMDA items in its database. (49) The number of loans in the CRC database was sufficient to allow meaningful analysis of conventional home-purchase and refinance loans on one- to four-family, site-built, owner-occupied homes for two minority groups--blacks and Hispanic whites--and for non-Hispanic whites.

For pricing outcomes for conventional first-lien home-purchase loans from the eight lenders, as adjusted for borrower-related factors (that is, using only the HMDA data), the incidence of higher-priced lending differs between black and non-Hispanic white borrowers by 1.5 percentage points and between Hispanic white and non-Hispanic white borrowers by 0.3 percentage point (table 13). When additional factors available only in the CRC data are taken into account, the differences between black and non-Hispanic white borrowers and between Hispanic white and non-Hispanic white borrowers falls about one-third. For refinance loans, the 1.2 percentage point gap between black and non-Hispanic white borrowers that remains after controlling for HMDA data items is removed when additional factors in the CRC database are controlled for. With respect to mean APR spreads for these loan products, gross differences across racial or ethnic groups (that is, differences that exist in the raw, or unadjusted, data) are de minimus, a fact little changed by adding more controls from either the HMDA or the CRC data.

These results suggest that an analysis employing comprehensive information on specific loan products (for example, different types of adjustable-rate loans) from specific lenders--information unavailable in the HMDA or CRC data--would be required to draw firm conclusions about racial or ethnic differences in pricing.

USING THE EXPANDED HMDA DATA AS A SCREENING TOOL FOR FAIR LENDING ENFORCEMENT

Ensuring compliance with the nation's fair lending laws is one of the responsibilities of the federal and state agencies that regulate financial institutions. For some time, the Federal Reserve has been using a statistical analysis system that relies on the HMDA data to help assess fair lending compliance by high-volume mortgage lenders. The system identifies which supervised institutions and which loan products and geographic markets show meaningful differences in the denial rates of loan applications by the race, ethnicity, or sex of the borrower and thus warrant greater supervisory attention. The statistical analysis system measures differences in denial rates by comparing applications for a specific loan product filed by applicants who differ by race, ethnicity, or sex but who are matched on the basis of the limited set of items in the HMDA data. For example, the analysis can focus on denial rates of whites, Hispanic whites, and blacks by comparing the denials of applicants from each of those groups who sought the same loan product for about the same loan amount, are from the same metropolitan area, and have similar incomes, dates of application, and number of applicants in the transaction. The statistical analysis system also provides compliance examiners with a specific list of matched application files to review during the on-site part of an examination.

The expanded data provide opportunities to improve the statistical analysis system in two ways. First, some of the new data items can be used to refine the existing system of analyzing denial rates of loan applications by allowing more precise differentiation among loan products. Lien status and manufactured-home designation are prominent examples because both typically have significant roles in loan underwriting. Second, the new loan-pricing information provides opportunities to expand the statistical analysis beyond the disposition of applications to differences in loan pricing. The data can be reviewed for differences across groups in the incidence of higher-priced lending and in average spreads paid by borrowers with loans priced above the thresholds. The pricing data can also be reviewed for broader patterns that may indicate fair lending issues. For example, an institution's overall lending activity can be reviewed to identify geographic variations in pricing that may be associated with neighborhood racial or ethnic population characteristics. Each of these approaches will improve the fair lending analyses conducted by examiners.

Screening Using the 2004 HMDA Data

As of this writing, the Federal Reserve has modified its statistical analysis system for fair lending examinations to incorporate the new information available in the expanded HMDA data. To examine the potential utility of the enhanced system, we used a streamlined version of the system to conduct a review of the lending activity of the 8,853 institutions reporting 2004 HMDA data. The approach here and in the earlier sections of this article are related, but unlike the earlier sections, which involved an analysis of aggregate patterns that included a control for lender, this exercise uses the data to identify patterns in the lending of individual institutions.

The streamlined analysis starts by evaluating the statistical significance of differences across racial or ethnic lines in the unadjusted (or gross) incidence of denial rates, incidence of higher-priced lending, and average spreads paid by those with higher-priced loans for each lender separately. This procedure produces a series of lender-product combinations. For each lender-product combination, further analysis matches each minority applicant (or borrower) with nonminority applicants (or borrowers) on the basis of a variety of factors available in the HMDA data, including loan product, borrower income and loan amount, geographic market (for example, specific MSA), and number of applicants (one or more than one).

Adjusted differences are computed by comparing the denial rates, incidence of higher-priced lending, and average APR spreads of minorities with those of the nonminorities matched to them. This procedure is designed to remove the effects of these other factors from the calculations of the differences. The adjusted differences are an estimate of the expected differences in outcomes if a minority and a nonminority with the same income, loan amount, and number of applicants applied for the same loan product, at the same institution, in the same market.

The streamlined analysis used for purposes of this article relies on publicly available data as well as data filed under HMDA that are not subject to public disclosure (the dates an application was filed and acted on). A picture of the outcomes of the statistical analysis system with regard to the analysis of denial rates or pricing can be conveyed by reviewing the number of loan products for which the system indicates a statistically significant difference between blacks and Hispanics (as a group) and non-Hispanic whites in denial rates, in the incidence of higher-priced lending, or in the mean differences in pricing for those with higher-priced loans. (50) The focus of the streamlined analysis is on the eight owner-occupied product areas that account for the vast majority of owner-occupied loans in the 2004 HMDA data. (51)

In total, there are 13,260 lender-product combinations for the 8,853 HMDA reporting institutions that have at least one black or Hispanic borrower who can be matched (for comparison) to at least one non-Hispanic white borrower. Of the 13,260 lender-product combinations, 2,418 have at least fifty black or Hispanic borrowers and at least fifty non-Hispanic white borrowers, numbers that provide a more meaningful basis for comparison.

For the black and Hispanic group and the non-Hispanic white group, we have calculated, for the 13,260 lender-product combinations, the distribution across the categories of the statistical significance of three indicators--the difference in denial rates (table 14), in the incidence of higher-priced lending (table 15), and in the mean spreads for loans above the threshold (table 16). We differentiate between situations in which the black and Hispanic group has an indicator (denial rate, incidence of higher-priced loans, or mean spread) that is greater than that for the non-Hispanic white group and situations in which an indicator is lower for the black and Hispanic group than for the non-Hispanic white group. Differences are presented in two ways: (1) as the distribution of the statistical significance of the unadjusted rate or incidence and (2) for each unadjusted category, as the distribution of the statistical significance that remains after the statistical analysis system has been applied. The categories of statistical significance are, from highest to lowest level of significance, 1 percent, 5 percent, 10 percent, and not statistically significant.

Denial Rates

Of the 13,260 lender-product combinations, 3,075, or 23 percent, are those in which the minority group has an unadjusted denial rate that is different from that of the non-Hispanic white group by a statistically significant amount. In almost all of these cases, the black and Hispanic denial rate is higher, although the reverse holds in 4 percent of the cases. Eleven percent of the lender-product combinations show a statistically significant difference in denial rates after the matching procedure is employed (6 percent at the 1 percent significance level), and only 2 percent of the 11 percent show a lower denial rate for blacks and Hispanics.

Incidence of Higher-Priced Loans

Of the 13,260 lender-product combinations, 1,148, or 9 percent, have a statistically significant difference between the minority and nonminority groups in the unadjusted incidence of higher-priced lending. Most of the significant differences show a higher incidence for the black and Hispanic group, although about 8 percent show a lower incidence. Employing the matching process to control for differences in income, loan amount, and other HMDA factors reduces the number of statistically significant differences by more than one-half. Of the lender-product combinations that are statistically significant at the 1 percent level when unadjusted differences are evaluated, fewer than one-half (2 percent of the total number of lender-product combinations) are statistically significant at the 1 percent level after adjustment. A similar reduction occurs in the number of lenders with at least one loan product with a statistically significant minority-nonminority difference--the matching procedure reduces the number of lenders almost 50 percent (data not shown in tables).

The effect of employing the statistical procedures is particularly dramatic for those products for which the black and Hispanic group shows an unadjusted incidence that is lower by a statistically significant amount. Only 7 percent of such cases remain statistically significant after the matching procedures are employed--and none are at the 1 percent significance level.

The reduction in the number of statistically significant differences results primarily from a reduction in the magnitude of the differences. For example, on average, the minority-nonminority difference in the incidence of higher-priced loans for those lender-product combinations in which the unadjusted difference is positive and statistically significant falls 4.5 percent, or about one-fourth, after matching.

Mean Pricing Spreads

The matching procedure yields a similar reduction in the number of lender-product combinations showing a statistically significant difference in the mean spreads for blacks and Hispanics versus those for non-Hispanic whites. Of the 13,260 lender-product combinations, 5 percent show a statistically significant minority-nonminority difference in the mean spread; about one-third of these show a lower spread for blacks and Hispanics. The number of statistically significant differences for lender-product combinations is reduced 60 percent after the matching procedure is employed, and about one-sixth of those show a lower spread for blacks and Hispanics. The number of lenders with a statistically significant difference in mean spread for at least one product also falls about 60 percent when adjustments are made.

Overall Patterns

A high degree of overlap exists among lenders with statistically significant adjusted differences between the minority and nonminority groups in denial rates, the incidence of higher-priced lending, and mean spreads. For example, 60 percent of the lenders with a difference in the adjusted mean spread that is statistically significant at the 1 percent level also have a statistically significant difference in the incidence of higher-priced lending for at least one product. Almost 80 percent of the lenders with an adjusted difference in the incidence of higher-priced lending that is statistically significant at the 1 percent level also have a statistically significant difference in the denial rates for at least one product. However, the presence of statistically significant differences in multiple outcome measures does not necessarily imply a more serious fair lending concern; it may simply reflect differences in the distribution of credit characteristics of the minority and nonminority populations served by the lender.

We emphasize that the Federal Reserve's statistical analysis system is only a screening tool. The HMDA data alone, no matter how much they are manipulated, cannot be used to conclude whether a particular applicant was treated adversely on the basis of a prohibited factor regarding either the disposition of the application or the pricing of the loan. The data reveal little about an individual's financial circumstances and nothing about the condition or value of the property offered as collateral. Furthermore, the data reveal nothing about the underwriting standards used by a lender to assess the creditworthiness of an individual or to set loan price. Moreover, the data do not reveal how a lender's credit decisions relate to its overall business strategy. For example, the data do not account for the possibility that an institution's outreach efforts may attract a larger proportion of applicants with weaker credit profiles than do other institutions. Consequently, the data do not provide a final basis on which to draw conclusions regarding either the existence or the absence of fair lending violations.

The Use of Screening in the Enforcement Process

As implemented in the Federal Reserve's bank supervisory process, the statistical analysis system is used as a screening procedure to identify those institutions and their specific products that warrant closer review for fair lending concerns. Examiners familiar with the procedures and products of a given institution conduct the analysis and tailor it to the specific circumstances relevant to the institution in ways that reflect the institution's product offerings, its compliance risk-management systems, and the Federal Reserve's overall supervisory experience with the institution. Thus, the system used as a screening tool to analyze fair lending compliance for a particular institution is more complex than the streamlined analysis used in this article.

If an institution is targeted for more-intensive review, follow-up procedures can take one or more forms, including soliciting more information from the institution regarding its lending and underwriting procedures; gathering additional loan-level data, such as credit scores and loan-to-value ratios; performing detailed reviews of loan-file data; and conducting interviews with current or past bank personnel or borrowers. The follow-up can be integrated into the normal consumer examination cycle or can become a special review of fair lending compliance. The Federal Reserve has already applied the expanded statistical management system that includes the 2004 HMDA data to many of the institutions it supervises and has contacted those that exhibited relatively large pricing differences by race, ethnicity, or sex to learn more about their lending practices and to improve compliance oversight.

In addition, a review of the 2004 data by other agencies is under way. The Federal Reserve is sharing the screening procedures with other agencies so that, if they wish, they may integrate them into their supervisory programs. It is also responding to agency requests for additional, more detailed analysis of individual institutions that may be of concern to the agencies.

Follow-Up Procedures When Unexplained Differences Are Found

Experience with fair lending reviews indicates that widely used and largely noncontroversial, objective underwriting factors, such as credit scores and loan-to-value ratios, can often account for some or perhaps all of the pricing differences by race, ethnicity, and sex that are not explained by the HMDA data. Thus, generally the first step in a compliance examination in which pricing differences are at issue is to gather additional information on the factors that are used in underwriting and pricing but that are not included in the HMDA data. These factors can vary from institution to institution and from product to product. This step is generally taken after consultation with the lending institution and after a review of its underwriting policies and procedures. If accounting for these objective factors explains all racial or ethnic pricing differences remaining after controlling for HMDA-related factors, the examination will typically be closed unless other pricing issues remain. If, however, not all differences can be explained by controlling for these factors, further steps will typically be taken.

Ordinarily examiners will ask the institution to provide any evidence it has for the remaining differences. Supervisory experience shows that these differences frequently arise in institutions that employ discretionary pricing programs. Lenders who indicate to examiners that pricing differences are the result of either (1) the use of discretionary pricing to adjust for varied market factors, such as a competitor's pricing or individualized credit-risk or pricing-related factors not encompassed in a rate sheet, or (2) differences in the extent to which borrowers negotiate for the best available pricing on their loans should expect to be asked to provide credible evidence to support such explanations. Such evidence could include contemporaneous documentation from loan files, credible statements by participating loan personnel, and nondiscriminatory underwriting policies and procedures, such as internal audits of discretionary pricing patterns or training that focuses on a loan officer's responsibility to avoid setting pricing overages according to the perceived susceptibility of a given group to such pricing.

A particularly complex arena in which evidence of pricing differences may arise involves institutions having multiple loan origination channels, particularly channels that involve indirect loans (for example, those supplied by brokers or wholesalers). These channels can include multiple origination sources within a particular institution, including the institution's own loan officers along with those of its affiliates or subsidiaries, as well as indirect lending in which an institution originates loans referred to it by brokers or loan correspondents or purchases loans or pools of loans from unaffiliated, third-party originators. Such channels can encompass significant variety in the nature of the relationships between a given lender and the affiliates, brokers, or third-party originators that deliver loans to the lender. Indeed, such relationships may range across a spectrum from a prime-rate lender that also operates a subprime business in the same geographic market through a direct subsidiary to a regional bank in one section of the country that makes fully "arm's length" purchases of closed loans from an unaffiliated mortgage company operating solely in a different area.

From a fair lending perspective, a lender whose different channels of lending serve either borrowers or geographic areas that differ by race, ethnicity, or other prohibited characteristic is likely to be further reviewed. That will certainly be the case if these different channels produce loan pricing that also differs by race, ethnicity, or other prohibited characteristic. (52) A full review of fair lending compliance in multiple-channel situations will turn on complex factual analyses that are beyond the scope of this article. It is sufficient to note here that such analyses will cover the type and degree of pricing differences, the nature of the various channels, the lender's legal and business relationships with other entities (for example, affiliates, brokers, correspondents, or wholesalers), the lender's business or economic basis for operating through those channels, and the lender's explanation for the pricing differences.

The process of examining an institution for which pricing differences based on race, ethnicity, or sex are statistically significant and for which purely objective pricing factors, such as credit scores or loan-to-value ratios, cannot explain the differences, will include a review of loan files; discussions with management or loan personnel about possible reasons for the differences; a review of evidence put forth to support their explanations; interviews with customers, where necessary, regarding their experiences with the lender; and a careful vetting of an institution's policies and procedures and actual practices. If, after conducting an examination, there is no credible nondiscriminatory explanation for such differences, examiners will consider what supervisory action will be appropriate to address the issue. Moreover, a lender that cannot account for differences in pricing across groups may also be exposed to private rights of action under applicable fair lending laws and may face adverse effects on its reputation.

HOEPA Enforcement

For the agencies that evaluate compliance with HOEPA, the expanded HMDA data provide the first opportunity to readily identify which lenders extend home loans subject to that law and to measure the extent of their involvement in such lending. The new information also provides examiners with the data needed to efficiently select samples of loan files for review. The data can also be used to examine patterns of HOEPA-related lending across borrowers and neighborhoods, arrayed by their racial and ethnic profiles. Such analysis may reveal possible fair lending issues and may indicate communities where credit counseling activities could be targeted.

The Federal Reserve's statistical analysis system has been augmented to include several screens to aid HOEPA-related enforcement. These screens include the identification of HOEPA loans that are potentially unaffordable given a comparison of the applicant's income and the estimated monthly loan payments, the identification of loans with APR spreads that would appear to have triggered HOEPA coverage but were not reported as such, and the calculation of differences across racial and ethnic groups in the incidence of HOEPA lending. (53)

SUMMARY AND CONCLUSIONS

In 2002 the Federal Reserve Board amended its Regulation C to expand the types of information that lenders covered by HMDA must disclose to the public about their home-lending activities. The amendments are intended to improve the quality, consistency, and utility of the reported data and to keep the regulation in step with recent developments in home-loan markets. Data reported for 2004 are the first to reflect the changes in the reporting rules.

As anticipated, the expanded data provide new opportunities to assess home-lending activity. Newly available information on lien status and on whether a loan is for a site-built or manufactured home, as well as more uniformity in the information on home-improvement and refinancing loans, allows analyses that are more relevant to the current state of the market. Most prominently, the new information provides the first publicly available loan-level information on loan pricing in the higher-priced segment of the home-loan market, a segment that was virtually nonexistent a decade or so ago but is now an important part of the overall home-loan market.

This article presents an analysis of the 2004 HMDA data. The analysis is conducted with the national HMDA database and is designed to provide an understanding of the overall patterns in the data rather than patterns that pertain to any individual market or lender. Much of the presentation focuses on the new items in the data. On balance, the analysis suggests that the information on lien status, manufactured homes, requests for pre-approval, and refined product definitions provides a much improved basis for describing lending activity and the disposition of applications for credit. Much of the initial public review of the data will, however, undoubtedly focus on loan pricing and particularly on the incidence of higher-priced lending and the comparison of prices paid by borrowers grouped by race, ethnicity, and sex.

The most likely initial public focus will be on the incidence of higher-priced lending among minorities (particularly blacks) and among non-Hispanic whites. In the raw data, the differences between these two groups in the incidence of higher-priced lending are generally more than 20 percentage points for various loan products. Our analysis shows, however, that more than two-thirds of the aggregate difference in the incidence of higher-priced lending between black and non-Hispanic white borrowers can be explained by differences in the groups' distributions of income, loan amounts, other borrower-related characteristics included in the HMDA data, and the choice of lender. Further, analysis at the level of individual lenders suggests that about 2 percent of the 8,853 lenders covered by HMDA exhibited a statistically significant difference in the incidence of higher-priced loans between black and Hispanic borrowers, on the one hand, and non-Hispanic white borrowers, on the other, after accounting for factors included in the HMDA data.

Thus, we see a sizable narrowing, at both the aggregate and institution levels, in the unexplained differences in the the incidence of higher-priced lending between minority and nonminority groups. This narrowing suggests that controlling for credit-related factors not found in the HMDA data, such as credit history scores and loan-to-value ratios, might further reduce unexplained racial or ethnic differences. Whether controlling for such additional factors will completely account for all remaining differences is unclear. In that regard, our collaborative study with the Credit Research Center on the lending activities of eight large subprime lenders, reported here, suggests that controlling for credit-related factors not included in the data can make a difference. Our analysis demonstrated that for some products the racial or ethnic differences were fully accounted for, whereas for other products, unexplained differences remained. Clearly, reaching convincing conclusions about whether institutions treat individuals differently on a prohibited basis requires institution-specific analysis.

Hence, our analysis strongly indicates that the raw data alone can lead to inaccurate conclusions, which in turn may be unfair to particular institutions and may lead to unnecessary restrictions on the availability of loans to less-creditworthy applicants. Risk-based pricing has greatly expanded the availability of home loans to borrowers who, because of weaknesses in their credit profiles, had previously been unable to qualify. It would be unfortunate if unwarranted accusations of illegal bias, stemming from improperly analyzed pricing differences, discouraged lenders from participating in this segment of the market.

The primary responsibility for ensuring compliance with fair lending laws falls on lenders. HMDA data may help lenders analyze and monitor their lending patterns. In addition, the regulatory agencies use the data for screening purposes to identify individual lenders that warrant heightened scrutiny regarding their loan-pricing activities. Where warranted, such reviews include gaining a fuller understanding of the institution's loan-pricing practices, analyzing loan-level data, and interviewing appropriate personnel to determine whether pricing differences identified through the HMDA screening process are explained by controlling for these additional data or by other objective factors.

To improve its fair lending examination capabilities, the Federal Reserve has modified its statistical analysis tool to use the new data to screen institutions for significant differences in lending outcomes across borrowers grouped by race, ethnicity, or sex. The Federal Reserve has already applied this expanded statistical management system to many of the institutions it supervises. It has also contacted those institutions that exhibit relatively large pricing differences to learn more about their lending practices and to improve its compliance oversight. Moreover, a review of the 2004 data by other agencies is under way, and the Federal Reserve is sharing the screening procedures with other agencies to facilitate their efforts.

Institution-specific evaluations, which are not possible with the HMDA data alone, are essential to determining whether loan-pricing differences in fact reflect discriminatory treatment of minority groups. However, the aggregate data can nonetheless provide valuable, broader insights into the experience of such groups in the home-loan market. For example, black and Hispanic borrowers taken together are much more likely than non-Hispanic white borrowers to obtain credit from institutions that report a higher incidence of higher-priced loans. On the one hand, this pattern may be benign and reflect a sorting of individuals into different market segments by their credit characteristics. On the other hand, it may be symptomatic of a more serious issue. Lenders that report a lower incidence of higher-priced products may be either less willing or less able to serve minority neighborhoods. More troubling, these patterns may stem, at least in part, from borrowers being steered to lenders or to loans that offer higher prices than the credit characteristics of these borrowers warrant. Reaching accurate determinations among these alternative possible outcomes is one goal of the supervision system. Moreover, we hope that future research using the new HMDA data will provide insights that will facilitate this process.

APPENDIX: ENHANCED DATA ON PRIVATE MORTGAGE INSURANCE

Historically, mortgage lenders have required prospective borrowers to make a down payment of at least 20 percent of a home's value before they will extend a home-purchase loan. Such down payments are required because experience has shown that homeowners with little equity are substantially more likely to default on their mortgage. Private mortgage insurance (PMI) emerged as a response to both creditors' concerns about the elevated credit risk of lending backed by little equity in a home and the difficulties that some consumers encounter in accumulating sufficient savings to meet required down-payment and closing costs.

PMI protects a lender if a borrower defaults on a loan: It reduces a lender's credit risk by insuring against losses associated with default up to a contractually established percentage of the claim amount. The costs of the insurance are typically paid by the borrower through a somewhat higher interest rate on the loan.

In 1993 the Mortgage Insurance Companies of America (MICA) asked the Federal Financial Institutions Examination Council (FFIEC) to process data from PMI companies on applications for mortgage insurance and to produce disclosure statements for the public based on the data. (54) The PMI data largely mirror the types of information submitted by lenders covered by HMDA. However, because the PMI companies do not receive all the information about a prospective loan from the lenders seeking insurance coverage, some HMDA items are not included in the PMI data. In particular, loan-pricing information, requests for pre-approval, and HOEPA status are unavailable in the PMI data.

For 2004 the seven PMI companies that were writing private mortgage insurance submitted data to the FFIEC through MICA. In total, these companies acted on nearly 2 million applications for insurance: 1.3 million to insure mortgages for purchasing homes and about 650,000 to insure mortgages for refinancing existing mortgages. PMI companies approved more than 90 percent of the applications they received. Approval rates are high because lenders are familiar with the underwriting standards used by PMI companies and generally submit applications for insurance only if the applications are likely to be approved.

NOTE: Gregory Elliehausen, of the Credit Research Center of Georgetown University, prepared a special analysis for this article.

(1.) See Home Mortgage Disclosure Act (12 U.S.C. 2801), Regulation C (12 C.F.R. pt. 203), and the staff commentary accompanying Regulation C (12 C.F.R. pt. 203, Supp. I). The Board's revisions to Regulation C that are the focus of this article were issued in 2002. See the following issues of Federal Register (2002), vol. 67: February 15, p. 7222; May 8, p. 30771; and June 27, p. 43218; and Federal Register (2003), vol. 68 (May 28), p. 31589.

(2.) The Federal Reserve has adopted no specific definition of predatory lending, but the term is often considered to encompass a variety of lending practices involving fraud, deception, or unfairness. Some predatory lending practices are illegal; others, although legal, are still considered abusive in certain circumstances. Some of the practices considered questionable or in some cases illegal include (1) making loans that are based on the asset value of the collateral but are unaffordable given the consumer's ability to pay the obligation; (2) inducing repeated refinancing accompanied by high fees that provide no material benefit to the consumer (sometimes referred to as "loan flipping"); (3) inducing the consumer, through deception or fraud, to accept loan add-ons, such as single-premium credit insurance; (4) "steering" borrowers qualified for lower-rate loans into higher-priced loans; and (5) purposely overestimating the value of the collateral to overstate available equity or induce a consumer to pay an inflated price for a home.

(3.) For additional information, see Board of Governors of the Federal Reserve System, Department of Housing and Urban Development, Federal Deposit Insurance Corporation, National Credit Union Administration, Office of the Comptroller of the Currency, and Office of Thrift Supervision (2005), "Agencies Announce Answers to Frequently Asked Questions about New HMDA Data," press release, March 31, www.federalreserve.gov/boarddocs/press/all/2005/.

(4.) Although coverage of financial institutions under HMDA is limited to those with offices in metropolitan statistical areas, covered institutions must report on all their home-lending activities whether the properties involved in the loan are located in a metropolitan or nonmetropolitan area.

(5.) For the details of the coverage rules and for additional information about the data collection and reporting requirements, see A Guide to HMDA Reporting: Getting It Right! published annually by the Federal Financial Institutions Examination Council (www.ffiec.gov/ hmda/guide.htm).

(6.) Reporting institutions had been allowed to choose from four scenarios in deciding which refinancings to report. The new rules define a refinancing simply as a secured home loan that satisfies and replaces another secured home loan by the same borrower. The reporting of home equity lines of credit (extended for any purpose) is voluntary.

(7.) One technique used to identify loans backed by junior liens was to assume that all loans below a given amount were junior-lien loans. This approach is flawed because some homes, including many manufactured homes, have low prices and purchasers of these properties often need only a small loan. Similarly, some borrowers make substantial down payments when they buy a home, and in such circumstances, the amount of the first-lien loan may be small.

(8.) During some fair lending reviews, lenders have provided examiners with information that has allowed the separation of first and junior liens.

(9.) A manufactured home is a single-family house constructed under a federal building code administered by the U.S. Department of Housing and Urban Development (HUD). The Federal Manufactured Home Construction and Safety Standards (commonly known as the HUD code) took effect on June 15, 1976. The result of federal regulation was to more clearly define mobile homes as buildings rather than vehicles--although the HUD code imposes standards to make sure the units can be transported by truck to the placement site. The Housing Act of 1980 officially adopted this change, mandating that, for homes built under the HUD code, the term manufactured housing (factory-built homes) replace the term mobile homes in all federal law and literature.

(10.) In the past, loans on manufactured homes were identified using information about the lender's main line of business. This proxy is helpful only for lenders focused mainly on manufactured-home lending. A large number of other lenders also extend such credit, but because manufactured-home lending does not constitute their main line of business, determining which of their loans involve manufactured homes has been impossible. See www.huduser.org/datasets/ manu.html.

(11.) One method of estimating the annual volume of subprime loans is based on a list of subprime lenders that was developed by HUD and has been released each year since 1993. The number of loans in the HMDA data originated by lenders on the HUD list has been used as an estimate of subprime lending volume. The list has also been used to support other analyses of subprime lending activity. Of the 224 lenders on the list for 2003, 191 appear under the same name and identification number as reported in the HMDA filings for 2004.

(12.) Estimates pertain to home loans backed by one- to four-family homes; see Inside Mortgage Finance Publications (2005), Mortgage Market Statistical Annual 2005, 2 vols. (Bethesda, Md.: IMFP).

(13.) Regulation Z, 226.3(a). The Federal Reserve Board's Regulation Z (12 C.F.R. pt. 226) implements the Truth in Lending Act.

(14.) For such calculation, the rule directs creditors to use the fifteenth day of a given month for any loan on which the interest rate was set on or after that day through the fourteenth day of the next month. The relevant date is when the interest rate on the loan was determined, which is often, but not always, set pursuant to a lock-in agreement between the borrower and the lender. The APR used in the calculations is the one determined and disclosed to the consumer under section 226.6 or section 226.18 of Regulation Z. To ease reporting burdens and to help ensure high-quality data, the FFIEC makes available a "Rate Spread Calculator" that lenders can use to determine whether they must report the spread on a given loan and, if so, what the spread is; see www.ffiec.gov/ratespread/default.aspx.

(15.) See the U.S. Census Bureau's 1998 and 1999 Annual Housing Surveys, the Federal Reserve Board's 2001 Survey of Consumer Finances, the Federal Housing Finance Board's 1999 Mortgage Interest Rate Survey, and data on subprime lending from the Credit Research Center of Georgetown University.

(16.) See Board of Governors of the Federal Reserve System (2002), "Regulatory Analysis of Proposed Amendments to Regulation C," staff memorandum, Division of Research and Statistics, January 15.

(17.) Unlike lines of credit, closed-end loans are amortizing--they require fixed monthly payments against both principal and interest-- and are thus scheduled to close at the end of a given term to maturity, when the balance will reach zero. A balloon payment might be involved if the amortization schedule leaves a relatively large balance owed at the end of the loan's term. In contrast, a home equity line of credit (HELOC) is a revolving account that permits borrowing from time to time at the account holder's discretion, up to the amount of the credit line. Under a HELOC, a consumer may repeatedly pay the balance down to zero and then redraw against the line.

(18.) For further details, see section 226.32 of Regulation Z.

(19.) The only pre-approval programs covered by HMDA are those in which the decision to grant or deny the request is based on a comprehensive credit underwriting process in which a lender collects and reviews the information it typically considers in making credit decisions in a traditional application (that is, an application for a specific property). For a pre-approval program to be covered, the lender must issue binding written commitments (subject only to very limited conditions) for consumers whose requests were granted. Because requests for pre-approval typically do not identify a specific home for purchase, the HMDA data do not show the property location for pre-approvals that do not ultimately result in an application for credit related to a specific property.

(20.) See Office of Management and Budget (1997), "Revisions to the Standards for the Classification of Federal Data on Race and Ethnicity," Federal Register, vol. 62 (October 30), pp. 58782-90.

(21.) The OMB changed MSA boundaries to encompass 288 previously rural counties and to exclude 46 counties previously in MSAs.

(22.) Because of the reporting rules, asset size is generally a meaningful measure of size only for depository institutions. Assets are measured as of December 31, 2004.

(23.) Among the loan originations in the 2004 data, about 470,000 were reported as being initiated through a pre-approval program. This figure likely understates the number of originations that began in pre-approval programs because the transition rules did not require the reporting of this item on applications taken before January 1, 2004.

(24.) The HMDA data do not include a code indicating whether the junior-lien home-purchase loan reported in the data is associated with any particular first-lien loan. The junior-lien loan may be in the reported data, but the first-lien loan may not be. This distinction can arise if, for example, the lender extending the first-lien loan is not covered by HMDA. We estimate that about 62 percent of the junior-lien home loans used to purchase owner-occupied homes in the reported data are likely associated with a reported first-lien home loan of the same lender--these "paired loans" were extended by the same lender in the same census tract to borrowers of the same sex, race, and ethnicity, and the actions taken on each loan in the loan pair were within a couple of days of each other.

(25.) An investment property is a nonowner-occupied dwelling that is intended to be continuously rented. Some nonowner-occupied units--vacation homes and second homes--are for the primary use of the owner and would thus not be considered investment properties. The HMDA data do not, however, distinguish between these two types of nonowner-occupied dwellings.

(26.) The HMDA data tend to undercount the volume of secondary-market sales somewhat. One reason is that, for example, some loans originated in 2004 will be sold to a secondary-market institution in 2005 or later and thus will never be reported as a sale. Another is that, as with other HMDA data, about 20 percent of home loans originated in 2004 were extended by lenders not covered by HMDA.

(27.) GSEs are privately owned institutions that blend the characteristics of public and private institutions. They receive certain benefits from government sponsorship in exchange for their advancement of certain public policy goals such as homeownership among lower-income households and in targeted communities.

(28.) See www.huduser.org/datasets/manu.html.

(29.) Unless otherwise indicated, information in this section is derived from the following sources: U.S. Bureau of the Census (for HUD), American Housing Survey (formerly the Annual Housing Survey) and the Residential Finance Survey www.huduser.org/ datasets/pdrdatas.html; Manufactured Housing Institute 2004, www.manufacturedhousing.org; and U.S. Census Bureau 2000 census, www.census.gov/main/www/cen2000.html. Information on the default experience regarding loans secured by manufactured homes and on the credit history scores of individuals were derived from the June 30, 2003, credit records of a nationally representative sample of approximately 300,000 individuals (with all personal identifying information removed); the sample was obtained by the Federal Reserve Board from one of the three national credit-reporting agencies. See Robert B. Avery, Paul S. Calem, and Glenn B. Canner (2004), "Credit Report Accuracy and Access to Credit," Federal Reserve Bulletin, vol. 90 (Summer), pp. 297-322.

(30.) See American Bankers Association, Consumer Credit Delinquency Bulletin, www.aba.com.

(31.) To facilitate this discussion, we have adjusted the credit history scores assigned to the individuals in the Federal Reserve sample of 300,000 credit records (see text note 30) to match the distribution of the more-familiar FICO credit history scores developed by Fair Isaac Corporation, for which information is publicly available. See www.myfico.com/myfico/CreditCentral/ScoringWorks.asp.

(32.) In recent years, the manufactured-home lending industry has been adversely affected by the excessive production of units in the late 1990s and the reliance on the relaxed credit underwriting that accompanied the sales of these units. See Neil J. Morse (2004), "Manufacturing the Dream," Mortgage Banking (August), pp. 50-56.

(33.) As noted, the transition rules regarding the reporting of data pose difficulties for evaluating the 2004 HMDA data for manufactured homes. Consequently, applications governed by the transition rules are excluded from tables 5 and 6. Despite the reporting exceptions created by the rules, some lenders chose to report information on manufactured-home status for applications submitted before January 1, 2004. However, it is not clear whether these lenders identified all, or only some, of the pre-2004 applications for loans on manufactured homes, and so we exclude these additional data from the analysis.

The 2004 data include information on applications or loans related to manufactured homes from an additional 400 or so lenders--about 4,800 in all--which indicates that some institutions chose to identify manufactured homes on applications taken during the transition period (before January 1, 2004).

(34.) The income category of a purchaser is relative to the median family income of the area (MSA or statewide non-MSA) in which the property being purchased is located, and the income category of a census tract is the median family income of the tract relative to that of the area (MSA or statewide non-MSA) in which the tract is located: "Low" is less than 50 percent of the median; "moderate" is 50 percent to 79 percent (in this article, "lower income" encompasses the low and moderate categories); "middle" is 80 percent to 119 percent; and "higher" is 120 percent or more.

(35.) For loans with two or more applicants, HMDA-covered lenders report data on only two. Income for two applicants is reported jointly. Although, as of 2004, applicants may choose more than one race as well as one of two ethnicities, applications are placed for the purposes of table 6 and tables 9 through 13 under only one category for race and ethnicity, generally according to the race and ethnicity of the person listed first on the application. However, under race, the application is designated as joint if one applicant reported the single designation of white and the other reported one or more minority races. If the application is not joint but more than one race is reported, the following designations are made: If at least two minority races are reported, the application is designated as two or more minority races; if the first person listed on an application reports two races, and one is white, the application is categorized under the minority race.

(36.) Pre-approvals were subject to the transition rules; consequently, these numbers and the others in this section exclude applications submitted before 2004.

(37.) Applications in which the lender reported that the race, ethnicity, and sex of the applicant or co-applicant were "not applicable" were assumed to have been made by businesses (including trusts) rather than by individuals.

(38.) Reasons for denial are not provided for requests for pre-approvals that are denied.

(39.) Analysis of the data is further complicated because loans not subject to the Federal Reserve's Regulation Z--that is, business loans--are reported with the same code as loans with spreads below the threshold. Some, perhaps most, of these loans are identifiable, however, because, as explained in text note 37, an application can be identified as being from a trust or other organization rather than from a person.

(40.) See www.freddiemac.com.

(41.) Reporting institutions are required to report all their lending in MSAs as well as in the nonmetropolitan portions of states. However, because institutions operating exclusively in nonmetropolitan areas are not covered by HMDA, loans in nonmetropolitan areas are underrepresented in the data. For this reason, the geographic analysis here is focused on MSAs.

(42.) In November 2004, the OMB redesignated this Texas MSA as McAllen-Edinburg-Mission (see www.whitehouse.gov/omb/bulletins/ fy05/b05-02_attachment.pdf).

(43.) For example, if a homeowner takes out a HOEPA-covered loan to pay off outstanding credit card debt or some other type of consumer credit, and the loan does not involve the refinancing of an existing home loan or home improvement, then the loan is not covered by Regulation C and is thus not required to be part of an institution's HMDA reporting.

(44.) The HMDA data also include information on loans purchased by covered institutions during 2004: Among purchased loans, about 2,700 were designated as HOEPA loans.

(45.) The analysis was not conducted for unsecured loans because pricing data were not collected for these loans. Eleven other product areas were not used because they accounted for so few loans that matching was difficult.

(46.) The action date on an application is used to determine the reporting year for HMDA data. The gap between the application date and the action date is generally shorter for denied applications than for originated loans. For example, applications received and acted upon in December (and therefore reported in HMDA for that year) are more likely to be denials than acceptances. Similarly, applications acted upon in January but received in the previous year are more likely to be acceptances. In analyzing denial rates for 2004, excluding applications covered by the transition period (that is, applications received before 2004) is therefore likely to disproportionately exclude acceptances. This can be seen in the memo item in table 7, where the denial rates are lower for applications filed during the transition period. For this reason, transition-period applications are included in the denial-rate analysis in tables 9 and 14.

(47.) Larger depository institutions covered by the CRA (generally those with assets of $250 million or more) are required to identify the census tracts in their CRA assessment areas as of the end of each calendar year. That information was used to determine which loans in the HMDA data were for properties within the lenders' CRA assessment areas. When lenders were part of a bank holding company, the combined assessment areas of all banks in the holding company were used for the analysis. For a definition of "assessment area," see note 4 to box "Reasons for Loan Price Variation."

The identification of specialists in high-priced lending was based solely on the incidence of higher-priced loans for non-Hispanic whites. This restriction prevented the identification from being affected by differences between the pricing outcomes of blacks and Hispanics, on the one hand, and those of non-Hispanic whites, on the other.

(48.) The eight subprime mortgage lenders are subsidiaries of large financial institutions. The 2004 data from these lenders consist of about 626,000 loans that they originated; more than 60 percent of the loans in that group are higher-priced. The data from these subprime lenders have been used in various research initiatives and public policy deliberations.

(49.) Only loans with complete information on all relevent factors were used in the analysis. Loans with missing information for any factor had a somewhat lower incidence of higher-priced lending than did the loans used in the analysis.

(50.) Black and Hispanic borrowers were selected for this review because these groups generally showed the greatest differences from non-Hispanic whites. The groups were combined to have sufficient numbers for a meaningful statistical comparison. The "Hispanic" category used here includes all borrowers designated as Hispanic regardless of their race. This definition differs from that used in the previous section, which restricted the category of "Hispanics" to white Hispanics.

(51.) The eight products are virtually the same as those portrayed in table 9 with the modifications of (1) adding refinance and home-improvement first liens to the manufactured-housing product area and (2) combining the home-improvement and junior-lien refinancing products and expanding the category to include government-backed loans. Combining products for this exercise was done purely for the purpose of paralleling the way products are grouped in the Federal Reserve's statistical analysis system. The key product areas are identical to those used for tables 7 and 8. The product groupings do not affect the actual matching procedure because minority borrowers are always matched with non-Hispanic white borrowers with exactly the same product as defined using all the information available in HMDA. Transition-period applications were used in the denial-rate comparison but not in the comparisons for pricing.

(52.) The expanded HMDA data can be used to roughly differentiate the pricing of an institution's retail lending operations from the pricing of loans obtained from other channels by comparing the locations of borrowers with the locations of an institution's assessment areas. Loans outside an institution's assessment areas are more likely to have been initiated by third-party brokers or through other indirect channels.

(53.) The estimated monthly loan payment is derived from HMDA data using the reported loan amount and an estimated APR that assumes a fixed-rate thirty-year loan.

(54.) Founded in 1973, MICA is the trade association for the PMI industry. The FFIEC prepares disclosure statements for each of the PMI companies. The statements are available at the corporate headquarters of each company and at a central depository in each MSA in which HMDA data are held. The central depository also holds aggregate data for all the PMI companies active in that MSA. In addition, the PMI data are available from the Federal Reserve Board through its HMDA Assistance Line (202-452-2016).

Robert B. Avery and Glenn B. Canner, of the Division of Research and Statistics, and Robert E. Cook, of the Division of Consumer and Community Affairs, prepared this article. Shannon C. Mok and Caitlin G. Coslett provided research assistance. Patricia J. Dykes and Sylvia A. Freeland assisted in preparing the 2004 HMDA data for analysis.

1. Distribution of home lenders covered by HMDA,
by type of institution, 2004

         Type            Number   Percent

Depository institution
  Commercial bank         3,946      44.6
  Savings institution     1,017      11.5
  Credit union            2,030      22.9
  All                     6,993      79.0

Mortgage company
  Independent             1,464      16.5
  Affiliated (1)            396       4.5
  All                     1,860      21.0

All institutions          8,853       100

(1.) Subsidiary of a depository institution or an affiliate
of a bank holding company.

SOURCE: In this and subsequent tables except as noted, Federal
Financial Institutions Examination Council, data reported under the
Home Mortgage Disclosure Act (www.ffiec.gov/hmda).

2. Home loan and reporting activity of home lenders covered
under HMDA, 1990-2004

Number

       Applications received for home loans, and home
           loans purchased from other lenders
                      (millions)

                      Applications

         Home                    Home
Year   purchase   Refinance   improvement   Total (1)

1990     3.27       1.07         1.16          5.51
1991     3.26       2.11         1.18          6.55
1992     3.54       5.24         1.23         10.01
1993     4.52       7.72         1.40         13.64
1994     5.20       3.80         1.69         10.69

1995     5.51       2.70         1.75          9.96
1996     6.33       4.54         2.14         13.01
1997     6.75       5.39         2.16         14.30
1998     7.96      11.42         2.04         21.43
1999     8.43       9.37         2.05         19.85

2000     8.28       6.54         1.99         16.81
2001     7.69      14.29         1.85         23.83
2002     7.40      17.48         1.53         26.41
2003     8.15      24.60         1.51         34.26
2004     9.79      16.10         2.20         28.13

        Applications received
       for home loans, and home
         loans purchased from
       other lenders (millions)

            Loans                             Disclosure
Year      purchased   Total       Reporters   reports (2)

1990        1.15       6.66         9,332       24,041
1991        1.36       7.91         9,358       25,934
1992        1.98      12.00         9,073       28,782
1993        1.80      15.44         9,650       35,976
1994        1.48      12.17         9,858       38,750

1995        1.28      11.24         9,539       36,611
1996        1.82      14.83         9,328       42,946
1997        2.08      16.38         7,925       47,416
1998        3.23      24.65         7,836       57,294
1999        3.01      22.86         7,832       56,966

2000        2.40      19.21         7,713       52,776
2001        3.77      27.59         7,631       53,066
2002        4.83      31.24         7,771       56,506
2003        7.23      41.49         8,121       65,808
2004        5.14      33.27         8,853       72,246

NOTE: Here and in all subsequent tables except tables 3 and 8,
applications exclude requests for pre-approval that were denied by the
lender or were accepted by the lender but not acted upon by the
borrower.

In this article, applications are defined as being for a loan on a
specific property; they are thus distinct from requests for
pre-approval, which are not related to a specific property.

(1.) Applications for multifamily homes are included only in the
"total" column; for 2004 these applications numbered about 62,000.

(2.) A report covers the mortgage lending activity of a lender in a
single metropolitan statistical area in which it had an office during
the year.

3. Distribution of home lenders covered by HMDA, by type of lender
and the number of applications they receive, 2004

                                         Number of applications

                                                 1-99

           Type of lender,
           and subcategory
     (asset size in millions of        Percent of      Percent of
      dollars, or affiliation)           type (1)   subcategory (2)

Depository institution
  Commercial bank
    Less than 250                          80.2            56.7
    250-999                                16.4            24.5
    1,000 or more                           3.4            13.4
    All                                   100              42.8

  Savings institution
    Less than 250                          85.1            40.9
    250-999                                10.7             7.2
    1,000 or more                           4.1             6.5
    All                                   100              23.8

  Credit union
    Less than 250                          96.9            62.9
    250-999                                 3.0             7.6
    1,000 or more                            .1             1.0
    All                                   100              49.1

  All depository institutions
    Less than 250                          86.3            57.1
    250-999                                11.4            17.7
    1,000 or more                           2.3            10.0
    All                                   100              41.9

Mortgage company
  Independent                              78.9            21.5
  Affiliated                               21.1            21.2
  All                                     100              21.4

All lenders                                 ...            37.6

MEMO
All applications, by number reported
  by lender                                 ...              .5

                                         Number of applications

                                               100-249

           Type of lender,
           and subcategory
     (asset size in millions of        Percent of      Percent of
      dollars, or affiliation)          type (1)    subcategory (2)

Depository institution
  Commercial bank
    Less than 250                         71.4            31.6
    250-999                               25.9            24.2
    1,000 or more                          2.7             6.6
    All                                  100              26.8

  Savings institution
    Less than 250                         71.2            35.7
    250-999                               25.3            17.8
    1,000 or more                          3.6             5.8
    All                                  100              24.9

  Credit union
    Less than 250                         84.7            27.5
    250-999                               14.7            18.4
    1,000 or more                           .6             3.0
    All                                  100              24.5

  All depository institutions
    Less than 250                         75.1            30.6
    250-999                               22.7            21.8
    1,000 or more                          2.2             5.9
    All                                  100              25.9

Mortgage company
  Independent                             75.0            10.7
  Affiliated                              13.1            13.1
  All                                     88.1            11.2

All lenders                                ...            22.8

MEMO
All applications, by number reported
  by lender                                ...             1.2

                                         Number of applications

                                                250-999

           Type of lender,
           and subcategory
     (asset size in millions of        Percent of      Percent of
      dollars, or affiliation)          type (1)    subcategory (2)

Depository institution
  Commercial bank
    Less than 250                         31.3            11.0
    250-999                               57.2            42.4
    1,000 or more                         11.5            22.6
    All                                  100              21.2

  Savings institution
    Less than 250                         30.6            20.4
    250-999                               60.8            57.1
    1,000 or more                          8.6            18.8
    All                                  100              33.1

  Credit union
    Less than 250                         35.9             9.3
    250-999                               58.0            58.3
    1,000 or more                          6.0            24.0
    All                                  100              19.6

  All depository institutions
    Less than 250                         32.3            11.5
    250-999                               58.2            48.5
    1,000 or more                          9.5            22.0
    All                                  100              22.5

Mortgage company
  Independent                             83.3            26.2
  Affiliated                              16.7            19.4
  All                                    100              24.8

All lenders                                ...            23.0

MEMO
All applications, by number reported
  by lender                                ...             3.6

                                         Number of applications

                                              1,000-4,999

           Type of lender,
           and subcategory
     (asset size in millions of        Percent of      Percent of
      dollars, or affiliation)          type (1)    subcategory (2)

Depository institution
  Commercial bank
    Less than 250                          6.3              .7
    250-999                               35.7             8.6
    1,000 or more                         58.1            37.3
    All                                  100               6.9

  Savings institution
    Less than 250                         11.0             2.8
    250-999                               48.0            17.0
    1,000 or more                         40.9            33.8
    All                                  100              12.5

  Credit union
    Less than 250                          3.1              .3
    250-999                               47.7            15.4
    1,000 or more                         49.2            63.0
    All                                  100               6.3

  All depository institutions
    Less than 250                          6.6              .8
    250-999                               41.6            11.6
    1,000 or more                         51.8            40.3
    All                                  100               7.5

Mortgage company
  Independent                             81.8            27.3
  Affiliated                              18.2            22.5
  All                                    100              26.2

All lenders                                ...            11.5

MEMO
All applications, by number reported
  by lender                                ...             7.6

                                         Number of applications

                                             5,000 or more

           Type of lender,
           and subcategory
     (asset size in millions of        Percent of      Percent of
      dollars, or affiliation)          type (1)    subcategory (2)

Depository institution
  Commercial bank
    Less than 250                          1.1              .0
    250-999                                4.4              .4
    1,000 or more                         94.4            20.1
    All                                  100               2.3

  Savings institution
    Less than 250                          1.7              .2
    250-999                                5.2              .8
    1,000 or more                         93.1            35.1
    All                                  100               5.7

  Credit union
    Less than 250                           .0              .0
    250-999                               10.0              .3
    1,000 or more                         90.0             9.0
    All                                  100                .5

  All depository institutions
    Less than 250                          1.3              .1
    250-999                                5.1              .4
    1,000 or more                         93.7            21.8
    All                                  100               2.3

Mortgage company
  Independent                             69.2            14.4
  Affiliated                              30.8            23.7
  All                                    100              16.4

All lenders                                ...             5.2

MEMO
All applications, by number reported
  by lender                                ...            87.1

                                         Number of applications

                                                  Any

           Type of lender,
           and subcategory
     (asset size in millions of        Percent of      Percent of
      dollars, or affiliation)          type (1)    subcategory (2)

Depository institution
  Commercial bank
    Less than 250                         60.6             100
    250-999                               28.7             100
    1,000 or more                         10.8             100
    All                                  100               100

  Savings institution
    Less than 250                         49.6             100
    250-999                               35.3             100
    1,000 or more                         15.1             100
    All                                  100               100

  Credit union
    Less than 250                         75.6             100
    250-999                               19.5             100
    1,000 or more                          4.9             100
    All                                  100               100

  All depository institutions
    Less than 250                         63.3             100
    250-999                               27.0             100
    1,000 or more                          9.7             100
    All                                  100               100

Mortgage company
  Independent                             78.7             100
  Affiliated                              21.3             100
  All                                    100               100

All lenders                                ...             100

MEMO
All applications, by number reported
  by lender                                ...             100

                                                MEMO

           Type of lender,
           and subcategory
     (asset size in millions of        Number of     Percent of
      dollars, or affiliation)          lenders    applications

Depository institution
  Commercial bank
    Less than 250                        2,391           1.1
    250-999                              1,131           1.8
    1,000 or more                          424          19.6
    All                                  3,946          22.5

  Savings institution
    Less than 250                          504            .4
    250-999                                359            .9
    1,000 or more                          154          11.9
    All                                  1,017          13.2

  Credit union
    Less than 250                        1,534            .6
    250-999                                396            .9
    1,000 or more                          100           1.1
    All                                  2,030           2.5

  All depository institutions
    Less than 250                        4,429           2.2
    250-999                              1,886           3.5
    1,000 or more                         678           32.6
    All                                  6,993          38.3

Mortgage company
  Independent                            1,464          44.2
  Affiliated                               396          17.6
  All                                    1,860          61.7

All lenders                              8,853         100

MEMO
All applications, by number reported
  by lender                               ...          100

NOTE: See table 1, note 1, and general note to table 2.

(1.) Distribution sums vertically.

(2.) Distribution sums horizontally.

... Not applicable.

4. Distribution of home loan applications and home loans,
by purpose, lien status, and type of loan and by type and
occupancy status of home, 2004

                                Application

                         One- to four-family home

                                Site built

                         Owner occupied

   Loan category                            Nonowner
   (purpose and       Percent               occupied,
   lien status)       of loan    Percent     percent
   and loan type      category   of loan     of loan
(government-backed      (1)      type (2)   category
 or conventional)                              (1)

Home purchase
First lien
  Government backed
    FHA                   8.3       94.2             .1
    VA                    2.3       97.8         *
    FSA/RHS                .4       98.9         *
  Conventional           89.0       94.3           99.9
  Total                 100         94.4          100

MEMO: Number                 6,899,878      1,156,788

Junior lien
  Government backed
    FHA                    .1       96.1         *
    VA                   *          99.4         *
    FSA/RHS              *         100           *
  Conventional           99.8       99.6          100
  Total                 100         99.6          100

MEMO: Number                 1,134,740         83,626

Refinance
First lien
  Government backed
    FHA                   2.5       97.2             .6
    VA                    1.0       98.8             .6
    FSA/RHS              *          96.4         *
  Conventional           96.5       98.1           98.8
  Total                 100         98.1          100

MEMO: Number                13,695,847        975,583

Junior lien
  Government backed
    FHA                    .1       98.2             .1
    VA                   *         100           *
    FSA/RHS              *         100           *
  Conventional           99.9       99.1           99.9
  Total                 100         99.1          100

MEMO: Number                 1,003,316         24,991

Home improvement
First lien
  Government backed
    FHA                    .5       95.2             .1
    VA                     .1       97.8         *
    FSA/RHS              *         100           *
  Conventional           99.4       97.4           99.9
  Total                 100         97.4          100

MEMO: Number                   727,677         58,664

Junior lien
  Government backed
    FHA                    .6       99.7             .9
    VA                   *         100           *
    FSA/RHS              *         100           *
  Conventional           99.4       98.7           99.1
  Total                 100         98.7          100

MEMO: Number                   950,082         17,049

Unsecured
  Government backed
    FHA                    .4       96.0             .1
    VA                   *         100           *
    FSA/RHS              *         100           *
  Conventional           99.6       97.2           99.9
  Total                 100         97.2          100

MEMO: Number                   372,385         21,780

MEMO: Total number          24,783,925      2,338,481

                                Application

                         One- to four-family home

                               Manufactured

                         Owner occupied

   Loan category                            Nonowner
   (purpose and       Percent               occupied,
   lien status)       of loan    Percent     percent
   and loan type      category   of loan     of loan
(government-backed      (1)      type (2)   category
 or conventional)                              (1)

Home purchase
First lien
  Government backed
    FHA                   8.6        5.8             .1
    VA                     .9        2.2         *
    FSA/RHS                .1        1.2         *
  Conventional           90.5        5.7           99.9
  Total                 100          5.6          100

MEMO: Number                   411,500         26,640

Junior lien
  Government backed
    FHA                   1.5        3.9             .2
    VA                   *            .6         *
    FSA/RHS              *          *            *
  Conventional           98.5         .4           99.8
  Total                 100           .4          100

MEMO: Number                     4,126

Refinance
First lien
  Government backed
    FHA                   3.7        2.8             .3
    VA                     .6        1.2             .3
    FSA/RHS              *           3.6         *
  Conventional           95.7        1.9           99.4
  Total                 100.0        1.9          100

MEMO: Number                   266,184         10,020

Junior lien
  Government backed
    FHA                    .2        1.9            1.1
    VA                   *          *            *
    FSA/RHS              *          *            *
  Conventional           99.9         .9           98.9
  Total                 100           .9          100

MEMO: Number                     8,703            267

Home improvement
First lien
  Government backed
    FHA                   1.0        4.9             .4
    VA                     .1        2.2         *
    FSA/RHS              *          *            *
  Conventional           98.9        2.6           99.6
  Total                 100          2.7          100

MEMO: Number                    19,784          1,275

Junior lien
  Government backed
    FHA                    .1         .3            1.1
    VA                   *          *            *
    FSA/RHS              *          *            *
  Conventional           99.9        1.3           99.0
  Total                 100          1.3          100

MEMO: Number                    12,702            191

Unsecured
  Government backed
    FHA                    .6        4.0             .3
    VA                   *          *            *
    FSA/RHS              *          *            *
  Conventional           99.5        2.8           99.7
  Total                 100          2.8          100

MEMO: Number                    10,785            713

MEMO: Total number             733,784         39,556

                                        Application

                          One- to four-family home

                                  Total

                            Owner occupied

   Loan category                            Nonowner      Multifamily
   (purpose and       Percent               occupied,        home,
   lien status)       of loan    Percent     percent        percent
   and loan type      category   of loan     of loan        of loan
(government-backed      (1)      type (2)   category       category
 or conventional)                              (1)            (1)

Home purchase
First lien
  Government backed
    FHA                   8.3      100               .1           .5
    VA                    2.2      100           *             *
    FSA/RHS                .4      100           *             *
  Conventional           89.1      100             99.9         99.5
  Total                 100        100            100          100

MEMO: Number                 7,311,378      1,183,428       28,345

Junior lien
  Government backed
    FHA                    .1      100           *                .1
    VA                   *         100           *             *
    FSA/RHS              *         100           *             *
  Conventional           99.8      100            100           99.9
  Total                 100        100            100          100

MEMO: Number                 1,138,866         84,076          795

Refinance
First lien
  Government backed
    FHA                   2.6      100               .6           .8
    VA                    1.0      100               .6        *
    FSA/RHS              *         100           *             *
  Conventional           96.4      100             98.8         99.2
  Total                 100        100            100          100

MEMO: Number                13,962,031        985,603       27,558

Junior lien
  Government backed
    FHA                    .1      100               .1           .2
    VA                   *         100           *             *
    FSA/RHS              *         100           *             *
  Conventional           99.9      100             99.9         99.8
  Total                 100        100            100          100

MEMO: Number                 1,012,019         25,258          887

Home improvement
First lien
  Government backed
    FHA                    .6      100               .1        *
    VA                     .1      100           *             *
    FSA/RHS              *         100           *             *
  Conventional           99.4      100             99.9        100
  Total                 100        100            100          100

MEMO: Number                   747,461         59,939        2,751

Junior lien
  Government backed
    FHA                    .6      100               .9           .1
    VA                   *         100           *             *
    FSA/RHS              *         100           *      *
  Conventional           99.4      100             99.1         99.9
  Total                 100        100            100          100

MEMO: Number                   962,784         17,240        1,070

Unsecured
  Government backed
    FHA                    .4      100               .1        *
    VA                   *         100           *             *
    FSA/RHS              *         100           *             *
  Conventional           99.6      100             99.9        100
  Total                 100        100            100          100

MEMO: Number                   383,170         22,493          489

MEMO: Total number          25,517,709      2,378,037       61,895

                                  Loans

                         One- to four-family home

                                 Site built

                      Owner occupied

   Loan category                            Nonowner
   (purpose and       Percent               occupied,
   lien status)       of loan    Percent     percent
   and loan type      category   of loan     of loan
(government-backed      (1)      type (2)   category
 or conventional)                              (1)

Home purchase
First lien
  Government backed
    FHA                   8.8       95.1             .1
    VA                    2.6       97.9         *
    FSA/RHS                .5       99.0         *
  Conventional           88.2       97.5           99.9
  Total                 100         97.3          100

MEMO: Number                 4,654,243        811,816

Junior lien
  Government backed
    FHA                    .1       96.8         *
    VA                   *          99.1         *
    FSA/RHS              *         100           *
  Conventional           99.8       99.7          100
  Total                 100         99.7          100

MEMO: Number                   735,361         50,362

Refinance
First lien
  Government backed
    FHA                   3.2       97.3             .5
    VA                    1.5       98.7             .7
    FSA/RHS              *          96.8         *
  Conventional           95.4       98.8           98.8
  Total                 100         98.7          100

MEMO: Number                 6,405,770        578,753

Junior lien
  Government backed
    FHA                  *          98.0         *
    VA                   *         100           *
    FSA/RHS              *         100           *
  Conventional           99.9       99.4          100
  Total                 100         99.4          100

MEMO: Number                   461,649         11,430

Home improvement
First lien
  Government backed
    FHA                    .7       96.4             .1
    VA                     .1       98.5         *
    FSA/RHS              *         100           *
  Conventional           99.3       97.7           99.9
  Total                 100         97.7          100

MEMO: Number                   352,066         35,835

Junior lien
  Government backed
    FHA                   .6        99.6             .6
    VA                   *         100           *
    FSA/RHS              *         100           *
  Conventional           99.4       98.8           99.3
  Total                 100         98.8          100

MEMO: Number                   393,226          7,058

Unsecured
  Government backed
    FHA                    .4       95.4             .1
    VA                   *         100           *
    FSA/RHS              *         100           *
  Conventional           99.6       97.1           99.6
  Total                 100         97.1          100

MEMO: Number                   147,510          6,302

MEMO: Total number          13,149,825      1,501,556

                                   Loans

                         One- to four-family home

                               Manufactured

                         Owner occupied

   Loan category                            Nonowner
   (purpose and       Percent               occupied,
   lien status)       of loan    Percent     percent
   and loan type      category   of loan     of loan
(government-backed      (1)      type (2)   category
 or conventional)                              (1)

Home purchase
First lien
  Government backed
    FHA                  16.4        4.9             .1
    VA                    2.0        2.1         *
    FSA/RHS                .2        1.0         *
  Conventional           81.5        2.5           99.9
  Total                 100          2.7          100

MEMO: Number                   129,150         15,272

Junior lien
  Government backed
    FHA                   1.4        3.2         *
    VA                   *            .9         *
    FSA/RHS              *          *            *
  Conventional           98.6         .3          100
  Total                 100           .3          100

MEMO: Number                     2,510            285

Refinance
First lien
  Government backed
    FHA                   6.7        2.7             .2
    VA                    1.5        1.3             .4
    FSA/RHS              *           3.2         *
  Conventional           91.8        1.2           99.5
  Total                 100          1.3          100

MEMO: Number                    83,946          5,980

Junior lien
  Government backed
    FHA                    .1        2.0             .7
    VA                   *          *            *
    FSA/RHS              *          *            *
  Conventional          99.9          .6           99.3
  Total                100            .6          100

MEMO: Number                     2,814            138

Home improvement
First lien
  Government backed
    FHA                   1.1        3.6             .4
    VA                     .1        1.5         *
    FSA/RHS              *          *            *
  Conventional           98.9        2.3           99.6
  Total                 100          2.3          100

MEMO: Number                     8,249            787

Junior lien
  Government backed
    FHA                    .2         .4            2.7
    VA                   *          *            *
    FSA/RHS              *          *            *
  Conventional           99.8        1.2           97.3
  Total                 100          1.2          100

MEMO: Number                     4,599             75

Unsecured
  Government backed
    FHA                    .6        4.6         *
    VA                   *          *            *
    FSA/RHS              *          *            *
  Conventional           99.4        2.9          100
  Total                 100          2.9          100

MEMO: Number                     4,399            267

MEMO: Total number             235,667         22,804

                                        Loans

                          One- to four-family home

                                 Total

                        Owner occupied

   Loan category                            Nonowner      Multifamily
   (purpose and       Percent               occupied,         home,
   lien status)       of loan    Percent     percent        percent
   and loan type      category   of loan     of loan        of loan
(government-backed      (1)      type (2)   category       category
 or conventional)                              (1)            (1)

Home purchase
First lien
  Government backed
    FHA                   9.0      100               .1           .5
    VA                    2.5      100           *             *
    FSA/RHS                .4      100           *             *
  Conventional           88.0      100             99.9         99.5
  Total                 100        100            100          100

MEMO: Number                 4,783,393        827,088       22,247

Junior lien
  Government backed
    FHA                    .2      100           *                .2
    VA                   *         100           *             *
    FSA/RHS              *         100           *             *
  Conventional           99.8      100            100           99.8
  Total                 100        100            100          100

MEMO: Number                   737,871         50,647          625

Refinance
First lien
  Government backed
    FHA                   3.2      100               .5           .8
    VA                    1.5      100               .7        *
    FSA/RHS              *         100           *             *
  Conventional           95.3      100             98.8         99.2
  Total                 100        100            100          100

MEMO: Number                 6,489,716        584,733       21,703

Junior lien
  Government backed
    FHA                  *         100           *                .3
    VA                   *         100           *             *
    FSA/RHS              *         100           *             *
  Conventional           99.9      100             99.9         99.7
  Total                 100        100            100          100

MEMO: Number                   464,463         11,568          580

Home improvement
First lien
  Government backed
    FHA                    .7      100               .1        *
    VA                     .1      100           *             *
    FSA/RHS              *         100           *             *
  Conventional           99.3      100             99.9        100
  Total                 100        100            100          100

MEMO: Number                   360,315         36,622        2,058

Junior lien
  Government backed
    FHA                    .6      100               .7           .2
    VA                   *         100           *             *
    FSA/RHS              *         100           *             *
  Conventional           99.4      100             99.3         99.9
  Total                 100        100            100          100

MEMO: Number                   397,825          7,133          653

Unsecured
  Government backed
    FHA                    .4      100               .1        *
    VA                   *         100           *             *
    FSA/RHS              *         100           *             *
  Conventional           99.6      100             99.9        100
  Total                 100        100            100          100

MEMO: Number                   151,909          6,569          284

MEMO: Total number          13,385,492      1,524,360       48,150

NOTE: For one- to four-family homes, excludes applications for which
occupancy status was missing.

FHA Federal Housing Administration VA Veterans Administration
FSA/RHS Farm Service Agency and Rural Housing Service

(1.) Distribution sums verticaly.

(2.) Distribution sums horizontally.

* Less than 0.05 percent.

5. Distribution of loans on manufactured homes, by type of loan
and type of home lender, 2004

                                       Home purchase

                             First lien            Junior lien

                         Percent    Percent    Percent    Percent
                            of         of         of         of
    Type of lender        loan      lender      loan      lender
                         type (1)   type (2)   type (1)   type (2)

                                     Government backed

Depository institution
  Commercial bank          28.1        5.2        18.8       *
  Savings institution       7.5       10.1        12.5       *
  Credit union               .1         .3        *          *

Mortgage company
  Independent              45.6       11.2        59.4       *
  Affiliated               18.8       18.0        9.4        *

All lenders               100          8.6      100          *

MEMO
Number of loans                20,909                  32

                                         Refinance

                             First lien            Junior lien

                         Percent    Percent    Percent    Percent
                            of         of         of         of
    Type of lender         loan      lender      loan      lender
                         type (1)   type (2)   type (1)   type (2)

                                     Government backed

Depository institution
  Commercial bank          22.5        1.2        *          *
  Savings institution       7.9        3.1       25.0        *
  Credit union               .2         .1       25.0        *

Mortgage company
  Independent              55.3        3.9       50.0        *
  Affiliated               14.1        3.8        *          *

All lenders               100          2.5      100          *

MEMO
Number of loans                 5,940                   4

                                     Home improvement

                             First lien            Junior lien

                         Percent    Percent    Percent    Percent
                            of         of         of         of
    Type of lender         loan      lender      loan      lender
                         type (1)   type (2)   type (1)   type (2)

                                     Government backed

Depository institution
  Commercial bank          45.5        *          50.0       *
  Savings institution       4.6        *          30.0       *
  Credit union              *          *          20.0       *

Mortgage company
  Independent              18.2        *          *          *
  Affiliated               31.8        .1         *          *

All lenders               100          *         100         *

MEMO
Number of loans                  88                    10

                          Home improvement

                             Unsecured

                         Percent    Percent
                            of         of
    Type of lender         loan      lender
                         type (1)   type (2)

                          Government backed

Depository institution
  Commercial bank          92.9        *
  Savings institution       *          *
  Credit union              3.6        *

Mortgage company
  Independent               3.6        *
  Affiliated                *          *

All lenders               100          *

MEMO
Number of loans                  28

                                      Home purchase

                             First lien            Junior lien

                         Percent    Percent    Percent    Percent
                            of         of         of         of
    Type of lender         loan      lender      loan      lender
                         type (1)   type (2)   type (1)   type (2)

                                       Conventional

Depository institution
  Commercial bank          48.4       49.1       24.2        .6
  Savings institution       6.9       50.9        3.4        .6
  Credit union              3.0       44.6       17.8       6.2

Mortgage company
  Independent              35.9       48.2       51.2       1.6
  Affiliated                5.8       30.2        3.4        .4

All lenders               100         47.1      100         1.1

MEMO
Number of loans                114,021                2,653

                                        Refinance

                             First lien            Junior lien

                         Percent    Percent    Percent    Percent
                            of         of         of         of
    Type of lender         loan      lender      loan      lender
                         type (1)   type (2)   type (1)   type (2)

                                       Conventional

Depository institution
  Commercial bank          45.8       31.6       41.9       1.1
  Savings institution       5.8       29.0        6.2       1.2
  Credit union              3.5       35.9        7.0       2.7

Mortgage company
  Independent              33.4       30.5       36.6       1.3
  Affiliated               11.5       40.9        8.3       1.1

All lenders               100         32.0      100         1.2

MEMO
Number of loans                77,571                 2,908

                                     Home improvement

                             First lien            Junior lien

                         Percent    Percent    Percent    Percent
                            of         of         of         of
    Type of lender         loan      lender      loan      lender
                         type (1)   type (2)   type (1)   type (2)

                                       Conventional

Depository institution
  Commercial bank          61.6       4.8        63.9       2.6
  Savings institution       5.5       3.1         3.9       1.2
  Credit union              3.5       4.0         6.2       3.8

Mortgage company
  Independent              21.2       2.2        19.0       1.0
  Affiliated                8.2       3.3         7.0       1.5

All lenders               100         3.6       100         1.9

MEMO
Number of loans                 8,739                 4,633

                          Home improvement

                             Unsecured

                         Percent    Percent     MEMO
                            of         of      Number
    Type of lender         loan      lender      of
                         type (1)   type (2)    loans

                            Conventional

Depository institution
  Commercial bank          89.7       3.7      112,385
  Savings institution       2.5        .8       15,391
  Credit union              4.0       2.5        7,632

Mortgage company
  Independent                .3        *        84,940
  Affiliated                3.5        .8       21,849

All lenders               100         1.9      242,197

MEMO
Number of loans                 4,661          242,197

NOTE: Excludes transition-period loans (those for which the
application was submitted before 2004). For definition of
manufactured home, see text note 9. See also table 1, note 1.

(1.) Distribution sums vertically.

(2.) Distribution sums horizontally.

* Less than 0.05 percent.

6. Distribution of home-purchase loans for one- to four-family
owner-occupied homes, by characteristic of borrower and of
census tract and by type of home, 2004

                                 Site built            Manufactured

                            Percent of                  Percent of
     Characteristic          charac-      Percent of     charac-
       and status          teristic (1)   status (2)   teristic (1)

      BORROWER (3)

Income ratio (percent
  of area median)
Less than 50                     4.7          93.3          12.2
50-79                           18.9          95.8          29.1
80-119                          29.4          97.2          30.1
120 or more                     46.9          98.3          28.7
  Total (4)                    100            97.3         100

Race
American Indian or                .8          96.1           1.2
  Alaska Native
Asian                            4.9          99.7            .5
Black or African                 7.1          97.9           5.4
  American
Native Hawaiian or                .5          98.3            .3
  other Pacific Islander
White                           74.9          96.9          84.4
Two or more minority              .1          97.9            .1
  races
Joint                            1.4          98.0           1.0
Missing (5)                     10.4          98.1           7.2
  Total                        100            97.3         100

Ethnicity
Hispanic or Latino              11.2          98.2           7.4
Not Hispanic or Latino          76.4          97.1          82.1
Joint (6)                        1.3          97.9           1.0
Missing (5)                     11.1          97.7           9.5
  Total                        100            97.3         100

Minority status
Minority                        26.0          98.3          15.9
Non-Hispanic white              62.6          96.8          73.7
Missing (5)                     11.4          97.5          10.3
  Total                        100            97.3         100

CENSUS TRACT OF PROPERTY

Income ratio (percent
  of area median)
Less than 50                     1.6          99.1            .5
50-79                           13.2          96.7          16.8
80-119                          49.2          96.2          71.4
120 or more                     36.0          99.1          11.4
  Total (4)                    100            97.3         100

Racial or ethnic
  composition
  (minorities as
  percentage of
  population)
Less than 10                    32.5          96.5          43.4
10-19                           22.9          97.6          21.1
20-49                           27.7          97.5          25.7
50-79                           10.2          98.0           7.6
80-100                           6.6          99.1           2.2
  Total (4)                    100            97.3         100

Location
Central city                    38.5          98.8          17.1
Noncentral city                 52.7          97.6          47.3
Rural or only state              8.8          90.0          35.6
    known
  Total (4)                    100            97.3         100

                            Manufac-
                             tured               Total

                                         Percent of
     Characteristic        Percent of     charac-      Percent of
       and status          status (2)   teristic (1)     status

      BORROWER (3)

Income ratio (percent
  of area median)
Less than 50                  6.8            4.9          100
50-79                         4.2           19.2          100
80-119                        2.8           29.5          100
120 or more                   1.7           46.4          100
  Total (4)                   2.8          100            100

Race
American Indian or            3.9             .8          100
  Alaska Native
Asian                          .3            4.7          100
Black or African              2.1            7.0          100
  American
Native Hawaiian or            1.7             .5          100
  other Pacific Islander
White                         3.1           75.2          100
Two or more minority          2.1             .1          100
  races
Joint                         2.0            1.4          100
Missing (5)                   1.9           10.3          100
  Total                       2.8          100            100

Ethnicity
Hispanic or Latino            1.8           11.1          100
Not Hispanic or Latino        2.9           76.6          100
Joint (6)                     2.1            1.3          100
Missing (5)                   2.4           11.1          100
  Total                       2.8          100            100

Minority status
Minority                      1.7           25.7          100
Non-Hispanic white            3.2           62.9          100
Missing (5)                   2.5           11.4          100
  Total                       2.8          100            100

CENSUS TRACT OF PROPERTY

Income ratio (percent
  of area median)
Less than 50                   .9            1.5          100
50-79                         3.4           13.3          100
80-119                        3.8           49.8          100
120 or more                    .9           35.3          100
  Total (4)                   2.7          100            100

Racial or ethnic
  composition
  (minorities as
  percentage of
  population)
Less than 10                  3.5           32.8          100
10-19                         2.5           22.9          100
20-49                         2.5           27.7          100
50-79                         2.0           10.2          100
80-100                        0.9            6.5          100
  Total (4)                   2.7          100            100

Location
Central city                  1.2           38.0          100
Noncentral city               2.4           52.5          100
Rural or only state          10.0            9.5          100
    known
  Total (4)                   2.7          100            100

     Characteristic          MEMO
       and status           Number

      BORROWER (3)

Income ratio (percent
  of area median)
Less than 50                 205,771
50-79                        798,602
80-119                     1,227,091
120 or more                1,933,772
  Total (4)                4,165,236

Race
American Indian or            36,650
  Alaska Native
Asian                        206,716
Black or African             305,432
  American
Native Hawaiian or            23,246
  other Pacific Islander
White                      3,280,354
Two or more minority           3,282
  races
Joint                         59,524
Missing (5)                  447,970
  Total                    4,363,174

Ethnicity
Hispanic or Latino           483,253
Not Hispanic or Latino     3,341,979
Joint (6)                     55,914
Missing (5)                  482,028
  Total                    4,363,174

Minority status
Minority                   1,120,646
Non-Hispanic white         2,745,937
Missing (5)                  496,591
  Total                    4,363,174

CENSUS TRACT OF PROPERTY

Income ratio (percent
  of area median)
Less than 50                  65,777
50-79                        575,070
80-119                     2,149,842
120 or more                1,524,643
  Total (4)                4,315,332

Racial or ethnic
  composition
  (minorities as
  percentage of
  population)
Less than 10               1,417,201
10-19                        988,062
20-49                      1,193,394
50-79                        438,175
80-100                       279,509
  Total (4)                4,316,341

Location
Central city               1,642,184
Noncentral city            2,272,738
Rural or only state          411,822
    known
  Total (4)                4,326,744

NOTE: Excludes transition-period loans (those for which the application
was submitted before 2004). For definition of income categories for
borrower and census tract, see text note 34. Census tract is for the
property securing the loan. Categories for race and ethnicity reflect
the revised standards established in 1997 by the Office of Management
and Budget (OMB); for details, see text discussion. The term minority
means Hispanic or Latino ethnicity or any race other than white.
Census-tract data reflect the 2000 decennial census; they also reflect
definitions for metropolitan statistical areas established by the OMB
in June 2003 and used in HMDA for the first time in the 2004 data
(see note 2 in main-text box "Distribution of HMDA Data and Pre-2004
Requirements of Regulation C").

(1.) Distribution sums vertically.

(2.) Distribution sums horizontally.

(3.) For details on the identification of borrower income, race,
and ethnicity, see text note 35.

(4.) Excludes loans for the information for the characteristic was
missing on the application.

(5.) Information for the characteristic was missing on the
application.

(6.) On the applications for these loans, one applicant reported
"Hispanic or Latino," and the other reported "not Hispanic or Latino."

7. Disposition of applications for home loans, and origination
and pricing of loans, by type of home and type of loan, 2004

                                            Applications

                                             Acted upon by lender

                                  Number                   Number
    Type of home and loan       submitted      Number      denied

     ONE- TO FOUR-FAMILY
   Nonbusiness related (4)
       Owner occupied

Site built
Home purchase
  Conventional
    First lien                   5,559,099    4,938,892     737,756
    Junior lien                  1,072,726      964,662     164,750

  Government backed
    First lien                     652,281      583,299      79,253
    Junior lien                      1,563        1,254         171

Refinance
  Conventional
    First lien                  12,261,720    9,641,212   2,973,609
    Junior lien                    954,842      785,067     270,594

  Government backed
    First lien                     427,105      347,785      51,661
    Junior lien                        766          451         172

Home improvement
  Conventional
    First lien                     706,594      619,012     224,727
    Junior lien                    915,901      784,857     332,508

  Government backed
    First lien                       3,876        3,361         820
    Junior lien                      5,505        4,899       2,372

  Conventional or government-
    Backed, unsecured              364,947      348,629     182,505

Manufactured
Conventional, first lien
  Home purchase                    359,129      347,524     186,618
  Refinance                        239,999      201,876     104,276
Other                               99,144       88,765      33,661

    Nonowner occupied (5)

Conventional, first lien
  Home purchase                  1,112,330    1,003,071     156,925
  Refinance                        937,424      808,515     193,158
Other                              234,450      208,729      70,590

    Business related (4)

Conventional, first lien
  Home purchase                     54,944       50,213       3,062
  Refinance                         55,051       47,590       5,952
Other                               29,115       26,444       3,638

       MULTIFAMILY (6)

Conventional, first lien
  Home purchase                     24,593       22,599       2,372
  Refinance                         23,424       21,619       2,306
Other                                5,662        5,067         954

Total                           26,102,190   21,855,392   5,784,410

                                Applications       Loans originated

                                                              Loans
                                                               with
                                                              annual
                                 Acted upon                 percentage
                                 by lender                  rate (APR)
                                                              spread
                                                            above the
                                                            threshold
                                                               (1)

                                  Percent
    Type of home and loan          denied        Number       Number

     ONE- TO FOUR-FAMILY
   Nonbusiness related (4)
       Owner occupied

Site built
Home purchase
  Conventional
    First lien                      14.9        3,745,490      432,364
    Junior lien                     17.1          701,078      270,688

  Government backed
    First lien                      13.6          479,498        6,298
    Junior lien                     13.6            1,036           29

Refinance
  Conventional
    First lien                      30.8        5,708,965      884,108
    Junior lien                     34.5          439,495      120,500

  Government backed
    First lien                      14.9          269,349        4,084
    Junior lien                     38.1              268           12

Home improvement
  Conventional
    First lien                      36.3          339,836       74,584
    Junior lien                     42.4          376,785       65,185

  Government backed
    First lien                      24.4            2,350           90
    Junior lien                     48.4            2,142        1,133

  Conventional or government-
    Backed, unsecured               52.3          143,856          ...

Manufactured
Conventional, first lien
  Home purchase                     53.7           98,864       56,498
  Refinance                         51.7           71,508       34,171
Other                               37.9           48,565        9,807

    Nonowner occupied (5)

Conventional, first lien
  Home purchase                     15.6          760,796       92,715
  Refinance                         23.9          539,758       75,537
Other                               33.8          122,321       36,442

    Business related (4)

Conventional, first lien
  Home purchase                      6.1           45,339        4,244
  Refinance                         12.5           38,922        3,997
Other                               13.8           21,427        1,952

       MULTIFAMILY (6)

Conventional, first lien
  Home purchase                     10.5           19,294          861
  Refinance                         10.7           18,468          886
Other                               18.8            3,942          279

Total                               26.5       13,999,352    2,176,464

                                   Loans originated

                                   Loans with annual
                                  percentage rate (APR)
                                    spread above the
                                     threshold (1)

                                              Percent
                                           distribution,
                                           by percentage
                                             points of
                                            APR spread

    Type of home and loan       Percent   3-3.99   4-4.99

     ONE- TO FOUR-FAMILY
   Nonbusiness related (4)
       Owner occupied

Site built
Home purchase
  Conventional
    First lien                   11.5      58.0     27.5
    Junior lien                  38.6       ...      ...

  Government backed
    First lien                    1.3      58.3     24.4
    Junior lien                   2.8       ...      ...

Refinance
  Conventional
    First lien                   15.5      23.9     28.1
    Junior lien                  27.4       ...      ...

  Government backed
    First lien                    1.5      69.5     19.8
    Junior lien                   4.5       ...      ...

Home improvement
  Conventional
    First lien                   21.9      49.0     25.9
    Junior lien                  17.3       ...      ...

  Government backed
    First lien                    3.8      48.9     21.1
    Junior lien                  52.9       ...      ...

  Conventional or government-
    Backed, unsecured             ...       ...      ...

Manufactured
Conventional, first lien
  Home purchase                  57.1      22.9     21.8
  Refinance                      47.8      32.9     27.3
Other                            20.2      24.3     13.4

    Nonowner occupied (5)

Conventional, first lien
  Home purchase                  12.2      59.0     26.8
  Refinance                      14.0      53.4     27.4
Other                            29.8      10.7      5.9

    Business related (4)

Conventional, first lien
  Home purchase                   9.4      49.9     24.7
  Refinance                      10.3      45.7     29.0
Other                             9.1       6.3      4.7

       MULTIFAMILY (6)

Conventional, first lien
  Home purchase                   4.5      60.2     23.2
  Refinance                       4.8      58.8     23.5
Other                             7.1      14.0     10.8

Total                            15.5      41.4     21.3

                                     Loans originated

                                Loans with annual percentage
                                  rate (APR) spread above
                                     the threshold (1)

                                   Percent distribution,
                                   by percentage points
                                       of APR spread

    Type of home and loan       5-6.99   7-8.99   9 or more

     ONE- TO FOUR-FAMILY
   Nonbusiness related (4)
       Owner occupied

Site built
Home purchase
  Conventional
    First lien                    13.2      1.2          .2
    Junior lien                   76.2     21.7         2.1

  Government backed
    First lien                    12.7      4.3          .3
    Junior lien                   69.0     17.2        13.8

Refinance
  Conventional
    First lien                    15.7      2.2          .2
    Junior lien                   58.3     28.9        12.9

  Government backed
    First lien                     9.4      1.0          .3
    Junior lien                   58.3     41.7           0

Home improvement
  Conventional
    First lien                    19.4      4.5         1.1
    Junior lien                   41.5     27.9        30.7

  Government backed
    First lien                    21.1      7.8         1.1
    Junior lien                   23.1     29.3        47.6

  Conventional or government-
    Backed, unsecured              ...      ...         ...

Manufactured
Conventional, first lien
  Home purchase                   32.4     16.8         6.0
  Refinance                       28.9      8.2         2.8
Other                             30.5     16.4        15.4

    Nonowner occupied (5)

Conventional, first lien
  Home purchase                   11.6      1.8          .8
  Refinance                       16.4      2.4          .5
Other                             50.9     25.6         6.8

    Business related (4)

Conventional, first lien
  Home purchase                   20.7      3.5         1.2
  Refinance                       21.6      3.2          .5
Other                             64.3     20.9         3.8

       MULTIFAMILY (6)

Conventional, first lien
  Home purchase                   13.9      2.1          .6
  Refinance                       16.0      1.4          .3
Other                             57.7     15.1         2.5

Total                             27.2      7.7         2.5

                                    Loans originated

                                    Loans with annual
                                  percentage rate (APR)
                                    spread above the
                                     threshold (1)

                                  APR spread
                                 (percentage
                                   points)

                                                Number of
                                                 HOEPA-
                                                 covered
    Type of home and loan       Mean   Median   loans (2)

     ONE- TO FOUR-FAMILY
   Nonbusiness related (4)
       Owner occupied

Site built
Home purchase
  Conventional
    First lien                   4.1    3.8           ...
    Junior lien                  6.4    6.2           ...

  Government backed
    First lien                   4.2    3.9           ...
    Junior lien                  7.1    6.0           ...

Refinance
  Conventional
    First lien                   4.2    3.9         7,249
    Junior lien                  7.3    6.7         3,987

  Government backed
    First lien                   3.9    3.6           496
    Junior lien                  6.7    6.6             2

Home improvement
  Conventional
    First lien                   4.4    4.0         1,965
    Junior lien                  8.0    7.6         5,046

  Government backed
    First lien                   4.7    4.0             4
    Junior lien                  9.2    8.9         1,002

  Conventional or government-
    Backed, unsecured            ...    ...           ...

Manufactured
Conventional, first lien
  Home purchase                  5.7    5.2           ...
  Refinance                      5.0    4.6         1,830
Other                            6.4    5.5           904

    Nonowner occupied (5)

Conventional, first lien
  Home purchase                  4.1    3.8           ...
  Refinance                      4.2    3.9           612
Other                            6.4    6.1           218

    Business related (4)

Conventional, first lien
  Home purchase                  4.4    4.0           ...
  Refinance                      4.4    4.1           104
Other                            6.2    6.1            29

       MULTIFAMILY (6)

Conventional, first lien
  Home purchase                  4.1    3.7           ...
  Refinance                      4.1    3.8            29
Other                            5.9    5.8             7

Total                            4.8    4.3        23,484

                                              MEMO
                                 Transition-period applications
                                 (those submitted before 2004)

                                 Number     Number     Percent
    Type of home and loan       submitted   denied    denied (3)

     ONE- TO FOUR-FAMILY
   Nonbusiness related (4)
       Owner occupied

Site built
Home purchase
  Conventional
    First lien                    490,846    41,115       9.9
    Junior lien                    47,351     5,230      13.4

  Government backed
    First lien                     85,896     8,172      11.5
    Junior lien                       226        16       9.8

Refinance
  Conventional
    First lien                    813,761   106,316      18.5
    Junior lien                    36,965     6,184      21.6

  Government backed
    First lien                     49,849     6,008      15.9
    Junior lien                        63         6      21.4

Home improvement
  Conventional
    First lien                     13,773     1,733      14.6
    Junior lien                    21,962     4,342      24.0

  Government backed
    First lien                        391        74      25.0
    Junior lien                       273        67      30.9

  Conventional or government-
    Backed, unsecured               4,881     1,209      27.1

Manufactured
Conventional, first lien
  Home purchase                     9,595     1,177      13.8
  Refinance                        12,252     2,370      25.0
Other                               6,679       722      12.5

    Nonowner occupied (5)

Conventional, first lien
  Home purchase                    84,952     7,864      10.8
  Refinance                        82,569    11,682      17.3
Other                               8,684     1,120      15.7

    Business related (4)

Conventional, first lien
  Home purchase                    87,425    11,520      14.9
  Refinance                       117,852    25,941      25.4
Other                              42,414     7,973      22.0

       MULTIFAMILY (6)

Conventional, first lien
  Home purchase                     3,607       169       5.2
  Refinance                         3,920       262       7.3
Other                                 689        35       5.6

Total                           2,026,875   251,307      15.7

                                              MEMO
                                 Transition-period applications
                                 (those submitted before 2004)

                                Loans originated

                                             Percent
                                               with      Number of
                                            APR spread    HOEPA-
                                              above       covered
    Type of home and loan        Number     threshold    loans (2)

     ONE- TO FOUR-FAMILY
   Nonbusiness related (4)
       Owner occupied

Site built
Home purchase
  Conventional
    First lien                    303,881       5.1             ...
    Junior lien                    26,475      23.1             ...

  Government backed
    First lien                     56,693        .7             ...
    Junior lien                       134       2.2             ...

Refinance
  Conventional
    First lien                    333,550      10.7             322
    Junior lien                    16,479      23.7             177

  Government backed
    First lien                     23,485        .8             106
    Junior lien                        19         0             ...

Home improvement
  Conventional
    First lien                      7,912      14.6              63
    Junior lien                    10,646      13.0             158

  Government backed
    First lien                        189       3.7             ...
    Junior lien                        88      43.2              17

  Conventional or government-
    Backed, unsecured               2,132       ...             ...

Manufactured
Conventional, first lien
  Home purchase                     5,003      22.2             ...
  Refinance                         4,590      21.2              41
Other                               4,308       4.9              24

    Nonowner occupied (5)

Conventional, first lien
  Home purchase                    53,019       7.8             ...
  Refinance                        42,145       9.1              34
Other                               4,739      24.9               6

    Business related (4)

Conventional, first lien
  Home purchase                    50,885       5.7             ...
  Refinance                        54,936      10.5              36
Other                              21,444       9.8             121

       MULTIFAMILY (6)

Conventional, first lien
  Home purchase                     2,838       3.1             ...
  Refinance                         3,060       3.0               3
Other                                 548       2.7               2

Total                           1,029,198       8.4           1,110

NOTE: Excludes transition-period applications (those submitted before
2004) and transition-period loans (those for which the application
was submitted before 2004).

(1.) APR spread is the difference between the APR on the loan and
the yield on a comparable-maturity Treasury security. The threshold
for first-lien loans is a spread of 3 percentage points; for
junior-lien loans, it is a spread of 5 percentage points.

(2.) Loans covered by the Home Ownership and Equity Protection Act of
1994, which does not apply to home-purchase loans (for details, see
text).

(3.) Number denied divided by number (not shown) acted upon.

(4.) Business-related applications and loans are those for which the
lender reported that the race, ethnicity, and sex of the applicant or
co-applicant are "not applicable"; all other applications and loans
are nonbusiness related.

(5.) Includes applications and loans for which occupancy status was
missing.

(6.) Includes business-related and nonbusiness-related applications
and loans for owner-occupied and nonowner-occupied properties.

... Not applicable

8. Home-purchase lending that began with a request for pre-approval:
Disposition and pricing, by type of home, 2004

                              Request for pre-approval

                            Number     Number     Percent
      Type of home         submitted   denied    denied (2)

  ONE- TO FOUR-FAMILY
  NONBUSINESS RELATED
     Owner occupied

Site built
Conventional
  First lien                 684,306   153,773         22.5
  Junior lien                 88,793    15,423         17.4

Government backed
  First lien                 100,118    26,682         26.7
  Junior lien                    147        35         23.8

Manufactured
Conventional, first lien      39,791    23,838         59.9
Other                          3,714       935         25.2

   Nonowner occupied

Conventional, first lien      84,763    15,252         18.0
Other                          7,003     1,067         15.2

    BUSINESS RELATED

Conventional, first lien       3,667       350          9.5
Other                          1,540       117          7.6

      MULTIFAMILY

Conventional, first lien         226        29         12.8
Other                             16         4         25.0

Total                      1,014,084   237,505         23.4

                           Applications preceded by
                           requests for pre-approval (1)

                                          Acted upon
                                          by lender

                            Number               Number
      Type of home         submitted   Number    denied

  ONE- TO FOUR-FAMILY
  NONBUSINESS RELATED
     Owner occupied

Site built
Conventional
  First lien                 448,771   396,998   34,665
  Junior lien                 67,757    61,771    4,728

Government backed
  First lien                  71,632    64,214    7,218
  Junior lien                    118        99       13

Manufactured
Conventional, first lien      37,592    35,700   22,039
Other                          2,912     2,351      419

   Nonowner occupied

Conventional, first lien      58,101    51,548    5,457
Other                          5,953     5,040      487

    BUSINESS RELATED

Conventional, first lien       3,317     2,791      244
Other                          1,419     1,208       90

      MULTIFAMILY

Conventional, first lien         207       186       14
Other                             14        13        2

Total                        697,793   621,919   75,376

                           Loan originations whose applications were
                           preceded by requests for pre-approval

                                     Loans with annual percentage
                                           Rate (APR) spread
                                        Above the threshold (3)

                                                           Percent
                                                        distribution
                                                        by percentage
                                                        points of APR
                                                            spread

      Type of home         Number    Number   Percent   3-3.99   4-4.99

  ONE- TO FOUR-FAMILY
  NONBUSINESS RELATED
     Owner occupied

Site built
Conventional
  First lien               332,804   27,340       8.2     49.2     28.4
  Junior lien               52,671   14,865      28.2   ...      ...

Government backed
  First lien                53,527      662       1.2     34.1      7.4
  Junior lien                   83        7       8.4      ...      ...

Manufactured
Conventional, first lien     7,430    4,735      63.7     19.6     23.4
Other                        1,787       94       5.3     20.2        0

   Nonowner occupied

Conventional, first lien    41,564    3,478       8.4     48.0     21.8
Other                        4,000    1,307      32.7        0        0

    BUSINESS RELATED

Conventional, first lien     2,417      364      15.1     36.8     19.8
Other                        1,079      149      13.8      0.7        0

      MULTIFAMILY

Conventional, first lien       167       19      11.4     47.4     31.6
Other                           11        6      54.5        0        0

Total                      497,540   53,026      10.7     31.0     18.4

                           Loan originations whose applications were
                           preceded by requests for pre-approval

                           Loans with annual percentage rate (APR)
                               spread above the threshold (3)

                              Percent distribution         APR spread
                            by percentage points of       (percentage
                                   APR spread               points)

      Type of home         5-6.99   7-8.99   9 or more   Mean   Median

  ONE- TO FOUR-FAMILY
  NONBUSINESS RELATED
     Owner occupied

Site built
Conventional
  First lien                 20.0      2.0          .5    4.3      4.0
  Junior lien                81.3     16.7         2.0    6.2      6.0

Government backed
  First lien                 29.5     28.5          .5    5.6      6.3
  Junior lien                71.4     14.3        14.3    6.4      5.3

Manufactured
Conventional, first lien     33.5     15.2         8.3    5.8      5.3
Other                        77.7      2.1           0      5      5.3

   Nonowner occupied

Conventional, first lien       17      7.7         5.4    4.8      4.0
Other                        53.8     38.5         7.7      7      6.8

    BUSINESS RELATED

Conventional, first lien       31      8.2         4.1    5.1      4.6
Other                        82.6     11.4         5.4    6.4      6.3

      MULTIFAMILY

Conventional, first lien     10.5        0        10.5    5.1      4.1
Other                          50       50           0    6.8      6.9

Total                        39.5        9         2.1    5.1      5.0

                                   MEMO
                              Applications with
                           transition-period requests
                           for pre-approval (request
                             submitted before 2004)

                                    Number     Number
      Type of home         Number   denied   denied (2)

  ONE- TO FOUR-FAMILY
  NONBUSINESS RELATED
     Owner occupied

Site built
Conventional
  First lien               20,444      731          4.3
  Junior lien               1,269       37          3.4

Government backed
  First lien                5,866      351          7.2
  Junior lien                  11        1         10.0

Manufactured
Conventional, first lien      172       23         18.4
Other                         508       26          6.5

   Nonowner occupied

Conventional, first lien    2,406      107          5.4
Other                         184        8          5.8

    BUSINESS RELATED

Conventional, first lien    1,800       75          4.6
Other                         570       21          4.2

      MULTIFAMILY

Conventional, first lien       16        0            0
Other                           1        0            0

Total                      33,247    1,380            5

                                      MEMO
                               Applications with
                           transition-period requests
                           for pre-approval (request
                             submitted before 2004)

                            Loans originated

                                    Percent
                                    with APR
                                    spread
                                      above
      Type of home         Number   threshold

  ONE- TO FOUR-FAMILY
  NONBUSINESS RELATED
     Owner occupied

Site built
Conventional
  First lien               13,382         3.8
  Junior lien                 919         9.6

Government backed
  First lien                3,770         1.4
  Junior lien                   9           0

Manufactured
Conventional, first lien       91        22.0
Other                         335           0

   Nonowner occupied

Conventional, first lien    1,493         4.8
Other                          81        18.5

    BUSINESS RELATED

Conventional, first lien    1,393         3.2
Other                         450         4.7

      MULTIFAMILY

Conventional, first lien       15         6.7
Other                           1           0

Total                      21,939         3.8

NOTE: Excludes transition-period requests for pre-approval (those
submitted before 2004). See also notes 4, 5, and 6 of table 7
for details on business-related, nonowner-occupied, and
multifamily properties and general note to table 2.

(1.) These applications are included in the total of 26,102,190
reported in table 7.

(2.) Number denied divided by number (not shown) acted upon.

(3.) See table 7, note 1.

... Not applicable.

9. Unadjusted and adjusted denial rates on applications for loans on
one- to four-family homes, by type of loan and by race and ethnicity
and sex of applicant, 2004

A. Home purchase, owner-occupied site-built home

Percent except as noted

                                       Conventional

                                        First lien

Race and ethnicity                               Adjusted denial rate,
     and sex                                     by adjustment factor
                        Number      Unadjusted
                          of          denial                 Borrower-
                     applications      rate      Borrower-    related
                                                  related      plus
                                                              lender

Race and ethnicity
American Indian or
  Alaska Native           42,460       21.1        20.4        15.7
Asian                    288,060       13.5        12.7        12.9
Black or African
  American               402,090       24.7        22.3        18.2
Native Hawaiian or
  other Pacific
  Islander                30,866       17.5        15.3        15.4
Two or more
  minority races           3,768       13.8        12.3        15.7
Joint                     64,744       12.0        15.0        12.4
Race missing             696,276       21.8        20.1        15.1
Hispanic white           423,395       18.4        15.8        14.8
Non-Hispanic white     3,309,353       10.9        10.9        10.9

Sex
One male               1,636,413       16.8        16.8        16.8
One female             1,217,287       16.0        15.3        15.8
Two males                 55,264       15.6        15.6        15.6
Two females               45,100       15.2        14.0        14.4

                                     Conventional

                                    Junior lien

Race and ethnicity                               Adjusted denial rate,
     and sex                                     by adjustment factor
                        Number      Unadjusted
                          of          denial                 Borrower-
                     applications      rate      Borrower-    related
                                                  related      plus
                                                              lender

Race and ethnicity
American Indian or
  Alaska Native          11,211        22.7        22.6        18.3
Asian                    48,970        18.6        17.4        16.3
Black or African
  American               96,741        21.9        21.2        18.8
Native Hawaiian or
  other Pacific
  Islander                9,312        18.3        17.2        15.6
Two or more
  minority races            975        12.5        11.8        15.5
Joint                    12,224        13.1        15.3        13.8
Race missing            156,504        20.4        19.6        17.7
Hispanic white          124,483        20.0        18.0        16.0
Non-Hispanic white      532,260        13.7        13.7        13.7

Sex
One male                344,552        18.7        18.7        18.7
One female              235,350        18.3        18.2        18.4
Two males                11,311        18.3        18.3        18.3
Two females               8,591        17.5        16.4        16.1

                            Government backed, first lien

Race and ethnicity                               Adjusted denial rate,
     and sex                                     by adjustment factor
                        Number      Unadjusted
                          of          denial                 Borrower-
                     applications      rate      Borrower-    related
                                                  related      plus
                                                              lender

Race and ethnicity
American Indian or
  Alaska Native           6,425        15.5        14.4        14.7
Asian                     7,645        12.4        12.2        13.2
Black or African
  American               85,845        17.2        16.9        16.0
Native Hawaiian or
  other Pacific
  Islander                3,629        13.8        12.9        11.7
Two or more
  minority races            534        15.4        15.8        12.1
Joint                    12,807        9.6         12.5        11.8
Race missing             63,179        20.0        17.0        16.6
Hispanic white           59,172        15.6        13.8        13.5
Non-Hispanic white      396,856        10.4        10.4        10.4

Sex
One male                195,671        14.0        14.0        14.0
One female              133,456        14.4        13.2        13.5
Two males                13,409        11.2        11.2        11.2
Two females               9,121        12.1        10.9        11.8

B. Refinance, owner-occupied site-built home

Percent except as noted

                                       Conventional

                                       First lien

Race and ethnicity                               Adjusted denial rate,
     and sex                                     by adjustment factor
                        Number      Unadjusted
                          of          denial                 Borrower-
                     applications      rate      Borrower-    related
                                                  related      plus
                                                              lender

Race and ethnicity
American Indian or
  Alaska Native           93,068       38.2        40.1        32.3
Asian                    321,978       18.7        25.7        27.3
Black or African
  American               897,836       41.9        39.8        32.4
Native Hawaiian or
  other Pacific
  Islander                55,859       29.7        33.5        31.2
Two or more
  minority races           8,340       28.2        32.9        31.3
Joint                    108,225       25.7        33.0        27.3
Race missing           2,084,368       43.9        47.3        32.3
Hispanic white           639,075       29.6        31.3        28.6
Non-Hispanic white     5,916,294       24.3        24.3        24.3

Sex
One male               2,683,328       35.1        35.1        35.1
One female             2,161,057       32.2        31.0        32.8
Two males                 69,113       26.9        26.9        26.9
Two females               78,731       30.1        27.0        26.1

                                      Conventional

                                     Junior lien

Race and ethnicity                               Adjusted denial rate,
     and sex                                     by adjustment factor
                        Number      Unadjusted
                          of          denial                 Borrower-
                     applications      rate      Borrower-    related
                                                  related      plus
                                                              lender

Race and ethnicity
American Indian or
  Alaska Native           6,529        40.8        38.6        34.9
Asian                    20,194        33.1        36.1        35.2
Black or African
  American               54,132        44.5        42.2        37.3
Native Hawaiian or
  other Pacific
  Islander                4,518        36.3        36.3        37.3
Two or more
  minority races            704        31.5        37.3        39.8
Joint                     9,382        28.9        33.9        31.7
Race missing            193,182        43.8        41.1        35.2
Hispanic white           38,892        37.9        37.0        33.8
Non-Hispanic white      475,471        28.4        28.4        28.4

Sex
One male                179,812        37.4        37.4        37.4
One female              131,172        37.9        35.5        36.0
Two males                 5,471        32.7        32.7        32.7
Two females               5,370        34.4        32.8        34.2

                               Government backed, first lien

Race and ethnicity                               Adjusted denial rate,
     and sex                                     by adjustment factor
                        Number      Unadjusted
                          of          denial                 Borrower-
                     applications      rate      Borrower-    related
                                                  related      plus
                                                              lender

Race and ethnicity
American Indian or
  Alaska Native           3,222        15.7        13.8        13.3
Asian                     3,698        15.0        15.4        16.1
Black or African
  American               69,607        17.6        18.5        17.2
Native Hawaiian or
  other Pacific
  Islander                2,204        12.2        12.0        14.1
Two or more
  minority races            442        13.6        13.6        14.3
Joint                     7,857        11.0        13.8        14.0
Race missing             55,907        21.4        17.8        17.1
Hispanic white           29,182        15.8        15.1        16.2
Non-Hispanic white      206,851        12.4        12.4        12.4

Sex
One male                 91,609        16.1        16.1        16.1
One female               71,666        16.1        15.2        16.3
Two males                 4,995        17.4        17.4        17.4
Two females               4,740        15.5        14.6        14.7

C. Home improvement, conventional loan, owner-occupied site-built home

Percent except as noted

                                         First lien

       Race and ethnicity             Number      Unadjusted
            and sex                    of           denial
                                   applications      rate

Race and ethnicity
American Indian or Alaska Native        8,843        47.1
Asian                                  15,204        27.8
Black or African American              67,098        49.1
Native Hawaiian or other
  Pacific Islander                      4,402        38.3
Two or more minority races                677        41.5
Joint                                   6,832        32.6
Race missing                          102,716        53.0
Hispanic white                         54,319        37.5
Non-Hispanic white                    360,959        28.6

Sex
One male                              159,817        43.4
One female                            143,959        38.4
Two males                               5,024        39.1
Two females                             5,788        38.4

                                        First lien

       Race and ethnicity           Adjusted denial rate,
          and sex                   by adjustment factor

                                   Borrower-    Borrower-
                                    related      related
                                               plus lender

Race and ethnicity
American Indian or Alaska Native     47.1         40.3
Asian                                34.8         33.4
Black or African American            47.2         41.7
Native Hawaiian or other
  Pacific Islander                   42.7         38.9
Two or more minority races           43.0         40.7
Joint                                40.2         34.1
Race missing                         56.0         39.5
Hispanic white                       37.7         36.2
Non-Hispanic white                   28.6         28.6

Sex
One male                             43.4         43.4
One female                           37.0         39.9
Two males                            39.1         39.1
Two females                          37.5         34.6

                                         Junior lien

       Race and ethnicity             Number      Unadjusted
            and sex                     of          denial
                                   applications      rate

Race and ethnicity
American Indian or Alaska Native       7,712         54.7
Asian                                 16,276         46.1
Black or African American             62,045         60.1
Native Hawaiian or other
  Pacific Islander                     4,538         50.1
Two or more minority races               656         62.0
Joint                                 10,842         40.6
Race missing                         159,984         49.6
Hispanic white                        39,529         51.4
Non-Hispanic white                   494,636         36.3

Sex
One male                             189,244         51.1
One female                           143,026         50.0
Two males                              5,588         46.3
Two females                            6,261         44.3

                                        Junior lien

       Race and ethnicity           Adjusted denial rate,
            and sex                 by adjustment factor

                                                Borrower-
                                   Borrower-     related
                                    related    plus lender

Race and ethnicity
American Indian or Alaska Native     48.2         47.3
Asian                                45.7         44.3
Black or African American            55.6         52.0
Native Hawaiian or other
  Pacific Islander                   46.4         43.8
Two or more minority races           57.0         48.8
Joint                                47.1         42.2
Race missing                         48.4         42.1
Hispanic white                       46.4         45.1
Non-Hispanic white                   36.3         36.3

Sex
One male                             51.1         51.1
One female                           47.3         49.4
Two males                            46.3         46.3
Two females                          42.9         44.9

D. Manufactured housing, conventional loan, first lien,
owner-occupied home

Percent except as noted

                                        Home purchase

                                      Number      Unadjusted
       Race and ethnicity               of          denial
            and sex                applications      rate

Race and ethnicity
American Indian or Alaska Native        4,785        57.3
Asian                                   1,736        44.4
Black or African American              28,363        64.2
Native Hawaiian or other
  Pacific Islander                      1,351        57.3
Two or more minority races                121        68.6
Joint                                   3,398        53.5
Race missing                           33,385        60.9
Hispanic white                         20,820        55.1
Non-Hispanic white                    238,698        48.0

Sex
One male                              101,964        52.7
One female                             82,536        54.9
Two males                               4,483        50.3
Two females                             5,766        56.7

                                       Home purchase

                                    Adjusted denial rate,
                                    by adjustment factor

                                                Borrower-
       Race and ethnicity          Borrower-     related
            and sex                 related    plus lender

Race and ethnicity
American Indian or Alaska Native     55.9         52.7
Asian                                44.1         40.2
Black or African American            62.4         57.7
Native Hawaiian or other
  Pacific Islander                   55.2         50.2
Two or more minority races           63.5         74.7
Joint                                58.3         57.5
Race missing                         58.4         54.7
Hispanic white                       55.9         52.5
Non-Hispanic white                   48.0         48.0

Sex
One male                             52.7         52.7
One female                           52.3         52.1
Two males                            50.3         50.3
Two females                          53.4         49.9

                                           Refinance

                                      Number      Unadjusted
       Race and ethnicity               of          denial
            and sex                applications      rate

Race and ethnicity
American Indian or Alaska Native        1,670        58.0
Asian                                     641        49.5
Black or African American               9,535        63.9
Native Hawaiian or other
  Pacific Islander                        346        59.8
Two or more minority races                 73        56.2
Joint                                   1,549        50.7
Race missing                           39,237        64.2
Hispanic white                          6,760        51.0
Non-Hispanic white                    149,366        46.2

Sex
One male                               54,294        52.8
One female                             39,726        52.7
Two males                               1,271        43.2
Two females                             1,792        50.0

                                        Refinance

                                    Adjusted denial rate,
                                    by adjustment factor

                                                Borrower-
       Race and ethnicity          Borrower-     related
            and sex                 related    plus lender

Race and ethnicity
American Indian or Alaska Native     57.9         51.5
Asian                                48.4         51.6
Black or African American            61.6         53.8
Native Hawaiian or other
  Pacific Islander                   60.3         56.5
Two or more minority races           68.4         50.2
Joint                                50.4         47.9
Race missing                         68.5         52.7
Hispanic white                       51.3         49.3
Non-Hispanic white                   46.2         46.2

Sex
One male                             52.8         52.8
One female                           52.1         52.9
Two males                            43.2         43.2
Two females                          51.4         43.3

E. Nonowner-occupied site-built home

Percent except as noted

                                 Conventional, first lien

                                     Home purchase

                                              Adjusted denial rate,
                                              by adjustment factor

                                                          Borrower-
Race and ethnicity    Number     Unadjusted   Borrower-    related
     and sex         of appli-     denial      related      plus
                      cations       rate                   lender

Race and ethnicity
American Indian or
  Alaska Native          6,881         37.5        28.5        16.2
Asian                   48,246         15.0        13.3        12.9
Black or African
  American              80,051         24.1        19.6        17.1
Native Hawaiian or
  other Pacific
  Islander               5,127         18.2        15.6        13.3
Two or more
  minority races           600         14.8        13.1        10.8
Joint                   11,839         11.1        13.4        12.6
Race missing           123,905         18.2        17.2        14.2
Hispanic white          56,912         18.7        14.9        14.2
Non-Hispanic white     659,022         11.0        11.0        11.0

Sex
One male               318,461         17.2        17.2        17.2
One female             164,107         17.5        16.3        16.3
Two males               25,564         10.7        10.7        10.7
Two females              8,813         12.2         9.6        10.8

                               Conventional, first lien

                                       Refinance

                                                          Borrower-
Race and ethnicity    Number     Unadjusted   Borrower-    related
     and sex         of appli-     denial      related      plus
                      cations       rate                   lender
                                                          lender
Race and ethnicity
American Indian or
  Alaska Native          5,674         30.5        30.5        27.0
Asian                   30,317         20.4        22.3        22.2
Black or African
  American              87,234         30.6        28.6        25.4
Native Hawaiian or
  other Pacific
  Islander               4,401         24.8        25.2        26.4
Two or more
  minority races           661         24.4        28.3        30.6
Joint                    7,957         18.1        23.0        20.1
Race missing           132,928         32.5        30.9        26.2
Hispanic white          44,640         26.8        24.4        22.7
Non-Hispanic white     484,333         17.8        17.8        17.8

Sex
One male               264,494         24.5        24.5        24.5
One female             143,972         25.5        24.0        23.8
Two males               13,387         16.2        16.2        16.2
Two females              5,283         21.5        17.6        14.9

                               Conventional, first lien

                                          Other

                                              Adjusted denial rate,
                                              by adjustment factor

                                                          Borrower-
Race and ethnicity    Number     Unadjusted   Borrower-    related
     and sex         of appli-     denial      related      plus
                      cations       rate                   lender
                                                          lender
Race and ethnicity
American Indian or
  Alaska Native          2,183         49.7        44.5        38.9
Asian                    6,716         30.0        32.1        28.4
Black or African
  American              26,017         44.8        38.4        34.1
Native Hawaiian or
  other Pacific
  Islander               1,611         40.5        33.7        32.4
Two or more
  minority races           214         42.1        33.5        46.2
Joint                    2,293         24.7        31.9        27.5
Race missing            31,499         40.3        37.6        31.5
Hispanic white          14,508         37.9        34.9        31.8
Non-Hispanic white     133,098         25.5        25.5        25.5

Sex
One male                74,353         33.6        33.6        33.6
One female              43,267         39.3        34.6        33.6
Two males                4,006         24.0        24.0        24.0
Two females              1,881         35.3        29.5        24.6

NOTE: Includes transition-period applications (those submitted before
2004); for explanation, see text note 46. For explanation of
adjustment factors, see text. For method of allocation into racial
and ethnic categories and definitions of categories, see general
note to table 6 and text note 35. Applications made jointly by a male
and female are not tabulated here because they would not be directly
comparable with applications made by one applicant or by two applicants
of the same sex.

(1.) "Other" consists of government-backed loans of all types, junior
liens, home-improvement loans, and unsecured loans.

10. Unadjusted and adjusted incidence of higher-priced lending for
loans on one- to four-family homes, by type of loan and by race and
ethnicity and sex of borrower, 2004

A. Home purchase, owner-occupied site-built home

Percent except as noted

                                   Conventional

                                   First lien

                                              Adjusted incidence
                                              by adjustment factor

                                                          Borrower-
                       Number                             related
Race and ethnicity       of      Unadjusted   Borrower-    plus
  and sex              loans     incidence    related     lender

Race and ethnicity
American Indian or
  Alaska Native         28,107       18.1       17.2        11.8
Asian                  199,359        5.9        7.4         8.1
Black or African
  American             232,688       32.4       26.7        15.7
Native Hawaiian or
  other Pacific
  Islander              20,293       15.7       16.3        11.1
Two or more
  minority races         2,613       22.9       22.2        12.2
Joint                   47,299        6.9       10.8         9.4
Race missing           390,136       13.4       16.8        11.1
Hispanic white         301,915       20.3       16.6        11.6
Non-Hispanic white   2,476,255        8.7        8.7         8.7

Sex
One male             1,129,781       15.3       15.3        15.3
One female             850,213       15.3       14.4        15.0
Two males               38,170        9.5        9.5         9.5
Two females             31,083       10.4        9.0         9.8

                                  Conventional

                                      Junior lien

                                             Adjusted incidence
                                             by adjustment factor

                                                        Borrower-
                     Number                             related
Race and ethnicity     of      Unadjusted   Borrower-     plus
  and sex            loans     incidence    related      lender

Race and ethnicity
American Indian or
  Alaska Native        7,618      50.2        47.5        33.6
Asian                 32,444      34.4        30.8        29.9
Black or African
  American            62,434      61.9        58.3        37.1
Native Hawaiian or
  other Pacific
  Islander             6,195      49.6        44.9        33.5
Two or more
  minority races         699      43.2        44.3        26.6
Joint                  9,090      29.3        36.7        31.3
Race missing          96,140      40.7        43.5        33.2
Hispanic white        87,612      58.0        49.9        35.1
Non-Hispanic white   394,357      30.4        30.4        30.4

Sex
One male             237,097      46.1        46.1        46.1
One female           162,680      47.8        46.9        46.7
Two males              7,879      34.4        34.4        34.4
Two females            6,064      37.3        34.5        35.2

                               Government backed, first lien

                                             Adjusted incidence,
                                             by adjustment factor

                                                        Borrower-
                     Number                             related
Race and ethnicity     of      Unadjusted   Borrower-     plus
  and sex             loans    incidence     related     lender

Race and ethnicity
American Indian or
  Alaska Native        4,751       4.9         4.1         1.1
Asian                  5,402       2.5         1.7          .8
Black or African
  American            59,275       1.5         1.3         1.1
Native Hawaiian or
  other Pacific
  Islander             2,373       2.3         1.6         1.4
Two or more
  minority races         400      20.5        17.0         1.2
Joint                 10,035       1.1         1.1         1.4
Race missing          35,547       1.4         1.3         0.9
Hispanic white        47,055       2.1         1.1         1.1
Non-Hispanic white   304,809       1.1         1.1         1.1

Sex
One male             145,275       1.3         1.3         1.3
One female            98,428       1.6         1.5         1.3
Two males             10,094       1.2         1.2         1.2
Two females            6,716       1.2         1.3         1.1

B. Refinance, owner-occupied site-built home

Percent except as noted

                                       Conventional

                           First lien

                                               Adjusted incidence,
                                              by adjustment factor

                                                          Borrower-
Race and ethnicity    Number     Unadjusted   Borrower-    related
     and sex            of       incidence     related      plus
                       loans                               lender

Race and ethnicity
American Indian or
  Alaska Native         44,503      20.2        21.0        14.7
Asian                  207,114       5.9         9.7        12.1
Black or African
American               391,524      34.6        29.5        17.6
Native Hawaiian or
  other Pacific
  Islander              31,381      16.4        18.6        14.5
Two or more
  minority races         5,089      21.1        22.4        15.0
Joint                   67,199      10.4        14.7        13.5
Race missing           827,590      19.3        25.4        15.3
Hispanic white         378,826      19.3        18.5        14.3
Non-Hispanic white   3,698,309      12.9        12.9        12.9

Sex
One male             1,360,350      18.6        18.6        18.6
One female           1,173,835      19.8        18.5        18.7
Two males               40,012      12.1        12.1        12.1
Two females             43,208      17.3        14.5        13.4

                                     Conventional

                         Junior lien

                                            Adjusted incidence,
                                            by adjustment factor

                                                        Borrower-
Race and ethnicity   Number    Unadjusted   Borrower-    related
     and sex           of      incidence     related      plus
                      loans                              lender

Race and ethnicity
American Indian or
  Alaska Native        2,981      26.8        24.4        26.1
Asian                 10,519      25.4        24.9        24.3
Black or African
American              24,292      45.2        41.3        26.4
Native Hawaiian or
  other Pacific
  Islander             2,267      35.8        33.2        24.6
Two or more
  minority races         394      19.3        25.4        24.7
Joint                  5,609      23.3        26.5        24.8
Race missing          82,329      32.7        42.1        24.8
Hispanic white        20,687      35.1        28.7        25.2
Non-Hispanic white   285,505      23.9        23.9        23.9

Sex
One male              90,991      34.0        34.0        34.0
One female            67,266      32.6        32.6        33.9
Two males              3,024      23.6        23.6        23.6
Two females            2,887      31.7        26.6        22.6

                            Government backed, first lien

                                            Adjusted incidence,
                                            by adjustment factor

                                                        Borrower-
Race and ethnicity   Number    Unadjusted   Borrower-    related
     and sex           of      incidence     related      plus
                      loans                              lender

Race and ethnicity
American Indian or
  Alaska Native        2,216       4.4         4.2         1.0
Asian                  2,348       3.9         2.0         1.0
Black or African
American              46,603       1.0         1.2         1.5
Native Hawaiian or
  other Pacific
  Islander             1,547       6.1         4.1         1.4
Two or more
  minority races         347      18.7        19.6         2.3
Joint                  6,100        .9         1.1         1.2
Race missing          30,603       2.0         2.0         1.3
Hispanic white        21,804       1.9         1.2         1.4
Non-Hispanic white   152,619       1.3         1.3         1.3

Sex
One male              63,536       1.4         1.4         1.4
One female            49,282       1.7         1.6         1.4
Two males              3,103       2.2         2.2         2.2
Two females            3,053       1.8         1.5         1.6

C. Home improvement, conventional loan, owner-occupied site-built
home

Percent except as noted

                                   First lien

                                              Adjusted incidence
                                              by adjustment factor

                                                        Borrower-
                      Number                            related
Race and ethnicity     of      Unadjusted   Borrower-    plus
  and sex             loans    incidence    related     lender

Race and ethnicity
American Indian or
  Alaska Native        3,775     23.4         25.8       22.5
Asian                  8,907      8.4         16.3       18.8
Black or African
  American            27,677     42.5         39.2       24.6
Native Hawaiian or
  other Pacific
  Islander             2,237     17.7         22.4       21.9
Two or more
  minority races         341     22.3         22.6       18.6
Joint                  3,894     15.5         20.7       19.2
Race missing          36,877     21.5         32.2       21.4
Hispanic white        28,395     26.7         27.6       21.0
Non-Hispanic white   221,213     29.5         19.5       19.5

Sex
One male              74,062     26.1         26.1       26.1
One female            74,479     27.0         25.8       26.5
Two males              2,545     17.2         17.2       17.2
Two females            3,053     23.9         18.5       21.2

                                Junior lien

                                             Adjusted incidence
                                             by adjustment factor

                                                        Borrower-
                     Number                             related
Race and ethnicity     of      Unadjusted   Borrower-     plus
  and sex            loans     incidence    related      lender

Race and ethnicity
American Indian or
  Alaska Native        2,628      18.8        17.4        17.5
Asian                  6,467      13.9        15.1        16.6
Black or African
  American            19,520      37.9        31.3        18.0
Native Hawaiian or
  other Pacific
  Islander             1,749      26.5        22.6        16.9
Two or more
  minority races         193      16.6        15.1        14.1
Joint                  5,280      15.2        18.3        16.3
Race missing          61,005      17.4        23.1        15.9
Hispanic white        15,464      21.0        20.0        16.5
Non-Hispanic white   259,771      15.6        15.6        15.6

Sex
One male              71,519      24.4        24.4        24.4
One female            57,771      22.3        21.4        23.6
Two males              2,390      19.1        19.1        19.1
Two females            2,807      19.0        18.1        19.2

D. Manufactured housing, conventional loan, first lien, owner-occupied
home

Percent except as noted

                                          Home purchase

                                                  Adjusted incidence,
                                                 by adjustment factor

                                                              Borrower-
                            Number      Un-                    related
    Race and ethnicity        of     adjusted    Borrower-       plus
         and sex            loans    incidence    related       lender

Race and ethnicity
American Indian or
  Alaska Native               1,168     68.2        58.3        54.8
Asian                           537     62.8        62.6        61.9
Black or African American     5,175     78.5        64.8        60.0
Native Hawaiian or other
  Pacific Islander              315     70.8        69.2        65.3
Two or more minority
  races                          21     47.6        51.2        57.0
Joint                           918     58.3        56.8        60.3
Race missing                  6,535     57.2        56.6        55.4
Hispanic white                5,598     67.5        65.2        60.7
Non-Hispanic white           76,324     54.7        54.7        54.7

Sex
One male                     28,100     61.0        61.0        61.0
One female                   22,177     60.5        60.2        59.8
Two males                     1,566     65.1        65.1        65.1
Two females                   1,547     65.0        64.6        65.1

                                            Refinance

                                                  Adjusted incidence,
                                                  by adjustment factor

                                                             Borrower-
                            Number                             related
    Race and ethnicity       of     Unadjusted   Borrower-      plus
         and sex            loans    incidence    related      lender

Race and ethnicity
American Indian or
  Alaska Native                513     51.9         53.5        50.6
Asian                          210     42.9         47.8        56.0
Black or African American    2,316     67.7         56.1        50.1
Native Hawaiian or other
  Pacific Islander             102     48.0         44.5        45.5
Two or more minority
  races                         26     42.3         54.3        -9.8
Joint                          568     41.2         49.3        48.7
Race missing                 7,420     59.0         51.6        46.3
Hispanic white               2,274     46.2         56.7        50.0
Non-Hispanic white          56,713     46.0         46.0        46.0

Sex
One male                    17,181     48.6         48.6        48.6
One female                  12,772     49.8         49.0        49.2
Two males                      560     56.3         56.3        56.3
Two females                    622     47.3         50.8        53.6

E. Nonowner-occupied site-built home

Percent except as noted

                            Conventional, first lien

                                  Home purchase

                                             Adjusted incidence,
                                            by adjustment factor

                                                        Borrower-
                     Number                              related
Race and ethnicity     of      Unadjusted   Borrower-     plus
     and sex          loans    incidence     related     lender

Race and ethnicity
American Indian or
  Alaska Native        3,576      16.9        15.6        11.1
Asian                 33,592       7.4         8.5         8.6
Black or African
  American            46,260      38.6        24.2        14.7
Native Hawaiian or
  other Pacific
  Islander             3,407      15.8        14.6        11.2
Two or more
  minority races         424      25.7        23.5        11.4
Joint                  8,971       7.1        11.0         9.9
Race missing          76,434      11.5        14.1        10.6
Hispanic white        40,643      16.8        15.4        10.9
Non-Hispanic white   502,030       9.4         9.4         9.4

Sex
One male             219,700      17.7        17.7        17.7
One female           111,338      17.4        16.9        16.9
Two males             19,437       8.3         8.3         8.3
Two females            6,428       7.6         6.8         7.9

                            Conventional, first lien

                                    Refinance

                                             Adjusted incidence,
                                            by adjustment factor

                                                        Borrower-
                     Number                              related
Race and ethnicity     of      Unadjusted   Borrower-     plus
     and sex          loans    incidence     related     lender

Race and ethnicity
American Indian or
  Alaska Native        3,166      19.7        21.5        14.1
Asian                 18,809       7.0        10.9        11.3
Black or African
  American            46,900      33.5        27.7        16.7
Native Hawaiian or
  other Pacific
  Islander             2,666      14.6        17.3        13.0
Two or more
  minority races         424      24.8        27.5        15.9
Joint                  5,421       8.2        13.5        11.3
Race missing          64,341      15.5        22.2        13.5
Hispanic white        28,011      17.9        18.1        13.3
Non-Hispanic white   330,979      10.9        10.9        10.9

Sex
One male             161,011      17.9        17.9        17.9
One female            85,278      19.8        19.8        18.4
Two males              9,083       7.9         7.9         7.9
Two females            3,307      12.0         9.3        10.2

                                   Other (1)

                                                       Borrower-
                     Number                             related
Race and ethnicity     of     Unadjusted   Borrower-     plus
     and sex         loans    incidence     related     lender

Race and ethnicity
American Indian or
  Alaska Native         828      34.8        25.2        31.2
Asian                 3,785      37.5        32.1        30.4
Black or African
  American           11,099      46.8        43.2        33.6
Native Hawaiian or
  other Pacific
  Islander              731      40.8        33.5        29.5
Two or more
  minority races         97      38.1        19.9        55.4
Joint                 1,466      27.2        31.8        28.6
Race missing         13,877      41.1        40.0        32.0
Hispanic white        7,522      49.4        40.1        34.5
Non-Hispanic white   81,408      30.2        30.2        30.2

Sex
One male             39,283      42.5        42.5        42.5
One female           20,416      43.0        42.5        43.1
Two males             2,526      27.2        27.2        27.2
Two females             984      35.6        33.5        32.0

(1.) "Other" consists of government-backed loans of all types,
junior liens, home-improvement loans, and unsecured loans.

NOTE: Excludes transition-period loans (those for which the
application was submitted before 2004). For definition of higher-priced
lending and explanation of adjustment factors, see text. For method of
allocation into racial and ethnic categories and definitions of
categories, see general note to table 6 and text note 35. Loans taken
out jointly by a male and female are not tabulated here because they
would not be directly comparable with loans taken out by one
borrower or by two borrowers of the same sex.

11. Unadjusted and adjusted incidence of higher-priced lending for
conventional, first-lien loans on site-built, owner-occupied, one-
to four-family homes, by selected race and ethnicity for selected
characteristics of borrower, property, and lender, 2004

Percent except as noted

                                            Home purchase

                                                            Adjusted
                                                           incidence,
                                                               by
                                                           adjustment
                                                             factor

   Selected characteristic,        Number
     by race and ethnicity           of       Unadjusted   Borrower-
          of borrower               loans     incidence     related

      INCOME OF BORROWER

Lower
Black or African American            87,841      39.2         35.2
Hispanic white                       83,642      23.6         21.2
Non-Hispanic white                  637,019      12.9         12.9

Middle
Black or African American            66,997      33.9         31.1
Hispanic white                       84,988      22.1         19.5
Non-Hispanic white                  648,853       9.9          9.9

Higher
Black or African American            70,247      23.9         19.8
Hispanic white                      118,728      17.4         12.5
Non-Hispanic white                1,099,544       5.8          5.8

    INCOME OF CENSUS TRACT

Lower
Black or African American            69,713      40.9         34.7
Hispanic white                       85,791      26.0         25.0
Non-Hispanic white                  254,231      15.0         15.0

Middle
Black or African American           107,177      32.0         29.0
Hispanic white                      137,291      20.5         17.7
Non-Hispanic white                1,229,899      10.3         10.3

Higher
Black or African American            55,401      22.7         18.2
Hispanic white                       78,431      13.6         10.5
Non-Hispanic white                  984,982       5.2          5.2

      CRA ASSESSMENT AREA

Property not in assessment area
Black or African American            78,331      31.9         25.5
Hispanic white                      100,290      16.7         14.3
Non-Hispanic white                1,010,824       7.4          7.4

Property in assessment area
Black or African American            43,680      12.7          9.9
Hispanic white                       68,101       9.6          5.6
Non-Hispanic white                  620,870       2.8          2.8

          DISPOSITION

Sold to GSE (1)
Black or African American            43,683       9.0          6.7
Hispanic white                       67,768       3.9          3.2
Non-Hispanic white                  808,454       1.6          1.6

Sold to others
Black or African American           137,679      40.5         33.3
Hispanic white                      160,188      26.3         24.0
Non-Hispanic white                1,065,255      13.8         13.8

Retained
Black or African American            51,326      30.7         25.6
Hispanic white                       73,959      22.2         17.5
Non-Hispanic white                  602,546       9.3          9.3

        TYPE OF LENDER,
       BY PERCENTAGE OF
       LENDER'S LOANS TO
      NON-HISPANIC WHITES
    THAT ARE HIGHER PRICED

Less than 10
Black or African American           121,500      10.6          8.5
Hispanic white                      182,503       5.7          4.8
Non-Hispanic white                1,993,900       2.3          2.3

10-49
Black or African American            69,894      44.3         40.9
Hispanic white                       78,695      31.0         33.0
Non-Hispanic white                  366,709      25.3         25.3

50 or more
Black or African American            41,189      76.6         75.6
Hispanic white                       40,393      65.4         70.7
Non-Hispanic white                  115,646      66.5         66.5

                                     Home
                                   purchase           Refinance

                                   Adjusted
                                  incidence,
                                      by
                                  adjustment
                                    factor

   Selected characteristic,        Borrower-     Number
     by race and ethnicity          related        of       Unadjusted
          of borrower             plus lender     loans     incidence

      INCOME OF BORROWER

Lower
Black or African American            21.7         161,762      42.1
Hispanic white                       16.3         120,253      22.5
Non-Hispanic white                   12.9         961,571      19.3

Middle
Black or African American            18.1         112,007      34.8
Hispanic white                       13.3         118,699      20.8
Non-Hispanic white                    9.9         989,420      14.9

Higher
Black or African American            10.9         104,662      26.0
Hispanic white                        7.8         123,101      16.9
Non-Hispanic white                    5.8       1,539,336       9.0

    INCOME OF CENSUS TRACT

Lower
Black or African American            22.7         141,851      42.5
Hispanic white                       18.8         126,577      23.3
Non-Hispanic white                   15.0         383,904      21.7

Middle
Black or African American            17.8         175,784      33.4
Hispanic white                       13.0         167,456      19.2
Non-Hispanic white                   10.3       1,936,163      15.0

Higher
Black or African American             9.8          73,571      22.2
Hispanic white                        7.0          84,151      13.6
Non-Hispanic white                    5.2       1,370,653       7.4

      CRA ASSESSMENT AREA

Property not in assessment area
Black or African American            17.7         121,253      28.3
Hispanic white                       11.1         105,634      13.6
Non-Hispanic white                    7.4       1,419,662       8.8

Property in assessment area
Black or African American             6.6          79,403      13.3
Hispanic white                        4.0         113,714       7.7
Non-Hispanic white                    2.8         964,166       3.9

          DISPOSITION

Sold to GSE (1)
Black or African American             5.2          71,439       1.1
Hispanic white                        2.9          87,929        .4
Non-Hispanic white                    1.6       1,200,878        .4

Sold to others
Black or African American            20.6         196,055      43.6
Hispanic white                       17.5         179,549      27.3
Non-Hispanic white                   13.8       1,427,539      20.0

Retained
Black or African American            13.2         124,030      39.6
Hispanic white                       10.5         111,348      21.4
Non-Hispanic white                    9.3       1,069,892      17.4

        TYPE OF LENDER,
       BY PERCENTAGE OF
       LENDER'S LOANS TO
      NON-HISPANIC WHITES
    THAT ARE HIGHER PRICED

Less than 10
Black or African American             8.4         171,423       6.6
Hispanic white                        4.5         212,021       2.6
Non-Hispanic white                    2.3       2,589,887       1.8

10-49
Black or African American            35.9         139,090      46.1
Hispanic white                       31.2         111,898      32.1
Non-Hispanic white                   25.3         763,119      27.3

50 or more
Black or African American            74.8          80,743      74.4
Hispanic white                       71.6          54,639      58.2
Non-Hispanic white                   66.5         345,303      64.2

                                       Refinance

                                    Adjusted incidence,
                                    by adjustment factor

   Selected characteristic,                    Borrower-
     by race and ethnicity        Borrower-     related
          of borrower              related    plus lender

      INCOME OF BORROWER

Lower
Black or African American           39.6         25.9
Hispanic white                      24.2         20.4
Non-Hispanic white                  19.3         19.3

Middle
Black or African American           33.8         20.1
Hispanic white                      21.5         16.6
Non-Hispanic white                  14.9         14.9

Higher
Black or African American           22.8         12.7
Hispanic white                      14.6         10.4
Non-Hispanic white                   9.0          9.0

    INCOME OF CENSUS TRACT

Lower
Black or African American           37.5         27.4
Hispanic white                      25.6         22.9
Non-Hispanic white                  21.7         21.7

Middle
Black or African American           31.2         19.5
Hispanic white                      19.6         16.0
Non-Hispanic white                  15.0         15.0

Higher
Black or African American           18.1         10.2
Hispanic white                      11.2          8.2
Non-Hispanic white                   7.4          7.4

      CRA ASSESSMENT AREA

Property not in assessment area
Black or African American           23.3         14.8
Hispanic white                      12.6         10.3
Non-Hispanic white                   8.8          8.8

Property in assessment area
Black or African American           10.9          6.4
Hispanic white                       5.3          4.6
Non-Hispanic white                   3.9          3.9

          DISPOSITION

Sold to GSE (1)
Black or African American            1.0          .7
Hispanic white                       .4           .4
Non-Hispanic white                   .4           .4

Sold to others
Black or African American           36.7         25.9
Hispanic white                      27.9         22.1
Non-Hispanic white                  20.0         20.0

Retained
Black or African American           34.3         21.0
Hispanic white                      21.7         18.5
Non-Hispanic white                  17.4         17.4

        TYPE OF LENDER,
       BY PERCENTAGE OF
       LENDER'S LOANS TO
      NON-HISPANIC WHITES
    THAT ARE HIGHER PRICED

Less than 10
Black or African American            5.1          4.3
Hispanic white                       2.5          2.3
Non-Hispanic white                   1.8          1.8

10-49
Black or African American           42.2         36.9
Hispanic white                      32.9         30.6
Non-Hispanic white                  27.3         27.3

50 or more
Black or African American           71.8         71.7
Hispanic white                      65.8         67.1
Non-Hispanic white                  64.2         64.2

NOTE: For discussion of CRA assessment areas and the identification
of higher-priced lenders, see text note 47. See also general note
to table 10.A.

(1.) Government-sponsored enterprise; virtually all were to Fannie
Mae and Freddie Mac.

12. Unadjusted and adjusted mean APR spreads for higher-priced loans
on one- to four-family homes, by type of loan and by race and
ethnicity and sex of borrower, 2004

A. Home purchase, owner-occupied site-built home

Percent except as noted

                               Conventional

                                First lien

                                                Adjusted
                                               mean spread,
                                              by adjustment
                                                 factor

                     Number
                       of      Unadjusted               Borrower-
Race and ethnicity   higher-      mean      Borrower-    related
     and sex         priced      spread      related      plus
                      loans                              lender

Race and ethnicity
American Indian or
  Alaska Native        5,101      4.0          4.1         4.1
Asian                 11,771      3.8          4.0         4.0
Black or African
  American            75,427      4.2          4.2         4.2
Native Hawaiian or
  other Pacific
  Islander             3,186      4.0          4.1         4.1
Two or more
  minority races         598      4.1          4.3         4.1
Joint                  3,242      4.0          4.1         4.1
Race missing          52,094      4.1          4.1         4.1
Hispanic white        61,248      3.9          4.0         4.1
Non-Hispanic white   216,409      4.1          4.1         4.1

Sex
One male             173,166      4.0          4.0         4.0
One female           130,250      4.1          4.0         4.0
Two males              3,632      4.1          4.1         4.1
Two females            3,246      4.1          4.0         4.1

                                 Conventional

                                 Junior lien

                                                Adjusted
                                               mean spread,
                                              by adjustment
                                                 factor

                     Number
                       of      Unadjusted               Borrower-
Race and ethnicity   higher-      mean      Borrower-    related
     and sex         priced      spread      related      plus
                      loans                              lender

Race and ethnicity
American Indian or
  Alaska Native        3,828      6.3          6.4         6.4
Asian                 11,164      6.2          6.4         6.4
Black or African
  American            38,657      6.6          6.5         6.5
Native Hawaiian or
  other Pacific
  Islander             3,070      6.3          6.5         6.5
Two or more
  minority races         302      6.4          6.5         6.5
Joint                  2,662      6.4          6.5         6.5
Race missing          39,130      6.4          6.5         6.5
Hispanic white        50,817      6.3          6.4         6.4
Non-Hispanic white   119,879      6.4          6.4         6.4

Sex
One male             109,257      6.4          6.4         6.4
One female            77,785      6.4          6.4         6.4
Two males              2,711      6.3          6.3         6.3
Two females            2,261      6.4          6.3         6.3

                          Government backed, first lien

                                                Adjusted
                                               mean spread,
                                              by adjustment
                                                 factor

                     Number
                       of      Unadjusted               Borrower-
Race and ethnicity   higher-      mean      Borrower-    related
     and sex         priced      spread      related      plus
                      loans                              lender

Race and ethnicity
American Indian or
  Alaska Native          235      4.2          4.0         4.1
Asian                    137      4.3          4.2         4.2
Black or African
  American               911      4.1          4.0         4.2
Native Hawaiian or
  other Pacific
  Islander                55      4.3          4.3         4.5
Two or more
  minority races          82      3.9          3.6         4.1
Joint                    114      4.2          4.1         4.2
Race missing             515      4.2          4.2         4.1
Hispanic white           973      4.7          4.2         4.2
Non-Hispanic white     3,214      4.1          4.1         4.1

Sex
One male               1,924      4.2          4.2         4.2
One female             1,555      4.2          4.3         4.3
Two males                126      4.1          4.1         4.1
Two females               83      4.0          3.8         3.9

Note: For definition of APR spread, see table 7, note 1.
See also note to table 10.A.

B. Refinance, owner-occupied site-built home

Percent except as noted

                                    Conventional

                                     First lien

                                                   Adjusted
                                                  mean spread,
                                                 by adjustment
                                                    factor

                        Number
                          of      Unadjusted               Borrower-
 Race and ethnicity     higher-      mean      Borrower-    related
       and sex          priced      spread      related      plus
                         loans                              lender

Race and ethnicity
American Indian or
  Alaska Native           8,977      4.1          4.2         4.1
Asian                    12,250      3.9          4.1         4.1
Black or African
  American              135,467      4.3          4.3         4.3
Native Hawaiian or
  other Pacific
  Islander                5,153      4.1          4.2         4.2
Two or more
  minority races          1,072      4.0          4.1         4.1
Joint                     6,973      4.1          4.2         4.2
Race missing            159,741      4.2          4.2         4.2
Hispanic white           73,181      4.0          4.1         4.2
Non-Hispanic white      476,034      4.2          4.2         4.2

Sex
One male                252,618      4.1          4.1         4.1
One female              232,583      4.2          4.2         4.1
Two males                 4,833      4.2          4.2         4.2
Two females               7,479      4.3          4.2         4.2

                                     Conventional

                                      Junior lien

                                                   Adjusted
                                                  mean spread,
                                                 by adjustment
                                                    factor

                        Number
                          of      Unadjusted               Borrower-
 Race and ethnicity     higher-      mean      Borrower-    related
       and sex          priced      spread      related      plus
                         loans                              lender

Race and ethnicity
American Indian or
  Alaska Native             800      7.2          6.8         7.1
Asian                     2,675      6.5          6.9         7.1
Black or African
  American               10,974      7.5          7.2         7.2
Native Hawaiian or
  other Pacific
  Islander                  811      6.8          7.1         7.1
Two or more
  minority races             76      6.8          7.0         7.3
Joint                     1,308      7.0          7.1         7.1
Race missing             26,915      7.6          7.3         7.2
Hispanic white            7,255      6.9          7.0         7.1
Non-Hispanic white       68,211      7.1          7.1         7.1

Sex
One male                 30,905      7.1          7.1         7.1
One female               21,919      7.1          7.1         7.1
Two males                   714      7.0          7.0         7.0
Two females                 914      7.2          6.9         7.0

                           Government backed, first lien

                                                   Adjusted
                                                  mean spread,
                                                 by adjustment
                                                    factor

                        Number
                          of      Unadjusted               Borrower-
 Race and ethnicity     higher-      mean      Borrower-    related
       and sex          priced      spread      related      plus
                         loans                              lender

Race and ethnicity
American Indian or
  Alaska Native              98      4.1          4.0         3.8
Asian                        91      4.4          3.7         3.8
Black or African
  American                  456      3.7          3.8         3.8
Native Hawaiian or
  other Pacific
  Islander                   94      5.1          3.7         3.6
Two or more
  minority races             65      3.9          3.9         3.9
Joint                        56      3.9          3.8         3.6
Race missing                603      3.8          3.9         3.9
Hispanic white              408      4.4          4.0         3.9
Non-Hispanic white        2,002      3.8          3.8         3.8

Sex
One male                    895      3.9          3.9         3.9
One female                  823      3.9          3.9         3.8
Two males                    68      4.1          4.1         4.1
Two females                  55      3.9          4.0         4.3

Note: See note to table 12.A.

C. Home improvement, conventional loan, owner-occupied site-built home

Percent except as noted

                                   First lien

                                                  Adjusted
                                                 mean spread,
                                                by adjustment
                                                   factor

                     Number
                       of      Unadjusted                Borrower-
Race and ethnicity   higher-      mean      Borrower-     related
     and sex         priced      spread      related    plus lender
                      loans

Race and ethnicity

American Indian or
  Alaska Native          883      4.5          4.6          4.4
Asian                    752      3.9          4.3          4.4
Black or African      11,770      4.6          4.6          4.5
  American
Native Hawaiian or
  other Pacific
  Islander               397      4.2          4.4          4.5
Two or more
  minority races          76      4.4          4.3          4.3
Joint                    605      4.4          4.5          4.4
Race missing           7,920      4.5          4.6          4.5
Hispanic white         7,572      4.3          4.3          4.4
Non-Hispanic white    43,209      4.4          4.4          4.4

Sex
One male              19,340      4.4          4.4          4.4
One female            20,104      4.4          4.4          4.4
Two males                438      4.3          4.3          4.3
Two females              729      4.6          4.4          4.1

                                   Junior lien

                                                  Adjusted
                                                 mean spread,
                                                by adjustment
                                                   factor

                     Number
                       of      Unadjusted                Borrower-
Race and ethnicity   higher-      mean      Borrower-     related
     and sex         priced      spread      related    plus lender
                      loans

Race and ethnicity

American Indian or
  Alaska Native          493      7.7          7.6          7.8
Asian                    875      7.5          7.8          7.9
Black or African       7,389      8.4          8.3          8.1
  American
Native Hawaiian or
  other Pacific
  Islander               463      7.9          8.0          7.7
Two or more
  minority races          32      7.7          7.9          6.6
Joint                    800      7.8          8.1          8.2
Race missing          10,633      7.9          8.2          8.0
Hispanic white         3,255      7.7          8.1          8.0
Non-Hispanic white    40,410      7.9          7.9          7.9

Sex
One male              17,425      8.2          8.2          8.2
One female            12,873      8.1          8.1          8.1
Two males                457      7.9          7.9          7.9
Two females              532      8.0          8.0          7.9

D. Manufactured housing, conventional loan, first lien,
owner-occupied home

Percent except as noted

                                  Home purchase

                                                  Adjusted
                                                 mean spread,
                                                by adjustment
                                                   factor

                     Number
                       of      Unadjusted                Borrower-
Race and ethnicity   higher-      mean      Borrower-     related
     and sex         priced      spread      related    plus lender
                      loans

Race and ethnicity
American Indian or
  Alaska Native          796      6.0          6.0          5.9
Asian                    337      5.4          5.5          5.5
Black or African
  American             4,060      6.0          5.9          5.9
Native Hawaiian or
  other Pacific
  Islander               223      5.8          6.1          6.6
Two or more
  minority races          10      5.6          6.8          5.3
Joint                    535      5.6          5.8          5.7
Race missing           3,738      5.9          5.9          5.7
Hispanic white         3,777      6.0          6.1          6.0
Non-Hispanic white    41,725      5.6          5.6          5.6

Sex
One male              17,145      5.8          5.8          5.8
One female            13,413      5.7          5.7          5.7
Two males              1,019      5.9          5.9          5.9
Two females            1,005      5.7          5.6          6.0

                                       Refinance

                                                  Adjusted
                                                 mean spread,
                                                by adjustment
                                                   factor

                     Number
                       of      Unadjusted                Borrower-
Race and ethnicity   higher-      mean      Borrower-     related
     and sex         priced      spread      related    plus lender
                      loans

Race and ethnicity
American Indian or
  Alaska Native          266      5.4          5.3          4.9
Asian                     90      4.9          5.7          5.7
Black or African
  American             1,567      5.4          5.3          5.3
Native Hawaiian or
  other Pacific
  Islander                49      4.8          5.0          5.1
Two or more
  minority races          11      3.9          4.0          3.2
Joint                    234      4.9          5.1          5.1
Race missing           4,379      4.8          4.7          4.9
Hispanic white         1,051      5.0          5.0          5.1
Non-Hispanic white    26,104      5.0          5.0          5.0

Sex
One male               8,350      5.0          5.0          5.0
One female             6,363      5.1          5.0          5.0
Two males                315      5.4          5.4          5.4
Two females              294      5.4          5.3          5.7

NOTE: See note to table 12.A.

E. Nonowner-occupied site-built home

Percent except as noted

                           Conventional, first lien

                                Home purchase

                                                  Adjusted
                                                 mean spread,
                                                by adjustment
                                                   factor

                     Number
                       of      Unadjusted               Borrower-
Race and ethnicity   higher-      mean      Borrower-    related
     and sex         priced      spread      related      plus
                      loans                              lender

Race and ethnicity
American Indian or
  Alaska Native          603      3.9          4.1         4.1
Asian                  2,476      3.9          3.9         4.0
Black or African
  American            17,841      4.2          4.1         4.1
Native Hawaiian or
  other Pacific
  Islander               539      4.0          4.0         3.9
Two or more
  minority races         109      4.1          4.4         4.1
Joint                    637      4.0          4.0         3.9
Race missing           8,816      4.1          4.1         4.1
Hispanic white         6,831      3.9          4.0         4.0
Non-Hispanic white    47,211      4.0          4.0         4.0

Sex
One male              38,917      4.0          4.0         4.0
One female            19,349      4.0          4.0         4.0
Two males              1,605      4.1          4.1         4.1
Two females              489      4.1          4.2         5.4

                              Conventional, first lien

                                       Refinance

                                                  Adjusted
                                                 mean spread,
                                                by adjustment
                                                   factor

                     Number
                       of      Unadjusted               Borrower-
Race and ethnicity   higher-      mean      Borrower-    related
     and sex         priced      spread      related      plus
                      loans                              lender

Race and ethnicity
American Indian or
  Alaska Native          623      4.1          4.3         4.3
Asian                  1,319      3.9          4.0         4.1
Black or African
  American            15,731      4.3          4.3         4.3
Native Hawaiian or
  other Pacific
  Islander               388      4.2          4.4         4.8
Two or more
  minority races         105      3.9          3.9         3.7
Joint                    444      4.2          4.2         4.0
Race missing           9,995      4.2          4.3         4.2
Hispanic white         5,002      4.1          4.2         4.2
Non-Hispanic white    35,920      4.1          4.1         4.1

Sex
One male              28,824      4.2          4.2         4.2
One female            16,904      4.2          4.2         4.2
Two males                718      4.1          4.1         4.1
Two females              396      4.4          4.1         4.0

                                        Other (1)

                                                  Adjusted
                                                 mean spread,
                                                by adjustment
                                                   factor

                     Number
                       of      Unadjusted               Borrower-
Race and ethnicity   higher-      mean      Borrower-    related
     and sex         priced      spread      related      plus
                      loans                              lender

Race and ethnicity
American Indian or
  Alaska Native          288      6.5          6.2         6.3
Asian                  1,419      6.5          6.2         6.1
Black or African
  American             5,197      6.5          6.3         6.2
Native Hawaiian or
  other Pacific
  Islander               298      6.7          6.3         6.2
Two or more
  minority races          37      6.7          6.2         6.7
Joint                    399      6.5          6.4         6.0
Race missing           5,698      6.7          6.3         6.3
Hispanic white         3,718      6.3          6.2         6.1
Non-Hispanic white    24,622      6.1          6.1         6.1

Sex
One male              16,711      6.4          6.4         6.4
One female             8,774      6.3          6.3         6.4
Two males                688      6.3          6.3         6.3
Two females              350      6.5          6.2         6.4

Note: See note to table 12.A.

(1.) See table 10.E., note 1.

13. Loan pricing by eight subprime specialists, conventional
first-lien loans on owner-occupied one- to four-family homes, by
type of loan and by selected race and ethnicity of borrower, 2004

A. Unadjusted and adjusted incidence of higher-priced lending

Percent except as noted

                                    Home purchase

                                              Adjusted incidence
                                              by adjustment facor

                                                         Borrower-
                                                          related
                      Number    Unadjusted   Borrower-      plus
Race and ethnicity   of loans    incidence    related     non-HMDA
                                                           credit
                                                          factors

Black or African
  American             6,369       85.8         84.4        83.9
Hispanic white         6,110       72.6         83.2        83.1
Non-Hispanic white    30,224       82.9         82.9        82.9

                                      Refinance

                                              Adjusted incidence
                                              by adjustment factor

                                                         Borrower-
                                                          related
                      Number    Unadjusted   Borrower-      plus
Race and ethnicity   of loans    incidence    related     non-HMDA
                                                           credit
                                                          factors

Black or African
  American             37,354      82.9         81.9        80.7
Hispanic white         12,800      71.7         78.8        80.9
Non-Hispanic white    135,667      80.7         80.7        80.7

NOTE: Includes transition-period loans (those for which the
application was submitted before 2004). For details on higher-priced
lending, the subprime lenders, and the adjustment factors, see text
and text note 49. For method of allocation into racial and ethnic
categories and definitions of categories, see general note to table
6 and text note 35.

Source: Credit Research Center, Georgetown University.

B. Unadjusted and adjusted mean APR spreads for higher-priced loans

Percentage points except as noted

                                    Home purchase

                                              Adjusted spread
                                            by adjustment factor

                                                        Borrower-
                      Number                             related
                        of     Unadjusted   Borrower-      plus
Race and ethnicity   higher-     spread      related     non-HMDA
                      priced                              credit
                      loans                              factors

Black or African
  American             5,463       5.1         5.1         5.0
Hispanic white         4,437       4.7         4.9         4.9
Non-Hispanic white    25,053       4.9         4.9         4.9

                                     Refinance

                                              Adjusted spread
                                            by adjustment factor

                                                        Borrower-
                      Number                             related
                        of     Unadjusted   Borrower-      plus
Race and ethnicity   higher-     spread      related     non-HMDA
                      priced                              credit
                      loans                              factors

Black or African
  American            30,959       5.1         5.1         5.0
Hispanic white         9,183       4.7         4.8         4.9
Non-Hispanic white   109,524       5.0         5.0         5.0

NOTE: For definition of APR spread, see table 7, note 1. See also
notes to table 13.A.

14. Distribution of the difference between denial rates on
applications by black and Hispanic applicants as a group and such
denial rates for non-Hispanic white applicants, by lender-product
combination, 2004

Percent except as noted

                                           Distribution of percentage
                                                of lender-product
                                                combinations after
                                                   adjustment

                                           Black and Hispanic denial
                                               rate is higher and
                                          difference is statistically
                                           significant, by degree of
                                                  significance

                            Number of
  Type of unadjusted     lender-product      1         5         10
      difference          combinations    percent   percent   percent
                             before
                           adjustment

Black and Hispanic
denial rate is higher
and difference is
statistically
significant, by degree
of significance
  1 percent                   1,642         43.8      16.4      8.6
  5 percent                     758          4.1      12.4      9.5
  10 percent                    557           .7       5.7      8.8

Difference is not
  statistically
  significant                10,185           .3      48.0      1.0

Non-Hispanic white
denial rate is higher
and difference is
statistically
significant, by degree
of significance
  1 percent                      48          2.1       2.1       0
  5 percent                      34           0         0        0
  10 percent                     36           0         0        0

                        Distribution of percentage of lender-product
                               combinations after adjustment

                                          Non-Hispanic
                                          white denial
                          Difference     rate is higher
 Type of unadjusted         is not       and difference
     difference         statistically   is statistically   Total
                         significant       significant
                                         at least at the
                                        10 percent level

Black and Hispanic
denial rate is higher
and difference is
statistically
significant, by degree
of significance
  1 percent                  31.1                0           100
  5 percent                  74.0                0           100
  10 percent                 84.7                0           100

Difference is not
  statistically
  significant                98.1               .4           100

Non-Hispanic white
denial rate is higher
and difference is
statistically
significant, by degree
of significance
  1 percent                  45.8             50.0           100
  5 percent                  94.1              5.9           100
  10 percent                 80.6             19.4           100

NOTE: Includes transition-period applications (those submitted
before 2004); for explanation, see text note 46. The adjustment
factors and the racial and ethnic categories differ from those in
tables 9 through 13. For explanation of adjustment factors and for
method of allocation into racial and ethnic categories and
definitions of categories, see text and text notes 50 and 51.

15. Distribution of the difference between incidence of
higher-priced loans for black and Hispanic borrowers as
a group and the incidence for non-Hispanic white borrowers,
by lender-product combination, 2004

Percent except as noted

                                                 Distribution of
                                                  percentage of
                                                  lender-product
                                                 combinations after
                                                     adjustment

                                                 Black and Hispanic
                                                 incidence is higher
                                                  and difference is
                                                   statistically
                                                   significant, by
                                                      degree of
                                                    significance

                                  Number of
                                lender-product
Type of unadjusted difference    combinations    1 percent   5 percent
                                    before
                                  adjustment

Black and Hispanic incidence
is higher and difference is
statistically significant,
by degree of significance
  1 percent                          532           43.0        15.4
  5 percent                          285            4.9        11.9
  10 percent                         241            1.2         4.6

Difference is not
statistically significant           12,112          .1          .3

Non-Hispanic white incidence
is higher and difference is
statistically significant,
by degree of significance
  1 percent                           33             0           0
  5 percent                           26             0           0
  10 percent                          31             0           0

                                 Distribution of
                                   percentage of
                                  lender-product
                                 combinations after
                                    adjustment

                                Black and
                                Hispanic
                                incidence
                                is higher
                                   and
                                difference
                                is statis-
                                 tically
                                 signifi-
                                cant, by
                                degree of
                                 signifi-
                                  cance

                                              Difference
                                                is not
Type of unadjusted difference   10 percent   statistically
                                              significant

Black and Hispanic incidence
is higher and difference is
statistically significant,
by degree of significance
  1 percent                        7.5           34.0
  5 percent                        8.4           74.7
  10 percent                       9.1           85.1

Difference is not
statistically significant           .4           99.0

Non-Hispanic white incidence
is higher and difference is
statistically significant,
by degree of significance
  1 percent                         0            93.9
  5 percent                         0            96.2
  10 percent                        0            90.3

                                   Distribution of
                                    percentage of
                                   lender-product
                                  combinations after
                                      adjustment

                                  Non-Hispanic
                                white incidence
                                   is higher
                                 and difference    Total
Type of unadjusted difference   is statistically
                                  significant
                                at least at the
                                10 percent level

Black and Hispanic incidence
is higher and difference is
statistically significant,
by degree of significance
  1 percent                            0            100
  5 percent                            0            100
  10 percent                           0            100

Difference is not
statistically significant              .2           100

Non-Hispanic white incidence
is higher and difference is
statistically significant,
by degree of significance
  1 percent                           6.1           100
  5 percent                           3.8           100
  10 percent                          9.7           100

Note: Excludes transition-period loans (those for which
the application was submitted before 2004). The adjustment
factors and the racial and ethnic categories differ from
those in tables 9 through 13. For explanation of adjustment
factors and for method of allocation into racial and
ethnic categories and definitions of categories, see text
and text notes 50 and 51.

16. Distribution of the difference between mean APR spread of prices
for loans above the threshold for black and Hispanic borrowers as a
group and the mean spread for non-Hispanic white borrowers, by
lender-product combination, 2004

Percent except as noted

                                                Distribution of
                                                  percentage of
                                                 lender-product
                                               combinations after
                                                   adjustment

                                                 Black and Hispanic
                                               mean spread is higher
                                                 and difference is
                                                   statistically
                                                    significant,
                                                    by degree of
                                                   significance

                                Number of
                              lender-product    1      5      10
    Type of unadjusted         combinations    per-   per-   per-
        difference                before       cent   cent   cent
                                adjustment

Black and Hispanic mean
spread is higher and
difference is statistically
significant, by degree of
significance
  1 percent                          172       28.5   18.6    6.4
  5 percent                          145        5.5    8.3    9.0
  10 percent                         127        4.7    6.3    7.1

Difference is not
  statistically                   12,611         .1     .2     .3
  significant

Non-Hispanic white mean
spread is higher and
difference is statistically
significant, by degree of
significance
  1 percent                           75        1.3    1.3      0
  5 percent                           57        1.8    1.8      0
  10 percent                          73          0      0    1.4

                                   Distribution of percentage of
                                 lender-product combinations after
                                           adjustment

                                                Non-Hispanic
                                                 white mean
                               Difference     spread is higher
                                 is not        and difference
    Type of unadjusted        statistically   is statistically   Total
        difference             significant      significant
                                              at least at the
                                              10 percent level

Black and Hispanic mean
spread is higher and
difference is statistically
significant, by degree of
significance
  1 percent                       40.7                 0          100
  5 percent                       77.2                 0          100
  10 percent                      81.9                 0          100

Difference is not
  statistically                   99.3                .2          100
  significant

Non-Hispanic white mean
spread is higher and
difference is statistically
significant, by degree of
significance
  1 percent                       84.0              13.3          100
  5 percent                       91.2               5.3          100
  10 percent                      97.3               1.4          100

NOTE: For definition of APR spread, see table
7, note 1. See also note to table 15.