Statement by Lawrence B. Lindsey, Member, Board of Governors of the Federal Reserve System, before the Committee on Banking, Housing, and Urban Affairs, U.S. Senate, November 4, 1993.

Statement by Lawrence B. Lindsey, Member, Board of Governors of the Federal Reserve System, before the Committee on Banking, Housing, and Urban Affairs, U.S. Senate, November 4, 1993

I am pleased to appear before your committee today to present the results of the 1992 Home Mortgage Disclosure Act (HMDA) data. I will also make some remarks about the Federal Reserve's fair lending enforcement efforts.

Discrimination tears at the fabric of our democratic society. For the Federal Reserve, no single consumer issue is of greater concern than ensuring that the credit-granting process in the institutions that we regulate is free of unfair bias. Fairness in the assessment of credit applications is absolutely critical to our nation's well-being. Racial discrimination in particular--no matter how subtle, and whether intended or not--cannot, and will not, be tolerated.

The Federal Reserve's primary responsibility with respect to the HMDA data is to provide the data processing services for all the agencies under the auspices of the Federal Financial Institutions Examination Council (FFIEC) as a matter of operational convenience.

The responsibility for gathering the HMDA information, and ensuring that institutions follow fair lending practices, is allocated by law to six federal agencies. Of the more than 9,000 institutions that reported HMDA data in 1992, the Federal Reserve supervised approximately 600. For fair lending compliance--which applies not just to the institutions that file HMDA data but to all depositories--we supervise about 1,000 of the almost 13,000 banks and thrift institutions.

GENERAL DATA DESCRIPTION

The most striking feature of the HMDA data for 1992 is the enormous rise in the total number of housing loans applied for, compared with earlier years. The HMDA data show that more than 10 million such loans were applied for, compared with fewer than 7 million in 1991 and just 5.2 million in 1990. There is no question that a combination of lower interest rates and an improving and expanding economy in 1992 were the primary reasons for this growth.

The primary source of the growth in the volume of reported home lending activity was a dramatic increase in home refinancing. In 1992, 5.2 million applications for home refinancing were reported, compared with just 2.1 million in the previous year. The total number of home purchase loan applications also rose nearly 300,000. In addition, the number of applications for home improvement loans rose modestly. Not only was the number of applications up, but so was the number of approvals. More than 4 million home refinancing loans were approved, 77.7 percent of the total applied for, compared with roughly 1.5 million and a 73.2 percent approval rate in 1991. Home purchase approval rates for conventional loans were also up modestly, from 71.2 percent in 1991 to 72.9 percent in 1992. Approval rates for government-backed loans also rose.

This higher approval rate benefited both black and white applicants. Conventional home purchase loan approval rates rose 1.4 percentage points for blacks and 1.9 percentage points for whites. Government-backed mortgage approval rates rose 2.0 percentage points for blacks and 3.0 percentage points for whites. Of those individuals refinancing their homes, black approval rates rose roughly 6 percentage points, while white approval rates rose 4 percentage points. I would point out that these rises in approval rates for refinancings are particularly striking given that the number of applications for both groups more than doubled. And finally, with regard to home improvement loans, black approval rates rose 3.5 percentage points, while white approval rates rose 1.9 percentage points.

Approval rates also rose across the board for all income groups. Home re financing loan approval rates rose roughly 4 percentage points for each major income group, while home purchase loan approval rates rose most dramatically for low-income borrowers. The approval rate for applicants with incomes less than 80 percent of the metropolitan statistical area (MSA) median income went from 59.8 percent in 1991 to 68.9 percent in 1992 for conventional loans. For government-backed loans, the same group experienced a rise in approval rate from 66.2 percent to 74.8 percent. Approval rates for other income groups, on the other hand, were up roughly 1 to 2 percentage points.

The disparities between black and white approval and denial rates persist. For example, looking at conventional home purchase loans, about 36 percent of black applicants and 27 percent of Hispanic applicants were denied credit, compared with 16 percent of white applicants and 15 percent of Asian applicants--roughly the same as in 1991, although a slight improvement for black applicants. This matter continues to be of great concern.

Before going on, it is important to stress what conclusions can be drawn from the HMDA data. There is no question that the differential denial rates and approval rates for different income groups are troubling. However, the denial rates for applicants categorized by their race or national origin reflect a variety of factors. One factor relates to differences in the proportion of each group with relatively low incomes. In 1992, 21.0 percent of the white applicants for conventional home purchase loans had incomes that were less than 80 percent of the median family income for their MSA. The comparable percentages for blacks, Hispanics, and Asians were 37.1 percent, 27.6 percent, and 16.1 percent respectively.

Although the distribution of applicants by income may account for some variation among racial groups in loan disposition rates, looking at the 1992 HMDA data, other factors account for most of the difference. Differences in income do not completely explain it. This conclusion is evident because, after controlling for income, white applicants for conventional home loans in all income groupings have lower rates of denial than do black and Hispanic applicants. In fact, the denial rate of 21.1 percent for whites in the lowest income category (less than 80 percent of the MSA median family income) is the same as for blacks in the highest income category (more than 120 percent of the MSA median family income). [1]

Differential treatment on the basis of race and national origin may contribute to the variation, but it too does not fully explain the disparities in denial rates across racial and ethnic groups. For example, the study by the Federal Reserve Bank of Boston of lending patterns in Boston concluded that, after controlling for all known financial factors, race and national origin appeared to account for differences in denial rates among applicants. At the same time, the study also concluded that differences in income, together with other financial characteristics, alone would have caused black and Hispanic applicants to be denied credit at nearly twice the rate of white applicants.

The Boston study highlighted the limitations of interpreting the HMDA data. Such limitations do not in any way diminish the importance of ensuring equal access to credit for all Americans. The data merely point out the problems with relying on purely statistical analysis in reaching conclusions about the fairness of lending decisions. As I will note later in my remarks, the approach taken by the Federal Reserve and other agencies in developing new analytic techniques for investigating lending bias strikes a balance between traditional investigative techniques and computer-assisted statistical analysis. In particular, we use statistics to identify specific loan files that are suspicious and require further investigation. However, statistics alone can never, and should never, be used as the sole criterion for determining whether discrimination exists in a particular institution.

THE DISCLOSURE PROCESS

Under HMDA, most mortgage lenders that have offices in metropolitan areas, including independent mortgage companies, disclose information on the disposition of home loan applications and on the race or national origin, gender, and annual income of loan applicants and borrowers. Lenders also disclose, for loans originated or purchased during a year, the loans they sold, classified by the type of secondary market purchaser, and may indicate the reasons for denial of other applications. [2]

Covered institutions record separately, for each loan application acted on and each loan purchased, the items of information required by the Federal Reserve Board's Regulation C. Lenders submit this information to their respective federal regulator, which then sends the data to us for processing. Acting through the Federal Reserve, the FFIEC produces disclosure statements for each covered lender to make available to the public, plus an aggregate report for each metropolitan statistical area. These reports show the overall lending activity for covered lenders in each MSA and, together with the individual disclosure statements for lenders active in a given MSA, are available to the public at central data depositories. This information is also made available to the public in libraries throughout the United States.

Besides the print versions of the disclosure statements and aggregate reports, the FFIEC makes HMDA data available to the public in other forms. For instance, the HMDA reports or underlying data are available on microfiche, computer tape, and PC diskette and soon will be provided on CD ROM. The CD ROM format should be much more manageable than paper and microfiche for many users--especially those who view the data at central depositories--and will offer selections for viewing the data by MSA or by institution.

QUALITY OF THE DATA

I would like to say a few words about the quality of the HMDA data. Over the years, we and the other agencies that process HMDA data have had concerns about errors in the data that are submitted to us. By and large, errors can be traced to the data submitted (such as a lender's recording incorrect census tract numbers), although a few may arise during the agencies' data entry of loan register data submitted in hard copy. In the past three years, we have improved our capability to identify errors. As a result, we have succeeded in reducing the data errors in computer records from roughly 5 percent in 1990 and 1991 to less than one-half of 1 percent now.

Other types of errors cannot be identified at the processing stage. It is difficult to know, for instance, whether a financial institution has incorrectly identified the race of the applicant or has entered a census tract number that is valid but that is not correct for the property location to which the loan relates. Such errors evade our centralized data quality checks. Our examiners have stepped up their efforts to detect these problems during bank examinations, and we require institutions to correct and resubmit their HMDA data when we find errors. Financial institutions are strongly encouraged to ensure that they report accurate information; we help them by providing software with edit-check capabilities and through distribution of the FFIEC publication A Guide to HMDA Reporting: Getting it Right?

DETAILED RESULTS OF THE. 1992 HMDA DATA COLLECTION

The 1992 HMDA data reflect information submitted by 9,073 lenders, including 5,468 commercial banks, 1,395 savings and loan associations, 1,706 credit unions, and 504 mortgage companies (of which 224 were unaffiliated with a depository institution). The number of lenders disclosing data fell about 3 percent from 1991, a reflection of acquisitions, mergers, and failures. [3] Although the number of reporting institutions fell, the total number of applications and loans reported increased more than 50 percent, from 7.89 million in 1991 to 12.01 million in 1992. Much of the increase was due to refinancing activity.

Volume of Applications and Loans

In 1992, lenders covered by HMDA acted on roughly 10.03 million home loan applications--3.54 million for purchasing, 5.22 million for refinancing, and 1.24 million for improving dwellings for one to four families, and the balance for loans on multifamily dwellings for five or more families. [4] Nearly 78 percent of the reported applications for home purchase loans were for conventional mortgage loans; the remainder were for government-backed forms of credit--loans insured or guaranteed by the Federal Housing Administration (FHA), the Veterans Administration (VA), or the Farmers Home Administration (FMHA). The predominant reason for the substantial increase in volume of home loan applications reported in 1992 was the growth in refinancing activity. Spurred primarily by lower interest rates, the volume of applications to refinance an existing mortgage loan increased in 1992 almost 150 percent over the previous year. The growth in refinancings also reflects innovations in the marketplace, including the greater availability of "no-fee" loans and more efficient processing of applications that helped reduce closing costs. [5] Among the different racial and ethnic groups, the increase in 1992 applications for conventional loans by Asians was 5 percent, by blacks 22 percent, by Hispanics 8 percent, and by whites 17 percent. Applications for government-backed loans decreased roughly 5 percent for each group.

The conventional mortgage share of all reported home purchase loan applications increased roughly 4 percent from 1991 to 1992. This change in market share reflects a substantial decline in FHA activity. In 1991 the FHA accounted for 20.4 percent of all purchase loan applications and 20.5 percent of all home purchase loans. In 1992 these shares were 15.7 percent and 16.3 percent respectively. Recent increases in the cost to homebuyers using FHA loans, and greater availability of conventional loan products designed to reach low- and moderate-income homebuyers, likely account for the reduced reliance on FHA loans.

Despite this decline, the FHA program is favored by many thousands of households, particularly among first-time homebuyers. For instance, in 1992 almost half of the homebuyers using section 203(b) FHA loans (the principal type of FHA single-family mortgage loan program) were first-time homebuyers. The proportion had been even higher in 1991, when 57 percent of the FHA borrowers were first-time homebuyers. [6] On the other hand, the program is used infrequently to refinance existing home loans. Historically, FHA loans have accounted for only 3 percent to 4 percent of the refinancings annually. In 1992, FHA loans accounted for 3.7 percent of the 3.95 million refinancing loans reported by lenders covered by HMDA. One can surmise that households refinancing a loan often have accumulated sufficient equity in the home and no longer need the FHA's low-downpayment feature.

Use of Various Loan Products for Home Purchase

In 1992, 33.4 percent of home purchase loan applicants with low incomes (income less than 80 percent of the median family income for their MSA) applied for government-backed loans, compared with 13.2 percent of applicants with high incomes (income more than 120 percent of the median family income for their MSA). The greater reliance of lower-income households on government-backed loans reflects several factors. For instance, low-income households are much more likely to have limited money available to meet downpayment and closing cost requirements; hence, they are much more likely to use government-backed home loan programs. Conversely, the maximum limits on FHA loan insurance make this program less useful to households seeking to buy expensive properties.

Among the racial groups, blacks are much more likely than other groups to seek government-backed home purchase loans. In 1992, 41.2 percent of black applicants who applied for a home purchase loan sought government-backed loans; the comparable figures for Hispanics, whites, and Asians were 31 percent, 20.9 percent, and 10.6 percent respectively. These differences among racial groups are not entirely attributable to differences in income. For instance, among low-income loan applicants, 53.3 percent of blacks sought FHA or VA loans, whereas only 40.4 percent of Hispanic applicants, 31.2 percent of white applicants, and 21.7 percent of Asian applicants applied for a government-backed loan.

Disposition of Loan Applications

The 1992 HMDA data continue to show that lenders approve most home loan applications, particularly for buying a home or refinancing an existing loan. In regard to home purchase loans, lenders approved roughly 72.9 percent of applications for conventional financing and 74.1 percent of applications for government-backed financing. For refinancings, they approved 77.7 percent of the applications.

A comparison of the 1991 and 1992 HMDA data indicates that, nationally, denied applications for conventional home purchase loans declined somewhat, dropping from 18.9 percent in 1991 to 17.8 percent in 1992. Denial rates were also slightly lower in 1992 for applications for government-backed home purchase loans and for home improvement loans. For refinancings, on the other hand, denial rates dropped significantly--from 15.9 percent in 1991 to 12.4 percent in 1992. In general, low interest rates in 1992 coupled with relatively stable home values made homeownership more affordable in 1992 than in 1991 and may account for the lower denial rates. In addition, innovative mortgage loan programs by many lenders and greater use of affordable home loan programs sponsored by secondary market institutions may also have contributed to the decline in denial rates.

Disposition Rate for Different Groups of Applicants

The rates of approval and denial vary considerably among home loan applicants grouped by their income and racial characteristics. Nationwide in 1992, 80.5 percent of the applicants for conventional home purchase loans who were in the highest income grouping were approved for loans, compared with 68.9 percent for the lowest income grouping. A similar relationship between approval rates and applicant income is found for other types of home loans, including government-backed home purchase loans and loans for refinancing and for home improvement.

As in previous years, the 1992 HMDA data show that greater proportions of black and Hispanic loan applicants than of Asian and white applicants are turned down for credit. Consistent with these findings, the data also indicate that the rate of loan denial generally increases as the proportion of minority residents in a neighborhood increases.

Nationwide, for conventional home purchase loans, 35.9 percent of black applicants, 27.3 percent of Hispanic applicants, 15.9 percent of white applicants, and 15.3 percent of Asian applicants were denied credit in 1992. By comparison, the denial rates nationwide in 1991 for conventional loans were 37.4 percent for blacks, 26.5 percent for Hispanics, 14.9 percent for Asians, and 17.3 percent for whites.

The numbers for government-backed loans reflect somewhat lower rejection rates than for conventional loans. In 1992, 23.8 percent of black applicants, 18.5 percent of Hispanic applicants, 13.5 percent of Asian applicants, and 12.8 percent of white applicants were denied credit. In 1991, by comparison, the rates of loan denial were 26.4 percent for blacks, 18.9 percent for Hispanics, 16.3 percent for whites, and 12.5 percent for Asians.

Changes in the Amount of Lending by Income and Race

In recent years, lenders have targeted low- and moderate-income households and those seeking to buy homes in low- and moderate-income neighborhoods. Often such applicants have the necessary income to purchase homes in the price range they seek but lack the money to meet traditional downpayment and closing cost requirements. In some special programs, such as those sponsored by Fannie Mae and Freddie Mac, loan underwriting guidelines have been made more flexible. For example, these agencies' Community Homebuyers Programs have reduced the amount that must come from the applicant's own funds to cover the downpayment and closing costs, and lenders may take into account rent and utility payment records in lieu of other credit history information. [7] Other lender programs also target households with low asset levels and help keep monthly payments within the borrower's reach by waiving the usual requirements for private mortgage insurance on these very low downpayment loans.

It is difficult to gauge how much these targeted loan programs have increased homebuying opportunities for low- and moderate-income households. Our analysis of the 1992 HMDA data does, however, reveal a 27.1 percent increase in conventional home purchase loans to applicants from the two lowest income groupings (borrowers whose incomes were below the median family income for their MSA). The number of conventional loans to borrowers from the two highest income groupings (borrowers whose incomes were equal to or greater than the median family income for their MSA) also increased but by a more modest 12.3 percent rate.

We have seen some change in the volume of conventional home purchase loans to different racial groups from 1991 to 1992. Blacks had the largest growth in the number of loans received, increasing 25.9 percent from 1991 to 1992. The increase in loans extended to white households was a substantial 20.5 percent; the increases for Hispanics and Asians were a more modest 7.6 percent and 5.6 percent respectively. The number of loans made to minorities is not necessarily large, however. For example, out of a total of 1,896,000 conventional loans made in 1992 to the four largest racial or ethnic groups, whites received 1,582,030, Asians received 68,416, Hispanics received 66,995, and blacks received 56,516.

For each group, the largest percentage gains in conventional home purchase loans occurred among homebuyers with incomes below the median family income for their MSA. For example, among blacks whose incomes were below the median, the increase was 33.9 percent. The percentage changes for whites, Hispanics, and Asians in this income group were 28.2 percent, 25.4 percent, and 42.2 percent respectively.

Continuing Efforts to Eliminate Lending Discrimination

The HMDA reports reveal that credit history problems and excessive debt levels relative to income are the reasons most frequently given for credit denials. But specific information for applicants--on their level of debt, debt repayment record, employment experience, and other factors pertinent to an assessment of credit risk--is not available from the HMDA data. Nor do the HMDA data tell us about the specific underwriting standards used to assess prospective borrowers' applications. A popular tendency assumes that high denial rates are the result of unfair bias. In fact, the HMDA data by themselves do not give us a sufficient basis for assessing the fairness of the loan process or whether fair lending laws have been violated. The HMDA data do, however, provide a valuable tool to begin the inquiry into this question.

If you read the HMDA data on denial rates for minority applicants as synonymous with lending discrimination, then the similarities in each year's HMDA data would suggest that lending discrimination may be intractable. I do not believe that to be the case. But it will take new and increased measures to prevent, root out, and eliminate the problem. Such measures to deal with the problem, both directly and indirectly, are under way--among all the regulatory agencies-through enhancing examiner capabilities for detecting fair lending violations by financial institutions, increasing public information about discrimination in lending, and reforming the Community Reinvestment Act regulation.

Fair Lending Enforcement

In our program for enforcing fair lending, the Federal Reserve follows a coordinated approach. It focuses on examining for compliance with fair lending laws and more broadly on ensuring that credit is made available to low- and moderate-income areas, including those with substantial minority populations. Our approach also encompasses an aggressive program to investigate consumer complaints, provide consumer and creditor education, and gain insight through research.

Let me describe each segment briefly. In the research area, the study by the Federal Reserve Bank of Boston is well known. In my view, that study, released in October 1992, has done more than any other single effort to advance our understanding about fair mortgage lending and to suggest ways for us to attack the problem. It served to shift the focus, I believe, from an ongoing debate on whether unlawful discrimination exists in the mortgage markets to a concerted effort on the part of financial institutions, the regulatory agencies, and members of the public to search for ways to eliminate discriminatory practices.

Other research pieces--on HMDA data, household debt, credit shopping practices, the secondary market, and other related subjects--also have advanced our knowledge. And last week, the Federal Reserve released a comprehensive report to the Congress that compares the risks and returns of lending in low-income, minority, and distressed neighborhoods with those in other communities.

In regard to enforcement, the Federal Reserve System has oversight responsibility for approximately 1,000 state member banks. We have a comprehensive program of consumer compliance examinations, established in 1977, that are carried out by specially trained examiners. The scope of these examinations includes the Equal Credit Opportunity and Fair Housing Acts, and from the beginning our examiners have been trained to place special emphasis on problems involving potential discrimination of the kind prohibited by those statutes.

The Federal Reserve examines every state member bank at periodic intervals and on a regular basis. On average, about two-thirds of state member banks are examined each year for compliance with the fair lending and consumer protection laws. In general, examinations are scheduled every eighteen months for banks with a satisfactory record. For a limited number of banks with exceptional records, examinations take place every two years. Those banks with less than satisfactory records are examined every six months or every year, depending on the severity of their problems.

The examination procedures focus primarily on comparing the treatment of members of a minority or protected class with other loan applicants. First, the examiner reviews the bank's loan policies and procedures by looking at bank documents and interviewing lending personnel. The examiner seeks to determine, among other things, the bank's credit standards, and then--using a sample of actual loan applicants--to determine whether bank personnel have applied those standards uniformly. Special note is taken of applications received from minorities, women, and others whom the fair lending laws were designed to protect. The examiner looks at the same information the bank used to make its credit decision, including credit history, income, and total debt burden. If the bank's credit standards appear not to have been followed, or not applied consistently, these findings are discussed with lending personnel and a more intensive investigation is undertaken. Finally, an overall analysis of the bank's treatment of applications from minorities, women, and others within protected classes is conducted to identify any patterns or individual instances that might indicate applicants were treated less favorably than other loan applicants. When we find violations through any of these techniques, we will require correction by the institution, notification to the applicant, and referral of the matter to the Department of Justice or Department of Housing and Urban Development in appropriate cases.

Another important part of the examination involves talking with people in the community who are knowledgeable about local credit needs. Federal Reserve examiners routinely ask members of the community, local government officials, and the like about perceptions of credit availability for minorities and low- and moderate-income persons. The answers may suggest that a particular area of the bank needs additional scrutiny and may provide insights into how the bank is serving the credit needs of its local community, particularly among those protected by the anti-discrimination statutes.

But as you know, even with these procedures, it is difficult for our examiners to find evidence that we can be sure proves racial discrimination. Consequently, we have been searching for ways to provide them with better detection tools. Recently, the Federal Reserve System developed a computerized statistical model for using HMDA data in the fair lending portion of the examination, and we have shared this tool with the other financial regulators. I believe the model we have developed has the potential to be a substantial step forward, although we are still making adjustments to make sure it works as we want it to.

Starting with the HMDA data, the model allows the examiner to more expeditiously select a sample of loans for review. Ultimately, it enables us to match minority and nonminority pairs of applicants with similar credit characteristics, but different loan outcomes, for a more intensive fair lending review than would otherwise be possible for the examiner to make. Once the pairs are selected, examiners reexamine the credit files for the individual applicants to determine if discrimination may have played a part in reaching different outcomes. Our field tests of this "regression analysis" program have demonstrated its promise. We are working to refine the model, reduce the level of examiner resources that have been needed in some examinations, and implement the program throughout the Federal Reserve examination system. Although such comparisons of minority and majority applicants have always been a part of the Federal Reserve's fair lending examination, we believe that this computerized selection process will enable examiners to better focus their efforts and spend their time more effectively on the actual fair lending review of loan files.

Besides this "micro" use of the HMDA data, the Federal Reserve has developed, after discussions with the FFIEC constituent agencies, a computerized system for analyzing the expanded data collected under HMDA. The system is versatile and allows the data to be segmented by demographic characteristics such as race, gender, and income levels, or geographic boundaries. Examiners can now sort through vast quantities of data to focus attention on data for specific lending markets and to compare an individual HMDA reporter's performance against that of all other lenders in the area. They can more readily determine whether a bank is effectively serving, through mortgage and home improvement lending, all segments of its market, including low- and moderate-income and minority neighborhoods. And examiners can use this information to get a profile of the bank before they begin their examination, which gives them a head start in their investigation. We have been holding HMDA training sessions on how to use this system around the country for our examiners, as well as those from other agencies.

The Federal Reserve has also developed the capability to map by computer the geographic location of a bank's lending products, including mortgage loans. The mapping integrates demographic information for the bank's local community. We believe that this type of analysis and presentation will enhance our ability to assess a bank's CRA performance in meeting the credit needs of its local community, including minority areas. The mapping should also be helpful in evaluating a bank's geographic delineation of its local CRA service area to ensure that it does not exclude low- and moderate-income neighborhoods.

As you know, at President Clinton's behest, the financial regulatory agencies are also currently at work revising the regulations that implement the Community Reinvestment Act. One of our main goals with CRA reform is to make the standards used to judge lenders' performance more clear and objective. We are also trying to make sure that unwarranted paperwork and unnecessary regulatory burden are eliminated and that the focus of our efforts is clearly placed on the lending results achieved. The CRA obligates financial institutions to ensure that they are helping to meet the credit needs of their entire community, including low- and moderate-income areas. They cannot effectively meet this standard under the CRA if they discriminate against some segment of their community in making loans. It is our hope that reforming and strengthening the administration of CRA will result in greater investment in communities that may have suffered from disinvestment and discrimination.

The Federal Reserve's consumer complaint program is another element in our overall effort to enforce fair lending laws. Our procedures provide special guidance for investigating complaints alleging loan discrimination. Such complaints can prompt an on-site investigation by Reserve Bank personnel at the state member bank accused of discrimination. We also have a referral agreement with HUD for mortgage complaints and have sent a number of complaints to them for investigation. As in our examinations area, we are devoting considerable attention to strengthening our complaint processing system by increasing oversight, tightening deadlines for investigation, ensuring more personal contact with complainants, and making the public more aware of our procedures.

Public education also plays a role in our fair lending enforcement. We have distributed a pamplet entitled Home Mortgage Lending and Equal Treatment, A Guide for Financial Institutions to all the institutions we supervise. It identifies lending standards and practices that may produce unintended discriminatory effects, and it cautions lenders about their use. The pamphlet focuses on race and includes examples of subtle forms of discrimination, such as unduly conservative appraisal practices in changing neighborhoods; property standards such as size and age that may exclude homes in older neighborhoods; and unrealistically high minimum-loan amounts.

More recently, a comprehensive booklet was published and widely circulated by the Federal Reserve Bank of Boston, entitled Closing The Gap: A Guide To Equal Opportunity Lending. This significant and informative pamphlet is designed to straightforwardly address lending discrimination and what can be done to avoid it. It challenges lenders to reconsider every aspect of their lending operations, from the hiring of loan officers to the treatment and evaluation of applicants, to ensure that loan decisions are not made on the basis of race or ethnicity. The publication has been widely distributed, with more than 50,000 copies in circulation. In an effort to reach even more people with the information in Closing the Gap, the Reserve Banks of Boston, Chicago, and San Francisco are developing a videotape patterned on the pamphlet for use by banks in their in-house fair lending training. We hope that the training tape will be available for use in early 1994. We have also published a brochure entitled Home Mortgages: Understanding the Process and Your Right to Fair Lending to inform consumers about the mortgage application process and about their rights under fair lending and consumer protection laws.

Several public notices by the financial regulatory agencies recently also have stressed the need for financial institutions to provide credit on a nondiscriminatory basis. For example, the joint statements on credit availability discussed equal credit lending obligations. Also, a recent letter from Chairman Greenspan and the heads of the other supervisory agencies to the chief executive officers of all financial institutions stressed the importance of compliance with fair lending laws, and it provided guidance on how each institution could improve its performance.

One suggestion, which the letter recommended as a useful way to minimize the opportunity for bias in the evaluation of loan applications, is the so-called "second review" procedure. This procedure was suggested to address a concern raised by the Federal Reserve Bank of Boston study, which indicated that among marginally qualified applicants, white applicants were more likely to benefit from a lender's discretion in approving loans than black or Hispanic applicants. A second review would involve a financial institution's simply taking a second look at all the applications it expects to deny, as well as some loan approvals, to ensure that its existing credit standards were applied fully and fairly. We understand that the procedure provides lenders with greater comfort that they have made credit decisions in an unbiased manner. It can serve as another useful tool for lenders, suggesting adjustments in institutional behavior to correct racially disparate loan practices that may be occurring despite the institution's policies to the contrary. It also should assure borrowers who are aware of the procedure that an institution seeks to treat all applicants fairly.

The Board believes the goal of ensuring fair access to credit also can be advanced by focusing on positive actions that a lender may take. Through our Community Affairs program, the Federal Reserve conducts outreach and provides educational and technical assistance to help financial institutions and the public understand and address community development and reinvestment issues. We have increased resources to Community Affairs activities at the Reserve Banks--now staffed with more than fifty people--to enable the Federal Reserve System to respond to the growing number of requests for information and assistance from banks and others on the Community Reinvestment Act, fair lending, and community development topics. Efforts have been expanded to work with financial institutions, banking associations, governmental entities, businesses, and community groups to develop community lending programs that help finance affordable housing, small and minority business, and other revitalization projects. Overall, the Reserve Bank's Community Affairs programs sponsor or cosponsor about a hundred programs a year, involving thousands of participants, as a way to encourage economic development and ensure fair lending.

CONCLUSION

The 1992 national HMDA data continue to show, like the data in preceding years, relatively high rates of denial of home mortgage applications for minorities. They remain a troubling cause for concern about racial discrimination in mortgage lending. For us and for the other regulatory agencies, the data provide a starting point for in-depth analyses of the mortgage lending practices of individual institutions. We are engaged in an aggressive effort in our fair lending examinations to identify any violations of the fair lending laws for corrective action, referral to the Department of Justice, or both.

Fairness in assessing credit applications, without regard to race, sex, or other prohibited bases, is absolutely critical to our nation's well-being. Let there be no misunderstanding on that point. Racial discrimination cannot and will not be tolerated. We are committed to its elimination to the best of our ability.

[1]. In the highest income category, the denial rate was 8.8 percent for whites in 1992; the denial rate for blacks in the lowest income category was 36.0 percent.

[2]. Expanded data collection was required pursuant to amendments to HMDA in the Financial Institutions Reform, Recovery and Enforcement Act of 1989 (FIRREA). The expansion in coverage of mortgage companies came with FIRREA and with the amendments to HMDA in the Federal Deposit Insurance Corporation Improvement Act of 1991.

[3]. The total number of reporters will be higher for 1993, given the increased number of independent mortgage companies that will report lending activity as a consequence of changes in coverage that took effective January 1, 1993.

[4]. Besides applications, lenders also reported data on 1.98 million loans they purchased during 1992 from other institutions.

[5]. "No-fee" loans are those in which the consumer incurs no out-of-pocket expense to pay either closing costs or discount points on the loan. Such loans are often written with a higher interest rate to compensate.

[6]. U.S. Department of Housing and Urban Development, Characteristics of FHA Single-Family Mortgage: Selected Sections of National Housing Act, (HUD, 1991).

[7]. Other changes in the underwriting guidelines pertain to the treatment of nontaxable income and income from seasonal part-time or second jobs, income continuity and job stability, debt-to-income ratios, the appraiser's neighborhood and home improvement analyses, and property condition.