November 8, 2018
Reforming the Community Reinvestment Act Regulatory Framework
Docket ID OCC-2018-0008
To Whom it May Concern:
The National Community Reinvestment Coalition (NCRC) maintains that the Community Reinvestment Act (CRA or Act) has been one of the most valuable laws for increasing access to capital and credit for low- and moderate-income (LMI) communities. NCRC and our 600 community-based member organizations frequently comment on CRA exams and merger applications because the process of public input is key to directing banks and federal agencies to addressing unmet needs in LMI communities and communities of color. Federal Reserve Governor Brainard and former FDIC Chairman Martin Gruenberg have recently remarked on the critical role of community input and stakeholder collaboration encouraged by locally-based CRA exams.
Any CRA reform effort needs to tread carefully and build upon CRA’s success in order to ensure that the progress in reinvestment continues. The genius of CRA’s law is democratic input and accountability on a local level. When Congress enacted CRA in 1977, it sought to remediate redlining and other market failures that impeded lending and investing in LMI communities and communities of color. As envisioned by the CRA, the antidote to redlining was CRA exams scrutinizing lending on a state and local level.
The Office of the Comptroller of the Currency’s (OCC’s) proposed reforms and concepts would impede the effectiveness of public input by diminishing the importance of evaluations of bank records of meeting needs in local areas. The direction in which OCC seems to be going is contrary to the foundational goals and requirements of the CRA.
CRA needs an update, not an overhaul that removes the emphasis on rating and evaluating performance on a local level. CRA has been enormously successful in motivating banks to increase their lending, investing, and services in LMI communities, but the full potential of CRA has not been realized due to growing gaps in CRA’s coverage.
NCRC believes that addressing these gaps include the following:
- The geographical areas on CRA exams called assessment areas need to be updated to include areas with substantial amounts of lending and other business activity in addition to areas with bank branches.
- Banks now have the option of including affiliates on CRA exams, instead of affiliates being automatically covered. This treatment results in significant amounts of lending not being scrutinized by CRA exams.
- A third gap in CRA coverage is the lack of evaluation of lending to communities of color, although the CRA hearings in 1977 leading up to the passage of CRA featured extensive testimony and concern about the lack of lending in communities of color.
- Other needed reforms to address gaps and imperfections in the CRA regulations include reforms to the ratings, improvements in consistency and benchmarks, improved data, and consideration of community benefits agreements on merger applications and CRA exams.
The challenge and opportunity in CRA reform is successfully addressing the gaps in CRA coverage while not disturbing the core mechanisms of public input, transparency, and local accountability. The OCC’s proposals would significantly erode these core mechanisms of CRA. In particular:
- The OCC suggests interpreting the statutory reference to “community” in a far more expansive way that would include within the scope of “CRA-qualifying activities” credit for more aggregated activities without an examination of local credit needs (“non-local”) and for more activities not focused LMI borrowers and communities (“non-LMI”).
- The OCC then suggests determining a bank’s CRA rating based on one ratio composed of an inflated numerator capturing a redefined “community” with far more non-local and non-LMI activities that will, by definition, reduce a bank’s accountability for performance at the neighborhood level and mute differences in performance across local areas.
OCC Redefinition of Community Contravenes Language of CRA Statute
The OCC contemplates a redefinition of “community” that contravenes the plain language of the CRA and the Congressional intent to combat redlining.
At the root of the agency’s “transformational approach to the CRA regulatory framework” are assertions and questions about the agency’s longstanding interpretation of “community” and how the existing regulatory framework implements the law’s references to it. “Community” is currently implemented through an assessment area framework – with banks delineating assessment areas within their branch and deposit-taking ATM footprint. Examiners then evaluate banks’ performance and assign one of four ratings to their record of meeting various credit needs – lending for all banks, and community development/investment and services for others as well.
Regardless of a bank’s structure and scope, local community assessments must occur
The concept of a local community evaluation is connected repeatedly to mandatory sections of the law regarding the scope of agency assessments. The neighborhood focus is required in each of the geographic areas of a bank’s footprint. Aggregation is certainly not expressed, nor is it implied. To the contrary, the language drives toward assessments in localities by use of the word “community” and the inclusion of “neighborhoods” in the context of “entire communities.” “Separate written evaluation[s]” within states and multistate metropolitan areas are “required” pursuant to sections a, b and c (See 12 USC 2906(d) (1) and (2)). Sections 2903 and 2906 mandate community and neighborhood assessments. The Congressional lens was focused on bank activity in small geographical areas.
The structure of the Act precludes the use of aggregating CRA qualifying activities in a manner that would diminish the local analyses
The only exception to the local community assessment contemplated by the act is set forth in
12 USC 2902(4) which states:
- (4) A financial institution whose business predominantly consists of serving the needs of military personnel who are not located within a defined geographic area may define its “entire community” to include its entire deposit customer base without regard to geographic proximity.
Thus, geographic proximity is required in all other instances of the “entire community” as used in the Community Reinvestment Act pursuant to the legal principle of Expresio unius est exclusio alterus (to express or include one thing implies the exclusion of another). In the case of the CRA, the law taken as a whole reveals Congressional intent to focus on the local community as the subject of the mandatory assessment of meeting the needs of LMI neighborhoods.
The local focus also hews most closely to Congressional intent – to ferret out and remediate bank redlining of specific neighborhoods
U.S. Senator William Proxmire, the principal author of the Act, made clear a key purpose of the law was “to eliminate the practice of redlining by lending institutions.” He described redlining as the actual or figurative drawing of a red line on a map around the areas of a city, sometimes in the inner city, sometimes in the older neighborhoods, sometimes ethnic and sometimes black, but often encompassing a great area of their neighborhood.
- The data provided by that act (HMDA) remove any doubt that redlining indeed exists, that many credit-worthy areas are denied loans. This denial of credit, while it is certainly not the sole cause of our urban problems, undoubtedly aggravates urban decline.
Any reform that results in exams that do not carefully scrutinize access to credit in underserved neighborhoods across local communities circumvents the clear intention of Congress to evaluate and rectify the redlining of communities. The CRA requires each of the regulators “to use its authority when examining financial institutions, to encourage such institutions to help meet the credit needs of the local communities in which they are chartered…” The statute goes on to define exactly how regulators are to use their authority consistent with CRA, including to: “assess the institution’s record,” “prepare a written evaluation,” state the “agency’s conclusions,” “discuss the facts and data supporting such conclusions,” and assign a rating with “a statement describing the basis for the rating.”
The statute and the law’s drafters contemplated that regulators would evaluate and discuss why they concluded that the bank was doing more than meeting their charter obligations in bank law “to demonstrate their deposit facilities serve the convenience and needs of the communities in which they are chartered to do business.” CRA was enacted to require regulators to come to a conclusion about and rate whether the bank is “meeting the credit needs of its entire community, including low- and moderate-income neighborhoods.” In other words, the regulators must determine that the bank is not redlining neighborhoods in the communities they are serving, plain and simple, and that they are making affirmative and ongoing efforts to help meet the credit needs of LMI neighborhoods. And, deeming bank activities as CRA-qualifying “in the aggregate, at the bank level” would not allow them to reach a conclusion on whether that is occurring based on any facts and data.
Based on this analysis, it would be improper to define “community” to include aggregated, non-local activities or include them in a metric-driven, one-ratio or other approach to banks’ core CRA rating.
Agency and industry proposals to diminish CRA examination and focus on LMI communities and people would dilute CRA’s effect and undermine its purpose
A major theme in the advocacy to broaden CRA in a manner that is contradictory to the goals of the law is to include more non-LMI activities or general community building activities that benefit middle- and even upper-income communities, but may have some marginal benefit for LMI communities. This theme is inconsistent with the purpose of the Act or the economic rationales behind it.
Clearly, redlining motivated CRA’s passage, but also the concern about high capital export from local communities where banks were taking deposits. The law is designed to give banks an incentive to look for lending and investing opportunities in their local markets and to overcome the “market failures” that limited lending and investing in LMI areas. Banks were bypassing profitable lending opportunities locally in favor of far off investments, to the detriment of local housing, small business, and small farm credit needs. Senator Proxmire cited that newspapers in New York City had an elaborate series of stories on the amount of disinvestment in the City, pointing out that only 11 percent of the money deposited in Brooklyn remained in the borough. In the District of Columbia, researchers found that about 90 percent of the deposits are loaned and invested outside of the community. Chicago, Los Angeles, and St. Louis also experienced this high degree of deposit flight.
A 2000 baseline report on CRA by the U.S. Treasury Department captured the Congressional thinking around capital export and local revitalization efforts:
- Congress intended the CRA to increase credit access and revitalize inner-city communities that were deteriorating with the movement of investment and development elsewhere. In addition, Congress recognized that the success of federal community development, housing assistance and mortgage insurance programs enacted at the same time as the CRA…would depend in large part upon the availability of private capital, particularly as made available through local financial institutions.
The CRA was designed to strengthen these public economic development efforts by facilitating “efforts between private investments and federal grants and insurance in order to increase the viability of our urban communities.”
Importantly, the law requires the regulators to examine whether banks are overcoming the market failures and informational externalities associated with the lack of investment in LMI communities.
CRA’s regulatory framework must be targeted to and focused on LMI borrowers and communities to correct for market failures and externalities
The U.S. Treasury explains well the positive, negative, and related informational externalities that exist when banks, which decide where to lend and invest, do not bear the full costs nor reap the full benefits of their actions. As Treasury states:
- There are significant positive externalities, for example, associated with lending in areas where there are frequent numbers of transactions, such as middle-class or relatively affluent neighborhoods. These transactions produce a steady stream of information about market values for other lenders (and appraisers) to consider when making their credit decisions. A larger number of transactions increases confidence in the appraised value of individual properties, and increases the liquidity of other investments in the neighborhood, thus improving the values of properties that serve as collateral for mortgages. This process lowers lenders’ transactions costs, thereby lowering the cost of credit for all borrowers in the area.
- The reverse is true for neighborhoods where there are relatively few transactions. In particular, to the extent that lenders do conduct appraisals in LMI neighborhoods, these appraisals can be more costly and less accurate because “comparable” transactions and appraisers familiar with such neighborhoods are not available. Loans in these areas are therefore riskier, and lenders will compensate by charging higher rates of interest or requiring larger down payments. The stiffer terms on such loans can cause some borrowers either to borrow less or to drop out of the market altogether. For LMI neighborhoods, the end result can be a downward spiral – less lending, fewer appraisals, even less lending, and so forth – producing an effect that resembles redlining.
Being the first institution to enter a new or previously underserved market or investing in an innovative but high-impact LMI project, when other lenders are unwilling to lend, present related informational externalities and can result in delays and perceived risk. Lender expectations of this sort can cause a potentially viable market to suffocate from lack of credit. In the process, borrowers who may otherwise be credit-worthy will be denied credit because of the absence of entry by competitive lenders.
- Therefore, the CRA can be understood as a vehicle for facilitating coordination and for assuring banks that they will not be the lone participants in thinly-traded markets…the Act can produce positive information externalities and allow all lenders – both those covered by the CRA and those not covered by the CRA – to better assess and price for risk.
Congress recognized these market failures and externalities. Accordingly, CRA requires regulators to examine the data and assess whether banks are, in fact, overcoming these failures in a “continuing and affirmative” way. Since the law was enacted, banks have made good strides to doing just that. They have taken numerous steps, including entering loan pooling arrangements, undertaking lending consortiums, partnering with local groups, community development corporations, community development financial institutions (CDFIs) and others to break down the barriers that impede the efficient flow of capital into LMI communities.
The CRA was designed and the existing regulatory framework to-date has executed the law in a way to overcome redlining and other market failures. Regulators must not deviate from these important policy goals, but could strengthen their focus on and targeting to LMI communities in banks physical and digital footprint.
Research Supports Updating Current CRA Regulations, Not Transforming Them
Evidence supports the effectiveness of the current regulatory approach (Questions 1-6 of the ANPR). A large body of research has demonstrated overwhelmingly that CRA has leveraged significant increases in lending and reinvestment in LMI communities. Recently, Lei Ding and colleagues at the Philadelphia Federal Reserve Bank concluded that when census tracts lose CRA eligibility, the number of home purchase loans decreases between 10 to 20 percent. Similarly, Raphael Bostic, President of the Federal Reserve Bank of Atlanta, and colleagues show that CRA increases small business lending in CRA-eligible census tracts compared to tracts with income levels just over the CRA-income thresholds. In addition, a number of Federal Reserve studies show that CRA eligible home mortgage lending was significantly safer and sounder than high cost lending with high default rates issued by independent mortgage companies in the years preceding the financial crisis.
The existing assessment area framework increases safe and sound lending and investing in LMI areas
Studies have concluded that the CRA assessment areas described in the regulation increase lending. In commemoration of the 25th anniversary of CRA in 2002, the Joint Center for Housing Studies at Harvard University conducted a major research study on CRA, finding that banks make a higher percentage of their home purchase loans to LMI borrowers and census tracts in their assessment areas than outside of their assessment areas from 1993 through 2000. In addition, rejection rates for LMI applications were 8 percentage points lower in assessment areas than outside assessment areas.According to Harvard, the positive impact of assessment areas on lending was the equivalent to a 1.3 percentage point reduction in unemployment. In other words, CRA scrutiny of lending in assessment areas was equivalent to a significant nationwide reduction in unemployment in terms of increasing lending to LMI people and communities.
Likewise, Federal Reserve economist Daniel Ringo examined the impacts of changes in the boundaries of metropolitan areas on assessment area lending. In 2003, the Office of Management of Budget changed metropolitan area boundaries for a number of metropolitan areas in the United States, causing a significant number of census tracts to be either deleted or added to bank assessment areas. In the newly added LMI census tracts to bank assessment areas, Ringo found that lending by even a single bank increased overall lending in one LMI census tract by two to four percent from 2003 to 2004. Also, bank lending increased further over time as banks intensified their efforts in these new assessment area tracts. Moreover, Ringo found that the impact was greatest for low-income borrowers, those with less than 50 percent of area median income, than for moderate-income borrowers, those with between 50 to 80 percent of area median income. He hypothesizes that banks face less competition in extending loans to low-income borrowers than to moderate-income borrowers, so efforts to increase lending to these customers–prompted by the addition of census tracts to banks’ assessment areas–were more effective at filling unmet demand.
Research suggests losses of local lending and investing without transparency and scrutiny by regulators and the public
In one of the few studies to measure community development (CD) lending and investing, NCRC finds that intermediate small banks (ISB) issue an annual level of CD lending and investing of $3 billion, which about equals that of the Department of Housing and Urban Development’s (HUD’s) Community Development Block Grant program. Moreover, the levels of CD finance issued by ISB’s with Outstanding, Satisfactory, and the failed ratings are statistically significantly different, suggesting that ratings are effectively differentiating among differences in performance and thus likely motivating improvements in performance.
The positive outcomes documented by this study are driven by the public accountability mechanisms of CRA. CRA is a powerful motivator for banks to increase lending because a bank’s performance in LMI communities in its assessment areas is evaluated and rated in a transparent manner subject to public comment.
Extrapolating Ding’s estimates of losses in home mortgage lending of 10 to 20 percent due to a removal of CRA scrutiny, NCRC estimates that the loss nationally in home mortgage lending would be of a magnitude of $44 to $89 billion over five years. Bostic’s research suggests that a large loss would also occur in small business lending, which NCRC’s estimates to be between $8 and $16 billion. The total loss in home and small business lending over a five year time period would range from $52 to $105 billion. Moreover, the losses would be dramatic on a state and Congressional district level as documented in NCRC’s recent study. In sum, eliminating the local evaluation of CRA lending in assessment areas would significantly reduce CRA lending over several years. The impact would be lasting and severe.
Updating Geographic Scope of CRA Assessment Areas: Bank Branches and Digital/Business Footprint
Because of the effectiveness of assessment areas, the current procedure of delineating assessment areas for geographical areas containing bank branches must be retained. In order to rigorously evaluate non-traditional banks that make loans via non-branch means, assessment areas also must be established for areas in which these banks make a considerable number of loans and/or engage in a significant amount of business activity.
Most bank lending is still conducted in assessment areas. Research by Federal Reserve economist Neil Bhutta finds that assessment areas captures about 70 percent of home purchase lending for large banks. Likewise, in examining the 100 largest banks, NCRC found that assessment areas captured a great majority of their lending (91 percent of their home and small business lending).NCRC’s report relied upon the percentage of loans in assessment areas reported by CRA exams, which do not consider lending by affiliates when calculating the portion of loans in an assessment area. Therefore while NCRC’s percentage might be an over-estimate, it is consistent with Bhutta’s finding that for the largest banks, current assessment area procedures capture the great majority of their lending.
Since the current procedures capture the majority of traditional bank lending, reforms should adopt an additive approach instead of implementing wholesale changes. In particular, reforms should focus on non-traditional banks that are making large volumes of loans using non-branch means including brokers and the internet. For several years, NCRC has urged the agencies to update assessment area procedures to expand the number of assessment areas to account for lending beyond branches. This is a straightforward approach that retains assessment areas where branches are located and adds assessment areas to encompass geographical areas where banks lack branches but are engaged in significant lending or other business activity (see Question 13 of the ANPR which asks how to expand assessment areas). This approach does not nationalize the scope of the evaluation of banks’ CRA performance (which would result in dilution of the data) because it retains the required local focus in the new assessment areas.
Both the OTS and the OCC have examined non-branch lending by supplementing the existing assessment area framework
A number of CRA exams have adopted the NCRC approach. The former Office of Thrift Supervision (OTS) supervised several lenders without traditional branch networks. The OTS relied upon the Interagency Question and Answer (Q&A) document allowing examination of retail lending outside of assessment areas provided the retail lending inside the assessment areas has adequately responded to needs. However, good lending performance to LMI borrowers outside of the assessment areas would not compensate for poor lending performance in the assessment areas according to the Q&A.
The OTS 2009 CRA exam of Citicorp, a non-traditional thrift located in Wilmington, DE that made loans through 77,000 agents located throughout the country, included analyses of 10 metropolitan areas and three non-metropolitan areas with the largest percentage of lending outside of the Wilmington assessment area. Likewise the OTS examined Capital One’s lending in 20 areas beyond its one assessment area. These 20 areas comprised 25 percent of the thrift’s lending.
A more recent exam of the Bank of the Internet further develops and refines procedures for considering loans outside of assessment areas. Since assessment area lending in San Diego accounted for 1 percent of total lending activity, an examiner with the OCC evaluated retail lending in six states outside the San Diego assessment area. Bank of the Internet’s activity in these six states accounted for 30 percent of total deposits and 56 percent of home mortgage and small business lending. The retail lending in the states outside of the San Diego assessment area was factored into the rating for the lending test.
NCRC illustrates how assessment areas would work for the Lending Club, an online lender (fintech) with no branches that has made loan data by geographical area available on its webpage. In Congressional testimony, NCRC calculated that more than two thirds of Lending Club’s lending during 2012 and 2013 was in 15 states, making it feasible to designate those states as assessment areas. NCRC also used Texas as a case study of designating local assessment areas. NCRC found five metropolitan areas with more than 1,000 loans each and one area, North Texas that could possibly be considered a rural area. The five metropolitan areas ranged in size and location across the state and included Houston, Austin, Ft. Worth, Dallas, and San Antonio. El Paso is the seventh largest area by loan volume with more than 500 loans. Using Lending Club as an example, NCRC demonstrated that it is feasible to select assessment areas for states, metropolitan areas, and rural counties where a substantial amount of lending occurs.
Capturing the great majority of retail lending via assessment area coverage is imperative for rigorous grading.
To address CRA grade inflation, assessment area reform should ensure the vast majority of bank lending and investing is examined
NCRC has found that when the great majority of lending is not captured by assessment areas, CRA ratings are inflated. In our study of the one hundred largest banks, NCRC revealed that inflation was the largest concern when assessment areas covered less than 50 percent of retail lending but also occurred when assessment areas covered less than 75 percent of lending. When exams allow banks to focus more intently on a lower percentage of their loans, it is easier to score well on the lending test. While this may benefit assessment areas containing the minority of a bank’s loans, it results in less lending overall in LMI communities as demonstrated by the Federal Reserve-sponsored studies cited above. When ratings are inflated due to inadequacies in assessment area procedures, local needs will remain unmet. An approach that elevates the one ratio and diminishes the weight assessment areas contribute to the overall CRA rating will result in local needs being ignored and unmet.
Inexplicably, the ANPR does not seem to build upon the precedent of Bank of the Internet and other exams that incorporate supplemental areas covering substantial amounts of lending executed via non-branch means. Instead, it discusses how lending, investing, or other activities outside of areas with bank branches can be considered in “the aggregate.” In other words, if a bank made a $5 million investment and issued $10 million of loans to LMI borrowers outside of its branch network, the $15 million could be added to a one ratio. In contrast, if the $15 million of loans and investments occurred in one state, that state could be designated as an assessment area, and the lending and investing could be judged in the traditional manner to see how responsive it was to the needs in that state. An “aggregate” consideration of activity outside of branch networks would be inadequate and would fail to determine how responsive the activity was to local needs, as called for by the CRA.
A one ratio focused exam would also negate the recent attention the OCC has devoted to concerns that rural areas and smaller cities receive primarily “limited scope” or cursory consideration on CRA exams. The OCC stated in a bulletin issued over the summer that assessment areas could become “full scope” areas with more weight if “the bank’s activity represents a significant portion of total industry activity in an assessment area relative to deposit or loan market shares.” Also, public comments about a bank’s performance in a specific assessment area can be a factor elevating the area to full scope status. These changes plus the OCC adoption of a practice of rotation of assessment areas as full scope areas provide more opportunities for often overlooked geographical areas to be full scope. However, these advances will be under-cut if the OCC subsequently diminishes the importance of assessment areas on CRA exams.
The wrong approach on Varo Bank
On August 31, the OCC granted preliminary conditional approval of Varo’s application to become a nationally chartered bank. Varo is all mobile financial technology company (fintech) with no branches and offers deposit accounts, credit cards, home equity loans, and consumer loans nationally. Varo is seeking a charter as a national bank, which will make it subject to CRA. That is distinct from fintech’s approved under the OCC’s new fintech charter, which have CRA-like financial inclusion obligations. In its application, Varo designated one assessment area, the Salt Lake City-Provo-Orem, Utah combined statistical area (CSA), where its headquarters office is located. The approval order states that the Salt Lake City CSA is the most “expansive” assessment area that can be delineated under existing CRA regulations. We disagree. Varo should be required to delineate locally-based assessment areas and be examined in areas where their lending and other business activities constitute the great majority of their lending and other activities.
A Fairer Way to Consider Activities Outside of Assessment Areas
Even if the OCC were to adopt NCRC’s suggestions for expanding the number of assessment areas, banks will still seek to satisfy needs, particularly for community development lending and investment, in areas beyond their assessment areas. If a bank still sought to engage in community development financing beyond its assessment areas, objective guidelines could be developed to assess whether a bank was adequately meeting needs in its assessment area in at least a satisfactory manner (See ANPR question 14).
The guidelines could include consulting a number of quantitative measures. First, the previous CRA exams could be consulted to benchmark bank current performance in its assessment areas compared to its past levels of lending and investing. If a bank had performed in at least a Satisfactory manner on its previous exam and its current level of community development lending and investment per its deposits was the same or higher as its previous level, then it is likely on the right path towards meeting needs in its assessment area. Second, per our data recommendations below, annual community development data would be available for a bank’s peers in the bank’s assessment area. The agencies would therefore check to make sure the bank was at or above peer levels of community development financing.
As a final check, NCRC urges examiners to listen to community groups and local residents comment on whether a bank is meeting needs in assessment areas for the types of lending and investing it seeks to offer outside of assessment areas. If a high number of comments indicate a bank is not meeting needs, on an absolute basis and/or in reference to peers, then the bank should finish meeting needs in its assessment area first. The agencies could issue formal determinations in between exam cycles about whether a bank can venture beyond its assessment areas, which could be publicly available so that community stakeholders can approach banks that are allowed to satisfy needs beyond assessment areas. The Treasury Department recommended this type of determination process, the OCC talked about it with stakeholders, but for some unexplained reason did not bring it up in the ANPR.
The agencies could also expand the number of distressed areas that banks could be permitted to serve after they have met needs in assessment areas. The agencies could develop a list of underserved and distressed counties on a national basis that banks could receive consideration for serving (see Question 16 of the ANPR). The model for such a list is the current list for rural counties that are distressed and underserved. In developing a national list of counties, factors to be considered could include economic factors such as poverty and unemployment rates. In addition, measures of bank activity could be included such as home or small business loans per capita. Counties with loans per capita significantly lower than national medians would qualify as undeserved. In addition, NCRC urges the agencies to develop a database of community development lending and investing on a county level based on data collected during CRA exams. Such a database could also identify underserved counties based on levels of community development financing per capita.
NCRC believes that our approach to allowing banks to engage in activities beyond assessment areas is more objective than the current ad hoc and opaque system. It makes it more possible for community groups and other local stakeholders serving distressed and underbanked areas to reach out to banks and be confident that community development financing can be secured from them. It could alleviate the stress that arises over the current ambiguity and thus augment the current system rather than needlessly overhauling the examination regime and making it less responsive to local needs.
The U.S. Treasury on the Right Track: Automatic Inclusion of Affiliates on CRA Exams
In its memo to the federal banking agencies in the spring of 2018, the U.S. Treasury Department stopped short of calling for mandatory inclusion of affiliates but urged the agencies to evaluate their “approach to affiliates in order to ensure that performance evaluations accurately reflect the CRA-eligible activity of the overall bank.” CRA exams allow banks to either include or exclude their mortgage company affiliates on CRA exams. And, it is hard to think of a process that is not more prone to abuse. The natural tendency is for affiliates to be included on evaluations if they are responsibly lending to LMI borrowers and neighborhoods and to be excluded from exams if they are not.
An example of optional inclusion enabling abusive practices is Suntrust Mortgage Company, which Suntrust excluded from its CRA exam of 2013. The U.S. Justice Department, HUD, and the Consumer Financial Protection Bureau (CFPB) reached a $1 billion settlement with the mortgage company over widespread abuses associated with underwriting FHA mortgages and mortgage servicing that occurred in the time period covered by the CRA exam. Yet, because of the optional treatment of affiliates, Sunstrust’s CRA exam did not consider the mortgage company’s lending practices and whether these practices should result in a ratings downgrade. The optional treatment is inconsistent with the interconnectedness of affiliates and their parents. Suntrust’s CRA exam states, “SunTrust Mortgage Company is the primary originator of home purchase and refinance loans for the organization.” In most cases, the affiliates’ activities are inextricably connected with the banks. The optional treatment must end.
Alternatively, regulators could adjust the CRA rating of the bank if the affiliate is engaged in activity that is at wide variance from the bank such as abusive lending or not lending to modest income populations while the bank is offering a higher percentage of their loans to these populations. Disparate patterns like this could very well reflect gaming exams.
Conduct CRA exams at the holding company level
Currently, CRA exams are conducted separately for banks that are affiliates and are under control of the same holding company. These exams can be confusing for the general public and often do not completely reflect the extent to which the entire holding company responds to community needs. An exam at a holding company level may encourage the holding company to strategically plan how their individual banks could better complement each other’s efforts and the company’s total commitment to community reinvestment. The same reasoning applies in the case of non-bank affiliates that are often retail lender specialists while the bank affiliates can complement their retail lending with strategic community development loans or investments in targeted neighborhoods.
Within Statutory and Constitutional Parameters, CRA Exams Must Address Dearth of Credit in Minority Communities
Persistent and glaring racial disparities in lending manifest themselves year after year and decade after decade. The current trend is that minorities receive disproportionately few loans, while in the years leading up to the financial crisis, minorities received disproportionately high levels of subprime loans.
Last year, the journal Reveal released a well-publicized report documenting ongoing racial disparities in lending across several metropolitan areas. The HMDA data has also shown a multi-year stagnation in lending to minorities. For example, lenders have issued between 5 to 6 percent of their home purchase loans to African-Americans in each of the last ten years although African-Americans are about 13 percent of the population. Likewise, a couple of years ago, NCRC released Home Mortgage and Small Business Lending in Baltimore and Surrounding Areas. After controlling for neighborhood characteristics such as the percent of owner-occupied units, median family income, median home value, and unemployment levels, NCRC found that the percentage of African-Americans in neighborhoods was significantly and negatively correlated with the number of loans from 2011 through 2013.
An earlier NCRC study, Foreclosure in the Nation’s Capital: How Unfair and Reckless Lending Undermines Homeownership, shows that after controlling for various neighborhood characteristics, loan characteristics, and borrower characteristics including creditworthiness and payment-to-income ratios, minorities were still more likely to receive subprime loans and experience foreclosures than white borrowers. During the height of subprime lending, NCRC released annual reports called Income is No Shield against Racial Disparities in Lending which showed that even when controlling for income, people of color disproportionately received high cost loans.
We believe the regulators could do more within a CRA regulatory framework to facilitate the flow of more safe and sound credit to minority borrowers and communities.
In 1977, Congress sought to address what still plagues the nation today – gaping racial wealth inequality stemming from redlining and other market failures
The 1977 hearings featured extensive testimony documenting lack of lending in communities of color. For example, representatives of neighborhood-based Advisory Neighborhood Commissions (ANCs) in Washington DC testified that people of color had difficulty obtaining conventional bank loans and that the down payment requirements were often 25 percent of the loan amount. As a result, homebuyers had to rely on mortgage companies offering higher fee Federal Housing Administration (FHA) loans with higher default rates. In addition, the ANCs protested the application of American Security and Trust Company to establish a branch in Washington DC, asserting that the institution would make loans west of Rock Creek Park in predominantly white neighborhoods but would redline communities of color east of the park. They reported that bank branch personnel in communities of color would often state that they lack the authority to make loans. Lastly, they reported that a survey of small business owners found that just 12 percent had loans though 49 percent applied.
Additional witnesses reported the difficulties of communities of color in accessing loans. Gale Cincotta of National People’s Action reported that 93 percent of the homebuyers in the Logan Square neighborhood with a high percentage of Hispanics had high down payment FHA loans while 90 percent of the buyers in an adjacent neighborhood had low down payment conventional loans.Similarly, the Massachusetts Commissioner of Banks confirmed that lower income neighborhoods had to rely on more expensive mortgage companies. Finally, the Connecticut Commissioner of Banks reported that in eight cities banks provided evening and weekend hours in their branches in suburban neighborhoods but not urban neighborhoods. When the banking department placed a moratorium on bank applications, more than 20 institutions changed their practices.
While the 1977 CRA hearings effectively documented conditions of redlining and unequal access to lending in the 1970s, the phenomenon of redlining extends decades prior to the 1930s. The Federal government under the Roosevelt Administration’s Home Owner’s Loan Corporation (HOLC) drew maps of city neighborhoods and differentiated them according to risk as perceived by industry professionals working for the Federal government. The highest risk and “hazardous” neighborhoods were overwhelmingly minority and lower income. With federal government approval, these neighborhoods were then systematically redlined by lending institutions for decades. In a recent report, NCRC found that the neighborhoods classified as “hazardous” have remained predominantly minority and lower income.
Stronger fair lending reviews and specific standards for community development consideration should better account for race
Before the reforms to CRA regulations in the mid-1990s, lending to people of color was often considered in Factors D and F on CRA exams. For example, the 1996 Federal Reserve Bank of Richmond exam of Signet Bank analyzed the percentage of applications to minorities for the bank, its affiliated mortgage company, and all lenders in assessment areas in Maryland, Virginia, and Washington DC. Denial rates to minorities to whites were also analyzed. The fair lending review conducted a statistical analysis of rejected white and minority loan applications in order to assess whether denials were due to illegal discrimination. Similarly a 1995 OTS CRA exam of CenFed Bank examined applications and originations of the bank compared to demographics, concluding that while the agency could not find evidence of discrimination, the bank’s percent of applications and originations to Hispanics was considerably below the percent of the population that was Hispanic.
The agencies can and must improve examination of lending to people and communities of color on CRA exams. NCRC’s preference would be for lending, investments, and services to people of color and communities of color analyzed in a manner similar to lending, investments, and services to LMI borrowers and communities on CRA exams. Consideration of lending to minorities on CRA exams would facilitate bank compliance with fair lending laws by also requiring an affirmative obligation to serve minorities. Lenders would be less likely to engage in redlining and other racially discriminatory practices, lessening compliance costs for lenders and creating a more robust and competitive lending market in communities of color.
The agencies should include an “other populations or area category” described in Question 16 of the ANPR that would consist of underserved census tracts in the banks’ assessment areas. These tracts would be identified via data analysis as those with lowest loans per capita or and/or deposits per capita. Many of these tracts would likely be predominantly minority. CRA exams should also include a more robust fair lending section along the lines of exams conducted before the changes to the CRA regulations in the 1990s.
HMDA and other data have consistently shown racial disparities in access to lending and banking. During the foreclosure crisis, predominantly African-American middle income communities such as those in Prince George’s County, MD were afflicted with predatory, non-bank lending because they experienced a shortage of traditional bank lending.
CRA Grade Inflation has Undermined the Power of the Law
Since 1995, 98 percent of banks, on average, have passed their exams. If anything, CRA grading needs to be more vigorous with a success rate lower than 98 percent. If the pass rate was not this high, CRA would be even more effective in motivating increases in loans, investments, and services to LMI communities. In response to ANPR Questions 3, the agencies can be considerably more creative to make the ratings more objective, fair, and transparent.
The ratings distribution makes it difficult to discern significant differences in CRA performance when almost all the banks in the country are essentially receiving an A or B on their CRA exams. Distinctions in performance are more apparent on the component tests and on a state level, since banks can also receive the ratings of High and Low Satisfactory in addition to Outstanding, Needs-to-Improve, and Substantial Noncompliance.
The agencies have discretion to alter their existing point system to provide further distinctions among bank performance. Currently, the agencies use a scale of one to 24 as a point system. It is quite difficult to understand how a point system of one to 24 was developed to capture bank performance on various exam components. Instead, a point scale of 100 would make much more intuitive sense and would correspond to exam weights. For example, since the large bank exam weighs the lending test at 50 percent, the lending test score can have a possible range of one to 50. Likewise, since the investment test and service test are each weighed at 25 percent, they can have a range of one to 25 points each.
A 100 point scale would better differentiate between the four assigned ratings
The overall ratings would be more differentiated if they were accompanied by a publicly released score. For example, an Outstanding rating could be achieved if a bank had a score of 90 to 100, while a Satisfactory rating could be achieved if a bank had a score of 70 to 90. An Outstanding rating accompanied by a score of 90 would not be as remarkable as an Outstanding rating accompanied by a score of 99. Likewise, a Satisfactory rating accompanied by a score of 70 is just barely passing while a Satisfactory rating accompanied by a score of 89 is essentially a High Satisfactory rating.
The 100-point scale has the potential to add significantly more meaning and nuance to CRA ratings. In addition, the 100-point scale could be applied to ratings assigned to states and multi-state metropolitan areas in order to more effectively point out state and metropolitan areas where a bank performs well in addition to areas where it must make efforts to improve its performance.
The overall objective for ratings reform is not only to reduce the unjustifiably high pass rate but also to avoid a certain rating category from capturing the great majority of banks. It does not make sense that 90 percent of banks have been rated Satisfactory over the last several years. This bucket should be split into at least two since common sense would dictate that it is unlikely that 90 percent of banks would perform in the same manner. More refinement to the rating system would result in buckets that have less lopsided distribution and would result in even more responsible loans, investments, and services to LMI borrowers and communities. If for instance, 45 percent of the banks with Satisfactory ratings found themselves in a lower point category, a number of these would not want this lackluster rating and would be motivated to perform better.
Benchmarks in the CRA Tests Need to be More Objective, Promote Consistency, and Enable Better Comparison among Peers
Since the last major reform of the CRA regulations in 1995, the agencies have shied away from clearer benchmarks within the CRA component tests. Perhaps, they fear that benchmarks will be interpreted as an attempt at credit allocation. However, benchmarks can be clearer if they are established in reference to aggregate industry performance and demographic data. These benchmarks would not be a form of credit allocation since they assess a bank’s performance relative to its peers rather than establishing a certain percentage or quota. Moreover, they would make ratings more objective, consistent, and fair (ANPR Question 3).
Numerical guidelines with some qualification could correspond to the lending test’s ratings
On the retail lending test, a common performance measure is the percent of loans to LMI borrowers. The CRA exams provide extensive data on this performance measure and usually report whether the bank’s percentage of loans to LMI borrowers is above or below the aggregate percentage (all lenders in the geographical area). The reader is left with the impression that when a bank has a higher percentage of loans to LMI borrowers than the aggregate, its performance on this criterion is at least Satisfactory or is Outstanding. Likewise, when the percentage is below the aggregate percentage, the performance would be Low Satisfactory or worse. The shortcoming, however, is that the exams establish no range for performance that correspond to the various ratings. For example, if a bank’s percentage of loans to LMI borrowers was higher than one standard deviation from the aggregate percentage, perhaps this should correspond to High Satisfactory or Outstanding. Also, perhaps Outstanding performance would be reserved for banks that are not only above one standard deviation but are within a certain number of percentages points to a demographic statistic such as the percent of households that are LMI in the area.
Currently, Appendix A to the regulations does not provide clear guidance as to how performance on the various criteria correspond to various ratings. For instance, for the criterion of lending to LMI borrowers, Outstanding is described as “an excellent distribution, particularly in its assessment area(s), of loans among individuals of different income levels and businesses (including farms) of different sizes, given the product lines offered by the bank.” High Satisfactory corresponds to“a good distribution, particularly in its assessment area(s), of loans among individuals of different income levels.” These vague descriptions do not provide sufficient guidance to either banks or community groups.
Numerical guidelines should not be regarded as a guarantee that a certain rating on a criterion would be attained. Numerical guidelines such as those for LMI borrowers described above would be qualified with an explanation that final judgment of performance on a criterion also depended on assessment of performance context. However, stakeholders could still be reasonably confident that numerical ranges for the various ratings would apply most of the time. Also, stakeholders would be confident that if a bank landed in a range that would be normally reserved for Outstanding performance, it would most likely not score lower than Satisfactory. A deviation from an expected rating would occur rarely and would only apply in outlier situations such as for geographical areas where housing was quite affordable, meaning that several lenders were able to make a high percentage of loans to LMI borrowers.
Among other steps, investment test ratios should be benchmarked to peer performance with quantitative and qualitative criteria appropriately weighted in the rating
On the investment test and the community development lending part of the lending test, CRA examiners currently use ratios such as investments compared to assets but will not benchmark these ratios to peer performance. The lack of benchmarking makes it difficult for a reader of CRA exams to verify that performance merits the rating received. Alternatively, for the quantitative criterion of the dollar amount of investments, the rating could be determined in relation to peer performance (ANPR Question 8).
For instance, NCRC computed the level of community development (CD) financing (loans and investments) on the most recent CRA exams of the top 25 banks ranked by asset size. We found that the median level of CD financing per Tier 1 capital was 17.9 percent and median for CD financing per assets was 1.2 percent. The top quartile of either of these ratios could be considered Outstanding, the second quartile could be Satisfactory, and the lower two quartiles could be the failed ratings for the quantitative criterion. Then the qualitative criteria, including responsiveness to needs and the level of innovation, would be judged based on performance context analysis and comments of community groups and other stakeholders submitted to the CRA examiner. The final rating for a community development test would be determined by weighing the quantitative and qualitative criteria. The weights could be 65 percent for the quantitative criteria and 35 percent for the qualitative criteria (the qualitative criteria would be more subjective so it is probably sensible for it to be weighed less than 50 percent).
In another change to introduce more efficiency and objectivity on CRA exams, the agencies should consider converting the investment test on the large bank exam to a community development test. The community development test would consider both community development loans and investments. The community development test already exists for the intermediate small bank exam. Moreover, the form of community development financing should not depend on whether a bank needs to scramble for more investments or community development loans in order to pass its current lending or investment tests. The form of financing should be dictated by efficiency and affordability, not by which test a bank needs to pass. In order to ensure that a bank is not completely shirking either community development lending or investments, lending and investments must be analyzed separately on the test. Finally, the service test on a large bank exam should still analyze community development services since CD services are distinct from loans and investments.
The OCC must align with the Federal Reserve Board and FDIC so that CRA rating, examinations, and remedial standards are consistent
ANPR Question 2 asks whether the current CRA regulations are applied consistently. Consistency has been undermined by the OCC’s recent unilateral changes to CRA implementation and would be further undermined without coordination with the other regulators on CRA rulemaking. Over the last year, the agency has stretched out CRA exams for large banks and weakened fair lending and merger reviews of banks. The OCC must rescind these changes and align any future changes with the Federal Reserve Board and the Federal Deposit Insurance Corporation (FDIC) in a manner that comports with the CRA.
In a recent bulletin, the OCC describes a 48 month cycle for large banks with more than 30 rating areas and 36 months for those with less than 30 areas. This is a stretch out for OCC banks with more than 30 areas. The Treasury Department documented that OCC time periods for all large banks was previously 36 months. This was already on the high end compared to the other agencies that had time periods ranging from 12 to 36 months for large banks. Less frequent CRA exams reduce accountability and incentives for banks to perform in a consistently vigorous manner in fulfilling their CRA obligations.
The OCC has also issued a memo that dilutes the negative impact of discrimination and violation of consumer protection law on a bank’s CRA rating. The memo states that if a legal violation occurred in a type of lending not considered on a CRA exam, the violation will not impact the CRA rating. If such a violation is widespread, however, it must result in a ratings downgrade because the bank, overall, is not meeting credit needs in a responsible manner. Also, the memo states that double downgrades will be rare and reserved for particularly egregious violations. The language is too restrictive and does not place enough discretion for examiners to determine when and how violations should result in double downgrades.
The OCC has also made it easier for banks with failed CRA ratings to grow through mergers with other financial institutions or acquiring branches. Previously, a bank with a failed rating would be presumed not to be able to secure approval for its merger or branch application. The recent OCC memo provides such a bank with a potential path to approval if it can point out vague benefits associated with approval. Since only about two percent of banks fail on an annual basis, these institutions are exceptionally poor CRA performers. Currently, the only penalty for a failed CRA grading is the possibility of denial of merger or branch applications; one of only a few sticks that motivate banks to pass their CRA exams. A presumption that applications will be denied for failed CRA performance must remain the regulatory practice. The only exception that NCRC would support would be in cases where the bank seeks to open a new branch in an LMI census tract, or tract that is majority people of color, in order to facilitate increased lending and services to the underserved.
Better Data Reporting Would Reduce Grade Inflation and Improve Consistency
NCRC believes that improvements in data reporting would reduce grade inflation overall and improve the consistency of performance measures and benchmarks on CRA component tests. (ANPRQuestions 29 and 30). The most rigorous parts of the CRA exam currently are those that rely on annual HMDA and CRA small business loan submissions. The annual submissions improve the ability of CRA examiners to track performance over time and consider in their ratings decisions whether performance was constant or inconsistent over time. In addition, the public availability of the HMDA and small business loan data allows community groups and other stakeholders to provide their analysis of lending trends to the CRA examiners, which most likely results in examiners catching aspects of performance that they may have overlooked and thus results in more rigorous ratings.
Community development data should be collected annually and publicly disseminated
Similar to HMDA and small business data, the community development lending and investment data must be submitted annually and publicly by banks on a census tract level, a county level, and for the assessment areas. The community development data should also be reported separately for the major categories of community development including affordable housing, community services, economic development, and activities that revitalize and stabilize LMI census tracts. CRA exams often contain tables breaking out community development financing into the major categories. Community development loans, investments, and grants should be reported separately since each of these types of financing respond to different needs and contain different levels of explicit or implicit subsidies. Finally, the number of applications and denials of community development loans should be publicly reported. NCRC member organizations have described instances of arbitrary denials, which might be reduced if transparency is improved via data disclosure.
With annual data broken out by geographical area and purpose, examiners, community groups, and banks can track bank performance on a timelier basis and correct areas of weaknesses considerably before CRA exams. Annual submission of community development finance data would also facilitate determinations of whether banks are meeting community development needs in their assessment areas and thus whether they would receive consideration for community development financing outside of their assessment areas. In addition, annual submission would enable the regulatory agencies to create a database that could show which counties are well served and which are underserved based on the dollar amount of community development financing per capita. This would help establish a list of underserved counties across the country that banks would be allowed to serve provided they have met needs in their assessment areas. Finally, asset levels and Tier One capital levels must be reported annually so that the dollar amount of community development financing can be compared to bank capacity in a timely manner.
As any community development financing data reporting is implemented, the agencies must carefully oversee data collection and community development activities to ensure that the financing is not displacing or harming LMI people. For example, in high cost areas of the country, abusive multifamily lending in LMI tracts has facilitated the displacement and eviction of LMI tenants. In response to concerns raised by NCRC members and others, banks have implemented reforms to their multifamily lending practices and state agencies have issued guidelines to ensure responsible multifamily lending. For example, New York state advises banks to conduct due diligence of landlords and property owners, assess if appraisals are accurate, and analyze loan terms and conditions to make sure that current rents would not have to increase substantially in order for property owners to repay loans. CRA examiners must monitor banks and penalize them on CRA exams if they are financing abusive activities in LMI census tracts and also disallow community development data being reported that include predatory financing.
Regulators should make some improvement to small business lending data now notwithstanding the CFPB forthcoming Section 1071 rulemaking
The small business loan part of the lending test is not as rigorous as the home lending section because the small business data is not as refined. Some stakeholders complain that home lending receives priority over small business lending. This perception is fueled by the limited amount of small business lending data available to examiners. To improve the rigor of the small business part of the test and to improve the ability of examiners to determine if banks are serving the smallest businesses, small business data must be improved to include more categories rather than just above and below $1 million in revenue. Adopting the categories used by the Census Bureau, the revenue categories reported should include businesses with annual revenues $50,000 and below, $50,000 to $100,000, $100,000 to $500,000, $500,000 to $1 million, $1 million to $5 million, and $5 million and above. Improved data on the revenue size of the small business would enable examiners and the public to know whether the bank was lending to the smallest businesses as well as serving businesses in LMI tracts (which is another criterion on the lending test).
When the revenue of the small business is unknown, it must be reported that way instead of the current method of reporting that does not differentiate loans in which revenue is unknown and known. The ratios currently on exams that express loans to businesses with less than one million dollars in revenues as a percent of all loans can be quite inaccurate when banks make substantial amounts of loans to businesses with unknown revenues. There is no way presently to subtract the number of loans with revenues unknown from the denominator.
Similarly, originations must be reported separately from refinances so that examiners can separately analyze whether banks are meeting needs for originations and refinances. Lastly, credit card loans must be reported separately from non-credit card loans so that examiners and community groups can assess whether banks are satisfying the needs for larger and lower cost loans in addition to shorter term credit card loans with higher interest rates. The federal bank agencies should not wait for the implementation of Section 1071 of the Dodd-Frank Wall Street Reform and Consumer Protection Act to make refinements to the small business data and improve its ability to assess bank performance. CRA is significantly handicapped in its ability to assess banks’ responsiveness to the credit needs of small businesses. This retards economic growth in underserved communities and is also unfair to banks that specialize in small business lending and want exams that showcase their responsiveness to the needs of small businesses.
Community Benefits Agreements and Conditional Approvals Must Be Considered on CRA Exams and No Safe Harbors
The CRA regulation must formally recognize the relationship between the merger application process and CRA exams. A bank merger can dramatically change a bank’s ability to meet credit needs. Bank merger law expressly seeks to avoid reductions in banks’ abilities to meet credit needs by requiring banks to provide a public benefit as a result of their merger. In order to avoid outcomes in which mergers decrease banks’ abilities to meet credit needs, CRA exams must regularly evaluate whether banks are meeting goals established in either conditional merger approvals or community benefit agreements (CBA). CRA exams are a natural follow-up examination to determine whether banks are meeting their commitments made during mergers.
NCRC appreciates that the OCC’s June 2018 memo instructs CRA examiners to determine whether banks are meeting the goals in CRA plans required in conditional merger approvals. Conditional merger approvals issued by the OCC and the other federal agencies require plans with goals in cases in which the merging banks need to improve their performance in lending, investing, and/or providing services. Consequently, compliance with CRA plans should logically influence the bank’s rating on its next CRA exam.
CRA examiners must also assess bank compliance with CBAs. CBAs serve the same purpose as conditional approvals in that they ensure that bank CRA performance improves instead of regressing. In some cases, CBAs can avoid the need for conditional merger approvals. Community groups and banks negotiate CBAs when banks are seeking to acquire other banks or improve their CRA ratings. CBAs are forward-looking commitments establishing specific lending, investment, and service goals over a multi-year time period. Since 2016, NCRC and banks have negotiated $86 billion in CBAs.
The Treasury Department recognizes CBAs as an “effective tool” to “demonstrate how the approved application would benefit the communities served.” However, the Treasury Department also seemed to confine its recommendation regarding the value of CBAs to banks with Low Satisfactory or below ratings. To the contrary, CBAs are beneficial for a wide array of banks. One of the key benefits of CBAs is that they encourage banks to engage in partnerships with community organizations and other stakeholders in an effort to bolster its lending and investment in LMI communities.
The agencies should codify that CBAs and conditional merger approvals will receive follow-up evaluation in CRA exams. The CBAs and conditional merger approvals must contain verifiable goals targeting LMI people and people of color and LMI communities and communities of color. NCRC and our member organizations have had experience with so-called voluntary commitments and even conditional merger approvals in which the dollar amounts appear impressive but the income targeting is unclear (no income definitions were used or the income definitions did not correspond with CRA’s income definitions). Also, half or more of the lending and investing promised were not directed at LMI communities.
A CRA rating must not become a safe harbor providing expedited merger approvals or automatic approvals. Periodically, proposals will surface that Outstanding ratings should confer an easy merger approval process since they indicate impressive CRA performance. This position blithely ignores the mandate in merger application law that a bank provide a prospective or future public benefit arising from the merger. Congress created this requirement because mergers can dramatically change the capacity of a bank to lend and provide services to communities. In the wake of significant institutional change, Congress wanted to insure that mergers provided public benefits in the form of more, not less, lending and bank services to the public.
A past rating does not ensure that future performance will remain at the same level, particularly when mergers can have profound impacts on bank management and structure. Secondly, even a bank with an Outstanding rating can have inconsistencies in its performance on component tests or across assessment areas that must be addressed during mergers. Thirdly, performance could have declined since the previous CRA exam that can be several years old. This necessitates data analysis and public input in order to most accurately assess the performance of a bank. Any proposal to short circuit the application process or limit public comment will violate the OCC’s responsibilities under bank merger law.
Instead of so-called safe harbors, low ratings such as Low Satisfactory in any assessment area should trigger a public hearing requirement so that all parties have time to thoughtfully consider how bank performance can improve after a merger. Under a point system that NCRC proposes above, a point range such as 70 to 80 suggesting Low Satisfactory as a rating overall or in any assessment area should also trigger public hearings.
The OCC’s Metric-Based Framework Is a Misnomer and the One Ratio Would be Complex, Burdensome, and Not an Adequate Measure of Responsiveness to Local Needs
The OCC describes a one-ratio concept as a metric-based approach to CRA exams. This clever sounding phraseology makes it appear that the one ratio would be somehow more objective than the performance measures currently used on CRA exams. However, the one-ratio would make CRA exams less objective by replacing a set of performance measures that capture variations in local performance in meeting credit needs with reliance on one measure that fails to capture the complexity and nuances of bank performance across its assessment areas (Questions 7 through 9 in the ANPR).
The One Ratio is a Step Closer to Credit Allocation – An Approach Rejected by Senator Proxmire
During 1977 when Senator Proxmire was promoting CRA on the Senate floor and in hearings, he was insistent that CRA encourages banks to respond to local needs and was not intended as method of credit allocation. The one- ratio would be a step towards credit allocation by expressing a ratio of the dollar amount of CRA activity divided by Tier I capital or some other denominator reflecting bank capacity. If the OCC then established ranges of ratios corresponding to the various ratings, this would resemble a quota. The one ratio would be just one measure upon which most of the rating would depend and would earn a bank a certain rating depending on what ratio the bank attained. In contrast, the current CRA exams rate banks on several measures in several assessment areas. The measures are relative comparing banks to each other and to demographic benchmarks. The measures enable banks to compensate for poor performance in certain areas or activities with better performance elsewhere. A bank rating is not dependent on one test or quota but on how well it performs, in the aggregate, on several measures across local areas.
The first version of the CRA legislation contained specific requirements to serve primary service areas. The primary service area was defined as a compact area covering branches from which a bank expected to obtain more than one half of its deposit customers. Then, when a bank applied to merge or open a branch, it would have been required to indicate what proportion of customer deposits would be reinvested in primary service areas. Some Senators and bank representatives criticized this provision as cumbersome and bureaucratic. They also stated that it would prevent banks from addressing other unmet needs in other areas including rural communities (that presumably would have lower levels of deposits). In response, Senator Proxmire re-worked the legislation and introduced it without this provision. Instead, agencies were required to examine banks and rate them based on how well they met credit needs. The one ratio would be reverting back towards the original discarded ratio, which was replaced by a more flexible array of measures and metrics.
Throughout the pivotal year of 1977, Senator Proxmire repeated his mantra that CRA was not credit allocation but a requirement to meet local credit needs. He states:
- It (CRA) does not provide for credit allocation. To criticize reinvestment incentives as a form of credit allocation is disingenuous. It would not allocate credit, nor would it require any fixed ratio of deposits to loans. But it would provide that a bank charter is indeed a franchise to serve local convenience and needs, including credit needs.
- A public charter conveys numerous economic benefits and in return it is legitimate for public policy and regulatory practice to require some public purpose, without the need for costly subsidies, or mandatory quotas, or a bureaucratic credit allocation scheme.
The Numerous Flaws of a One-Ratio
Like a ratio requirement in the first version of the CRA legislation, the one ratio has some deceptive appeal based on its surface simplicity. The one ratio would consist of the dollar amount of a bank’s CRA activities (loans, investments, and services to LMI borrowers and communities) divided by the bank’s assets. The ratio is supposed to reflect CRA effort compared to a bank’s capacity and would influence a bank’s CRA rating. The same level of effort or dollar amount of CRA activities would not earn a bank with a larger amount of resources the same CRA rating as a bank with a smaller amount of resources.
The OCC has promoted the one ratio as a means to inject more certainty and clarity into CRA exams. The idea behind the one ratio is that it will immediately signal to banks whether they are in compliance with CRA and can expect to pass their next CRA exam. However, before applying this concept to CRA, we must pause and remember the purpose of the law.
The CRA statute reminds us that banks “have continuing and affirmative obligations to help meet the credit needs of the local communities in which they are chartered.” The key word is local. One ratio cannot tell an examiner, a bank, or a member of the public how responsive a bank is to its various service areas. Current CRA exams ensure banks are responsive to local needs by establishing assessment areas where branches take deposits. CRA exams scrutinize to what extent a bank makes loans and investments and offers services to LMI people and communities in its various assessment areas.
When reaching conclusions about performance, exams also assess to what extent a bank responds to different needs in its assessment areas. For instance, preserving affordable housing is a priority need in a metropolitan area experiencing rapid housing price increases whereas financing small businesses and job creation is a priority need for a metropolitan area with high unemployment. If a bank does well in job creation initiatives in the high unemployment metro area, but not so well in financing affordable housing in the expensive metro area, it would probably receive higher marks for its performance in the area with high unemployment than the expensive area. The exam then tallies performance across assessment areas to develop an overall rating. Differences in responsiveness to local needs therefore gets factored in exams with assessment areas.
Assessment areas allow examiners to conduct performance context analysis and determine to what extent banks are responding to differing priority needs. A CRA exam focused on the one ratio is incompatible with performance context analysis (ANPR Question 9). Moreover, an exam focused on the one ratio would not be effective in considering public input regarding local needs (ANPR Question 11). Examiners are currently required to consider community comments on local needs and how well banks are responding to them. Examiners take these comments into account when reaching conclusions about bank performance in assessment areas. A one-ratio focused exam, in contrast, would not explicitly factor community input into the conclusions of performance for each assessment area. At the very least, trying to factor in community group comments about bank performance and needs in several assessment areas into a one ratio analysis is an extraordinary difficult, cumbersome, and convoluted exercise that is hard to fathom. There is no better way to determine if banks are fulfilling their responsibilities under CRA than to listen carefully to local residents and consider their perspectives when reaching conclusions and ratings for various assessment areas. The one ratio-centered exam, in short, violates the intent and purpose of CRA to require banks to respond to local needs as articulated by community residents.
The OCC suggests that the ratio could be adjusted to provide more weight to activities that are particularly responsive to distressed communities with high needs for credit. For instance, an investment of $1 million in a distressed community can be weighted by a factor of two, meaning it will count for $2 million in the numerator of the ratio (ANPR Question 10).While this may sound appealing, consider how complicated and subjective it would be to do this weighting for banks, particularly large banks, which serve upwards of 20 states and hundreds of counties.
The other downside is that generous and frequent weighting (multiplying loans and investments by 2 or more) could easily result in half or less the dollar amount of loans and investments. The current system, while not perfect, can better adjust for responsiveness by weighing the importance of performance in each assessment area, including distressed areas. This avoids crude outcomes like one half the number of loans and investments equaling the same ratio due to weighting.
Another shortfall of the one ratio is that banks are likely to find the easiest loans and investments to undertake regardless of how well they respond to needs. They will search for lowest risk and highest yield. They will search for larger dollar loans and investments to bump up their numerator. Banks would likely gravitate to the 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 one ratio would divert attention from the harder-to-serve local needs. Instead of working closely with community groups and other stakeholders to meet needs in assessment areas, the banks will be mostly engaged in a mathematical exercise to increase their numerator.
The one ratio will lead to a preoccupation with assigning dollar figures to activities instead of measuring their effectiveness. One illustration of this fixation is the notion of converting community development services such as financial counseling into dollar amounts by applying a labor rate to the number of bank employee hours devoted to community development services (ANPR Question 12).This would artificially inflate these services. What is more important than calculating dollar values is measuring outcomes such as number of people opening bank accounts.
The approach outlined in Question 12 of the ANPR would counter the recent revisions to the Interagency Q&A document that were geared towards measuring outcomes of community development services and alternative service delivery. Measuring dollars could likely result in less impressive outcomes because it would emphasize large scale programs such as one hour introductory lectures on financial counseling in which a large number of bank employees can log labor hours which would be converted to dollars. In contrast, more labor intensive counseling such as working one-on-one to raise client credit scores involves fewer and more highly skilled employees and hence less dollars for the one ratio. However, this type of counseling has a more desirable outcome in terms of qualifying people for loans.
While the ANPR suggests that assessment areas will remain on CRA exams, the OCC seems to emphasize weights in ratio calculations, which suggests diminished importance of assessment areas. NCRC is not opposed to metrics or ratios on exams. NCRC adapts a ratio from existing CRA exams of community development loans and investments compared to assets when we propose community benefit agreement targets to banks. But this is just one quantitative measure we use in CBAs which contain a series of other performance measures and targets. While not opposed to metrics or ratios, NCRC is opposed to a single metric like the one ratio being determinative of the score on CRA exams. The OCC is developing a false dichotomy when it suggests that a one ratio focused exam is based on metrics while the current system is not.
NCRC Analysis of the Proposed One Ratio and Percent of Loans in Assessment Area
NCRC analyzed the relationship between the one ratio and the percent of loans in CRA assessment areas. NCRC computed the one ratio for home and small business lending. Specifically, we computed the dollar amount of home and small business loans made to LMI borrowers and census tracts and divided this dollar amount by the dollar amount of Tier One Capital. We then compared the one ratio to the percent of loans in CRA assessment areas of the top 20 banks ranked by asset size.
The one ratio would need to be consistent with assessment area analysis if the one ratio is not to undercut the local needs focus of CRA exams. If the one ratio is to be used as an indicator on CRA exams, it would be better if larger one ratios correspond to high percentages of loans in assessment areas. CRA exams evaluate lending in assessment areas, therefore, high percentages of loans in assessment areas means that CRA exams are analyzing most of a bank’s lending to determine the extent to which lending is being made to LMI borrowers and census tracts.
In contrast, if large one ratios correspond to smaller percentages of loans in assessment areas, then a large one ratio is accompanying CRA exams that do a poor job of analyzing how well loans meet local needs of LMI borrowers and communities. When low percentages of loans are in assessment areas, CRA exams are analyzing a small sample of loans in determining whether banks are serving LMI borrowers and communities.
In addition, a relatively high dollar amount of loans to LMI borrowers and census tracts in the numerator of the one ratio does not automatically translate to a high percentage of loans to LMI borrowers and census tracts. It could also be the case that a bank with relatively low percentages of loans in assessment areas is concentrating its lending to LMI borrowers and census tracts in assessment areas and neglecting LMI borrowers and census tracts in other areas.
In the NCRC sample of twenty very large banks, the one ratio ranged from 12.3 percent to under 1 percent with a median of 6.2 percent. The median of the percentage of home loans in assessment areas was 89.2 percent and the median of the percentage of small business loans in assessment areas was 92.1 percent.
The NCRC analysis reveals that a development of a one ratio measure along with current CRA exam procedures for delineating assessment areas would be a cause for concern. For the top quartile of banks sorted by the one ratio, the median percentage of home loans in assessment areas was only 65.8 percent (see tables below). Likewise for the top quartile of banks sorted by the one ratio, the median percentage of small business loans in assessment areas was the lowest (at 84.4 percent) of any other quartile. In contrast, the lowest quartile for the one ratio had a median percentage of small business loans in assessment areas of 93 percent (see tables below). Particularly for small business lending, larger one ratios correspond to lower percentages of loans in assessment areas (either due to purchases of loans made outside of assessment areas or non-branch lending not captured by assessment areas).
If the agencies were to significantly diminish the importance of assessment areas in favor of the one ratio, this analysis reveals that the one ratio is not an effective proxy for performance in lending to LMI borrowers and communities. While larger one ratios generally reflect higher dollar amounts of loans to LMI borrowers and communities in this analysis, they do not necessarily reflect higher percentages of loans to LMI borrowers and communities. This is likely the case since larger one ratios are associated with lower percentages of loans in assessment areas and therefore lower percentages of loans being evaluated by CRA exams.
In addition, some lenders in lower cost parts of the country could have low one ratios because the dollar amount of their home lending is lower due to the lower cost of housing. However, they could be offering a higher percentage of their loans to LMI borrowers and census tracts.
A use of the one ratio would need to be accompanied by reforms in designating assessment areas in order to provide confidence that CRA exams were effectively evaluating lending to LMI borrowers and communities. Assessment areas would need to capture not only lending in geographical areas with bank branches but also geographical areas without bank branches where the particular bank’s lending accounted for a significant volume of the market’s total lending or was a significant volume of the bank’s total loans. Furthermore, the importance of assessment area analysis for determining CRA ratings must not be diminished in order to preserve confidence that CRA exams are effectively evaluating lending to LMI borrowers and communities.
If the OCC were to significantly diminish the importance of analysis of performance in assessment areas in favor of the one ratio, NCRC’s analysis leads to the conclusion that a one ratio focused exam would encourage banks to adopt national models of high volume, standardized lending using correspondents, brokers, and/or the internet. The percentage of loans in assessment areas with branches would diminish, which would reduce the amount and percentage of lending that is most responsive to local needs (with product features that respond to local needs). A high touch localized model of lending that LMI borrowers find most comfortable and conducive to their success in paying the loan would be replaced by a model that likely results in less overall lending to LMI borrowers and lending that is less safe and sound (see below discussion on branch based lending).
Asset Thresholds for CRA Exams: CRA Exams Must not Reduce Requirements for Any Category of Banks
A technically arcane but vitally important issue that is likely to be raised in industry comments in response to the ANPR is asset thresholds on CRA exams. The ANPR invites these comments by stating that “some stakeholders have expressed the view that asset thresholds have not kept pace with bank asset sizes.” In addition, Question 7 of the ANPR asks whether alternative methods exist for evaluating CRA performance, taking into account banks’ business models and asset levels. The American Bankers Association (ABA) advocated eliminating the Intermediate Small Bank (ISB) category altogether. Small banks with assets below $313 million have a CRA exam that looks solely at retail lending. ISB banks with assets between $313 million and $1.3 billion have a retail lending test and a community development (CD) test. The CD test scrutinizes CD lending and investing such as construction loans for affordable housing or investments in Small Business Investment Corporations (SBIC).
Eliminating the ISB category means that the ISB banks would just have a retail test. NCRC estimates that ISBs finance about $3 billion annually in CD projects or about the same amount of annual funding as the Community Development Block Grant (CDBG) program. If the CD test is eliminated for ISB banks, their CD financing would plummet by 50 percent or more according to NCRC’s analysis. ISB banks would lack incentives to engage in CD financing. LMI communities cannot afford to lose a valuable source of financing for community development, a source of financing that is equal to CDBG, which is the major community development funding program of the Department of Housing and Urban Development (HUD).
In addition, reducing the number of banks subject to the large bank exam eliminates the service test for more banks. The service test examines the distribution of branches in LMI census tracts and the bank services and deposit products provided to LMI customers. The ISB exam reduces the number of service criteria on the CRA exam while the small bank exam eliminates the service test altogether. Reducing or eliminating the service test for a large number of lenders will result in significantly less access to bank branches, deposit products, and loans for LMI customers and census tracts. Research described below illustrates the vital importance of branches in increasing access to loans and deposit products for LMI customers and communities. While these notions of reducing or eliminating the service test may make exams easier for banks, it will result in reduced abilities of banks to serve the convenience and needs of communities, in violation of the intent of CRA.
The OCC needs to further contemplate the regional and state level impacts on any proposals to adjust asset thresholds. Smaller banks that would be most impacted by asset changes are disproportionately located in rural counties. Easier exams would reduce the amount of community development financing and bank branching in the region of the country already experiencing shortages of these vital activities.
The exams for the various categories of banks are sensible and do not need thorough overhaul (see ANPR Question 6). The large bank exam’s division into lending, investment, and services components logically examines different types of activities. Above, NCRC suggested changing the investment test to a community development test, but this is a tweak, not an overhaul. The ISB exam structure is likewise reasonable though the agencies made a serious mistake when deleting the service test and regular analyses of branching patterns for ISB banks. The strategic plan is a good option for banks that want to develop tailored CRA plans with community input. The wholesale and limited purpose test is acceptable for banks that are not in the retail business but is grossly inappropriate for credit card lenders for the reasons stated below related to the importance of examining consumer lending.
Keep CRA-Qualifying Activities Focused on and Targeted to LMI, But Clarify What Counts
Several of the ANPR questions contemplate more non-LMI and non-local CRA qualifying activities. Instead of broadening what qualifies for CRA credit, the OCC could clear up some of the confusion about what qualifies for CRA credit by providing more real-time determinations of what would count on the exams. As stated above, the Treasury Department recommended a system of early determinations.
Over the years, the Interagency Q&A document has refined an approach to focusing on LMI communities or people adopted by the 1995 final rule implementing the current CRA regulations. The preamble to the 1995 final rule states:
- Under the rule, community development includes activities outside of low- and moderate-income areas if the activities provide affordable housing for, or community services targeted to low- or moderate-income individuals or if they promote economic development by financing small businesses and farms.
- The final rule also requires that, in order to be community development loans or services or qualified investments, activities must have community development as their primary purpose. Activities not designed for the express purpose of revitalizing or stabilizing low- or moderate-income areas, providing affordable housing for, or community services targeted to, low- or moderate-income persons, or promoting economic development…are not eligible.
It is clear from the 1995 rule that either LMI communities or LMI people in non-LMI communities are to benefit from CRA activities. The focus is on LMI people and communities.
Developed and refined after the 1995 final rule, the Interagency Q&A provides careful exceptions to the focus on LMI people such as including mixed-income housing as an example of community development. This is as it should be; exceptions need to be narrowly targeted to promote positive outcomes like economic mobility that might provide low-income residents affordable housing in high-opportunity areas. Broad based exceptions would result in LMI communities being neglected in favor of easier-to-serve affluent communities. A second example in the Interagency Q&A is a supermarket on the edge of a middle-income census tract but adjacent to a LMI tract. In this case, the Q&A says that the supermarket development helps stabilize the LMI tract by providing “needed shopping services that are otherwise unavailable in the low-income community.” Again, the point is that it is acceptable for CRA activities to not exclusively benefit LMI areas but activities must have a significant benefit for LMI areas or LMI people.
Overall, however, the focus is on LMI people and communities. Additional Q&As in the Interagency Q&A document suggest that generally the substantial majority of recipients of community development activities should be LMI persons (the language varies between “majority” and “primarily” recipients being LMI).
In contrast to CRA regulation and the Interagency Q&A, the American Bankers Association (ABA) believes that a focus on LMI people and communities is “inconsistent with the law that requires banks to meet the credit needs of the community, not just particular parts of the community.” The ABA further elaborates that, “the regulation has caused examiners to exclude appropriate consideration of the revitalization or stabilization activities in middle- or higher-income neighborhoods.
Arguing that CRA consideration for revitalization and stabilization should not be restricted to LMI communities, the ABA states, “A loan to expand a new repair shop creates jobs regardless of whether the repair shop is located downtown or in the suburbs.” Firstly, while this statement about creating jobs on its face may be true, if the repair shop was in a non-LMI tract and part of a huge corporate chain that was easily creditworthy, CRA consideration would have gone towards a loan that would have been made in any case. The opportunity cost would be significant in that CRA would allow banks to fill up their CRA exams with loans to corporate chains in non-LMI areas that would likely be creating jobs for mostly non-LMI people. CRA would cease motivating banks to make loans to the smallest businesses that are creditworthy but which need more counseling and innovative underwriting approaches to receive loans. Secondly, the statement is an oversimplification because the suburbs also have significant numbers of LMI census tracts that are underserved.
Earlier this year, a trade paper quoted an official of the ABA as advocating for CRA consideration of financing a hospital no matter where it was constructed. While hospitals are vital institutions, new hospitals in affluent parts of a city without ready access to transit have limited utility for LMI people.
In sum, the statutory language, the regulatory history over 41 years, and economic reality in LMI communities suggest that CRA exams providing positive consideration for diffuse activities spread across LMI and non-LMI areas is not consistent with the intention of CRA. The insertion of LMI neighborhoods in the statute suggests a focus and a priority of LMI people and neighborhoods which have the greatest unmet needs due to redlining and other factors. If the OCC proposes and adopts a final rule not reaffirming LMI people and communities as the target of CRA activities, it would be in violation of the purpose of the statute and how the agencies have interpreted the statute over 40 years.
CRA-qualifying activities must remain focused on credit and community development needs
The second theme, allowing CRA to provide consideration for activities that do not respond to the needs for credit, has been picked up by the OCC in the ANPR. For example, the ANPR asks whether internships at banks for LMI young adults should be considered on CRA exams (ANPR Question 17). While these types of initiatives are desirable, the original purposes of CRA must guide final determinations of what counts. Congress enacted CRA to increase access to loans, deposit accounts, and other banking services. Consideration for meritorious activities that do not directly combat redlining and/or lack of access to banking will result in a regulation that frustrates the purpose of CRA to revitalize credit starved communities.
As desirable as it is for bank employees to volunteer as laborers for nonprofit affordable housing developments, construction work is not a community development service which is defined according to the CRA regulation as “related to the provision of financial services.” This type of volunteer service should not be accorded favorable consideration on the CRA service test. Doing so would reduce the amount of financial education provided by banks as they send out their employees to do construction work instead of providing financial education. It would be a misallocation of talent; bank staff have expertise in financial education whereas the skill and desire to construction work is more broadly distributed across the entire population.
CRA-qualifying activities should be those with measurable outcomes
NCRC does not support automatic CRA consideration of financing of projects, programs, or organizations because they have “a mission or purpose of community development,” as ANPR Question 15 suggests, without additional analysis. Instead, the programs and organizations must have data to verify that they are engaged in community development with specifics about how many affordable housing units developed or community services provided to LMI people. Relying just on mission statements can invite abuse and support for organizations not engaged in community development. In addition, federal, state, and tribal programs could be considered as community and economic development if they target people or communities that are LMI.
Community development projects that promote the Fair Housing Act’s goals of affirmatively furthering fair housing (AFFH) are consistent with CRA credit. Examples could be community development financing that: supports fair housing choice, promotes, economic diversity, facilitates the mobility of low-income residents from low-opportunity areas to high-opportunity areas,preserves affordable housing, or finances very small businesses in gentrifying census tracts.
In response to ANPR Question 16, Congressional intent would be consistent with a regulatory interpretation that identified underserved and disadvantaged census tracts that would be eligible for community development activities. Data analysis conducted by the FFIEC would identify tracts with characteristics such as high denial rates, low numbers of loans per capita or per owner-occupied housing units, and other indicators of economic distress such as high foreclosure rates or low property values. In addition to LMI tracts, these underserved and distressed tracts would qualify for community development. This expansion of areas for community development would provide more regulatory clarity by identifying community development opportunities for CRA-qualifying activities as well as targeting areas in true need for credit and community development.
Financial education currently receives favorable consideration on CRA exams because it is a community development service that helps LMI people access bank products. CRA exams, however, can do a better job in evaluating and considering financial education. Financial education should receive favorable consideration if it demonstrates solid outcomes such as considerable numbers of people opening and successfully maintaining bank accounts. In contrast, CRA exams must not consider programs that do not deliver outcomes but instead merely take clients’ money and/or inflate the ratings of the service test. A considerable body of research has identified critical elements of successful programs that include classroom and one-on-one counseling. The agencies should provide this research (and guidance on how to use it) to CRA examiners so that the examiners can assess whether the financial education programs are beneficial to LMI clients. The CRA examiners should also seek data on outcomes from the banks on the outcomes of the programs.
In addition, the ANPR suggests that digital literacy could be considered on CRA exams (Question 19). Digital literacy, while important, does not directly increase the chances of accessing responsible credit like financial education so it should not receive consideration as a stand-alone activity on CRA exams. Digital literacy could be a component of financial education programs that receive favorable consideration on CRA exams. Care must be taken to consider outcomes, which could be neglected in a one ratio focused exam that is fixated on converting financial education efforts to dollar amounts for the numerator of the ratio.
Questions 17 and 20 of the ANPR ask whether community development should be expanded to consider support for other populations such as the disabled or small and disadvantaged service providers. In the case of populations like the disabled or veterans, if they are LMI, then community development activities should be targeted to them. It is likely that LMI members of these populations currently do not receive adequate access to credit and banking service so including their more affluent counterparts would divert community development initiatives to where it is needed most. The OCC needs to engage in more research about needs before expanding the income range of disabled and veteran recipients eligible for community development.
The OCC should parse the term, small and disadvantaged service providers, further about what it means. In a procurement context, it often refers to small and disadvantaged businesses that are minority- and women-owned. NCRC would support CRA consideration for programs that follow well established criteria for providing contracting opportunities for smaller women- and minority-owned entrepreneurs.
Regarding the internship question on Question 17, NCRC would not support indiscriminate CRA consideration for internships or apprenticeships. If the internship is part of a thoughtful workforce and job training initiative that meaningfully trains young LMI adults for well-paying career opportunities, then it could merit CRA consideration.
Core Retail Lending Evaluations Must Remain on CRA Exams
ANPR Question 21 is a loaded and puzzling question. It asks whether CRA evaluations should consider home mortgage, small business, and small farm lending and whether these loans should only be considered if they are made to LMI borrowers and LMI areas or other identified areas. Even asking whether home lending, small business lending, and small farm lending should still be considered on CRA evaluations betrays the essential purpose of CRA in combating redlining and shortages of credit in LMI neighborhoods and communities of color. If the OCC seriously is entertaining this possibility, the agency needs to carefully review the 1977 CRA hearings.
On the floor of the Senate on January 24, 1977 Senator Proxmire stated:
- Although communities depend on their local institutions to supply capital for economic development, home mortgage loans, consumer credit, and a variety of other needs, the “convenience and needs” criteria as applied by regulators in their allocation of charters and branch approvals, have focused almost exclusively on deposit services.
The Senator was not dismissing the needs for deposit services but stating that the merger application process at that time did not usually review the need for credit. Foremost in the Senator’s mind was that CRA would fix the lack of attention paid to access to credit. Major retail products offered by banks over the decades include home, small business, and small farm lending. Eliminating these from CRA exams directly contradicts Senator Proxmire’s intentions.
The analysis of retail lending must also consistently consider affordability and sustainability. If a bank is making a disproportionate amount of high-cost retail lending compared to its peers, it should receive a lower rating, particularly if the high-cost bank is serving a similar percentage of LMI borrowers. This would suggest the high-cost bank is most likely price gouging or seeking to make more profit rather than adjusting price because it is serving a different population with more risky borrower profiles. More cost consciousness on the part of CRA exams would also be consistent with the desires of Senator Proxmire. During the CRA hearings, Senator Proxmire agreed with a witness that asserted that CRA exams should differentiate among banks that are offering retail products with differing costs. He says at one point to the witness, “I think you make a good point. In other words, if a bank is doing a vigorous job of promoting credit cards in the local community at 18-percent interest that is one thing; if they are promoting mortgages that is something different.”
The other part of Question 21 asks whether retail lending should be considered on CRA evaluations only when it is made to LMI borrowers and communities. On the criterion of lending by income of borrowers, retail lending should be counted only to LMI borrowers. Common performance measures on that criterion include the percentage of loans to LMI borrowers. The criterion of lending to LMI communities is more complicated since a portion of the loans will be made to middle- and upper-income borrowers. It is desirable to have an income mix of borrowers in LMI census tracts in order to promote economic diversity. However, in census tracts that are rapidly gentrifying, an imbalance of lending with the majority of loans being granted to middle- and upper-income borrowers should be penalized. This would only exacerbate the number of LMI residents being displaced and facing inferior housing and living opportunities outside of their current neighborhoods. Researchers at the agencies could identify census tracts with hot markets that are rapidly gentrifying by using data on property values, dollar amounts of home loans, and demographic trends using census data.
CRA Exams Must Consider Consumer Lending and Ensure it is Affordable and Complies with Consumer Protection Law
A noteworthy part of the Senator’s quote above was his mention of consumer credit. In response to ANPR Question 22, regular consideration of consumer credit on CRA exams would be aligned with the purposes envisioned by the founders of CRA. Currently, CRA exams infrequently examine consumer credit and only do so upon request of a bank or when consumer lending constitutes the substantial majority of a bank’s loans. Since consumer lending remains a significant credit need in LMI communities, a lack of regular examination of consumer lending is not only contrary to Senator Promire’s intention but contributes to an over-reliance on high-cost and abusive lending such as payday lending in LMI communities.
The current regulatory guidelines limiting examination of consumer loans are arbitrary. They are not grounded on an assessment of a bank’s suite of loan products or the credit needs of communities. Moreover, the current designation of credit card lenders as limited purpose and wholesale lenders not subject to a retail lending test is an abrogation of the responsibility of federal bank agencies to assess whether credit needs are being met in a safe and sound manner. As a result of these stunted procedures regarding examination of consumer lending, consumer lending is examined infrequently. A recent GAO study found that only 25 percent of large bank exams looked at consumer lending while only 3 percent of intermediate small bank and 6 percent of small bank exams evaluated consumer lending.
A plethora of needs for unmet consumer lending exists ranging from the needs for small dollar loans to finance emergency expenses, automobile loans (particularly for rural counties and other parts of the country without regular transit service), credit cards, and student loans. It is also critically important that CRA exams carefully assess whether this consumer lending is responsible, affordable, compliant with consumer protection laws, and not unfair or deceptive. The Interagency Q&A over the years has recognized the importance of low-cost products mentioning cost considerations in the provision of remittances, deposit accounts, check cashing services, and alternative service delivery. These instructions must be codified in the regulation and implemented rigorously in CRA exams to avoid another round of abusive and unsustainable lending afflicting vulnerable communities. The agencies should also consider guidelines such as an annual percentage rate (APR) limit of 36 percent, which is the maximum APR that can be charged to military personnel under the Military Lending Act.
Evaluations of Small Business Lending Must be Focused on the Smallest Businesses
In response to an ANPR Question 23 about small business lending, NCRC believes that the current definition of small business lending as non-residential loans of $1 million is basically sufficient but could be updated annually to take inflation into account. A recent GAO report finds that the $1 million limit for a small business loan should be updated to $1.6 million to account for inflation. It would be reasonable to update this limit to account for inflation since the costs of equipment and other needs has increased for small businesses.
At this time, however, the data does not support a further increase beyond inflation. NCRC conducted research to investigate dollar amounts of loans to the largest businesses with revenues above $1 million during 2016, the most recent year for which data was available for this analysis. NCRC excluded lenders with average loan amounts of $10,000 or less as these are banks that focus on smaller credit card loans. The sample was constructed to identify the loans of the largest dollar amounts. The average loan amount was $70,611 and the average loan amount for the quartile of loans with the largest amounts was $343,469. These loans do not come close to the $1 million limit. Therefore, according to the publicly available data, NCRC does not observe that banks are scraping up against the $1 million limit or are constrained by it.
The ANPR has an example in Question 23 of including “all loans to businesses that meet the Small Business Administration (SBA) standards.” NCRC believes that SBA standards are not consistent with the focus of CRA on underserved small businesses and LMI communities. For example, the SBA considers a brewery with 1,250 employees, and a television broadcast company with average annual receipts of $38.5 million, to be small businesses. NCRC believes that it makes sense to use these standards in evaluating contracting opportunities as the SBA does, but that it is unacceptable to apply them to a bank’s CRA performance, particularly on the lending test criterion of lending to small businesses with revenues under $1 million. The CRA evaluates how well financial institutions are meeting capital needs of underserved businesses, and businesses with annual revenue of $38.5 million do not face the same obstacles at securing financing as businesses with less than $1 million.
In response to ANPR Question 24, small business loans should continue to be considered on the retail portion of the lending test. If they are considered on the community development portion of the lending test, double counting results. Home loans are solely considered on the retail portion of the lending test although they also indirectly support job creation in the housing industry just as small business loans support job creation by the small business. Hence, small business loans should be considered on the retail portion of the test just as home loans. Loans designated as falling above the threshold of dollar amounts, whether that remains at $1 million or is adjusted, can certainly be considered to have a community development purpose. Those loans in LMI census tracts can be classified as revitalizing and stabilizing LMI areas. The agencies should also favorably consider the innovation and responsiveness of these community development loans if they finance businesses that are paying a livable wage as defined in reference to local and/or state minimum wage laws. Finally, loans to a nonprofit organization that is engaged in community development activities should be considered on the community development part of the test.
Loan Purchases and MBS Investments
Regarding loan purchases and secondary market activity asked in ANPR Questions 18 and 25, NCRC recognizes the need for banks to purchase mortgages from larger entities as they attempt to identify qualified CRA assets in the community. The share of loans made by private mortgage companies now constitute the majority of home loans made in the country. Large mortgage companies are able to sell mortgages through efficiency and volume that undercuts the ability of CRA regulated institutions in serving LMI geographies. Often the loans being purchased are securitized because these non-bank institutions cannot portfolio those loans. Thus it becomes important for banks, particularly those of small and medium-sized, to have access to purchasing those loans.
Purchases receiving full consideration as originations should also include purchases of loans made by Community Development Financial Institutions (CDFIs), B corporations, and intermediaries whose primary role is to make or find LMI and CRA qualified loans for lenders that have invested in those entities.
The practice of “churning” in the banking business is a whole other matter. Churning is the practice of one bank purchasing loans made to LMI borrowers for CRA exam consideration. After the CRA exam is over, the bank then sells the loans to another bank seeking to purchase loans for favorable CRA consideration. The Interagency Q&A contains an answer to a question in which it bans this practice.
The enhancements to HMDA data required under the Dodd-Frank Wall Street Reform and Consumer Protection Act that include unique loan identifications available to federal agencies should make it easier to identify and penalize churning. Banks must not receive CRA consideration for secondary market activity if this activity enables and facilitates abusive lending. Instead, the activity, depending on its extent, must either result in an overall downgrade or a lower rating on one of the component tests. Examples of financing abusive activity are purchasing loans from a fintech or nonbank that are high interest rate and that do not include proper disclosures of loan terms and conditions to borrowers. In addition, financing abusive payday lending or investing in abusive payday lenders must be penalized.
Loans Held in Portfolio Should be Evaluated Carefully
In response to ANPR Question 26, if loans held in portfolio are made under a special affordable loan program that does not use standardized secondary market underwriting, then these loans can receive consideration under the innovative and flexible part of the retail lending test. This part of the current lending test could be improved. This part of the test usually just lists the special affordable programs and sometimes reports the volumes of loans made under these programs. However, the analysis usually stops there and does not consider what portion of the bank’s loan portfolio is represented by these special affordable loans.
If the affordable loan subset is large, then it would boost the rating on the lending test. However, examiners should also take into consideration that a bank making large volumes of special affordable loans would also likely be making high percentages of loans to LMI borrowers and census tracts. Thus, the bank would score well on those criteria on the lending test and would likely have a good rating. Consequently, the innovative and flexible part of the lending test should not have too high a weight because it would inflate a bank’s rating that is already likely doing well.
Evaluations of Branches and Banking Services Must Remain on CRA Exams
Current large bank exams include a service test that assesses a bank’s branching patterns and provision of deposit accounts as well as a bank’s financial counseling efforts. Some stakeholders have asserted that branching is becoming less important as on-line lending has increased. However, despite the increase in on-line lending and banking, branches remain important for customers of all income groups. A recent report by Celent, an industry consultant, reports that a large majority of adults of all income groups (77 percent) prefer in-person interactions at a bank branch when conversations are needed for complicated matters such as applying for loans.
This finding of consumer preferences is consistent with FDIC research revealing that branch density or branches per capita actually increased over the decades. The number of branches more than doubled from 1970 through 2014 while population growth in the United States was 50 percent. Even in more recent years with the increase of mobile and on-line technology, branch density was higher than in previous years. In the last week of October, the FDIC released its 2017 unbanked and underbanked survey finding that 73 percent of banked households used tellers in the last year. Moreover, 81 percent of banked households that use mobile banking as their primary means of account access visited a bank branch in the last year. A quarter of these households used a branch 10 or more times in the last year.
The OCC asks a startling ANPR Question 27 of whether branching patterns, particularly in LMI census tracts, should continue to be evaluated on CRA exams. This should not even be a question. Branches are critical for helping LMI people obtain home loans, small business loans, and basic banking services. Lending transactions, particularly buying a home, are among the most complicated financial transactions most Americans will undertake. Lower income consumers, in particular, need one-on-one counseling for qualifying and applying for loans. Academic research, including NCRC’s, has revealed that home and small business lending increases in LMI neighborhoods with bank branches. In contrast, when branches close, lending decreases for several years, especially small business lending. This is likely due to the non-automated, labor intensive method of lending to and counseling traditionally underserved populations.
Ergungor finds that in low-income census tracts, greater access to branches is associated with more loans and lower interest rates on the loans. A one standard deviation increase in branch access is associated with a $592 increase in mortgage originations per household. In addition, borrowers who obtain mortgages from banks with branches in their neighborhoods are less likely to default than homebuyers using banks without branches or mortgage brokers. This is particularly true for borrowers with low credit scores. In contrast, there is no branch impact on the number of loans or loan performance in higher income tracts.
If CRA exams de-emphasize branches and do not analyze branching patterns by income level of census tract, the exams would likely result in less lending to LMI borrowers and census tracts. Instead of decreasing emphasis on branching, CRA exams should contain more data on the income levels of customers using deposit products so that exams can better judge the effectiveness of banks in serving LMI customers. More data on incomes of deposit customers is especially important for judging on-line banks.
A recent GAO study suggests that the CRA service test has played an important role in promoting access to branches for LMI census tracts but that the provision of deposit products remains uneven. The GAO gathered a large amount of data on bank branches and census tract demographics to conclude that residents of LMI tracts have as much access to branches as residents in middle- and upper-income tracts in rural areas and large metropolitan areas but not in small metropolitan areas with population of less than 100,000. At the same time, however, LMI customers are less likely than middle- and upper-income customers to have a bank account and more likely to use alternative financial service providers for services like check cashing or obtaining small dollar consumer loans.
The unequal access to bank services is probably influenced by a lack of consistent analysis by the service test of the provision of bank services. The GAO report included a survey of 219 CRA exams of banks of all sizes to see how often different aspects of banking services were examined. The report revealed that just 46 percent of large bank exams scrutinized bank provision of deposit products. However, the GAO did not probe further and investigate whether the exams merely mentioned the products or actually looked at how many were offered to LMI people. Better data on bank services and products would promote more rigorous and consistent CRA service test evaluation.
De-emphasizing branches and not analyzing their distribution in LMI tracts will significantly reduce access to loans and banking services. Instead of diminishing the service test, it must be enhanced through improvements to data on bank services. Moreover, de-emphasizing bank branching would be inconsistent with Congressional intent as reflected in the comments of CRA’s chief sponsor, Senator Proxmire. As stated above, the CRA hearings presented evidence of fewer branches and more limited hours and services at those branches in LMI census tracts. Hence, CRA needed to examine branches in LMI neighborhoods and needs to continue to do so today.
Summary of Recommendations
The summary of NCRC’s recommendations is the following:
Assessment Areas (AAs) Must be Expanded and Activities Outside of AAs Must be Considered Carefully
- AAs must remain the foundation of CRA exams. AAs must include areas with bank branches and ATMs. AAs must also be expanded to include areas where banks are making significant number of loans beyond bank branches and/or engaged in other significant business activity including collecting deposits.
- AAs must include the great majority of lending and other business activity.
- CRA consideration for community development lending and investing outside of AAs should be based on rigorous data analysis. If a bank’s current level of CD financing in AAs is similar to previous CRA exams in which it passed and also at or above peer levels of CD financing, then it can venture outside of its AAs. As a final check, CRA examiners must consult members of the public to assess whether the bank has met needs in AAs.
- When banks engage in CD financing outside of AAs, the agencies could develop a list of underserved and distressed counties based on objective data analysis that could be priority recipients of the CD financing.
Reconsider How Affiliates and Holding Companies are Evaluated
- Non-bank affiliates must be automatically included on CRA exams.
- CRA exams should be conducted on the holding company level.
People and Communities of Color Must be on CRA Exams
People and communities of color must be considered on CRA exams. The agencies have several options:
- They can consider activities to serve people and communities of color in a manner similar to LMI people and communities;
- They can examine activity in underserved census tracts identified via data analysis (the underserved tracts are likely to include many predominantly minority tracts);
- They can also resurrect the fair lending reviews in CRA exams that examined lending to people of color and explained in detail the statistical methodology that was employed to test for discrimination in lending in use before the 1995 regulatory changes.
CRA Ratings Reform and Improving Benchmarks
- The ratings could be accompanied by a point scale of 1 to 100 that would reveal more gradations in performance.
- Improve robustness of CRA exam analysis by developing guidelines for how to evaluate performance on various criteria such as percent of loans to LMI borrowers. The agencies should develop ranges of performance of these criteria that correspond to various ratings.
- On the investment test or CD part of the lending test, develop measures like CD financing compared to assets and compare these to peers. Develop ranges on these measures based on the distribution of banks’ performance that would correspond to various ratings.
Data Improvements Will Lead to More Robust Exams
- Develop a database of CD lending and investing that would consist of annual data on the major categories of CD finance. The agencies must exclude activities from the database that displace LMI people such as financing slumlords and must penalize this predatory financing.
- Improve the small business loan data to include additional revenue categories of small businesses receiving loans, improve the tracking of loans with revenue of business unknown, separate originations from refinances, and credit card from non-credit card lending.
Merger Application Process and CRA Exams
- CRA exams must consider conditional merger approvals and community benefit agreements (CBAs). The merger approvals and CBAs must have meaningful goals for LMI people and communities.
- The agencies must not establish so-called safe harbors expediting merger approvals and instead must hold more hearings when banks score poorly on CRA exams overall or in assessment areas.
The One Ratio Must not be Determinative
- The one ratio concept as a determinative measure on CRA exams will violate the purpose of CRA to require responsiveness to local needs. The one ratio is also a step closer to credit allocation which was rejected by Senator Proxmire.
- One ratio focused exams will frustrate community input and performance context analysis. The one ratio will further inflate CRA ratings through frequent weighting which applies multiples to actual loan and investment levels. The one ratio will favor easier and larger dollar value activities as opposed to more complex activities that are more responsive to local needs.
Retain Current Asset Thresholds and Exam Structure
- Asset thresholds must not be adjusted to qualify more banks as small banks as opposed to ISB banks or to qualify more large banks as ISB banks. The result will be significant declines in lending, CD financing, branches and services to LMI people and communities.
- The current categories of CRA exams for banks of various asset sizes and business models are generally reasonable. However, the agencies must not grant wholesale and limited purpose status to credit card banks and other banks with significant retail lending and/or services.
- Consider replacing the investment test of the large bank exam with a CD test. Move CD lending to the CD test.
LMI People and Communities Must Remain the Priority
- LMI people and communities must remain the focus on CRA exams. Allowing banks to satisfy their CRA requirements by neglecting LMI people and communities in favor of more affluent communities would violate the intent of CRA as revealed by the Congressional record leading up to passage of CRA. CRA is intended to address the most underserved communities that are redlined and/or in most need of credit and community development.
- The agencies must use CRA exams in a more creative and sophisticated manner to support economic diversity and promoting the goals of the Fair Housing Act, including toaffirmatively furthering fair housing (AFFH) (such as by placing affordable housing for LMI people in high-opportunity areas ) and prevent displacement of LMI people and small businesses that can accompany gentrification.
Activities Must Remain Focused on Credit Needs and Community Development
- The agencies must not broaden the range of CRA activities and divert CRA away from activities that directly address credit needs or community development. This would be a violation of CRA’s intent and would frustrate efforts to reinvest in and revitalize LMI communities.
- Home and small business lending must remain core activities evaluated on CRA exams.
- CRA exams should also carefully scrutinize consumer lending to ensure it is affordable and complies with fair lending and consumer protection statutes.
- The agencies must not adopt SBA definitions of small business which would result in significant diversions of lending away from the smallest and most underserved small businesses.
- Loan originations must count more than loan purchases. The agencies should adopt a tiered approach to considering loan purchases and MBS investments with priority to secondary market activities that support B corporations, CDFIs, and nonprofits.
- Evaluations of branching and services must remain on CRA exams. If these criteria on the service test are diminished, LMI people and communities will experience considerable less lending and access to deposit accounts and other bank services.
OCC Must Rescind its Unilateral Actions and Align with the Other Agencies
- The OCC must rescind its harmful bulletins and directives that stretch out CRA exams for the largest banks, diminish penalties for harmful and discriminatory activities, and provide allowance for banks failing their CRA exams to merge.
The challenge and opportunity in CRA reform is successfully addressing the gaps in CRA coverage while not disturbing the core mechanisms of public input, transparency, and local accountability. NCRC has maintained that increasing the effectiveness of CRA would entail increasing opportunities for public input, improving the quality of data on CRA exams, mandating the inclusion of affiliates on CRA exams, evaluating bank lending, investment, and service to people and communities of color, and expanding assessment areas to consider non-branch lending.
In contrast, the OCC proposes a number of reforms that would make CRA exams more convenient for banks but would reduce banks’ incentives to meet convenience and needs of the communities in which they conduct business. The one ratio approach is a prime example of a concept with surface appeal for banks but that would dramatically reduce CRA exams’ effectiveness in holding banks accountable to meeting local needs. A one ratio focused CRA exam would reduce opportunities for public input and would be more complex to administer than the current approach.
Combined with the recent actions taken by the OCC to reduce the impact of fair lending violations on CRA ratings, stretching out CRA exams, and reducing the consequences of failed ratings, the ANPR’s concepts would, if implemented, would significantly reduce lending, investment, and services to LMI people and communities. Moreover, the lending, investment, and services would be less responsive to local needs. All of this would be in direct conflict with the intentions of Senator Proxmire and Congress in 1977 that sought an end to redlining in modest income communities and communities of color.
Thank you for your serious consideration of our comments on these important matters. If you have any questions, please contact us or Josh Silver, Senior Advisor, on 202-628-8866.
President and Founder, NCRC
Jesse Van Tol
Chief Executive Officer, NCRC
The organizations listed below support the views expressed in this letter:
NCRC Board Members
Robert Dickerson, Jr., Chairperson
Birmingham Business Resource Center
Ernest Hogan, Vice Chairperson
Pittsburgh Community Reinvestment Group
Jean Ishmon, Vice Chairperson
Northwest Indiana Reinvestment Alliance
Catherine Hope Crosby, Secretary
Chief of Staff, Mayor’s Office, Toledo, Ohio
Ernest (Gene) E. Ortega, Treasurer
Rural Housing, Inc.
Bethany Sanchez, Immediate Past Chairperson
Metropolitan Milwaukee Fair Housing Council
Chicanos Por La Causa Inc.
CASA of Oregon
Housing Education & Economic Development
National Trust for Historic Preservation
Arkansas Fair Housing Commission
Hawaii Alliance for Community-Based Economic Development
Inner City Press
Maryellen J. Lewis
Michigan Community Reinvestment Coalition
Housing Assistance Council
Metropolitan St. Louis Equal Housing & Opportunity Council
Community Reinvestment Alliance of South Florida
California Reinvestment Coalition
Tennessee Human Rights Commission
Marceline A. White
Maryland Consumer Rights Coalition
Center for Responsible Lending
Grounded Solutions Network
Leadership Conference on Civil and Human Rights
National Coalition for Asian Pacific American Community Development (CAPACD)
National Alliance of Community Economic Development Association (NACEDA)
National NeighborWorks Association
National Urban League
 Federal Reserve Governor Lael Brainard, Keeping Community at the Heart of the Community Reinvestment Act, May 18, 2018, https://www.federalreserve.gov/newsevents/speech/brainard20180518a.htm and Remarks by Martin J. Gruenberg, Member, Board of Directors, Federal Deposit Insurance Corporation on The Community Reinvestment Act: Its Origins, Evolution, and Future at Fordham University, October 29, 2018, Lincoln Center Campus; New York, New York, https://www.fdic.gov/news/news/speeches/spoct2918.html
 Among other provisions: 12 U.S.C§ 2903. Financial institutions; evaluation
(a) In general. In connection with its examination of a financial institution, the appropriate Federal financial supervisory agency shall-
(1) assess the institution’s record of meeting the credit needs of its entire community, including low- and moderate-income neighborhoods, consistent with the safe and sound operation of such institution;
 “Where Congress explicitly enumerates certain exceptions to a general prohibition, additional exceptions are not to be implied, in the absence of a contrary legislative intent.” Andrus v. Glover Const. Co.,446 U.S. 608, 616-17 (1980) (citing Continental Casualty Co. v. United States, 314 U.S. 527, 533 (1942) (“Generally speaking, a ‘legislative affirmative description’ implies denial of the non-described powers.”).
 123 Cong. Rec. 17,604 (1977).
 Congressional Record – Senate, June 6, 1977, p. 17630.
 Congressional Record – Senate, June 6, p. 17630.
 Robert E. Litan, Nicolas P. Retsinas, Eric S. Belsky, Susan White Haag, The Community Reinvestment Act After Financial Modernization: A Baseline Report, U.S. Treasury Dept., April 2000.
 Ibid; Joint Explanatory Statement of the Managers of the Committee of Conference, Public Law 95-128, Subcommittee on Housing and Community Development of the Committee on Banking, Finance and Urban Affairs, House of Representatives, 95th Cong., 1st Sess. Washington, D.C.: U.S. Government Printing Office (October 1977), p. 92.
 Id. at note 7 U.S Treasury Baseline Report.
 Lei Ding and Leonard Nakamura, Don’t Know What You Got Till It’s Gone: The Effects of the Community Reinvestment Act (CRA) on Mortgage Lending in the Philadelphia Market, Working Paper No. 17-15, June 19, 2017,https://www.philadelphiafed.org/-/media/research-and-data/publications/working-papers/2017/wp17-15.pdf
 Raphael W. Bostic and Hyojung Lee, Small Business Lending Under the Community Reinvestment Act, in Cityscape – Department of Housing and Urban Development publication, Volume 119, No. 2, Chapter 6, November 2017, https://www.huduser.gov/portal/periodicals/cityscpe/vol19num2/ch6.pdf
 Elizabeth Laderman and Carolina Reid, Federal Reserve Bank of San Francisco, “CRA Lending during the Subprime Meltdown” in Revisiting the CRA: Perspectives on the Future of the Community Reinvestment Act, a Joint Publication of the Federal Reserve Banks of Boston and San Francisco, February 2009. Also see, Neil Bhutta and Daniel Ringo, Assessing the Community Reinvestment Act’s Role in the Financial Crisis, Feds Notes, May 2015, https://www.federalreserve.gov/econresdata/notes/feds-notes/2015/assessing-the-community-reinvestment-acts-role-in-the-financial-crisis-20150526.html
 Joint Center for Housing Studies of Harvard University, The 25th Anniversary of the Community Reinvestment Act: Access to Capital in an Evolving Financial Services System, March 2002, http://www.jchs.harvard.edu/research/publications/25th-anniversary-community-reinvestment-act-access-capital-evolving-financial
 Joint Center for Housing Studies, pg. 61-72.
 Daniel Ringo, Federal Reserve Board, Mortgage Lending, Default, and the Community Reinvestment Act, June 15, 2017, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2585215, pg. 4 and 13.
 Adam Dettelbach, Josh Silver, Bruce C. Mitchell, Intermediate Small Banks: The Forgotten but Significant Resource for Affordable Housing and Community Development, NCRC, October 2017, https://ncrc.org/intermediate-small-banks-forgotten-significant-resource-affordable-housing-community-development/
 NCRC Forecast: Weakening the Community Reinvestment Act Would Reduce Lending by Hundreds of Billions of Dollars, September 2018, https://ncrc.org/ncrc-forecast-weakening-the-community-reinvestment-act-would-reduce-lending-by-hundreds-of-billions-of-dollars/
 Neil Bhutta, Jack Popper, and Daniel R. Ringo, The 2014 Home Mortgage Disclosure Act Data, in the Federal Reserve Bulletin, https://www.federalreserve.gov/pubs/bulletin/2015/articles/hmda/2014-hmda-data.htm, Figure 13 and accompanying narrative.
 Josh Silver, The Community Reinvestment Act and Geography, How Well do CRA Exams Cover the Geographical Areas that Banks Serve, NCRC, April 2007, https://ncrc.org/wp-content/uploads/2017/05/cra_geography_paper_050517.pdf, p. 9.
 See the CRA regulation, https://www.ffiec.gov/cra/regulation.htm, §25.22 lending test pertaining to consideration of affiliate lending and that affiliate lending is not considered in portion of loans in assessment areas.
 The consideration of lending outside of assessment areas is described in Q&A § __ .22(b)(2) & (3)—4, see Federal Register, Vol. 81, No. 142, Monday, July 25, 2016, page 48538. The OCC reiterated this procedure in a recent bulletin, https://www.occ.gov/news-issuances/bulletins/2018/bulletin-2018-17.html#ft6
 OTS 2009 Citicorp CRA exam, see http://www.occ.gov/static/cra/craeval/OTS/CRAE_14470_20091109_64.pdf
 OTS 2005 Capital One CRA exam, see http://www.occ.gov/static/cra/craeval/OTS/CRAE_13181_20050718_64.pdf
 OCC 2016 CRA exam of Bank of the Internet, https://www.occ.gov/static/cra/craeval/nov16/716456.pdf.
 Bank of Internet exam, p. 11.
 Bank of Internet exam, p. 7.
 Bank of Internet exam, p. 1.
 NCRC Congressional Testimony on Fintech Oversight, February, 2018, https://ncrc.org/ncrcs-congressional-testimony-fintech-oversight/
 NCRC, The Community Reinvestment Act and Geography, p. 9.
 Office of the Comptroller of the Currency, Reforming the Community Reinvestment Act Regulatory Framework Advance Notice of Proposed Rulemaking (ANPR), Federal Register, Vol. 83, No. 172, Wednesday, September 5, 2018, Proposed Rules p. 45057, https://www.regulations.gov/document?D=OCC-2018-0008-0001
 OCC Bulletin, Description: Supervisory Policy and Processes for Community Reinvestment Act Performance Evaluations, 2018-17, https://www.occ.gov/news-issuances/bulletins/2018/bulletin-2018-17.html
 Section 802 (b) of the CRA statute, see FDIC version, see https://www.fdic.gov/regulations/laws/rules/6000-1500.html#fdic6000hcda1977
 Memorandum for the Office of the Comptroller of the Currency, the Board of Governors of the Federal Reserve System, the Federal Deposit Insurance Corporation from the Department of Treasury, Community Reinvestment Act – Findings and Recommendations, April 2018, p. 8, https://home.treasury.gov/sites/default/files/2018-04/4-3-18%20CRA%20memo.pdf
 Treasury memorandum, p. 24.
 Department Justice, Federal Government and State Attorneys General Reach Nearly $1 Billion Agreement with SunTrust to Address Mortgage Loan Origination as Well as Servicing and Foreclosure Abuses, June 2014, https://www.justice.gov/opa/pr/federal-government-and-state-attorneys-general-reach-nearly-1-billion-agreement-suntrust
CRA Exam of Suntrust Bank, Federal Reserve Bank of Atlanta, March 2013, https://www.frbatlanta.org/-/media/Documents/banking/cra_pes/2013/675332.pdf, p. 2.
 The Interagency Q&A recognizes possible gaming and cherry-picking using affiliate loans and imposes a prohibition on such behavior, see §_.22(c)(2)(ii)—1 and –2, pgs. 48539 and 48540, Community Reinvestment Act, Interagency Question and Answers Regarding Community Reinvestment Guidance,Federal Register, Vol. 81, No. 142, Monday, July 25, 2016, Rules and Regulations https://www.gpo.gov/fdsys/pkg/FR-2016-07-25/pdf/2016-16693.pdf
 Aaron Glantz and Emmanuel Martinez, For People of Color, Banks are Shutting the Door on Homeownership, February 15, 2018, https://www.revealnews.org/article/for-people-of-color-banks-are-shutting-the-door-to-homeownership/
 See Table 2A of the Consumer Financial Protection Bureau, Data Point: 2017 Mortgage Market Activity and Trends, May 7, 2018, https://www.consumerfinance.gov/data-research/research-reports/cfpb-data-point-mortgage-market-activity-and-trends/
 NCRC, Home Mortgage and Small Business Lending in Baltimore and Surrounding Areas, November 2105, https://ncrc.org/home-mortgage-and-small-business-lending-in-baltimore-and-surrounding-areas/
 NCRC, Foreclosure in the Nation’s Capital: How Unfair and Reckless Lending Undermines Homeownership, April 2010, https://ncrc.org/foreclosure-in-the-nations-capital-how-unfair-and-reckless-lending-undermines-homeownership/
 NCRC, Income is No Shield Against Racial Disparities in Lending II: A Comparison of High-Cost Lending In America’s Metropolitan and Rural Areas, July 2008, https://ncrc.org/wp-content/uploads/2008/07/income%20is%20no%20shield%20ii.pdf
 Hearings before the Committee on Banking, Housing, and Urban Affairs, United States Senate, Ninety-Fifth Congress, First Session, S. 406, March 23, 24, 25 1977, pages 44-46.
 Banking Committee Hearings on S. 406, March 1977 page 47.
 Banking Committee Hearings on S. 406, March 1977, page 48.
 Banking Committee Hearings on S. 406, March 1977, page 136.
 Banking Committee Hearings on S. 406, March 1977, page 172.
 Banking Committee Hearings on S. 406, March 1977, pages 174-177.
 Federal Reserve Bank of Richmond, CRA Exam of Signet Bank, January 1996, pgs. 18-20, https://www.federalreserve.gov/dcca/cra/1996/460024.pdf
 Office of Thrift Supervision CRA Exam of CenFed Bank, November 1995, p. 9, https://www.occ.gov/static/cra/craeval/OTS/CRAE_01788_19951127_60.pdf
 NCRC analysis of ratings on the FFIEC webpage. In contrast, the pass rate in the first five years (1990-1994) of publicly available ratings was between 90 to 95 percent.
 Large Institution CRA Examination Procedures: OCC, FRB, and FDIC, April 2014, see page 17, via https://www.federalreserve.gov/bankinforeg/caletters/CA_14-2_attachment_1_Revised_Large_Institution_CRA_Examination_Procedures.pdf
 Aggregate = all lenders in the market. For the retail lending test, NCRC has not observed significant differences in the percentage of loans to LMI borrowers or communities by banks with various asset levels. Hence, NCRC recommends using the aggregate percentage. Also, banks should be encouraged to perform at least as good as or better than the average non-bank so we favor including non-banks in the aggregate percentage. It also appears to be common CRA exam procedure to use all lenders or the aggregate on the retail portion of the lending test. If the agencies wish to change aggregate comparisons to peer comparisons grouping banks by similar asset ranges, they should produce a comprehensive and objective analysis and share it for public comment.
 On the investment test and community development part of the lending test, ratios of CD lending to assets or investments to assets should probably compare a bank to its peers by asset class. For example, the FFIEC search engine appears to have reasonable asset categories for large banks of approximately $1 to $10 billion, $10 billion to $100 billion, and $100 billion and more. It is possible that banks with larger assets will have higher ratios. An analysis on CRA exams would not want to compare banks with less capacity for CD financing to banks with more capacity for financing. The agencies should conduct additional analysis to refine the asset categories used for comparative purposes.
 OCC Bulletin 2018-17, June 15, 2018, https://www.occ.gov/news-issuances/bulletins/2018/bulletin-2018-17.html
 OCC Bulletin 2018-23, August 15, 2018, https://www.occ.gov/news-issuances/bulletins/2018/bulletin-2018-23.html
 OCC, Impact of CRA Ratings on Licensing Applications, November 2017, https://www.occ.gov/publications/publications-by-type/other-publications-reports/ppms/ppm-6300-2.pdf
 See Blog Posts of the Association of Housing and Neighborhood Development, https://anhd.org/blog/bad-boy-carveoutand https://anhd.org/blog/new-york-state-lenders-you-are-accountable-multifamily-displacement-lending. Also see Kevin Stein, Banks Should do More to Prevent Renters from Being Displaced, American Banker, August 14, 2018.
 New York State Department of Financial Services, DFS Advises State Chartered Banks of Their Responsibilities in Lending to Landlords of Rent-Stabilized or Rent Regulated Multifamily Residential Buildings, September 25, 2018, https://www.dfs.ny.gov/about/press/pr1809251.htm
 NCRC, Small Business Data: Recommendations to the Consumer Financial Protection Bureau for Implementing Section 1071 of the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010, August 2014, p. 21, https://ncrc.org/recommendations-to-cfpb-implementing-section-1071-dodd-frank/
 OCC, Description: Supervisory Policy and Processes for Community Reinvestment Act Performance Evaluations, OCC Bulletin 2018-17, June 2018, https://www.occ.gov/news-issuances/bulletins/2018/bulletin-2018-17.html
 Treasury Memorandum, Community Reinvestment Act, April 2018, p. 22
 Banking Committee Hearings on S. 406, March 1977, page 6.
 Banking Committee Hearings on S. 406, March 1977, page 7
 Banking Committee Hearings on S. 406, March 1977, Testimony of the American Bankers Association, p. 315 and Senator Garn, page 324.
 Banking Committee Hearings on S. 406, March 1977, page 2.
 Congressional Record – Senate, January 24, p. 1958.
 Office of the Comptroller of the Currency (OCC), Advance Notice of Proposed Rulemaking (ANPR), Federal Register, Vol. 83, No. 172, Wednesday, September 5, 2018, https://www.gpo.gov/fdsys/pkg/FR-2018-09-05/pdf/2018-19169.pdf, pgs. 45056 and 45057.
 Section 802(a)(3) of the CRA statute.
 Interagency Q&A, Federal Register, July 2016,§_.24(d)(3)—1, p. 48542
 At the time of NCRC’s analysis, we used CRA activity for the year 2016, which is the latest year that had a complete account (Business and Mortgage lending data) of the CRA activity. Tier 1 data from 3/31/2018 was used and is available via https://www.usbanklocations.com/bank-rank/tier-1-core-risk-based-capital.html
Banks in this analysis are the following: Capital One, Manufacturers and Traders Trust Company, KeyBank, Branch Banking and Trust Company, Santander Bank, N.A., Wells Fargo Bank, BMO Harris Bank, Regions Bank, U.S. Bank, The Huntington National Bank, JPMorgan Chase Bank, MUFG Union Bank, Fifth Third Bank, SunTrust Bank, Bank of America, TD Bank, Citibank, Citizens Bank, HSBC Bank USA, Ally Bank
 OCC, ANPR, Federal Register, September 2018, page 45055.
 American Bankers Association, Second Published Request for Comments Under the Economic Growth and Regulatory Paperwork Reduction Act of 1996 (February 13, 2015), p. 7, https://www.regulations.gov/document?D=FFIEC-2014-0001-0077
 NCRC, Intermediate Small Banks, the Forgotten but Significant Resource for Affordable Housing and Community Development, https://ncrc.org/intermediate-small-banks-forgotten-significant-resource-affordable-housing-community-development/
 Housing Assistance Council, CRA in Rural America, The Community Reinvestment Act and Mortgage Lending in Rural America, January 2015, http://www.ruralhome.org/storage/documents/publications/rrreports/rrr-cra-in-rural-america.pdf
 Federal Register, May 4, 1995, Volume 60, Number 86 https://www.fdic.gov/regulations/community/community/crapreamb.txt
 Interagency Q&A, Federal Register, July 2016, Q&A, §_.12(h)—8:, p. 48530.
 Interagency Q&A, Federal Register, July 2016, Q&A, §__.12(g)—2, p. 48525.
 Interagency Q&A, Federal Register, July 2016, p. 48526 §.12(g)(2) and §.12(g)(3)—1:
 ABA comment letter, Community Reinvestment Act Regulation Hearings, August 2010, p. 12, https://www.aba.com/archive/Comment_Letter_Archive/Comment%20Letter%20Archive/cl-CRAHearings2010Aug31.pdf
 Ibid, p. 13.
 Rachel Witkowski, Will CRA Finally Get its Makeover, American Banker, March 9
 American Bankers Association, CRA Modernization, Meeting Community Needs and Increasing Transparency, December 2017, p. 4, https://www.aba.com/Advocacy/Documents/CRA-WhitePaper2017.pdf#_ga=2.192150499.839944790.1512674294-422164602.1512674294
 See 2015 HUD’s final rule on AFFH; see also NCRC on AFFH can retrieved here: https://ncrc.org/ncrc-comment-letter-defending-huds-affirmatively-furthering-fair-housing-rule/; NCRC on FHFA’s Duty to Serve rule can be retrieved at: https://ncrc.org/wp-content/uploads/2016/03/ncrc_fhfa_duty_to_serve_web.pdf.
 Marvin M. Smith, Danile Hocberg, William H. Greene, Federal Reserve Bank of Philadelphia, The Effectiveness of Pre Purchase Homeownership Counseling and Financial Management Skills,April 2014, https://www.philadelphiafed.org/-/media/community-development/homeownership-counseling-study/2014/homeownership-counseling-study-042014.pdf?la=en. An example of an evaluation of foreclosure prevention programs is the Urban Institute, National Foreclosure Mitigation Counseling Program Evaluation, Final Report, Rounds 3 Through 5, September 2014, https://www.neighborworks.org/Documents/HomeandFinance_Docs/Foreclosure_Docs/ForeclosureCounseling(NFMC)_Docs/2014_NFMC_UrbanInstituteReport.aspx
 Congressional Record – Senate, January 24, 1977, p. 1958.
 Banking Committee Hearings on S. 406, March 1977, p. 226.
 See CRA regulation, §25.22 (a) lending test
 Government Accountability Office, Community Reinvestment Act: Options for Treasury to Consider to Encourage Services and Small-Dollar Loans When Reviewing Framework, GAO-18-244: Published: Feb 14, 2018. Publicly Released: Mar 16, 2018, https://www.gao.gov/products/GAO-18-244
 Interagency Q&A, Federal Register, July 2016, p.48542, §l_.24(a)—1:.
 CFPB, What are My Rights Under the Military Lending Act, October 7, 2016, https://www.consumerfinance.gov/ask-cfpb/what-are-my-rights-under-the-military-lending-act-en-1783/
 U.S. Small Business Administration Table of Small Business Size Standards Matched to North American Industry Classification System Codes. Version 2017. Available online at https://www.sba.gov/sites/default/files/2018-07/NAICS%202017%20Table%20of%20Size%20Standards.pdf.
 Interagency Q&A, Federal Register, July 2016, p. 48537, §l_.22(a)(2)—6:
 Robert Meara, Delivering Excellent Customer Service: When and How Consumers Prefer Face-to-Face Engagement and What it Means for Banks, Celent, May 16, 2018.
 FDIC, Brick and Mortar Banking Remains Prevalent in an Increasingly Virtual World,
FDIC Quarterly, 2015, Volume 9, No. 1, https://www.fdic.gov/bank/analytical/quarterly/2015-vol9-1/fdic-4q2014-v9n1-brickandmortar.pdf
 FDIC, 2017: FDIC National Survey of Unbanked and Underbanked Households, p. 5, https://www.fdic.gov/householdsurvey/2017/2017report.pdf
 For a literature review of the impact of branches and assessment areas, see Josh Silver, The Importance of CRA Assessment Areas and Bank Branches, NCRC, June 2018, https://ncrc.org/the-importance-of-cra-assessment-areas-and-bank-branches/
 Hoai-Luu Q. Nguyen, Do Bank Branches Still Matter? The Effect of Closings on Local Economic Outcomes, December 2014, http://economics.mit.edu/files/10143
 Ozgur Emre Ergungor, Bank Branch Presence and Access to Credit in Low? to Moderate?Income Neighborhoods, Journal of Money, Credit, and Banking, September 2011, p. 1336, https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1538-4616.2010.00343.x
 O. Emre Ergungor and Stephanie Moulton, Do Bank Branches Matter Anymore?,Economic Commentary, August 4, 2011, http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.645.1823&rep=rep1&type=pdf
 GAO, Community Reinvestment Act: Options for Treasury to Consider, op cit.