Analysis of the OCC’s Final CRA Rule

The Office of the Comptroller of the Currency (OCC) released their Community Reinvestment Act (CRA) final rule on May 20. It will lessen the public accountability of banks to their communities by enacting performance measures on CRA exams that will be complex and opaque, while at the same time over-simplifying how to measure a bank’s responsiveness to local needs. Public input into this unworkable evaluation framework will be more difficult and limited. The result will be significantly fewer loans, investments and services to low- and moderate-income (LMI) communities most in need of more credit and capital.

Major components of the final rule address what activities count on CRA exams, how activities count (revising performance measures), where activities count (geographical areas on CRA exams) and accountability mechanisms including publicly available data. Components of the final rule interact with each other in complex ways. For example, the range of activities that count on CRA exams boost scores on performance measures, which consist of summing the dollar amount of activities. In addition, if banks are able to count activities over a larger geographical area, they are also able to boost their scores on the performance measures.

The core of the final rule is the same as the proposed rule:

  • an unjustified expansion of activities that count and which include many activities that do not directly benefit low- and moderate-income (LMI) people and communities,
  • confusing and contradictory rules for multiplying the dollar amount of certain activities that will likely inflate ratings,
  • a flawed CRA evaluation measure that encourages banks to focus on the largest scale financing and to neglect local community needs including those for smaller dollar retail lending in LMI communities,
  • an unworkable reform regarding the geographical areas on CRA exams and too much leniency regarding banks seeking activities across the country without a focus on the areas in greatest need,
  • changes which reduce the public accountability of banks including reducing the amount of data publicly available, lessening the frequency of exams and eliminating consideration of bank affiliates in most cases.

This analysis will examine the major components of the changes including what activities count, how activities count, where activities count and public accountability mechanisms.

What Activities Count

Significant dilution of focus on LMI people and communities

In 1977, Congress passed CRA in response to redlining of LMI communities and communities of color. Since congressional passage, the federal bank agencies have focused CRA on meeting the needs of LMI families and communities. The current regulations direct exams to analyze home and small business lending to LMI borrowers and communities, investments in affordable housing and economic development in LMI communities and services including branches and deposit products for LMI borrowers and communities. An interagency Question and Answer (Q&A) document likewise focuses attention on LMI communities with some narrow and carefully crafted exceptions such as mixed-income housing that encourage desirable policies such as promoting integration.[1]

The focus on LMI consumers and communities responds to congressional intent that was described in detail during the March 1977 hearings on CRA. Senator Proxmire (D-WI), the architect of CRA, and numerous witnesses cited analysis of the then new Home Mortgage Disclosure Act (HMDA) data and surveys of bank services to describe systematic disinvestment in LMI communities.[2] To remedy disinvestment and discrimination, CRA imposed an affirmative and continuing mandate to serve LMI communities. A driving force of disinvestment was market failure or a lack of data and knowledge of borrower creditworthiness and housing stock characteristics and economic conditions in LMI communities. By imposing an affirmative obligation on banks to serve LMI communities, a central purpose of CRA was to help overcome market failures by requiring banks to increase their knowledge of LMI communities so that they could better judge creditworthiness of borrowers and economic conditions in LMI communities and thus make more safe and sound loans and investments in LMI communities.[3]

Some stakeholders attempt to diminish CRA’s focus on LMI communities by referring to statutory language that states banks are to meet needs in all communities, including LMI communities. However, it is clear that the congressional architects had the needs of redlined LMI communities foremost in their minds when they passed CRA. Moreover, over forty years, the federal regulatory agencies have interpreted the congressional mandate as a focus on LMI communities with some carefully crafted exceptions as noted above. The economic ecosystem in LMI communities remain fragile and market failure can easily reappear leading to a new round of disinvestment if the agencies significantly dilute the primacy of serving LMI communities.

Essential infrastructure and community facilities rise in importance and will dilute focus on neighborhoods since they can partially benefit LMI people and communities

As proposed, the final rule in § 25.04(c)(5) & (6) adds essential infrastructure and community facilities to the regulatory definition of community development.[4] In § 25.03, the definition of essential infrastructure is broad, including roads, bridges, tunnels, mass transit, water supply and distribution and sewage treatment. Likewise, the term community facilities is wide-ranging and includes hospitals and public safety facilities.[5] In contrast, the current regulation does not explicitly include infrastructure or facilities, which means that the final OCC rule elevates the importance of these items in contrast to the current regulation. Granted, some of these infrastructure projects already receive favorable consideration on CRA exams, but the projects tend to be focused on directly benefiting LMI neighborhoods since they have to demonstrate a primary benefit to LMI neighborhoods.

In contrast, consistent with the proposed rule, the final rule in § 25.03 removes the emphasis on primary (more than 50% of the dollars benefiting LMI people or communities)[6] and instead allows community development activities to partially benefit (less than a majority of the dollars) LMI people and communities. In practice, this could mean more frequent financing of large scale infrastructure projects such as bridges since bridges are now an explicit regulatory category and since they can partially benefit LMI people.

The OCC did not place a floor on a partial benefit; the agency says it doubts banks will string together projects that have a minimal dollar benefit such as 5% of the total dollars for LMI communities.[7] However, without a floor, the temptation is too great since 5% of an interstate bridge is a large dollar amount (perhaps a bank will claim that 5% of the cars will have LMI passengers). Moreover, this infrastructure will often not be in LMI neighborhoods or directly benefit LMI neighborhoods. At the very least, the OCC should have specified a floor such as 30% of the benefit must be for LMI people and neighborhoods. Moreover, the project should have a demonstrable relationship to a LMI community such as being built or rehabbed in a LMI community or running through it. Too much diversion away from LMI neighborhoods will dilute too much activity from LMI communities and will re-introduce market failures and disinvestment in them.

The final rule corrected a flaw in the proposal by adding back the criterion of economic development. In § 25.04(c)(4), economic development involves the creation of jobs for LMI people, the provision of job training and workforce development and financing infrastructure such as incubators that help small businesses start and grow.[8] In addition, the OCC melded to the definition of economic development the old criterion of revitalize and stabilize, which includes activities and financing that retain businesses and residents to LMI communities.[9] While these changes improve upon the proposal, the elevation of infrastructure and the allowance for partial benefit nevertheless threaten to divert financing away from economic development and revitalization activities that directly benefit LMI communities. The CRA evaluation measure discussed below will make large scale infrastructure projects attractive. In addition, the definition of economic development will include government programs that assist small businesses as those businesses are defined in the public sector programs. Some public sector programs could assist businesses with revenue sizes larger than CRA-targeted ones.

Total impact of rules regarding various population sub-groups difficult to assess because of inconsistencies in focus on LMI communities

The final rule has a series of changes that affect whether activities benefiting various population subgroups or businesses count under CRA. The sum total of these particular changes regarding whether financing for LMI communities would ultimately increase or decrease is hard to decipher because for every change that focuses more on LMI communities, another change dilutes attention away from LMI communities. Consider the following:

  • As the OCC proposed, the final rule assumes rental housing is affordable if lower-income people can afford to pay the rent without verifying that lower-income people will be the actual tenants. NCRC had suggested that the OCC require banks to obtain covenants from borrowers that the tenants would be LMI, a suggestion offered by industry stakeholders as well.[10] The OCC declined to adopt this, citing a vague increase in regulatory burden.[11]
  • As proposed, the final rule codifies that financial education can now benefit people of all income levels instead of just people with LMI. The OCC is not concerned that this will divert resources away from LMI consumers although LMI consumers are more likely to be un- or under-banked according to the FDIC.[12] The OCC rationalizes this change as saying CRA is to benefit all communities, not only LMI.[13] The OCC did not bother to identify any other specific communities that are underserved. This would include people of color that are also more likely to be un- or underbanked according to the FDIC.[14]
  • While the OCC removed the egregious example of large-scale stadium financing from the CRA illustrative list,[15] the agency did not specify that activities in LMI qualified opportunity zones need to conform to the definition of community development that focuses on affordable housing and economic development for small businesses. Projects in Opportunity Zones have featured developments that do not benefit LMI families such as luxury housing. The OCC referenced responsiveness to needs of LMI families in Opportunity Zones, however, without guardrails such as meeting the definition of community development, CRA examiners can adopt elastic definitions of what benefits LMI families (as they have in the past).[16]
  • In the proposed rule, the OCC had listed community development financing for family farms of up to $10 million in revenue. NCRC pointed out that this revenue size limit is too high since the United States Department of Agriculture (USDA) found that 76% of family farms have revenues less than $50,000.[17] In the final rule, the OCC removed the $10 million example. However, the agency did not specify any revenue limit stating that the USDA does not use revenue limits to define family farms. However, the agency could have easily said that it would adopt the USDA categories of small and midsize family farms with gross cash farm income (GCFI) of $1 million as the ceiling. By omitting GCFI size limits, the OCC invites abuse since there are more than 5,800 family farms with GCFI greater than $5 million.[18]

On the positive side, the final rule improved upon a few aspects of the proposed rule in terms of targeting certain underserved populations and businesses:

  • In contrast to the proposed rule, the final rule reduced the revenue size limits of small businesses and farms from $2 million to $1.6 million in § 25.03.[19] However, this is still a move in the wrong direction since the current revenue size limit is $1 million and the CFPB documented that 95% of small businesses have less than $1 million in revenues.[20]
  • The OCC deleted offering CRA consideration for middle-income housing in high-cost areas. The OCC states that upon reflection, “CRA credit for affordable housing should be focused on LMI individuals and families.”[21] NCRC‘s research concluded that this proposal would have diverted resources from affordable housing for LMI families in counties with 43 million people.[22]
  • In § 25.04(c)(3), the OCC added workforce development and job training as part of its definition of community support services in response to comments that the agency was diminishing the importance of job training.[23] The agency added a number of examples of activities including workforce development, financing education and sidewalk retrofits that benefited people with disabilities, in response to comments that the proposal has overlooked this population.[24]

Communities of Color

NCRC and several of our member organizations pressed the OCC about more explicit inclusion of communities of color on CRA exams, citing congressional concern with redlining in communities of color during the 1977 passage of CRA and continuing racial disparities in lending today. The OCC’s final rule stated that its new definition of distressed and underserved areas would benefit communities of color but it did not calculate how many predominantly minority census tracts would be considered distressed and underserved areas, making it hard for the public to assess the OCC’s claims. In § 25.03, the OCC’s new definition of distressed tracts would be middle-income tracts that exhibited high levels of unemployment and poverty. Underserved areas would be measured by a scarcity of branches.[25]

In contrast, NCRC proposed a definition of underserved tracts based on lending rates per capita and NCRC identified that the tracts in the lowest quintile of lending per capita would include a high percentage of predominantly minority tracts.[26] The OCC justified its decision to refrain from similar research about the impacts of its definition by stating, “It is clear from the comments received that commenters were able to look at the definitions included in the proposal and to use public data related to those definitions to analyze and comment on the proposal.”[27] However, the OCC missed the point. When an agency makes a proposal subject to public comment and states it is likely to benefit a particular population, it is incumbent for the agency to engage in data analysis to back up that claim instead of referring to third-party research (NCRC’s) that tested another definition of underserved areas.

In § 25.04(c)(4), the OCC’s distressed and underserved areas would benefit from community development financing but not be included on the retail lending test.[28] In contrast, NCRC had proposed that underserved tracts as we defined them would benefit from examination of retail lending in addition to community development financing.

CRA deserts as discretionary underserved areas subject to abuse

In a putative response to concerns about underserved areas, the OCC introduced a new concept of CRA deserts that was not part of its proposal. The OCC describes CRA deserts as areas with few branches, less community development financing than would be expected based on demographic factors and a lack of community organizations or infrastructure.[29] With these guidelines, the OCC in § 25.06(b) will allow banks to ask for designation of areas as CRA deserts.[30] The agency will maintain an illustrative list of areas.

The loose definition of CRA deserts and the application procedure invite abuse. First, it is unclear what “less community development financing than would be expected based on demographic factors” means? The OCC is not proposing an objective measure such as the dollar amount of community development financing on a per capita basis on a county level. Instead, the agency’s criterion invites subjective judgments by the banks and OCC examiners. Second, a procedure allowing banks to apply for designation of CRA deserts as preferential CRA areas creates a procedure that is closed off to the general public. This will create a secretive and rubber stamp process that results in approving several areas as CRA deserts that are not truly underserved areas. This procedure turns CRA on its head. Under CRA, banks are to respond to community needs. The best way to respond to community needs is to create a process in which the community articulates those needs as a bank is designing its products and marketing approach. Under this proposal, the bank is identifying needs in a process that excludes members of the public. In addition, the designation of underserved areas must derive from data analysis that objectively identifies communities in need as described in NCRC’s analysis discussed above and shared with the OCC in our comment letter.

Banks will have a significant incentive to designate areas as CRA deserts. Under the final rule in § 25.08(b)(3), the OCC will multiply the dollar value of qualified activities conducted in CRA deserts by up to four. After the multiplier is applied, the dollar amount is added to the numerator of the CRA evaluation measure discussed below. Since CRA deserts will be designated under a subjective and secretive process, the multiplier will greatly inflate CRA ratings.[31]

Multipliers Remain in Final Rule and Introduce Subjectivity in the Ratings

Despite the comments of NCRC and other stakeholders, the OCC retained multipliers in the final rule. In addition to expanding what counts in a manner that diverts activity away from LMI communities and populations, the OCC’s final rule will introduce a new level of arbitrariness by favoring some activities over others based on subjective judgments. The agency thus subverts the overall goal of its final rule of making the CRA examination process more objective.

Section § 25.08(b) describes a byzantine set of multipliers that will be applied to certain activities in an illogical manner.[32] For example, several categories of community development loans will be multiplied by two before being added to the numerator of the CRA evaluation measure, except for community development loans that support economic development.[33] The OCC does not discuss why community development loans for economic development would be less valuable for local communities than community development loans supporting affordable housing.

In addition, the OCC would multiply retail loans by two that originate from branches in LMI tracts, but will not add to the numerator of the CRA evaluation measure, the dollar value of home loans in LMI tracts.[34] These two valuations of loans in LMI tracts are confusing and work at cross-purposes. The OCC seeks to recognize the importance of branches in LMI tracts by the multiplier but then deletes the current service test, which more effectively assesses the provision of branches. The OCC does not want to add the dollar amount of home loans in LMI tracts to the numerator of the CRA evaluation measure, because it does not want to encourage high-dollar lending to upper-income people that might facilitate gentrification in LMI tracts. In some cases, however, banks will have a hard time disentangling loans made out of branches in LMI tracts from home loans in LMI tracts. This adds a needless level of complexity.

Another instance of subjectivity that is introduced by the final rule in § 25.08 (b)(4) is multiplying by four community development loans and investments considered responsive to needs, innovative and complex by examiners.[35] This multiplier is the OCC’s attempt to re-incorporate qualitative factors in exams that had been greatly diminished by the proposed rule. However, a multiplier of four is a huge boost to the dollar value of community development activities. Banks will pressure examiners to grant generously these multipliers.

A preferable approach suggested by NCRC would have been to retain the qualitative factors but to make them count for 20% to 30% of the ratings on component tests. In this manner, they would have been important and can make the difference in some cases between ratings on component tests. However, abusing qualitative factors under the NCRC approach is less likely than under the OCC approach. In its final rule, the OCC states that NCRC would favor the retention of a high degree of subjectivity regarding qualitative measures in order to “enable” advocacy rather than increasing money invested in communities.[36] Yet, it is the OCC’s proposed multiplier for qualitative factors that would be uncertain and subjective and could introduce rampant grade inflation rather than increasing lending and investing in LMI communities.

On the positive side, the OCC removed the penalty for originating loans and selling them within 90 days. Loans sold within 90 days to the Government Sponsored Enterprises (GSEs) or other investors would have counted as only one fourth of their dollar value. The OCC responded to stakeholder concern and conducted data analysis which revealed that this lending practice is a significant part of the marketplace. While this is an improvement, the multipliers retained from the proposal and the new highly problematic multipliers will overwhelm this improvement.[37]

CRA illustrative list

The OCC retained the CRA illustrative list and the ad hoc basis for adding activities to the list despite NCRC’s request for a formalized process. The OCC ruled that banks and any interested party can request an interpretation of whether an activity is a CRA-qualified activity. Under the proposal, only banks could make this request. In the final § 25.05 (b), the addition of the public is an improvement but does not compensate for the overall flawed process that builds the list.[38]

The OCC will render its judgment to the party and will add activities deemed to be CRA-qualified activities to the list on an annual basis.[39] NCRC suggested that while the agency can receive suggestions informally, it should hold a notice of proposed rulemaking annually (at least in the early years) regarding proposed additions to the list. In this manner, all parties have an equal opportunity to opine regarding the desirability of proposed additions to the list. The OCC declined this suggestion and stated that it had no legal duty to allow notice and comment on the illustrative list.[40] However, although the list is not exhaustive, it will have a profound impact on which activities banks will pursue. A lack of an even handed approach providing equal opportunity for all parties to comment risks allowing activities on the list that are not responsive to community needs and excluding some that are.[41]

The OCC in § 25.05(d) will engage in a public notice and comment process regarding the CRA illustrative list once every five years.[42] This is too long an interval to provide timely and equitable input into the development of the list.

An example of a problematic activity on the illustrative list (illustrating § 25.04(c)(4)(iii)) that was finalized under a one-sided process (the OCC added this activity in the final rule but had not proposed it) is the following:

A loan to a business to finance the development of workforce housing located in an underserved area that is within close proximity to a warehouse owned by a multinational conglomerate.[43]

This example conjures up the company towns of the early twentieth century that exploited workers through low wages and high rents for flats. It would seem like this CRA eligible activity was suggested by a large corporation to the OCC without input from the public.

Another example (illustrating§ 25.04(c)(8)(ii)) is investment in tax increment bonds to finance infrastructure improvements.[44] These types of bonds have been known to finance activities in gentrifying areas that displace lower-income residents. The OCC added this to the final list without proposing it first. Community stakeholders would have likely raised objections.

How it Counts

The OCC largely left intact its proposed evaluation method although it acknowledged that the majority of commenters on its proposal disagreed with the framework.[45] Its evaluation method, for many banks, will consist of a CRA evaluation measure which is an oversimplified ratio and a pass/fail retail lending test that will count for considerably less weight than the current retail test.

The CRA evaluation measure is the dominant and determinative factor on CRA exams

As proposed, the final rule in § 25.11 creates a ratio that would consist of the dollar amount of CRA activities divided by bank deposits. It would be the major factor in CRA exams.[46] It would be computed at the bank level and for each assessment area (AA) or geographical area that is examined. The CRA evaluation measure will favor larger and easier transactions and likely decrease bank small dollar home and small business lending. The CRA measure would most likely diminish new lending and investment since banks receive credit for the financing as long as it remains on their balance sheets.

The OCC scrapped the thresholds for its CRA evaluation measure, admitting it does not have data to support this core part of its proposal. The OCC had proposed specific ratios of 6% for Satisfactory and 11% for Outstanding. According to the OCC, the thresholds are to be developed through additional rulemaking. This was a major part of their proposal and its abandonment implicitly acknowledges that the agency did not have much evidence to back up its thresholds. NCRC and Urban Institute research had indicated that the proposed thresholds would have inflated CRA ratings.[47] The OCC states:

Although the OCC was not limited in its ability to leverage the existing data, the agency agrees that the existing data was limited, rendering the agencies’…choice of thresholds uncertain. While the proposed thresholds for each of the three components of the objective evaluation framework were reasonable, the agency believes it would be appropriate to gather more information and further calibrate the benchmarks, thresholds, and minimums. In addition, although the OCC issued a Request for Information (RFI) to gather additional information to assist in revising the thresholds and benchmarks in the proposal as appropriate, the data that the OCC gathered in response was too limited to reliably calibrate these measures for all banks subject to the general performance standards. Accordingly, the final rule does not contain benchmarks for the CRA evaluation measure, a specific CD minimum, or thresholds for the retail lending distribution tests.[48]

By abandoning the thresholds and not replacing the proposed tests with another approach, the OCC retained a flawed test framework for which it is impossible to collect data. Banks do not currently collect CRA data in the manner required for the OCC’s measures. Moreover, even if data was available, predicting appropriate thresholds for the future would be mere guesswork since the OCC dramatically expanded the range of eligible activities that banks have not yet pursued.

NCRC had commented that instead of specified ratios such as 55% and 65% on the retail distribution test, the OCC should award ratings on the performance measures based on how well a bank performs compared to its peers. Specifically, NCRC suggested that on a performance measure, a bank is awarded ratings in the following manner based on peer median performance:

Outstanding: greater than 100% because the bank would be better than its peers.

High Satisfactory: 80% to 99% because the bank would be approximately in line with its peers.

Low Satisfactory: 60% to 79% because the bank would be below its peers, but not so far below to be considered not satisfactory.

Needs to Improve: 40% to 60% because the bank would be at approximately half the level of its peers.

Substantial Noncompliance: 39% and lower because the bank would be far below the level of its peers.

For example, if a bank’s peers had issued 25% of their home loans to LMI borrowers and a bank issued 27%, it would score Outstanding on that performance measure. The ratings on the performance measures could be averaged (most likely using weighted averages that reflect the relative importance of the different performance measures) to derive an overall rating. The qualitative measures would also be considered and add 20% to 30% to a final rating. The qualitative measures would consider to what extent activities were responsive to pressing needs such as foreclosure prevention or combating high unemployment.

In contrast to the OCC, NCRC’s approach has clear thresholds capturing how a bank performs relative to its peers and to what extent its activities respond to local needs. Moreover, they do not consist of specific ratios such as 11% for Outstanding on the CRA evaluation measure or 65% to pass the peer measures on the lending test. The OCC’s thresholds would need to be adjusted frequently in a laborious process that is unlikely to keep pace with changes in economic conditions. On the contrary, NCRC’s measures automatically adjust to economic conditions. For example, the aggregate percent of home loans to LMI borrowers would fall in recessions and increase in expansions, reflecting the improved purchasing power of borrowers.

The OCC states that NCRC misinterpreted witness testimony regarding concerns of credit allocation during the congressional hearings leading up to the passage of CRA in 1977. The witnesses were discussing a proposed ratio similar to the OCC’s, which the architect of CRA, Senator William Proxmire, scrapped. This measure would have required banks to predict the future portion of deposits that would be reinvested into communities.[49]Claiming an erroneous NCRC interpretation, the OCC suggests that its ratio does not predict future performance like the discarded ratio of the first draft of the CRA bill. Instead, the OCC claims its ratio would assess past performance. However, the OCC would establish ratio thresholds that banks would adhere to at the beginning of their CRA evaluation cycles. In other words, the agency would be setting thresholds that banks would have to hit in the future at the end of their CRA exam cycles. NCRC and the majority of commenters stated that because future economic conditions and demographic characteristics cannot be forecasted with confidence, a threshold used to assess future performance is problematic.

Limited consideration of bank branches, no consideration of bank deposit accounts 

As the OCC proposed, the final rule eliminates the current large bank service test and examination of basic banking accounts for LMI customers instead of making the service test better. As the OCC proposed, bank branches in LMI communities are devalued considerably under the CRA evaluation measure. For example, the OCC calculated that the median percentage of branches in LMI tracts for banks with at least $2.5 billion in assets (or those banks required to use the CRA evaluation measure) was 28%. This percentage would be multiplied by .02. While this was an increase of .01 from the proposal, it would still undervalue branches.

For a typical bank, the branch percentage in LMI tracts would contribute .56 to the numerator of the CRA evaluation measure. The discarded threshold of 6% was required for a Satisfactory rating on the CRA evaluation measure. For a typical bank, the branch distribution would likely count for 9.3% of the threshold for Satisfactory (.56/6). However, now that thresholds are discarded pending another rulemaking, it is unclear what the importance of branch distribution would be. Nevertheless, it is likely that branch distribution would count less than under the current service test. The OCC estimates that the weight of branches in the current test is 12.5%, which is more than the weight for the typical bank under the OCC’s change. However, the OCC admits its estimation of branch impact on the current service test is uncertain. It is most likely an underestimate since the current service test devotes considerable attention to the percentage of branches in LMI tracts.[50]

Moving to the OCC’s approach will greatly diminish the importance of bank branches in CRA compliance, which will likely lead to significant branch loss in LMI communities and a decrease in lending as documented in NCRC’s comment letter.[51] Also, the OCC chose not to include an evaluation of the provision of bank accounts to LMI customers saying “these aspects of a bank’s business do not lend themselves to quantification.”[52]As the OCC acknowledged, commentators on the proposal stated that an analysis of bank services provides an incentive for banks to offer these services and compete against the usurious and high fee services of non-banks present in LMI communities.[53] NCRC also illustrated in our comment letter how to better quantitatively evaluate the provision of bank accounts, which current OCC CRA exams do contrary to the OCC’s claim that quantifying bank accounts is difficult.[54] The OCC responded that evaluation of bank services will be part of performance context analysis, which is hard to fathom and which the OCC does not explain how it would do this.[55]

Despite NCRC and other stakeholder comments, the OCC retained its proposal to quantify community development services and to count all types of volunteering, not only that related to the provision of financial services.[56] While volunteering at a museum or a construction site is laudable, it does not help prepare LMI consumers to enter mainstream banking nor helps increase access to credit. On a large enough scale, particularly for the largest banks, quantifying an expansive definition of volunteerism will further inflate the CRA evaluation measure.

Retail lending test is supplemental, counts less.

As the OCC proposed, the final rule in § 25.12 creates a lending test that looks at home mortgage, small business and consumer lending. It is now pass or fail instead of generating ratings.[57] It will count for much less towards the rating in each AA and overall. Previously, the lending test for large banks counted for 50% of the overall rating.

The OCC tweaked a few proposed items regarding the lending test. It had included all types of consumer lending but deleted overdraft products and credit card lending in the final rule, responding to community group concerns of encouraging high cost products and bank concerns about difficulties associated with data collection.[58] As described in § 25.03, other types of consumer lending like automobile lending will be part of the lending test.[59] The OCC did not include sufficient safeguards to ensure that the consumer lending is safe, sound and affordable. NCRC had encouraged the OCC to use the innovative and flexible criterion of the current lending test to examine the sustainability of consumer lending.[60] The OCC’s general response regarding qualitative criteria is the misguided multiplier discussed above and an assertion that these considerations will be part of performance context analysis.[61] Currently, performance context analysis is not well developed and it is doubtful, based on the slipshod nature of the final rule, that the OCC will implement examination procedures to incorporate adequately qualitative considerations into performance context analysis.

As stated above, the OCC removed its proposed thresholds that indicated passing performance on the retail test. For example, the OCC had proposed that a bank would pass its peer comparison if it made at least 65% of the peers’ portion of home loans to LMI borrowers. For example, if all lenders in the area made 25% of their home loans to LMI borrowers, a bank would pass if it made 16.25% of its loans to LMI borrowers. Often, a bank is not this far behind its peers on CRA exams, suggesting that the 65% threshold would result in inflated pass rates. NCRC research summarized in our comment letter further illustrated the groundless nature of these thresholds.[62]

Removing arbitrary thresholds is not sufficient. Astonishingly, the OCC issued a final rule without a major part of it settled and said it would gather more data and issue another proposal regarding the test thresholds. This leaves the public in the dark regarding the impact of a major part of the OCC’s rule. This is not an advance; it is a giant leap into a realm of subjectivity and uncertainty. At the very least, the OCC should have refrained from issuing a final rule until settling the threshold issue through a re-proposal.

The OCC had proposed to eliminate home mortgage lending in LMI communities as an exam criterion.[63] The OCC restored this criterion, but the diminished weight of the retail test is likely to lead to less retail lending to LMI people and in LMI communities.

Passing the CRA evaluation measure and the retail test

In a change from the proposal, the final rule requires large banks to pass their CRA exams in 80% of their AAs, up from 50%. While this appears to be an improvement, if the tests remain easy (which is likely but unclear since the OCC will be proposing more refinements to test thresholds in the coming weeks), the requirements for passing in more AAs will not make the overall exam more rigorous.

Where it Counts

Impacts of assessment area (AA) reform proposal is unclear

As the OCC proposed, the final rule in § 25.09(c) established new AAs for banks with significant levels of deposits outside of their branch networks. The thresholds remain the same: a bank must gather more than 50% of its deposits outside of its branch networks and more than 5% of its deposits from an area for that area to be designated as a deposit-based AA.[64]

However, a lack of data on deposit taking outside of branch networks makes it impossible to estimate the impact of this change in terms of how many banks would be impacted and the number of new geographical areas covered as even the OCC acknowledged in the final rule.[65] As NCRC and industry stakeholders urged, the OCC should have considered lending outside of branch networks to determine new AAs because of the availability of data on lending outside of branch networks.[66] In addition, the thresholds are too high as NCRC stated in its letter and would result in too few new AAs being designated, leaving out many other areas in which a bank has a significant presence and in which it needs to be held accountable.

The OCC also issued a final decision on the new deposit-based AAs that is puzzling in terms of facilitating responsiveness to needs by banks. Some banks asserted that although they had gathered a concentration of deposits in an area in which they did not have branches, they would not be familiar with the area or the particular needs of the area. Unless all of these deposits are gathered remotely by an internet platform, it is baffling to understand how a bank would have no familiarity with an area in which it engaged in significant business activity. Common sense would suggest that a bank had engaged in some targeted marketing or worked with partners on the ground to generate the business activity, enabling it to gain some knowledge of the area.[67]

In response to the industry concerns, the OCC will now allow a bank to designate an area larger than the smallest area in which it gathered 5% of deposits as a deposit-based AA. For example, a metropolitan area encompassing a county with 5% of deposits can be an AA or the entire state can be an AA. The OCC reasoned that this would give banks flexibility and would include areas with unmet needs. It does not make sense, however, that a bank would be more familiar with the larger area since it was the smaller area in which it generated a concentration of business. This procedure provides too much discretion for a bank to neglect needs in the smaller area and instead opt to serve easier areas in the state to conduct CRA activities, contrary to the OCC’s assertion that a bank would focus on areas with unmet needs on a state level.

As the OCC proposed, the final rule allows credit to be given for CRA activity outside of bank AAs regardless of performance in the banks’ AAs. This will encourage banks to gravitate to the largest deals anywhere in the country instead of executing smaller deals more responsive to local needs. Instead, NCRC had advocated for supplementing current procedures for considering activities outside of AAs with a procedure for designating underserved counties that also could be served provided the banks had met needs inside their AAs. The OCC did not adopt NCRC’s suggestion, instead retaining its proposal for national consideration coupled with the procedures on CRA deserts, which are arbitrary and one-sided as discussed above.

Public input harder because CRA exams will be less transparent

The core of CRA should be facilitating public comments on bank CRA performance since CRA requires banks to meet community needs. The OCC adopted a blinkered and truncated procedure for considering public comments. Public comments will be further inhibited by the OCC opting to make publicly available data on lending and investing limited despite public comments asking for more transparency.  

As the OCC proposed, the final rule emphasizes only public comments about local area needs and opportunities that examiners will use when conducting performance context analysis. Oddly, the OCC does not explicitly state in § 25.16 (b)(4) whether an examiner will consider comments on the CRA performance of a bank, which, after all, is the bottom line in terms of helping to determine the rating for a bank.[68]

The OCC also opted against improving public data availability. Although community development lending and investing will be important aspects of a bank’s CRA performance, the OCC stated in § 25.27 it will not report data on a census tract or county level as requested by NCRC but only at the total level for a bank.[69] This will frustrate the public’s ability to use data on annual community development financing that would facilitate the public’s ability to comment on bank CRA performance. The OCC says this data will be on CRA exams but CRA exam data can be three years old or older. The public needs more recent and annual data to comment on future CRA exams.

In response to comments about facilitating public input into exams, the OCC says it will explore technological and other approaches for doing so.[70] It is currently difficult for a member of the public to determine how to comment on a CRA exam; the OCC website advises the public to file a comment with the bank or the appropriate “supervisory” OCC office without providing information of which office that would be and any email or contact information for that office. [71]

The exams are likely to be less transparent, making it harder for the public to discern and comment on bank performance. What types of data tables will be on exams; will exams mainly show the ratio associated with the CRA evaluation measure and the few ratios that are on retail test? The OCC does not discuss this important issue in any detail.

Finally, while creating barriers to comment upon and review CRA exams, the OCC maintained its approach of asking banks to calculate the CRA evaluation measures, presumptive ratings, results of retail distribution tests and provide their calculations to the examiners.[72] In essence, the examination process now becomes bank-driven rather than examiner-driven. Examiners are reduced to more of an auditor role of verifying numbers and performance measure calculations. At the same time, the CRA examination process becomes less friendly to the input of the public.

Less Accountability for Banks: Outstanding Rated Banks Examined once Every Five Years and Affiliates Get a Pass

As the OCC proposed, the final rule indicates that banks with an Outstanding rating will be evaluated once every five years instead of once every two or three years. Banks with a five-year exam cycle will relax their efforts in the early years of the cycle. Banks will also have less accountability to maintaining acceptable recent CRA performance when they seek permission to merge with other banks.

The OCC reasons, “The agency notes that the concern about inconsistent performance is mitigated by the fact that the final rule incentivizes banks to consistently meet the needs of their communities by using the average on-balance sheet value of many qualifying loans and investments. The agency also emphasizes that, while it is maintaining the expectation of a general five-year evaluation cycle for banks rated outstanding, the final rule requires banks to report data annually, including banks with outstanding ratings that would be evaluated every five years.”[73]

The OCC’s logic is flawed. A balance sheet approach places inordinate weight on financing offered in the past and thus allows banks to offer less financing in the years closer to the exam. In addition, annual data disclosure is not an effective accountability mechanism because much of the data will be at the bank-level and will not enable stakeholders to monitor bank performance in their localities. Thus, instead of compensating for a stretched-out exam cycle, the factors mentioned by the OCC will exacerbate the harm of less frequent exams.

In a change that is worse than the proposal, the OCC generally excluded affiliates from consideration on CRA exams[74] unless they are involved in CRA-related financing as an “intermediary.” Now, inclusion of affiliates is optional. NCRC argued that inclusion must be mandatory to avoid affiliates from engaging in abusive lending and other behavior.[75] The strange inclusion of affiliates only when they are intermediaries in a CRA-qualified activity makes accountability harder for members of the public who are unlikely to know when affiliates will be involved as intermediaries.

Smaller banks get a pass and are not subject to tests that consider community development or services

The OCC worsened its proposal in terms of providing a greater number of banks with streamlined tests that do not scrutinize community development financing or services. Under the proposal, banks with assets between $321 million and $500 million would have been considered small banks instead of the intermediate small banks (ISB) that they are now classified.[76] As ISB banks, these institutions had a lending test and a community development test that looked at community development lending and investing. Under the proposal, they would have had just a lending test. NCRC commented that losing the community development test for these banks would have meant losing a significant amount of community development financing for smaller cities and rural areas where these banks tend to be headquartered.[77]

Instead of responding to these concerns, the OCC in § 25.03 increased the small bank threshold to $600 million.[78] About 181 banks that are currently ISBs become small banks.[79] The OCC responded to the concern about the loss of a community development test by saying that small banks can opt to have their community development financing considered.[80] This is nonresponsive since consideration of community development financing becomes optional instead of mandatory for these banks. Most of these banks will probably not opt to have community development activities considered. At the very least, it is safe to say that fewer will have community development activities scrutinized thoroughly.

The other part of this change involves re-classifying 69 large banks as intermediate (the OCC replaced the ISB label with intermediate). The OCC raised the endpoint of the intermediate asset threshold from $1.284 billion to $2.5 billion in assets. These banks opting for the intermediate test will no longer have a service test or any test that considers their branching in LMI tracts. As stated above, branches are critical for access to loans and services in LMI communities. No consideration of branches will likely lead to decreases in branches in LMI communities and less lending.

In sum, about 89% of OCC banks or those with less than $2.5 billion in assets (small and intermediate) will not be subject to the new exams but can opt to continue with their current streamlined exams that will either delete consideration of branches or community development financing. In sum, the OCC engaged in a massive rulemaking process that will likely apply the new ill-conceived tests to just 11% of its banks.


The core of OCC’s flawed proposal remains the same in the final rule. As warned in NCRC’s comment and many others, these changes weaken CRA enforcement and lead to less lending, investment and services most needed by LMI communities.

The CRA evaluation measure will still generate a presumptive rating and will be the dominant factor on CRA exams. Coupled with a boost for financing large infrastructure, which is explicitly added as part of the regulatory definition of community development, the evaluation measure will encourage banks to pursue large deals, many of which will have citywide or regional benefits but not targeted directly to LMI neighborhoods. The balance sheet approach that does not adequately distinguish newer financing from financing several years ago will encourage banks to relax on their CRA-related financing if they can land some large deals. At the same time, the retail lending test remains diminished as a pass/fail test replacing the current large bank lending test that counts for 50% of the rating. A bank will now be required to pass in 80% of their assessment areas instead of the proposed 50% but with lenient examinations, this requirement will be easy to meet.

The OCC made a subjective examination regime not backed by data analysis even more uncertain and subjective. The OCC said it lacked data to establish thresholds for its evaluation measure and will thus need to collect more data. It is baffling to contemplate how the agency thinks it will succeed in collecting more data since only six banks supplied data or were responsive to the agency’s request for information (RFI).[81] The OCC will fail again to rigorously test for appropriate thresholds because no forecasting model or data analysis can accurately establish acceptable thresholds since the OCC aggressively expanded the range of CRA-eligible activities that banks heretofore have not regularly claimed as CRA-eligible. Since it is impossible for the OCC to determine adequate thresholds for its flawed CRA evaluation measures, it was irresponsible to finalize a rule that established the evaluation measures.

In addition, the OCC directly contradicts its aim to make the CRA evaluation system more objective and predictable by introducing subjective elements in the rule that were not proposed. The newly introduced and supercharged multipliers for CRA deserts and innovation and flexibility will be subjective, inconsistently applied and will result in grade inflation. The entire CRA evaluation system will be a closely held affair, conducted in a veiled manner between banks and examiners with banks presenting their own ratings to examiners for examiners to verify rather than determine. The final rule does not facilitate public input. The regulatory language regarding public input inappropriately narrows its scope to comments on needs and opportunities instead of banks’ CRA performance. The whole rationale of CRA – determining how banks can respond to communities’ needs – is violated by constraining community voice and input on CRA exams.

In sum, this final rule signals an abrogation of the OCC’s statutory responsibilities to ensure that banks continually and affirmatively respond to credit needs of LMI borrowers and in LMI communities. It must be rescinded or overturned. 



[1] NCRC Comments Regarding Notice Of Proposed Rulemaking (Docket ID OCC–2018-0008 And RIN 3064-AF22), April 7, 2020, pp. 18-20,

[2] Josh Silver, The purpose and design of the Community Reinvestment Act (CRA): An examination of the 1977 hearings and passage of the CRA, NCRC, June 2019,

[3] NCRC comment letter, pp. 18-20.

[4] Office of the Comptroller of the Currency (OCC), Final Rule, Community Reinvestment Act Regulations, Docket ID OCC–2018–0008, Federal Register, Vol. 85, No. 109, Friday, June 5, 2020, p. 34794,

[5] Ibid, p. 34794.

[6] Ibid, p. 34795.

[7] Ibid, pp. 34754.

[8] Ibid, p. 34796.

[9] Ibid, p. 34743 and p. 34796.

[10] NCRC comment letter, p. 28.

[11] OCC final rule, p. 34742.

[12] NCRC comment letter, p. 29 and FDIC National Survey of Unbanked and Underbanked Households, 2017, p. 19,

[13] OCC final rule, p. 34746.

[14] See FDIC National Survey, p. 19.

[15] OCC CRA Illustrative List illustrating § 25.04(c)(11), p. 21, Also, see discussion is preamble of OCC final rule, p. 34747.

[16] OCC Final Rule, pp. p. 34747.

[17] NCRC comment letter p. 33.

[18] America’s Diverse Family Farms, 2018 Edition, United States Department of Agriculture, p. 3,

[19] OCC Final Rule, pp. 34741 and p. 34794 .

[20] NCRC comment letter, p. 33.

[21] OCC Final Rule, p. 34742.

[22] NCRC, For The Trump Administration: Affordable Housing Means Middle-Income Housing For Counties With 43 Million People, January 2020,

[23] OCC final rule, pp. 34796 and 34743.

[24] OCC final rule, p. 34748 of rule & CRA Illustrative list p. 11 (illustrating § 25.04(c)(3)) , p. 20 (illustrating § 25.04(c)(9)), and p. 17 (illustrating § 25.04(c)(6)(ii))

[25] OCC final rule, pp. 34794 and 34795.

[26] Adding Underserved Census Tracts As Criterion On CRA Exams by Bruce Mitchell, PhD. and Josh Silver, January 2020,

[27] OCC final rule, p. 34747.

[28] OCC final rule, p. 34796.

[29] OCC Final Rule, p. 34748.

[30] OCC Final Rule, p. 34797.

[31] OCC Final Rule, p. 34798.

[32] Ibid.

[33] Ibid.

[34] OCC final rule, p. 34740, p. 34796  (§ 25.04(a)(3), and p. 34798 (§ 25.08(b)(ii)(5)).

[35] OCC final rule, p. 34798.

[36] OCC final rule, pp. 34763 .

[37] OCC final rule, p. 34753 and p. 34796.

[38] OCC final rule, p. 34749 and p. 34797.

[39] Ibid.

[40] OCC final rule, pp. 34750.

[41] Ibid.

[42] OCC final rule, p. 34797.

[43] CRA Illustrative list, p. 12.

[44] CRA Illustrative list, p. 19.

[45] OCC Final Rule, p. 34738.

[46] OCC Final Rule, p. 34799.

[47] NCRC Supplemental CRA Comment Using Federal Reserve CRA Data, April 8, 2020, and Urban Institute comment letter to OCC,

[48] OCC Final Rule, p. 34774.

[49] OCC final rule, p. 34768.

[50] OCC final rule, p. 34771.

[51] NCRC Comment Letter, p. 66.

[52] OCC Final Rule, p. 34751.

[53] Ibid.

[54] NCRC comment letter, pp. 43-44.

[55] OCC Final Rule, p. 34751 and p. 34780.

[56] OCC CRA illustrative list (illustrating § 25.04(c)(1)(ii)) , p. 9 and OCC Final Rule, p. 34751.

[57] OCC Final Rule, p. 34800.

[58] OCC Final Rule, p. 34740.

[59] OCC Final Rule, p. 34794.

[60] NCRC Comment Letter, pp. 38-39.

[61] OCC Final Rule, p. 34751.

[62] NCRC Comment Letter, pp. 60-61.

[63] OCC Final Rule, pp. 34740 .

[64] OCC Final Rule, p. 34798.

[65] OCC Final Rule, The OCC says it will rely on supervisory experience rather than data in establishing deposit-based AAs. “The OCC acknowledges that there are limited data on deposits because the current reporting framework attributes deposits to a branch location,rather than the account holder’s address, and uses a definition of deposits different than the proposed definition of deposits. Additionally, not all banks geocode deposits by customer address. Accordingly, the OCC was unable to estimate the number of deposit-based assessment areas that banks would be required to delineate under the proposed framework. However, despite these limitations, given the importance of the CRA and the agency’s policy of ensuring that banks engage in CRA activities in communities where they receive deposits, consistent with the CRA statute, the OCC believes it is appropriate to rely on its experience, knowledge of the banking system, and supervisory judgment to establish initial thresholds. The thresholds in the final rule reflect the agency’s supervisory experience,” p.34759.

[66] NCRC Comment Letter, p. 47 and Comment letter of the Independent Community Bankers of America, April 8 2020,  p. 11,—april-8.pdf?sfvrsn=bd572b17_0

[67] OCC Final Rule discussion of deposit-based AAs, pp. 34757.

[68] OCC Final Rule, p. 34775; p. 34802.

[69] OCC Final Rule, p. 34807 and pp. 34781.

[70] OCC Final Rule, p. 34775.

[71] The CRA section of the OCC webpage,

[72] OCC Final Rule, p. 34781.

[73] OCC Final Rule, p. 34783.

[74] OCC Final Rule, p. 34748.

[75] NCRC Comment Letter, p. 74.

[76] For asset thresholds in 2019, see

[77] NCRC Comment Letter, pp. 68-69.

[78] OCC Final Rule, p. 34793.

[79] NCRC tabulations from FDIC data obtained via As of the May 28, the most current data on bank asset sizes was from December 2019.

[80] OCC Final Rule, p. 34764.

[81] OCC Final Rule, p. 34786.

Print Friendly, PDF & Email