|Remarks by Governor Edward M. Gramlich
Second Annual Robert J. Lampman Memorial Lecture, University of Wisconsin, Madison
June 16, 1999
A Policy in Lampman's Tradition: The Community Reinvestment Act
It is a pleasure to speak at the Lampman symposium. I first met Bob nearly three decades ago. At the time I was a young economist in the process of switching jobs from working on macroeconomics at the Federal Reserve Board to working on poverty problems at the Office of Economic Opportunity (OEO). In those days OEO was the main funder of Wisconsin's Institute for Research on Poverty (IRP), and Bob was of course the intellectual mainstay of IRP. Time has marched on, and I am now back at the Fed, ironically working on both macro and poverty issues at the same time.
Bob used to promote his anti-poverty agenda in two parts. The first was to ensure "high levels of employment and increasing average product per worker." This is, of course, one of the fundamental mandates of the Fed, and we are trying to achieve the objective daily. One often reads about these efforts in the newspapers, and there is little more about them that I want to say today. The second was "special private and public programs for those groups who do not readily share in the benefits of economic progress."1 It turns out that the Fed administers a very important program of this type as well, the Community Reinvestment Act (CRA). This is the program I would like to speak about today. Although to my knowledge Bob never studied the program, I think he would be very interested in the act.
CRA began back in 1977 as a little-known bank law, responding to reports of redlining and other credit market distortions. Based on the charters of banks and savings and loan associations, which require these financial institutions to meet the convenience and needs of the communities they serve, the CRA specifically encourages institutions to serve the credit needs of low- and moderate-income borrowers in their communities. The financial regulatory agencies grade financial institutions on their CRA performance, the grades are made public, and the institutions' CRA records are to be considered by the regulators in assessing proposed mergers and acquisitions.
The actual number of mergers denied by the regulators because of CRA problems is small, but the act has acquired much more importance and visibility than numbers alone would suggest. On the political side, it has become a major bone of contention in discussions about new financial reform legislation. On the economic side, it seems to have been responsible for about $120 billion annually in loans to low- and moderate- income homeowners, small businesses, small farms, and community development activities. Though some of these loans would no doubt have been made in the absence of CRA, and some may not benefit low-income groups disproportionately, $120 billion is a lot of money, far more than was ever contemplated for anti-poverty programs back in Bob's heyday.
Most poverty scholars are studying changes in the income support system in this country, but the impact of credit market programs such as CRA could be just as important in influencing the long-term economic prospects of low- and moderate-income groups. The University of Wisconsin has been a leader in studying social and poverty problems for a long time now, beginning with John R. Commons and Edmund Witte, continuing through Lampman and then up to the other leading present day IRP researchers. The ultimate point of my talk here today is to encourage you modern-day poverty scholars to get interested in credit market programs such as CRA. The impact of this program is likely to be significant, and social scientists interested in fields such as anti-poverty policy, urban development, and social policy should focus research attention on it.
How CRA Works
The law is administered by the federal regulatory agencies that already supervise the relevant financial institutions--the Fed, the Office of Comptroller of the Currency (OCC), and the Federal Deposit Insurance Corporation (FDIC) for banks and the Office of Thrift Supervision (OTS) for savings and loan associations. These regulatory agencies examine and rate financial institutions every two or three years, giving grades of outstanding, satisfactory, needs improvement, and substantial noncompliance. There has been evolution of the testing criteria and grade distribution over time, but today the following grade distribution has become characteristic:
Staff from the regulatory agencies meet frequently to try to standardize their CRA criteria and ratings as much as possible.
All CRA ratings are made public, and low ratings can be used by the regulatory agencies to block mergers and acquisitions, providing the real teeth in the CRA law. In addition to the formal CRA ratings, there is a public comment period for mergers and acquisitions, and in many cases a public meeting. If some point about the applicant's CRA record is raised in either the comment period or a public meeting, the agencies try to assess the validity of the claim objectively. Given the rating distribution above, it is rare that mergers are actually blocked because of low CRA ratings or negative public comments, but the publicity is certainly not good, and the great majority of financial institutions try hard to establish good CRA records--probably the main reason for the high CRA grades. It also, of course, means that we might expect CRA to be responsible for a significant amount of lending activity.
The biggest component of CRA loans is for mortgages to low- and moderate- income borrowers. There are detailed reporting requirements for mortgages made to these and other borrowers under the Home Mortgage Disclosure Act (HMDA). In 1997, the 10,000 commercial banks and savings and loan associations subject to this act made about 1.1 million new loans to low- and moderate-income homeowners or to those living in low- and moderate-income Census tracts in their communities. According to Fed staff research, the aggregate volume of these mortgage loans made under CRA in 1997 was about $58 billion, a total that has remained roughly stable through the 1990s.
The other main component of CRA lending is for small businesses and farms. The formal statistics from the Federal Financial Institutions Examination Council (FFIEC) peg the aggregate amount of new small business loans in 1997 at $159 billion, but the Fed staff estimates that only about one-fourth of these loans, $41 billion, were made as a result of CRA. Adding in another $18 billion of community development loans, we get an aggregate of $117 billion of these loans made under CRA for 1997, with similar totals for other recent years.
Nobody has studied these totals carefully enough to generate counterfactual evidence--that is, what share of these loans would have been made in the absence of CRA. Nevertheless, the fact that the CRA law was passed in response to fears of redlining and other barriers to credit in low- and moderate-income areas, and the eagerness of financial institutions to receive CRA credit for their low-income lending, suggests that a large share of the loans probably would not have been made in the absence of CRA.2 The econometrics are still unclear, but the law certainly could have been responsible for a sizable amount of new lending, dwarfing the expenditures on any other poverty program--in Lampman's day and today.
Perhaps as significant as the numbers themselves are the activities that take place. While there is a wide range of such activities, the prototype CRA project features a community group supplying the entrepreneurship and organizational capability. This group may procure some cheap vacant land from the city, obtain other grants or funds for construction or rehabilitation of housing units, and then sell the units to lower-income homeowners. The mortgages on the properties will be made by banks or savings and loan associations, which get CRA credit for these loans. The community group will take funds from the sale of the new homes and roll them over into a revolving loan fund. The bank or savings and loan will often sell the mortgage to some secondary buyer, often a government-sponsored enterprise such as Fannie Mae or Freddie Mac. These transactions take place roughly at market prices, however, so it makes sense to give the originating unit the CRA credit for the loan. Fannie Mae and Freddie Mac will then securitize the mortgages and sell them privately.
The repayment experience has generally been very good for CRA loans. Confidential data for one large but anonymous mortgage lender give a cumulative three-year foreclosure rate of 2.3 percent for loans that were likely to be made under CRA.3 This is an overestimate of the annual loss rate on CRA loans because the foreclosure rate is for the first three years after the loan and because on average only about half the value of the loan to the lender is lost at foreclosure. The Fed staff estimates that annual loss rates on CRA loans over the 1993-97 period are on the order of one-third of 1 percent. The aggregate annual loss rate on the $282 billion CRA mortgage loans over this five-year period is .0039; the aggregate annual loss rate on the $191 billion in CRA small business loans over this period is .0035. There is no evidence that the loss rate for CRA small business loans is any higher than the loss rate for other (non-CRA) small business loans, though there is evidence that the loss rate on CRA mortgage loans is slightly higher than the extremely low loss rate on non-CRA mortgage loans.
These low loan loss rates, when combined with some slight loan subsidies by many financial institutions, imply that CRA loans are nearly as profitable as other loans. A survey of large residential mortgage lenders showed that 98 percent of these lenders found CRA loans profitable, with 24 percent finding them as or more profitable than other loans.4 Hard evidence is difficult to come by, but the profitability of CRA loans could be rising over time as lenders learn more about screening and credit-scoring for this segment of the population and as new opportunities arise for secondary market sales.
Given the rapid rise in bankruptcy filings in recent years, it is natural to wonder if CRA-induced expansions of credit are not partly to blame. The evidence weighs in against the hypothesis. Again, the Fed staff has looked into the matter and attributes, at most, 3 to 4 percent of overall bankruptcy losses to CRA loans. Moreover, the time patterns differ. While overall bankruptcy filings and chargeoffs for consumer credit cards and other loans have increased sharply since 1995, CRA loan losses have been low and stable throughout the 1990s.
Questions Regarding CRA
Structure of the Law
Since about 95 percent of these institutions generally receive passing CRA grades --outstanding or satisfactory--one would expect about 90 percent of the mergers between two randomly chosen institutions to be approved without a detailed CRA review. In fact, more than 99 percent were. The Treasury Department reports that of the 86,000 merger applications filed since 1985, there were only 755 CRA protests.5 In these protested cases, 690 mergers were approved, and only 65 were denied on CRA grounds, a rejection rate of .0008.
One can get more detailed information from the Federal Reserve, one of the regulatory agencies included in the above statistics. Over 1993-97, the Fed considered an average of 1,100 merger cases a year in which the merging institution was subject to CRA. In 1,030 of these cases the merger was approved without a CRA protest. In the remaining 70 cases, issues serious enough to trigger a detailed CRA review were raised. Sometimes there were CRA protests raised in writing or in public meetings held to discuss the potential merger. Sometimes the issue involved the CRA rating itself. Sometimes the issue involved an alleged fair-lending problem or some other compliance matter. In each instance the Board staff would analyze the case and bring a recommendation to the Board. During this period there was an average of only one denial per year based primarily on CRA, but in another seven cases per year the application was withdrawn, perhaps for CRA-related reasons. In factual terms, then, mergers were disapproved in somewhere between one and eight cases out of 1,100.
These numbers do not make the CRA test look very formidable, but that impression is belied by other considerations. One is the simple amount of lending, and the share of banks with good CRA records. Both sets of facts suggest either that the CRA test is stronger than it looks or that banks are increasingly finding it in their own interest to do CRA-type lending. Another consideration is the organizational impact of CRA. Many banks devote significant resources to their CRA efforts, often creating special divisions to manage their CRA projects. Bank personnel work very hard on their CRA submissions, and regulatory personnel work very hard on their examinations. Often, potentially merging banks make substantial prospective CRA lending commitment agreements with community groups, even though the Fed and other regulatory agencies have often stated that they go only by the past record, not by any prospective CRA agreements.
These CRA agreements have caused some problems in and of themselves because there are allegations that community groups "hold up" banks by threatening to file protests or to speak against them at public meetings unless they sign CRA agreements. Given the announced stance of the regulatory agencies--that the regulators look only at past CRA records and not at future agreements--it is puzzling that community group threats, if indeed they are made, should have such power over banks.
In the end, coming to a judgment about the structure of the CRA law is difficult. On one side, the present structure does seem to have been effective at getting a wide range of financial institutions to pay serious attention to low- and moderate-income credit needs, buttressed by government examinations but not by a costly CRA legal enforcement process. The law has certainly opened up a dialogue between financial institutions and community groups, and it seems to have opened up profitable new lending opportunities to low- and moderate-income borrowers--opportunities that financial institutions may not have discovered without the push of federal legislation. On the other side, there may well be more focus than is socially optimal on banks in position to merge, as opposed to being in position to improve lending practices to the benefit of low- and moderate-income groups. There may also be gaps in the coverage of CRA across lending markets--not all markets may be served by financial institutions subject to CRA. Scholars interested in government regulations should study CRA to see what can be learned about desirable ways to regulate financial activities.
Because competition in providing credit to low- and moderate-income borrowers is still limited, it may be possible for unscrupulous lenders to give superficially attractive terms on loans. Loans for home repairs or debt consolidation might give cash up front to cash-poor borrowers but carry high interest rates, high fees, unnecessary insurance, and repayment terms that make the repayment schedules difficult to meet down the road. Loans may also feature balloon payments that end up forcing borrowers into default or refinancing, at still more unfavorable terms.
Many such dubious practices are reported. Loan sales people are reported to search neighborhoods for low-income but high-equity borrowers, often elderly, starved for cash and barely literate. There are reports of deceptive advertising, excessive fees, sales of useless insurance (credit life insurance for borrowers with no dependents, disability insurance for retirees, credit insurance that exceeds the loan balance), high penalty payments, and even of outright forgeries on loan forms. These reports indicate that borrowers are unwittingly led to rapid turnover of home equity loans--a process known as "loan flipping"--with high finance charges every time. As borrowers pay the fees and get even more cash-starved, the equity they have built up in their homes over a long period of time gets reduced--a process known as "equity stripping."
While there are reports of outright frauds and forgeries, the more common problem seems to be plain old deception. If borrowers were careful, financially astute, and had good counseling, these loans would rarely be entered into, and fraudulent lending should not be much of a problem. But this is a segment of the borrowing population that can be trusting, is often illiterate about credit matters, and is potentially vulnerable.
Knowing what to do about the problem is a major public policy challenge. Often the reported abuses revolve around refinancing of homes and balloon payments. There are plenty of borrowers out there who greatly benefit from such credit arrangements, so it makes no sense to outlaw the practices altogether. Earlier there were attempts to limit high loan charges such as interest rates with usury laws, but these just prevent mutually advantageous transactions between knowledgeable lenders and borrowers and cause credit to dry up when general levels of interest rates rise. Better disclosure laws might help, but disclosure forms are already complex with much fine print, and the supposed readers of these forms often cannot understand them and sometimes cannot even read.
In response to earlier reports of fraudulent lending, the Congress in 1994 passed the Home Ownership Equity Protection Act (HOEPA). HOEPA prevented balloon payments in the first five years of a loan--hence limiting early loan-flipping but perhaps taking even further advantage of myopic borrowers. It also defined a class of "high cost" loans, loans that charge closing fees of 8 points or more, or have an annual percentage interest rate 10 percentage points above prevailing Treasury rates. For these HOEPA-protected loans there are thorough disclosure requirements and prohibitions of some practices that often lead to abuses. What has happened is that many lenders have skated just below the HOEPA requirements, there has continued to be rapid growth of subprime lending, and there are continued reports of predatory lending.
The connection between CRA and predatory lending is potentially quite interesting and, at this point, is not very clear. There may be very little connection. Or there is a possibility that CRA might inadvertently foster predatory lending because institutions subject to CRA can receive credit for lending to low- and moderate-income borrowers regardless of the terms of the loan. Or CRA could limit predatory practices, by bringing more competition to low- and moderate-income credit markets. Predatory lending would be held in check if there were true competition in these markets, where even low-income and vulnerable borrowers had a real choice of borrowing possibilities and clear explanations of lending terms. By encouraging banks to get into these markets--banks that are subject to regulation, that use standard loan packages, that are used to working with community groups, and that have a CRA service test--CRA might be a force for the good.
But this whole set of issues has not been studied much by researchers. It would first of all be desirable to find out just how widespread the predatory practices are: The only remotely quantifiable information on this score is now from data on legal actions.6 It would be helpful to understand the low- and moderate-income lending process better, to determine the conditions under which predatory lending can and cannot thrive. It would be helpful to identify the role of competition, or the lack of it, in the predatory lending process. And it would be helpful to determine the role of CRA in bringing about better competition. Different geographic areas could be studied, with different concentrations of vulnerable borrowers and different intensities of CRA activities, to answer some of these questions.
The Overall Impact of CRA
A first question is the relatively straightforward one of determining the impact of CRA on low- and moderate-income lending. CRA seems to have had a large effect, but perhaps not as much as the gross figures would indicate because some of this lending would have taken place without CRA. The growth of subprime lending by institutions not covered by CRA is one such indication.7 On the other side, having been shown the way by banks, a number of nonfinancial corporations are now getting into the community lending business, a factor that might be counted in favor of CRA. Researchers could try to understand these loan statistics better. How much new lending is truly due to CRA? Has CRA inspired nonbank lending, or has bank and nonbank subprime lending grown for similar reasons? What are the properties of this lending, and what is the impact of CRA on underserved segments of the credit market?
A second question involves the impact of CRA on lending discrimination. CRA was passed largely in response to allegations of discrimination, redlining, and more general concerns about unequal access to credit in different areas. It is now more than twenty years since that time, and four years since the CRA testing criteria were fine-tuned to focus particularly on low- and moderate-income lending in low- and moderate- income areas. What has happened? There is evidence that the amount of loans to low- and moderate-income borrowers has risen, but is there evidence that lending discrimination is reduced, or that housing neighborhoods are more integrated?8 It would again seem that CRA has been around long enough that these types of evaluation studies could be done.
A third question involves the impact of CRA on financial evolution. Many CRA arrangements are very elaborate, with either outright gifts of urban properties from the city or city funds for land acquisition and development, funds from foundations and private developers, and CRA-type lending from financial institutions. Community groups that formerly would have been either agitating or running small local programs have become financial entrepreneurs--putting together deals, working with city bureaucracies and banks, even wearing charcoal grey suits. Similarly, city bureaucracies have changed and know how to work with these groups, and many banks now have community development divisions. It would seem that this sort of financial institution-building would be profoundly important, but at this point the changes and their effects are largely observed anecdotally. Putting rigorous form on all of these developments is a very challenging research task, but there may be big institutional changes out there, and one would like to think that social science researchers would have something to say about them.
A fourth question involves the impact of CRA on overall credit markets. Lenders are now perfecting analytic techniques such as credit-scoring that permit and facilitate identification of good credit risks among the low-income population. Fannie Mae and Freddie Mac have developed their own financial models to provide a functioning secondary market for the mortgage loans, without which the primary lenders would do much less lending. Researchers at the Fed and government-sponsored enterprises have already begun to understand the effects of these changes on credit markets.9 But there is more to do, both in understanding the impact of credit-scoring and in developing improved scoring techniques.
The fifth question seems most important of all, the impact of CRA on economic development. As one tours the country and visits projects partially or fully funded under CRA, there are clear suggestions of success. In some cities there seems to be a visual difference between neighborhoods with and without CRA projects--the houses look better, housing occupancy is more complete, the neighborhoods are neater. But these impressions are superficial and nonsystematic. One would think it possible to examine Census tract housing records, or city property value records, to do formal comparison studies of CRA and non-CRA activity areas in the same city, or pre- and post-CRA activities. Again, many of these projects are old enough now that if there were differences in property values, home ownership, school attendance, crime rates, teen pregnancy rates, or whatever, these differences should be showing up by now. The law is ultimately designed to improve urban and rural life on these dimensions, and at some point scholars should take on the question of ascertaining whether it really does so.
1 Both quotes are taken from Robert J. Lampman, The Low Income Population and Economic Growth, Study Paper No. 12, U.S. Congress, Joint Economic Committee (Washington, D.C.: Government Printing Office, 1959).
2 Suggests, but does not prove. There is a developing debate on the issue. On one side, Jeffery W. Gunther, Kelly Klemme, and Kenneth J. Robinson, "Redlining or Red Herring," Southwest Economy, May/June 1999, Issue 3, pp. 8-13, argue that credit market barriers to low-income lending have been breaking down anyway, and that CRA has not played much of a role in the process. On the other side, Douglas O. Evanoff and Lewis M. Segal, "CRA and Fair Lending Regulations: Resulting Trends in Mortgage Lending," Economic Perspectives, Federal Reserve Bank of Chicago, 1997, pp. 19-43, find evidence that the recent increase in mortgage lending to low- and moderate-income groups was due to CRA.
3 Michael LaCour-Little, "Does the Community Reinvestment Act Make Mortgage Credit More Widely Available? Some Evidence Based on Performance of CRA Mortgage Credits," Mimeo, May 4, 1998. LaCour-Little also attributes a large role to CRA in the growth of low- and moderate-income mortgage lending.
4 Larry Meeker and Forest Myers, "Community Reinvestment Act Lending: Is it Profitable?" Financial Industry Perspectives, Federal Reserve Bank of Kansas City, 1996, pp. 13-45. This finding is supported by Robert B. Avery, Raphael W. Bostic, Paul S. Calem, and Glenn B. Canner, "Credit Risk, Credit-Scoring, and the Performance of Home Mortgages," Federal Reserve Bulletin, July 1996, pp. 621-48.
5 Treasury Department release, May 4, 1999.
6 Luxman Nathan, "Borrower Beware: Equity Strippers Are Preying on Elderly Homeowners," Compensation and Banking, Federal Reserve Bank of Boston, Spring 1999, pp. 6-18.
7 Gunther, Klemme, and Robinson, "Redlining or Red Herring." The Shadow Committee on bank regulatory policy has the same skepticism--see George J. Benston, "Discrimination in Mortgage Lending: Why HMDA and CRA Should Be Repealed," Journal of Retail Banking Services, Volume 19, No. 3, Autumn 1997, pp. 47-57.
8 Helen Ladd, "Evidence on Discrimination in Mortgage Lending," The Journal of Economic Perspectives, Spring 1998, Volume 12, No. 2, pp. 41-62, summarizes the recent evidence on housing discrimination.
9 Robert B. Avery, Raphael W. Bostic, Paul S. Calem, and Glenn B. Canner, "Credit Risk, Credit-Scoring, and the Performance of Home Mortgages."