This is the second in a series of articles reviewing public comments submitted in response to the Federal Reserve’s Advance Notice of Proposed Rulemaking to revise Community Reinvestment Act rules. See the first here.
In a previous article, I described how some big issues on reforming the Community Reinvestment Act (CRA) exposed a considerable difference of views among community groups and the banking industry. These issues involved assessment areas, asset thresholds and data collection. The federal agencies will have to work hard to find equitable compromises that increase reinvestment in communities but also respond to concerns about ease of implementation.
However, there was some convergence of views on a series of other important issues, including performance measures and ratings categories. On these issues, the remaining differences appear to be manageable.
Performance measures should not combine income groups or products
Currently, CRA exams assess retail lending performance to low-income borrowers and communities separately from moderate-income borrowers and communities. The Federal Reserve Board (board) proposed in its Advance Notice of Proposed Rulemaking (ANPR) to combine low- and moderate-income borrowers and communities in the evaluation of lending performance to determine if a bank passed on its retail lending subtest. The board reasoned this would result in fewer calculations and would simplify the CRA retail test. However, simplicity should not always be the ultimate objective. The Consumer Banker Association (CBA) wanted to make sure that changing the lending subtest “does not degrade levels of CRA performance in low-income communities.”
Agreeing with this concern, NCRC noted in our comment letter that combining the income categories might encourage banks to focus on lending to moderate-income borrowers in order to pass. Since it is easier for banks to lend to moderate-income borrowers than low-income borrowers, combining the categories could skew lending to moderate-income borrowers.
The National Association of Home Lenders (NAAHL) supported combining income categories saying that lending to lower-income borrowers would be difficult in higher-cost markets. This potential issue, however, is dealt with in current CRA exams that do not combine income categories. In higher-cost markets, the industry, as a whole, will not be making as high a percentage of loans to low-income borrowers. Thus, the industry percentage of loans to low-income borrowers will be lower, which, in a sense, automatically adjusts for the difficulty of lending in higher-cost markets. The lower industry percentage of loans is then compared to the bank that is being examined. In NCRC’s view, the CBA’s view wins out on this issue.
Relatedly, the board proposed that product categories be combined to determine if a bank passes on its retail lending test. Again, the CBA pointed out that different product categories have different percentages of loans to various income groups. For a variety of reasons, lower-income borrowers do not use Home Equity Lines of Credit (HELOCs) as much as upper-income groups. Combining HELOCs with other types of loan products such as home purchase loans will therefore not provide as accurate a picture of how well banks are reaching lower-income groups with different products. Current CRA exams scrutinize product performance separately and this should remain the practice.
Presumption of Satisfactory, performance ranges and number of ratings need to be designed carefully to avoid grade inflation
In addition to simplifying exams, the board sought to increase clarity and consistency through its proposals regarding the presumption of satisfactory and performance ranges. On the retail lending test, the board proposed to clarify when a bank passed with a satisfactory rating. For example, if the bank issued a percentage of home loans to low- and moderate-income borrowers (LMI) that was 70% of the industry percentage, it would be presumed to have a satisfactory rating on that part of the home lending test. NCRC observed that this could be a low bar since our work analyzing lending data tends to show a clustering around the industry average percent of loans to LMI borrowers. The board did not consider NCRC’s observation nor estimate how its proposal would affect ratings.
Some of the trade associations picked up on this concern, particularly when considering the board’s proposal to reduce the number of ratings from five to four on the subtests such as the retail lending subtest. While the overall rating for a CRA exam is one of four ratings, the subtests have five currently: Outstanding, High Satisfactory, Low Satisfactory, Needs-to-Improve and Substantial Noncompliance. The board would replace High and Low Satisfactory on the subtests with Satisfactory in order to conform to the overall rating categories.
NAAHL was concerned that if the board establishes its threshold on presumption of satisfactory, the four ratings will create such a low bar that banks would “race to the bottom,” engage in minimal lending and still pass exams. The CBA had a similar concern, stating, “The designation (of five possible ratings) helps to incentivize banks to take on more innovative or impactful CRA activities as they seek to improve from a “Low Satisfactory” rating to a “High Satisfactory” rating or beyond.”
Moreover, the board’s performance ranges would work better with five subtest ratings in order to reveal meaningful distinctions of performance. For example, if further data analysis confirms that the proposed 70% threshold described above is not a difficult benchmark, it could be part of a range for Low Satisfactory. For example, the bank would be at Low Satisfactory if it was in a range of 70% to 89% of the industry benchmark. It would be in the higher rating categories if it was above 90% of the industry benchmark.
Re-structuring of CRA exams has some agreement and some differences
The current CRA exams for large banks with assets more than $1.3 billion have an investment test that examines community development investments like Low Income Housing Tax Credits (LIHTC) and investments in Small Business Investment Companies (SBICs). The lending test looks at retail lending but also has a component called community development lending that assesses construction loans for affordable rental housing and other loans for projects that have a neighborhood-wide impact.
Over the years, some stakeholders have advocated for the creation of a community development test that would replace the investment test and would consider both community development lending and investments. NCRC agrees with this proposal since banks sometimes structure financing as either a community development loan or investment to make sure they have a minimal amount of that type of financing in order to pass either the lending or investment tests. With both considered in one test, there is less of a possibility of tailoring financing to pass a test instead of pursuing the type of financing that is most efficient for the project. In addition, with community development lending as part of the current lending test, NCRC has found that performance in retail lending does not correlate as well as it could with ratings on the lending test. In other words, combining retail lending and community development lending hindered the ability of the lending test to accurately assess retail lending performance.
The bank trades supported the community development test but differed in some respects from community groups on how the community development test would assess activities. One possible shortcoming of a community development test is that banks end up making considerably fewer investments than community development loans. Stakeholders are concerned that the new test would further exacerbate the current trend of fewer investments than loans, and thus lead to a shortage of funding for instruments like LIHTC or SBICs.
A way to resolve this is not only to consider investment and lending together but to also consider separate totals for community development lending and investments. The board would then decide how much to weigh the separate community development lending and investment subtotals. In contrast to the weighting approach that NCRC supports, the Independent Community Bankers Association (ICBA) asked the board to apply a multiplier of 2 to investments and 10 to grants in order to compensate for the lower dollar amounts of this financing. The shortcoming of multipliers is that the financing being multiplied could decrease since banks figure they would receive credit for an artificially higher level of the financing than they actually offered. Separate evaluation measures for various types of financing avoid this problem.
The board seemed to elevate simplicity as the priority objective in too many instances. While simplicity is desirable, it cannot overrule accuracy and make CRA exams less able to assess whether banks are meeting various credit, capital or deposit needs. In addition, clarity and consistency can help make CRA exams more understandable to all stakeholders without sacrificing rigor; simplicity usually entails reducing rigor. Current CRA exams are too inconsistent in how they evaluate performance. CRA reform, through the increased use of thresholds and performance ranges, can increase consistency and clarity.
While the board and the other regulatory agencies will have to sort out a thicket of conflicting views on assessment areas, data collection and asset thresholds, the areas of agreement and narrower areas of disagreement on some of the other important issues can perhaps serve as a springboard to getting all the parties closer together and also making the harder decisions on the thornier questions.
Josh Silver is a senior policy advisor at NCRC.