The Community Reinvestment Act in the Age of Fintech and Bank Competition

When the CRA came into force in 1977, banks were the main source of loans for home buyers and small businesses, and restrictions on bank branching posed a high barrier to competition. Today’s competitive environment is much changed. The removal of branching restrictions has allowed banks to expand and consolidate — leading to a 77 percent increase in the number of bank offices since the CRA’s passage. Furthermore, a growing share of mortgage and small-business lending now comes from financial institutions that are not subject to the CRA. In fact, LMI borrowers represent a larger share of these institutions’ borrowers than they do for banks, which are subject to the CRA.

Conversely, mounting evidence suggests the CRA is either ineffective or damaging. Before the financial crisis, community groups touted the act’s influence in lowering lending standards. Empirical research also shows that banks’ risk taking increases ahead of their CRA evaluations — contravening the CRA’s requirement that lending be consistent with bank safety and soundness. In cases where CRA lending is not riskier, evidence suggests that banks may be “skimming the top” — lending to high-income residents of low-income communities, thus meeting their regulatory mandate but failing to reach the people the CRA intends to help.

There is a strong case for repealing the CRA in favor of alternative policies that better achieve its goals. It would be a mistake to expand the CRA to cover online (fintech) lenders and credit unions, which already serve LMI borrowers as well as, or better than, many lenders that are subject to the act. If the CRA remains in place, its regulations should change to allow banks to trade their CRA obligations in order to encourage lender specialization and efficiency.

Introduction

The Community Reinvestment Act (CRA) is a 42-year-old statute that requires depository institutions “to demonstrate that their deposit facilities serve the convenience and needs of the communities in which they are chartered to do business ... consistent with the safe and sound operation of such institutions.”² The CRA ostensibly seeks to improve the welfare of low- and moderate-income (LMI) Americans by assessing and rating depository institutions on the basis of how much they lend to, invest in, and serve the communities in which LMI Americans live. Racial minorities were, and continue to be, disproportionately represented among LMI communities, so the CRA is considered part of the anti-discrimination legislation of the late 1960s and 1970s.³

For its first 18 years of existence, the CRA was “a vague statement of principle without much real-world effect.”⁴ Notably, a series of investigative articles in the Atlanta Journal-Constitution in 1988 documented large and persistent differences in the amount of bank credit extended to majority black communities compared to majority white ones.⁵ The reports uncovered evidence of redlining: the denial of services to poor and minority geographic regions.⁶ It was only after 1995, when changes to CRA enforcement shifted the focus from banks’ ex ante lending commitments to actual lending outcomes, that bank lending and other activities in LMI communities appeared to increase.⁷

But whether this increase was consistent with the safe and sound operation of banks is unclear. There is evidence that CRA-regulated institutions engage in significantly riskier lending in advance of CRA assessments, compromising their safety and soundness.⁸ This paper shows that, because such risky lending results in higher rates of default and harms the financial well-being of the borrowers who struggle to repay their loans, it is not clear that LMI borrowers benefit from the CRA.

There are still other reasons to question the present-day usefulness of the CRA. The 1977 act was inspired by a long-standing American tradition of bank localism, which has since ceased to characterize the U.S. banking system. With the removal of branching restrictions and statutory ceilings on savings and demand deposits, the concern that motivated the CRA’s drafters to mandate local credit extension, namely that potential borrowers would face few alternative suppliers, has become moot. Additionally, the use of CRA ratings in regulators’ deliberations on bank mergers creates incentives for inefficient lending and distracts attention from more important factors for consumer welfare, such as local bank competition.⁹

This paper argues for repealing the CRA, making the case that the act remains ill-defined in its policy objectives, arbitrary in its assessment practices, and is liable to harm borrowers and bank depositors. By contrast, reductions in regulatory barriers to branching and the growth of online lenders have significantly increased LMI Americans’ access to banking services, indicating that competitive markets can more efficiently achieve the CRA’s goals of serving these communities. The evidence presented here indicates that the case for outright repeal of the act is quite strong: short of repeal, its current system of ambiguous and bureaucratic assessment should at least be replaced with a system of tradable obligations related to the lending, investment, and service provision that the CRA seeks to encourage.¹⁰

A Brief Overview of the Community Reinvestment Act

Metrics and Requirements

The CRA applies to all insured depository institutions except credit unions.¹¹ It is enforced by a depository institution’s primary regulator, which may be the Office of the Comptroller of the Currency (OCC), the Board of Governors of the Federal Reserve System (FRB), or the Federal Deposit Insurance Corporation (FDIC).¹² As of 2018, the FDIC was the primary regulator for 3,617 depository institutions out of the 5,644 institutions subject to the CRA. The OCC and FRB conducted CRA examinations of 1,210 and 817 institutions, respectively.¹³ Since 1990, regulators’ CRA reports have been available to the public, enabling activist groups to use CRA ratings to oppose bank expansions and mergers on the grounds that the bank has failed to satisfy its obligation to fulfill the needs of the communities where it conducts business.¹⁴

CRA examiners use multiple measures to evaluate a bank’s lending to low- and moderate-income borrowers within a given assessment area.¹⁵ Assessment areas consist of one or more metropolitan statistical areas or metropolitan divisions where an institution subject to the CRA has its main office, branches, and deposit-taking automatic teller machines (ATMs), as well as surrounding areas where the institution has originated or purchased a substantial portion of its loans.¹⁶ LMI borrowers are those whose incomes fall below 80 percent of the median income in the metropolitan area where a bank branch is located, as well as those who live in census tracts with an income that is 80 percent or less than the area median, as determined by the Census Bureau.¹⁷

In 1995, CRA regulations underwent a series of significant revisions that shifted the focus of its assessments from processes — how a bank planned to increase lending to LMI communities — to outcomes — how much of its lending and other activities actually went to LMI borrowers and communities.¹⁸ The revised regulations also created separate assessment tiers for banks of different asset sizes. In 2005, those tiers were indexed to the consumer price index.¹⁹ As of January 1, 2019, institutions with less than $321 million in assets qualified as “small banks” for CRA examination purposes. Those with more than $321 million but less than $1.284 billion in assets were designated “intermediate small” banks.²⁰ Currently, examinations for these small and intermediate small banks are intended to be less onerous, and less frequent, than those for larger banks. Additionally, banks that receive high CRA examination grades (regardless of size) are rewarded with longer periods between examination cycles. The time between examinations can thus range from anywhere between 12 and 60 months, depending on the bank.²¹

Since the 1995 revisions, these examinations have taken the form of a one-, two-, or three-pronged test, graduated according to banks’ size. For banks with assets above the intermediate-small threshold, CRA regulators use the full three-pronged test, which evaluates the lending, investment, and services that banks provide to LMI customers and communities. The lending test evaluates the volume and distribution of an institution’s loans across borrowers’ income levels and geographic regions.²² The investment test examines the institution’s community development investments, such as activities that revitalize low-income geographic regions and disaster areas.²³ The service test evaluates the geographic distribution of a bank’s branches and ATMs, as well as how effectively the bank’s services promote community development.²⁴ There is some overlap in the activities that each of the three tests is meant to evaluate, and CRA regulations recognize this overlap by excluding activities counted under the lending or service tests from consideration in the investment test.²⁵ The three tests apply only to depository institutions above the regulatory thresholds for small and intermediate small banks. Small banks are evaluated according to their lending performance only; intermediate small banks are assessed on both their lending and community development activities.²⁶ Depository institutions subject to the CRA receive a rating according to their performance on each of the relevant tests. Each rating, in turn, is based on a qualitative assessment of the institution’s performance on the test’s different dimensions. For example, an institution’s rating is “outstanding” if, among other behaviors, it exhibits “excellent responsiveness to credit needs in its assessment area.”²⁷ However, institutions are downgraded a notch, to “high satisfactory,” if regulators deem their responsiveness to local credit needs as just good.²⁸ Perhaps unsurprisingly, the Treasury has criticized CRA ratings for lacking clear guidelines and leaving unexplained the criteria by which a bank can meet each level of performance.²⁹

Table 1 reproduces the number of points awarded by CRA regulators for a given level of performance under each of the assessment tests. The lending test is the most heavily weighted and therefore the most important: for each level of performance, it counts at least as much as the other two categories combined.³⁰

Table 2 shows how each aggregate point score translates into an overall CRA rating. The preponderance of the lending test means that no institution can receive an overall “satisfactory” rating unless it scores at least “low satisfactory” on the lending test.³¹

The CRA’s Flawed Foundations

The Community Reinvestment Act was one of several measures — the others being the Fair Housing Act (FHA, 1968), Equal Credit Opportunity Act (ECOA, 1974), and Home Mortgage Disclosure Act (HMDA, 1975) — aimed at reducing credit discrimination against poor and minority communities and otherwise improving those communities’ access to financial services.³²

The policymakers who supported the CRA in 1977 worried about financial institutions engaging in “capital export.”³³ This refers to the practice of lending deposits outside of the communities where those deposits are collected (and where the depositors themselves typically reside). Proponents of the CRA argued that depository institutions, most of which have enjoyed federal deposit insurance since 1933, had an obligation to lend within the communities from which they received their deposits.³⁴ Sen. William Proxmire (D-WI), then chairman of the Senate Banking Committee and the CRA’s sponsor, argued that:

Federal and state authorities limited entry into the U.S. banking system, argued Proxmire, so it was only fair to require banks to lend in the communities from which they were able to extract rents — profits in excess of what the banks could earn in a competitive market — courtesy of government regulation.

The idea that bank deposits should remain in the areas where the depositors live is a long-standing one in American banking. Small agricultural interests were early supporters of this type of localism and were happy to support granting bankers local monopoly charters in exchange for the bankers’ commitment to lend to them in both good times and bad.³⁶ Localism also informed the persistence of regulatory restrictions on intrastate and interstate branching well into the 1980s.³⁷ Indeed, the prevalence of unit banking — that is, a statutory prohibition on operating more than one bank office — was deeply rooted in popular culture for much of American history. For example, the 1946 movie It’s a Wonderful Life — “one of the most beloved in American cinema”³⁸ — presents small-town thrift banker George Bailey (Jimmy Stewart) as the community’s bulwark against evil big business.³⁹

Yet economically, the practice of confining savings to the localities where they are collected is very costly. A useful function of banks is to pool depositor savings and to deploy those funds as loans. Pooling facilitates beneficial diversification: acting on their own, individuals can only commit funds to one or a handful of projects, exposing themselves to the specific risks of those borrowers each time they do so. Banks, on the other hand, can allocate funds among hundreds of thousands of different lending opportunities. Diversification therefore both reduces portfolio risk for a given level of returns and helps depositors earn more at lower risk.⁴⁰ Furthermore, trade in credit — that is, borrowing and lending funds — is like other forms of trade in that it is mutually beneficial, enabling the depositor to earn a satisfactory rate of return and enabling the borrower to secure capital for consumption and investment.⁴¹ Just as restricting trade in goods is harmful to the welfare of consumers and producers, restricting trade in credit (for instance, by requiring the borrower to reside in the same location as the depositor) can only reduce profitable opportunities for credit extension.

Moreover, bank branching enables banks to diversify their loan portfolios across assets and places, which in turn makes bank failure less likely.⁴² Banks with multiple branches can offset losses in some hard-hit areas with earnings from relatively less affected areas. For example, when the Great Depression began, California had the most developed bank branch network in the United States.⁴³ Branch banking in that state not only made the banks that operated branches more stable; it also made the unit banks in competition with branch banks more stable than unit banks in places without branch banks, suggesting that competitive pressure provided a healthy check on inefficient unit bank practices.⁴⁴ Another example is the Canadian banking system — characterized by a small number of large banks — which has exhibited a great deal of stability over 150 years, without detriment to depositor rates of return.⁴⁵

Although restrictions against branching were the main impediment to bank loan diversification at the time of the CRA’s passage, the CRA further undermined geographic diversification, which helps to mitigate risk, by implicitly requiring that a share of deposits be lent out where those deposits are collected.⁴⁶ Because only certain types of lending — mainly mortgages and small-business loans — and investment count for CRA credit, the CRA also reduces a bank’s opportunities for asset diversification. Furthermore, and contrary to the philosophy that informed the CRA, the deposit-taking activities of banks benefit communities quite apart from any related lending operations. Deposit facilities give their holders access to the benefits of loan diversification. They also contribute to depositors’ credit histories, facilitating future use of other credit products. Finally, deposit accounts give depositors a valuable reward for their funds, in the form of low-cost banking and payments, and, for some demand deposits, regular interest payments.

At the 1977 Senate hearings, Proxmire argued that the CRA would ultimately assist bank diversification by “alleviating fears that a more liberal branching policy would be inimical to community welfare,” thereby making the removal of branching restrictions more politically palatable.⁴⁷ However, his suggestion overlooked the fact that, if branching were in fact permitted, the CRA would limit both its attractiveness to banks and banks’ ability to take full advantage of it.

It is difficult to picture a scenario, absent considerable information asymmetries, in which political direction of bank lending could improve upon the allocation resulting from market prices: if attractive projects in the local community can yield an adequate return for a given risk, there is no need for a political directive to mandate lending. If there are better opportunities elsewhere, such a mandate is harmful to depositor returns and banks’ safety and soundness. Proponents of the CRA in the 1970s and 1980s claimed that bank redlining warranted political intervention. As argued below, however, CRA regulations may be undermining LMI communities’ efforts toward financial inclusion in today’s much-changed banking environment.

The regulators charged with enforcing the CRA (the OCC, the FRB, and the FDIC) voiced some of these concerns in 1977. Then comptroller of the currency Robert Bloom warned that the CRA would harm credit institutions “established primarily to serve the needs of a particular segment of the United States population nationwide,” using as an example the case of an American Indian bank that aimed to offer banking services to that group on a nationwide basis.⁴⁸ Fed Chairman Arthur Burns worried that mandating “standards for setting the proportion of total loans that an institution should allocate to local credit would necessarily be arbitrary.”⁴⁹ FDIC Chairman Robert Barnett raised the concern that the CRA could “discourage financial institutions from making applications for offices in neighborhoods where funds are badly needed because of the reexamination that this would entail in [the] areas where they already have offices.” Barnett also worried about increased concentration of bank branches in affluent areas, and the duplicative reporting burden on institutions that were already subject to the Home Mortgage Disclosure Act.⁵⁰

The Changing U.S. Banking Landscape

Two structural trends in U.S. banking since 1977 further strengthen the case for reconsidering the CRA: bank consolidation as a result of the removal of branching restrictions, and the growing market share of online (fintech) lenders.⁵¹ This section considers the merits of current CRA assessments in light of the rise of branch banking. A later section argues that the rise of fintech lending bolsters the case for repealing the CRA altogether.

Many of the anti-competitive restrictions that Proxmire cited to justify the CRA in 1977 have since been removed, improving the welfare of bank customers and weakening the case for the CRA’s implicit local lending mandates. The most important of these policy changes has been the steady liberalization of bank branching, that is, the ability of a single bank to operate multiple offices within states and beyond their home state. In 1970, only 13 states allowed banks to operate branches, and no state allowed out-of-state banks to operate branches within its borders.⁵² From the 1970s onward, however, a growing number of states authorized in-state and out-of-state branching, so that by 1990, all but five states allowed intrastate branching, and the same number (although not the same states) permitted interstate branching.⁵³ This process of steady liberalization culminated in 1994 with the passage of the Riegle-Neal Act, which removed federal restrictions on branching.⁵⁴

Branching deregulation ushered in rapid bank consolidation, with the average annual number of bank mergers more than doub­ling between the 1960s and the 1990s.⁵⁵ The number of FDIC-insured commercial banks peaked at 14,496 in 1984 (see Table 3). It stood at 10,453 by the passage of the Riegle-Neal Act, dropping to 7,279 on the eve of the financial crisis, and to 4,918 by the end of 2017. The number of branches, on the other hand, had expanded from 42,731 in 1984 to 79,163 by 2017, only slightly lower than its 2009 peak of 83,130.⁵⁶ This means that the number of bank offices (headquarters plus branches) is much higher today than at any time before the consolidation trend started — albeit below the number of bank offices in operation just before the financial crisis.

The CRA was passed during a period of extensive branching restrictions. At the time, there was a worry that without strict regulation, communities where locally headquartered banks did not lend would struggle to find a competing supplier. In addition, between 1933 and 1986 the Federal Reserve set an interest rate ceiling on bank savings deposits through Regulation Q.⁵⁷ This regulation also banned interest on demand deposits until 2011. By restricting the interest that banks could offer to depositors, Regulation Q subsidized bank funding, creating rents for banks above the return they would earn in a competitive market. The weakened competition, both from Regulation Q’s subsidies and from branching restrictions, arguably strengthened the case for local lending mandates that forced banks to share some of their regulatory rents with customers.⁵⁸ However, these rents were a product of interest rate controls and anti-competitive regulations, not market factors, so they could have been better addressed by repealing Regulation Q and liberalizing bank branching sooner. At any rate, the rationale for community reinvestment that Regulation Q provided disappeared with its repeal.⁵⁹

Branching deregulation had several beneficial effects. First, it increased the efficiency of the banking sector by facilitating the expansion of the best-performing institutions and removing anti-competitive protections for the worst-performing ones. Greater competition in turn lowered both the share of bad loans on bank balance sheets and the average loan interest rate.⁶⁰ Economic growth increased as states liberalized bank branching.⁶¹ Furthermore, thanks to branching liberalization, there are more banks serving any given individual community today than there were at the height of branching restrictions.⁶² The increased banking options now available to consumers (regardless of income level) have made deposit rates more competitive, increased loan options, and enabled consumers to benefit from large fee-free networks across the United States.⁶³ One way to illustrate this expansion of choice and its effects is by examining the long-term increase in the average distance between small business borrowers and their lenders — from 100 miles in 1996 to 250 miles by 2016.⁶⁴ When prospective borrowers have access to more distant lenders, the local bank’s willingness to lend can no longer determine whether borrowing will occur.

Yet the CRA remains in place, restraining further competition and growth by limiting where and to whom institutions can lend. Indeed, today’s CRA regulations do not just require banks to lend in the communities where they take deposits, they also ban branches deemed “primarily for the purpose of deposit production.”⁶⁵ In other words, despite the fact that banks have had the ability to open branches outside their home state since the mid-1990s, regulators have the authority to close any branches they believe to be conducting insufficient local lending.⁶⁶ While CRA enforcement authorities do not appear to have yet used this power,⁶⁷ even before they acquired it there were reports of delayed and abandoned bank mergers because of pending CRA examinations and concerns that a bank’s low CRA rating might pose an obstacle to the merger.⁶⁸

Contemporary Problems with the CRA

The CRA Encourages Banks to Make Riskier Loans

Although it is clear that the CRA places constraints on the way in which banks allocate credit, the act’s proponents have long argued that CRA loans are too small a share of total lending to constitute a prudential risk, and that there is no evidence that CRA loans are riskier or less profitable than other loans.⁶⁹

The experience of the financial crisis suggests otherwise. In fact, pre-crisis testimony from community organization representatives explicitly pointed to the CRA as one cause of overly lenient underwriting standards. In a 2007 report, for example, the National Community Reinvestment Coalition touted “higher debt-to-equity ratios than ... conventional loans,” “flexible underwriting standards,” “low or no down payment[s],” “commitments by secondary market institutions [notably, the government-sponsored enterprises (GSEs)] to purchase loans,” and “second review” of denied applications as consequences of the CRA.⁷⁰ Raising loan-to-value ratios, relaxing borrower standards, and pushing secondary market institutions to buy more bank loans all make lending riskier.

The financial crisis cast a bright light upon the extent to which many households, particularly LMI ones, had taken on large housing debts.⁷¹ There is considerable evidence that the regulatory push to extend mortgage lending to LMI communities, which accelerated in the mid-1990s, and the accompanying promise that the GSEs (Freddie Mac and Fannie Mae) would buy those mortgages, drove that debt increase.⁷² The extent to which the CRA is responsible for unprofitable lending to LMI households remains a matter of debate, but the $4.5 trillion in CRA commitments between 1992 and 2007 tracks closely with the excess in affordable housing loans made by the GSEs (relative to their historical norm) during that same period.⁷³

Even if the CRA was not the main contributor to bad mortgage credit growth in the run-up to the financial crisis, it may have enabled the proliferation of credit by giving aggressive lending practices the respectable cover of community reinvestment. Pre-crisis accounts of the CRA’s “success” support this hypothesis, as they focus on the growth of LMI lending and homeownership, rather than the CRA’s suitability for borrowers or its implications for bank safety and soundness.⁷⁴ Even before the financial crisis, CRA supporters recognized the difficulty of attributing increases in low-income lending to the act, since both high rates of economic growth and other government policies — such as the loosening of GSE standards — could better explain the observed increase in lending to those communities.⁷⁵

However, proving that the CRA was a success requires showing that it led to higher lending volumes without compromising the lenders’ safety and soundness. Pre-crisis evidence of the CRA’s impact, even when it suggested significant growth in LMI lending by depository institutions, failed to show that such credit was sound.⁷⁶ The crisis and its aftermath, on the other hand, showed that mortgage lending on lenient terms could harm financial institutions and borrowers alike. It is not surprising that institutions subject to the CRA, especially those looking to grow and merge with others, would increase their LMI lending, since regulators take CRA ratings into account when approving bank expansions.⁷⁷ Such lending may even have benefited banks and their managers in the short term. But was it good for borrowers, bank shareholders, taxpayers, and the economy in the long run?

Some of the evidence says no. A 2012 National Bureau of Economic Research (NBER) paper looking at CRA lending between 1999 and 2009 finds that banks significantly increased their lending around the time of CRA examinations, and that such loans were riskier. Specifically, lending volume increased by 5 percent and default rates increased by 15 percent in the quarters surrounding a bank’s examination.⁷⁸ The increase in lending is particularly large and significant for banks with more than $50 billion in assets, which is consistent with the hypothesis that larger institutions, being more likely to expand and merge, will have a greater incentive to strive for high CRA ratings in hopes of having their mergers approved.⁷⁹ The study’s authors also find that the increase in risky lending became more pronounced in the later years of the housing boom.⁸⁰ Their finding agrees with the contention made by, among others, former Federal Reserve governor (and 1990s CRA reform architect) Lawrence Lindsey that, to avoid a CRA rating downgrade, before the crisis banks increasingly reached out to riskier borrowers as the demand from more creditworthy borrowers was satisfied.⁸¹

The NBER paper has been criticized for focusing on the quarters surrounding CRA examinations, thus failing to recognize that these examinations themselves evaluate lending in periods well before those dates.⁸² The authors counter, however, that depository institutions have an incentive to concentrate their CRA lending close to the exam so as to minimize recorded default rates, which might fall foul of the CRA’s requirement that lending be consistent with safety and soundness.⁸³

In contrast, a 2013 Federal Reserve bulletin found that LMI loan delinquency rates in banks’ CRA assessment areas were lower than those outside their assessment areas, suggesting — according to the authors — that the impact of the CRA on financial fragility, if any, was comparably minor.⁸⁴ However, the study cited only one year of evidence; furthermore, it showed that credit scores of LMI borrowers within the surveyed banks’ assessment areas were higher, and those borrowers much less likely to be subprime, than in the case of LMI borrowers outside CRA assessment areas.⁸⁵ More creditworthy borrowers are less likely to default. Their preponderance among the LMI borrower cohorts of banks’ assessment areas suggests that banks might be “skimming the top”: lending to the most creditworthy borrowers in LMI areas to fulfill their CRA requirements while also minimizing risk.⁸⁶ Such behavior may satisfy regulators, but it contradicts the assertion that CRA loans serve the marginal borrowers and communities that the statute ostensibly targets.

Furthermore, evidence that CRA-motivated lending was less risky than other types of lending to LMI borrowers does not prove that the CRA was beneficial, or even neutral, for bank balance sheets and the health of the wider banking system. Indeed, as recently as 2006, regulators issued draft rules to exempt CRA-related equity investments — such as providing capital and employment for community development purposes — from higher Basel II capital charges.⁸⁷ Pro-CRA activists encouraged this move, which made it more attractive to make CRA investments at the expense of bank safety and soundness.⁸⁸

Yet another problem with citing increases in LMI lending as evidence for the economic gains associated with the CRA is that the opportunity costs of CRA-induced lending may exceed the benefits. Gross growth rates of LMI loans ignore opportunity costs. Consider the scenario in Table 4, where a bank with $10 million worth of available funds faces a choice of four projects to finance.

In the absence of the CRA, and assuming for simplicity that all prospective borrowers face a similar interest rate, the bank would pick the projects with the highest likelihood of repayment; that is, Projects A, B, and C. Under the CRA, however, if the bank believes that its loan to Project B will not suffice to get the bank a high CRA rating, it may choose Project D over Project C because the loan applicant in D (with income below 80 percent of the area median) qualifies for LMI status, whereas the applicant in C does not.

While the philosophy of the CRA implies that lending to applicant D over applicant C has positive benefits beyond its return to the bank, it is important to note that rejecting applicant C comes with costs: first, to the bank’s shareholders, who will receive a lower expected return on capital; second, to applicant C, who, while not low-income by the regulatory definition, is not much better off than applicant D and still places below the median income of the bank’s assessment area. In fact, a 2000 Fed survey showed that 44 percent of respondent banks found their CRA mortgage loans to be less profitable than their other mortgage loans.⁸⁹ Furthermore, 52 percent of respondents indicated that CRA-related mortgage loans were costlier to originate, on a per dollar basis, than non-CRA loans.⁹⁰ These results suggest that, for many institutions, CRA lending involves higher costs than non-CRA lending — and these costs are passed on to other borrowers and shareholders as well.

The 2007–2009 financial crisis illustrated the harm that a single-minded drive to increase mortgage lending could do to vulnerable communities. It was a surfeit of politically induced housing credit, rather than a scarcity of it, that left households badly exposed when the crisis hit.⁹¹ Yet the CRA continues to assess depository institutions primarily on their lending to LMI areas, despite evidence that such lending is riskier and costlier to underwrite.

Compliance with the CRA Is Unnecessarily Burdensome

There are four levels of CRA performance: “outstanding,” “satisfactory,” “needs to improve,” and “substantial noncompliance.”⁹² Between 2006 and 2014, no more than 3.5 percent of depository institutions subject to the CRA received an overall rating below satisfactory in any year. More than 90 percent of institutions received a satisfactory rating in 2014.⁹³ These statistics have caused some analysts to conclude that compliance with the CRA is not a burden on depository institutions.⁹⁴ Their assumption is that, if the CRA were onerous, more institutions would fail their CRA exams.

In fact, things are not so simple. An institution’s CRA ratings tell us only that it passed the evaluation; they say nothing about the resources it dedicated toward doing so. Just as the low number of bank failures prior to the 2007–2009 financial crisis did not imply the absence of financial fragility, one cannot conclude that a high pass rate means the CRA is not burdensome.⁹⁵ The resources a bank dedicates to merit a satisfactory or outstanding evaluation can be substantial, and these often come with further direct and indirect costs to banks, shareholders, and consumers. In fact, the CRA is responsible for 7.2 percent of community banks’ compliance costs, according to a Federal Reserve Bank of St. Louis survey.⁹⁶ That is despite the fact that more than 80 percent of community banks, that is, those with less than $10 billion in assets, are subject to the less burdensome small or intermediate small CRA assessment protocols. As compliance costs represent between 5 and 10 percent of community banks’ noninterest expenses overall, the CRA can make a perceptible dent on bank operating margins, particularly on those of smaller banks that have lately seen higher compliance costs and lower rates of return.⁹⁷

The CRA Fails to Promote Financial Inclusion

At the time of the CRA’s passage, there was a concern that certain depository institutions would systematically refuse to lend to minority communities, even when doing so would not mean taking on undue credit risk.⁹⁸ Investigative reporting, notably by the Atlanta Journal-Constitution, continued to expose this practice of redlining in the years immediately after the CRA went into effect.⁹⁹ However, 42 years later, the barriers to financial inclusion for low-income and minority communities are different. The CRA not only fails to address those barriers; it may contribute to the difficulty of overcoming them.

According to the FDIC, as of 2017 there were 8.4 million U.S. households (6.5 percent of households) without a bank account. Another 24.2 million have only limited access to banking services and must instead resort to alternative — usually costlier — providers.¹⁰⁰ Unbanked rates are much higher for minorities: 16.9 percent of black households and 14 percent of Hispanic ones are unbanked, compared to 3 percent of white ones. Additionally, more than half of black and Hispanic households with incomes below $30,000 report no mainstream source of credit.¹⁰¹

Two commonly cited reasons for lacking a bank account are not having enough money to deposit and account fees being too high.¹⁰² Regulatory compliance costs are a principal driver of both account fees and minimum deposit requirements to avoid those (and other) fees. As mentioned earlier, CRA loans are costlier to originate than other loans, and CRA compliance costs account for 7.2 percent of all community bank compliance costs.¹⁰³ While this is lower than the share of bank costs related to the Bank Secrecy and Truth in Lending Acts, it is still significant, especially considering that overall bank compliance costs have increased in recent years.¹⁰⁴ Thus, while it might appear that the CRA’s low-income lending mandates promote financial inclusion among lower-income borrowers, its indirect impact on account charges likely reduces access to deposit and credit services among the very populations the CRA is meant to serve.

The decline of small banks (despite a steady rise in the number of bank offices), further bank consolidation since the financial crisis, and rising compliance costs have all contributed to the phenomenon of so-called banking deserts. These are census tracts with no bank branches within a 10-mile radius of their centers.¹⁰⁵ As of 2016, 3.7 million Americans lived in banking deserts, while another 3.9 million lived in areas that may soon lose their last bank office.¹⁰⁶ Banking deserts are mostly rural and therefore do not account for a large share of the unbanked population.¹⁰⁷ Nevertheless, some states with a high population share living in banking deserts, such as Arizona, Nevada, and New Mexico, also have above-average unbanked rates (Table 5).

Median household incomes in banking deserts are lower than they are in nondeserts. Even for potential banking deserts, median household income sits at 10 to 20 percent below the nationwide median.¹⁰⁸ The population of banking deserts is therefore not much different in its socioeconomic characteristics from the groups that the CRA aims to help.

Growing regulatory compliance expenses contribute significantly to the rising cost of operating bank branches, increasing the likelihood of branch closures and contributing to the spread of banking deserts.¹⁰⁹ CRA branching restrictions worsen this problem by discouraging the establishment of bank branches and ATMs in sparsely populated areas. Without the implicit lending requirements of the CRA, depository institutions might be more willing to take deposits in, and to serve, low-density geographies. CRA regulations also discourage banks from expanding to areas with few lending opportunities by making it costlier to operate a branch, thus reducing banks’ incentive to open new branches or maintain old ones in marginally profitable locations.

The CRA Has Not Resolved “Rational Redlining”

Even some critics believe that the CRA may be justified, if informational asymmetries specific to LMI communities lead banks to ration credit more with them than they do among other populations — what one author calls “rational redlining.”¹¹⁰ For example, credit rationing can occur if a community has witnessed only a limited number of local property transactions, generating insufficient data on home prices and complicating banks’ capacity to make accurate loan appraisals.¹¹¹ Uncertain appraisals, in turn, raise the down payments demanded by banks, further dampening loan volumes.¹¹²

However, even if rational redlining is a problem in some present-day LMI communities, CRA regulations can be of only limited help. The CRA’s assessment policy links lending obligations to deposit-taking. But that cannot resolve rational redlining: if banks take deposits within a community, they will have access to information about its credit quality, economic conditions, and property values. On the other hand, banks lacking such information are unlikely to operate or take deposits in that community precisely because they face uncertainty regarding available lending opportunities. CRA regulations target institutions that already have operations in local communities and have information about local market conditions. A better way to facilitate greater credit extension is to increase competition by attracting new lenders into communities, thereby helping to correct any information failures. As argued below, the CRA works against these efforts.

The CRA Is a Harmful Industrial Policy

One way that the CRA encourages bank compliance is by directing regulators to take a bank’s CRA rating into account when evaluating its application for a deposit facility — that is, whenever that bank wants to set up a branch or merge with another bank.¹¹³ On the one hand, such use of CRA ratings encourages regulators and activist groups to oppose the applications of banks that they perceive as having underperformed in their lending to LMI communities.¹¹⁴ Indeed, groups such as the National Community Reinvestment Coalition have explicitly linked periodic waves of bank mergers to increases in those banks’ commitments to lend, suggesting that the merger would not have taken place without the banks’ lending promises.¹¹⁵ On the other hand, a bank’s positive CRA record can lead regulators to overlook other concerns related to its activities, such as the impact a large bank merger could have on local credit spreads — the difference between the interest banks charge borrowers and what they pay to depositors. Because the credit spread is a proxy measure for competition within a local banking market, it is often a more economically significant indicator of a merger’s impact on consumer welfare than the merging banks’ CRA ratings.¹¹⁶

The CRA has thus become a tool of industrial policy, rewarding institutions for meeting political goals and threatening to punish those perceived to have fallen short. This facet of the CRA raises three important concerns. First, it may reduce efficiency by blocking consolidation that would lower bank operating costs and increase loan diversification, which, as discussed above, promotes safety and soundness. There has been a steady trend of bank consolidation since passage of the Riegle-Neal Act, but many small banks remain: for example, as of the third quarter of 2018, the FDIC reported 1,335 supervised institutions with assets below $100 million, with an average return on equity 3 percentage points below that for larger banks.¹¹⁷ Both figures suggest that some gains from economies of scale remain to be grasped in U.S. banking. Foreclosing such efficiency-enhancing mergers would harm depositor returns and undermine bank safety and soundness.

Second, the CRA may reduce competition if a bank merger gives the resulting institution sufficient market power to raise prices. The United States has a long history of monopolistic and oligopolistic local banking markets. Competition only started to become the norm from the late 1980s onward, and the evidence suggests it has had positive effects on economic growth and consumer well-being.¹¹⁸ The CRA, by making it costlier to establish branches and expand operations, can have a deleterious impact on local bank competition.¹¹⁹ This possibility is particularly worrying in the post-crisis U.S. banking landscape, which is characterized by very few new banks¹²⁰ and strong restrictions on the number of new charters issued by the FDIC.¹²¹

Finally, CRA regulations weaken the incentive for banks to guard against unprofitable lending, if banks perceive the benefits from easier consolidation to outweigh the losses incurred from CRA loans.¹²² Between 1992 and 2007, cumulative CRA lending commitments increased 500-fold, suggesting that banks were willing to spend heavily to please their regulators once branching liberalization increased merger and expansion opportunities.¹²³ As discussed earlier, the evidence also suggests that loans timed to coincide with banks’ CRA evaluations are riskier than other loans.¹²⁴

Credit volumes are an imperfect proxy, and certainly not a substitute, for the welfare of communities and households. In 1977, the CRA focused on LMI lending because there was evidence of widespread redlining, abetted by nationwide restrictions on bank branching. Four decades later, the CRA, as currently enforced, raises many concerns regarding the effectiveness of its LMI lending mandates, its compliance cost to banks, and its impact on prudential standards. It also fails to address the contemporary issues facing LMI communities, such as the high rate of unbanked households and the growth of banking deserts.

Better Ways to Promote Community Development

If the goal of the CRA is to raise the real incomes of LMI communities, then a more diverse array of policies offers greater promise for achieving it. These alternative policies would also have fewer adverse consequences for bank safety and soundness than the CRA’s implicit lending mandate. They include liberalizing zoning laws to lower the cost of housing, reducing low-income tax burdens, curbing occupational licensing to facilitate employment and entrepreneurship, lowering tariffs on food and clothing imports, and relaxing overly strict childcare regulations.¹²⁵ The high cost of living in many urban areas hurts LMI communities in particular, but that is a problem that neither banks nor financial regulation can readily solve.

Of course, improving LMI households’ access to credit can also improve the well-being of those households. But there are better ways to facilitate LMI access to credit than by imposing CRA mandates. The most effective alternative is to ease banks’ entry into the lending business.

Facilitate Lender Entry

Regulators should make entry into local lending markets easier by issuing more charters and reducing regulatory barriers for nondepository institutions. The rate of new bank creation has slowed dramatically, from an average of more than 100 banks per year between 1990 and 2008 to just 13 for the eight years between 2010 and 2018.¹²⁶ While low interest spreads and higher post-crisis rates of regulatory burden account for some of the decline, the FDIC also toughened its capital and supervisory regime for new banks in 2009, discouraging newcomers’ entry into lending markets.¹²⁷ Since then, the stabilization of the financial system and expansion of the economy, together with new leadership at the FDIC, have created an opportunity to ease new charter policy for the benefit of depositors and borrowers.¹²⁸

In the meantime, the growth of online lending has further reduced the loan market share of CRA-subject depository institutions.¹²⁹ The volume of CRA lending has thus become less representative of overall credit conditions in LMI communities. Online lenders have devised ways to allocate credit profitably and competitively without an established relationship with prospective borrowers. Indeed, recent evidence suggests that online lenders can allocate credit more efficiently — with higher loan volumes at lower interest rates — than depository institutions.¹³⁰ Online lending has therefore reduced the potential for asymmetric information to lead to rationing in local credit markets.¹³¹ Other research suggests that online lenders tend to serve communities with a small number of banks and bank branches, which increases competition and credit availability in areas to which banks may not previously have fully catered.¹³²

The OCC’s proposed special-purpose national bank charter for fintech firms promises to make nationwide operations by nonbank lenders easier and less expensive (currently, the cost of state-by-state licensing and examinations can reach up to $30 million).¹³³ Comptroller Joseph Otting has previously estimated that as many as 30 to 40 online lenders could apply for a fintech charter.¹³⁴ Unfortunately, state-level legal challenges to the charter have led to policy uncertainty, discouraging firms from taking up the OCC’s offer for the time being.¹³⁵

Branching liberalization and the advent of online lending have allowed for freer local bank entry, substantially reducing the likelihood of persistently low lending rates in LMI communities. For example, as Table 6 shows, recent Home Mortgage Disclosure Act data reveal that on average, 26.2 percent of mortgages originated by the largest nonbank lenders (including fintech) are issued to LMI borrowers. Among those same lenders, 23.9 percent of all mortgage loans are issued to minorities. By comparison, LMI borrowers and minorities account for 20 percent and 22.2 percent of mortgages from the largest banks, respectively.¹³⁶ (Together, the top 25 banks and nonbanks — including mortgage companies and credit unions — account for 33.6 percent of all loan originations.¹³⁷) In short, market developments are already solving the primary issues that the CRA has spent the past 42 years trying to address.

Additionally, racial desegregation of many inner-city neighborhoods, itself a welcome development, has weakened the link between geography and CRA-targeted populations. For example, CRA lending in the historically black Philadelphia neighborhood of Point Breeze now seems to be reaching mostly newer (and better-off) white residents.¹³⁸ Because CRA regulations evaluate a census tract’s LMI status by comparing its median income with the median income of the metropolitan area, loans to better-off borrowers in LMI tracts still count for CRA assessment purposes.¹³⁹ Urban desegregation, perversely, has undermined the CRA’s effectiveness in promoting lending to vulnerable communities.

Let Fair Lending Laws Help

The objectives of the CRA remain vague and ill-defined. Regulators should clarify these objectives, both among themselves and to eligible institutions and community organizations. Is the goal of the CRA to fight lending discrimination, to increase lending in LMI communities, to raise living standards in LMI communities, or to achieve other public-interest goals?

If the CRA is supposed to fight discrimination, then the Fair Housing Act and the Equal Credit Opportunity Act are better tools, as they specifically address the disparate treatment of prospective borrowers according to race, gender, age, marital status, or other protected characteristics because they prohibit lending discrimination in the mortgage and small-business lending markets that the CRA targets.¹⁴⁰ There are concerns that the Consumer Financial Protection Bureau has been overbroad in its interpretation of the ECOA’s meaning in recent enforcement actions.¹⁴¹ Yet, unlike the CRA, the FHA and the ECOA focus on preventing the unfair treatment of individual vulnerable borrowers. Also unlike the CRA, they explicitly ban discriminatory practices and instruct financial regulators to prosecute violations.¹⁴² These are more efficient means of achieving public-policy goals than the CRA’s implicit requirement that banks lend in specific locations or risk having future expansion or merger applications rejected.

Increasing lending to poor communities is not a sound policy goal on its own, as it can encourage unprofitable loans that end up harming borrowers and bank balance sheets. There was a time when banks could profitably ration credit and exclude vulnerable populations due to branching restrictions and interest-rate caps. But the liberalization of bank branching in the 1980s and 1990s and the growth of nonbank lenders have increased competition in local banking markets and given consumers a more diverse set of credit options. Today, ensuring that public policies do not drive credit to borrowers who can ill afford it is as important as enabling financial institutions to serve all communities.

How Should the CRA Change if It Remains in Place?

There is reason to believe that the CRA is outdated and ill-suited to the current needs of LMI communities. Given how radically banking and credit markets in the United States have changed since 1977, Congress should strongly consider repealing the act. If the CRA remains in force, however, it would be a mistake to expand its mandate to cover nonbanks, such as fintech lenders and credit unions. Short of repealing the act, Congress and regulators should consider more efficient ways for depository institutions to discharge their CRA duties, such as by establishing a system of tradable lending obligations.

Allow Fintech Firms to Remain Exempt

The growth of online lending has led proponents of the CRA to call for the act’s extension to “branchless” fintech lenders.¹⁴³ Such an extension would not be possible under the present CRA evaluation framework, which defines eligible assessment areas as those where institutions have offices, branches, or ATMs.¹⁴⁴ Moreover, the rationale for mandating community reinvestment by banks — that they enjoy government deposit insurance — fails to apply to fintech and other nonbank lenders.¹⁴⁵

The proposed extension would also have negative practical effects. Nonbank fintech lenders have gained a substantial foothold in mortgage lending over the last decade, owing to their technological advantage as well as new regulations placed on banks.¹⁴⁶ Indeed, critics of the CRA — as well as some regulators — have warned of the act’s potentially adverse impact on depository institutions’ ability to compete with nonbank lenders.¹⁴⁷ The answer to these criticisms, however, is not to subject fintech lenders to the same CRA regulations: that would discourage their participation in marginal lending markets, which are the very markets that fintech lenders are more likely to serve than traditional lenders. Their withdrawal would have a disproportionately adverse impact on credit conditions and welfare in those communities.¹⁴⁸

In 1977, politicians justified the CRA by claiming that the government was underwriting bank credit risk and granting economic privileges to banks through restricted charters and federal deposit insurance.¹⁴⁹ That argument does not apply to fintech lenders, who neither hold charters nor enjoy the benefits of a taxpayer-guaranteed public deposit insurance scheme.¹⁵⁰ Extending the CRA to nondepository institutions such as fintech firms would therefore not create a level playing field. Rather, it would broaden the scope of a statute whose policy efficacy is already in doubt to institutions for which it was never intended.

Allow Credit Unions to Remain Exempt

Credit unions, which are exempt under the CRA’s current provisions, have recently become the target of similar calls for the act’s expansion. A bill introduced by Sen. Elizabeth Warren (D-MA) in September 2018 would have made credit unions, as well as nonbank lenders, subject to the CRA,¹⁵¹ although Warren subsequently revised her bill to remove credit unions from the set of institutions covered.¹⁵² The American Bankers Association, in a recent public filing with the OCC, likewise called for applying CRA regulations to credit unions.¹⁵³

Subjecting credit unions to CRA regulations would be counterproductive. In order to meet the conditions of the Federal Credit Union Act (FCUA), credit unions are already subject to restrictions on their activity that make them fundamentally different, for the CRA’s purposes, than banks.¹⁵⁴ The FCUA restricts credit union membership to groups that share a “common bond of occupation or association,” and to “persons or organizations within a well-defined community, neighborhood, or rural district.”¹⁵⁵ These common-bond provisions are at once redundant and incompatible with the CRA. Both acts are similar in that they aim to ensure that lending institutions serve their constituents. Yet the FCUA’s provisions would make enforcing the CRA among credit unions impossible: whereas CRA compliance relates to a bank’s lending activities within a given geographic region, the common bond that credit union members share under the FCUA may be professional, social, or demographic instead of geographic. Thus, credit unions are an example of the type of institutions that Comptroller Bloom, during the 1977 hearings, feared the CRA would undermine.

Additionally, there is evidence that credit unions already serve CRA-targeted populations. Since the financial crisis, the share of mortgages originated by credit unions has increased steadily, rising from 2.6 percent in 2007 to 8.7 percent as of mid-2018.¹⁵⁶ Recent HMDA data also show that credit unions originate a larger share of their mortgage loans to LMI borrowers than small banks do: 13.4 percent versus 12.5 percent.¹⁵⁷ Several factors could be behind these findings. For one, credit unions securitize a smaller portion of their loans than other mortgage originators, which may make them more sensitive to portfolio risk and lead them to spend more resources screening for creditworthy LMI borrowers.¹⁵⁸ Indeed, credit unions reject a larger share of mortgage loan applicants than do other institutions, which is consistent with the hypothesis of tighter screening owing to increased risk sensitivity.¹⁵⁹ Moreover, perhaps the increase in mortgage lending regulation has affected small banks more than credit unions, or perhaps banks are more vulnerable to nonbank lender competition than credit unions are. Finally, it could be that the FCUA’s common-bond provisions facilitate risk management by giving credit unions information about the credit quality of their borrowers that other institutions, whose customers need not share similar characteristics, cannot easily observe.

Credit unions appear to be achieving the CRA’s policy goals without being subject to its regulations. Applying the CRA to credit unions would impose substantial new compliance costs that are both unnecessary and incompatible with the nature of credit unions themselves. If policymakers are concerned about the changing business model of credit unions — particularly larger ones — the appropriate route to address such concerns is to revise the FCUA.¹⁶⁰

A Quantitative Score Has Clear Advantages — but Also Problems

The Treasury’s April 2018 memorandum on improving the CRA recommended “an approach to ... CRA that incorporates less subjective evaluation techniques.”¹⁶¹ The memorandum pointed out that relevant performance indicators in CRA assessments, “such as ‘excellent,’ ‘substantial,’ and ‘extensive,’ are undefined.”¹⁶² Other analysts have raised similar concerns on the use of “innovativeness” and “complexity” in the CRA investment test.¹⁶³ The OCC has subsequently suggested the use of a metric-based framework for CRA performance assessments.¹⁶⁴ While the details of a quantitative approach remain unclear, it would likely involve assigning CRA ratings based on the share of CRA-eligible loans, investments, and services in bank deposits, assets, or capital.¹⁶⁵

There are clear advantages to a metric-based approach. It would make assessments less arbitrary and provide greater certainty to institutions regarding the expectations of regulators. Quantitative assessment would also make it easier to compare performance between institutions and time periods. In these ways, a metric-based approach could reduce the administrative and compliance costs of the CRA.

Yet a metric-based approach also raises new concerns. Banks have warned that a quantitative method might not account for differences in business context across assessment areas.¹⁶⁶ Furthermore, in practice a metric-based approach would resemble a quota system for bank lending, investment, and services, unless regulators linked quantitative scores to qualitative judgements. Quotas, however, contradict the spirit of the CRA, which — in the words of Senator Proxmire — should not involve “costly subsidies, or mandatory quotas, or a bureaucratic credit allocation scheme.”¹⁶⁷

The advantage of a metric-based approach is that it would make it easier for banks to understand how best to demonstrate their LMI lending to regulators and estimate their performance in advance of an evaluation. But if regulators want to move toward quantitative forms of CRA assessment, there are more efficient ways to do so than a rigid quota scheme. Instead of fixed quotas, regulators should quantify the aggregate amount of CRA lending they expect to see in each assessment area and allow the most productive institutions to bid for their fulfillment.

Make CRA Obligations Tradable

If the CRA remains in place, there is a better way to encourage banks to improve the quality of the lending and other financial services they provide to LMI communities: create a market for tradable CRA obligations.¹⁶⁸ Under this system, the regulator would define the specific lending, investment, and services obligations among banks within a given assessment area. Obligations could be allocated in various ways, but for consistency with present CRA practice — which ties lending obligations to deposit-taking — it might be easiest to determine them according to an institution’s local deposit-market share. Lenders, including nondepository institutions such as fintech firms and community development financial institutions, would be able to bid for the obligation to fulfill CRA lending in exchange for a fee from banks.

A system of tradable CRA obligations would have several advantages over the current CRA enforcement regime. First, it would ask regulators to quantify the lending, investment, and services needs of the various LMI communities, thus introducing rigor into the assessment process and making explicit the community obligations of banks. As a result of trading in CRA obligations, a price would emerge to reflect the cost of fulfilling the act’s requirements. Importantly, the price of individual CRA obligations would vary according to the difficulty of profitably fulfilling them. For example, since it becomes more difficult to find creditworthy borrowers the more a community’s credit needs are satisfied, the price of a CRA lending obligation would rise as local CRA lending increased, serving as a useful bellwether for excessive lending in a particular community.

Second, a system of tradable obligations would encourage specialization and competition among lenders while ensuring that CRA obligations continued to be met.¹⁶⁹ The CRA as currently enforced deters specialization by requiring banks to lend roughly proportionately wherever they take deposits.¹⁷⁰ Under a tradable obligation regime, lenders from outside the assessment area, including those not subject to the CRA, would have an incentive to participate in the market if they could lend efficiently to local LMI communities. Given the role that fintech lenders play in providing credit to lower-income communities, for instance, their participation in such a trading scheme could increase the efficiency of CRA lending.¹⁷¹ As competition increased, the cost — that is, the market price for a representative obligation — of complying with the CRA would decline.

Third, tradable obligations would give CRA-subject banks increased opportunities for portfolio diversification. Because of the capital export rationale underpinning the 1977 act, the CRA currently forces banks to restrict some of their lending to the communities where they operate branches, even if they would like to lend elsewhere. For small banks in particular, the CRA’s local bias can impair geographic diversification. A system of tradable obligations, on the other hand, would enable depository institutions to lend in the locations best suited to their expertise and overall loan portfolio, while compensating other institutions for fulfilling CRA obligations on their behalf.

Fourth, a system of tradable obligations would reduce assessment uncertainty for depository institutions. Instead of grappling with a mounting list of ill-defined objectives, banks would discharge their obligations either by lending and investing directly, or by paying more efficient competitors to do so on their behalf. This would reduce compliance costs for CRA-eligible firms and reduce evaluation costs for the regulator. Meanwhile, communities would get what they need — or at least, what regulators think they need.

In fact, quantifying CRA commitments based on the needs of individual communities would require regulators to face an important challenge. The difficulty of ascertaining the level of unmet, yet profitable, credit demand is precisely why developed financial systems largely rely on markets — not regulators — to make determinations about credit allocation.¹⁷² As of now, the CRA forecloses this possibility. If regulators are required to set individual CRA obligations by region, however, they will likely have to do so in conjunction with banks and community groups in order to enlist their local market knowledge. This form of decisionmaking is still imperfect, as it would leave such obligations, and therefore regulatory compliance, vulnerable to interest-group pressures. But that is already the case under the current CRA’s uncertain and bureaucratic assessment regime. Moving to a system of tradable obligations would facilitate the benefits listed above while revealing the opportunity costs of the CRA — in terms of both foregone loans and overall safety and soundness.

Conclusion

The lending landscape in the United States has changed substantially since the 1977 enactment of the CRA. Written for what was then a competitive environment shaped by branching restrictions, the act took no account of the possibility that technological innovation would expand the opportunities for financial services provision. Today, the CRA is ill-suited to address the problems of unequal access to banking and credit as they currently affect low- and moderate-income borrowers.

In today’s landscape of widespread branching and diverse lending sources, the CRA has become a law in search of a public policy role. Congress should consider whether the benefits of preserving it justify the costs, or whether the act’s original goals can be (and already are) more effectively fulfilled through other channels. Fair lending laws can better prevent financial exclusion. Supply-side policies outside of financial regulation stand a greater chance of improving living standards in LMI communities. Perhaps the time of the CRA has simply passed.

However, if the CRA remains in place, policy­makers should take steps to make compliance less arbitrary and costly for banks. Implementing a system of tradable obligations that can be fulfilled by the most efficient lender at a market-determined rate combines the benefits of a clearly defined, quantitative approach with the flexibility and choice that America’s highly diverse credit market demands. Such a system would increase the efficiency of CRA enforcement and finally recognize that U.S. retail credit markets are much changed, and in many ways much improved, from the landscape that prevailed 42 years ago.

Notes

This paper is an expanded version of the author’s public filing with the Office of the Comptroller of the Currency. See Diego Zuluaga, “Reforming the Community Reinvestment Act Regulatory Framework,” Docket ID OCC–2018–0008, November 28, 2019.

¹ “Community Credit Needs: Hearings on S. 406, Before the Senate Committee on Banking, Housing, and Urban Affairs,” 95th Cong. (January 24, 1977)(statement of Robert Bloom).

² Community Reinvestment Act of 1977, 12 U.S.C. § 2901 (1977).

³ Raymond H. Brescia, “Part of the Disease or Part of the Cure: The Financial Crisis and the Community Reinvestment Act,” University of South Carolina Law Review 60 (2009): 627–28.

⁴ Jonathan R. Macey and Geoffrey P. Miller, “The Community Reinvestment Act: An Economic Analysis,” Virginia Law Review 79, no. 2 (March 1993): 292.

⁵ Bill Dedman, “The Color of Money,” Atlanta Journal-Constitution, May 1–4, 1988.

⁶ Redlining is “the practice of denying services, either directly or through selectively raising prices, to residents of certain geographies.” See Department of the Treasury, “Memorandum for the Office of the Comptroller of the Currency, the Board of Governors of the Federal Reserve System, the Federal Deposit Insurance Corporation: Community Reinvestment Act — Findings and Recommendations,” April 3, 2018.

⁷ Richard Marsico, “Democratizing Capital: The History, Law, and Reform of the Community Reinvestment Act,” New York Law School Review 49 (2004–2005): 723.

⁸ Sumit Agarwal, Efraim Benmelech, Nittai Bergamn, and Amit Seru, “Did the Community Reinvestment Act (CRA) Lead to Risky Lending?,” Kreisman Working Papers Series in Housing Law and Policy no. 8, October 2012, pp. 14–19.

⁹ Charles W. Calomiris and Stephen H. Haber, in Fragile by Design: The Political Origins of Banking Crises and Scarce Credit (Princeton: Princeton University Press, 2014), pp. 208–11, 216–26, document the ways in which the CRA influenced regulators’ decisions on bank mergers.

¹⁰ Michael Klausner, “A Tradable Obligation Approach to the Community Reinvestment Act,” in Revisiting the CRA: Perspectives on the Future of the Community Reinvestment Act (Boston/San Francisco: Federal Reserve Banks of Boston and San Francisco, September 2009).

¹¹ Office of the Comptroller of the Currency, “Fact Sheet: Community Reinvestment Act,” March 2014.

¹² 12 U.S.C. § 2901.

¹³ Department of the Treasury, “Memorandum,” April 3, 2018, p. 28.

¹⁴ Macey and Miller, “The Community Reinvestment Act,” pp. 322–23.

¹⁵ However, financial institutions face uncertainty about individual LMI loans’ eligibility for CRA credit. See ABA Banking Journal, “Bonus Podcast: Key Points on CRA Modernization,” American Bankers Association, November 14, 2018, podcast audio, 3:45, https://bankingjournal.aba.com/2018/11/bonus-podcast-key-points-on-cra-modernization/.

¹⁶ 12 C.F.R. 25.41 (1995, amended 2004).

¹⁷ Ben Horowitz, “Defining ‘Low- and Moderate-Income’ and ‘Assessment Area,’” in Community Dividend (Minneapolis: Federal Reserve Bank of Minneapolis, March 8, 2018).

¹⁸ Brescia, “Part of the Disease or Part of the Cure,” p. 634.

¹⁹ Darryl E. Getter, “The Effectiveness of the Community Reinvestment Act,” Congressional Research Service, January 7, 2015, p. 5.

²⁰ Federal Financial Institutions Examination Council (FFIEC), “Explanation of the Community Reinvestment Act Asset-Size Threshold Change.”

²¹ Community Reinvestment Act of 1977, 12 U.S.C. § 2908 (1977); and William C. Apgar and Mark Duda, “The Twenty-Fifth Anniversary of the Community Reinvestment Act: Past Accomplishments and Future Regulatory Challenges,” Federal Reserve Bank of New York Economic Policy Review (June 2003): 174.

²² 12 C.F.R. 25.22.

²³ 12 C.F.R. 25.23. The regulatory definition of community development includes affordable housing, community services aimed at LMI individuals, local economic development activities, and activities that revitalize or stabilize distressed areas. See 12 C.F.R. 25.12.

²⁴ 12 C.F.R. 25.24.

²⁵ 12 C.F.R. 25.23.

²⁶ 12 C.F.R. 25.26.

²⁷ 12 C.F.R. 25 Appendix A.

²⁸ 12 C.F.R. 25 Appendix A.

²⁹ Department of the Treasury, “Memorandum,” pp. 9–10.

³⁰ FFIEC, “Community Reinvestment Act: Interagency Questions and Answers Regarding Community Reinvestment,” 66 Fed. Reg. 36639, 33640 (July 12, 2001).

³¹ Apgar and Duda, “The Twenty-Fifth Anniversary of the CRA,” p. 174.

³² Brescia, “Part of the Disease or Part of the Cure,” p. 628.

³³ Brescia, “Part of the Disease or Part of the Cure,” p. 630.

³⁴ “Community Credit Needs: Hearings on S. 406,” Senate Committee on Banking, Housing, and Urban Affairs, 95th Cong. (January 24, 1977)(statement of William Proxmire).

³⁵ Proxmire, “Community Credit Needs,” pp. 9–10.

³⁶ Charles A. Calomiris, U.S. Bank Deregulation in Historical Perspective (New York: Cambridge University Press, 2000), pp. 61–62.

³⁷ Calomiris and Haber, Fragile by Design, pp. 183–95.

³⁸ “It’s a Wonderful Life,” Wikimedia Foundation, last modified March 3, 2019, https://en.wikipedia.org/w/index.php?title=It%27s_a_Wonderful_Life&oldid=886040649.

³⁹ Macey and Miller, “The Community Reinvestment Act,” pp. 303–11.

⁴⁰ Ross Levine, “Finance and Growth: Theory and Evidence,” in Handbook of Economic Growth, vol. 1A, eds. Philippe Aghion and Steven Durlauf (Amsterdam: North Holland, 2005).

⁴¹ Macey and Miller, “The Community Reinvestment Act,” pp. 307–10.

⁴² Calomiris, U.S. Bank Deregulation in Historical Perspective, pp. 22–28.

⁴³ Mark Carlson and Kris James Mitchener, “Branch Banking as a Device for Discipline: Competition and Bank Survivorship during the Great Depression,” Journal of Political Economy 117, no. 2 (April 2009): 169.

⁴⁴ Carlson and Mitchener, “Branch Banking as a Device for Discipline,” pp. 201–03.

⁴⁵ Michael D. Bordo, Hugh Rockoff, and Angela Redish, “The U.S. Banking System from a Northern Exposure: Stability versus Efficiency,” The Journal of Economic History 54, no. 2 (June 1994): 325–41.

⁴⁶ Marsico, in “Democratizing Capital,” pp. 724–25, argued that CRA regulators should assess compliance by comparing a bank’s share of loans to low-income communities with its competitors’ shares. Such a reform would significantly increase the role of regulation in credit allocation.

⁴⁷ Proxmire, “Community Credit Needs,” p. 11.

⁴⁸ Proxmire, “Community Credit Needs,” p. 13.

⁴⁹ Proxmire, “Community Credit Needs,” p. 14.

⁵⁰ Proxmire, “Community Credit Needs,” pp. 15–16.

⁵¹ Regulators have explicitly cited these two trends as reasons to review CRA enforcement. See Office of the Comptroller of the Currency, “Reforming the Community Reinvestment Act Regulatory Framework,” September 5, 2018, p. 45054.

⁵² Jith Jayaratne and Philip E. Strahan, “The Benefits of Branching Deregulation,” Regulation 22, no. 1 (Spring 1999): 10.

⁵³ Jayaratne and Strahan, “The Benefits of Branching Deregulation,” p. 10. The states that did not allow intrastate branching as of 1990 were Arkansas, Colorado, Iowa, Minnesota, and New Mexico. The states that still forbade interstate branching as of that year were Hawaii, Iowa, Kansas, Montana, and North Dakota.

⁵⁴ Riegle-Neal Interstate Banking and Branching Efficiency Act of 1994, H.R. 3841, 103rd Cong. (1993–1994).

⁵⁵ Margaret Z. Clarke, “Geographic Deregulation of Banking and Economic Growth,” Journal of Money, Credit and Banking 36, no. 5 (October 2004): 931–32.

⁵⁶ Federal Deposit Insurance Corporation, “Number of Institutions, Branches and Total Offices,” Historical Bank Data, 2019, https://www.fdic.gov/bank/statistical/. The number of FDIC-supervised commercial banks had dropped to 4,774 as of September 30, 2018. See Federal Deposit Insurance Corporation, “FDIC Statistics at a Glance,” September 2018, https://www.fdic.gov/bank/statistical/stats/2018sep/industry.pdf.

⁵⁷ Paul Calem, “The New Bank Deposit Markets: Goodbye to Regulation Q,” Federal Reserve Bank of Philadelphia Business Review (November/ December 1985), p. 20.

⁵⁸ R. Alton Gilbert, “Requiem for Regulation Q: What It Did and Why It Passed Away,” Federal Reserve Bank of St. Louis, February 1986.

⁵⁹ 76 Fed. Reg. 42015.

⁶⁰ Jayaratne and Strahan, “The Benefits of Branching Deregulation,” p. 11.

⁶¹ Clarke, in “Geographic Deregulation,” pp. 938–940, reports statistically significant increases in income growth of 1.2 percent as a result of branching deregulation, using the size of banks’ geographic market as a proxy.

⁶² Esteban Rossi-Hansberg, Pierre-Daniel Sarte, and Nicholas Trachter report a decreased local market concentration in the finance, insurance, and real estate sector that has accompanied the increase in concentration at the national level. See Esteban Rossi-Hansberg, Pierre-Daniel Sarte, and Nicholas Trachter, “Diverging Trends in National and Local Concentration,” NBER Working Paper no. 25066, National Bureau of Economic Research, Cambridge, Massachusetts, September 2018.

⁶³ Katherine Ho and Joy Ishii, “Location and Competition in Retail Banking,” International Journal of Industrial Organization 29, no. 5 (September 2011): 537–46; Astrid A. Dick, “Nationwide Branching and Its Impact on Market Structure, Quality, and Bank Performance,” Journal of Business 79, no. 2 (March 2006): 591.

⁶⁴ João Granja, Christian Leuz, and Raghuram Rajan, “Going the Extra Mile: Distant Lending and Credit Cycles,” NBER Working Paper no. 25196, National Bureau of Economic Research, Cambridge, Massachusetts, October 2018. Other scholars have noted a similar increase in the distance between borrower and lender in mortgage markets. See James Charles Smith, “The Structural Causes of Mortgage Fraud,” Syracuse Law Review 60 (2010): 473–503.

⁶⁵ 12 C.F.R. 25.61 (1997). This provision was added to the CRA with passage of the 1994 Riegle-Neal Act (H.R. 3841), which removed restrictions on interstate bank branching.

⁶⁶ 12 C.F.R. 25.65 (1997).

⁶⁷ Author’s private correspondence with OCC officials.

⁶⁸ Macey and Miller, “An Economic Analysis,” pp. 322–24.

⁶⁹ See, for instance, Raymond H. Brescia, “The Community Reinvestment Act: Guilty, but Not as Charged,” St. John’s Law Review 88, no. 1 (Spring 2014): 2–3; Michael S. Barr, “Credit Where It Counts: The Community Reinvestment Act and Its Critics,” New York University Law Review 80 (May 2005): 516; and Neil Bhutta and Daniel Ringo, “Assessing the Community Reinvestment Act’s Role in the Financial Crisis,” FEDS Notes, Board of Governors of the Federal Reserve System, May 26, 2015.

⁷⁰ National Community Reinvestment Coalition (NCRC), “CRA Commitments,” September 2007, pp. 21–22.

⁷¹ Edward J. Pinto, “Government Housing Policies in the Lead-Up to the Financial Crisis: A Forensic Study,” American Enterprise Institute discussion draft, February 5, 2011, pp. 5–6, http://www.aei.org/wp-content/uploads/2010/10/Pinto-Government-Housing-Policies-in-the-Lead-up-to-the-Financial-Crisis-Word-2003-2.5.11.pdf.

⁷² Calomiris and Haber, Fragile by Design, pp. 231–246; and Pinto, “Government Housing Policies,” p. 15.

⁷³ Pinto, “Government Housing Policies,” p. 14.

⁷⁴ Barr, “Credit Where It Counts,” pp. 566–67.

⁷⁵ Barr, “Credit Where It Counts,” pp. 568, 74.

⁷⁶ Barr, “Credit Where It Counts,” pp. 560–80. Barr gives a comprehensive review of econometric evidence on the effects of the CRA. Even the studies that find the CRA increased lending do not address the question of the loans’ impact on bank soundness.

⁷⁷ Community Reinvestment Act of 1977, 12 U.S.C. § 2903 (1977). See also NCRC, “CRA Commitments,” September 2007, p. 5, which recounts how CRA commitments by financial institutions ebbed and flowed with the waves of bank mergers in the late 1990s and 2000s.

⁷⁸ Agarwal et al., “Did the CRA Lead to Risky Lending?,” p. 3.

⁷⁹ Agarwal et al., “Did the CRA Lead to Risky Lending?,” p. 17.

⁸⁰ Agarwal et al., “Did the CRA Lead to Risky Lending?,” p. 22.

⁸¹ Lawrence B. Lindsey, “The CRA as a Means to Provide Public Goods,” in Revisiting the CRA: Perspectives on the Future of the Community Reinvestment Act (Boston/San Francisco: Federal Reserve Banks of Boston and San Francisco, September 2009), p. 164.

⁸² Bhutta and Ringo, “Assessing the CRA’s Role in the Financial Crisis.”

⁸³ Agarwal et al., “Did the CRA Lead to Risky Lending?,” p. 15.

⁸⁴ Neil Bhutta and Glenn B. Canner, “Mortgage Market Conditions and Borrower Outcomes: Evidence from the 2012 HMDA Data and Matched HMDA–Credit Record Data,” Federal Reserve Bulletin 4, no. 99 (November 2013): 42.

⁸⁵ Bhutta and Canner, “Mortgage Market Conditions and Borrower Outcomes,” p. 34.

⁸⁶ Anecdotal evidence that banks in CRA-eligible communities prefer to lend to newer, wealthier residents is consistent with the hypothesis that banks are “skimming the top.” See Aaron Glantz and Emmanuel Martinez, “Gentrification Became Low-Income Lending Law’s Unintended Consequence,” RevealNews.org, February 16, 2018, https://www.revealnews.org/article/gentrification-became-low-income-lending-laws-unintended-consequence/.

⁸⁷ Office of the Comptroller of the Currency, Board of Governors of the Federal Reserve System, and Federal Deposit Insurance Corporation, “Risk-Based Capital Standards: Advanced Capital Adequacy Framework and Market Risk; Proposed Rules and Notices,” 71 Fed. Reg. 55895, no. 185 (September 25, 2006). The Basel Committee on Banking Supervision is an international body that promulgates standards for the prudential regulation of banks. There is no economic reason why prudential standards should be laxer for CRA-related equity investments by banks than for their other investments.

⁸⁸ “Poverty, Public Housing and the CRA: Have Housing and Community Investment Incentives Helped Public Housing Families Achieve the American Dream?,” Subcommittee on Federalism and the Census of the Committee on Government Reform, U.S. House of Representatives (June 20, 2006)(statement of Judith A. Kennedy), p. 58.

⁸⁹ “The Performance and Profitability of CRA-Related Lending,” report by the Board of Governors of the Federal Reserve System, submitted to the Congress pursuant to section 713 of the Gramm-Leach-Bliley Act of 1999, July 17, 2000, p. 45.

⁹⁰ “The Performance and Profitability of CRA-Related Lending,” p. 51.

⁹¹ Pinto, “Government Housing Policies in the Lead-Up to the Financial Crisis,” p. 26ff.

⁹² FFIEC, “CRA: Interagency Questions and Answers,” p. 36639.

⁹³ Getter, “The Effectiveness of the CRA,” p. 9.

⁹⁴ Kenneth H. Thomas, “Dear Regulators: Don’t Take CRA’s Revamp Too Far,” American Banker, editorial, October 30, 2018.

⁹⁵ For a year-by-year summary of bank failures since 2001, see FDIC, “Bank Failures in Brief,” May 31, 2019, https://www.fdic.gov/bank/historical/bank/.

⁹⁶ Federal Reserve Bank of St. Louis, “Compliance Costs, Economies of Scale and Compliance Performance: Evidence from a Survey of Community Banks,” April 2018, p. 5.

⁹⁷ Federal Reserve Bank of St. Louis, “Compliance Costs, Economies of Scale and Compliance Performance,” p. 9. For the comparably poor performance of small banks, see, for example, FDIC, “Quarterly Banking Profile: Third Quarter 2018,” p. 7. FDIC-supervised institutions with fewer than $100 million in assets have an average return on equity of 8.28 percent, compared to 11 to 13 percent for larger banks.

⁹⁸ Raymond H. Brescia, “The Community Reinvestment Act: Guilty, but Not as Charged,” St. John’s Law Review 88, no. 1 (Spring 2014): 5–6.

⁹⁹ Dedman, “The Color of Money.”

¹⁰⁰ FDIC, “National Survey of Unbanked and Underbanked Households, 2017,” October 2018, pp. 17, 19, 24.

¹⁰¹ FDIC, “National Survey of Unbanked and Underbanked Households,” p. 11. A household is considered to have used mainstream credit if it used a credit card; a personal loan or line of credit from a bank; a store credit card; an auto loan; a student loan; a mortgage, home equity loan, or home equity line of credit (HELOC); or other personal loans or lines of credit from a company other than a bank in the past 12 months. The FDIC’s definition of mainstream credit does not include alternative financial services (AFS), such as money orders, check cashing, international remittances, payday loans, refund anticipation loans, rent-to-own services, pawn shop loans, and auto title loans (see p. 39).

¹⁰² FDIC, “National Survey of Unbanked and Underbanked Households,” p. 4.

¹⁰³ Federal Reserve Bank of St. Louis, “Compliance Costs, Economies of Scale and Compliance Performance,” p. 5.

¹⁰⁴ Federal Reserve Bank of St. Louis, “Compliance Costs, Economies of Scale and Compliance Performance,” p. 13.

¹⁰⁵ Drew Dahl and Michelle Franke, “‘Banking Deserts’ Become a Concern as Branches Dry Up,” Federal Reserve Bank of St. Louis, Regional Economist, Second Quarter 2017, pp. 20–21.

¹⁰⁶ Michelle Franke, “Who Would Be Affected by More Banking Deserts?,” Federal Reserve Bank of St. Louis, On the Economy (blog), July 17, 2017.

¹⁰⁷ Donald P. Morgan, Maxim Pinkovskiy, and Davy Perlman, “The ‘Banking Desert’ Mirage,” Federal Reserve Bank of New York, Liberty Street Economics (blog), January 10, 2018.

¹⁰⁸ Franke, “Who Would Be Affected by More Banking Deserts?”

¹⁰⁹ This is both because higher fixed compliance costs induce consolidation and higher regulatory costs raise the required return on a bank branch. See Julie Stackhouse, “Why Are Banks Shuttering Branches?,” Federal Reserve Bank of St. Louis, On the Economy (blog), February 26, 2018.

¹¹⁰ Michael Klausner, “Market Failure and Community Investment: A Market-Oriented Alternative to the Community Reinvestment Act,” University of Pennsylvania Law Review 143 (1995): 1565–68.

¹¹¹ Barr, “Credit Where It Counts,” p. 516.

¹¹² William W. Liang and Leonard I. Nakamura, “A Model of Redlining,” Journal of Urban Economics 33, no. 2 (March 1993): 223–34.

¹¹³ 12 U.S.C. § 2903.

¹¹⁴ Lindsey, “The CRA as a Means to Provide Public Goods,” p. 160.

¹¹⁵ NCRC, “CRA Commitments,” p. 6.

¹¹⁶ Calomiris and Haber, in Fragile by Design, pp. 216–17, cite the Fleet Financial-BankBoston merger of 1999 as an example of a time when good CRA performance caused the Fed to approve a merger despite concerns about its competitive effect.

¹¹⁷ FDIC, “Quarterly Banking Profile: Third Quarter 2018,” FDIC Quarterly 12, no. 4 (2018): 7.

¹¹⁸ Jayaratne and Strahan, “The Benefits of Branching Deregulation,” p. 14.

¹¹⁹ Lawrence J. White, “The Community Reinvestment Act: Good Goals, Flawed Concept,” December 18, 2008, p. 5.

¹²⁰ American Bankers’ Association, “ABA Data Bank: Economic Recovery Leaving De Novo Banks Behind,” ABA Banking Journal (website), September 28, 2018.

¹²¹ FDIC, “Enhanced Supervisory Procedures for Newly Insured FDIC-Supervised Depository Institutions,” FIL-50-2009, August 28, 2009.

¹²² Macey and Miller, “The Community Reinvestment Act,” p. 323.

¹²³ Pinto, “Government Housing Policies in the Lead-Up to the Financial Crisis,” p. 15.

¹²⁴ Agarwal et al., “Did the CRA Lead to Risky Lending?,” p. 3.

¹²⁵ For a comprehensive discussion, see Ryan Bourne, “Government and the Cost of Living: Income-Based vs. Cost-Based Approaches to Alleviating Poverty,” Cato Institute Policy Analysis no. 847, September 2018.

¹²⁶ American Bankers’ Association, “ABA Data Bank.”

¹²⁷ FDIC, “Enhanced Supervisory Procedures for Newly Insured FDIC-Supervised Depository Institutions.”

¹²⁸ FDIC Chairman Jelena McWilliams recently indicated interest in easing de novo bank entry. See Back to Basics, Federal Reserve Bank of Chicago 13th Annual Community Bankers Symposium, Chicago (November 16, 2018)(remarks of Jelena McWilliams).

¹²⁹ Apgar and Duda, “The Twenty-Fifth Anniversary of the CRA,” p. 180.

¹³⁰ Julapa Jagtiani and Catharine Lemieux, “The Roles of Alternative Data and Machine Learning in Fintech Lending: Evidence from the LendingClub Consumer Platform,” Federal Reserve Bank of Philadelphia Working Paper 18–15, April 2018, pp. 12–13.

¹³¹ Klausner, “Market Failure and Community Investment.” Klausner cites the canonical credit-rationing model in Joseph E. Stiglitz and Andrew Weiss, “Credit Rationing in Markets with Imperfect Information,” American Economic Review 71, no. 3 (June 1981): 393–410.

¹³² Julapa Jagtiani and Catharine Lemieux, “Do Fintech Lenders Penetrate Areas That Are Underserved by Traditional Banks?,” Federal Reserve Bank of Philadelphia Working Paper 18–13, March 2018, p. 12.

¹³³ U.S. Government Accountability Office, “Financial Technology: Additional Steps by Regulators Could Better Protect Consumers and Aid Regulatory Oversight,” Report to Congressional Requesters, March 2018, p. 45.

¹³⁴ Lalita Clozel (@laliczl), “OCC’s Otting,” Twitter post, February 7, 2019, 12:35 p.m., https://twitter.com/laliczl/status/1093563942050455553.

¹³⁵ Nutter, McClennen & Fish LLP, “Fintech in Brief: OCC Fintech Charter Continues to Face Legal Challenges,” January 30, 2019.

¹³⁶ Author’s calculations based on Bureau of Consumer Financial Protection, “Data Point: 2017 Mortgage Market Activity and Trends,” May 2018, pp. 70–72.

¹³⁷ Bureau of Consumer Financial Protection, “Data Point: 2017 Mortgage Market Activity and Trends,” May 2018, p. 64.

¹³⁸ Glantz and Martinez, “Gentrification Became Low-Income Lending Law’s Unintended Consequence,” RevealNews.org, February 16, 2018.

¹³⁹ Horowitz, “Defining ‘Low- and Moderate-Income’ and ‘Assessment Area.’”

¹⁴⁰ 15 U.S C. § 1691.

¹⁴¹ Daniel Press, “The CFPB and the Equal Credit Opportunity Act,” Competitive Enterprise Institute, On Point (blog), May 15, 2018.

¹⁴² 15 U.S.C. § 1691c.

¹⁴³ Kenneth H. Thomas, “Why Fintechs Should Be Held to CRA Standards,” American Banker, editorial, August 24, 2018.

¹⁴⁴ 12 C.F.R. 25.41.

¹⁴⁵ Macey and Miller, “The Community Reinvestment Act,” p. 313.

¹⁴⁶ Greg Buchak, Gregor Matvos, Tomasz Piskorski, and Amit Seru, “Fintech, Regulatory Arbitrage, and the Rise of Shadow Banks,” NBER Working Paper no. 23288, National Bureau of Economic Research, Cambridge, Massachusetts, September 2018. The authors find that 60 percent of the growth of “shadow banks” is due to regulation, whereas 30 percent is due to technology.

¹⁴⁷ Macey and Miller, “The Community Reinvestment Act,” pp. 312–13; Bloom, “Community Credit Needs,” pp. 15–16.

¹⁴⁸ Jagtiani and Lemieux, “Do Fintech Lenders Penetrate Areas That Are Underserved by Traditional Banks?,” p. 10.

¹⁴⁹ Proxmire, “Community Credit Needs,” pp. 9–10.

¹⁵⁰ Additionally, as discussed earlier, deposit-taking institutions no longer operate local monopolies or oligopolies, because of the removal of branching restrictions.

¹⁵¹ American Housing and Economic Mobility Act of 2018, S. 3503, 115th Cong. (2018).

¹⁵² American Housing and Economic Mobility Act of 2019, H.R. 1737, 116th Cong. (1st Sess. 2019).

¹⁵³ Krista Shonk, “Reforming the Community Reinvestment Act Regulatory Framework,” American Bankers Association comment letter to the Comptroller of the Currency, November 15, 2018, p. 33, https://www.aba.com/Advocacy/commentletters/Documents/cl-CRA20181115.pdf.

¹⁵⁴ Federal Credit Union Act of 1934, 12 U.S.C. §§ 1752–1775 (1934).

¹⁵⁵ Federal Credit Union Act of 1934, 12 U.S.C. § 1759 (1934).

¹⁵⁶ Mortgage Bankers’ Association (MBA) and Credit Union National Association (CUNA) data. The author is grateful to Mike Schenk from CUNA for sharing it.

¹⁵⁷ James DiSalvo and Ryan Johnston, “Credit Unions’ Expanding Footprint,” Banking Trends (Philadelphia: Federal Reserve Bank of Philadelphia, First Quarter 2017), p. 20.

¹⁵⁸ Securitization rates are around 35 percent for credit unions and 70 percent for all mortgage originators. See CUNA data (note 161) and Urban Institute, “Housing Finance at a Glance: A Monthly Chartbook,” research report, June 2018.

¹⁵⁹ DiSalvo and Johnston, “Credit Unions’ Expanding Footprint,” pp. 19–20.

¹⁶⁰ Aaron Klein, “Banklike Credit Unions Should Follow Bank Rules,” American Banker, editorial, June 25, 2018.

¹⁶¹ Department of the Treasury, “Memorandum for the OCC,” p. 11.

¹⁶² Department of the Treasury, “Memorandum for the OCC,” p. 9.

¹⁶³ Getter, “The Effectiveness of the CRA,” p. 8.

¹⁶⁴ Office of the Comptroller of the Currency, “Reforming the Community Reinvestment Act Regulatory Framework,” Advance Notice of Proposed Rulemaking, 83 Fed. Reg. 172 (September 5, 2018): 45053.

¹⁶⁵ Office of the Comptroller of the Currency; Shonk, “Reforming the CRA Regulatory Framework,” pp. 13–14.

¹⁶⁶ Shonk, “Reforming the CRA Regulatory Framework,” pp. 11–12.

¹⁶⁷ Proxmire, “Community Credit Needs,” p. 9.

¹⁶⁸ See Klausner, “Market Failure and Community Investment,” and Klausner, “A Tradable Obligation Approach,” for the original proposals that inform the approach outlined in this section.

¹⁶⁹ Klausner, “Market Failure and Community Investment,” pp. 1586–88.

¹⁷⁰ Klausner, “Market Failure and Community Investment,” pp. 1575–76.

¹⁷¹ Jagtiani and Lemieux, “Do Fintech Lenders Penetrate Areas That Are Underserved by Traditional Banks?,” p. 10.

¹⁷² F. A. Hayek, “The Use of Knowledge in Society,” American Economic Review 35, no. 4 (September 1945): 519–30.