Initial NCRC Analysis of the FDIC and OCC Notice of Proposed Rulemaking Concerning the Community Reinvestment Act

In December 2019, the Office of the Comptroller of the Currency (OCC) and the Federal Deposit Insurance Corporation (FDIC) issued a notice of proposed rulemaking (NPRM) that would considerably weaken the regulations implementing the Community Reinvestment Act (CRA), a law designed to combat redlining by requiring banks to affirmatively and continually meet community needs for credit and banking services in low- and moderate-income (LMI) communities. The law and regulation would be diluted in significant ways:

• It would broaden what bank activities are CRA-qualifying so that CRA dollars may look like they are increasing but may actually decrease for people that are LMI and LMI communities.
• It would significantly dilute focus of bank activities on LMI consumers and communities.
• It contains a proposal for expanding geographical areas on CRA exams whose impact cannot be assessed due to a lack of publicly available data.
• It sets arbitrary performance thresholds based on data absent from the NPRM or not available – unclear whether banks will get better CRA ratings for more, less or the same level of activities.
• Banks could fail in almost half their assessment areas and still pass.
• It limits consideration of bank branches more than under the current CRA service test. Banks may respond by closing more branches in LMI communities.
• Retail lending analysis would count for much less under the new proposed exams.
• It would reduce consideration of qualitative criteria that measure responsiveness to local needs.
• Small banks would have an option for a streamlined exam while the agencies suggest that they can perform well on the new exams.
• Banks that achieve an Outstanding rating would be examined once every five years, an increase from the two to three year cycle today. They would have less incentive to perform well in the early years of the new exam cycle.

Instead of updating CRA, the agencies would lessen the public accountability of banks to their communities by enacting performance measures on CRA exams that would simultaneously be complex and opaque while at the same time over-simplifying how to measure bank’s responsiveness to local needs. The result will be significantly fewer loans, investments and services to LMI communities most in need of more credit and capital.

As FDIC board member Martin Gruenberg stated, “This is a deeply misconceived proposal that would fundamentally undermine and weaken the Community Reinvestment Act.”^[1]

The NPRM is divided into various sections regarding: what activities are counted, where activities are counted, and how activities are counted. In brief, the OCC and FDIC would expand which activities are counted and in doing so would divert the focus away from LMI communities. Regarding where activities are counted, the OCC and FDIC propose an expansion of assessment areas (AAs) or geographical areas scrutinized on exams that may appear to capture more bank activity but relies on data that is not yet collected. Its feasibility is thus in question. Thirdly, regarding how activities are measured, the OCC and FDIC continue to insist on a one ratio measure that will make it much harder to determine if banks are responding adequately to various needs in their AAs. The OCC and FDIC preserve a retail lending test, but its weight on the exam has been dramatically reduced. In addition, the NPRM would allow a bank to fail in up to 50% of its AAs and still pass overall.^[2]

The major parts of the proposal are now considered in turn:

What Counts

The original impetus of CRA was to combat redlining and systematic discrimination against LMI communities and communities of color. As a result of decades of public and private sector discrimination, lending markets in many LMI communities were not functioning. Banks were not familiar with the creditworthiness of borrowers, the quality of the housing stock or the viability of small businesses in these communities. To overcome the market failures caused by discrimination, CRA imposed an affirmative obligation on all banks to serve LMI communities. By requiring banks to seek out business in LMI communities, CRA sought to break down barriers to information that contributed to a scarcity of lending. All banks were required to learn about and collect information and data about LMI communities so that they could make safe and sound loans. As a result of this affirmative obligation, Federal Reserve research has concluded that CRA increased home lending by about 20% in LMI census tracts and also increased small business lending.^[3]

The FDIC and OCC’s proposal will diminish if not halt this progress by allowing banks to turn their attention away from LMI communities. A major change in the proposed regulations regards the definition of community development. In addition to retail lending (home and small business lending), LMI communities need community development financing which supports affordable housing, job creation, small business growth, facilities like child care and health clinics and larger scale projects like the development of commercial corridors. Home loans or small business loans will not succeed in creating an economically thriving and vibrant community if the community lacks community development financing. If the agencies reduce the focus of community development on LMI communities, the regeneration of lending and housing markets will be halted and market failure will re-emerge.

The current definition says community development is affordable housing for LMI households, economic development focused on small businesses under $1 million in revenue, community facilities and the revitalization and stabilization of LMI communities. The OCC and FDIC would delete the criteria of economic development and revitalization and stabilization from the definition of community development. In practice, this means that banks would reduce the financing of infrastructure such as business incubators that target start-up businesses and revitalization and stabilization activities such as reclamation of abandoned housing or instituting foreclosure prevention programs. These activities are vital to the continued development of vibrant housing and lending markets in LMI communities.

In addition to deleting economic development and revitalization/stabilization, the agencies introduced criteria that will dilute the focus on LMI communities. For example, a new criterion will be essential infrastructure. This refers to major projects like roads and bridges. The difficulty is that these projects are not necessarily located in LMI communities.

The NPRM discusses a public hospital as a project eligible for CRA consideration under the community development definition. The NPRM, however, does not discuss whether such a hospital needs to be in a LMI community or within easy access of it.^[4] If the definition of community development encourages banks to favor financing of infrastructure which is needed by a city as a whole (a broad definition of community) instead of infrastructure targeted to LMI communities, then the revitalization of LMI communities financed by CRA will slow down. The NPRM suggests that, “The addition (of infrastructure) also would recognize that essential infrastructure projects are often community-wide projects for which it is not feasible to allocate the benefit to specific populations or areas.”^[5]

The FDIC and OCC also would include community development financing that partially instead of primarily benefiting LMI households as eligible for CRA consideration. A pro rata procedure would apply. For example, if a mixed-income housing development has 40% of the units reserved for LMI households, the total dollar amount would be multiplied by 40% to determine the dollar amount considered on CRA exams. The pro rata procedure currently is applied in circumstances like this. Nevertheless, the agencies need to clarify that it will continue to be applied in these circumstances; the list of qualifying activities introduce confusion since it lists affordable housing with various percentages of the units being affordable for LMI households, and does not indicate clearly that a pro rata procedure will be applied.^[6]

In addition, the introduction of large-scale infrastructure may make this procedure less reliable. When a pro rata percentage is applied, how would the agencies calculate a pro rata share for LMI populations of a major road or bridge used by millions of people? The result of back-of-the envelope guesses of pro rata share in these cases would likely be overestimating LMI benefit.

Other aspects of what counts that dilute attention to LMI households and communities include:

• Deposit accounts and other bank services would no longer be considered qualifying activities and hence not even considered by CRA exams. Deposit accounts provide a safe and affordable place for LMI customers to store their money and is the first essential step to establishing a banking relationship. CRA exams need better data and more consideration for deposit accounts and banking services. Discarding consideration of bank accounts and services would be a major step backwards for underserved communities in that banks would deemphasize bank services in LMI communities, making them more dependent on check cashers and other high cost fringe services.
• Counting financial education that benefits all income groups, even though people with LMI is the population group that is disproportionately unbanked or underbanked, according to the FDIC’s own studies.^[7]
• Excluding consideration of home lending in LMI communities. The Comptroller has emphasized that he does not want lending to middle- and upper-income people in LMI communities to displace LMI residents through gentrification. However, creating diversity and integration by income is desirable. NCRC has proposed other methods to deal with avoiding displacement in gentrifying communities (see NCRC’s ANPR letter).^[8]
• Raising the threshold for what counts as a small business from $1 million to $2 million in revenue.^[9] While adjusting the loan size that counts as a small business loan to account for inflation is reasonable, an adjustment to the revenue size of the small business is not justified by research. The CFPB, in fact, finds that the great majority of small businesses have revenues under $1 million.^[10] An example in the list of qualifying activities is a loan to a family farm with revenues of $10 million; this seems too high and would invite additional examples to be added to the list including farms with higher annual revenues.^[11]
• The definition of community development services has been amended to allow for all volunteer activities such as hammering nails or being a docent at a museum.^[12] The current definition is a service related to the provision of financial products for people with LMI. It is more desirable to direct the expertise of bank staff to financial counseling than general volunteer projects in the community.
• The proposal applies a multiplier to almost all community development activities expect purchasing Mortgage Bank Securities (MBS) and municipal bonds. A multiplier can result in less financing, which even a question on the NPRM acknowledges.^[13] A better approach than multipliers would be to apply less weight to activities deemed less responsive to need. NCRC also believes that multipliers should not be applied to banks that have decreased their annual levels of community development, at both the overall bank level and assessment area level, in order to prevent banks using multipliers to simply decrease their community development activity.
• Allowance for affordable housing to middle-income people in high cost areas. Currently, an Interagency Q&A discusses this activity. Elevating this to part of the regulatory definition will divert too much financing away from LMI affordable housing, especially in high cost areas. The list of qualifying activities includes housing produced under inclusionary zoning in which 30% of the units are for middle-income households in high-cost areas.^[14] The remaining units can be for upper-income households. Limits need to be established as to the portion of units for middle-income households or else it is possible that virtually all of the “affordable” housing in several high-cost areas financed by CRA will be for middle-income households.
• The agencies will now allow naturally occurring affordable housing (NOAH) to be considered as affordable housing. NOAH is rental housing in which LMI households would have to pay no more than 30% of 80% of area median income on monthly rent. The agencies need to consider how to make sure LMI households actually occupy the units.^[15]
• The definition of community development must not be relaxed in the case of Opportunity Zones, including when the census tracts are low-income. The Opportunity Zone program lacks documentation or data to specify who benefits from the financing. Thus, if CRA financing is not constrained to meet the definition of community development, instead of affordable housing, the financing could be for luxury condominiums. The NPRM does not seem to include these safeguards. In fact, the NPRM would allow “athletic stadiums” to be financed in LMI tracts.^[16]
• The agencies created a new category of tracts called underserved areas that are middle-income tracts with low population levels and without easy access to branches.^[17] The agencies conduct no research to justify the creation of these new tracts eligible for CRA activities. NCRC has a proposal that focuses on low levels of lending and is more effective in identifying lower-income and middle-income as well as communities of color that are underserved.
• The list of qualifying activities only must be updated in a transparent manner so the public can have knowledge of all updates. When an agency adds an activity in response to a bank request, the agency must publish a notice to the public that includes a rationale for inclusion. Most of the updating should be conducted via public notice and comment periodically, perhaps every two years, instead of three years as mentioned by the agencies.  

Where it Counts – Assessment Areas

Presently, CRA exams define assessment areas as geographical areas including bank branches and where a substantial amount of lending activity occurs. The OCC and FDIC would retain this definition. They would call these assessment areas (AAs) “facility-based” AAs. In addition, the FDIC and OCC would establish “deposit-based” AAs. These AAs would apply to either internet-based banks or traditional banks that also conduct much business over the internet. Under the procedures for designating deposit based AAs, the bank would be required to establish these AAs if more than 50% of the bank’s deposits were collected beyond branch networks. Also, the bank would be required to designate additional geographical areas as AAs if 5% or more of their deposits came from these area(s). The bank would be required to use the smallest geographical area(s) (whether it be a county, metropolitan area or state) that generated 5% or more of their deposits.^[18]

While it is commendable that the FDIC and OCC recognized the need to update AAs for banks that conduct a significant amount of their business on-line, the proposal is not fully developed and may exacerbate credit deserts. Firstly, data is not currently collected in an accurate manner concerning deposits that are generated through non-branch means. Banks now arbitrarily assign deposits collected via the internet to branches. The OCC and FDIC would need to implement a rulemaking to establish procedures for accurately collecting and disseminating this data.

The FDIC and OCC offer no data analysis estimating how many banks the new AA procedure would apply to and how many additional AAs would be created. Discussing this aspect of the rule in the NPRM, the FDIC concludes, “It is difficult to accurately quantify these aspects of the proposed rule with the information currently available to the FDIC.”^[19]

The 50% is probably too high of a threshold. Currently, AAs capture about 70% of bank home lending so it would stand to reason that AAs also capture a similar percentage of their deposits.^[20] Given this, it may only be a handful of banks that would need to establish AAs. The threshold should be based on clearly described research and reporting of the impacts.

Another issue is that 5% of a bank’s deposits is too high of a bar. In particular, many rural counties or smaller cities would not qualify for AA status, particularly for very large banks. Thus, this threshold would likely exacerbate the problem of credit and branching deserts for the underserved communities for which the OCC and FDIC express concern. A better threshold would be the bank’s market share of deposits, which would likely create more AAs for large banks in underserved areas. A market share threshold would need to be established by research and data analysis.

Finally, using HMDA and CRA small business data is currently a more reliable method to establish AAs outside of branch networks because this data exists and is accurate for the purposes of establishing AAs outside of branch networks. It also reflects where banks are engaged in significant business activity.

In a careless manner, the NPRM also appears to open up “credit for qualifying activities conducted outside of their assessment areas.”^[21] This is a complex issue that deserves more development than one sentence in a NPRM. The agencies have already developed a procedure for banks to receive credit outside of their AAs in statewide and regional areas after they have demonstrated they have been responsive to needs in their AAs. While the current approach would benefit from additional objectivity regarding how banks demonstrate responsiveness, giving credit outside of AAs in an unspecified manner is not the solution. NCRC has suggested adding underserved counties as areas outside of AAs that a bank can serve with community development finance. We also advocate the use of annual community development data (which would be collected per the NPRM) in helping to determine if a bank has adequately served its AA.^[22]

The NPRM’s approach to allowing banks to serve areas outside of their AAs will encourage banks to gravitate to national level intermediaries that can put together the largest deals instead of working with locally-based CDFIs and other nonprofits on smaller deals that are more directly responsive to the needs in their localities. The OCC and FDIC may respond by saying that this concern is addressed by their suggested retail lending test and established minimums of community development activity at the AA level, but it’s important to recall that under this new proposal, banks would only need to reach satisfactory CRA performance in half of their assessment areas.

How it Counts – Performance Measures

The agencies set up two examination standards:  small bank performance standards and general performance standards.

• Small bank performance standards apply to smaller institutions with $500 million or less in assets opting to be examined under current CRA regulations for small banks.
• General bank performance standards apply to institutions with over $500 million in assets

Comments submitted to the OCC in response to last year’s Advance Notice of Proposed Rulemaking (ANPR) overwhelmingly opposed the creation of the one ratio, which would be a performance measure that would be the major determinant of a bank’s CRA rating. Yet, last week’s NPRM continued to propose a one ratio measure that would be “dominant determinant,” according to the one dissenting FDIC board member, Martin Gruenberg.^[23]

CRA exams would now have two major components for evaluations at each AA and overall at the bank level. The one ratio component is now called the evaluation measure and the other component is now called the retail test.

General Performance Standards

The CRA evaluation measure – one ratio

The one ratio or evaluation measure would continue to be the dollar amount of qualified CRA activities divided by the bank’s quarterly average for retail deposits for each AA and at the bank level.^[24] It would be the determinative factor on a CRA exam as reflected by its name “presumptive” rating.

The numerator would also include a measure capturing bank branches in LMI areas and other underserved areas; this would be the percentage of all branches that are in the specified areas multiplied by .01. This branch measure is supposed to provide some weight to the importance of branches by increasing the ratio by as much as one percentage point. Yet, the current service test of the large bank exam now counts for 25% of the rating. This would be replaced by using branches as part of the numerator of the service test which in most cases will count for considerably less than 25% of the one ratio, and thus will undermine the importance of branches.

Under the new formula for branches, a bank with 30% of their branches in LMI census tracts, which would be a relatively high percentage of branches in LMI census tracts, would only receive a branch score of .3 percentage points in the CRA evaluative measure. Moving to this approach will greatly diminish the importance of bank branches in CRA compliance, which will likely lead to significant branch loss in LMI communities. Recent Federal Reserve research documented that the current service test that explicitly examines branch distribution in census tracts with different incomes prevented the closures of economically viable branches in LMI tracts.^[25] A test that is more opaque regarding consideration of branches is likely to be less successful in preserving branches that remain a vital means of access to populations with limited incomes and mobility.

The OCC and FDIC have established specific ratios for the evaluation measure that would correspond to ratings. A ratio of 11% would correspond to Outstanding, 6% to Satisfactory, 3% to Needs to Improve and less than 3% for Substantial Noncompliance.

The agencies briefly describe their data using the time period of 2011 through 2018 to establish these ratios but do not release their research as an appendix or accompanying paper. The agencies admit the data was incomplete and that they had to use some assumptions.^[26] The OCC and the FDIC have repeatedly stated that their goal in this process is to increase CRA activity, and that moving to this approach would help accomplish that. The OCC and the FDIC must immediately release their research on the current levels of CRA activity in order for the public to evaluate whether the proposed thresholds would actually motivate increases in CRA activity, or would merely legitimize current rates of CRA activity.

To compound the difficulty, actual ratios can fluctuate widely during recessions and expansions. Using a long time period that included the Great Recession as well as the ensuing recovery compromises the robustness of this ratio. Overall, the ratio could be set too low for expansions and too high for recessions, especially if it is based on incomplete data and too long a time period. The agencies state that they will adjust the empirical benchmarks every three years, but it is not clear why they did not do so with their initial benchmarks.

Banks may have to engage in riskier financing, such as stretching underwriting criteria too far, to meet the ratio during recessions and may have too easy a time meeting it in expansions. Finally, the one ratio is not adjusted or computed separately for banks of different asset classes. This would mean that regional or state banks are being compared against the largest banks in the country on a determinant performance measure. This not only creates competitive inequities among banks but may also encourage the smaller banks to engage in riskier financing to achieve the ratios of their larger brethren.

Other major difficulties of the one ratio remain:

• It may encourage an over-reliance on the largest and easiest deals in order for banks to hit the ratio. The FDIC and OCC worsened this problem by opting against a single transaction limit that they had contemplated, such as one deal can be no more than 10% of the numerator, in order to help mitigate this over-reliance on larger deals.
• The one ratio and the qualification of large-scale infrastructure projects discussed above will likely bias banks towards the large deal to the neglect of small dollar business and home mortgage lending needed in many communities.
• Banks may neglect partnerships between nonprofit and public sector entities since they may feel that they can more easily meet CRA obligations with a few large deals in each AA.
• Community development lending and investments are now measured in a manner similar to a one ratio measure. It is more appropriate to measure community development lending and investment in a manner like this than to also include retail and service activities. Moreover, the existing CD lending and investments tests have a ratio measure as one measure and not the determinative one.
• The one ratio discards important qualitative considerations such as the responsiveness of activities to community need. The agencies say that they will retain performance context analysis that assesses a bank’s responsiveness to needs. Their description of performance context is brief and it is not clear how this analysis would adjust ratings determined by the evaluation measure. By generously expanding upon how banks are to describe their capacities and limitations in reinvestment opportunities in local areas, it appears that performance context guidelines would be used mainly to excuse banks from not hitting CRA performance thresholds.^[27]

Retail Lending Distribution Tests & Community Development Financing Minimum

For approximately half of a bank’s AAs, a bank would need to pass a retail lending distribution test and ensure 2% of deposits finance community development.^[28]

The retail test would apply to each major retail lending product lines that comprises 15% or more of the bank’s overall dollar volume of retail originations. In an AA, the retail product line would have to have at least 20 loan originations during the exam time period to be evaluated in that AA.^[29]

The retail test would look at the distribution of borrowers for home and consumer lending. It would not consider home or consumer lending in LMI tracts under the general performance standards because of concerns of causing displacement as discussed above (yet NCRC believes there are better ways to deal with this issue). Small banks opting to be examined under small bank performance standards would continue to be evaluated based on geographic distribution, however. ^[30] The retail test would look at the distribution of lending by borrower for small business and farm lending and also at the geographical distribution of lending for those loan types.

In order to pass the borrower and geographic distribution tests, the bank, for all major retail product lines, would need to meet or exceed a minimum threshold associated with either the demographic or with peer comparator in that assessment area.^[31] These are:

• Demographic comparator: A bank’s percentage of lending for LMI borrowers or small businesses or farms would need to be at least 55% of the percentage of LMI households or small businesses or small farms in an AA. A bank’s percentage of small loans to businesses or farms in LMI census tracts would need to be at least 55% of the percentage of businesses or farms in LMI census tracts in the AA.
• Peer comparator: A bank’s percentage of lending for LMI borrowers or small businesses or farms would need to be at least 65% of the percentage of loans to LMI borrowers or small businesses or small farms originated by all banks in the AA under the general performance standards. A bank’s percentage of small loans to businesses or farms in LMI census tracts would need to be at least 65% of the percentage of small loans to businesses or farms in LMI tracts originated by all banks in the AA under the general performance standards.

While it is an advance to establish benchmarks for the demographic and peer comparators, the NPRM does not describe any rationale for the 55% and 65% benchmarks. If they conducted data analysis and produced averages across metropolitan and rural areas, the reader does not know that. Also, these benchmarks would likely have to be computed for additional areas such as high-cost metropolitan areas, lower-cost areas and possibly smaller metropolitan areas. In particular, the 55% demographic benchmark might be too low in lower-cost areas where it is easier to lend to LMI borrowers or too high in higher-cost areas where it is harder to lend to LMI borrowers. The proposal also does not explain how 2% of deposits was determined to be a significant level of community development activity.^[32]

• The OCC and FDIC requires a bank to pass both the borrower and geographic distribution tests. However, the bank only needs to pass either the demographic or peer comparator to pass the borrower or geographic distribution tests.^[33] Also, the general public has no idea how easy or hard it will be to pass on both tests because the agencies provide no information on estimates regarding pass rates compared to current pass rates. Moreover, a ratings system with at least five ratings provides more weight and importance to the retail test. Five ratings also provides more information to the public about gradations in performance.
• Since the retail test would be only pass or fail, the evaluation measure would have much more weight. At the very least, all components of an exam at an AA level should have ratings and have equal or similar weight.

Summing Up the Ratings: Can Fail in Almost 50% of AAs and Still Pass

Banks can fail in up to one half of AAs on their evaluation measure and retail distribution test, and still receive an overall Satisfactory or even Outstanding rating. This can exacerbate banking and credit deserts since banks can focus on passing in AAs where they consider it easier to conduct business. This is likely to be the larger areas with more population, higher employment levels, income levels and more of a well-established infrastructure to facilitate banking activities.

Carve-Out for Smaller Banks

The FDIC and OCC provide an option for small banks with assets under $500 million to be evaluated under a streamlined small bank exam that only has a lending test or to be evaluated under the proposed tests. As of the most recent call report data (June 2019), 84.6% of all banks or 4,455 banks would be designated as small banks. The asset level will be adjusted annually to take inflation into account. It is justified by the so-called burden and costs of collecting new data required for the new tests. However, the agencies say that available data and their own analysis conclude the small banks would perform better on the new tests than their larger counterparts.^[34] This is possible because while the small banks probably do not engage in as much community development financing as their larger counterparts, a higher percentage of their retail lending might be for LMI borrowers, small businesses and farms.

It is not justified to provide an option for small banks to be excluded from the new exams if the agencies think they will do as well or better than their larger counterparts of the exam. The agencies are failing in their obligation to promote banks meeting community needs. In this case, the agencies are missing an opportunity to preserve the small bank retail lending while increasing their community development financing. To make matters worse, the agencies ask in one of their questions whether the threshold for small banks should be increased to $1 billion. This would surely lead to the loss of hundreds of millions of annual community development financing as documented in a NCRC study.^[35]

Data Collection and Availability

The NPRM would require banks to collect new data on deposits by customer location and data on community development lending and investing. It refers to “certain” data being available to the public.^[36]In the proposed regulations, the agencies would provide to the public data on a county level of retail lending such as consumer lending but not data on community development financing (this data would only be available at the bank level).^[37] The new data required to be collected also must be disseminated to the public so that the community can hold banks accountable for making deposit accounts available to LMI customers and engaging in community development lending and investments.

Public comment

The NPRM refers to public comments only a few times and only in reference to needs in AAs.^[38] The NPRM leaves open the question about whether the OCC and FDIC will continue receiving community group comments on the performance of banks in adhering to their CRA obligations.

In addition, the FDIC and OCC state that banks need to keep data and information in their CRA public file and that they can place this information on the internet. People without internet can obtain paper copies from the bank but the bank can charge people for the information. The public should not have to pay for information regarding whether banks are complying with the laws including CRA.^[39]

Stretch-Out for Outstanding Ratings

The NPRM states that banks receiving Outstanding ratings would be subject to CRA exams once every five years, as opposed to the current schedule of once every two to three years. Five years is too long a time period between exams; it fails to hold banks to the statutory requirement of “continual” and “affirmative” obligation to meet community needs since banks will relax their commitment to CRA, especially in the first year or two at the beginning of a five year time cycle. Also, CRA ratings are an integral part of merger reviews, which will be compromised by stale exams that are less likely to reflect recent past CRA performance.