Reading the comment letters from industry and community groups regarding the Federal Reserve Board’s (board) Advanced Notice of Proposed Rulemaking on the Community Reinvestment Act (CRA), I am reminded of the childhood tale of the “Little Engine that Could,” in which a line of rail cars get caught at a bottom of a big hill. A little blue engine was the only engine that volunteered to pull the cars and did so by repeating the mantra: “I think I can, I think I can.”
One could characterize the comments reacting to the board’s ideas for changing CRA this way: the industry kept repeating, “No, we can’t,” while community groups affirmed that “Yes, you can” reinvest and meet the challenges of revitalizing communities of color and low- and moderate-income (LMI) communities disproportionately afflicted by the pandemic.
CRA reform is prompted, in part, by technological innovations and the emergence of online banking. Community groups are confident that these advances can be used to promote reinvestment in underserved communities, and to hold banks accountable to a higher standard. Banks, in contrast, worry that they do not have the ability to handle more data collection requirements or geographical areas on their exams in an effort to make more loans, investments and services in underserved communities.
Banks want asset thresholds raised to create streamlined exams
An immediate indication of the divergent views on bank capacity involves the wonky-sounding issue of asset thresholds. This refers to asset levels that determine whether a streamlined small bank exam, an intermediate small bank (ISB) exam or a more comprehensive large bank exam is used to evaluate a bank’s CRA performance. Currently, a bank receives a small bank exam that looks only at retail lending if it has assets of $330 million or less (adjusted annually for inflation). It undergoes an ISB exam if it has assets between $330 million and $1.3 billion; such an exam looks at retail lending and community development financing, but not branching. A bank is subject to a large bank exam if it has more than $1.3 billion in assets; this exam considers retail lending, community development financing, bank services and branching.
The board proposed an adjustment to the small bank asset level to either $750 million or $1 billion to mark the difference between a small and large bank, while the ISB category would be eliminated. In affirming that “Yes, banks can,” NCRC documented that raising the current threshold would cost smaller metropolitan areas and rural counties precious community development financing dollars because they are disproportionately served by ISB banks. Every three years, these areas could lose between $5.6 billion to $7.8 billion in community development financing depending on which asset level threshold is chosen, according to NCRC’s analysis. Moreover, these ISB banks can continue to provide critical community development financing, which they have done for several years.
The American Bankers Association (ABA) and the Independent Community Bankers Association (ICBA) called for the board to exceed its proposed adjustment to the small bank threshold. The ABA and ICBA stated that about 20% of banks were ISB banks and 70% were small banks in 2005, when the agencies first established the ISB category. In order to maintain these percentages, the small bank asset threshold would be raised to $600 million and the ISB threshold would be raised to $2.5 billion. The ICBA would like to go even further and set the ISB threshold at $10 billion so ISB banks are not compared to the largest banks in the country.
There is nothing inviolate in the original 70/20 split that should guide CRA policy in perpetuity. More pertinent is whether banks have the capacity to perform under their CRA exams. In this case, as shown by the high levels of community development financing by ISB banks, the answer is “Yes, they can.” In the aftermath of a pandemic, agencies should preserve a precious source of community development financing that is of similar magnitude as federal housing and community development programs operated by the Department of Housing and Urban Development.
Furthermore, extending ISB status for banks with assets up to $10 billion would eliminate the service test and examination of branches, which the Consumer Bankers Association admitted “are often vital to LMI communities,” in terms of access to loans and banking services. Comparing CRA exams of banks of up to $10 billion in assets and those of the largest banks in the country would convince most impartial observers that the exams of the smaller large banks are not as sweeping in terms of the number of geographical areas examined or in the expectations regarding levels of retail lending or community development financing. The agencies can also tailor expectations or numerical benchmarks to avoid any unrealistic comparisons in the future. Given the alternatives, extending ISB status for banks with assets up to $10 billion would be the incorrect solution and would be costly to communities.
Banks want little new data collection despite technological advances
Asset thresholds trigger data collection requirements. If the agencies designate more banks as large as opposed to small based on asset levels, the banks will have more data collection requirements including submitting CRA data on small business and farm lending. The ABA objected, saying that by eliminating the ISB category “…hundreds of [former] Intermediate Small Banks would be subject to the same new data collection, maintenance and reporting requirements as the largest banks in the country.”
In reply to this “No, we can’t,” assertion, NCRC maintains that the banks can readily report small business and farm data. NCRC conducted a study that used some of the last years of small business loan data for small banks before the agencies eliminated the data reporting requirement in 2005 for the Appalachian Regional Commission (ARC). NCRC found that these banks had an important niche in small business lending in rural counties. Without this data now, it is much more difficult for the public to ascertain if small and ISB banks retain this role, which is important to understand for policymakers and reinvestment practitioners. Moreover, a segment of small banks (upwards of 200 annually) voluntarily submit this data to the federal bank regulators in order to have it considered on their CRA exams.
The ICBA asserted that data reporting requirements and other compliance burdens fueled the consolidation of banks and caused smaller banks to merge with their larger brethren. This assertion is not supported by empirical evidence. The Consumer Financial Protection Bureau recently examined the cost issue and concluded that data collection costs for small business lending were modest for smaller banks. If the banks charged borrowers more to cover these costs, the additional fees paid by borrowers would be about $5 per loan.
If CRA reform is to hold banks accountable for more retail lending and community development financing, the data collected needs to be enhanced, not subtracted. There is a great need for better small business loan and community development financing data as well as deposit data. With advances in technology and careful development of data reporting requirements by the agencies, much-improved data can be collected in a cost-effective matter. The CBA, which represents larger banks, was not quite as adamant against data collection as the ICBA but its watchword was “optionality;” for the board to make most of this new reporting optional for banks that want to earn bonus points on CRA exams. The downfall of this approach, however, is creating inconsistent and subjective exams.
The National Association of Affordable Home Lenders (NAAHL) affirmed that important data enhancements are desirable and practical. NAAHL maintained, “Using depositor addresses is technically achievable and integral to modernizing CRA as the banking industry continues to evolve beyond a traditional branch-based business model.” NAAHL continued, “Collection and reporting of community development (CD) data at the county level will be necessary to provide data to the public, especially since levels of CD activity will, appropriately, vary among counties within an MSA.” Both deposit and CD data will need to be reported in this manner if the proposed CD test is to be effective in leveraging more CD financing, particularly in areas of need.
Banks want few new assessment areas and national evaluations that overlook local needs
While the views about data collection were mixed among the bank trade associations, they were unanimous in disliking the board’s proposals to increase the number of geographical areas on exams. Despite the touted benefits of online and branchless banking in reaching underserved populations, the industry comments seemed to suggest that reinvestment obligations should not be bolstered, although bank abilities to meet obligations will improve in future years due to technological and other enhancements.
NCRC is clear that new assessment areas (AAs), which are what geographical areas on CRA exams are called, should be designated to include areas with significant numbers of loans and/or deposits for branchless banks or hybrid banks with both branches and non-branch delivery methods. The board would call these new AAs lending- or deposit-based AAs. NCRC recently wrote a paper describing how the metric loans per capita could be used to develop a feasible number of AAs for branchless or hybrid banks that include smaller metropolitan areas and rural counties as well as larger metropolitan areas.
The ABA responded to the board’s AA proposal by stating, “Moreover, in updating the assessment area construct, regulators should take great care not to dilute a bank’s overall CRA impact in communities by creating numerous new assessment areas.” Such willy-nilly creation of numerous AAs also may interfere with safe and sound lending by requiring banks to be examined for their performance in serving LMI borrowers, according to the CBA.
Such AA creation is just not feasible, maintained NAAHL. Lending- and deposit-based AAs are the “wrong paradigm” for evaluating activity that is not “inherently local.”
The “No, we can’t” chorus seemed to contradict the lofty praises that industry stakeholders have heaped upon fintechs, which are branchless lenders. Some stakeholders extol the ability of fintechs to reach underserved populations. Now, others suggest that fintechs do not have the technological ability to submit data for use in CRA exams or to serve LMI borrowers in a safe and sound manner in new AAs, as CBA claimed.
However, underwriting mortgage loans involves well-understood metrics and standardization (thanks to the Dodd-Frank Act’s rules on anti-predatory lending and the presence of the government sponsored enterprises, Fannie Mae and Freddie Mac). In addition, some fintech lenders have successfully employed their proprietary underwriting models to more accurately predict and account for risk in consumer lending. Risk and non-branch delivery should be monitored and regulated closely, but it rings hollow to wave the safety and soundness flag in claiming that non-branch delivery is ill-suited to serve LMI borrowers especially after all the hype associated with this approach.
NAAHL and the ABA asked the board to investigate whether the branchless banks should have a nationwide evaluation or use tailored metrics for distinct geographical markets. The ABA and NAAHL were also sensitive to competitive disadvantages that may arise should the branchless banks be evaluated on a national level and traditional banks continue to have local AAs. NAAHL stated that the national model may not be the most “fair” but is “simpler.”
If technology and digital marketing can help branchless lenders reach LMI borrowers and communities, then local AAs are feasible and desirable. Finally, if branchless banks do not perform as well in these communities (some preliminary evidence from NCRC research suggested this is the case), CRA exams can weigh comparisons among branchless banks more than comparisons among branchless banks and the industry as a whole in retail lending.
Lastly, the board was not blameless for the industry blowback against its proposal for lending- and deposit-based AAs. The prose was clumsy and incomplete in describing the implications of its proposal. The board stated that “only” 157 banks would have to create additional AAs under one of its proposals, suggesting that it was disappointed that its proposal did not include more banks. The ABA took the board to task, saying the goal should not be to maximize the number of banks complying with a new AA requirement but rather should be updating the regulation to account for non-branch lending. The description prompted ABA to discuss worries that the board’s ambition was to add several AAs for several banks.
The board’s lack of quantification set up this response. NCRC estimated, on average, these 157 banks would be required to establish between 10 to 12 AAs while many would designate less. Overall, 157 banks would be a small minority of all banks and the number of new AAs is feasible. The board needs to more fully describe its proposal if it wants to save its important concept of trying to hold non-branch lending accountable for serving LMI borrowers and communities. A national approach to AAs would allow banks to “cherry-pick” the easiest areas in which to serve LMI borrowers and communities, neglecting urgent needs in underinvested areas.
We can and must reinvest
CRA exams must remain robust and comprehensive in terms of the tests for all sizes of banks, the data collected to be evaluated on tests, and the numbers of geographical areas on exams. Unjustified shortcuts in any of these areas will result in fewer loans, investments and services for underserved communities trying to recover from the pandemic and will exacerbate racial inequities. The shortcuts would be driven in response to the “No, we can’t” mantra. However, the little blue engine has been tugging and reinvesting for decades. This little engine can undergo more robust CRA tests as it has been upgraded in terms of its technology and power. CRA reform should and must affirm, that “yes, we can and must reinvest!”
Josh Silver is a senior policy advisor at NCRC.