220 Week 1 A / for WIZARD KIM

199Cityscape: A Journal of Policy Development and Research • Volume 19 Number 2 • 2017 U.S. Department of Housing and Urban Development • Office of Policy Development and Research

Cityscape

Commentary: What Can We Learn From Government Attempts To Modify the Allocation of Mortgage and Consumer Credit in the United States? Anthony Yezer George Washington University

Introduction This commentary considers the Community Reinvestment Act (CRA) in historical context. CRA reflects one of many government attempts to influence the allocation of mortgage and consumer credit, but many of these interventions have had adverse outcomes. This commentary is written in the hope that those who are aware of history will stop repeating it. Specifically, I argue that lawmakers have been too quick to succumb to political pressures and have failed to follow basic economic principles when creating mortgage market policies. As a result, expanding access to credit has been prioritized over the safety and soundness of the housing and mortgage markets.

The Legacy of Past Housing Policies Until the 1990s, restrictions on banks limited their geographical expansion. The policy of not al- lowing interstate branching was formalized in the McFadden Act1 of 1927 and strengthened by the Bank Holding Company Act2 of 1956. Although these restrictions seem absurd today, they had im- portant implications for mortgage finance. Because they could not branch across state lines, banks held local mortgages in their portfolios and were forced to take substantial geographic risk that could not easily be diversified away. The lack of portfolio diversification was magnified because deposits were also local. As a result, a downturn in the local economy could result in bank failure, because customers would withdraw deposits and loan performance would deteriorate. Banks could not market the poorly performing local loans, and liquidity problems would turn into insolvency.

1 Pub. L. 69–639. 2 Pub. L. 84–511, 70 Stat. 133.

200

Yezer

The CRA Turns 40

Given the political unpopularity of branching, the answer to geographic risk diversification was to get mortgages out of the portfolios of the depository institutions that underwrote and endorsed them. The National Housing Act3 of 1934, which established the Federal Housing Administration (FHA) and introduced mortgage insurance to make mortgages more marketable, accomplished this goal. In the beginning, insurable mortgages had a maximum term to maturity of 20 years and a maximum loan-to-value (LTV) ratio of 80 percent, based on strict appraisals and required property inspections. The founding of the Federal National Mortgage Association (Fannie Mae) in 1938 to purchase both FHA-guaranteed and conventional mortgages was the second answer to the problem of diversifying geographic risk. Fannie Mae enabled housing to be financed by ultimate lenders who held a well-diversified portfolio. In many cases, banks, which could not diversify geographically due to statutory limits, purchased the mortgage-backed securities back from Fannie Mae.

The prohibition against branching provided a justification for federal involvement to diversify geographic risk, but it introduced other problems. Initially, FHA Section 203(b) mutual mortgage insurance was seen as a success. FHA was designed to protect homebuyers and taxpayers, but the limits on both maturity and LTV ratio crept upward as house prices rose and memories of the Great Depression faded. Redlining—which was designed to manage FHA’s risk by avoiding neigh- borhoods where house prices were likely to decline—came under attack for discriminating against minority neighborhoods.

Yet another policy was added in response: Section 235 of the Fair Housing Act4 of 1968. It relaxed lending criteria, reduced property inspection requirements (increasing the risk that mortgages were made on flawed units), and provided interest rate subsidies. The next 5 years were marked by scandal; more than 240,000 units went into default, resulting in a foreclosure rate five times that of FHA insurance. The effects of dilapidated and abandoned structures on neighborhoods turned residents against FHA and raised demands that the private sector become more involved in financing higher-risk loans. In my opinion, the primary impetus for passage of the Home Mortgage Disclosure Act5 (HMDA) of 1975 and the Community Reinvestment Act6 (CRA) of 1977 was the complete failure of Section 235, which was in turn a reaction to deficiencies in FHA Section 203(b) mutual mortgage insurance.

Given the failures of these FHA programs, public policy turned to the thrift industry to provide mortgage credit to lower-income borrowers and, again, set up policy conditions that worked against sound economic principles. Thrifts were given valuable competitive advantages. First, Regulation Q interest-rate ceilings were set to keep the cost of capital for thrifts artificially low but high enough to give them an advantage over commercial banks in attracting small savers. Second, restrictions on branching, particularly convenience and advantage regulations, gave thrifts some degree of local market power. However, a combination of rising interest rates and financial innova- tion that provided small savers access to market returns through Money Market Mutual Funds prompted disintermediation and destroyed the thrift business model. Economists had forecast these effects, but regulators ignored them.

3 Pub. L. 73–479. 4 Pub. L. 90–284, 82 Stat. 73. 5 Pub. L. 94–200, 89 Stat. 1124. 6 Pub. L. 95–128, 91 Stat. 1147, Title VIII.

Commentary: What Can We Learn From Government Attempts To Modify the Allocation of Mortgage and Consumer Credit in the United States?

201Cityscape

The thrift crisis of the 1980s gave the banking system a reprieve from the regulatory effects of HMDA and CRA, as the government’s problem was not how to finance more housing, but how to dispose of all the mortgages and properties acquired in the financial crisis. Eventually, about 750 insolvent institutions with assets of $800 billion (in 2016 dollars) closed. The Resolution Trust Corporation, established under the Financial Institutions Reform, Recovery, and Enforcement Act7 of 1989, was involved in disposing of defaulted housing assets not unlike that which followed the demise of the Section 235 program. Once again, history repeated itself.

The Post-1990 Public Policy Record Since 1990, a steady technological transformation of mortgage and consumer credit markets has taken place. Brick-and-mortar branches are closing. Lending is accomplished on the internet. Property appraisals and tax assessments are automated. When HMDA passed in 1975, property records were recorded on paper and filed in local courthouses. Now, property-transfer records are available on the web, easily scraped, and matched with HMDA records, so that today there is virtually no privacy in HMDA data. Indeed, the publication of HMDA data is inconsistent with U.S. Census Bureau standards for preserving privacy.8 CRA has also failed to respond to technological change. CRA is based on the presumption that deposit insurance is so valuable to banks with brick-and-mortar branches that they will assume substantial examination and compliance costs and will adjust lending, investment, service practices, or a combination of the three to achieve an “outstanding” or “satisfactory” CRA rating. That presumption, however, is technologically obsolete and financially unsound for both borrower and lender. Equally troubling is that economists have been unsuccessful in determining that having institutions with high CRA ratings makes a signifi- cant difference in overall community economic performance. Given that 97 percent of institutions examined achieve high ratings, the opportunity to study the effects of unsatisfactory performance on local economies is scarce.9 Paradoxically, it may be that CRA has actually discouraged branching that could expand the definition of market area. Given that branches have been closing rapidly, perhaps having CRA impose extra burdens on banks that branch is not a good idea.10

Despite the fact that the financial landscape had changed in a way that made CRA’s original premises invalid, the 1990s instead saw a move to combine CRA with the now revitalized and recapitalized government-sponsored enterprises (GSEs; that is, Fannie Mae and the Federal Home Loan Mortgage Corporation, or Freddie Mac) to provide high-risk mortgage credit. The Federal Housing Enterprise Safety and Soundness Act11 of 1992 enabled HUD to set mortgage purchase

7 Pub. L. 101–73, 103 Stat. 183. 8 For example, Gerardi and Willen (2008) were able to identify more than 70 percent of HMDA respondents by matching census tract, lender, and loan amount with readily available property records purchased from a commercial firm. This finding contrasts sharply with CFPB’s (2016) assurance to the public regarding privacy of their HMDA data, which stated, “This provides enough information about the location to be useful, but still provides protections for individual privacy.” 9 For an excellent discussion of the nature of CRA examinations and attempts to find economic effects, see Getter (2015). 10 For example, Bank of America’s annual reports indicate that it has 4,600 branches today compared with 6,100 in 2009. Of course, institutions can earn CRA points by selective branching, so the net effect of CRA on branching is difficult to determine. The intent of CRA is a complete reversal of previous public policy that discouraged branching. Indeed, under convenience and advantage regulation, banks were allowed to branch only into fast-growing, higher-income areas, because the concern was to preserve the safety and soundness of the banking system. 11 Pub. L. 102–550, 106 Stat. 3672, Title XIII.

202

Yezer

The CRA Turns 40

goals for the GSEs and established the Office of Federal Housing Enterprise Oversight (OFHEO) to monitor their safety and soundness. Again, sound economic principles were ignored. Initially, OFHEO calibrated a stress test using GSE mortgages acquired between 1979 and 1997 and was required to publish the results.12 However, OFHEO never updated it. Frame, Gerardi, and Willen (2015) found that, if the stress test had been updated, by 2004 it would have been apparent that the GSEs had a capital adequacy problem. Yet, in 2004, HUD raised the GSE affordable housing goals at precisely the time when the GSEs should have been contracting. In mid-2007, the GSEs reported $65.5 billion in book value of equity against $1.7 trillion in assets (3.9-percent ratio). In June 2008, they reported $54 billion in equity supporting $1.8 trillion in assets. On September 7, 2008, they were put into receivership. OFHEO and the Federal Reserve had allowed them to expand, rather than contract, based on faulty modeling and political pressure. Impartial economic analysis would have curtailed their operations years earlier, but political forces always triumph over economic analysis in mortgage market policy.

During the 1990s, another episode occurred in which political pressure caused large FHA losses in a policy initiative at least as flawed as the Section 235 program: the seller-funded downpayment program. Originally designed to expand access to homeownership, the seller-funded downpayment program enabled sellers to “voluntarily” contribute the downpayment for FHA-insured mortgages to an approved nonprofit organization, which used part of the contribution to help finance a downpayment. The Housing and Economic Recovery Act13 of 2008 finally terminated the program due to high default rates. Hard experience demonstrated the economic unsoundness of assuming that sellers would voluntarily contribute funds to a third party to pay the downpayment without raising the asking price by the amount of the contribution. Simple economic analysis would have demonstrated the fallacy of the program’s expectations and prevented the high rates of default and foreclosure.

Since the housing crisis of the mid-2000s, the gap between economic analysis and public policy toward credit markets has only grown wider. For example, the Dodd-Frank Wall Street Reform and Consumer Protection Act14 of 2010 (Dodd-Frank Act) limits the fees that mortgage brokers can charge to take, underwrite, and endorse mortgages. Fees are now expected to be uniform; no yield spread premium can be paid, regardless of whether the applicant applies online and has perfect financial records, a large downpayment, high credit score, and ample income or whether the applicant is unable to use a computer, keeps financial records in a shoe box, and has a low credit score.15 Discovering the lack of creditworthiness of individuals whose financial records are in a shoe box can be very expensive. The response to regulation of fees for brokerage services has been a predictable decline in mortgage brokers serving low-income, less-educated borrowers.

12 The requirement to publish a stress test is problematic and shows a profound misunderstanding of the problem of bank regulation because the test, once published, invites institutions to engage in regulatory capital arbitrage. The process is like giving students the questions on the final examination at the start of a course. They are likely to learn the answers and nothing else. 13 Pub. L. 110–289, 122 Stat. 2654. 14 Pub. L. 111–203, 124 Stat. 1376. 15 The Federal Reserve proposed the original broker compensation rule in July 2009. It is now part of the Dodd-Frank Act. Section 129 B of the Truth in Lending Act was modified, adding section (c), which prohibits a mortgage originator from receiving and no person from paying compensation based on the terms of the loan except for the amount of a residential mortgage loan.

Commentary: What Can We Learn From Government Attempts To Modify the Allocation of Mortgage and Consumer Credit in the United States?

203Cityscape

GSE regulation also continues to be problematic, for example, when political pressure does not allow for GSEs to price mortgages based on geographic risk. In 2008, the GSEs asked Congress to be allowed to require more equity in declining housing markets, a logical step to protect them against default risk. As Hurst et al. (forthcoming) explained—

The declining market policy was announced in December of 2007 and was implemented in mid-January of 2008. After receiving large amounts of backlash from a varied set of constituents, the policy was abruptly abandoned in May of 2008. Consumer advocacy groups rallied against the policy, arguing that it was a form of space-based discrimination. Real estate trade organizations used their political clout to protest the policy because it was hurting business. For example, the Wall Street Journal summarized the GSEs’ aban- doning the declining market policy by saying, “The change comes in response to protests from vital political allies of the government sponsored provider of funding for mortgages, including the National Association of Realtors, the National Association of Home Builders, and organizations that promote affordable housing for low-income people.” The Washington Post reported, “Critics, including the National Association of Realtors and consumer advocacy groups, had charged that Fannie Mae’s policy further served to depress sales and real estate values in the areas tainted as declining.

A further attempt in 2014 by the GSEs to add a 25-basis-point origination fee that would help cover additional losses in five states where foreclosure delays were long was also defeated by the same political forces.

Recently, Luan (2017) demonstrated that the spread between interest rates on jumbo mortgages (loans above the GSE and FHA ceiling) and conventional (GSE and FHA) rates can be used to pre- dict future changes in house prices. This important finding shows that the jumbo market responds effectively to differences in local risk. In contrast, these risks are ignored in pricing conforming mortgages. The result is that the normal role of interest rates in attenuating housing price bubbles is short-circuited by the GSE and FHA pricing policy. How important was this in the housing bubble that precipitated the mortgage crisis? That is difficult to determine, but Tian reported that a 1-standard-deviation change in the jumbo- and conforming-mortgage spread is associated with a 2.5-percent lower rate of subsequent house price appreciation. Clearly, current federal housing policy accommodates local housing bubbles, with CRA contributing to the dynamic by encourag- ing banks to make loans even in areas where rising house prices threaten housing affordability.

Contrasting an Economic Perspective and Current Regulatory Policy The fundamental basis for disagreement between economic analysis and current regulatory policy is that economists believe in using prices—that is, interest rates—to ration credit among alternative uses and users. Economists believe that raising interest rates is the proper response to a local hous- ing bubble, but public policy views higher interest rates as a bad thing because they make housing less affordable just as prices of housing units are rising.

204

Yezer

The CRA Turns 40

CRA is based on the notion that the banking system is responsible for making housing affordable, even when market prices are rising. Economists believe that, if housing is less affordable, that is a housing market issue. If housing costs $300 or more per square foot, the banking system is not responsible for making it affordable for low- or moderate-income households. Economists say that lawmakers should examine policies influencing the supply (zoning, building codes, transportation access, and so on) and demand (tax treatment of mortgage interest and property taxes and the exclusion from capital gains taxation) for housing rather than the local bank branch.

In addition to insisting on geographic uniformity of mortgage interest rates, public policy tends to oppose variation in rates across individuals. Borrowers are urged to shop for credit, presumably in the hope of lowering cost. However, current examination practices penalize lenders if interest rates are determined to be correlated with borrower demographic characteristics. Given that demo- graphic characteristics are correlated with financial and numerical literacy, the correlation of rates with demographic characteristics is inevitable if borrowers earn positive returns from shopping in credit markets, something that an economist would see as a positive.

Also, attitudes toward applicant competence differ between the regulatory environment and economics. Applicants face two challenges: (1) finding investment opportunities that are attractive, and (2) determining the best debt instrument to use in financing the opportunity. Most economists would say that the first challenge is greater and that many borrowers make bad investment choices. Indeed, one function of responsible lending is to stop borrowers from making bad decisions by denying them credit or making the credit so expensive that the problematic nature of the investment decision is clear. CRA does not credit lenders who prevent financial failure by denying applications or offering credit only at high rates. Loan denials are not seen as a valuable deterrent to overly optimistic loan applicants.

Another problem concerns the economic view of appropriate mortgage instruments. Advocates who focus on expanding access to credit often discount the risks and costs of making mortgages; some expect the financial system to provide homebuyers with 30-year, self-amortizing mortgages with no prepayment penalties, no deficiency judgments, debt-to-income ratios of 0.43, and LTV ratios of 0.95 or more. Credit scores of 620 (even 600) should be adequate, and an interest rate within 200 basis points of the 10-year treasury rate is expected. These mortgages may be economi- cally sound as long as nominal housing prices are rising, but if house prices fall, as happens periodically, this mortgage product could result in substantial losses. Accordingly, the financial system is generally not willing to hold such mortgages as an asset in significant quantity unless the government offers some form of guarantee that causes losses to fall on taxpayers or to be passed on to future homebuyers in the form of higher guarantee fees.

Finally, the premise of CRA is inconsistent with modern economic thought. The idea that local depository institutions should reinvest local deposits by holding local liabilities is completely inconsistent with modern portfolio theory and sound banking practice. Modern technology has broken the link between local deposit taking and lending. The great mystery about CRA is how such a flawed view of banking has survived so long in a country that leads the world in both internet technology and financial economics.

Commentary: What Can We Learn From Government Attempts To Modify the Allocation of Mortgage and Consumer Credit in the United States?

205Cityscape

Acknowledgments

The author acknowledges suggestions and encouragement provided by Carolina Reid, Mark Shro- der, Edgar Olsen, and Steve Malpezzi.

Author

Anthony Yezer is a professor of economics in at George Washington University.

References

Consumer Financial Protection Bureau (CFPB). 2016. “About HMDA.” Transcript of a video on the Home Mortgage Disclosure Act. http://www.consumerfinance.gov/data-research/hmda/learn-more.

Frame, W. Scott, Kristopher Gerardi, and Paul S. Willen. 2015. The Failure of Supervisory Stress Testing: Fannie Mae, Freddie Mac, and OFHEO. Working Paper 15-4. Boston: Federal Reserve Bank of Boston.

Gerardi, Kristopher S., and Paul S. Willen. 2008. Subprime Mortgages, Foreclosures, and Urban Neighborhoods. Working Paper 08-6. Boston: Federal Reserve Bank of Boston.

Getter, Darryl E. 2015. The Effectiveness of the Community Reinvestment Act. Congressional Research Report 7-5700. Washington, DC: Congressional Research Service.

Hurst, Erik, Benjamin J. Keys, Amit Seru, and Joseph S. Vara. Forthcoming. Regional Redistribu- tion Through the U.S. Mortgage Market, American Economic Review.

Luan, Tian. 2017. Did Investors Price the Housing Bubble? Evidence From the Jumbo/Conforming Rate Spread. Working paper. Washington, DC: George Washington University, Institute for Interna- tional Economic Policy.

Reproduced with permission of copyright owner. Further reproduction prohibited

without permission.