NAAHL Community Reinvestment Act Comments

The interaction of three policies is likely to determine how motivated banks will be to maximize reinvestment.

1. 1. Weighting between Retail and CD
      Retail test performance in the NPR would drive a bank’s rating because the Retail Financing Test and the Retail Services and Products Test comprise 60 percent of a bank’s overall rating. Unless a bank achieves Outstanding retail performance, its CD performance is highly unlikely to affect the bank’s overall rating. A bank that is Satisfactory on retail is likely to receive an overall Satisfactory rating regardless of whether its CD performance is Outstanding or even Needs to Improve. ¹ If a bank cannot reasonably expect to achieve an Outstanding retail performance, CRA will provide little motivation for CD activity. Especially because CRA drives so much CD activity, such an outcome would be a major setback for the entire CD financing field. We urge the Agencies to weight retail and CD activity equally for large banks, so they are motivated to maximize performance on both. We note that, per the NPR, for intermediate banks, retail and CD would be weighted equally.
   2. Retail Lending Test performance thresholds
       The risk that CD could become incidental to an overall rating is greater because the Retail Lending Test’s Outstanding market metric performance threshold is so high (125 percent of industry-wide performance) that the Agencies estimate that, based on existing and available data, no bank with assets over $50 billion would achieve an Outstanding rating on the Retail Lending Test.² More broadly, we estimate that only 2 percent of the banking system’s assets would reside in banks projected to achieve Outstanding performance.³ Retail Lending Test performance is critical because it accounts for 75 percent of retail performance and 45 percent of overall performance. A bank would rationally determine that an overall Outstanding rating is beyond reach on this basis. It would then logically follow that the bank should seek only to avoid a Needs to Improve rating, a performance level well below the industry average. Should these dynamics take hold, they could easily feed a decline in lending. We take little solace in the fact that the Agencies would also publish numerical performance scores in addition to ratings. The headline rating will be what matters. This prospect would be a highly counter-productive result for communities. We therefore urge the Agencies to lower the Outstanding threshold to a point where it is more realistically achievable and therefore worth striving for.In addition, the High Satisfactory market metric performance threshold (110 percent of median industry performance) means that about 60 percent of banking assets will be in banks with Low Satisfactory (or lower) performance. Achieving even a High Satisfactory performance will be challenging. It is counter-intuitive that a bank should out-perform its competitors and still be Low Satisfactory. We urge the Agencies to consider lowering this threshold.It bears noting that these performance distributions are unlikely to change much, regardless of whether overall industry performance rises or falls, since the market performance benchmark moves along with the overall industry performance. If industry performance improves, it will become increasingly unsustainable for a bank to significantly out-perform the rising market benchmark and maintain even a High Satisfactory level of lending. This dynamic would change only if the community benchmarks become binding at some point.
   3. Differentiating between High Satisfactory and Low Satisfactory performance scores
       The NPR would assign 10 points for Outstanding performance, 7 for High Satisfactory, 6 for Low Satisfactory, and 3 for Needs to Improve. Because this system minimizes the difference between High Satisfactory and Low Satisfactory performance, it is unlikely to motivate High Satisfactory performance. Because perhaps 80 percent of all banks are likely to receive Satisfactory ratings, it is essential to differentiate performance within this vast majority. High Satisfactory performance should tell a bank it can reach the Outstanding level if it stretches, and a Low Satisfactory performance should concern a bank that it risks slipping to a Needs to Improve. Instead, a bank might reasonably infer that, under the proposed points system, whether a bank is in the 15^th percentile of industry performance or the 85^th percentile will be of little consequence. The risk to community reinvestment is that enough banks will aim for the former instead of the latter, and thereby pull overall industry performance down rather than up, resulting in lower performance standards. We urge the Agencies to widen the point spread substantially between High Satisfactory and Low Satisfactory performance.

Taken together, these issues as presented in the NPR pose a significant risk that reinvestment and its impact might not rise, as we hope, and they could well decline. The good news is that a few refinements could make a decisive difference.

 ¹ Two scenarios illustrate the point. First, a bank with retail performance at the midpoint of the High Satisfactory range (7.5 points) and CD performance at the midpoint of the Outstanding range (9.25 points) would have an overall Satisfactory rating (7.5 X 60% plus 9.25 X 40% = 8.2 points). Second, a bank with retail performance at Low Satisfactory midpoint (5.5 points) and CD performance at the Needs to Improve midpoint (3.0 points) would still receive an overall Satisfactory rating (5.5 X 60% plus 3.0 X 40% = 4.5 points). It is mathematically possible for a bank to achieve an overall Outstanding rating if it combines retail performance at the top end of the High Satisfactory range with CD performance at the top end of the Outstanding range, but such scenarios are highly unlikely.

 ² 87 Fed. Reg. at 33,954 (Table 9 to Section_.22)

³ Even so, we expect the market benchmarks to be more achievable – and therefore more applicable – than the community benchmarks in most cases.

We enthusiastically support consolidating all CD activities within a single CD test separate from retail activities. We also applaud the Agencies’ recognition and banks’ responsibility for CD activity at the institution level, including outside assessment areas (AAs), which we believe would bring more CD activity to more communities, as well as support for CDFIs.

In addition to increasing the weight of CD in ratings, as discussed above, several other changes are important for CD to reach its potential.

1. Eliminate the CD Services Test
   The NPR cites three main CD service activities, but two – financial literacy and small business technical assistance – properly belong on the Retail Services and Products Test because they serve consumers and small businesses, respectively. Service on nonprofit boards is a legitimate CD service but it does not warrant a separate test worth 10 percent of the entire rating. This is especially the case because CD Financing would be worth only 30 percent of the overall rating, a mere 5 percent more than the current Investment Test, which does not include CD lending. True CD services should be considered qualitatively on a single, consolidated CD Test.
2. Affordable housing
   As described in much more detail in our responses to Questions 3- 10, we appreciate the Agencies’ recognition of naturally occurring affordable housing (NOAH) and their setting a clear standard for qualifying it. However, we believe distinguishing between government supported housing and NOAH is misconstrued. Some government programs, including FHA’s flagship production program, Section 221(d)(4), have a stated purpose of affordability but do not require affordability and are accordingly often used for non-affordable housing. Instead, we propose a uniform performance-based approach for all affordable rental housing, except for Low Income Housing Tax Credit (LIHTC) properties. Key elements include: (1) rent affordability at 80 percent of area median income (AMI), or the HUD Fair Market Rent (FMR) standard in the few most unaffordable markets where it is higher; (2) at least one other qualifying factor chosen from a menu similar to the NPR’s proposal for NOAH; and (3) periodic confirmation that the standard is being maintained as a condition of continuing to confer CRA credit for financing provided in prior years. It would also be important to: (1) recognize LIHTC and other CD equity investments as a CD Impact Factor so that moving from today’s Investment Test structure to a new CD Test does not disrupt the investment market; and (2) limit the recognition of mortgage-backed securities to no more than 25 percent of a bank’s nationwide CD activity.
3. CD
   The Agencies should recognize financing for the construction and rehabilitation of ownership homes in LMI CTs and distressed and underserved middle-income nonmetro CTs.
4. Impact Factors
   We support the Impact Factor approach and hope it will weigh significantly on the CD Test. More specifically, Impact Factors should additionally recognize: (1) as already noted, CD equity investments; (2) activities serving low-income (as distinguished from moderate-income) geographies, a logical counterpart to the proposed Impact Factor for activities serving low-income individuals; (3) activities in rural communities, whether within or outside Metropolitan Statistical Areas (MSAs);⁴ and (4) activities that integrate decarbonization, such as energy efficiency, separate and apart from disaster preparedness and recovery. We also encourage qualitative consideration of CD activity features that are responsive to CD needs but are not reflected in the Impact Factors.
5. Public disclosure
   We urge the Agencies to publish detailed CD financing data. If county level disclosures would raise privacy concerns, then the MSA/nonmetro statewide and institution level data would still be valuable. Banks will report CD financing data anyway, so their publication would impose no additional burden.

⁴ The U.S. Census Bureau reports that 54 percent of rural residents lived in MSAs in 2016. https://www.census.gov/content/dam/Census/library/publications/2019/acs/ACS_rural_handbook_2019_ch01.pdf

In addition to reconsidering the market benchmarks as already discussed, the Agencies should:

1. Focus the Retail Lending Test on home mortgages and small business/farm loans
   Home mortgages and small business/farm loans are fundamental to CRA and should always be evaluated if a bank routinely makes them in a Facility-Based AA (FBAA) or nationwide. These are primary drivers of wealth creation and community stability.
2. Evaluate other retail lending products elsewhere on the exam
   We discuss specific challenges with auto lending, multifamily lending, and open-end mortgages in more detail at Questions 59-61 and 66. These other products are incidental to wealth building or present significant benchmarking issues, and their routine inclusion would add unnecessary complexity to a rule that is already very complex. We are also concerned that their inclusion in calculating the 15 percent major retail products threshold could, in some cases, cause small business lending to drop out of the evaluation. Auto loans and open-end home mortgages should be evaluated qualitatively on the Retail Services and Products Test. Multifamily loans should be evaluated on the CD test and only if they support affordable housing. If auto, multifamily, open-end home mortgages, or another retail lending product, such as small consumer loans, represents a plurality of its lending, a bank should have the option of pursuing a Strategic Plan.
3. Setting market performance benchmarks before the start of each performance year.
   The NPR would require a bank to meet market benchmarks it cannot know until after its performance year is over. Such unpredictability and opacity undermine the Agencies’ stated goals and are unhelpful to banks and the community. Instead, these benchmarks should be set based on the most recently available data before the start of each performance year. For example, the market performance data for year 1 would become available in year 2 and should be used to benchmark a bank’s performance in year 3. Our next point addresses a problem that could arise with this recommendation if changing interest rates shift the blend of home purchase mortgages and refinance mortgages, because the LMI share of purchase mortgages tends to exceed the LMI share of refinancing, as research shows has recently been the case.
4. Separate evaluation for home purchase and home refinancing mortgages
   The timing problem discussed just above can be addressed by separately evaluating home purchase mortgages and refinancing mortgages. Although adding a retail lending product means more complexity, this can be offset by limiting the primary products to mortgages and small business lending in most cases and by our next point.
5. Combine low-income and moderate-income mortgages; and small business/farm loan size categories
   Although low-income mortgage lending and loans to businesses/farms with revenues below $250,000 are important, these loan volumes are quite small relative to moderate-income mortgages and loans to small businesses/farms with revenues $250,000-$1 million, respectively. Keeping income and business size categories separate would not affect performance conclusions and ratings because the categories are combined as part of conclusion computations. Combining the categories would reduce by half the number of measures that banks must track and seek to achieve, a welcome reduction of the considerable complexity of the evaluation, especially for banks with numerous AAs. A more impactful way to recognize low-income mortgages and loans to small businesses/farms with revenues under $250,000 would be to provide positive qualitative consideration for them on the Retail Lending Test (also see 8. below). Finally, while low-income CT lending could shed indirect light on lending to racial and ethnic minorities, the NPR’s more direct plan to publish data on each bank’s minority lending would achieve that important objective more directly.
6. Community benchmarks and multipliers
   The community benchmark for home mortgage borrowers should be the LMI share of homeowners in the AA, not the LMI share of families in the AA. Home prices are 10 times higher in San Jose ($1.64 million) than in Toledo ($158,500),⁵ so the opportunities for LMI mortgage borrowers vary greatly among local markets. LMI homeownership would be a better benchmark, as the Agencies have proposed as the corresponding community benchmark for LMI geographies. With respect to community multipliers, we see no rationale for why the same multipliers should apply to different loan products. We already know that purchase mortgages are more LMI-rich than refinancing mortgages, so the multipliers for setting conclusion thresholds should differ. The community benchmark multipliers for small business/farm loans should also be considered on their own terms as well.
7. Weighting product performance to compute Retail Lending Test conclusions
   We are very concerned that weighting different loan products to compute Retail Lending Test conclusions based solely on dollar volume could greatly devalue small business/farm lending because the average home mortgage amount is so much larger. In many cases, small business lending would have only an incidental weight, effectively turning the Retail Lending Test into a de facto home mortgage test. The Agencies should explore a weighting system based on a combination of dollar volume and the number of loans for each product.
8. Qualitative factors
   The qualitative elements of retail lending should be considered on the Retail Lending Test rather than on the Retail Services and Products Test. Qualitative factors could include special purpose credit programs, small-balance loans, and second- review protocols for preliminarily rejected loan applications. Considering quantitative and qualitative elements together, along with performance context factors, would provide a more integrated and complete picture of a bank’s retail lending performance.

⁵ https://cdn.nar.realtor/sites/default/files/documents/metro-home-prices-q1-2022-ranked-median-single-family- 2022-05-03.pdf

We would prefer that all retail lending beyond FBAAs be considered as part of an Outside AA analysis. However, if the Agencies pursue RLAAs, we strongly recommend that:

1. A more robust materiality threshold should establish RLAAs.
   In addition to the proposed minimum loan count, lending should exceed (1) 1 percent of the area’s total lending for the applicable loan product; and (2) 0.5 percent of the bank’s total lending for the applicable loan product. A bank with less than 1 percent local market share is incidental to the community. In addition, if a bank makes less than 0.5 percent of its loans in an area, that lending will be immaterial to the bank’s CRA rating. This more robust standard would significantly reduce the number of RLAAs, focus where they matter most, and substantially mitigate the additional complexity that numerous RLAAs would add to CRA evaluation. We affirm that all Outside AA lending would still be evaluated, so reducing the number of RLAAs would not reduce a bank’s accountability.
2. RLAAs should apply only to a lending product if that product meets the materiality test
   For example, if a bank’s home mortgage lending meets the materiality test to establish an RLAA, but its small business lending does not, then the RLAA analysis would apply only to the bank’s home mortgages and not to its small business lending.
3. RLAAs should not be counted in determining whether at least 60 percent of a bank’s AAs have at least Satisfactory performance
   It is much harder for banks to penetrate LMI lending if they have no local branch presence. The NPR’s Table 12 indicates that banks would have less than Satisfactory ratings in 34 percent of RLAAs, far more than in FBAAs. We note that a bank must achieve at least Satisfactory performance for retail lending at the institution level in order to receive an overall Satisfactory rating so, again, the bank would still be accountable.

We urge the Agencies to set a clear, standard benchmark for AA performance of wholesale and limited purpose banks. These banks are important because they collectively have more than $1.5 trillion in assets and serve a nationwide market. Absent a clear AA benchmark, it will be impossible for these banks to plan and deliver appropriate AA performance. This ambiguity has been confusing and deleterious to CD financing for many years. We greatly appreciate the Agencies’ commitment to setting a similarly clear benchmark at the institution level. If the AA benchmark should prove inappropriate for any given bank because of special circumstances, the bank can explain its performance context or pursue a Strategic Plan.

Performance context – for both banks and the communities they serve – is important to understanding the many metrics embedded in the CRA exam. We urge the Agencies to fully integrate performance context into all performance conclusions and ratings. Performance context will be more important than ever as banks and the Agencies encounter unanticipated issues and unintended consequences during and after the transition to an entirely new framework. We also urge the Agencies to have and train a specialized corps of CRA examiners that can discern the true impact of a bank’s CD and retail activity. A rating evaluation should mean much more than a numerical formula.

We would be remiss not to express regret that CRA is unable to address racial equity more directly within the examination structure. NAAHL believes racial equity and justice are foundational to such core American values as equal economic opportunity and justice, as well as to our nation’s prosperity, social cohesion, and world-wide reputation. Racial equity and justice are also at the heart of NAAHL’s mission “to expand economic opportunity through the responsible financing of affordable housing and inclusive neighborhood revitalization.” As noted earlier, NAAHL member banks provided $183 billion in financing for LMI people and communities in 2020. However, we also acknowledge our industry’s failure to meet other responsibilities to communities of color. We have sometimes redlined neighborhoods and otherwise denied credit on fair terms. We have missed opportunities to develop and deploy the financial products that communities need. We have insufficiently engaged the power and agency of Black, Latino, and all people and communities that have suffered under systemic racism.

We regret that CRA – a civil rights law – is so limited in its capacity to further our highest ideals. Income is only a weak proxy for race and ethnicity. We understand that evidence of illegal discrimination can result in a CRA rating downgrade, but racial equity requires the kind of pro- activity that we hoped CRA could do more to encourage. Publishing HMDA – and, in the future, section 1071 small business lending data – by race and recognition of special purpose credit programs would offer welcome but only limited progress.

Pro-rata consideration should be limited only to affordable housing. Mixed-income housing is uniquely beneficial because it offers LMI people greater opportunities. In LMI neighborhoods, mixed-income affordable housing contributes to income diversity, which is important to the community’s economic and social stability and to sustaining retail and other essential services.

In middle- and upper-income areas, mixed-income housing offers LMI renters proximity to good schools and a wider range of jobs. In rural communities, mixed-income housing is needed to accommodate properties of sufficient scale to achieve development and operating efficiencies. And, in a wide range of communities, mixed-income housing is preferred by neighbors and supports the properties’ financial sustainability and risk management.

Other types of CD activities should receive CRA consideration only if a majority of the beneficiaries are LMI. Since about 30 percent of the national population is LMI, many activities would generally achieve about that degree of LMI benefit as a matter of course without any targeting or intentionality. Conferring CRA credit in these cases would dilute CRA’s consideration of CD activities that primarily benefit LMI people and places.

We believe that 51 percent is the right minimum percentage to achieve CRA credit. We see no justification for denying CRA credit if most of the beneficiaries are LMI. While we do not support partial consideration for non-housing CD activities, if the Agencies should decide otherwise, then a minimum threshold of 30 percent would be helpful.

We believe creating separate affordable housing standards based on the presence or absence of governmental support would be a mistake. Governmental support varies too widely to be a reliable proxy of LMI benefit. At the federal level, for example, HUD/FHA primary multifamily mortgage insurance program for new construction and rehabilitation, Section 221(d)(4), has a stated purpose of serving moderate-income renters,⁶ but, in fact, the program does not require either affordable rents or LMI occupancy and it is often accordingly used for middle- and upper- income housing. In addition, many states and localities support “affordable” housing but standards vary widely. For example, New York City’s 421-a tax exemption applies to properties affordable at 130 percent of AMI;⁷ many inclusionary zoning programs require only 5-10 percent of the property to be affordable; and the duration of affordability varies widely. On the other hand, some entirely private affordable housing initiatives, such as the Washington Housing Initiative,⁸ require both affordable rents and documentation of LMI occupancy.

We strongly urge the Agencies to establish a uniform yet flexible performance-based standard that would apply to all non-FLIHTC affordable multifamily housing, regardless of whether it has governmental support. This approach combines elements of the NPR’s proposals for governmentally supported housing and NOAH. Such a standard should apply a universal rent affordability standard and also meet one additional standard from a menu of options, with periodic re-confirmation of compliance.

1. All affordable properties should have rents affordable at 80 percent of AMI, based on (1) a 30-percent-of-income rent standard, and (2) the rents used for underwriting after any planned rehabilitation or construction.
   • • In the absence of substantial public subsidies, setting affordable rents at 60 percent of AMI, as the NPR proposes for NOAH, is too restrictive to be workable. Please see our response to Question 6 for more details.
     • To address needs in the least affordable local markets, a bank should be permitted to use HUD FMRs where they exceed the 30 percent of 80 percent of AMI standard. See our response to Question 4 for more details.
2. Properties should be required to meet any one of the following requirements in addition to affordable rents.
   • • Location in a LMI CT, as included in the NPR. This has been construed as an informal rule of thumb under current policy for many years.
     • Location in a CT where the median renter is LMI, as the NPR offers to consider. Because most renters in the CT are LMI and because the rents for the specific property are LMI affordable, there is a reasonable likelihood that most of the occupants will be LMI. See our response to Question 6 for more details.
     • Nonprofit or CDFI ownership or control (e.g., where a nonprofit is the general partner of a limited partnership). The NPR provides for nonprofit sponsorship. Nonprofits and CDFIs have proven their commitment to maintaining affordability over several decades. Some CDFIs are not nonprofits, but all CDFIs must meet the same public purpose test.
     • Documented LMI occupancy. It is impractical to require that most NOAH meet an LMI occupancy standard. Banks and other lenders do not have access to this information. Accordingly, most of the properties electing this option are likely to involve government programs. A tenant’s income would be established at the time they first occupy the property, and no re-certification of income would be required. For properties that were occupied before the bank finances it, the bank should be able to rely on a tenant’s prior income documentation or have perhaps 12-18 months after the financing is provided to document a tenant’s income.
     • Owner commitment to maintain affordability for at least five years, as provided in the NPR.
3. Periodic confirmation that a property is continuing to meet the above standards should be required for a bank to receive continuing credit for financings made in previous years. We strongly support the Agencies’ proposal to provide continuing credit for CD loans made in prior years. Continuing consideration will encourage the kind of long-term financing that is so important in many cases. Yet, in some cases there may be legitimate concerns that affordability today does not guarantee affordability tomorrow. Housing markets and communities change in ways that are not always easy to anticipate, and these changes vary across markets. Instead of setting a policy that could prove both too narrow and too broad, depending on local circumstances, the Agencies can set a common-sense performance-based standard. How long a bank can continue to count housing as affordable should depend on whether it remains affordable.Fortunately, affordability can be easily confirmed based on a rent roll, which most responsible lenders regularly collect as a normal business practice. Properties located in LMI CTs (2a above) or in a CT where the median renter is LMI (2b above) could continue to meet this standard as without question, assuming that rent affordability is maintained. In cases where documented LMI occupancy (2d above) is the second qualifier, it should be subject to periodic confirmation.
4. Activities would qualify for full credit if more than 50 percent of the units meet the affordable housing standards. Pro-rata credit should be available for properties where 20-50 percent of the units meet the affordable housing requirements. Mixed-income housing is important because it promotes mixed-income neighborhoods that can sustain important services and amenities, as well as long-term property financial sustainability. However, affordability may be incidental and CRA immaterial to financing decisions where fewer than 20 percent of the apartments are affordable. We note that the LIHTC, tax-exempt multifamily bonds, and HUD’s HOME Investment Partnerships program all require a minimum of 20 percent affordability.

⁶ https://www.hud.gov/program_offices/housing/mfh/progdesc/rentcoophsg221d3n4

⁷ https://www1.nyc.gov/site/hpd/services-and-information/tax-incentives-421-a.page

⁸ https://www.washingtonhousinginitiative.com/

Please see our answer to Question 3, which is more comprehensive and sets a context for our response to this question. In general, we do not support special consideration for government- supported affordable housing, except the important case of LIHTC because it drives virtually all privately financed new construction and substantial rehabilitation.

Regarding middle-income housing, we urge the Agencies to allow consideration for housing at rents up to HUD’s FMR standard in the relatively few, particularly unaffordable markets where the FMR exceeds 30 percent of 80 percent of AMI, including New York, Los Angeles, Miami, and San Francisco. It is in these markets that middle-income people are more likely to be renters (because home prices also tend to be unattainably high) and to face rent burdens. HUD’s FMRs follow a careful, time-tested methodology pegged to rents at the 40th percentile of the local rental market. FMRs are also a fully transparent, national governmental standard. In addition, the Small Area FMRs are tailored to reflect rents at the zip code level, so they could be an especially effective way for the Agencies to encourage affordable housing in high-opportunity areas, where rents tend to be relatively high.

Yes. We appreciate the Agency’s apparent concern about the possibility that housing that is affordable when financing is first provided could become unaffordable over time. See our response to Question 3, which provides a more comprehensive and contextualized response. In addition, we offer two specific points.

First, we support the NPR’s requirement to consider the rents used for underwriting. If rehabilitation or construction are involved, the post-completion rent should be used. For underwriting purposes, borrowers are motivated to propose the use of the highest possible rents to justify larger loans.

Second, to ensure continuing affordability, CRA consideration for prior-year financings should be conditioned on periodic documentation that affordability is being continued. We applaud the NPR’s proposal to confer CRA credit for outstanding CD loans made in prior years, as current policy already provides for CD investments. Continuing consideration will encourage the kind of long-term financing that is so important in many cases. In return, it is appropriate to require that affordable housing financing that receives continuing CRA credit is also continuing to achieve the CD purpose.

Fortunately, periodic documentation is readily practical. In most cases, a rent roll can document affordability. Most responsible rental housing lenders routinely obtain and review rent rolls in their normal course of prudent business practice. In cases where, per our response to Question 3, a property is qualified based on renter incomes, we recommend requiring income determinations only when a renter first occupies the property (or within 18 months of financing for already occupied properties if incomes were not previously determined). It is our experience that resident incomes seldom rise greatly over time relatively to the AMI, so requiring re-certification of incomes is unduly burdensome.

For context, please see our response to Question 3, which provides a more consistent yet flexible approach to non-LIHTC affordable rental housing.

We strongly urge the Agencies to apply an affordability standard for NOAH and other non-LIHTC housing based on 80 percent of AMI. We applaud the Agencies for proposing clear standards for qualifying NOAH because only about 20 percent of all affordable rental housing is directly government subsidized. However, with specific reference to NOAH, a 60 percent of AMI rent standard offers insufficient opportunity for debt financing. It would be counter-intuitive for the Agencies to set a higher rent limit (80 percent of AMI) for properties that can use public support to target rents more deeply, but a lower rent limit (60 percent of AMI) for NOAH properties that do not benefit from public support.

1. • In most markets, rents below 60 percent of AMI generate insufficient net operating income (i.e., after expenses) to support significant mortgage amounts. The problem is worse for smaller properties, where the total mortgage amount is simply too small to offset transactional expenses. The problem is also worse in markets with low AMIs. Many rural markets would be doubly disadvantaged because they tend to have both smaller properties and very low AMIs. We note that LIHTC properties – which do limit rents to 60 percent of AMI – are fundamentally different because: (1) the tax credits support equity investments that typically cover 50-75 percent of development costs without requiring investor returns from cash flow; (2) LIHTC properties tend to have significant scale; (3) a bank may achieve efficiencies by combining permanent mortgages with equity investments or construction loans; and (4) many LIHTC projects do not need permanent mortgages from banks because borrowers can obtain permanent loans through tax-exempt bonds, Fannie Mae, Freddie Mac, and FHA.

1. • Rents rose nearly 18 percent in 2021,⁹ far outpacing incomes, so the number of low-rent properties is dwindling and the number of rent burdened households with incomes 60- 80 percent of AMI is growing. But this spike only accelerated a pre-existing trend. According to the Harvard Joint Center for Housing Studies, “Before the pandemic, cost burdens were highest for lower-income renter households and rising fastest among middle-income renters. This trend continued in the first year of the pandemic. Cost burdens increased by nearly 5 percentage points, to 80 percent, for households making less than $30,000 (Figure 1). Renters making $30,000–45,000 saw their cost burden rate jump by an astounding 9 percentage points to 58 percent, while those making $45,000– 75,000 posted rates of 30 percent, a 5-percentage-point rise from 2019. Higher-income households maintained a relatively low burden rate of 8 percent, less than 2 percentage points above 2019 levels.”¹⁰Apartments with rents affordable at 60-80 percent of AMI are therefore becoming a more important source of affordable housing. CRA should encourage and recognize bank financing for this precious resource. In 2019 – before this most recent spike in rents – the number of rental units affordable at 60-80 percent of AMI was 50 percent greater than the number of units affordable at 50-60 percent of AMI.¹¹ (As a rule of thumb, apartments affordable below 50 percent of AMI do not generate sufficient net operating income to support mortgage loans.)

• • While the Agencies are correct that a substantial minority of rental units affordable at 60-80 percent of AMI are occupied by renters with incomes greater than 80 percent of AMI, we believe many of those are located in large, extremely unaffordable markets with high rates of rentership, including New York, Los Angeles, Miami, and San Francisco. In these markets, many middle-income households are cost-burdened and have significant needs for affordable rentals. In our response to Question 3, we recommend allowing rents up to HUD’s FMRs where they exceed the rent affordable at 80 percent of AMI. In contrast, in more affordable markets, it is less likely that higher- income renters will occupy apartments affordable at 60-80 percent of AMI because homeownership is more affordable and more common among these households.

• • To address a concern that properties affordable at 60-80 percent of AMI when a bank provides financing may subsequently become unaffordable, the property should remain affordable for a bank to continue receiving CRA credit. As described in our response to Question 3, we support the Agencies’ proposal to confer continuing credit for prior-year financings for as long as they remain outstanding. Documenting continued compliance (e.g., with rent rolls) is administratively feasible and it should provide assurance that CRA credit will not apply to unaffordable housing.

• • The affordable multifamily housing goals set for Fannie Mae and Freddie Mac are also informative because, like CRA, they primarily address mortgage lending activity, and do not assume availability of the public subsidies needed for deep income targeting. “There are three multifamily housing goals: a goal for the total number of units affordable to low-income families (income no greater than 80 percent of area median income), a goal for the total number of units affordable to very low-income families (income no greater than 50 percent of area median income), and a goal for the total number of units in small (5- to 50-unit) multifamily properties affordable to low-income families.” CRA should similarly set 80 percent as the basic standard; the proposed Impact Factors provide additional consideration for targeting at or below 50 percent of AMI.¹²

We also strongly support CRA credit for affordable housing located in a CT where the median renter is LMI. Affordable housing in these predominantly middle-income neighborhoods offers important opportunities for LMI renters to live closer to good schools, employment, and other community amenities. If most renters in the CT are LMI and if the rents for the specific property are LMI affordable, there is a reasonable likelihood that most occupants will be LMI. If, for some reason, a property stops being affordable, CRA consideration for it should end. These safeguards should protect the integrity of CRA policy.

⁹ https://www.politico.com/news/2022/03/18/housing-costs-inflation-00015808

¹⁰ https://www.jchs.harvard.edu/blog/affordability-gaps-widened-renters-first-year-pandemic

¹¹ https://www.huduser.gov/portal/sites/default/files/pdf/Worst-Case-Housing-Needs-2021.pdf

¹² https://www.fhfa.gov/PolicyProgramsResearch/Programs/AffordableHousing/Pages/Affordable-Housing- FNMandFRE.aspx

Permanent mortgages for single-family rental housing are covered as part of the Retail Financing Test, so they should not receive consideration as CD. Construction financing for single-family rentals should qualify as CD if it meets the criteria outlined in our response to Question 3.

Financing the construction or rehabilitation of owner-occupied homes (including condominiums or cooperatives) should receive CRA consideration if: (1) the homes are located in a LMI CT or a distressed or underserved middle-income nonmetropolitan CT; and (2) the sales price does not exceed four times the AMI. Financing the rehabilitation or reconstruction of an already owner- occupied home (so no sale is involved) should qualify if the owner is either LMI or middle- income. Please also see our response to Question 16.

We recommend that mortgage-backed securities (MBS), which are usually guaranteed by Fannie Mae, Freddie Mac, or Ginnie Mae, be treated carefully, since these MBS comprise the world’s largest, most liquid investment market (except for certain sovereign obligations) and should not crowd out CD activities that add more value.

• • Only the portion of the MBS attributable to CRA-qualified loans should be considered. Loans not meeting CRA eligibility should be disregarded to avoid over-stating their volume. Single-family loans within an MBS pool would be considered individually. Multifamily loans within an MBS would be treated consistent with CRA policy – i.e., the entire loan would qualify if the property is at least 51 percent LMI.

• • Banks should be required to hold MBS for which CRA consideration is claimed for at least two years, measured annually on a weighted portfolio basis. Applying the test on a portfolio basis would allow banks some flexibility while discouraging short-term holdings. In particular, this approach would discourage banks from purchasing MBS at the end of a year or exam period unless it has held other MBS for sufficiently longer periods to maintain the two-year average holding period.

• • At the institution level, not more than 25 percent of a bank’s CD activity should be credited for MBS (excluding CDFI-issued MBS, which do not benefit from a deep liquid market). It may be necessary for a bank to rely more heavily on MBS in any given AA, since sufficient CD opportunities may not be available in any given AA in any given year. However, MBS should not be a primary way for a bank to fulfill its overall CD financing responsibilities at the institution level.

MBS issued by a CDFI should be treated the same as any other CDFI loan or investment.

For rental housing, we believe the standards outlined in our response to Question 3 are sufficiently flexible to accommodate complex and novel solutions. Additional consideration should be awarded as a CD Impact Factor.

Limiting placed-based activities to government support would be misguided in two respects. First, it would miss important opportunities to motivate and recognize private activities that do not have affirmative government support. The needs of targeted CTs far exceed what governments can meaningfully support. We appreciate that the Agencies recognize the importance of nongovernmental affordable housing activities. The same principle should also apply to place-based activities. Second, determining meaningful governmental support will be extremely difficult and will invite the contrivance of nominal support to attain CRA qualification, which the Agencies would be compelled to assess. For example, precisely what would a government plan have to contain to qualify an activity? Would the plan have to name specific neighborhoods or projects? Could it be aspirational rather than operational, and what would qualify it as operational? Would the plans that more than 1,200 jurisdictions submit under HUD’s CD Block Grant program qualify? Would a zoning variance be sufficient to establish public support and how would that be equitable for localities that have different thresholds for requiring variances? We believe such issues will be impossible for banks to navigate and for the Agencies to implement effectively.

Determinations of LMI resident benefit will probably vary according to the activity. In some cases, a bright-line criterion may be elusive and judgment may be necessary. The pre-approval process and illustrative list process would be helpful here.

For example, most neighborhood-scale essential community facilities in targeted CTs (e.g., childcare and libraries) may be presumed to benefit LMI residents, as should community health centers and other facilities expressly intended to serve LMI people. Other facilities, however, such as hospitals and various kinds of essential community infrastructure (e.g., broadband, telecom, water supply, mass transit, and flood control), may be designed to serve a broader market area. In such cases, additional evidence should be provided to show that LMI people will primarily benefit.

The Agencies should clearly articulate that the financing of retail services, including grocery stores, pharmacies, and other neighborhood-scale services, qualify as essential community facilities, regardless of the size of the occupant business. For example, a chain supermarket brings the wide array of high quality, healthy, fresh, and affordable food that is common in middle- and upper-income communities but missing in so many LMI communities. These facilities bring convenience, jobs, physical revitalization, and lower prices for consumers.

Indeed, the presence of quality shopping at major-brand grocery stores, pharmacies, fitness gyms, and other retailers is often experienced by residents and perceived by outsiders as a marker for community stability and attractiveness. Yet, financing these facilities is often complicated. Uncertainty and perceived risk – especially if comparable facilities are not already present – makes CRA consideration important to their financing.

Financing the construction or rehabilitation of owner-occupied homes (including condominiums and cooperatives) should receive CRA consideration if: (1) the homes are located in a LMI CT or a distressed or underserved middle-income nonmetropolitan CT; and (2) the sales price does not exceed four times the AMI. Financing the rehabilitation or reconstruction of an already owner-occupied home (where no sale is involved) should qualify if the owner is either LMI or middle-income.

The NPR would provide CRA credit on the Retail Lending Test for all home mortgages in LMI CTs, without regard to the owner’s income. Indeed, the Urban Institute has found that 60 percent of banks’ mortgages in LMI CTs went to middle- and upper-income borrowers. ¹³ Under the affordable housing definition, however, the NPR would apparently confer credit for home construction or rehabilitation financing only if the occupant is LMI, and without regard to the neighborhood’s median income. We are deeply concerned this latter approach fails to recognize the importance of home construction and rehabilitation to community stabilization and revitalization. Single-family homes comprise the primary land use in most LMI CTs, but many or most existing homes in these neighborhoods are old or in need of improvement, and empty lots (sometimes where dilapidated homes were demolished) are common. These communities typically have relatively low rates of homeownership and little chance of attracting or retaining homeowners unless quality homes can be built or rehabilitated. While many of these prospective homeowners may be middle-income, not LMI, they are important to sustaining the diversity of incomes that neighborhoods need to support retail activity and community institutions ranging from youth sports leagues to churches. In a rural context, we often hear that it is hard to keep or attract growing businesses because quality affordable homes are simply not available. Revitalizing both urban and rural communities is very difficult unless these problems can be addressed. CRA is needed and well justified to support the construction and rehabilitation of owner-occupied homes.

To avoid providing CRA credit for constructing or rehabilitating expensive homes that could contribute to gentrification and displacement, we recommend limiting CRA credit to homes that sold for a price not exceeding four times the AMI. This limitation would ensure that the homes are broadly affordable to middle-income homebuyers. In cases where an already owner- occupied home is being rehabilitated, the owner should be either LMI or middle-income. This approach is also consistent with requirements of the proposed Neighborhood Homes Investment Act, bipartisan legislation Congress is currently considering with the sponsorship of 24 Senators and 86 Representatives.

¹³ https://www.urban.org/research/publication/community-reinvestment-act-what-do-we-know-and-what-do-we- need-know

As noted in our response to Question 15, determinations of LMI resident benefit will vary according to the activity. In some cases, a bright-line criterion may be elusive and judgment may be necessary. The CD activity pre-approval process and illustrative list process would be helpful here.

For example, most neighborhood-scale essential community facilities in targeted CTs (e.g., childcare and libraries) may be presumed to benefit LMI residents, as should community health centers and other facilities expressly intended to serve LMI people. Other facilities, however, such as hospitals and various kinds of essential community infrastructure (e.g., broadband, telecom, water supply, mass transit, flood control), may be designed to serve a broader market area. In such cases, additional evidence should be provided to show that LMI people will primarily benefit.

This question is important because we are aware of cases where LMI people and communities have not received their fair share of disaster recovery resources. As a practical matter, residency in a LMI CT may be a proxy.

Rather than just relying on recovery and resiliency alone, the Agencies should also prioritize reducing carbon emissions and slow the rate of global warming as a step toward reducing the frequency and impact of climate related disasters. To align with national and international climate goals, we recommend focusing on three core priorities:

• • Mitigation – slow the rate of global warming through meaningful carbon reduction strategies.

• • Adaptation – take steps to live with the effects of climate change.

• • Resilience – become better able to withstand the effects of climate change and natural disasters.

Climate resiliency should certainly include energy efficiency but should also be broadened to include other decarbonization activities. For example, installing solar panels may not technically increase “energy efficiency” – i.e., the property may consume the same amount of energy – but it would promote decarbonization. Replacing energy-inefficient windows or water heaters would contribute to both energy efficiency and decarbonization.

• • Activities that finance stand-alone improvements but not an entire property – e.g., replacing energy inefficient windows or adding solar panels – should receive consideration under the applicable CD purpose (e.g., affordable housing or essential community facilities).

• • An Impact Factor should be established to recognize decarbonization features of otherwise qualified CD activities. For example, affordable housing may or may not involve decarbonization; but if it does, then it should receive additional consideration as an Impact Factor. We recommend including nationally recognized, third-party verified certification standards in CRA criteria to encourage reductions in greenhouse gas (GHG) emissions and energy consumption. These standards include LEED Zero, DOE Zero Energy Ready Homes and Living Building Challenge – Zero Energy. Net zero buildings with highly efficient building envelopes and systems allow for efficient use of energy to run emergency services and allow residents to comfortably shelter in place during major weather events making them naturally more resilient. Investments in mitigation and adaptation strategies and standards stretch recovery dollars further and preserve quality housing for residents for decades to come.

Yes. Please see our response to Question 20.

Such activities should qualify only if LMI individuals or targeted CTs will be the primary beneficiaries.

Such activities should qualify only if LMI individuals and targeted CTs will be the primary beneficiaries.

Such activities should qualify only if LMI individuals and targeted CTs will be the primary beneficiaries.

Activities with MDIs, WDIs, LICUs, and CDFIs

We strongly support CRA credit for banks that partner with CDFIs, MDIs, WDIs, and LICUs because they have a long record of reaching underserved people and places.

Applying this same reasoning, we also support extending this status to partnerships between banks and nonprofit organizations that hold a charter from NeighborWorks America.

NeighborWorks is a Congressionally chartered organization, and membership in the network for these mission-driven organizations requires rigorous financial and management assessments prior to receiving their charters and on an ongoing basis thereafter. Furthermore, membership in the NeighborWorks network is only available to organizations that demonstrate a commitment to resident leadership, ensuring that the organization continues to represent the interests of the communities in which it works. The accountability and oversight that NeighborWorks America provides to network organizations is akin to the stewardship of Treasury for certified CDFIs, ensuring that NeighborWorks organizations (NWOs) maintain their physical and financial health as well as their mission-driven focus. Activities relating to partnerships with NWOs, including loans and grants, should be explicitly included in the regulations describing qualified activities. Inclusion of this provision would strengthen community-based organizations’ ability to attract investment from financial institutions by providing the clarity and certainty that such investments would receive CRA consideration.

We also urge activity with a CDFI’s or NWO’s wholly owned or controlled subsidiary be treated as support for the CDFI or NWO itself. There are many good business reasons why CDFIs and NWOs establish subsidiaries. As just one example, CDFIs and NWOs often create limited partnerships they control as the general partner for specific projects or initiatives.

We urge the Agencies to establish a process for timely consideration – within 30 days of the request – so that banks and their partners can proceed with confidence. Community development projects are often complicated, and undue delay presents costly obstacles.

We urge the Agencies to create Impact Factors to address two critical CD needs beyond the scope of questions asked.

1. Equity investments. Although it does not address a specific question, the Agencies should recognize CD equity investments (including equity equivalent investments in CDFIs) as a CD Impact Factor. Equity investments are critically important to CD. For example, LIHTC equity investments are the primary driver of affordable rental housing production and comprise 50-75 percent of project financing. New Markets Tax Credit (NMTC) equity investments are likewise instrumental to many economic development activities; equity investments in an emerging group of affordable housing preservation funds are responsive to an important challenge; and equity-equivalent investments are an important source of flexible, enterprise-level capital for CDFIs. Banks are the primary source of CD equity investments, motivated by CRA. For example, banks provided 85 percent of all LIHTC investments in 2021, a level consistent with prior years.¹⁴

However, there is a significant risk that banks’ equity investments could diminish under a new CRA rule that does not recognize their unique value as an Impact Factor. A major proposed change in the exam structure is to replace the current Investment Test with a new CD Financing Test that combines loans and investments. While we support this change, banks may be motivated to shift their CD activities from equity investing to lending because equity investments: (1) require banks to hold more capital; (2) are less senior in the capital stack; (3) are less liquid; (4) are subject to the possibility that some banks may at some point have less capacity to use tax benefits associated with LIHTC and NMTC, as occurred during the Great Recession and could occur under a proposed global corporate minimum tax regime; (5) are a specialty product outside the range of commercial financing that banks routinely offer and require banks to have special in- house expertise; and (6) involve long-term investments, especially in the case of LIHTC.

We are not suggesting that banks would withdraw entirely from the LIHTC or NMTC market. However, among the 24 banks responding to a survey (sponsored by the Affordable Housing Investors Council, Affordable Housing Tax Credit Coalition, and NAAHL), 42 percent said the CRA proposal as currently drafted would incline them to reduce their LIHTC investment activity. Even a modest pull-back would have significant negative investment pricing effects that would: (1) make many projects infeasible; (2) create financing gaps that pull scarce public funds from other projects; or (3) require projects to reduce deep income targeting and forego other important features and amenities. Recognizing the value of equity investments as a CD Impact Factor is favored by 83 percent of the surveyed banks to ensure that there is no disruption to the LIHTC market.

1. Decarbonization. An Impact Factor should be established to recognize decarbonization features of otherwise qualified CD activities. This recommendation complements, but is separate, from disaster resilience. For example, affordable housing may or may not involve decarbonization; but if it does, then it should receive consideration as an Impact Factor. We recommend including nationally recognized, third-party verified certification standards to encourage reductions in greenhouse gas (GHG) emissions and energy consumption. These standards include LEED Zero, DOE Zero Energy Ready Homes and Living Building Challenge – Zero Energy. Net zero buildings with highly efficient building envelopes and systems allow for efficient use of energy to run emergency services and allow residents to comfortably shelter in place during major weather events making them naturally more resilient. Investments in mitigation and adaptation strategies and standards stretch recovery dollars further and preserve quality housing for residents for decades to come.

¹⁴ https://www.cohnreznick.com/-/media/resources/2022_housing-tax-monitor_march_2022.pdf

The most consistent and equitable approach would be to recognize activities serving low-income (as distinguished from moderate-income) CTs. This approach would be consistent with the proposed Impact Factor for activities serving low-income individuals. As with individuals, low-income CTs are far less numerous than their moderate-income counterparts, their needs are more pressing, and meeting these needs is more challenging.

We also prefer an income-based measure to a poverty-based measure because the former is more equitable. Low-income is set relative to the median income of each area, so every MSA and nonmetro statewide area should have an equitable share of CTs presenting reinvestment opportunities. However, because poverty is a national standard, areas with lower AMIs will have greater shares of high-poverty CTs than areas with higher AMIs. Because the cost of living varies similarly, the same dollar goes a lot farther in most low-AMI areas than in most high-AMI areas.

1. 1. We also recommend that activities in all rural areas (as distinguished from nonmetro areas) qualify as an Impact Factor. The disadvantages and challenges facing rural areas are well known, including the lower capacity of most rural governments, lower income levels, limited infrastructure, and the difficulty of financing the small-scale properties that many rural areas need and can support. Perhaps less appreciated is that

“nonmetro” is an inadequate proxy for “rural.” As the U.S. Census Bureau explains,

“Many counties classified as metropolitan include rural territory, while many counties outside of metropolitan and micropolitan statistical areas contain urban clusters. Based on the Census Bureau’s urban/rural classification, 54 percent of the rural population in 2016 resided within metropolitan counties. Rural areas within metropolitan counties encompass a wide variety of landscapes and settlement patterns—from sparsely populated desert lands within large metropolitan counties in the Southwest to small- town landscapes and “large-lot” (one-, three-, or five-acre) housing subdivisions on the

fringes of large metropolitan statistical areas.”¹⁵ For the purpose of setting an Impact Factor, we recommend the Federal Housing Finance Agency’s definition of “rural area” under its Duty to Serve regulation. ¹⁶

¹⁵ https://www.census.gov/content/dam/Census/library/publications/2019/acs/ACS_rural_handbook_2019_ch01.pdf

¹⁶ https://www.fhfa.gov/DataTools/Downloads/Pages/Duty-to-Serve-Data.aspx

Short-term deposits should not qualify for an Impact Factor. However, we see little reason to exclude other activities. As described above, we strongly urge the Agencies to establish a separate Impact Factor for equity investments. We also note with appreciation that the NPR would already reward long-term loans by conferring continuing consideration for CD loans made in prior years. Finally, we also appreciate the separate Impact Factor for grants to nonprofit CD organizations.

The Agencies should also add NeighborWorks network organizations (NWOs) to the list of favored bank partners. NeighborWorks is a Congressionally chartered organization, and membership in the network for these mission-driven organizations requires rigorous financial and management assessments prior to receiving their charters and on an ongoing basis thereafter. Furthermore, membership in the NeighborWorks network is only available to organizations that demonstrate a commitment to resident leadership, ensuring that the organization continues to represent the interests of the communities in which it works. The accountability and oversight that NeighborWorks America provides to network organizations is akin to the stewardship of Treasury for certified CDFIs, ensuring that NWOs maintain their physical and financial health as well as their mission-driven focus. Activities relating to partnerships with NWOs, including loans and grants, should be explicitly included in the regulations describing qualified activities. Inclusion of this provision would strengthen community-based organizations’ ability to attract investment from financial institutions by providing the clarity and certainty that such investments would receive CRA consideration.

We also urge activity with a CDFI’s or NWO’s wholly owned or controlled subsidiary be treated as support for the CDFI or NWO itself. There are many good business reasons why CDFIs and NWOs establish subsidiaries. As just one example, CDFIs and NWOs often create limited partnerships they control as the general partner for specific projects or initiatives.

We believe $250,000 is an appropriate threshold. It would also be consistent with the level proposed for the Retail Lending Test.

The FHFA definition is workable because it is limited to CTs with low poverty. Although, as discussed above, we generally prefer to avoid eligibility criteria based on the poverty rate criteria because it is a national standard that does not account varying AMIs, in this case, consistency with the FHFA standard justifies its use.

A bank should have the option to designate the portion of the MSA that excludes the FBAA as a new RLAA.

RLAA evaluations should apply only to the specific product – closed-end home mortgages or small business loans – for which the bank exceeds the required volume.

We would prefer that all retail lending beyond FBAAs be considered as part of an Outside AA analysis. Our reason is embedded in the final words of the Agencies’ question: areas where banks have no branches are not “its local communities,” but instead are outside them.

Nevertheless, if the Agencies move forward with RLAAs, we strongly recommend that:

1. 1. A more robust materiality threshold should be used to establish RLAAs. In addition to the proposed minimum loan count in a RLAA, lending should exceed both (1) 1 percent of the area’s total lending for the applicable loan product; and (2) 0.5 percent of the bank’s total lending for the applicable loan product. A bank with less than 1 percent local market share is not material to the community. Similarly, performance in an area will be immaterial to a bank’s CRA rating if the bank makes less than 0.5 percent of its loans in that area. This more robust standard would significantly reduce the number of RLAAs, focusing RLAAs where they matter most and substantially mitigating the additional complexity that numerous RLAAs would add to a bank’s management of its CRA responsibilities. Based on unpublished research conducted by Urban Institute, this more robust materiality test would reduce the total number of RLAAs for mortgage lending from 654 to 102 (84 percent) and the most RLAAs for any single bank from 121 to 20 (83 percent); and for small business lending from 826 to 121 (85 percent) and the most for any single bank from 233 to 41 (82 percent). Our approach would help reduce the considerable complexity of managing CRA compliance. We reiterate that all Outside AA lending would still be evaluated, so reducing the number of RLAAs would not reduce any bank’s responsibility.
   2. RLAAs should apply only to a lending product if that product meets the materiality test. For example, if a bank’s home mortgage lending meets the materiality test but its small business lending does not, then the RLAA analysis should apply only to the bank’s home mortgages and not to its small business lending.
   3. RLAAs should not be counted in determining whether at least 60 percent of a bank’s AAs have at least Satisfactory performance. It is much harder for banks to penetrate LMI lending if they have no local physical presence. The NPR’s Table 12 indicates that banks would have less than Satisfactory ratings in 34 percent of RLAAs, far more than in AAs more generally. We note that a bank must achieve at least Satisfactory performance for retail lending at the institution level in order to receive an overall Satisfactory rating so, again, the bank would still be accountable.

If the Agencies move forward with RLAAs, all large banks should be evaluated for mortgage or small business only where lending meets the materiality standard described in our response to Question 45. This materiality standard would screen out lending that is insignificant to the community or the bank. Additional exemptions would become unnecessary.

We greatly appreciate the Agencies’ proposal and strongly discourage any restriction or limitation on consideration for CD financing outside FBAAs. CRA’s current geographic restriction has seriously impeded the flow of CD capital to the places that need it most. The clearest evidence relates to LIHTC. Because banks provide 85 percent of all LIHTC investment, CRA policy significantly affects our nation’s ability to address the growing affordable housing crisis. LIHTC investment pricing, which determines the amount of equity invested, can vary dramatically, by $0.20 for each $1.00 of LIHTC, between what are commonly known as CRA hot markets and CRA deserts.¹⁷ In other words, properties with the least CRA demand receive 20 percent less equity for the same amount of LIHTCs as properties with the highest CRA demand, rendering many properties with low pricing financially infeasible. The Agencies’ proposal to remove geographic limitations for CRA consideration would mitigate this long-standing problem. Based on a survey jointly undertaken by the Affordable Housing Investors Council, the Affordable Housing Tax Credit Coalition, and NAAHL, two-thirds of the 24 respondent banks said the NPR’s new geographic flexibility would make them likely to make LIHTC investments in underserved CRA markets outside their current AAs.

We believe that a similar dynamic would apply to other forms of CD financing.

Finally, the CD Impact Factors would incent banks to focus their CD activity on underserved and other high-priority communities, so any restriction of consideration for activity outside FBAAs would be unnecessary as well as counter-productive.

¹⁷ https://www.housingonline.com/wp-content/uploads/2022/07/2022_Housing-Tax-Monitor_August_final.pdf

Yes. Communities need all the CD reinvestment they can attract.

CD Financing by a Consortium or a Third Party

We support the Agencies’ proposal to retain the current flexibility with respect to consideration for CD loans and investments by a consortium in which the bank participates or by a third party in which the bank has invested, including CDFIs and regional and nationwide LIHTC and other CD financing funds. For these purposes, the same activity could not be allocated to more than one bank and no bank may claim more than its percentage share of the sponsor’s total activity.

However, additional clarity is needed.

First, the Agencies should clarify that banks may rely on geographic allocations provided by the CDFI, consortium, or fund sponsor, such as through side letters, as has been common practice.

Second, it is common for a bank to provide financing that the recipient uses for its general purposes, rather than for passing through to specific activities. For example, a bank may provide working capital, a grant, an equity investment, or a CDFI equity-equivalent investment to support the recipient’s operations. Similarly, a bank may provide a line of credit that will be used to fund future activities that have not yet been identified. Such support is often critically important to the recipient’s effectiveness and impact. To accommodate and encourage this financing, the Agencies should clarify that the CDFI or other recipient may assign a geographic benefit on any reasonable basis, such as where it is located, has historically worked, or intends to work.

We believe that performance context information is very important to CRA and urge that examiners consider it seriously. Communities and banks differ widely. Understanding their performance context is essential to evaluating a bank’s responsiveness to community needs.

Home purchase mortgages and home refinance mortgages should be evaluated separately. These loan types serve different and important purposes. Combining them would provide less visibility into banks’ service to address LMI needs. In addition, these two loan types operate differently in high-priced and low-priced markets. In higher-priced markets, LMI home purchase opportunities will tend to be relatively limited, but LMI refinancing may make more sense for current LMI homeowners. In lower-priced markets, LMI home purchases tend to be more achievable but refinancing of smaller balance mortgages may not produce sufficient monthly payment reductions.

Fluctuating interest rates make refinancing significantly more or less voluminous relative to home purchases, as the last several years have made abundantly clear. This connects to a related issue. The market metric would compare a bank’s home mortgage performance to the industry’s performance – a benchmark impossible for banks to know before the year for which their lending is being evaluated, or even during the evaluation year. This timing problem undermines the transparency and predictability that the Agencies have sought to bring to CRA modernization.

Our remedy is to use the most recently available year of market data to set the benchmark for the activity year being evaluated. For example, year 1 data would become available in year 2 and should be used to benchmark year 3 performance. This approach would enable banks to understand their own performance on an ongoing basis and adjust strategies in time to implement them for evaluation. However, fixing this timing problem requires that home purchase mortgages and home refinancing be evaluated separately. Otherwise, the combination of changing interest rates with the different LMI mixes of purchase and refinance loans would distort the analysis.

Home improvement and other purpose closed-end home mortgage loans should be evaluated qualitatively under the Retail Services and Products Test. It is highly unlikely that such loans would meet the 15 percent major product lines threshold. Moreover, we believe the Retail Lending Test should limit its focus to home purchase, home refinance, small business, and (where significant) small farm loans.

Closed-end non-owner-occupied housing lending should be included in the closed-end home mortgage loan product category. These loans are typically part of lenders’ home mortgage retail product lines. The borrowers tend to be small-scale owners.

Open-end home mortgage loans should be evaluated qualitatively under the Retail Services and Products Test. It is highly unlikely that such loans would meet the 15 percent major product lines threshold. Moreover, we believe the Retail Lending Test should limit its focus to home purchase, home refinance, small business, and (where significant) small farm loans.

Multifamily lending should be evaluated only as part of CD, for several reasons.

• • Multifamily loans are commercial real estate loans, not retail loans. The two business lines are entirely different.

• • As the Agencies suggest, an inference that all multifamily loans in LMI CTs should always be viewed favorably and all such loans outside LMI CTs unfavorably, as the proposed metric implies, would be misguided on both counts. Some multifamily loans in LMI CTs finance high-rent properties that may contribute to gentrification and even displacement of LMI CT residents; and many multifamily loans outside LMI CTs provide beneficial affordable housing opportunities for LMI renters in middle- and even upper- income communities offering good schools and proximity to jobs.

• • HMDA data are too limited to support a reliable metric. The unit-count categories are too broad for meaningful use and using loan counts will defeat sound analysis of loans for properties of very different sizes.

• • Multifamily lending for most banks would not exceed the 15 percent major product lines test and, in any event, we believe the Retail Lending Test should limit its focus to home purchase, home refinance, small business, and (where significant) small farm loans.

• • CRA’s primary focus for multifamily lending should be on affordability. Affordable housing is rightly considered on a CD test.

It may well make sense for banks that are primarily multifamily lenders to be designated as limited purpose banks, except that such banks should also have a Retail Services and Products test if they operate multiple branches that take deposits from and otherwise serve the general public in their regular course of business. A Strategic Plan should also be an available option.

We generally agree that Section 1071 rules, including the size standards, should be the basis for small business and small farm lending. Section 1071 does not expressly reference purchased loans, and it is not yet clear whether the final rule for Section 1071 will include them. If it does not, the Agencies should give banks the option to report purchased loans for inclusion in the bank’s performance evaluation. In any case, the Agencies should provide an additional transition year to align CRA and Section 1071 reporting.

Current definitions for small business and small farm loans should sunset at the end of the calendar year after the new definitions take effect.

The Agencies should evaluate auto lending on a qualitative basis on the Retail Services and Products Test, along with other consumer loans.

• • Auto lending does not build wealth and it sometimes depletes wealth. Autos lose value from the moment of purchase, and sometimes faster than the outstanding loan balance. Moreover, auto lending is already plentiful.

• • Including auto lending on the Retail Lending Test would dilute CRA’s more important retail lending focus on home mortgage and small business/farm loans, which are crucial wealth-building tools. This could occur in two ways. In the context of the 15 percent major retail lending products test, the inclusion of auto lending could reduce small business/farm or mortgage lending below the threshold. Second, when retail lending performance combines the product lines, auto lending would dilute the weighting of small business/farm and mortgage lending.

• • A perverse consequence of including LMI auto lending would be to compel banks to aggressively seek to make subprime and otherwise problematic loans to compete against other lenders that operate in the LMI segment. These loans are not per se abusive, but neither do they tend to be consumer-friendly. For example, 84-month loans have become increasingly common because they are effective at reducing monthly payments, a critical selling point for consumers. However, Experian warns that these loans can be a “financial nightmare,” requiring consumers to pay thousands of dollars more in interest over the life of the loan, leaving consumers with little positive equity or even negative equity, and outlasting most warranties. As Experian says, “the tradeoff of lower monthly payments is rarely worth the risk of owing more than your car is worth, being tied to endless car payments or spending more than you can really afford.”¹⁸

• • Banks make auto loans both directly and indirectly through dealers. Banks have no control over the LMI share of these indirect loans. Some banks might withdraw from working with dealers of higher-priced brands because they make few LMI loans, unintentionally distorting the market.

• • A reliable market metric would be impossible to set. There is no comprehensive market- wide data regarding the income or neighborhood residence of auto loan borrowers. Banks provide only 31 percent of auto loans,¹⁹ and the Agencies propose to collect data only from those banks with assets exceeding $10 billion.

• • Even a community benchmark could present challenges in some markets with good public transit, such as New York, Chicago, and Boston, where LMI auto ownership is probably low.

Data collection and reporting for auto lending represent an unjustified administrative burden on banks. We affirmatively support requiring banks to collect and report more data on CD activity and the location of customer deposits because we believe their importance justifies the additional resources required of banks. However, we cannot say the same for collecting auto lending data, except perhaps for the few banks for which auto lending is their primary retail lending product.

¹⁸ https://www.experian.com/blogs/ask-experian/are-84-month-auto-loans-a-good-idea/

¹⁹ Experian, State of the Automotive Finance Market Q1 2022. https://www.experian.com/automotive/auto-credit- webinar-form

No. Access to consumer credit cards is plentiful. Focusing CRA on consumer credit cards would distract from more important wealth-building credit products like home mortgages and small business/farm loans. Including consumer credit card lending could incent banks to make more subprime credit cards and cards with terms unfavorable to consumers.

For some banks, consumer loans, including small dollar loans, comprise a plurality of the bank’s retail lending. These banks should have the option of having these loans considered both quantitatively – against a community benchmark – and qualitatively.

Instead of the 15 percent major product line approach, we urge the Agencies to evaluate home purchase loans, closed-end home refinance loans, small business loans, and small farm loans in every FBAA where a bank routinely makes more than a minimal number of such loans. These loan products should be the principal focus of CRA because they are the most critical to building individual and community wealth. Open-end home mortgages should be evaluated only qualitatively. Multifamily loans should be evaluated only on the CD test, and there only if they support affordable housing.

Anecdotal pre-testing indicates cases where including all six product lines would crowd out the consideration of small business loans.

As we explain in our response to Question 60, multifamily lending should be evaluated only on the CD test, and there only with regard to affordable housing. Per our response to Question 61, monoline or predominately multifamily lenders should, in some cases, be evaluated as limited purpose banks or have a Strategic Plan.

We urge the Agencies to avoid unnecessarily duplicative requirements under CRA, especially since the NPR would establish a much more complex and detailed evaluation structure. Accordingly, instead of the proposed retail lending screen, the Agencies should apply the Riegle-Neal Act’s section 109 host-state loan-to-deposit ratio, including to banks with branches within a single state.

Section 109 is very similar to the retail lending screen, albeit at the state level, not the AA level.

“To determine compliance with section 109, the appropriate agency first compares a bank’s estimated statewide loan-to-deposit ratio to the estimated host state loan-to- deposit ratio for a particular state. If the bank’s statewide loan-to-deposit ratio is at least one-half of the published host state loan-to-deposit ratio, the bank has complied with section 109. A second step is conducted if a bank’s estimated statewide loan-to- deposit ratio is less than one-half of the published ratio for that state or if data are not available at the bank to conduct the first step. The second step requires the appropriate agency to determine whether the bank is reasonably helping to meet the credit needs of the communities served by the bank’s interstate branches. A bank that fails both steps is in violation of section 109 and subject to sanctions by the appropriate agency.”²⁰

We note that host state loan-to-deposit ratios generally range from 61 percent to 97 percent, except for Delaware (51 percent), which has many limited purpose banks, Puerto Rico (49 percent), and the Virgin Islands (45 percent). Some special adjustment could be applied in these states.

Banks are already familiar with these standards, so using them in a CRA context would be a straight-forward simplification of the Agencies’ proposal. In any case, such an analysis should apply only to states that include a FBAA, since banks have no branches outside FBAAs and lending and deposit-taking are unrelated outside FBAAs. As the Agencies have proposed for the retail lending screen, if a bank that fails a section 109 review, then a review of specific performance context factors should determine whether there is an acceptable basis for such failure.

Bank Geographic Distribution Metrics and Borrower Distribution Metrics

The Agencies should combine the low-income and moderate-income mortgages and combine the small business/farm loan size categories. Although low-income mortgage lending and loans to businesses/farms with revenues below $250,000 are important, these loan volumes are quite small relative to moderate-income mortgages and loans to small businesses/farms with revenues $250,000-$1 million, respectively. Maintaining separate income and business size categories would not affect performance conclusions and ratings because the categories are combined as part of conclusion computations. Combining the categories would reduce by half the number of measures that banks must track and seek to achieve, a welcome reduction of the considerable complexity of the evaluation, especially for banks with numerous AAs. A more impactful way to recognize low-income mortgages and loans to small businesses/farms with revenues under $250,000 would be to provide positive qualitative consideration for them on the Retail Lending Test. Finally, while low-income CT lending could shed indirect light on lending to racial and ethnic minorities, it would also reflect lending to Whites. The NPR’s plan to publish data on each bank’s minority lending would achieve that important objective more directly.

²⁰ https://www.fdic.gov/news/press-releases/2021/pr21047a.pdf

Because lending will be assessed annually, the benchmarks should be set annually, based on the most recent year for which they are available. For example, data from year 1 would be available in year 2 and used to set benchmarks for year 3. This approach would solve the significant transparency and predictability problem under the NPR that banks cannot know at the start of any year what they will be expected to achieve in that year. This approach would also avoid applying different benchmarks to different banks, depending on the timing of their evaluation periods. The Agencies have sometimes changed the end date of exam periods long after they have begun, creating a possibility that setting benchmarks in the last year of an exam period would hold banks accountable for hitting lending targets set several years after the fact.

These metrics should average a bank’s annual performance over the exam period. Please also see our response to Question 76.

The community benchmark for home mortgage borrowers should be the LMI share of homeowners in the AA, not the LMI share of families in the AA. Home prices are 10 times higher in San Jose ($1.64 million) than in Toledo ($158,500),21 so the opportunities for LMI borrowers vary greatly among local markets. LMI homeownership would be a much better (if still imperfect) benchmark for LMI borrowers, just as the Agencies have proposed as the corresponding community benchmark for LMI geographies.

²¹ https://cdn.nar.realtor/sites/default/files/documents/metro-home-prices-q1-2022-ranked-median-single-family-2022-05-03.pdf  

We do not believe a reliable market benchmark for auto lending can be established. Limiting data collection to banks with assets exceeding $10 billion is only part of the problem. In addition, banks provide only 31 percent of all auto lending22 and many auto loans are made indirectly, though dealers. Indirect lending is largely beyond banks’ control; and the degree of indirect LMI auto lending depends on whether the dealer is selling higher-priced or lower-priced vehicles.

²² Experian, State of the Automotive Finance Market Q1 2022. https://www.experian.com/automotive/auto-credit-webinar-form

We are highly doubtful that the same community metric multiplier should apply to different products and metrics. For example, since closed-end mortgages and open-end mortgages tend to serve LMI borrowers and communities to different degrees, there is every likelihood that different community metric multipliers are needed to calibrate performance accurately. The same disparity applies to home purchase mortgages and refinance mortgages, the blend of which shifts with interest rates. If this is the case just within the home mortgage space, there is no reason to assume that the same community benchmark multiplier should also apply to small business, small farm, or multifamily multipliers. The issue should be less problematic for market metric multipliers, at least as they apply to home mortgage and small business/farm lending, because these are set relative to broad industry (not only banks’) performance in that AA.

The agencies should set thresholds based on the lower of the calibrated market benchmark or calibrated community benchmark.

As discussed in our response to Question 80, we have little confidence that the community metric multipliers are properly calibrated. As discussed in our response to Question 78, we also suspect the validity of the home mortgage borrower benchmark. Unless these concerns can be addressed, we strongly advise against using them as binding elements in performance measurement, such as in combination with the market benchmarks.

The best way to avoid market underservice is to motivate better performance, particularly by encouraging a race to the top. This can be best accomplished by a combination of elements, as we described at the outset of this comment letter.

The market metric for Outstanding performance is too high. Banks should see that they have a reasonable chance of attaining Outstanding performance. Setting the Outstanding market metric threshold at 125 percent of industry performance – a level the Agencies estimate that no bank with assets exceeding $50 billion would achieve and that we estimate would include banks with only 2 percent of banking system assets – would clearly signal to banks that an Outstanding rating is beyond reasonable reach. Instead, banks would rationally resign themselves to an overall Satisfactory rating, along with perhaps 80 percent of all banks. Moreover, since Retail Lending accounts for 45 percent of the overall rating, and since the market metric is highly likely to be binding in most cases, it is highly unlikely that a bank could achieve an overall Outstanding rating unless it is Outstanding in Retail Lending. If the great majority of banks expect to receive the same Satisfactory rating, there will be little motivation within a given bank to keep up with high performing competitors, and the bank will find no discomfort in the middle of the pack. The NPR’s approach could also make CD performance immaterial to most banks’ overall CRA rating, as we describe elsewhere, which would be a very serious concern.

If Outstanding is beyond reach a bank’s reach, its objective might rationally be only to avoid a Needs to Improve rating, suggesting the bank should be comfortable performing below industry average. Should these dynamics take hold, they could easily feed a decline in lending. We take little solace in the fact that the Agencies would also publish numerical performance scores in addition to ratings. The headline rating will be what matters. This prospect would be a highly counter-productive result for communities.

The Agencies should also reconsider the market metric for High Satisfactory retail lending performance. A High Satisfactory market threshold set at 110 percent of industry performance is problematic because it locks in the conclusion that an estimated 60 percent of the industry is destined to perform weakly. This is the case because when industry performance rises or falls, every percentage of that performance will also rise or fall. We infer that the Agencies propose to set high standards in hopes that banks will seek to exceed them, thereby serving their communities. We doubt this strategy will succeed.

A better way – indeed, a necessary way, in our view – to incent High Satisfactory performance would be to differentiate much more between the points awarded for High Satisfactory and for Low Satisfactory performance. As proposed, the difference is literally minimal: only one point. This signals there is little difference between banks performing at the 15th and 85th (or even 95th) percentiles of the industry. If, however, the points differential is wide enough, it will signal to banks that a High Satisfactory puts them on the road to Outstanding, while a Low Satisfactory warns they could slip to Needs to Improve. 39 Of course, performance conclusions and ratings will aggregate numerous metrics, especially for banks with multiple AAs. Most banks will receive a mix of performance scores. Because the long road to an overall rating starts with the performance scores for each metric, getting the metrics, multipliers, and assigned points right is essential to achieve the best outcomes for LMI people and communities.

In general, we support equal weighting of geographic and borrower metrics, but we would support a different blend in rural or nonmetro areas with few LMI CTs.

This is an important issue for which we have only limited visibility. Anecdotally, a few of our members have shared pre-testing results that suggest that home mortgages would have a dominant influence on Retail Lending Test conclusions because the average mortgage size is so much greater than the average small business/farm loan size. We cannot tell how much this balance might change when small business definitions change, since moving to a $5 million revenue threshold could result in a larger average loan size. However, the limited information we have is a source of significant concern. Even though “housing” is part of our organization’s name, we recognize the importance of small business lending to communities and would not want the Retail Lending Test to become a de facto home mortgage lending test. Switching from dollar volume alone to loan count alone would create the opposite problem, especially for banks that make many small business credit card loans. For this reason, we tentatively support using a combination of dollar volume and loan count.

It would be great to formulate a robust estimate of good market service. However, such a formula would be difficult to set. If the Agencies wish to set such a formula, they should propose it for public comment before adopting it.

The Agencies should consider additional information offered by banks and other stakeholders.

All large banks’ home mortgage and small business/farm lending outside AAs should be evaluated, with a possible exception of banks that make less than perhaps 1-2 percent of their loans there.

We prefer the tailored approach as the Agencies have proposed.

Yes, AA and outside retail lending area conclusions should be weighted by the average of a bank’s percentage of loans and deposits there.

We support the inclusion of special purpose credit programs as an example of loan product or program that facilitates home mortgage and consumer lending for low- and moderate-income individuals. Such programs can promote equal opportunities for racial and ethnic minorities and other disadvantaged groups and should be defined either demographically or geographically. This objective is entirely consistent with CRA, which is, after all, a civil rights law intended to help meet the credit needs of a bank’s entire community. We find that LMI is only a weak proxy for race, and that people and communities of color face long-standing disadvantages that justify targeted initiatives.

We support using the local benchmarks. Using a nationwide benchmark in AAs with low levels of CD financing is likely to be unduly demanding because CD financing opportunities in these areas tend to be limited. Conversely, a national benchmark would be unduly lenient in areas with higher levels of CD financing, where CD financing opportunities are likely more plentiful.

Setting performance expectations at the end of a performance period would defy basic management principles. Because CD financing activity will be assessed annually, the benchmarks should be set annually, based on the most recent year for which they are available. For example, data from year 1 would be available in year 2 and used to set benchmarks for year 3. This approach would solve the significant transparency and predictability problem under the NPR that banks cannot know at the start of any year what they will be expected to achieve in that year, thwarting their ability to manage their efforts effectively. This approach would also avoid applying different benchmarks to comparable banks, depending on the timing of their evaluation periods.

Yes.

We strongly support substantial emphasis on responsiveness, including the Impact Factors as well as other qualitative factors that reflect an activity’s responsiveness to local needs. However, it is difficult to comment further without a clearer understanding of how the Agencies intend to apply these considerations.

Publishing standard metrics would be informative. The Agencies should consider both the dollar value and number of activities, so that small-scale activities get due consideration. We are also mindful that some activities could qualify for two or more Impact Factors, such as support for a CDFI that works in a specially targeted geography and serves low-income individuals. Finally, we are not certain that a simple formula can entirely avoid unintended consequences.

We urge the Agencies to eliminate the CD Services Test. The NPR cites three main CD service activities, but two – financial literacy and small business technical assistance – properly belong on the Retail Services and Products Test because they serve consumers and small businesses, respectively. Service on nonprofit boards is a legitimate CD service but it does not warrant a separate test worth 10 percent of the entire rating. This is especially the case because CD Financing would be worth only 30 percent of the overall rating, a mere 5 percent more than the current Investment Test, which does not include CD lending. True CD services should be considered qualitatively within a single, consolidated CD Test.

On a separate point, as the diversity of bank models proliferates, opportunities are arising to apply banks’ specialized expertise and relationships to advance CD in ways far beyond more traditional roles such as nonprofit board service. For example, banks have started to help CDFIs and affordable housing nonprofits to access capital markets in powerful and exciting ways, such as by underwriting bonds or securitizing unconventional loans. The bank is not itself a lender or investor in these cases, so the activity would not qualify as financing. Nor is this (or should it be) volunteer work and it should not be measured by the number of hours provided. Other examples are likely to come from fintech banks, which could offer more efficient technology driven solutions to such challenges as developing better loan management systems for CDFIs.

The Impact Factors alone may be insufficient to recognize such forms of CD services appropriately.

We do not support consideration for volunteer activities unrelated to the provision of financial services. For example, if bank employees volunteer to help build a home or serve Thanksgiving dinner at a homeless shelter – to cite two plausible CD purposes – that is commendable but it should not qualify for CRA credit.

The administrative burden would be excessive for many banks. Such activities should be considered qualitatively. In addition, the Agencies should clarify that banks need not collect and support volunteer service hours, or other activities, for which they do not wish to claim CRA credit. CRA is not the motivation for many volunteer activities, so requiring banks to identify and document every activity that might be CRA-qualifying would be unproductive.

No.

Establishing AA benchmarks for wholesale and limited purpose banks is important. A CD financing benchmark at the FBAA level should be based on:

• • for banks with assets over $10 billion, the share of the bank’s deposits it collects from that FBAA, multiplied by the bank’s institution CD financing benchmark; and

• • for banks with assets up to $10 billion, the share of the U.S. population (or alternatively, the share of the U.S. LMI population) residing in that FBAA, multiplied by the bank’s institution CD financing benchmark.

Establishing institution benchmarks for wholesale and limited purpose banks is important. The nationwide CD financing benchmark for all large banks should be the starting point. However, assets are a better measure than deposits of a wholesale and limited purpose bank’s capacity. Accordingly, the assets of these banks should be used to set their institution benchmark.

Wholesale and limited purpose banks that wish to receive consideration for CD services should submit services to be reviewed on a qualitative basis at the institution level, without having to opt into the CD Services Test.

Although we would expect Strategic Plans to be used predominantly by wholesale banks, limited purpose banks, and banks that operate primarily outside FBAAs, it is entirely possible that Strategic Plans could be appropriate in other cases. For example, a branch-based bank whose primary retail lending product is consumer lending should be able to have a Strategic Plan weighted to emphasize that product, since the standard weighting between Retail Lending and Retail Services and Products would inaccurately reflect its business strategy.

The proposed approach does not sufficiently accommodate the full range of business models in the banking industry. For banks with unique business models, a flexible Strategic Plan option is needed to craft a CRA evaluation framework that reflects the bank’s particular structure, products, and delivery systems. We urge the Agencies to preserve the flexibility of the current rule’s approach to Strategic Plans, including flexibility with respect to AAs, in-scope products, test weights, and measurable goals. We would note that the requirements for public comment and regulatory approval serve as guardrails to ensure that banks do not attempt to use the Strategic Plan option to evade CRA responsibilities.

Depending on the thresholds for RLAAs, it is possible that a bank could have an unmanageable number of RLAAs, especially relative to its overall size. In such cases, the Agencies should have the flexibility to approve a Strategic Plan with fewer or no RLAAs.

Approved Strategic Plans should be easily available to the public, including postings on the Agency’s and bank’s websites.

Regarding large banks, not only intermediate banks, it will be essential that retail and CD performance receive equal weight in order to motivate the best performance for both activities. Large banks’ retail test performance in the NPR would drive a bank’s rating because the Retail Financing Test and the Retail Services and Products Test comprise 60 percent of a bank’s overall rating. Unless a bank achieves Outstanding retail performance, its CD performance is highly unlikely to affect the bank’s overall rating. A bank that is Satisfactory on retail is nearly certain to receive an overall Satisfactory rating whether its CD performance is Outstanding or even Needs to Improve.

Two scenarios illustrate the point. First, a bank with retail performance at the midpoint of the High Satisfactory range (7.5 points) and CD performance at the midpoint of the Outstanding range (9.25 points) would have an overall Satisfactory rating (7.5 X 60% plus 9.25 X 40% = 8.2 points). Second, a bank with retail performance at Low Satisfactory midpoint (5.5 points) and CD performance at the Needs to Improve midpoint (3.0 points) would still receive an overall Satisfactory rating (5.5 X 60% plus 3.0 X 40% = 4.5 points). It is mathematically possible for a bank to achieve an overall Outstanding rating if it combines retail performance at the top end of the High Satisfactory range with CD performance at the top end of the Outstanding range, but such scenarios are highly unlikely.

If a bank cannot reasonably expect to achieve an Outstanding retail performance, CRA will provide little motivation for CD activity. Especially because CRA drives so much CD activity, such an outcome would be a major setback for the entire CD field. We urge the Agencies to weight retail and CD activity equally for large banks, so they are motivated to maximize performance on both. We note that, per the NPR, for intermediate banks, retail and CD are weighted equally.

No. Please see our response to the previous question.

Only FBAAs should be subject to this standard. As data presented in the NPR indicates, it is much harder for banks to meet LMI credit needs where they do not have a local branch presence and must compete with banks that do have branches there. We support evaluation of bank performance outside FBAAs, but RLAAs should not be counted in the 60 percent Low Satisfactory (or better) standard.

We support the NPR’s proposal that banks with assets over $10 billion collect and report deposits data.

Large banks should collect and report activity-level CD data.

However, banks should at least have the option to measure CD activity data as of the end of each calendar year, not quarterly. Quarterly CD activity measurement would be excessively burdensome. Although year-end data might exclude some loans or investments repaid during that year, many banks would value the burden relief more.

Yes. We believe racial equity and justice are foundational to such core American values as equal economic opportunity and justice, as well as to the nation’s prosperity, social cohesion, and world-wide reputation. Racial equity and justice are also at the heart of NAAHL’s mission “to expand economic opportunity through the responsible financing of affordable housing and inclusive neighborhood revitalization.” As a national alliance of major banks and mission-based capital providers, we believe deeply in the power of community reinvestment to improve neighborhoods and to create opportunities for economic mobility. Quality affordable housing in vibrant neighborhoods and well-paying jobs are platforms for educational opportunity, financial security, health, and public safety – all of which are essential to racial equity. NAAHL member banks provided $183 billion in financing for LMI people and communities in 2020. However, we also acknowledge our industry’s failure to meet other responsibilities to communities of color. We have sometimes redlined neighborhoods and otherwise denied credit on fair terms. We have missed opportunities to develop and deploy the financial products that communities need. We have insufficiently engaged the power and agency of Black, Latinx, and all people and communities that have suffered under systemic racism.

We regret that CRA – a civil rights law – is so limited in its capacity to further these ideals. Publishing HMDA – and, in the future, section 1071 small business lending data – would offer a welcome if limited advance.

We urge the Agencies to publish CD financing activity data, preferably at the county level. If county-level data would compromise privacy, then data should be published at the MSA, nonmetro statewide, state, and institution levels. Aggregated data for all reporting banks should also be published at the county, MSA, nonmetro statewide, state, and nationwide levels.

It would be better to provide an additional 12 months after CFPB finalizes its rule, so banks have time to adjust their systems.