Community groups, think tanks and the residential mortgage lending industry took advantage of the Request for Input ("RFI") issued by the Federal Housing Finance Agency ("FHFA") on September 7, 2021, seeking public input and information to assist Fannie Mae and Freddie Mac (the "GSEs" or "Enterprises") in their respective preparation of the first of their three-year Equitable Housing Finance Plans (the "Plans"). The Plans must identify and address barriers to sustainable housing opportunities, including the Enterprises' goals and action plans to advance equity in housing finance. The Enterprises were scheduled to submit these Plans to FHFA by December 31, 2021.

Background on FHFA RFI

The RFI notes that "...a number of statutory and regulatory authorities that apply to FHFA and the Enterprises speak to the need to advance equity for homebuyers, homeowners, and tenants in the housing market" as part of the public purposes of the Enterprises. It itemizes many of these authorities. One statutory mandate is to provide ongoing assistance to the secondary market for residential mortgages, including mortgages for low- and moderate-income ("LMI") families involving a reasonable economic return that may be less than the return earned on other activities, and to promote access to mortgage credit in central cities, rural areas and underserved areas.1 Other authorities pertain to implementing Enterprise affordable housing goals2 and satisfying the Enterprises' statutory "Duty to Serve" affordable housing needs of certain underserved markets consisting of manufactured housing, affordable housing preservation and rural housing.3

Similarly, the RFI references Executive Order 13985, which President Biden issued on January 20, 2021. Titled "Advancing Racial Equity and Support for Underserved Communities Through the Federal Government," the executive order provides that "the Federal Government should pursue a comprehensive approach to advancing equity for all, including people of color and others who have been historically underserved, marginalized, and adversely affected by persistent poverty and inequality." 4 The Plans, according to the RFI, serve to supplement existing FHFA and Enterprise requirements, programs and plans "and are designed to ensure a continued focus on housing equity that is aligned with other critical objectives including safety and soundness and other mission activities."

Content of the Plans

The Plans, which are to become effective in February 2022 and updated annually, must include objectives, measurable goals and planned meaningful actions related to reducing:

  • The racial or ethnic homeownership gap
  • Underinvestment or undervaluation in formerly redlined areas that remain racially or ethnically

concentrated areas of poverty or otherwise underserved or undervalued Additionally, FHFA provided the Enterprises with a non-exclusive list of objectives and goals that could be undertaken in the Plans, including reducing:

  • Racial or ethnic disparities in acceptance rates for the Enterprises' respective automated underwriting systems ("AUS")
  • Racial or ethnic disparities in the share of loans acquired by the Enterprises compared to the overall mortgage market

Responses to the RFI also are expected to assist the Enterprises in their preparation and FHFA in its oversight of the Plans. The RFI posed the following questions:

  1. How should measurable goals be selected and set by the Enterprises? For example, is pursuing a small set of focused goals or a wide portfolio of goals better?
  2. What data, information or analyses would be helpful for the Enterprises to consider or use to support their plans?
  3. How should the Enterprises undertake setting objectives, measurable goals and meaningful actions to sustainably address the racial and ethnic homeownership gap?
  4. How should the Enterprises undertake setting objectives, measurable goals and meaningful actions for formerly redlined areas? How should such areas be defined?
  5. What other objectives and measurable goals should the Enterprises pursue in their plans?
  6. What constitutes a "meaningful" action, and what kinds of meaningful actions should be taken by the Enterprises under their plans?
  7. How can the Enterprises and FHFA ensure that actions taken under the plans provide sustainable housing opportunities and are consistent with safety and soundness?
  8. What should FHFA consider in overseeing the Enterprises' plans? Should FHFA provide a rating or some other public assessment? If so, how should the plans be assessed?
  9. How should the plans interact with Duty to Serve, Housing Goals or other requirements?
  10. Could special purpose credit programs (as defined in 12 CFR 1002.8) be included in the Enterprises' plans? How should such programs be structured?
  11. Are there additional or different required objectives and goals that FHFA should consider for future Enterprise plans?
  12. What communities and stakeholders should the Enterprises consult with in developing their plans?

Summary of General Comments to the RFI

Most of the responses to the RFI did not specifically answer each of the 12 questions but centered on common themes.

One theme focused on the process of how to define "Equitable Housing Finance" and establish measurable quantifiable goals toward achieving that objective. A second focused on what resources and expertise the Enterprises need in order to be better-functioning institutions, such as more experienced staffing, changes to the Enterprises' increased capital requirements and augmenting their fair lending supervision. Third was the use, continued sharing of and refinement of the Uniform Appraisal Dataset and National Mortgage Database, among other data collected by the Enterprises, with cross-market stakeholders and researchers.

A fourth and oft-repeated theme was the mismatch between supply and demand for homes for owner-occupancy and the litany of federal, state and local initiatives that could be undertaken to increase the supply of housing. More equitable loan servicing practices was a fifth theme. The failure of FHFA to include people with disabilities, instead focusing on racial and ethnic minority populations, was another theme, given that adults with disabilities are more likely to be poor and live in LMI neighborhoods than individuals and families without disabilities. But, for this Legal Update, we highlight another theme pertaining to the Enterprises' purchases of mortgage loans, including underwriting criteria, product types, loan size and loan pricing. Below we summarize some of these comments.

Summary of Comments Related to Loan Purchases

BORROWER ELIGIBILITY

Industry stakeholders provided a series of recommendations related to loan purchase eligibility based on borrower creditworthiness. For purposes of this Legal Update, we refer to these eligibility criteria as underwriting, while recognizing that, as loan purchasers, the Enterprises do not underwrite loans originated by lenders.

The Urban Institute's ("UI") comment reflected a common sentiment: "The Enterprises' underwriting box is narrow, a relic of the Great Recession that is likely exacerbated by lender overlays and Enterprise loan-level risk based pricing, impeding homeownership for many who might have been approved under a more `normal,' risk-tolerant set of underwriting guidelines." The comment letter further asserted that while the Enterprises' eligibility criteria for loan purchase "may seem to be race neutral, they aren't." Cutting to the chase, the UI wrote: "Despite increased reliance on automated and algorithmic underwriting, the three basic `C's' of mortgage lending--credit, collateral and capacity--still drive underwriting decisions today. While these criteria don't explicitly include race, relying on them necessarily perpetuates racial inequities." All current underwriting determinants, the UI wrote, "should be scrutinized through an equity lens to identify opportunities for safely expanding equitable access to housing."

Many commenters commended FHFA and the Treasury Department for suspending the prior presidential administration's amendments to the Preferred Stock Purchase Agreements to cap the support the Enterprises can provide for certain single-family mortgage loans. The single-family loan limit restricts the Enterprises' purchase of "high-risk" loans to 6 percent of their purchases and 3 percent of their refinance mortgages. "High-risk" loans are defined as loans with at least two of the following characteristics: greater than 90 percent loan-to-value ratios, greater than 45 percent debt-to-income ratios and a borrower credit score below 680. In order to make responsible mortgage credit available to underserved families of color, some commenters, such as the Center for Responsible Lending ("CRL"), urged the conversion of this suspension to an outright rescission of the prior amendments.

Automated Underwriting Systems

Some commenters recommended the modernization of the Enterprises' automated underwriting systems, in part, by relying more on alternative data. In addition to considering positive rental payment history, which the Enterprises recently have incorporated, assessing the viability of other data sets, such as car payments, utility bills and cash flow data from consumers' bank account records, was a common comment. Rocket Mortgage, for example, supports pilot programs to regularly assess the merits of incorporating alternative data sets into the Enterprises' automated underwriting systems, and the Housing Policy Council ("HPC") suggested that updates to the Enterprises' automated underwriting systems should explore ways to capture income variability, which is more common for LMI individuals. The National Consumer Law Center ("NCLC") (joined by several other consumer advocacy groups) also encouraged the use of certain alternative data, such as cash flow, but cautioned that "alternative data is a mixed bag, and the details matter."

One comment from the National Housing Conference ("NHC") asked the Enterprises to work with qualified fair housing organizations to identify racial and ethnic disparate impact in their automated underwriting systems and develop strategies that result in less discriminatory alternatives. It also encouraged the use of artificial intelligence and machine learning to develop less discriminatory predictive models. Similarly, the National Community Stabilization Trust ("NCST") believes that reducing the "accept-rate gaps" of the Enterprises' automated underwriting systems is a "critical and direct way" to improve racial and ethnic homeownership gaps calls for an examination of arbitrary cutoffs (such as credit score floors) and upfront fair lending testing of new releases before implementation. In contrast, the CRL cautioned in its response that the Enterprises should ensure that the broad use of machine learning, artificial intelligence and big data do not perpetuate discrimination and thwart inclusive lending. Similarly, the National Fair Housing Alliance ("NFHA") (on behalf of itself, the Consumer Federation of America, and other advocacy groups) encouraged the Enterprises to "embrace innovative technologies designed to significantly reduce, and ultimately eliminate, bias in algorithmic-based systems."

Credit Scores

Many commenters discussed the need for refinement of credit score models on which the Enterprises presently rely and likely will rely in the future. As the NHC comment letter explained: "Credit scores, while predictive of risk in the aggregate and an important component of sustainable lending, remain a barrier to mortgage credit for many borrowers of color and low income borrowers." The National Community Reinvestment Coalition ("NCRC"), for example, urged FHFA to ensure that the Enterprises only use credit scoring models that reduce barriers for people of color by using data elements that are more likely to include those with little or no credit history. The National Association of Realtors ("NAR") applauded the Enterprises' prior rulemaking providing for the development of a framework to evaluate new credit scoring models but expressed its frustration that "there has been little progress in adopting new scores." While the NAR acknowledged the value of FHFA's action to allow the Enterprises to use rental data in underwriting, it wrote that homebuyers must continue to qualify for mortgage loan assistance based on traditional credit score thresholds, limiting the benefits of underwriting with alternative data.

The National Housing Resource Center ("NHRC") instead simply stated that the Enterprises should lower the required credit scores or adjust their automated underwriting systems that ascribe higher risk to applicants with lower credit scores.

Debt-to-Income Ratios

Stakeholders also highlighted the need to reevaluate the debt-to-income ratio as a means to determine that the underwriting process had fewer discriminatory effects. For example, NFHA commented on the struggle of prospective borrowers of color to meet standard debt-to-income and loan-to-value ratios necessary for loan approval because of past discriminatory federal housing policies that prevented access to homeownership and the wealth building benefits it offers. While the NAR acknowledged the efforts of the Enterprises to reduce the effects of student loan payments in underwriting, the NAR focused on the higher burden of student debt on persons of color and the fact that student debt may make a renter's debt-to-income ratio exceed current eligibility thresholds. Interestingly, there were no public comments challenging the use of debt-to-income ratios as reliable indicators of the likelihood of default.

Loan-to-Value Ratios

According to a number of commenters, one of the main methods to ameliorate the underwriting process for marginalized borrowers is to reevaluate loan-to-value ratio requirements. Comments regarding required loan-to-value ratios generally fell into two categories that are inextricably tied together: appraisals and down payments. An undervalued property overstates the loan-to-value-ratio of a loan, potentially resulting in a loan denial or an increased down payment requirement. Consumer advocacy groups noted that, if a borrower is struggling to meet the initial down payment requirement, an undervalued property that results in an increased down payment requirement to meet the specified loan-to-value ratio will exacerbate the problem.

Appraisals

Bias in home loan appraisals is a hot topic these days, and FHFA received several comments on appraisals. Without regard to questions of bias, comments expressed the need to modernize and standardize the appraisal process. Additionally, some commented that the current appraiser shortage is causing unprecedented delays and costs to consumers while creating a disparate impact on communities of color.

CoreLogic wrote that the lack of supply of affordable housing in LMI neighborhoods is made worse because home builders who target these neighborhoods for development lack "market comparable properties" to use to establish value, resulting in the appraised value of new or rehabilitated affordable homeownerships units being "chronically lower than the market listing price of other homes in the neighborhood, which tend to be older, not in as good a condition, or use older technology materials that are less durable and energy efficient as the new/rehabilitated property targeted for development." It recommended the Enterprises create an "appraisal gap" subsidy program to cover the difference between appraised value and market listing price in neighborhoods that do not provide adequate market comparable properties.

The NAR urged the Enterprises to conduct more research on determining where and how valuation can reinforce discrimination so that safe and sound measures to correct any discrimination in the valuation process can be determined.5 NFHA said that the Enterprises should ensure that appraisal guidelines mitigate any potential fair lending risk and any potential harm to people and communities of color and specifically called out the sales comparison approach as an appraisal method that may perpetuate the history of undervaluing homes in communities of color. NCRC recommended that the Enterprises use appraisal data to identify geographic areas with the largest appraisal gaps as a first step toward providing more targeted solutions. Many commenters asked the Enterprises to release their appraisal data to the public on an aggregate basis to facilitate research regarding the impact of appraisals and other housing practices on communities of color. Referencing research from the Woodstock Institute, NCRC recommended that the Enterprises discontinue their policies of only using the first, lower appraisal when homeowners receive a second, higher appraisal because they believe the first appraisal was discriminatory.

Down Payment Assistance

Funding a house down payment is one of the largest barriers to homeownership. Many commenters highlighted the lack of generational wealth as a barrier for many African-Americans to make their required equity investment. The NAR highlighted the double-digit percentage increases in singlefamily home sale prices throughout the country, which raised required down payment amounts across the spectrum of buyers. It supports the creation of safe and sound down payment assistance programs, as long as such programs do not facilitate lending to borrowers who are not financially capable of ongoing mortgage payments. In addition, stakeholders, such as NFHA, encouraged the Enterprises to support targeting down payment assistance to first generation homebuyers.

CoreLogic, on the other hand, pointed out that there are thousands of down payment assistance programs available for LMI borrowers, but prospective homebuyers and lenders need "easy access to tools that match these homeowners to available programs in order to make sure the homeowner is eligible." Lenders also need the knowledge of "which prospective LMI homeowner populations are most likely to be in need of down payment assistance and where they are located so that down payment assistance programs can be targeted and allocated accordingly."

LOAN PRODUCT TYPES

Comments addressed loan product types in two different ways. Increasing the income thresholds on the Enterprise's affordable housing products (e.g., HomeReady and Home Possible), particularly in high-cost areas, is one way to expand the pool of minority borrowers eligible for Enterprise-backed financing, suggested the National Association of Home Builders ("NAHB") and the Mortgage Bankers Association ("MBA"). U.S. Mortgage Insurers ("USMI") suggested that FHFA reverse the changes to its affordable housing products that FHFA made under the prior presidential administration, such as the area median income eligibility, which resulted in a significant decrease in the use of these products. The idea is that limiting affordable housing products offered by the Enterprises to very low-, low-, and moderate-income borrowers shuts out minority borrowers with higher incomes but who need lower down payment requirements and flexible sources of down payment funds.

Another way pertains to the types of loan products. NCRC encouraged the Enterprises to expand their 20-year loan products to help eligible people of color build equity faster and adopt metrics regarding these products in their equity plans. The NAHB recommended allowing the Enterprises to create a secondary market for acquisition, development and construction loans, which has the dual benefit of increasing the supply of single-family homes and addressing the demand side of the equation. CoreLogic called for "retooling" the home improvement loan products of the Enterprises. The NAR recommended expanding mortgage products that allow homeowners to combine a home purchase and renovation, such as Fannie Mae's HomeStyle and Freddie Mac's CHOICERenovation loan products, and increasing loan limits. USMI also suggested implementing a conventional mortgage product similar to that proposed by Senator Mark Warner in his "Low-Income First Time Homebuyers Act." This product would be targeted to first-time, first-generation homebuyers with household incomes of less than 120 percent of the area median income (or less than 140 percent of area median income for properties located in high cost areas).

The Pew Charitable Trusts ("PCT") suggested that the Enterprises promote manufactured housing as a source of affordable housing through a personal property loan pilot program. It cited the work of the MBA, which had identified the lack of a secondary market for personal property loans as a barrier for lenders to finance manufactured home purchases due to liquidity constraints.

LOAN SIZE

Increasing purchases of small-dollar mortgages was another major topic. For example, NCLC, among others, urged FHFA to require the Enterprises to consider significantly expanding their purchase of, and support for, small balance mortgage loans in their Plans. Residential mortgage loans with original principal amounts of under $150,000 generally are thought to be small-dollar mortgages. Of course, lack of supply of affordable housing is one reason for the decline in small-dollar mortgages due in part to the increasing percentage of all cash purchases of single-family homes for investment as rental housing. But, assuming such loans are available to be made, one comment letter reported that the denial rate for minority borrowers of small-dollar loans is high. This is due in part to the denials often not relating to an applicant's credit profile but, rather, being based on a lender's inability to make small-dollar loans profitably. This is tied to the economics where origination fees and servicing fees are pegged to the principal amount of the loan but the fixed cost to originate or service does not depend entirely on the size of the loan.

The result, according to the PCT, is that many borrowers who otherwise would qualify for traditional residential mortgage loans instead may turn to non-mortgage alternative financing arrangements that are thought by some to be less beneficial to consumers than a traditional mortgage loan, such as land contracts and lease/purchase arrangements. Some stakeholders, however, asserted (i) it would be premature to conclude that non-mortgage alternative financing arrangements are universally harmful to consumers and (ii) more analysis is needed to understand why some buyers choose alternative arrangements and to what extent the perceived benefits of those arrangements materialize.

LOAN PRICING

Comments regarding loan pricing generally fell into three categories. One line focused on the enhanced capital requirements for the Enterprises that FHFA imposed last year through regulation. These comments generally noted that there is a direct relationship between higher capital requirements for the Enterprises and their pricing of loan purchases and recommended a relaxation of some of these requirements.

Other comments addressed pricing subsidies. Overall, the Enterprises should pool more of the mortgage risk and keep pricing as level as possible across credit categories without driving loans out of the Enterprises' market through adverse selection, wrote the CRL. But, more broadly, the HPC noted that "The [Enterprises'] pricing framework would benefit from transparency regarding the overall [Enterprise] cross-subsidization methodology."

Loan Level Pricing Adjustments ("LLPAs") generated the most comments about pricing. Introduced in 2008 to help partially offset the record losses resulting from the then-housing and financial crisis, LLPAs are payable by loan sellers to the Enterprises based on perceived higher risks but generally are passed through to borrowers through higher interest rates. Many argue that LLPAs may not be as necessary today. For example, NFHA wrote:

Notably, the key business justifications for the LLPAs do not hold up under scrutiny. The factors upon which the LLPAs rely (loan-to-value ratio and credit score) are not related to the key risk features of the majority of the loans that experienced massive defaults in the foreclosure crisis. Those loans failed due to the combination of poor underwriting, little or no documentation, high fees, exploding interest rates, risk layering, and negative amortization. The Truth in Lending Act's Ability to Pay/Qualified Mortgage rule already addresses most of these risks.

Regardless of the reasons for their initial adoption, LLPAs, many (such as NFHA) wrote, have a disparate impact on minority borrowers because the common metrics of a loan's loan-to-value ratio and the borrower's credit score generally work against minority borrowers due to a lack of generational wealth to fund down payments and materially lower average credit scores. The MBA recommended reversing credit score-related pricing adjustments on low down payment products in specific metropolitan areas as well as eliminating LLPAs entirely for first-generation homebuyers.

The NCLC went even further, commenting that the Enterprises should be required to abandon the LLPA framework. The NFHA declared that "LLPAs are not necessary for safety and soundness reasons or to recoup the GSEs' lost capital" and should be eliminated "as they are a barrier that unnecessarily increases the cost of homeownership." At a minimum, many asserted, LLPAs should be evaluated and likely reduced in order to align with reasonably expected credit losses to be suffered by the Enterprises. The HPC, for example, advocated that the Enterprises' upfront fees should be "appropriately calibrated to compensate the [Enterprises] for expected...credit losses on borrower defaults, without unintended, disparate impacts on protected classes."

Another common suggestion was to remove LLPAs on rate and term re-financings. Presumably, the seller paid a LLPA to the relevant Enterprise at the time the loan is sold to the Enterprise, and the borrower bore the economic brunt of that adjustment through a higher interest rate. Borrowers refinancing loans already acquired by Freddie Mac or Fannie Mae indirectly paid an LLPA when the loan was initially originated and sold to an Enterprise, and some commentators, such as America's Homeowner Alliance, asserted that this fee should not be paid more than once.

The HPC suggested perhaps adding more elements in the pricing matrix:

An additional way to achieve more balance between risk and access in loan pricing may be to add other loan attributes to the pricing matrix. We would encourage exploration of additional factors such as small-dollar mortgages (those with an unpaid principal balance of less than $150,000) loans in certain geographies (including historically redlined areas, areas of concentrated poverty, high opportunity areas, or rural areas), and low down payment mortgages that have 3rd party credit enhancements. It is possible that loans with these characteristics are potentially being mispriced due to their slower prepayment rates as well as their lower loss severities (particularly due to the increase in home prices for entry level properties). A study of this issue could uncover excess pricing for some of the most vulnerable borrowers and contributing to inequitable housing outcomes.

SPECIAL PURPOSE CREDIT PROGRAMS

One of the other major recommendations from industry stakeholders is the more effective use of Special Purpose Credit Programs ("SPCP"), which are authorized under the Equal Credit Opportunity Act ("ECOA").6 Notwithstanding ECOA's general prohibitions against discrimination in the extension of credit, ECOA provides that it is not discriminatory for creditors to establish targeted credit assistance programs for certain purposes, including an SPCP offered by a for-profit lender to meet special social needs and that meets standards prescribed in Regulation B.7

Some of the commenters to the RFI expressed their interest in the Enterprises' potential use of SPCPs as a means to target the racial or ethnic homeownership gap but referenced the continuing uncertainty about their legality. After the end of the RFI comment period, on December 6, 2021, the Office of General Counsel of the Department of Housing and Urban Development ("HUD") issued a legal opinion concluding that a for-profit's SPCP, designed and administered in accordance with ECOA and Regulation B, "generally" does not violate the Fair Housing Act. While not binding on courts, the HUD legal opinion perhaps opens the door for the Enterprises to design and implement targeted programs that are explicitly dependent on a common characteristic such as race or ethnicity. Lenders in turn could mimic the targeted eligibility criteria in an SPCP.

NFHA urged the Enterprises to create purchase programs for "race-conscious" SPCPs to ensure liquidity for the market. NCRC advocated that SPCPs should have a "vital role" in the Enterprises' Plans, explaining that "these programs can target formerly redlined neighborhoods or geographical areas with past and ongoing discrimination and can target the subgroup within people of color that are experiencing discrimination." The comment letter did not mention the legal uncertainty of this course of action, but it did identify the reluctance of lenders to adopt SPCPs because of liquidity concerns that the Enterprises might be able to alleviate through their purchase of loans originated through SPCPs.

LIMITED ENGLISH PROFICIENCY

Stakeholders also support additional efforts to promote access to mortgage credit for mortgage-ready limited English proficiency borrowers. According to commenters, there is a categorical need to provide translations into more languages and to continue working with lenders to reduce potential legal liability. The latter can be done by including disclosures on translated documents that state the English language version is the enforceable version of the documents. As the Enterprises do with lenders, stakeholders asserted the Enterprises should provide indemnification to housing providers that use the Enterprises' translations of the standardized documents if there is a conflict between the translation and the English language original. NCLC's comment letter went to great lengths to discuss the need for FHFA to take meaningful steps to promote language access in origination and servicing as well.

The stakeholders, such as NCRC, NCLC, NHRC and NFHA, encourage FHFA to restore the language preference question in the Uniform Residential Loan Application ("URLA"), particularly in the fields of the application that collect information on participation on pre-purchase education or counseling. In 2017, FHFA finalized a language preference question on the revised URLA, but it subsequently removed the question following industry concerns about legal risk. By the FHFA integrating the collection of language preference in mortgage origination, requiring that the language preference transfers at closing with the mortgage information to the servicer, and making the information available if servicing is sold, commenters argue that lenders and servicers would have the ability to communicate in a borrower's preferred language at any point in the relationship.

Conclusion

FHFA received a broad range of recommendations on ways that the Enterprises can help lessen the apparent and actual racial inequity in homeownership across the country. It will be interesting to see if and how, regardless of how the Enterprises incorporate these suggestions into the final Plans, the Plans will impact the conforming residential mortgage market.