2020-27567. Qualified Mortgage Definition Under the Truth in Lending Act (Regulation Z): General QM Loan Definition

AGENCY:

Bureau of Consumer Financial Protection.

ACTION:

Final rule; official interpretation.

SUMMARY:

With certain exceptions, Regulation Z requires creditors to make a reasonable, good faith determination of a consumer's ability to repay any residential mortgage loan, and loans that meet Regulation Z's requirements for “qualified mortgages” (QMs) obtain certain protections from liability. One category of QMs is the General QM category. For General QMs, the ratio of the consumer's total monthly debt to total monthly income (DTI or DTI ratio) must not exceed 43 percent. This final rule amends the General QM loan definition in Regulation Z. Among other things, the final rule removes the General QM loan definition's 43 percent DTI limit and replaces it with price-based thresholds. Another category of QMs consists of loans that are eligible for purchase or guarantee by either the Federal National Mortgage Association (Fannie Mae) or the Federal Home Loan Mortgage Corporation (Freddie Mac) (government-sponsored enterprises or GSEs), while operating under the conservatorship or receivership of the Federal Housing Finance Agency (FHFA). The GSEs are currently under Federal conservatorship. In 2013, the Bureau established this category of QMs (Temporary GSE QMs) as a temporary measure that would expire no later than January 10, 2021 or when the GSEs cease to operate under conservatorship. In a final rule released on October 20, 2020, the Bureau extended the Temporary GSE QM loan definition to expire on the mandatory compliance date of final amendments to the General QM loan definition in Regulation Z (or when the GSEs cease to operate under the conservatorship of the FHFA, if that happens earlier). In this final rule, the Bureau adopts the amendments to the General QM loan definition that are referenced in that separate final rule.

DATES:

This final rule is effective on March 1, 2021. However, the mandatory compliance date is July 1, 2021. For additional discussion of these dates, see part VII of the Supplementary Information section below.

FOR FURTHER INFORMATION CONTACT:

Waeiz Syed, Counsel, or Ben Cady, Pedro De Oliveira, Sarita Frattaroli, David Friend, Mark Morelli, Marta Tanenhaus, Priscilla Walton-Fein, or Steve Wrone, Senior Counsels, Office of Regulations, at 202-435-7700. If you require this document in an alternative electronic format, please contact [e-mail].

SUPPLEMENTARY INFORMATION:

I. Summary of the Final Rule

The Ability-to-Repay/Qualified Mortgage Rule (ATR/QM Rule) requires a creditor to make a reasonable, good faith determination of a consumer's ability to repay a residential mortgage loan according to its terms. Loans that meet the ATR/QM Rule's requirements for QMs obtain certain protections from liability. The ATR/QM Rule defines several categories of QMs.

One QM category defined in the ATR/QM Rule is the General QM category. General QMs must comply with the ATR/QM Rule's prohibitions on certain loan features, its points-and-fees limits, and its underwriting requirements. For General QMs, the consumer's DTI ratio must not exceed 43 percent. The ATR/QM Rule requires that creditors must calculate, consider, and verify debt and income for purposes of determining the consumer's DTI ratio using the standards contained in appendix Q of Regulation Z.

A second, temporary category of QMs defined in the ATR/QM Rule consists of mortgages that (1) comply with the same loan-feature prohibitions and points-and-fees limits as General QMs and (2) are eligible to be purchased or guaranteed by the GSEs while under the conservatorship of the FHFA. This final rule refers to these loans as Temporary GSE QMs, and the provision that created this loan category is commonly known as the GSE Patch. Unlike for General QMs, the ATR/QM Rule does not prescribe a DTI limit for Temporary GSE QMs. Thus, a loan can qualify as a Temporary GSE QM even if the consumer's DTI ratio exceeds 43 percent, as long as the loan is eligible to be purchased or guaranteed by either of the GSEs and satisfies the other Temporary GSE QM requirements. In addition, for Temporary GSE QMs, the ATR/QM Rule does not require creditors to use appendix Q to determine the consumer's income, debt, or DTI ratio.

In 2013, the Bureau provided in the ATR/QM Rule that the Temporary GSE QM loan definition would expire with respect to each GSE when that GSE ceases to operate under Federal conservatorship or on January 10, 2021, whichever comes first. The GSEs are currently under Federal conservatorship. Despite the Bureau's expectations when the ATR/QM Rule was published in 2013, Temporary GSE QM originations continue to represent a large and persistent share of the residential mortgage loan market. Without changes to the General QM loan definition, a significant number of Temporary GSE QMs would not be made or would be made at higher prices when the Temporary GSE QM loan definition expires. The affected loans would include loans for which the consumer's DTI ratio is above 43 percent or the creditor's method of documenting and verifying income or debt is incompatible with appendix Q. Based on 2018 data, the Bureau estimates that, as a result of the General QM loan definition's 43 percent DTI limit, approximately 957,000 loans—16 percent of all closed-end first-lien residential mortgage originations in 2018—would be affected by the expiration of the Temporary GSE QM loan definition. These loans are currently originated as QMs due to the Temporary GSE QM loan definition but would not be originated under the current General QM loan definition, and might not be originated at all, if the Temporary GSE QM loan definition were to expire.

On June 22, 2020, the Bureau released two proposed rules concerning the ATR/QM Rule; these proposed rules were published in the Federal Register on July 10, 2020. In one of the proposals—referred to in this final rule as the Extension Proposal—the Bureau proposed to extend the Temporary GSE QM loan definition until the effective date of a final rule issued by the Bureau amending the General QM loan definition.^[1] The other proposal concerned the issues addressed in this final rule. In that proposal—referred to in this final rule as the General QM Proposal or as the proposal—the Bureau proposed amendments to the General QM loan definition.^[2] In the General QM Proposal, the Bureau proposed, among other things, to remove the General QM loan definition's DTI limit and replace it with a limit based on the loan's pricing. The Bureau stated that it expected such amendments would allow most loans that currently could receive QM status under the Temporary GSE QM loan definition to receive QM status under the General QM loan definition if they are made after the Temporary GSE QM loan definition expires. Based on 2018 data, the Bureau estimated in the General QM Proposal that 943,000 conventional loans with DTI ratios above 43 percent would fall outside the QM definitions if there are no changes to the General QM loan definition before the expiration of the Temporary GSE QM loan definition but would fall within the General QM loan definition if it were amended as the Bureau proposed. The Bureau stated that, as a result, the General QM Proposal would help to facilitate a smooth and orderly transition away from the Temporary GSE QM loan definition.

On August 18, 2020, the Bureau issued a third proposal concerning the ATR/QM Rule. In that proposal—referred to in this final rule as the Seasoned QM Proposal—the Bureau proposed to create a new category of QMs (Seasoned QMs) for first-lien, fixed-rate covered transactions that meet certain performance requirements over a 36-month seasoning period, are held in portfolio until the end of the seasoning period, comply with general restrictions on product features and points and fees, and meet certain underwriting requirements.^[3]

In a final rule released on October 20, 2020 (the Extension Final Rule), the Bureau amended Regulation Z to replace the January 10, 2021 sunset date of the Temporary GSE QM loan definition with a provision stating that the Temporary GSE QM loan definition will be available only for covered transactions for which the creditor receives the consumer's application before the mandatory compliance date of final amendments to the General QM loan definition in Regulation Z. The Extension Final Rule did not amend the provision stating that the Temporary GSE QM loan definition expires with respect to a GSE when that GSE ceases to operate under conservatorship (the conservatorship clause). The Extension Final Rule did not affect the QM definitions that apply to Federal Housing Administration (FHA), U.S. Department of Veterans Affairs (VA), U.S. Department of Agriculture (USDA), or Rural Housing Service (RHS) loans.

In this final rule, the Bureau amends Regulation Z to replace the existing General QM loan definition with its 43 percent DTI limit with a price-based General QM loan definition. Under the final rule, a loan meets the General QM loan definition in § 1026.43(e)(2) only if the annual percentage rate (APR) exceeds the average prime offer rate (APOR) for a comparable transaction by less than 2.25 percentage points as of the date the interest rate is set. The final rule provides higher thresholds for loans with smaller loan amounts, for certain manufactured housing loans, and for subordinate-lien transactions. The final rule retains the existing product-feature and underwriting requirements and limits on points and fees. Although the final rule removes the 43 percent DTI limit from the General QM loan definition, the final rule requires that the creditor consider the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan, debt obligations, alimony, child support, and DTI ratio or residual income and verify the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan and the consumer's current debt obligations, alimony, and child support. The final rule removes appendix Q. To prevent uncertainty that may result from appendix Q's removal, the final rule clarifies the consider and verify requirements. The final rule preserves the current threshold separating safe harbor from rebuttable presumption QMs, under which a loan is a safe harbor QM if its APR does not exceed APOR for a comparable transaction by 1.5 percentage points or more as of the date the interest rate is set (or by 3.5 percentage points or more for subordinate-lien transactions).

The effective date of this final rule is March 1, 2021, and the mandatory compliance date is July 1, 2021. Creditors will have the option of complying with the revised General QM loan definition for covered transactions for which creditors receive an application on or after March 1, 2021, and before July 1, 2021. The revised regulations apply to covered transactions for which creditors receive an application on or after July 1, 2021.

II. Background

A. Dodd-Frank Act Amendments to the Truth in Lending Act

The Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank Act) ^[4] amended the Truth in Lending Act (TILA) ^[5] to establish, among other things, ability-to-repay (ATR) requirements in connection with the origination of most residential mortgage loans.^[6] The amendments were intended “to assure that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loans and that are understandable and not unfair, deceptive or abusive.” ^[7] As amended, TILA prohibits a creditor from making a residential mortgage loan unless the creditor makes a reasonable and good faith determination based on verified and documented information that the consumer has a reasonable ability to repay the loan.^[8]

TILA identifies the factors a creditor must consider in making a reasonable and good faith assessment of a consumer's ability to repay. These factors are the consumer's credit history, current and expected income, current obligations, DTI ratio or residual income after paying non-mortgage debt and mortgage-related obligations, employment status, and other financial resources other than equity in the dwelling or real property that secures repayment of the loan.^[9] A creditor, however, may not be certain whether its ATR determination is reasonable in a particular case.

TILA addresses this potential uncertainty by defining a category of loans—called QMs—for which a creditor “may presume that the loan has met” the ATR requirements.^[10] The statute generally defines a QM to mean any residential mortgage loan for which:

• The loan does not have negative amortization, interest-only payments, or balloon payments;
• The loan term does not exceed 30 years;
• The total points and fees generally do not exceed 3 percent of the loan amount;
• The income and assets relied upon for repayment are verified and documented;
• The underwriting uses a monthly payment based on the maximum rate during the first five years, uses a payment schedule that fully amortizes the loan over the loan term, and takes into account all mortgage-related obligations; and
• The loan complies with any guidelines or regulations established by the Bureau relating to the ratio of total monthly debt to monthly income or alternative measures of ability to pay regular expenses after payment of total monthly debt.^[11]

B. The ATR/QM Rule

In January 2013, the Bureau issued a final rule amending Regulation Z to implement TILA's ATR requirements (January 2013 Final Rule).^[12] The January 2013 Final Rule became effective on January 10, 2014, and the Bureau has amended it several times since January 2013.^[13] This final rule refers to the January 2013 Final Rule and later amendments to it collectively as the ATR/QM Rule or the Rule. The ATR/QM Rule implements the statutory ATR provisions discussed above and defines several categories of QMs.^[14]

1. General QMs

One category of QMs defined by the ATR/QM Rule consists of General QMs.^[15] A loan is a General QM if:

• The loan does not have negative-amortization, interest-only, or balloon-payment features, a term that exceeds 30 years, or points and fees that exceed specified limits; ^[16]

• The creditor underwrites the loan based on a fully amortizing schedule using the maximum rate permitted during the first five years; ^[17]

• The creditor considers and verifies the consumer's income and debt obligations in accordance with appendix Q; ^[18] and

• The consumer's DTI ratio is no more than 43 percent, determined in accordance with appendix Q.^[19]

Appendix Q contains standards for calculating and verifying debt and income for purposes of determining whether a mortgage satisfies the 43 percent DTI limit for General QMs. The standards in appendix Q were adapted from guidelines maintained by FHA when the January 2013 Final Rule was issued.^[20] Appendix Q addresses how to determine a consumer's employment-related income (e.g., income from wages, commissions, and retirement plans); non-employment related income (e.g., income from alimony and child support payments, investments, and property rentals); and liabilities, including recurring and contingent liabilities and projected obligations.^[21]

2. Temporary GSE QMs

A second, temporary category of QMs defined by the ATR/QM Rule, Temporary GSE QMs, consists of mortgages that (1) comply with the ATR/QM Rule's prohibitions on certain loan features and its limitations on points and fees ^[22] and (2) are eligible to be purchased or guaranteed by either GSE while under the conservatorship of the FHFA.^[23] Unlike for General QMs, Regulation Z does not prescribe a DTI limit for Temporary GSE QMs. Thus, a loan can qualify as a Temporary GSE QM even if the DTI ratio exceeds 43 percent, as long as the DTI ratio meets the applicable GSE's DTI requirements and other underwriting criteria, and the loan satisfies the other Temporary GSE QM requirements. In addition, income, debt, and DTI ratios for such loans generally are verified and calculated using GSE standards, rather than appendix Q. The January 2013 Final Rule provided that the Temporary GSE QM loan definition—also known as the GSE Patch—would expire with respect to each GSE when that GSE ceases to operate under conservatorship or on January 10, 2021, whichever comes first.^[24]

In the January 2013 Final Rule, the Bureau explained why it created the Temporary GSE QM loan definition. The Bureau observed that it did not believe that a 43 percent DTI ratio “represents the outer boundary of responsible lending” and acknowledged that historically, and even after the financial crisis, over 20 percent of mortgages exceeded that threshold.^[25] However, the Bureau stated that, as DTI ratios increase, the general ATR procedures, rather than the QM framework, are “better suited for consideration of all relevant factors that go to a consumer's ability to repay a mortgage loan” and that “[o]ver the long term . . . there will be a robust and sizable market for prudent loans beyond the 43 percent threshold even without the benefit of the presumption of compliance that applies to qualified mortgages.” ^[26]

At the same time, the Bureau noted that the mortgage market was especially fragile following the financial crisis, and GSE-eligible loans and federally insured or guaranteed loans made up a significant majority of the market.^[27] The Bureau believed that it was appropriate to consider for a period of time, and while the GSEs were under Federal conservatorship, that GSE-eligible loans were originated with an appropriate assessment of the consumer's ability to repay and therefore warranted being treated as QMs.^[28] The Bureau believed in 2013 that this temporary category of QMs would, in the near term, help to ensure access to responsible, affordable credit for consumers with DTI ratios above 43 percent, as well as facilitate compliance by creditors by promoting the use of widely recognized, federally related underwriting standards.^[29]

In making the Temporary GSE QM loan definition temporary, the Bureau sought to “provide an adequate period for economic, market, and regulatory conditions to stabilize” and “a reasonable transition period to the general qualified mortgage definition.” ^[30] The Bureau believed that the Temporary GSE QM loan definition would benefit consumers by preserving access to credit while the mortgage industry adjusted to the ATR/QM Rule.^[31] The Bureau also explained that it structured the Temporary GSE QM loan definition to cover loans eligible to be purchased or guaranteed by either of the GSEs—regardless of whether the loans are actually purchased or guaranteed—to leave room for non-GSE private investors to return to the market and secure the same legal protections as the GSEs.^[32] The Bureau believed that, as the market recovered, the GSEs and the Federal agencies would be able to reduce their market presence, the percentage of Temporary GSE QMs would decrease, and the market would shift toward General QMs and non-QM loans above a 43 percent DTI ratio.^[33] The Bureau's view was that a shift towards non-QM loans could be supported by the non-GSE private market—i.e., by institutions holding such loans in portfolio, selling them in whole, or securitizing them in a rejuvenated private-label securities (PLS) market. The Bureau noted that, pursuant to its statutory obligations under the Dodd-Frank Act, it would assess the impact of the ATR/QM Rule five years after the ATR/QM Rule's effective date, and the assessment would provide an opportunity to analyze the Temporary GSE QM loan definition.^[34]

3. Presumption of Compliance for QMs

In the January 2013 Final Rule, the Bureau considered whether QMs should receive a conclusive presumption (i.e., a safe harbor) or a rebuttable presumption of compliance with the ATR requirements. The Bureau concluded that the statute is ambiguous as to whether a creditor originating a QM receives a safe harbor or a rebuttable presumption that it has complied with the ATR requirements.^[35] The Bureau noted that its analysis of the statutory construction and policy implications demonstrated that there are sound reasons for adopting either interpretation.^[36] The Bureau concluded that the statutory language does not mandate either interpretation and that the presumptions should be tailored to promote the policy goals of the statute.^[37] The Bureau ultimately interpreted the statute to provide for a rebuttable presumption of compliance with the ATR requirements but used its adjustment authority to establish a conclusive presumption of compliance for loans that are not “higher-priced.” ^[38]

Under the ATR/QM Rule, a creditor that makes a QM is protected from liability presumptively or conclusively, depending on whether the loan is “higher-priced.” The ATR/QM Rule generally defines a “higher-priced” loan to mean a first-lien mortgage with an APR that exceeded APOR for a comparable transaction as of the date the interest rate was set by 1.5 or more percentage points; or a subordinate-lien mortgage with an APR that exceeded APOR for a comparable transaction as of the date the interest rate was set by 3.5 or more percentage points.^[39] A creditor that makes a QM that is not “higher-priced” is entitled to a conclusive presumption that it has complied with the ATR/QM Rule—i.e., the creditor receives a safe harbor from liability.^[40] A creditor that makes a loan that meets the standards for a QM but is “higher-priced” is entitled to a rebuttable presumption that it has complied with the ATR/QM Rule.^[41]

The Bureau explained in the January 2013 Final Rule why it was adopting different presumptions of compliance based on the pricing of QMs.^[42] The Bureau noted that the line it was drawing is one that has long been recognized as a rule of thumb to separate prime loans from subprime loans.^[43] The Bureau noted that loan pricing is calibrated to the risk of the loan and that the historical performance of prime and subprime loans indicates greater risk for subprime loans.^[44] The Bureau also noted that consumers taking out subprime loans tend to be less sophisticated and have fewer options and that the most abuses prior to the financial crisis occurred in the subprime market.^[45] The Bureau concluded that these factors warrant imposing heightened standards for higher-priced loans.^[46] For prime loans, however, the Bureau found that lower rates are indicative of ability to repay and noted that prime loans have performed significantly better than subprime loans.^[47] The Bureau concluded that if a loan met the product and underwriting requirements for QMs and was not a higher-priced loan, there are sufficient grounds for concluding that the creditor satisfied the ATR requirements.^[48] The Bureau noted that the conclusive presumption may reduce uncertainty and litigation risk and may promote enhanced competition in the prime market.^[49] The Bureau also noted that the litigation risk for rebuttable presumption QMs likely would be quite modest and would have a limited impact on access to credit.^[50]

The Bureau also noted in the January 2013 Final Rule that policymakers have long relied on pricing to determine which loans should be subject to additional regulatory requirements.^[51] That history of reliance on pricing continues to provide support for a price-based approach to the General QM loan definition. For example, in 1994 Congress amended TILA by enacting the Home Ownership and Equity Protection Act (HOEPA) as part of the Riegle Community Development and Regulatory Improvement Act of 1994.^[52] HOEPA was enacted as an amendment to TILA to address abusive practices in refinancing and home-equity mortgage loans with high interest rates or high fees.^[53] The statute applied generally to closed-end mortgage credit but excluded purchase money mortgage loans and reverse mortgages. Coverage was triggered if a loan's APR exceeded comparable Treasury securities by specified thresholds for particular loan types, or if points and fees exceeded 8 percent of the total loan amount or a dollar threshold.^[54] For high-cost loans meeting either of those thresholds, HOEPA required creditors to provide special pre-closing disclosures, restricted prepayment penalties and certain other loan terms, and regulated various creditor practices, such as extending credit without regard to a consumer's ability to repay the loan. HOEPA also created special substantive protections for high-cost mortgages, such as prohibiting a creditor from engaging in a pattern or practice of extending a high-cost mortgage to a consumer based on the consumer's collateral without regard to the consumer's repayment ability, including the consumer's current and expected income, current obligations, and employment.^[55] The Board implemented the HOEPA amendments at §§ 226.31, 226.32, and 226.33 ^[56] of Regulation Z (12 CFR part 226).^[57]

In 2001, the Board issued rules expanding HOEPA's protections to more loans by revising the APR threshold for first-lien mortgage loans and revising the ATR provisions to provide for a presumption of a violation of the rule if the creditor engages in a pattern or practice of making high-cost mortgages without verifying and documenting the consumer's repayment ability.

In 2008, the Board exercised its authority under HOEPA to extend certain protections concerning a consumer's ability to repay and prepayment penalties to a new category of “higher-priced mortgage loans” (HPMLs) ^[58] with APRs that are lower than those prescribed for high-cost loans but that nevertheless exceed the APOR by prescribed amounts. This new category of loans was designed to include subprime credit, including subprime purchase money mortgage loans. Specifically, the Board exercised its authority to revise HOEPA's restrictions on high-cost loans based on its conclusion that the revisions were necessary to prevent unfair and deceptive acts or practices in connection with mortgage loans.^[59] The Board concluded that a prohibition on making individual loans without regard to repayment ability was necessary to ensure a remedy for consumers who are given unaffordable loans and to deter irresponsible lending. The 2008 HOEPA Final Rule provided a presumption of compliance with the higher-priced mortgage ability-to-repay requirements if the creditor follows certain procedures regarding underwriting the loan payment, assessing the DTI ratio or residual income, and limiting the features of the loan, in addition to following certain procedures mandated for all creditors.^[60] However, the 2008 HOEPA Final Rule made clear that even if the creditor follows the required and optional criteria, the creditor obtained a presumption (not a safe harbor) of compliance with the repayment ability requirement. The consumer therefore could still rebut or overcome that presumption by showing that, despite following the required and optional procedures, the creditor nonetheless disregarded the consumer's ability to repay the loan.

C. The Bureau's Assessment of the ATR/QM Rule

Section 1022(d) of the Dodd-Frank Act requires the Bureau to assess each of its significant rules and orders and to publish a report of each assessment within five years of the effective date of the rule or order.^[61] In June 2017, the Bureau published a request for information in connection with its assessment of the ATR/QM Rule (Assessment RFI).^[62] These comments are summarized in general terms in part III below.

In January 2019, the Bureau published its ATR/QM Rule Assessment Report.^[63] The Assessment Report included findings about the effects of the ATR/QM Rule on the mortgage market generally, as well as specific findings about Temporary GSE QM originations.

The Assessment Report found that loans with higher DTI ratios have been associated with higher levels of “early delinquency” (i.e., delinquency within two years of origination), which the Bureau used as a proxy for measuring consumer repayment ability at consummation across a wide pool of loans.^[64] The Assessment Report also found that the ATR/QM Rule did not eliminate access to credit for consumers with DTI ratios above 43 percent who qualify for Temporary GSE QMs.^[65] On the other hand, based on application-level data obtained from nine large lenders, the Assessment Report found that the ATR/QM Rule eliminated between 63 and 70 percent of home purchase loans with DTI ratios above 43 percent that were not Temporary GSE QMs.^[66]

One main finding about Temporary GSE QMs was that such loans continued to represent a “large and persistent” share of originations in the conforming segment of the mortgage market.^[67] As discussed, the GSEs' share of the conventional, conforming purchase-mortgage market was large before the ATR/QM Rule, and the Assessment found a small increase in that share since the ATR/QM Rule's effective date, reaching 71 percent in 2017.^[68] The Assessment Report noted that, at least for loans intended for sale in the secondary market, creditors generally offer a Temporary GSE QM even if a General QM could be originated.^[69]

The continued prevalence of Temporary GSE QM originations is contrary to the Bureau's expectation at the time it issued the ATR/QM Rule in 2013.^[70] The Assessment Report discussed several possible reasons for the continued prevalence of Temporary GSE QM originations. The Assessment Report first highlighted commenters' concerns with the perceived lack of clarity in appendix Q and found that such concerns “may have contributed to investors'—and at least derivatively, creditors'—preference” for Temporary GSE QMs instead of originating loans under the General QM loan definition.^[71] In addition, the Bureau has not revised appendix Q since 2013, while other standards for calculating and verifying debt and income have been updated more frequently.^[72]

The Assessment Report noted that a second possible reason for the continued prevalence of Temporary GSE QMs is that the GSEs were able to accommodate the demand for mortgages above the General QM loan definition's DTI limit of 43 percent as the DTI ratio distribution in the market shifted upward.^[73] According to the Assessment Report, in the years since the ATR/QM Rule took effect, house prices have increased and consumers hold more mortgage and other debt (including student loan debt), all of which have caused the DTI ratio distribution to shift upward.^[74] The Assessment Report noted that the share of GSE home purchase loans with DTI ratios above 43 percent has increased since the ATR/QM Rule took effect in 2014.^[75] The available data suggest that the share of loans with DTI ratios above 43 percent has declined in the non-GSE market relative to the GSE market.^[76] The non-GSE market has constricted even with respect to highly qualified consumers; those with higher incomes and higher credit scores represent a greater share of denials.^[77]

The Assessment Report found that a third possible reason for the persistence of Temporary GSE QMs is the structure of the secondary market.^[78] If creditors adhere to the GSEs' guidelines, they gain access to a robust, highly liquid secondary market.^[79] In contrast, the Assessment Report noted that while private market securitizations had grown somewhat in recent years, their volume was still a fraction of their pre-crisis levels.^[80] There were less than $20 billion in new origination PLS issuances in 2017, compared with $1 trillion in 2005,^[81] and only 21 percent of new origination PLS issuances in 2017 were non-QM issuances.^[82] To the extent that private securitizations have occurred since the ATR/QM Rule took effect in 2014, the majority of new origination PLS issuances have consisted of prime jumbo loans made to consumers with strong credit characteristics, and these securities include a small share of non-QM loans.^[83] The Assessment Report noted that the Temporary GSE QM loan definition may itself be inhibiting the growth of the non-QM market.^[84] However, the Assessment Report also noted that it is possible that this market might not exist even with a narrower Temporary GSE QM loan definition, if consumers were unwilling to pay the premium charged to cover the potential litigation risk associated with non-QM loan (which do not have a presumption of compliance with the ATR requirements) or if creditors were unwilling or lack the funding to make the loans.^[85]

D. Effects of the COVID-19 Pandemic on Mortgage Markets

The COVID-19 pandemic has had a significant effect on the U.S. economy. In the early months of the pandemic, economic activity contracted, millions of workers became unemployed, and mortgage markets were affected. In recent months, the unemployment rate has declined and there has been a significant rebound in mortgage-origination activity, buoyed by historically low interest rates and by an increasingly large share of government and GSE-backed loans. However, origination activity outside the government and GSE-backed origination channels has declined, and mortgage-credit availability for many consumers—including those who would be dependent on the non-QM market for financing—remains tight. The pandemic's impact on both the secondary market for new originations and on the servicing of existing mortgages is described below.

1. Secondary Market Impacts and Implications for Mortgage Origination Markets

The early economic disruptions associated with the COVID-19 pandemic restricted the flow of credit in the U.S. economy, particularly as uncertainty rose in mid-March 2020, and investors moved rapidly towards cash and government securities.^[86] The lack of investor demand to purchase mortgages, combined with a large supply of agency mortgage-backed securities (MBS) entering the market,^[87] resulted in widening spreads between the rates on a 10-year Treasury note and mortgage interest rates.^[88] This dynamic made it difficult for creditors to originate loans, as many creditors rely on the ability to profitably sell loans in the secondary market to generate the liquidity to originate new loans. This resulted in mortgages becoming more expensive for both homebuyers and homeowners looking to refinance. After the actions taken by the Board in March 2020 to purchase agency MBS “in the amounts needed to support smooth market functioning and effective transmission of monetary policy to broader financial conditions and the economy,” ^[89] market conditions have improved substantially.^[90] This has helped to tighten interest rate spreads, which stabilizes mortgage rates, resulting in a decline in mortgage rates since the Board's intervention and in a significant increase in refinance activity.

However, non-agency MBS ^[91] are generally perceived by investors as riskier than agency MBS. As a result, private capital has remained tight and non-agency mortgage credit, including non-QM lending, has declined. Issuance of non-agency MBS declined by 8.2 percent in the first quarter of 2020, with nearly all the transactions completed in January and February before the COVID-19 pandemic began to affect the economy significantly.^[92] Nearly all major non-QM creditors ceased making loans in March and April 2020. Beginning in May 2020, issuers of non-agency MBS began to test the market with deals collateralized by non-QM loans largely originated prior to the pandemic, and investor demand for these securitizations has begun to recover. However, no securitization has been completed that is predominantly collateralized by non-QM loans originated since the pandemic began.^[93] Many non-QM creditors—which largely depend on the ability to sell loans in the secondary market in order to fund new loans—have begun to resume originations, albeit with tighter underwriting requirements.^[94] Prime jumbo financing also dropped nearly 22 percent in the first quarter of 2020.^[95] Banks increased interest rates and narrowed the product offerings such that only consumers with pristine credit profiles were eligible, as these loans must be held in portfolio when the secondary market for non-agency MBS contracts, and volume remains flat.^[96]

Despite the recent gains in both the agency and the non-agency mortgage sectors, the GSEs continue to play a dominant role in the market recovery, with the GSE share of first-lien mortgage originations at 61.9 percent in the third quarter of 2020, up from 45.3 percent in the third quarter of 2019. The FHA and VA share declined slightly to 17.4 percent from 19.5 percent a year prior, according to an analysis by the Urban Institute. Portfolio lending declined to 19.6 percent in the third quarter of 2020, down from 33.3 percent in the third quarter of 2019, and private label securitizations declined to 1 percent from 1.8 percent a year prior.^[97]

2. Servicing Market Impacts and Implications for Origination Markets

In addition to the direct impact on origination volume and composition, the pandemic's impact on the mortgage servicing market has downstream effects on mortgage originations as many of the same entities both originate and service mortgages. Anticipating that a number of homeowners would struggle to pay their mortgages due to the pandemic and related economic impacts, Congress passed and the President signed into law the Coronavirus Aid, Relief, and Economic Security Act (CARES Act) ^[98] in March 2020. The CARES Act provides additional protections for borrowers with federally backed mortgages, such as those whose mortgages are purchased or securitized by a GSE or insured or guaranteed by the FHA, VA or USDA. The CARES Act mandated a 60-day foreclosure moratorium for such mortgages, which has since been extended by the agencies until the end of 2020 or January 31, 2021 in the case of the GSEs.^[99] The CARES Act also allows borrowers with federally backed mortgages to request up to 180 days of forbearance due to a COVID-19-related financial hardship, with an option to extend the forbearance period for an additional 180 days.

Following the passage of the CARES Act, some mortgage servicers remain obligated to make some principal and interest payments to investors in GSE and Ginnie Mae securities, even if consumers are not making payments.^[100] Servicers also remain obligated to make escrowed real estate tax and insurance payments to local taxing authorities and insurance companies. While servicers are required to hold liquid reserves to cover anticipated advances, early in the pandemic there were significant concerns that higher-than-expected forbearance rates over an extended period of time could lead to liquidity shortages, particularly among many non-bank servicers. However, while forbearance rates remain elevated at 5.54 percent for the week ending November 22, 2020, they have decreased since reaching their high of 8.55 percent on June 7, 2020.^[101]

Because many mortgage servicers also originate the loans they service, many creditors, as well as several warehouse providers,^[102] initially responded to the risk of elevated forbearances and higher-than-expected monthly advances by imposing credit overlays—i.e., additional underwriting standards—for new originations. These new underwriting standards include more stringent requirements for non-QM, jumbo, and government loans.^[103] An “adverse market fee” of 50 basis points on most refinances became effective for new originations delivered to the GSEs on or after December 1, 2020, to cover projected losses due to forbearances, the foreclosure moratoriums, and other default servicing expenses.^[104] However, due to refinance origination profits resulting from historically low interest rates, the leveling off in forbearance rates, and actions taken at the Federal level to alleviate servicer liquidity pressure,^[105] concerns over non-bank liquidity and related credit overlays have begun to ease, though Federal regulators continue to monitor the situation.^[106] While the non-QM market has begun to recover, it is unclear how quickly non-banks that originate non-QM loans will fully return to their pre-pandemic level of operations and loan production.

III. Summary of the Rulemaking Process

The Bureau has solicited and received substantial public and stakeholder input on issues related to this final rule. In addition to the Bureau's discussions with and communications from industry stakeholders, consumer advocates, other Federal agencies,^[107] and members of Congress, the Bureau issued requests for information (RFIs) in 2017 and 2018 and in July 2019 issued an advance notice of proposed rulemaking regarding the ATR/QM Rule (ANPR). The Bureau released the Extension Proposal and the General QM Proposal on June 22, 2020, and the Seasoned QM Proposal on August 18, 2020. The Bureau issued the Extension Final Rule on October 20, 2020.

A. The Requests for Information

In June 2017, the Bureau published the Assessment RFI to gather information for its assessment of the ATR/QM Rule.^[108] In response to the Assessment RFI, the Bureau received approximately 480 comments from creditors, industry groups, consumer advocates, and individuals.^[109] The comments addressed a variety of topics, including the General QM loan definition and the 43 percent DTI limit; perceived problems with, and potential changes and alternatives to, appendix Q; and how the Bureau should address the expiration of the Temporary GSE QM loan definition. The comments expressed a range of ideas for addressing the expiration of the Temporary GSE QM loan definition. Some commenters recommended making the definition permanent or extending it for various periods of time. Other comments stated that the Temporary GSE QM loan definition should be eliminated or permitted to expire.

Beginning in January 2018, the Bureau issued a general call for evidence seeking comment on its enforcement, supervision, rulemaking, market monitoring, and financial education activities.^[110] As part of the call for evidence, the Bureau published requests for information relating to, among other things, the Bureau's rulemaking process,^[111] the Bureau's adopted regulations and new rulemaking authorities,^[112] and the Bureau's inherited regulations and inherited rulemaking authorities.^[113] In response to the call for evidence, the Bureau received comments on the ATR/QM Rule from stakeholders, including consumer advocates and industry groups. The comments addressed a variety of topics, including the General QM loan definition, appendix Q, and the Temporary GSE QM loan definition. The comments also raised concerns about, among other things, the risks of allowing the Temporary GSE QM loan definition to expire without any changes to the General QM loan definition or appendix Q. The concerns raised in these comments were similar to those raised in response to the Assessment RFI, discussed above.

B. The ANPR

On July 25, 2019, the Bureau issued the ANPR. The ANPR stated the Bureau's tentative plans to allow the Temporary GSE QM loan definition to expire in January 2021 or after a short extension, if necessary, to facilitate a smooth and orderly transition away from the Temporary GSE QM loan definition. The Bureau also stated that it was considering whether to propose revisions to the General QM loan definition in light of the potential expiration of the Temporary GSE QM loan definition and requested comments on several topics related to the General QM loan definition, including whether and how the Bureau should revise the DTI limit in the General QM loan definition; whether the Bureau should supplement or replace the DTI limit with another method for directly measuring a consumer's personal finances; whether the Bureau should revise appendix Q or replace it with other standards for calculating and verifying a consumer's debt and income; and whether, instead of a DTI limit, the Bureau should adopt standards that do not directly measure a consumer's personal finances.^[114] The Bureau requested comment on how much time industry would need to change its practices in response to any changes the Bureau might make to the General QM loan definition.^[115] The Bureau received approximately 85 comments on the ANPR from businesses in the mortgage industry (including creditors), consumer advocates, elected officials, individuals, and research centers. The General QM Proposal provided a summary of these comments, and the Bureau considered these comments in developing the proposal.

C. The Extension Proposal, General QM Proposal, and Seasoned QM Proposal

The Bureau issued the Extension Proposal and the General QM Proposal on June 22, 2020, and those proposals were published in the Federal Register on July 10, 2020. In the Extension Proposal, the Bureau proposed to replace the January 10, 2021 sunset date of the Temporary GSE QM loan definition with a provision that extends the Temporary GSE QM loan definition until the effective date of final amendments to the General QM loan definition in Regulation Z (i.e., a final rule relating to the General QM Proposal). The Bureau did not propose to amend the conservatorship clause. The comment period for the Extension Proposal ended on August 10, 2020.

In the General QM Proposal, the Bureau proposed, among other things, to remove the General QM loan definition's DTI limit and replace it with a limit based on the loan's pricing. Under the proposal, a loan would have met the General QM loan definition in § 1026.43(e)(2) only if the APR exceeds APOR for a comparable transaction by less than 2 percentage points as of the date the interest rate is set. The Bureau proposed higher thresholds for loans with smaller loan amounts and subordinate-lien transactions. The Bureau also proposed to retain the existing product-feature and underwriting requirements and limits on points and fees. Although the Bureau proposed to remove the 43 percent DTI limit from the General QM loan definition, the General QM Proposal would have required that the creditor consider the consumer's DTI ratio or residual income, income or assets other than the value of the dwelling, and debts and verify the consumer's income or assets other than the value of the dwelling and the consumer's debts. The Bureau proposed to remove appendix Q. To mitigate the uncertainty that may result from appendix Q's removal, the General QM Proposal would have clarified the consider and verify requirements. The Bureau proposed to preserve the current threshold separating safe harbor from rebuttable presumption QMs, under which a loan is a safe harbor QM if its APR does not exceed APOR for a comparable transaction by 1.5 percentage points or more as of the date the interest rate is set (or by 3.5 percentage points or more for subordinate-lien transactions).

Although the Bureau proposed to remove the 43 percent DTI limit and adopt a price-based approach for the General QM loan definition, the Bureau also requested comment on two alternative approaches: (1) Retaining the DTI limit and increasing it to a Specific threshold between 45 percent and 48 percent or (2) using a hybrid approach involving both pricing and a DTI limit, such as applying a DTI limit to loans that are above specified rate spreads. Under these alternative approaches, creditors would not have been required to verify debt and income using appendix Q.

The Bureau stated in the General QM Proposal that the proposed amendments would allow most loans that currently could receive QM status under the Temporary GSE QM loan definition to receive QM status under the General QM loan definition.^[116] The Bureau stated that, as a result, the General QM Proposal would help to facilitate a smooth and orderly transition away from the Temporary GSE QM loan definition. The Bureau proposed that the effective date of a final rule relating to the General QM Proposal would be six months after publication of the final rule in the Federal Register. The revised regulations would have applied to covered transactions for which creditors receive an application on or after this effective date. The comment period for the General QM Proposal ended on September 8, 2020. The Bureau received approximately 75 comments in response to the General QM Proposal from industry, consumer advocates, and others. The Bureau summarizes and responds to these comments in parts V through VIII below.

On August 18, 2020, the Bureau issued the Seasoned QM Proposal, which was published in the Federal Register on August 28, 2020. The Bureau proposed to create a new category of QMs for first-lien, fixed-rate covered transactions that have met certain performance requirements over a 36-month seasoning period, are held in portfolio until the end of the seasoning period, comply with general restrictions on product features and points and fees, and meet certain underwriting requirements.^[117] The Bureau stated that the primary objective of the Seasoned QM Proposal was to ensure access to responsible, affordable mortgage credit by adding a Seasoned QM definition to the existing QM definitions. The Bureau proposed that a final rule relating to the Seasoned QM Proposal would take effect on the same date as a final rule relating to the General QM Proposal. Under the Seasoned QM Proposal—as under the General QM Proposal—the revised regulations would apply to covered transactions for which creditors receive an application on or after this effective date. Thus, due to the 36-month seasoning period, no loan would be eligible to become a Seasoned QM until at least 36 months after the effective date of a final rule relating to the Seasoned QM Proposal. The comment period for the Seasoned QM Proposal ended on October 1, 2020.^[118] The Bureau is issuing the Seasoned QM Final Rule concurrently with this final rule.

D. The Extension Final Rule

The Bureau issued the Extension Final Rule on October 20, 2020. It was published in the Federal Register on October 26, 2020. The Extension Final Rule amended Regulation Z to replace the January 10, 2021 sunset date of the Temporary GSE QM loan definition with a provision stating that the Temporary GSE QM loan definition will be available only for covered transactions for which the creditor receives the consumer's application before the mandatory compliance date of final amendments to the General QM loan definition in Regulation Z. The Extension Final Rule did not amend the conservatorship clause.^[119]

IV. Legal Authority

The Bureau is issuing this final rule pursuant to its authority under TILA and the Dodd-Frank Act. Section 1061 of the Dodd-Frank Act transferred to the Bureau the “consumer financial protection functions” previously vested in certain other Federal agencies, including the Board. The Dodd-Frank Act defines the term “consumer financial protection function” to include “all authority to prescribe rules or issue orders or guidelines pursuant to any Federal consumer financial law, including performing appropriate functions to promulgate and review such rules, orders, and guidelines.” ^[120] Title X of the Dodd-Frank Act (including section 1061), along with TILA and certain subtitles and provisions of title XIV of the Dodd-Frank Act, are Federal consumer financial laws.^[121]

A. TILA

TILA section 105(a). Section 105(a) of TILA directs the Bureau to prescribe regulations to carry out the purposes of TILA and states that such regulations may contain such additional requirements, classifications, differentiations, or other provisions and may further provide for such adjustments and exceptions for all or any class of transactions that the Bureau judges are necessary or proper to effectuate the purposes of TILA, to prevent circumvention or evasion thereof, or to facilitate compliance therewith.^[122] A purpose of TILA is “to assure a meaningful disclosure of credit terms so that the consumer will be able to compare more readily the various credit terms available to him and avoid the uninformed use of credit.” ^[123] Additionally, a purpose of TILA sections 129B and 129C is to assure that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loans and that are understandable and not unfair, deceptive, or abusive.^[124] As discussed in the section-by-section analysis below, the Bureau is issuing certain provisions of this final rule pursuant to its rulemaking, adjustment, and exception authority under TILA section 105(a).

TILA section 129C(b)(2)(A). TILA section 129C(b)(2)(A)(vi) provides the Bureau with authority to establish guidelines or regulations relating to ratios of total monthly debt to monthly income or alternative measures of ability to pay regular expenses after payment of total monthly debt, taking into account the income levels of the borrower and such other factors as the Bureau may determine relevant and consistent with the purposes described in TILA section 129C(b)(3)(B)(i).^[125] As discussed in the section-by-section analysis below, the Bureau is issuing certain provisions of this final rule pursuant to its authority under TILA section 129C(b)(2)(A)(vi).

TILA section 129C(b)(3)(A), (B)(i). TILA section 129C(b)(3)(B)(i) authorizes the Bureau to prescribe regulations that revise, add to, or subtract from the criteria that define a QM upon a finding that such regulations are necessary or proper to ensure that responsible, affordable mortgage credit remains available to consumers in a manner consistent with the purposes of TILA section 129C; or are necessary and appropriate to effectuate the purposes of TILA sections 129B and 129C, to prevent circumvention or evasion thereof, or to facilitate compliance with such sections.^[126] In addition, TILA section 129C(b)(3)(A) directs the Bureau to prescribe regulations to carry out the purposes of section 129C.^[127] As discussed in the section-by-section analysis below, the Bureau is issuing certain provisions of this final rule pursuant to its authority under TILA section 129C(b)(3)(B)(i).

B. Dodd-Frank Act

Dodd-Frank Act section 1022(b). Section 1022(b)(1) of the Dodd-Frank Act authorizes the Bureau to prescribe rules to enable the Bureau to administer and carry out the purposes and objectives of the Federal consumer financial laws, and to prevent evasions thereof.^[128] TILA and title X of the Dodd-Frank Act are Federal consumer financial laws. Accordingly, the Bureau is exercising its authority under Dodd-Frank Act section 1022(b) to prescribe rules that carry out the purposes and objectives of TILA and title X and prevent evasion of those laws.

V. Why the Bureau Is Issuing This Final Rule

The Bureau concludes that this final rule's bright-line pricing thresholds strike the best balance between ensuring consumers' ability to repay and ensuring access to responsible, affordable mortgage credit. The Bureau is amending the General QM loan definition because retaining the existing 43 percent DTI limit would reduce the size of the QM market and likely would lead to a significant reduction in access to responsible, affordable credit when the Temporary GSE QM definition expires. The Bureau continues to believe that General QM status should be determined by a simple, bright-line rule to provide certainty of QM status, and the Bureau concludes that pricing achieves this objective. Furthermore, the Bureau concludes that pricing, rather than a DTI limit, is a more appropriate standard for the General QM loan definition. While not a direct measure of financial capacity, loan pricing is strongly correlated with early delinquency rates, which the Bureau uses as a proxy for repayment ability. The Bureau concludes that conditioning QM status on a specific DTI limit would likely impair access to credit for some consumers for whom it is appropriate to presume their ability to repay their loans at consummation. Although a pricing limit that is set too low could also have this effect, compared to DTI, loan pricing is a more flexible metric because it can incorporate other factors that may also be relevant to determining ability to repay, including credit scores, cash reserves, or residual income. The Bureau concludes that a price-based General QM loan definition is better than the alternatives because a loan's price, as measured by comparing a loan's APR to APOR for a comparable transaction, is a strong indicator of a consumer's ability to repay and is a more holistic and flexible measure of a consumer's ability to repay than DTI alone.

A loan's price is not a direct measure of ability to repay, but the Bureau concludes that it is an effective indirect measure of ability to repay. The final rule amends Regulation Z to provide that a loan would meet the General QM loan definition in § 1026.43(e)(2) only if the APR exceeds APOR for a comparable transaction by less than 2.25 percentage points as of the date the interest rate is set. The Bureau is finalizing a threshold of 2.25 percentage points, an increase from the proposed threshold of 2 percentage points. The Bureau concludes that, for most first-lien covered transactions, a 2.25-percentage-point pricing threshold strikes the best balance between ensuring consumers' ability to repay and ensuring continued access to responsible, affordable mortgage credit. The final rule provides higher thresholds for loans with smaller loan amounts and for subordinate-lien transactions. As described below, the final rule provides an increase from the proposed thresholds for some small manufactured housing loans to ensure consumers have continued access to responsible, affordable credit.

Consistent with the proposal, the Bureau is not amending the existing General QM loan product-feature and underwriting requirements and limits on points and fees. Under the final rule, creditors are required to consider the consumer's DTI ratio or residual income, income or assets other than the value of the dwelling, and debts and verify the consumer's income or assets other than the value of the dwelling and the consumer's debts. The final rule removes the 43 percent DTI ratio limit and appendix Q and clarifies the consider and verify requirements for purposes of the General QM loan definition.

The Bureau is preserving the current threshold separating safe harbor from rebuttable presumption QMs, under which a loan is a safe harbor QM if its APR exceeds APOR for a comparable transaction by less than 1.5 percentage points as of the date the interest rate is set (or by less than 3.5 percentage points for subordinate-lien transactions).

A. Overview of the Existing General QM Loan Definition and the DTI Requirement

TILA section 129C(b)(2) defines QM by limiting certain loan terms and features. The statute generally prohibits a QM from permitting an increase of the principal balance on the loan (negative amortization), interest-only payments, most balloon payments, a term greater than 30 years, and points and fees that exceed a specified threshold. In addition, the statute incorporates limited underwriting criteria that overlap with some elements of the general ATR standard, including prohibiting “no-doc” loans where the creditor does not verify income or assets. TILA does not require DTI ratios to be included in the definition of a QM. Rather, the statute authorizes, but does not require, the Bureau to establish additional criteria relating to monthly DTI ratios, or alternative measures of ability to pay regular expenses after payment of total monthly debt, taking into account the income levels of the consumer and other factors the Bureau determines relevant and consistent with the purposes described in TILA section 129C(b)(3)(B)(i).

In 2011, the Board proposed two alternative approaches to the General QM loan definition to implement the statutory QM requirements.^[129] Proposed Alternative 1 would have included only the statutory QM requirements and would not have incorporated the consumer's DTI ratio, residual income, or other factors from the general ATR standard.^[130] Proposed Alternative 2 would have included the statutory QM requirements and additional factors from the general ATR standard, including a requirement to consider and verify the consumer's DTI ratio or residual income.^[131]

In the January 2013 Final Rule, the Bureau adopted the General QM loan definition in § 1026.43(e)(2). The existing General QM loan definition includes the statutory QM factors and additional factors from the general ATR standard. The existing General QM loan definition also contains a DTI limit of 43 percent. In adopting this approach in the January 2013 Final Rule, the Bureau explained that it believed the General QM loan definition should include a standard for evaluating the consumer's ability to repay, in addition to the product feature restrictions and other requirements that are specified in TILA.^[132]

With respect to DTI, the January 2013 Final Rule noted that DTI ratios are widely used for evaluating a consumer's ability to repay over time because, as the available data showed, DTI ratio correlates with loan performance as measured by delinquency rate.^[133] The January 2013 Final Rule noted that, at a basic level, the lower the DTI ratio, the greater the consumer's ability to pay back a mortgage loan.^[134] The Bureau believed this relationship between the DTI ratio and the consumer's ability to repay applied both under conditions as they exist at consummation and under future changed circumstances, such as increases in payments for adjustable-rate mortgages (ARMs), future reductions in income, and unanticipated expenses and new debts.^[135]

To provide certainty for creditors regarding the loan's QM status, the January 2013 Final Rule contained a specific DTI limit of 43 percent as part of the General QM loan definition. The Bureau stated that a specific DTI limit also provides certainty to assignees and investors in the secondary market, which the Bureau believed would help reduce concerns regarding legal risk and promote credit availability.^[136] The Bureau noted that numerous commenters had highlighted the value of providing objective requirements determined based on information contained in loan files.^[137] To address concerns that creditors may not have adequate certainty about whether a particular loan satisfies the requirements of the General QM loan definition, the Bureau provided definitions of debt and income for purposes of the General QM loan definition in appendix Q.^[138]

The Bureau selected 43 percent as the DTI limit for the General QM loan definition. Based on analysis of data available at the time and comments, the Bureau believed that the 43 percent limit would advance TILA's goals of creditors not extending credit that consumers cannot repay while still preserving consumers' access to credit.^[139] The Bureau acknowledged that there is no specific threshold that separates affordable from unaffordable mortgages; rather, there is a gradual increase in delinquency rates as DTI ratios increase.^[140] Additionally, the Bureau noted that a 43 percent DTI ratio was within the range used by many creditors, generally comported with industry standards and practices for prudent underwriting, and was the threshold used by FHA as its general boundary at the time the Bureau issued the January 2013 Final Rule.^[141] The Bureau noted concerns about setting a higher DTI limit, including concerns that it could allow QM status for mortgages for which there is not a sound reason to presume that the creditor had a reasonable belief in the consumer's ability to repay.^[142] The Bureau was especially concerned about setting a DTI limit higher than 43 percent in the context of QMs that receive a safe harbor from the ATR requirements.^[143] The Bureau was also concerned that a higher DTI limit would result in a QM boundary that substantially covered the entire mortgage market. If that were the case, creditors might be unwilling to make non-QM loans, and the Bureau was concerned that the QM rule would define the limit of credit availability.^[144] The Bureau also suggested that a higher DTI limit might require a corresponding weakening of the strength of the presumption of compliance, which the Bureau believed would largely defeat the point of adopting a higher DTI limit.^[145]

The January 2013 Final Rule also acknowledged concerns about imposing a DTI limit. The Bureau acknowledged that the Board, in issuing the 2011 ATR/QM Proposal, found that DTI ratios may not have significant predictive power, once the effects of credit history, loan type, and loan-to-value (LTV) ratio are considered.^[146] Similarly, the Bureau noted that some commenters responding to the 2011 ATR/QM Proposal suggested that the Bureau should include compensating factors in addition to a specific DTI limit due to concerns about restricting access to credit.^[147] The Bureau acknowledged that a standard that takes into account multiple factors may produce more accurate ability-to-repay determinations, at least in specific cases, but was concerned that incorporating a multi-factor test or compensating factors into the General QM loan definition would undermine the certainty for creditors and the secondary market of whether loans were eligible for QM status.^[148] The Bureau also acknowledged arguments that residual income—generally defined as the monthly income that remains after a consumer pays all personal debts and obligations, including the prospective mortgage—may be a better measure of repayment ability.^[149] However, the Bureau noted that it lacked sufficient data to mandate a bright-line rule based on residual income.^[150] The Bureau anticipated further study of the issue as part of the five-year assessment of the Rule.^[151]

The Bureau acknowledged in the January 2013 Final Rule that the 43 percent DTI limit in the General QM loan definition could restrict access to credit based on market conditions. Among other things, the Bureau expressed concern that, as the mortgage market recovered from the financial crisis, there could be a limited non-QM market, which, in conjunction with the 43 percent DTI limit, could impair access to credit for consumers with DTI ratios over 43 percent.^[152] To preserve access to credit for such consumers while the market recovered, the Bureau adopted the Temporary GSE QM loan definition, which did not include a specific DTI limit. As discussed below, the Temporary GSE QM loan definition continues to play a significant role in ensuring access to credit for consumers.

B. Why the Bureau Is Adopting a Price-Based QM Definition To Replace the General QM Loan Definition's DTI Limit

The Bureau concludes that this final rule's price-based approach best balances consumers' ability to repay with ensuring access to responsible, affordable mortgage credit. The Bureau is amending the General QM definition because retaining the existing 43 percent DTI limit would reduce the size of the QM market and likely would lead to a significant reduction in access to responsible, affordable credit when the Temporary GSE QM definition expires. The Bureau continues to believe that General QM status should be determined by a simple, bright-line rule to provide certainty of QM status, and the Bureau concludes that pricing achieves this objective. The Bureau concludes that a price-based General QM loan definition is better than the alternatives because a loan's price, as measured by comparing a loan's APR to APOR for a comparable transaction, is a strong indicator of a consumer's ability to repay and is a more holistic and flexible measure of a consumer's ability to repay than DTI alone.

1. Considerations Related to the General QM Loan Definition's DTI Limit

The proposal described the Bureau's concerns about the 43 percent DTI limit and its potentially negative effect on access to credit. In particular, the Bureau is concerned that imposing a DTI limit under the General QM loan definition would deny QM status for loans to some consumers for whom it is appropriate to presume ability to repay at consummation and that denying QM status to such loans risks denying consumers access to responsible, affordable credit. The Bureau is concerned that the current approach to DTI ratios as part of the General QM loan definition is not the best approach because it would likely impair some consumers' ability to access responsible and affordable credit. These access-to-credit concerns are especially acute for lower-income and minority consumers.

The proposal noted that a DTI limit may unduly restrict access to credit because it provides an incomplete picture of the consumer's financial capacity. While the Bureau acknowledges that DTI ratios generally correlate with loan performance, as the Bureau found in the January 2013 Final Rule and as shown in recent Bureau analysis described below, the proposal noted that a consumer's DTI ratio is only one way to measure financial capacity and is not necessarily a holistic measure of the consumer's ability to repay. The proposal also noted that the Bureau's own experience and the feedback it has received from stakeholders since issuing the January 2013 Final Rule suggest that imposing a DTI limit as a condition for QM status under the General QM loan definition may be overly burdensome and complex in practice.

As described in the proposal, the Bureau's Assessment Report highlights the tradeoffs of conditioning the General QM loan definition on a DTI limit. The Assessment Report included specific findings about the General QM loan definition's DTI limit, including certain findings related to DTI ratios as probative of a consumer's ability to repay. The Assessment Report found that loans with higher DTI ratios have been associated with higher levels of “early delinquency” (i.e., delinquency within two years of origination), which, as explained below, may serve as a proxy for measuring whether a consumer had a reasonable ability to repay at the time the loan was consummated.^[153] For example, the Assessment Report notes that for all periods and samples studied, a positive relationship between DTI ratios and early delinquency is present and economically meaningful.^[154] The Assessment Report states that higher DTI ratios independently increase expected early delinquency, regardless of other underwriting criteria.^[155]

At the same time, findings from the Assessment Report indicate that the specific 43 percent DTI limit in the current rule has restricted access to credit, particularly in the absence of a robust non-QM market. The report found that, for consumers with DTI ratios above 43 percent who qualify for loans eligible for purchase or guarantee by the GSEs, the Rule has not decreased access to credit.^[156] However, the Assessment Report attributes the fact that the 43 percent DTI limit has not reduced access to credit for such consumers to the existence of the Temporary GSE QM loan definition. The findings in the Assessment Report indicate that there would be some reduction in access to credit for consumers with DTI ratios above 43 percent when the Temporary GSE QM loan definition expires, absent changes to the General QM loan definition. For example, based on application-level data obtained from nine large lenders, the Assessment Report found that the January 2013 Final Rule eliminated between 63 and 70 percent of non-GSE eligible home purchase loans with DTI ratios above 43 percent.^[157] The proposal noted the Bureau's concern about a similar effect for loans with DTI ratios above 43 percent when the Temporary GSE QM loan definition expires. The proposal acknowledged that the Assessment Report's finding, without other information, does not prove or disprove the effectiveness of the DTI limit in achieving the purposes of the January 2013 Final Rule in ensuring consumers' ability to repay the loan. If the denied applicants in fact lacked the ability to repay, then the reduction in approval rates is a consequence consistent with the purposes of the Rule. However, if the denied applicants did have the ability to repay, then these data suggest an unintended consequence of the Rule. This possibility is supported by the fact that other findings in the Assessment Report suggest that applicants for non-GSE eligible loans with DTI ratios above 43 percent are being denied, even though other compensating factors indicate that some of them may have the ability to repay their loans.^[158]

The current condition of the non-QM market heightens the access-to-credit concerns related to the specific 43 percent DTI limit, particularly if such conditions persist after the expiration of the Temporary GSE QM loan definition. The Bureau stated in the January 2013 Final Rule that it believed mortgages that could be responsibly originated with DTI ratios that exceed 43 percent, which historically includes over 20 percent of mortgages, would be made under the general ATR standard.^[159] However, the Assessment Report found that a robust market for non-QM loans above the 43 percent DTI limit has not materialized as the Bureau had predicted. Therefore, there is limited capacity in the non-QM market to provide access to credit after the expiration of the Temporary GSE QM loan definition.^[160] As described above, the non-QM market has been further reduced by the recent economic disruptions associated with the COVID-19 pandemic, with most mortgage credit now available in the QM lending space. The Bureau acknowledges the slow development of the non-QM market since the January 2013 Final Rule took effect and further acknowledges that the recent economic disruptions associated with the COVID-19 pandemic may significantly hinder its development in the near term.

At the time of the January 2013 Final Rule, the Bureau adopted the Temporary GSE loan definition to provide a period for economic, market, and regulatory conditions to stabilize and for a reasonable transition period to the General QM loan definition and non-QM loans above a 43 percent DTI ratio. However, contrary to the Bureau's expectations, lending largely has remained in the Temporary GSE QM space, and a sizable market to support non-QM lending has not yet emerged.^[161] As noted above, the Bureau acknowledges that the recent economic disruptions associated with the COVID-19 pandemic may further hinder the development of the non-QM market, at least in the near term. As noted in the proposal, the Bureau expects that a significant number of Temporary GSE QMs would not qualify as General QMs under the current rule after the Temporary GSE QM loan definition expires, either because they have DTI ratios above 43 percent or because their method of documenting and verifying income or debt is incompatible with appendix Q. Some alternative loan options would still be available to many consumers after the expiration of the Temporary GSE QM loan definition. The proposal, however, emphasized the Bureau's expectation that, with respect to loans that are currently Temporary GSE QMs and would not otherwise qualify as General QMs under the current definition, some would cost materially more for consumers and some would not be made at all.

Based on 2018 data, the Bureau estimated in the proposal that, as a result of the General QM loan definition's 43 percent DTI limit, approximately 957,000 loans—16 percent of all closed-end first-lien residential mortgage originations in 2018—would be affected by the expiration of the Temporary GSE QM loan definition.^[162] These loans are currently originated as QMs due to the Temporary GSE QM loan definition but would not be originated under the current General QM loan definition, and might not be originated at all, if the Temporary GSE QM loan definition were to expire. An additional, smaller number of loans that currently qualify as Temporary GSE QMs may not fall within the General QM loan definition after the expiration of the Temporary GSE QM loan definition because the method used for verifying income or debt would not comply with appendix Q.^[163] As explained in the Extension Final Rule, the Temporary GSE QM loan definition will expire on the mandatory compliance date of this final rule or when GSE conservatorship ends.

As explained in the proposal, the Bureau believes that many loans currently originated under the Temporary GSE QM loan definition would cost materially more or may not be made at all, absent changes to the General QM loan definition. After the Temporary GSE QM loan definition expires, the Bureau expects that many consumers with DTI ratios above 43 percent who would have received a Temporary GSE QM would instead obtain FHA-insured loans since FHA currently insures loans with DTI ratios up to 57 percent.^[164] The number of loans that move to FHA would depend on FHA's willingness and ability to insure such loans, whether FHA continues to treat all loans that it insures as QMs under its own QM rule, and how many loans that would have been originated as Temporary GSE QMs with DTI ratios above 43 percent exceed FHA's loan-amount limit.^[165] For example, the Bureau estimated in the proposal that, in 2018, 11 percent of Temporary GSE QM loans with DTI ratios above 43 percent exceeded FHA's loan-amount limit.^[166] Thus, the Bureau considers that at most 89 percent of loans that would have been Temporary GSE QMs with DTI ratios above 43 percent could move to FHA.^[167] The Bureau expects that loans that would be originated as FHA loans instead of under the Temporary GSE QM loan definition generally would cost materially more for many consumers.^[168] The Bureau expects that some consumers offered FHA loans might choose not to take out a mortgage because of these higher costs.

The proposal explained that it is also possible that some consumers with DTI ratios above 43 percent would be able to obtain loans in the private market.^[169] The number of loans absorbed by the private market would likely depend, in part, on whether actors in the private market would be willing to assume the legal or credit risk associated with funding loans—as non-QM loans or small-creditor portfolio QMs—that would have been Temporary GSE QMs (with DTI ratios above 43 percent) ^[170] and, if so, whether actors in the private market would offer lower prices or better terms.^[171] For example, the Bureau estimated that 55 percent of loans that would have been Temporary GSE QMs (with DTI ratios above 43 percent) in 2018 had credit scores at or above 680 and LTV ratios at or below 80 percent—credit characteristics traditionally considered attractive to actors in the private market.^[172] At the same time, the Assessment Report found there has been limited momentum toward a greater role for private market non-QM loans. It is uncertain how great this role will be in the future,^[173] particularly in the short term due to the economic effects of the COVID-19 pandemic. Finally, the proposal noted that some consumers with DTI ratios above 43 percent who would have sought Temporary GSE QM loans may adapt to changing options and make different choices, such as adjusting their borrowing to result in a lower DTI ratio.^[174] However, some consumers who would have sought Temporary GSE QMs (with DTI ratios above 43 percent) may not obtain loans at all.^[175] For example, based on application-level data obtained from nine large lenders, the Assessment Report found that the January 2013 Final Rule eliminated between 63 and 70 percent of non-GSE-eligible home purchase loans with DTI ratios above 43 percent.^[176]

As noted in the proposal, the Bureau also has particular concerns about the effects of the appendix Q definitions of debt and income on access to credit. The Bureau intended for appendix Q to provide creditors with certainty about the DTI ratio calculation to foster compliance with the General QM loan definition. However, based on extensive stakeholder feedback and the Bureau's own experience, the proposal recognized that appendix Q's definitions of debt and income are rigid and difficult to apply and do not provide the level of compliance certainty that the Bureau anticipated. Stakeholders have reported that these concerns are particularly acute for transactions involving self-employed consumers, consumers with part-time employment, and consumers with irregular or unusual income streams. The proposal expressed concern that the standards in appendix Q could negatively impact access to credit for these consumers, particularly after expiration of the Temporary GSE QM loan definition. The Assessment Report also noted concerns with the perceived lack of clarity in appendix Q and found that such concerns “may have contributed to investors'—and at least derivatively, creditors'—preference” for Temporary GSE QMs.^[177] Appendix Q, unlike other standards for calculating and verifying debt and income, has not been revised since 2013.^[178] The current definitions of debt and income in appendix Q have proven to be complex in practice. In the proposal, the Bureau expressed concerns about other potential approaches to defining debt and income in connection with conditioning QM status on a specific DTI limit.

The current approach to DTI ratios under the General QM loan definition may also stifle innovation in underwriting because it focuses on a single metric, with strict verification standards under appendix Q, which may constrain new approaches to assessing repayment ability. Such innovations include certain new uses of cash flow data and analytics to underwrite mortgage applicants. This emerging technology has the potential to accurately assess consumers' ability to repay using, for example, bank account data that can identify the source and frequency of recurring deposits and payments and identify remaining disposable income. Identifying the remaining disposable income could be a method of assessing the sufficiency of a consumer's residual income and could potentially satisfy a requirement to consider either DTI or residual income. This innovation could potentially expand access to responsible, affordable mortgage credit, particularly for applicants with non-traditional income and limited credit history. The proposal expressed concern that the potential negative effect of the current General QM loan definition on innovation in underwriting may be heightened while the market is largely concentrated in the QM lending space and may limit access to credit for some consumers with DTI ratios above 43 percent.

2. The Proposed Price-Based General QM Loan Definition

In light of these concerns, the Bureau proposed to remove the 43 percent DTI limit from the General QM loan definition in § 1026.43(e)(2)(vi) and replace it with a requirement based on the price of the loan. In issuing the proposal, the Bureau acknowledged the significant debate ^[179] over whether loan pricing, a consumer's DTI ratio, or another direct or indirect measure of a consumer's personal finances is a better predictor of loan performance, particularly when analyzed across various points in the economic cycle. In seeking comments on the proposal, the Bureau noted that it was not making a determination as to whether DTI ratios, a loan's price, or some other measure is the best predictor of loan performance. Rather, the analyses provided by stakeholders and the Bureau's own analysis show that pricing is strongly correlated with loan performance, based on early delinquency rates, across a variety of loans and economic conditions. The Bureau acknowledged that DTI is also predictive of loan performance and that other direct and indirect measures of consumer finances may also be predictive of loan performance. However, the Bureau weighed several policy considerations in selecting an approach for the proposal based on the purposes of the ATR/QM provisions of TILA.

In proposing a price-based General QM loan definition, the Bureau sought to balance considerations related to ensuring consumers' ability to repay and maintaining access to credit. As noted in the proposal, the Bureau views the relevant provisions of TILA as fundamentally about assuring that consumers receive mortgage credit that they can repay. However, the Bureau also stated its concern about maintaining access to responsible, affordable mortgage credit. The proposal noted the Bureau's concern that the current General QM loan definition, with a 43 percent DTI limit, would result in a significant reduction in the scope of the QM market and could reduce access to responsible, affordable mortgage credit after the Temporary GSE QM loan definition expires. The lack of a robust non-QM market enhances those concerns. Although it remains possible that, over time, a substantial market for non-QM loans will emerge, that market has developed slowly, and the recent economic disruptions associated with the COVID-19 pandemic may significantly hinder its development, at least in the near term.

With respect to ability to repay, the proposal focused on analysis of early delinquency rates to evaluate whether a loan's price, as measured by the spread of APR over APOR (herein referred to as the loan's rate spread), is an appropriate measure of whether a loan should be presumed to comply with the ATR provisions. The proposal noted that, because the affordability of a given mortgage will vary from consumer to consumer based upon a range of factors, there is no single recognized metric, or set of metrics, that can directly measure whether the terms of mortgage loans are reasonably within consumers' ability to repay. As such, consistent with the Bureau's prior analyses in the Assessment Report, the Bureau's analysis in the proposal used early distress as a proxy for the lack of the consumer's ability to repay at consummation across a wide pool of loans. Specifically, and consistent with the Assessment Report,^[180] the proposal measured early distress as whether a consumer was ever 60 or more days past due within the first two years after origination (referred to herein as the early delinquency rate). The Bureau's analysis focused on early delinquency rates to capture consumers' difficulties in making payments soon after consummation of the loan (i.e., within the first two years), even if these delinquencies do not lead to consumers potentially losing their homes (i.e., 60 or more days past due, as opposed to 90 or more days or in foreclosure), as early difficulties in making payments indicate a higher likelihood that the consumer may have lacked ability to repay at consummation. As in the Assessment Report, the Bureau assumed that the average early delinquency rate across a wide pool of mortgages—whether safe harbor QM, rebuttable presumption QM, or non-QM—is probative of whether such loans are reasonably within consumers' repayment ability. The Bureau acknowledged that alternative measures of delinquency, including those used in analyses submitted as comments on the ANPR, may also be probative of repayment ability.

In issuing the proposal, the Bureau reviewed available evidence to assess whether rate spreads can distinguish loans that are likely to have low early delinquency rates, and thus may be presumed to comply with the ATR requirements, from loans that are likely to have higher rates of early delinquency, for which a presumption of compliance with the ATR requirements would not be warranted. The proposal stated that the Bureau's own analysis and analyses published in response to the Bureau's ANPR and RFIs provide strong evidence of increasing early delinquency rates with higher rate spreads across a range of datasets, time periods, loan types, measures of rate spread, and measures of delinquency. The Bureau's delinquency analysis used data from the National Mortgage Database (NMDB),^[181] including a matched sample of NMDB and HMDA loans.^[182] As noted in the proposal, the analysis shows that delinquency rates rise with rate spread. The Bureau's delinquency analysis is described below. Table numbers in part V match those from the Bureau's proposal, except that Tables 7A and 8A in part V.B.5, below, did not appear in the proposal.

Table 1 shows early delinquency rates for 2002-2008 first-lien purchase originations in the NMDB, with loans categorized according to their approximate rate spread. The Bureau analyzed 2002 through 2008 origination years because the relatively fixed private mortgage insurance (PMI) pricing during these years allows for reliable approximation of this important component of rate spreads.^[183] The sample is restricted to loans without product features that would make them non-QM loans under the current rule. Table 1 shows that early delinquency rates increase consistently with rate spreads, from a low of 2 percent among loans with rate spreads below or near zero, up to 14 percent for loans with rate spreads of 2.25 percentage points or more over APOR.^[184] This sample includes loans originated during the peak of the housing boom and delinquencies that occurred during the ensuing recession, contributing to the high overall levels of early delinquency.

Table 1—2002-2008 Originations, Early Delinquency Rate by Rate Spread

Rate spread (interest rate + PMI approximation−PMMS 185) in percentage points	Early delinquency rate (%)
<0	2
0-0.24	2
0.25-0.49	4
0.50-0.74	5
0.75-0.99	6
1.00-1.24	8
1.25-1.49	10
1.50-1.74	12
1.75-1.99	13
2.00-2.24	14
2.25 and above	14

The proposal noted that analysis of additional data, as reflected in Table 2, also shows early delinquency rates rising with rate spread. Table 2 shows early delinquency statistics for 2018 NMDB first-lien purchase originations that have been matched to 2018 HMDA data, enabling the Bureau to use actual rate spreads over APOR rather than approximated rate spreads in its analysis.^[186] As with the data reflected in Table 1, loans with product features that would make them non-QM under the current rule are excluded from Table 2. However, only delinquencies occurring through December 2019 are observed in Table 2, meaning most loans are not observed for a full two years after origination. This more recent sample provides insight into early delinquency rates under post-crisis lending standards, and for an origination cohort that had not undergone (as of December 2019) a large economic downturn. The 2018 data are divided into wider bins (as compared to Table 1) to ensure enough loans per bin. As with Table 1, the proposal noted that Table 2 shows that early delinquency rates increase consistently with rate spreads, from a low of 0.2 percent for loans with rate spreads near APOR or below APOR, up to 4.2 percent for loans with rate spreads of 2 percentage points or more over APOR.^[187]

Table 2—2018 Originations, Early Delinquency Rate by Rate Spread

Rate spread over APOR in percentage points	Early delinquency rate (as of Dec. 2019) (%)
<0	0.2
0-0.49	0.2
0.50-0.99	0.6
1.00-1.49	1.7
1.50-1.99	2.7
2.00 and above	4.2

Given the specific DTI limit under the current rule, the Bureau also analyzed the relationship between DTI ratios and early delinquency for the same samples of loans in Tables 3 and 4. As described in the proposal, the Bureau's analyses show that early delinquency rates increase consistently with DTI ratio in both samples. In the 2002-2008 sample, early delinquency rates increase from a low of 3 percent among loans with DTI ratios at or below 25 percent, up to 9 percent for loans with DTI ratios between 61 and 70 percent.^[188] In the 2018 sample, early delinquency rates increase from 0.4 percent among loans with DTI ratios at or below 25 percent, up to 0.9 percent among loans with DTI ratios between 44 and 50.^[189] The difference in early delinquency rates between loans with the highest and lowest DTI ratios is smaller than the difference in early delinquency rates between the highest and lowest rate spreads during both periods. The proposal explained that, for these samples and bins of rate spread and DTI ratios, this pattern is consistent with a stronger correlation between rate spread and early delinquency than between DTI ratios and early delinquency.

Table 3—2002-2008 Originations, Early Delinquency Rate by DTI Ratio

DTI	Early delinquency rate (%)
0-20	3
21-25	3
26-30	4
31-35	5
36-40	6
41-43	6
44-45	7
46-48	7
49-50	8
51-60	8
61-70	9

Table 4—2018 Originations, Early Delinquency Rate by DTI

DTI	Early delinquency rate (as of Dec. 2019) (%)
0-25	0.4
26-35	0.5
36-43	0.7
44-48	0.9
49-50	0.9

The proposal further analyzed the strengths of DTI ratios and pricing in predicting early delinquency rates in Tables 5 and 6, which show the early delinquency rates of these same samples categorized according to both their DTI ratios and their rate spreads. Table 5 shows early delinquency rates for 2002-2008 first-lien purchase originations in the NMDB, with loans categorized according to both their DTI ratio and their approximate rate spread. For loans within a given DTI ratio range, those with higher rate spreads consistently had higher early delinquency rates. Loans with low rate spreads had relatively low early delinquency rates even at high DTI ratio levels, as seen in the 2 percent early delinquency rate for loans priced below APOR but with DTI ratios of 46 to 48 percent, 51 to 60 percent, and 61 to 70 percent. However, the highest early delinquency rates occurred for loans with high rate spreads and high DTI ratios, reaching 26 percent for loans priced 2 to 2.24 percentage points above APOR with DTI ratios of 61 to 70 percent. Across DTI bins, loans priced significantly above APOR had early delinquency rates much higher than loans priced below APOR.

Table 5—2002-2008 Originations, Early Delinquency Rate by Rate Spread and DTI Ratio

Rate spread (interest rate + PMI approx.−PMMS) in percentage points	DTI 0-20 (%)	DTI 21-25 (%)	DTI 26-30 (%)	DTI 31-35 (%)	DTI 36-40 (%)	DTI 41-43 (%)	DTI 44-45 (%)	DTI 46-48 (%)	DTI 49-50 (%)	DTI 51-60 (%)	DTI 61-70 (%)
<0	2	1	1	2	2	2	2	2	3	2	2
0-0.24	2	2	2	2	2	3	3	3	3	3	3
0.25-0.49	3	3	3	3	4	5	4	5	5	5	5
0.50-0.74	4	4	4	4	5	6	6	6	7	7	7
0.75-0.99	4	5	5	6	6	7	7	7	8	8	10
1.00-1.24	6	6	6	7	7	9	9	9	10	11	13
1.25-1.49	6	7	8	8	10	11	12	12	12	14	15
1.50-1.74	7	8	9	10	13	13	15	14	16	15	20
1.75-1.99	7	8	10	12	14	15	16	16	16	18	22
2.00-2.24	6	10	10	12	15	15	17	19	18	20	26
2.25 and above	7	9	10	13	15	16	16	18	19	20	25

Similarly, Table 6 shows average early delinquency statistics, with loans categorized according to both DTI and rate spread, for the sample of 2018 NMDB first-lien purchase originations that have been matched to 2018 HMDA data.^[190] For Table 6, the higher early delinquency rate for loans with higher rate spreads over APOR matches the pattern shown in the data from Table 5. Overall early delinquency rates are substantially lower, reflecting the importance of economic conditions in the likelihood of delinquency for any given consumer. However, the 2018 loans priced significantly above APOR also had early delinquency rates much higher than loans priced below APOR.

Table 6—2018 Originations, Early Delinquency Rate by Rate Spread and DTI Ratio

Rate spread over APOR in percentage points	DTI 0-25 (%)	DTI 26-35 (%)	DTI 36-43 (%)	DTI 44-50 (%)
<0	0.1	0.1	0.2	0.3
0-0.49	0.2	0.1	0.3	0.3
0.50-0.99	0.1	0.4	0.8	0.8
1.00-1.49	1.0	1.4	1.5	2.3
1.50-1.99		3.2	2.5	2.3
2.00 and above		4.4	3.9	4.2

The proposal noted that the high relative risk of early delinquency for higher-priced loans holds across samples, demonstrating that rate spreads distinguish early delinquency risk under a range of economic conditions and creditor practices. The proposal also highlighted that analyses published in response to the Bureau's ANPR and RFIs are consistent with the Bureau's analysis showing that early delinquency rates rise consistently with rate spread. For example, CoreLogic analyzed a set of 2018 HMDA conventional mortgage originations merged to loan performance data collected from mortgage servicers.^[191] The CoreLogic analysis found: (1) The lowest delinquency rates among loans with rate spreads that are below APOR, and (2) increased early delinquency rates for each sequentially higher bin of rate spreads up to 2 percentage points over APOR. In assessing the CoreLogic analysis, the Bureau noted that loans priced at or above 2 percentage points over APOR in the 2018 HMDA data are relatively rare and are disproportionately made for manufactured housing and smaller loan amounts and therefore may not be well represented in mortgage servicing datasets. However, the proposal noted that these loans also have relatively high rates of delinquency.^[192] CoreLogic found a similar, but more variable, positive relationship between rate spreads over APOR and delinquency in earlier cohorts (2010-2017) of merged HMDA-CoreLogic originations, a period in which rate spreads were only reported for loans priced at least 1.5 percentage points over APOR.^[193] The proposal also noted that analyses by the Urban Institute (using loan performance data from Black Knight) show a comparable positive relationship between rate spreads—measured there as the note rate over PMMS—and delinquency.^[194] The analysis found that the relationship holds across a range of loan types (conventional loans held in portfolio, in GSE securitizations, and in private securitizations; FHA loans; VA loans) and years (1995-2018). Additional analyses by the Urban Institute show the same positive relationship between rate spread and loan performance in Fannie Mae loan-level performance data.^[195]

The proposal stated that, collectively, this evidence suggests that higher rate spreads—including the specific measure of APR over APOR—are strongly correlated with early delinquency rates. Given that early delinquency captures consumers' difficulty making required payments, the proposal preliminarily concluded that rate spreads provide a strong indicator of ability to repay.

The proposal acknowledged that a test that combines rate spread and DTI may better predict early delinquency rates than either metric on its own. However, the proposal also noted that any rule with a specific DTI limit would need to provide standards for calculating the income that may be counted and the debt that must be counted so that creditors and investors can ensure with reasonable certainty that they have accurately calculated DTI within the specific DTI limit. As noted above, the current definitions of debt and income in appendix Q have proven to be complex in practice and may unduly restrict access to credit. The proposal expressed concerns about whether other potential approaches could define debt and income with sufficient clarity while at the same time providing flexibility to accommodate new approaches to verification and underwriting.

In addition to strongly correlating with loan performance, the proposal tentatively concluded that a price-based General QM loan definition is a more holistic and flexible measure of a consumer's ability to repay than DTI alone. The proposal explained that mortgage underwriting, and by extension, a loan's price, generally includes consideration of a consumer's DTI. However, the proposal explained that loan pricing also includes an assessment of additional factors that might compensate for a higher DTI ratio and that might also be probative of a consumer's ability to repay. One of the primary criticisms of the current 43 percent DTI ratio is that it is too limited in assessing a consumer's finances and, as such, may unduly restrict access to credit. Therefore, the proposal noted that a potential benefit of a price-based General QM loan definition is that a mortgage loan's price reflects credit risk based on many factors, including DTI ratios, and may be a more holistic measure of ability to repay than DTI ratios alone. Further, there is inherent flexibility for creditors in a rate-spread-based General QM loan definition, which could facilitate innovation in underwriting, including the use of emerging research into alternative mechanisms to assess a consumer's ability to repay. Such innovations include certain new uses of cash flow data and analytics to underwrite mortgage applicants. This emerging technology has the potential to accurately assess consumers' ability to repay using, for example, bank account data that can identify the source and frequency of recurring deposits and payments and identify remaining disposable income. Identifying the remaining disposable income could be a method of assessing the consumer's residual income and could potentially satisfy a requirement to consider either DTI or residual income, absent a specific DTI limit.

The proposal also noted that there is significant precedent for using the price of a mortgage loan to determine whether to apply additional consumer protections, in recognition of the lower risk generally posed by lower-priced mortgages. A price-based General QM loan definition would be consistent with these existing provisions that provide greater protections to consumers with more expensive loans. For example, TILA and Regulation Z use a loan's APR in comparison to APOR and as one trigger for heightened consumer protections for certain “high-cost mortgages” pursuant to HOEPA.^[196] Loans that meet HOEPA's high-cost trigger are subject to special disclosure requirements and restrictions on loan terms, and consumers with high-cost mortgages have enhanced remedies for violations of the law. Further, in 2008, the Board exercised its authority under HOEPA to require certain consumer protections concerning a consumer's ability to repay, prepayment penalties, and escrow accounts for taxes and insurance for HPMLs, which have APR spreads lower than those prescribed for high-cost mortgages but that nevertheless exceed APOR by a specified threshold.^[197] Although the ATR/QM Rule replaced the ability-to-repay requirements promulgated pursuant to HOEPA and the Board's 2008 rule,^[198] HPMLs remain subject to additional requirements related to escrow accounts for taxes and homeowners insurance and to appraisal requirements.^[199] The proposal also noted that the ATR/QM Rule itself provides additional protection to QMs that are higher-priced covered transactions, as defined in § 1026.43(b)(4), in the form of a rebuttable presumption of compliance with the ATR provisions, instead of a conclusive safe harbor.

Finally, the proposal preliminarily concluded that a price-based General QM loan definition would provide compliance certainty to creditors because creditors would be able to readily determine whether a loan is a General QM. Creditors have experience with APR calculations due to the existing price-based regulatory requirements described above, and for various other disclosure and compliance reasons under Regulation Z. Creditors also have experience determining the appropriate APOR for use in calculating rate spreads. As such, the proposal stated that the approach should provide certainty to creditors regarding a loan's status as a QM.^[200]

Although the proposal would have required creditors to consider the consumer's DTI ratio or residual income, income or assets other than the value of the dwelling, and debts, the proposal would not have mandated a specific DTI limit. The proposal would have removed appendix Q and instead would have provided creditors additional flexibility for defining income or assets other than the value of the dwelling and debts. The Bureau did not propose a single, specific set of standards equivalent to appendix Q for what must be counted as income or assets and what may be counted as debts. For purposes of the proposed requirement, income or assets and debts would be determined in accordance with proposed § 1026.43(e)(2)(v)(B), which would have required the creditor to verify the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan and the consumer's current debt obligations, alimony, and child support. The proposed rule would have provided a safe harbor to creditors using verification standards the Bureau specifies. The proposal noted that this could potentially include relevant provisions from Fannie Mae's Single Family Selling Guide, Freddie Mac's Single-Family Seller/Servicer Guide, FHA's Single Family Housing Policy Handbook, the VA's Lenders Handbook, and the Field Office Handbook for the Direct Single Family Housing Program and Handbook for the Single Family Guaranteed Loan Program of the USDA, current as of the proposal's public release. However, under the proposal, creditors would not have been required to verify income and debt according to the standards the Bureau specifies. Rather, the proposal would have provided creditors with the flexibility to develop other methods of compliance with the verification requirements.

The proposal would have provided that a loan meets the General QM loan definition in § 1026.43(e)(2) only if the APR exceeds APOR for a comparable transaction by less than 2 percentage points as of the date the interest rate is set. In proposing this threshold, the Bureau tentatively concluded that it would strike an appropriate balance between ensuring that loans receiving QM status may be presumed to comply with the ATR provisions and ensuring that access to responsible, affordable mortgage credit remains available to consumers. For these same reasons, the Bureau proposed higher thresholds for smaller loans and subordinate-lien transactions, as the Bureau was concerned that loans with lower loan amounts may be priced higher than larger loans, even if the consumers have similar credit characteristics and a similar ability to repay. For all loans, regardless of loan size, the Bureau did not propose to alter the current threshold separating safe harbor from rebuttable presumption QMs in § 1026.43(b)(4), under which a loan is a safe harbor QM if its APR exceeds APOR for a comparable transaction by less than 1.5 percentage points as of the date the interest rate is set (or 3.5 percentage points for subordinate-lien transactions). As such, under the proposal, first-lien loans that otherwise meet the General QM loan definition and for which the APR exceeds APOR by 1.5 or more percentage points (but by less than 2 percentage points) as of the date the interest rate is set would have received a rebuttable presumption of compliance with the ATR provisions.

Finally, the proposal provided analysis of the potential effects on access to credit of a price-based approach to defining a General QM. As indicated by the various combinations in Table 7 below, the proposal analyzed 2018 HMDA data and found that under the current rule—including the Temporary GSE QM loan definition, the General QM loan definition with a 43 percent DTI limit, and the Small Creditor QM loan definition in § 1026.43(e)(5)—90.6 percent of conventional purchase loans were safe harbor QMs and 95.8 percent were safe harbor QMs or rebuttable presumption QMs. Under the proposed General QM loan definition's rate-spread thresholds of 1.5 (safe harbor) and 2.0 (rebuttable presumption) percentage points over APOR, the proposal stated that 91.6 percent of conventional purchase loans would have been safe harbor QMs and 96.1 percent would have been safe harbor QM or rebuttable presumption QMs.^[201] Based on these 2018 data, the proposal stated that rate-spread thresholds of 1.0-2.0 percentage points over APOR for safe harbor QMs would have covered 83.3 to 94.1 percent of the conventional purchase market (as safe harbor QMs), while rate-spread thresholds of 1.5-2.5 percentage points over APOR for rebuttable presumption QMs would have covered 94.3 to 96.8 percent of the conventional purchase market (as safe harbor and rebuttable presumption QMs). As explained further in part V.B.5, the Bureau is providing in Table 7A revised estimates for the size of the QM market based on the higher thresholds for small loans and manufactured housing loans as adopted by this final rule and also to reflect a revised methodology to identify creditors eligible to originate loans as small creditors under § 1026.43(e)(5).

Table 7—Proposal's Estimated Share of 2018 Conventional First-Lien Purchase Loans Within Various Price-Based Safe Harbor (SH) QM and Rebuttable Presumption (RP) QM Definitions

[HMDA data]

Approach	Safe harbor QM (share of conventional purchase market)	QM overall (share of conventional purchase market)
Temporary GSE QM + DTI 43	90.6	95.8
Proposal (SH 1.50, RP 2.00)	91.6	96.1
SH 0.75, RP 1.50	74.6	94.3
SH 1.00, RP 1.50	83.3	94.3
SH 1.25, RP 1.75	88.4	95.3
SH 1.35, RP 2.00	89.8	96.1
SH 1.40, RP 2.00	90.5	96.1
SH 1.75, RP 2.25	93.1	96.6
SH 2.00, RP 2.50	94.1	96.8

Despite the expected benefits of a price-based General QM loan definition, the proposal noted concerns about the definition. In particular, the Bureau acknowledged that while the Bureau believes a loan's price may be a more holistic and flexible measure of a consumer's ability to repay than DTI alone, the Bureau recognized that there is a distinction between credit risk, which largely determines pricing relative to the prime rate, and a particular consumer's ability to repay, which is one component of credit risk. The Bureau also acknowledged that factors unrelated to the individual loan (e.g., institutional factors such as the competing policy considerations inherent in setting guarantee fees on GSE loans) can influence its price and that a price-based approach would incentivize some creditors to price some loans just below the threshold so that the loans will receive the presumption of compliance that comes with QM status. The proposal also acknowledged concerns about the sensitivity of a price-based General QM loan definition to macroeconomic cycles and that a price-based approach would likely be pro-cyclical, with a more expansive QM market when the economy is expanding, and a more restrictive QM market when credit is tight. The Bureau discusses these concerns below in part V.B.5.

As noted above, stakeholders providing feedback prior to the General QM Proposal suggested a range of options the Bureau should consider to replace the 43 percent DTI limit in the General QM loan definition. These options are discussed at length in the proposal.^[202] The Bureau considered these options in developing the proposal, but preliminarily concluded that the price-based approach in proposed § 1026.43(e)(2) would best achieve the statutory goals of ensuring consumers' ability to repay and maintaining access to responsible, affordable, mortgage credit. However, as explained in part V.B.3, below, the Bureau requested comment on whether an alternative approach that adopts a higher DTI limit or a hybrid approach that combines pricing and a DTI limit, along with a more flexible standard for defining income or assets and debts, could provide a superior alternative to the price-based approach.

The proposal also acknowledged that some stakeholders requested that the Bureau make the Temporary GSE QM loan definition permanent. The Bureau did not propose this alternative because of its concern that there is not a basis to presume for an indefinite period that loans eligible to be purchased or guaranteed by the GSEs—whether or not the GSEs are under conservatorship—have been originated with appropriate consideration of consumers' ability to repay.^[203] The Bureau also expressed concern that making the Temporary GSE QM loan definition permanent could stifle innovation and competition in private-sector approaches to underwriting. The Bureau also expressed concern that, as long as the Temporary GSE QM loan definition continues in effect, the non-GSE private market is less likely to rebound and that the existence of the Temporary GSE QM loan definition may be contributing to the limited non-GSE private market. As explained above, the Extension Final Rule extended the Temporary GSE QM loan definition to expire on the mandatory compliance date of this final rule or when GSE conservatorship ends.

3. Alternative to the Proposed Price-Based General QM Loan Definition: Retaining a DTI Limit

Although the Bureau proposed to remove the 43 percent DTI limit and adopt a price-based approach for the General QM loan definition, the Bureau requested comment on an alternative approach that would retain a DTI limit, but raise it above the current limit of 43 percent, and provide a more flexible set of standards for verifying income or assets and debts in place of appendix Q. The Bureau requested comment on this alternative proposal because of concerns about the price-based approach. In particular, the Bureau acknowledged the sensitivity of a price-based QM definition to macroeconomic cycles, including concerns that the price-based approach could be pro-cyclical, with a more expansive QM market when the economy is expanding, and a more restrictive QM market when credit is tight. The Bureau was especially concerned about these potential effects given the recent economic disruptions associated with the COVID-19 pandemic. The Bureau also acknowledged that a small share of loans that satisfy the current General QM loan definition would lose QM status under the proposed price-based approach due to the loans' rate spread exceeding the applicable threshold. Further, and as described above, the Bureau analyzed the relationship between DTI ratios and early delinquency, using data on first-lien conventional purchase originations from the NMDB, including a matched sample of NMDB and HMDA loans. That analysis, as shown in Tables 3 and 4 above, shows that early delinquency rates increase consistently with DTI ratio. For these reasons, the Bureau requested comment on whether an approach that increases the DTI limit to a specific threshold within a range of 45 to 48 percent and that includes more flexible definitions of debt and income would be a superior alternative to a price-based approach.^[204]

The Bureau also analyzed the potential effects of a DTI-based approach on the size of the QM market and on access to credit. As indicated in the proposal's Table 8, the proposal found that 2018 HMDA data show that with the Temporary GSE QM loan definition and the General QM loan definition with a 43 percent DTI limit, 90.6 percent of conventional purchase loans were safe harbor QMs and 95.8 percent were safe harbor QM or rebuttable presumption QMs. If, instead, the Temporary GSE QM loan definition were not in place along with the General QM loan definition (with the 43 percent DTI limit), and assuming no change in consumer or creditor behavior from the 2018 HMDA data, then the proposal found that only 69.3 percent of loans would have been safe harbor QMs and 73.6 percent of loans would have been safe harbor QMs or rebuttable presumption QMs. The proposal also noted that raising the DTI limit above 43 percent would increase the size of the QM market and, as a result, potentially increase access to credit relative to the General QM loan definition with a DTI limit of 43 percent. The proposal noted that the magnitude of the increase in the size of the QM market and potential increase in access to credit would depend on the selected DTI limit. A DTI limit in the range of 45 to 48 percent would likely result in a QM market that is larger than one with a DTI limit of 43 percent but smaller than the status quo (i.e., Temporary GSE QM loan definition and DTI limit of 43 percent). However, the proposal noted the Bureau's expectation that consumers and creditors would respond to changes in the General QM loan definition, potentially allowing additional loans to be made as safe harbor QMs or rebuttable presumption QMs. As explained further in part V.B.5, the Bureau is providing in Table 8A revised analysis of the size of the QM market to reflect a revised methodology to identify creditors eligible to originate loans as small creditors under § 1026.43(e)(5).

Table 8—Proposal's Estimated Share of 2018 Conventional Purchase Loans Within Various Safe Harbor QM and Rebuttable Presumption QM Definitions

[HMDA data]

Approach	Safe harbor QM (share of conventional market)	QM overall (share of conventional market)
Temporary GSE QM + DTI 43	90.6	95.8
Proposal (Pricing at 2.0)	91.6	96.1
DTI limit 43	69.3	73.6
DTI limit 45	76.1	80.9
DTI limit 46	78.8	83.8
DTI limit 47	81.4	86.6
DTI limit 48	84.1	89.4
DTI limit 49	87.0	92.4
DTI limit 50	90.8	96.4

The Bureau specifically requested comment on a specific DTI limit between 45 and 48 percent. The Bureau requested comment and data on whether increasing the DTI limit to a specific percentage between 45 and 48 percent would be a superior alternative to the proposed price-based approach, and, if so, on what specific DTI percentage the Bureau should include in the General QM loan definition. The Bureau requested comment and data as to how specific DTI percentages would be expected to affect access to credit and would be expected to affect the risk that the General QM loan definition would include loans that should not receive a presumption of compliance with TILA's ATR requirements. The Bureau also requested comment on whether increasing the DTI limit to a specific percentage between 45 to 48 percent would better balance the goals of ensuring access to responsible, affordable credit and ensuring that QM status is limited to loans for which it is appropriate to presume that consumers have the ability to repay. The Bureau also requested comment on the macroeconomic effects of a DTI-based approach, as well as whether and how the Bureau should weigh such effects in amending the General QM loan definition. In addition, the Bureau requested comment on whether, if the Bureau adopts a higher specific DTI limit as part of the General QM loan definition, the Bureau should retain the price-based threshold of 1.5 percentage points over APOR to separate safe harbor QMs from rebuttable presumption QMs for first-lien transactions.

The Bureau also requested comment on whether to adopt a hybrid approach in which a combination of a DTI limit and a price-based threshold would be used in the General QM loan definition. The proposal noted that one such approach could impose a DTI limit only for loans above a certain pricing threshold. Such an approach would be intended to reduce the likelihood that loans for which the consumer lacks ability to repay would receive a presumption of compliance with the ATR requirements, while avoiding the potential burden and complexity of a DTI limit for many lower-priced loans. The proposal explained that a similar approach might impose a DTI limit above a certain pricing threshold and also tailor the presumption of compliance with the ATR requirements based on DTI. For example, the proposal noted that the rule could provide that (1) for loans with rate spreads under 1 percentage point, the loan is a safe harbor QM regardless of the consumer's DTI ratio; (2) for loans with rate spreads at or above 1 but less than 1.5 percentage points, a loan is a safe harbor QM if the consumer's DTI ratio does not exceed 50 percent and a rebuttable presumption QM if the consumer's DTI is above 50 percent; and (3) if the rate spread is at or above 1.5 but less than 2 percentage points, the loan would be rebuttable presumption QM if the consumer's DTI ratio does not exceed 50 percent and a non-QM loan if the DTI ratio is above 50 percent.

The proposal explained another hybrid approach that would impose a DTI limit on all General QMs but would allow higher DTI ratios for loans below a set pricing threshold. For example, the rule could generally impose a DTI limit of 47 percent but could permit a loan with a DTI ratio up to 50 percent to be eligible for QM status under the General QM loan definition if the APR is less than 2 percentage points over APOR. This approach might limit the likelihood of providing QM status to loans for which the consumer lacks ability to repay, but also would permit some lower-priced loans with higher DTI ratios to achieve QM status.

With respect to the Bureau's concerns about appendix Q, the Bureau requested comment on an alternative method of defining debt and income to replace appendix Q in conjunction with a specific DTI limit. The Bureau expressed concern that the appendix Q definitions of debt and income are rigid and difficult to apply and do not provide the level of compliance certainty that the Bureau anticipated at the time of the January 2013 Final Rule. The proposal further noted that, under the current rule, some loans that would otherwise have DTI ratios below 43 percent do not satisfy the General QM loan definition because their method of documenting and verifying income or debt is incompatible with appendix Q. In particular, the Bureau requested comment on whether the approach in proposed § 1026.43(e)(2)(v) could be applied with a General QM loan definition that includes a specific DTI limit. As discussed in more detail in the section-by-section discussion of § 1026.43(e)(2)(v), proposed § 1026.43(e)(2)(v)(A) would have required creditors to consider the consumer's monthly DTI ratio or residual income; current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan; and debt obligations, alimony, and child support. Proposed § 1026.43(e)(2)(v)(B) and the associated commentary would have explained how creditors must verify and count the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan and the consumer's current debt obligations, alimony, and child support, relying on the standards set forth in the ATR requirements in § 1026.43(c). Proposed § 1026.43(e)(2)(v)(B) would have further provided creditors a safe harbor with standards the Bureau may specify for verifying debt and income, potentially including relevant provisions from the Fannie Mae Single Family Selling Guide, the Freddie Mac Single-Family Seller/Servicer Guide, FHA's Single Family Housing Policy Handbook, the VA's Lenders Handbook, and USDA's Field Office Handbook for the Direct Single Family Housing Program and Handbook for the Single Family Guaranteed Loan Program, current as of the proposal's public release. The Bureau also requested comments on potentially adding to the safe harbor other standards that external stakeholders develop.

The Bureau requested comment on whether the alternative method of defining debt and income in proposed § 1026.43(e)(2)(v)(B) could replace appendix Q in conjunction with a specific DTI limit. As noted above, the proposal expressed concern that this approach, which combines a general standard with safe harbors, may not be appropriate for a General QM loan definition with a specific DTI limit. The Bureau requested comment on whether the approach in proposed § 1026.43(e)(2)(v)(B) would address the problems associated with appendix Q and would provide an alternative method of defining debt and income that would be workable with a specific DTI limit. The Bureau requested comment on whether allowing creditors to use standards the Bureau may specify to verify debt and income—as would be permitted under proposed § 1026.43(e)(2)(v)(B)—as well as potentially other standards external stakeholders develop and the Bureau adopts, would provide adequate clarity and flexibility while also ensuring that DTI calculations across creditors and consumers are sufficiently consistent to provide meaningful comparison of a consumer's calculated DTI ratio to any DTI ratio threshold specified in the rule.

The Bureau also requested comment on what changes, if any, would need to be made to proposed § 1026.43(e)(2)(v)(B) to accommodate a specific DTI limit. For example, the Bureau requested comment on whether creditors that comply with manuals that have been revised but are substantially similar to the manuals specified above should receive a safe harbor, as the Bureau proposed. The Bureau also requested comment on its proposal to allow creditors to “mix and match” verification standards, including whether the Bureau should instead limit or prohibit such “mixing and matching” under an approach that incorporates a specific DTI limit. The Bureau requested comment on whether these aspects of the approach in proposed § 1026.43(e)(2)(v)(B), if used in conjunction with a specific DTI limit, would provide sufficient certainty to creditors, investors, and assignees regarding a loan's QM status and whether it would result in potentially inconsistent application of the General QM loan definition.

4. Comments on the Price-Based General QM Loan Definition

Numerous commenters supported the Bureau's proposal to move from a DTI-based General QM loan definition to one based on pricing. Commenters that supported the proposal included industry commenters, consumer advocate commenters, a research center commenter, joint industry and consumer advocate commenters, and two GSE commenters. Commenters who supported the proposed price-based approach generally supported the Bureau's rationale for the proposal, described in part V.B.2 above. With respect to measuring consumers' ability to repay, commenters supporting the proposal generally agreed with the Bureau's analysis showing that the price of a loan is strongly associated with its performance, measured by whether a consumer was 60 days or more past due during the first two years of the loan, and also agreed that price is a strong indicator of consumers' ability to repay.

A joint consumer advocate and industry comment letter generally supporting the proposal described its analysis of the relationship between delinquency rates and rate spread. The commenter's analysis used Fannie Mae Single-Family Loan Performance data and, like the Bureau's 2002-2008 delinquency analysis, approximated rate spreads using the sum of the mortgage interest rate and an estimated PMI premium, minus APOR. Unlike the Bureau's analysis, however, the commenter used a risk-based estimated PMI premium to approximate current PMI pricing practices. The commenter noted that using risk-based PMI pricing increases the variance of rate spread estimates for loans with PMI, such that low-risk consumers have lower premiums and high-risk borrowers have higher premiums. Like the Bureau's delinquency analysis, the joint commenter defined early delinquency as whether the consumer was ever 60 days delinquent during the first two years of the loan. The joint commenter's analysis looks at loans by rate spread, ranging from less than a 0.5 percentage point rate spread, up to 3.0 or more percentage points, in increments of 0.5 percentage points. The commenter provided results of this analysis for loans originated between 1999-2019, and also provided results for loans originated between 2013-2018. For both sets of loans, the analysis shows early delinquency rates rising with rate spread. For the 1999-2019 dataset, loans with rate spreads of less than 0.5 percentage points had an early delinquency rate of 1.0 percent, rising to 14.3 percent for rate spreads of 3 percentage points or more. For the 2013-2018 dataset, loans with rate spreads of less than 0.5 percentage points had an early delinquency rate of 0.5 percent, rising to 10.5 percent for rate spreads of 3 percentage points or more.

Similarly, a research center commenter generally supporting the proposal also provided analysis of loan performance by rate spread. The commenter looked at Fannie Mae Single-Family Loan Performance data and portfolio loans and loans in PLS channels in the Black Knight McDash database. The commenter measured loan performance by whether the consumer was ever 60 days or more delinquent, rather than by whether the consumer was 60 days or more delinquent in the first two years of the loan as in the Bureau's delinquency analysis. The commenter stated that its measure is more conservative in that it produces higher default rates. The commenter noted that its analysis found all measures of default to be highly correlated with rate spreads but also noted that defaults on loans originated after the financial crisis (defined by the commenter as 2013 to 2018 originations) are lower than for any other period in recent history. The commenter attributes this to improvements in mortgage underwriting. This commenter's analysis is discussed further below in the section-by-section analysis of § 1026.43(e)(2)(vi).

Some commenters supporting the proposal, including a research center and a joint consumer advocate and industry comment, argued that pricing is a stronger predictor of default than DTI. The joint consumer advocate and industry commenter noted that DTI is a particularly weak predictor of loan performance for near-prime loans. In support of that assertion, the commenter cited analysis finding that, for a thousand consumers with DTI ratios between 45 and 50 percent, only two additional consumers default compared to consumers with DTI ratios between 40 and 45 percent. That commenter also cited analysis showing that, for each year since 2011, the 90-day delinquency rate for loans with DTI ratios over 45 percent is less than that for loans with DTI ratios between 30 percent and 45 percent. The commenter asserts that this is counterintuitive to the idea that higher DTI ratios are a sound predictor of default.

Some commenters supporting the proposed price-based approach, including several industry commenters, specifically agreed with the Bureau's observation that pricing is a more holistic measure of a consumer's financial capacity than DTI alone. Generally, these commenters agreed with the Bureau's observation that pricing considers a broader set of factors, which results in a strong measure of ability to repay that is more complete than a DTI-based definition. A joint consumer advocate and industry commenter asserted that a DTI limit would curtail access to credit for creditworthy consumers, such as those who have demonstrated the ability to handle debt by regularly paying rent or who have compensating factors permitting them to exceed a particular DTI cutoff. That commenter also asserted that there are considerable challenges to the measurement of DTI, especially the income component, which are accentuated for non-traditional and non-salary employees, including many entrepreneurs and gig workers.

Commenters supporting the price-based approach, including a GSE commenter, also agreed with the Bureau's assertion that the price-based approach would maintain access to responsible, affordable mortgage credit after the expiration of the Temporary GSE QM loan definition. A research center commenter estimated the overall effect of the proposed changes on QM lending volumes using 2019 HMDA data to determine the number of loans that would not have been QMs in 2019 under the current rule but would be QMs under the proposal (using the General QM pricing thresholds in the proposal). The commenter found that there were 346,376 such loans that would have gained QM status under the proposal. The commenter further found that 49,200 loans would have been QMs in 2019 under the current rule but would be non-QM loans under the proposal (i.e., loans with DTI ratios of 43 percent or lower, but with pricing that exceeded the proposed rate-spread thresholds), resulting in a gain of approximately 297,000 QMs under the proposed thresholds. The commenter asserted that, while the creditors of these loans gaining QM status would receive legal protection due to the loans' QM status, the reduction in litigation risk would translate into better pricing for the consumer. A joint consumer advocate and industry commenter expressed concern about access to credit under a DTI-based approach, noting that “higher DTI” consumers above the threshold would likely pay substantially higher interest rates on potentially riskier products or may be unable to obtain financing. In support of that assertion, the commenter cited the Assessment Report findings that applicants for jumbo loans with DTI ratios above 43 percent (who were therefore ineligible for QMs under the General QM loan definition or the Temporary GSE QM loan definition) paid significantly higher interest rates and had reduced access to credit. The commenter further expressed concern that such effects would disproportionately affect low-income and low-wealth families, including families of color.

As compared to a DTI-based approach, some commenters indicated that the price-based approach would expand access to credit for certain underserved market segments, such as low-income and minority consumers. Conversely, some commenters, including a consumer advocate commenter, expressed concern that a price-based General QM loan definition would curtail access to credit to low-income and minority consumers. A research center commenter that supported the price-based approach also acknowledged that minority consumers are more likely to have higher rate spreads. This commenter stated that, for GSE loans, 6.2 percent and 5.0 percent of all purchase lending to Black and Hispanic households, respectively, had rate spreads above 1.5 percentage points, compared with 2 percent for non-Hispanic White households. The commenter stated that the disparity was wider in the non-GSE conventional channel, with 13.4 percent and 17.0 percent for Black and Hispanic households, respectively, compared with 5 percent for non-Hispanic White households. An industry commenter cited a 2019 study that found that, compared to similar borrowers, Hispanic and African-American borrowers are charged rates that are 7.9 basis points higher for purchase transactions and 3.6 basis points higher for refinance transactions by creditors using algorithmic-based pricing systems.^[205] However, this commenter suggested that the Bureau address this access-to-credit concern by adjusting the rate-spread threshold. As discussed below, many commenters supporting the proposed price-based approach requested that the Bureau increase either the proposed safe harbor threshold, the threshold separating QMs from non-QM loans, or both, to further ensure continued access to credit, including for minority consumers. A consumer advocate commenter also cited the 2019 study referenced above.

Commenters supporting the proposed price-based approach also generally supported removing the 43 percent DTI limit and appendix Q. With respect to appendix Q, a consumer advocate commenter specifically asserted that, even if the Bureau retained and revised appendix Q, those revisions would quickly become antiquated. Consistent with the Bureau's rationale for the proposal, some commenters also cited the historical precedent for a price-based threshold in Regulation Z, including the existing QM safe harbor threshold. Some commenters noted that a price-based approach would be simple to implement because rate spreads are already required to be calculated for other regulatory purposes.

Although many commenters supported the overall shift from a DTI-based General QM loan definition to one based on pricing, numerous commenters opposed the price-based approach. These commenters include individual commenters, an academic commenter, a research center commenter, industry commenters, and some consumer advocate commenters. Some commenters asserted that a loan's price is not an adequate indicator of a consumer's ability to repay. For example, some commenters that opposed the price-based approach argued that creditors do not necessarily consider individual ability-to-repay factors in deciding on the price of loans they offer to the consumer, that price may vary across creditors for reasons unrelated to the consumer, and that the price-based approach may favor some creditors or business models over others. Some commenters critical of the proposal noted that a loan's price is set by reference to factors that are not specific to the consumer, in some instances including prohibited factors such as race, and therefore is an inappropriate basis for the General QM loan definition. Similarly, some commenters argued that price is an inadequate indicator of a consumer's ability to repay because price is based on credit risk (i.e., risk of loss to the creditor or investor) rather than risk to the consumer. Some commenters asserted that creditors do not price risk accurately, with some commenters citing the experience of loans made prior to the financial crisis as support for this concern. Some commenters, including a research center commenter, asserted that creditors would use the price-based approach to manipulate APOR or adjust their prices to fit just under the rate-spread thresholds.

A consumer advocate commenter argued that LTV ratios, which may be one component of pricing, cannot form the basis of the QM definition. This commenter cited TILA section 129C(a)(3), which provides that the consumer's equity in the dwelling or real property that secures repayment of the loan cannot be considered as a financial resource of the borrower in determining a consumer's ability to repay. The commenter argued that, by extension, LTV ratios also cannot legally form any part of the basis of the QM definition. That commenter further asserted that creditors' reliance on LTV ratios in setting price does not reflect consumers' ability to repay because (1) consumers with substantial equity are likely to pay their mortgage regardless of the impact it may have on their overall finances; (2) consumers with substantial equity may have the option to refinance or sell their home and are therefore unlikely to default and allow their home to go into foreclosure; and (3) even if a consumer with substantial equity does go into foreclosure, the lower the LTV ratio, the more likely the creditor will be able to recover the unpaid principal balance from sale proceeds. The commenter contends that because pricing a loan involves consideration of the consumer's equity, a price-based approach to defining QM is impermissible.

One research center commenter asserted that the price-based approach does not capture risk accurately and criticized the Bureau's delinquency analysis, which focuses on average early delinquency rates by rate spread and DTI bins. That commenter analyzed 2018 HMDA data, which is described in the Bureau's Tables 2 and 4 provided in the proposal and above, and servicer data from CoreLogic's Loan Level Markets Analytics dataset through 2019, using a risk assessment matrix developed by the commenter that combines LTV ratios, DTI ratios, and credit scores. The commenter's analysis replicated the Bureau's definition of early delinquency of 60 days past due during the first two years of the loan. The commenter found that, for loans with identical rate spreads, early delinquency rates vary with other characteristics like LTV ratios, DTI ratios, and credit scores. Similarly, for loans with similar risk levels based on the commenters' risk assessment matrix, the rate spreads vary greatly. The commenter asserts that this is evidence that price does not capture risk accurately. The commenter further argued that the price-based approach is less accurate in predicting the likelihood of default for higher-risk loans. The commenter asserted that some higher-risk loans may be cross-subsidized, and further noted that pricing can be influenced by whether the consumer shopped for a loan and by “random luck.” Analyzing Optimal Blue rate data from the 2013-2018 timeframe, the research center commenter contended that the price-based approach would have signaled that market-wide risk declined, whereas other measures, including DTI and other industry risk metrics, would have signaled the opposite.

A consumer advocate commenter asserted that the price-based approach would grant QM status to loans where a sizeable percentage of consumers lack ability to repay and would create heightened risk of foreclosure. The commenter cited to the Bureau's delinquency analysis in Table 1 (provided in the proposal and above) that looked at loans originated between 2002 and 2008 and shows an early delinquency rate of 13 percent for loans priced between 1.75 and 1.99 percentage points over APOR. The commenter also cited Urban Institute analysis of loans from 2001 to 2004 and 2005 to 2008 and pointed to loans priced between 1.51 and 2.0 percentage points over APOR having 90-day delinquency rates of 20.4 percent and 29.2 percent, respectively.^[206] The commenter asserted that this undercuts the Bureau's theory that creditors accurately assess and price for risk throughout the business cycle and indicates that the proposal would extend a presumption of compliance with the ATR provisions to loans that are not affordable.

That consumer advocate commenter disagreed with the Bureau's analysis using 60-day delinquency rates during the first two years of the loan as a measure of ability to repay because the commenter asserted that consumers tend to forgo other expenses ^[207] and take extreme measures to make timely mortgage payments, even if the loan was not affordable at consummation. This commenter argued that TILA requires assessment of a consumer's ability to repay the mortgage and still meet other obligations and cover basic living expenses. The commenter argued that the fact that a consumer was not 60 days or more past due on their mortgage does not answer the question of whether the loan was affordable at consummation. The commenter requested that the Bureau examine correlations between mortgage originations and delinquencies on other types of credit obligations that are visible in credit reporting data to assess the extent to which mortgages at various price and DTI levels are consistent with an assessment of the consumer's ability to repay. That commenter further asserted that default has more to do with macroeconomic conditions than individual ability to repay.

An industry commenter asserted that the Bureau failed to examine the effect of a DTI limit on mortgage performance by property type. The commenter asserted that community association housing ^[208] is unique from other housing models in that homeowners are required to pay assessments for community operations and that consumers' DTI may increase if community association costs increase. The commenter provided analysis of the percentage of loans 180 days delinquent by DTI bin, using Fannie Mae Condominium Unit Mortgages from 2002-2008 and 2015-2019. The commenter asserted that the analysis shows that, within the sample, “high DTI” loans have higher 180-day delinquency rates and the difference in delinquency rate is significant. The commenter asserted that this is evidence that reasonable DTI requirements are important for condominium unit mortgages and urged the Bureau to study the relationship between high DTI ratios, property type, and delinquency prior to issuing the final rule or to expand its analysis to include property type as a variable in testing the effectiveness of pricing as a measure of ability to repay.

Some commenters, including a research center commenter, a consumer advocate commenter, and two academic commenters, raised concerns that the price-based approach would be pro-cyclical. Some commenters that criticized the proposal as pro-cyclical expressed concern that the price-based General QM loan definition could grant QM status to loans exceeding consumers' ability to repay during periods of economic expansion, lead to increased housing prices, and create systemic risk. Similarly, some commenters that criticized the proposed approach expressed concern that removing the DTI limit would remove a constraint on housing prices. These commenters generally asserted that increased housing prices could increase consumers' mortgage payments and thereby increase the likelihood that consumers would be unable to afford their loan. These commenters further asserted that increased housing prices would prevent some consumers from obtaining loans altogether. For these reasons, these commenters asserted that the price-based approach could have a negative effect on access to credit for some consumers. These commenters also asserted that the pro-cyclical nature of the price-based approach could disproportionately affect underserved borrowers, including minority consumers.

An academic commenter expressed concern that the Bureau's delinquency analysis does not reflect the full extent of rate compression. That commenter criticized the Bureau's delinquency analysis of 2002-2008 first-lien purchase originations in the NMDB (Tables 1, 3, and 5 in the proposal and above), asserting that the analysis incorrectly assumes that rate spreads remained constant during that seven year period. The commenter stated that the Bureau should analyze rate spreads and associated default risk by vintage year, citing analysis showing that rate spreads fell significantly between 2004 and 2006 and suggesting that the Bureau's analysis therefore underestimates early delinquency rates at the height of the subprime mortgage boom. The commenter also criticized the Bureau's delinquency analysis of 2018 HMDA data (Tables 2, 4, and 6 in the proposal and above) as not informative because they do not cover two full years and are not indicative of bubble conditions. Another academic commenter analyzed a dataset of primarily subprime loans that were securitized in private-label securitizations during the housing bubble of the 2000s. The commenter stated that, in that dataset, over half of the subprime loans made between 2003 to 2005 had rate spreads that would satisfy the proposed rate-spread test for QM status. The commenter asserts that the data show that pricing as a measure of ability to repay fails when there is a credit boom due to rate spread compression and urged the Bureau to retain a DTI limit and consider an LTV ratio requirement as well as part of the General QM loan definition.

Other commenters, including commenters that supported the proposed price-based approach, expressed concerns about fluctuations in rate spreads over time. An industry commenter and a research center commenter suggested that the Bureau evaluate the rate-spread thresholds periodically and on an as-needed basis to determine if adjustments to the thresholds may be necessary to accommodate changing market and economic conditions. These commenters cited the rapidly changing market conditions at the beginning of the COVID-19 pandemic as an example of why it may be necessary to periodically adjust rate spreads. A consumer advocate commenter urged the Bureau not to adopt a mechanism that would allow the Bureau to adjust the rate-spread thresholds in emergency situations without notice and comment rulemaking.

Some commenters that did not support the price-based approach argued that the approach would not achieve the Bureau's stated goals of maintaining access to responsible, affordable mortgage credit. A research center commenter cited the January 2013 Final Rule, including the General and Temporary GSE QM loan definitions, as pro-cyclically supporting the current home price boom by providing additional leverage to consumers to bid up home prices. The commenter stated that this disproportionately affects the housing markets for low-income households and entry-level homes, where the supply is the tightest and the increase in leverage has been the greatest. The commenter disagreed with the Bureau's assertion that a DTI limit would unduly restrict access to credit, as the commenter asserts that a DTI limit would provide friction during a housing boom, which would reduce demand and slow house price appreciation. The commenter stated that the proposed price-based approach would not achieve the Bureau's goal of expanding access to credit because it would be even more pro-cyclical, resulting in higher house price appreciation. The commenter asserted that the proposed price-based approach does not provide any friction to slow house price appreciation and would boost demand more than the current rule, including the Temporary GSE QM loan definition. The commenter stated that the average rate spread for 2018 GSE purchase loans was 0.51 basis points, and asserted that creditors can therefore loosen lending standards and increase rate spreads over the foreseeable future with the resulting loans remaining below the 1.5 percentage point safe harbor threshold. The commenter also noted concern that the proposal would lower the QM standard and fuel higher risk leverage.

Some commenters specifically expressed concerns that the proposed rule would disproportionately harm minority consumers. For example, one commenter asserted that by replacing the DTI requirement with a pricing threshold, the proposed rule would subject higher percentages of Black or Hispanic borrowers to higher default rates. Another commenter stated that the proposal would burden borrowers of color with higher mortgage costs without underwriting and repayment ability assessment protections. Some commenters suggested that the proposed rule is fundamentally flawed because it may subject minority borrowers to higher prices that are unrelated to their actual risk due to ongoing discrimination in the market. Commenters urged the Bureau to assess and empirically evaluate the extent to which there is fair lending risk created by and embedded in its proposed pricing thresholds for QMs before adopting any final rule. One commenter suggested the Bureau disaggregate its analysis to assess the extent to which, at any given price band (and especially at the margins), early delinquency rates are consistent for non-Hispanic White, Black, and Hispanic consumers.

Some commenters (including industry commenters, consumer advocate commenters, and two joint industry and consumer advocate commenters that supported the proposed price-based approach) expressed concern about the connection between the price-based General QM loan definition and fair lending laws, including the Equal Credit Opportunity Act ^[209] (ECOA) and the Fair Housing Act.^[210] These commenters stated that pricing discrimination contravenes the underlying tenet of the General QM Proposal that if a consumer is purely priced on the true level of risk and ability to repay, the rate charged to the consumer is an indicator of risk—in the event of discriminatory pricing on a prohibited basis, the rate charged to the consumer is not a true indicator of risk. The commenters urged the Bureau to (1) make clear that it will not tolerate pricing discrimination or other forms of bias in the lending process and (2) limit the ability of a financial institution to receive the QM safe harbor in instances where pricing discrimination has occurred. Some of these commenters asked the Bureau to articulate explicitly that the designation of a loan as a QM does not signify compliance with the Fair Housing Act, ECOA, or any other anti-discrimination law pertaining to mortgage lending. Other commenters further requested that the rule specifically condition a General QM's safe harbor status on compliance with ECOA. These commenters requested that the rule provide that a loan loses its QM safe harbor status if there is a confirmed instance of discriminatory pricing on a prohibited basis that is not self-reported and remedied by the creditor.

A research center commenter, as well as an individual commenter, argued that the proposed approach would disproportionately affect minority consumers, which the commenters asserted would be a violation of the Fair Housing Act. In particular, the commenters described analysis indicating that increased housing prices that occur during periods of economic expansion (which the commenters asserted would be exacerbated as a result of the price-based General QM loan definition) occur predominately in areas with lower-income consumers, with higher concentrations of minority consumers. The commenters further asserted that the price-based approach would stimulate greater availability of credit which, combined with increased home prices, would expose low-income households, especially minority consumers, to heightened risk of default through higher mortgage payments. The commenters asked the Bureau to implement a multi-factor approach that combines DTI ratio, LTV ratio, and credit score as the key regulatory component of the General QM loan definition. The commenters argued that this approach would narrow the differential in delinquency rates between Black or Hispanic consumers and non-Hispanic White consumers when compared to delinquency rates under the proposed price-based approach.

Most commenters that did not support the proposed price-based approach advocated for alternative approaches to the General QM loan definition, such as retaining a DTI-based definition, a hybrid approach based on DTI and pricing, or a multi-factor approach. Several commenters supported a DTI-based approach rather than an approach based on pricing. Some commenters, including an academic commenter, industry commenters, and consumer advocate commenters, asserted that DTI is more reflective of a consumer's ability to repay than a loan's price, which includes factors that are not related to the specific consumer. For example, an academic commenter argued that the rule should retain a DTI limit because a DTI limit is effective in containing default risk. This commenter asserted that the Bureau should increase the DTI limit above 43 percent, should further expand the DTI limit for GSE mortgage programs that have an established track record of safe loans, and should amend appendix Q to provide more flexible methods for determining DTI. Other commenters advocating for a DTI-based approach suggested that the Bureau raise the current 43 percent limit. An industry commenter advocating for a DTI-based approach suggested retaining the current 43 percent DTI limit. Another industry commenter suggested that the Bureau retain a DTI limit for General QMs and raise the threshold to 50 percent with compensating factors, such as allowances for lower LTV ratios and for verified assets. That commenter also suggested that residual income be permitted as a compensating factor for a high DTI ratio but did not favor allowing residual income as a substitute for a DTI determination. As described above, several commenters advocating for the price-based General QM loan definition criticized a DTI-based General QM loan definition.

Other commenters advocated for a hybrid approach to the General QM loan definition. Some commenters, including a consumer advocate commenter and industry commenters, advocated for an approach that would raise the DTI ratio limit and also would expand the General QM loan definition to include loans with higher DTI ratios if the loans are below a set pricing threshold. For example, an industry commenter suggested that the Bureau impose a DTI limit of 47 percent but allow a General QM to have a DTI ratio of up to 50 percent if the rate spread is less than 2 percentage points. Another industry commenter suggested a hybrid approach that would retain the current DTI-based approach for higher-priced loans. Commenters advocating for hybrid approaches generally asserted that such approaches would better balance ensuring consumers have the ability to repay with ensuring access to responsible, affordable mortgage credit than a General QM loan definition based on pricing alone. An industry commenter advocated for an alternative method of defining General QMs that would use a DTI limit of 45 to 48 percent, in addition to the price-based approach. As noted above, a research center commenter suggested the Bureau define General QMs by reference to a multi-factor approach that combines DTI ratio, LTV ratio, and credit score. Other commenters argued against hybrid approaches, including noting concerns about the complexity of such approaches and concerns generally related to retaining a specific DTI component to the rule.

Commenters also raised issues related to the timing of the rulemaking and the issuance of the final rule. Some consumer advocate commenters and an individual commenter requested that the Bureau pause the rulemaking in light of the COVID-19 pandemic. Consumer advocate commenters requesting the Bureau pause the rulemaking cited the turmoil and economic fallout from the pandemic and the rising calls for racial justice as reasons to pause the rulemaking. The individual commenter and consumer advocate commenters raising this issue suggested that the Bureau focus its efforts on assisting homeowners struggling due to the pandemic. An industry commenter asserted that the Bureau should extend the Temporary GSE QM loan definition while it undertakes a study of alternative measures to evaluate consumers' ability to repay, such as residual income or cash flow underwriting (e.g., using bank account data that can identify the source and frequency of recurring deposits and payments and identify remaining disposable income).

An academic commenter stated that the Bureau should not address the Temporary GSE QM loan definition until the final resolution of the GSEs' status. That commenter also expressed concerns that the elimination of the Temporary GSE QM loan definition would set off a housing crisis by making homeownership unattainable for some consumers and risky for others if the GSEs respond to the elimination of the Temporary GSE QM loan definition by retreating from a substantial segment of the market. Another industry commenter expressed concern about the provision of the Temporary GSE QM loan definition that provides that the definition expires with respect to a GSE when that GSE ceases to operate under conservatorship. The commenter recommended that the Bureau remove this conservatorship clause. The commenter noted that the status of the conservatorships is outside of the Bureau's control and stated that, if one or both conservatorships were to end on short notice, the sudden expiration of the Temporary GSE QM loan definition would create uncertainty in the market and reduce access to credit. The commenter stated that the Bureau should clarify in advance of the end of conservatorship what steps the Bureau would take with respect to the Temporary GSE QM loan definition if the conservatorships were to end.

A research center commenter suggested that the Bureau consider the proposed changes to the QM rule in conjunction with the more recent Seasoned QM Proposal. The commenter suggested that the Bureau should consider additional analysis to study the interplay between default rates, rate-spread thresholds, loan products, and seasoning periods. The commenter asserted that, to the extent the seasoning proposal has implications for the General QM loan definition (or vice versa), a combined evaluation of both proposals would be more accurate than assessing the proposals separately.

The Final Rule

The Bureau concludes that this final rule's bright-line pricing thresholds best balance consumers' ability to repay with ensuring access to responsible, affordable mortgage credit. The Bureau is amending the General QM loan definition because retaining the existing 43 percent DTI limit would reduce the size of the QM market and likely would lead to a significant reduction in access to responsible, affordable credit when the Temporary GSE QM definition expires. The Bureau continues to believe that General QM status should be determined by a simple, bright-line rule to provide certainty of QM status, and the Bureau concludes that pricing achieves this objective. Furthermore, the Bureau concludes that pricing, rather than a DTI limit, is a more appropriate standard for the General QM loan definition. While not a direct measure of financial capacity, loan pricing is strongly correlated with early delinquency rates, which the Bureau uses as a proxy for repayment ability. The Bureau concludes that conditioning QM status on a specific DTI limit would likely impair access to credit for some consumers for whom it is appropriate to presume their ability to repay their loans at consummation. Although a pricing limit that is set too low could also have this effect, compared to DTI, loan pricing is a more flexible metric because it can incorporate other factors that may also be relevant to determining ability to repay, including credit scores, cash reserves, or residual income. The Bureau concludes that a price-based General QM loan definition is better than the alternatives because a loan's price, as measured by comparing a loan's APR to APOR for a comparable transaction, is a strong indicator of a consumer's ability to repay and is a more holistic and flexible measure of a consumer's ability to repay than DTI alone.

Specifically, the final rule amends Regulation Z to remove the current 43 percent DTI limit and provides that a loan would meet the General QM loan definition in § 1026.43(e)(2) only if the APR exceeds APOR for a comparable transaction by less than 2.25 percentage points as of the date the interest rate is set. As described further below, the Bureau is finalizing a threshold of 2.25 percentage points, an increase from the proposed threshold of 2 percentage points, because the Bureau concludes that, for most first-lien covered transactions, a 2.25-percentage-point pricing threshold strikes the best balance between ensuring consumers' ability to repay and ensuring access to responsible, affordable mortgage credit. The final rule provides higher thresholds for loans with smaller loan amounts and for subordinate-lien transactions.^[211] The final rule provides an increase from the proposed thresholds for some small manufactured housing loans to ensure continued access to credit.^[212] The Bureau is preserving the current threshold separating safe harbor from rebuttable presumption QMs, under which a loan is a safe harbor QM if its APR exceeds APOR for a comparable transaction by less than 1.5 percentage points as of the date the interest rate is set (or by less than 3.5 percentage points for subordinate-lien transactions).

The final rule requires the creditor to consider the consumer's monthly DTI ratio or residual income. The final rule also requires the creditor to consider the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan and the consumer's debt obligations, alimony, and child support, as described in the section-by-section analysis of § 1026.43(e)(2)(v)(A). The final rule removes appendix Q and, as described further below in the section-by-section analysis of § 1026.43(e)(2)(v)(B), provides creditors additional flexibility for defining the consumer's income or assets and debts. As discussed below, these amounts must be determined in accordance with § 1026.43(e)(2)(v)(B), which requires the creditor to verify the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan and the consumer's current debt obligations, alimony, and child support. The final rule provides a safe harbor to creditors using verification standards the Bureau specifies. Under the final rule, this safe harbor includes relevant provisions from Fannie Mae's Single Family Selling Guide, Freddie Mac's Single-Family Seller/Servicer Guide, FHA's Single Family Housing Policy Handbook, the VA's Lenders Handbook, and the Field Office Handbook for the Direct Single Family Housing Program and Handbook for the Single Family Guaranteed Loan Program of the USDA, current as of the proposal's public release. However, creditors are not required to verify income and debt according to the standards the Bureau specifies. The final rule provides creditors with the flexibility to develop other methods of compliance with the verification requirements.

Consistent with the proposal, the Bureau is not amending the existing product-feature and underwriting requirements and limits on points and fees. The statutory QM protections prohibit certain risky loan terms and features that could increase the risk that loans would be unaffordable and also include limited underwriting criteria that overlap with some elements of the ATR requirements. However, the Bureau concludes, as it initially concluded in the January 2013 Final Rule, that the General QM criteria should include additional assurances of a consumer's ability to repay to ensure that loans that obtain QM status warrant a presumption of compliance with the ATR requirements. The Bureau also continues to believe that creditors should be able to determine whether individual mortgage transactions will be deemed QMs through a bright-line metric.

In the January 2013 Final Rule, the Bureau exercised its authority under TILA section 129C(b)(2)(A)(vi) to impose a specific DTI limit as part of the General QM loan definition. The Bureau concludes that retaining the existing 43 percent DTI limit after the Temporary GSE QM loan definition expires would significantly reduce the size of the QM market and likely would reduce access to responsible, affordable mortgage credit. For the reasons described in part V.B.1, the Bureau believes that many loans currently originated under the Temporary GSE QM loan definition would cost materially more or may not be made at all, absent changes to the General QM loan definition. In particular, based on 2018 data, the Bureau estimated in the proposal that, as a result of the General QM loan definition's 43 percent DTI limit, approximately 957,000 loans—16 percent of all closed-end first-lien residential mortgage originations in 2018—would be affected by the expiration of the Temporary GSE QM loan definition. These loans are currently originated as QMs due to the Temporary GSE QM loan definition but would not be originated under the current General QM loan definition, and might not be originated at all, if the Temporary GSE QM loan definition were to expire. An additional, smaller number of loans that currently qualify as Temporary GSE QMs may not fall within the General QM loan definition after expiration of the Temporary GSE QM loan definition because the method used for verifying income or debt would not comply with appendix Q.

After the Temporary GSE QM loan definition expires, the Bureau expects that many consumers with DTI ratios above 43 percent who would have received a Temporary GSE QM would instead obtain FHA-insured loans if the 43 percent DTI limit remained in place. The Bureau estimated in the proposal that, in 2018, 11 percent of Temporary GSE QMs with DTI ratios above 43 percent exceeded FHA's loan-amount limit.^[213] Thus, the Bureau considers that at most 89 percent of loans that would have been Temporary GSE QMs with DTI ratios above 43 percent could move to FHA.^[214] The Bureau expects that loans that would be originated as FHA loans instead of under the Temporary GSE QM loan definition generally would cost materially more for many consumers, and that some consumers offered FHA loans might choose not to take out a mortgage because of these higher costs. Some consumers with DTI ratios above 43 percent would be able to obtain loans in the private market. The number of loans absorbed by the private market would likely depend, in part, on whether actors in the private market would be willing to assume the legal or credit risk associated with funding loans—as non-QM loans or small-creditor portfolio QMs—that would have been Temporary GSE QMs (with DTI ratios above 43 percent) ^[215] and, if so, whether actors in the private market would offer more lower prices or better terms.^[216] Finally, some consumers with DTI ratios above 43 percent who would have sought Temporary GSE QMs may make different choices, such as adjusting their borrowing to result in a lower DTI ratio, if the 43 percent DTI limit remained in place.^[217] However, some consumers who would have sought Temporary GSE QMs (with DTI ratios above 43 percent) may not obtain loans at all.^[218] For example, based on application-level data obtained from nine large lenders, the Assessment Report found that the January 2013 Final Rule eliminated between 63 and 70 percent of non-GSE eligible home purchase loans with DTI ratios above 43 percent.^[219]

As described in the proposal and above, the Bureau is now adopting a price-based approach to replace the specific DTI limit in the General QM loan definition because the Bureau concludes that a loan's price, as measured by comparing a loan's APR to APOR for a comparable transaction, is a strong indicator of a consumer's ability to repay. A loan's price is not a direct measure of ability to repay, but the Bureau concludes that it is an effective indirect indicator for ability to repay. The Bureau's delinquency analysis, analysis provided by commenters, and other analysis published in response to the Bureau's requests for comment, provide strong evidence that rate spreads distinguish loans that are likely to have low early delinquency rates, and thus should receive a presumption of compliance with the ATR requirements, from loans that are likely to have higher rates of delinquency, which should not receive that presumption. The Bureau finds this to be the case across a range of datasets, time periods, loan types, measures of rate spread, and measures of delinquency.

The Bureau acknowledged in the proposal that there is significant debate over whether a loan's price, a consumer's DTI ratio, or another direct or indirect measure of a consumer's personal finances is a better predictor of loan performance, particularly when analyzed across various points in the economic cycle. Some commenters argued that DTI ratios are a better predictor of default than a loan's price and therefore provide a better indicator of a consumer's ability to repay. However, as noted in the proposal, the Bureau is not determining whether DTI ratios, a loan's price, or some other measure is the best predictor of loan performance. Rather, the Bureau sought to balance considerations related to ensuring consumers' ability to repay and maintaining access to responsible, affordable credit in selecting the price-based approach, consistent with the purposes of the ATR/QM provisions of TILA. As noted, the Bureau's delinquency analysis, along with other available evidence, provide strong evidence that rate spreads can distinguish loans that are likely to have low early delinquency rates from loans that are likely to have higher rates of early delinquency. Further, maintaining access to responsible, affordable mortgage credit after the expiration of the Temporary GSE QM loan definition is a critical policy goal, and the Bureau finds that the price-based approach would also further this goal.

The Bureau further concludes that the price-based approach is a more holistic and flexible measure of a consumer's ability to repay than DTI alone, as described above and in the proposal. Mortgage underwriting, and by extension, a loan's price, generally includes an assessment of additional factors, such as credit scores and cash reserves, that might compensate for a higher DTI ratio and that might also be probative of a consumer's ability to repay. In contrast, the Bureau finds that a DTI limit may unduly restrict access to credit because it provides an incomplete picture of the consumer's financial capacity. In particular, and as described above, the Bureau concludes that conditioning QM status on a specific DTI limit would likely impair access to credit for some consumers for whom it is appropriate to presume ability to repay their loans at consummation. Further, and as described above in part V.B.2, there is inherent flexibility for creditors in a price-based QM definition, which will facilitate innovation in underwriting, including use of emerging research into alternative mechanisms to assess a consumer's ability to repay, such as cash flow underwriting. The Bureau concludes that the price-based approach best balances ability-to-repay considerations with ensuring continued access to responsible, affordable mortgage credit.

The Bureau is also concerned that including a specific DTI limit in the General QM loan definition would be in tension with the changes to the debt and income verification requirements in this final rule. As described in the section-by-section analysis of § 1026.43(e)(2)(v)(B) below, the Bureau is finalizing a revised approach for verifying debt and income in § 1026.43(e)(2)(v)(B) that provides flexibility for creditors to adopt innovative verification methods while also providing greater certainty that a loan has QM status. The revised verification approach allows creditors flexibility to use any reasonable verification method and criteria, provided that the creditor verifies debt and income using reasonably reliable third-party records. The final rule provides a safe harbor for creditors that use specific versions of manuals listed in commentary and provides that creditors also obtain a safe harbor if they “mix and match” the verification standards in those manuals, or use revised versions of the manuals that are “substantially similar” to the versions listed in the commentary. The Bureau is concerned that this verification approach, which provides flexibility to creditors in verifying debt and income, could create uncertainty if it were used in conjunction with a specific DTI limit. In particular, the Bureau is concerned that it could lead to disagreement among market participants over whether the DTI ratio for a given loan is above or below the limit and therefore whether the loan is a QM, which could complicate the sale of loans into the secondary market and disrupt access to credit. The Bureau has not identified verification approaches that, if used in conjunction with a specific DTI limit, would provide sufficient certainty to creditors, investors, and assignees regarding a loan's QM status and also provide flexibility to creditors in order to preserve access to responsible, affordable mortgage credit.

The Bureau also concludes that the price-based approach will ensure continued access to responsible, affordable mortgage credit after the expiration of the Temporary GSE QM loan definition. As described above, the proposal provided analysis of the potential effects on access to credit of a price-based approach to defining a General QM using 2018 HMDA data to estimate the percentage of conventional first-lien purchase loans within various price-based safe harbor and General QM thresholds. The Bureau has adjusted that analysis for the final rule to account for the final rule's higher pricing threshold for some small manufactured home loans, discussed below in the section-by-section analysis of § 1026.43(e)(2)(vi). The Bureau has also adjusted its analysis to reflect a revised methodology to identify creditors eligible to originate QMs as small creditors under § 1026.43(e)(5). Specifically, the Bureau lacks data on assets for certain non-depository creditors. The revised methodology estimates that such lenders have assets over $2 billion if their volume of 2018 HMDA originations not reported as sold exceeds $400 million. This revised methodology slightly reduces the estimated number of creditors eligible to originate QMs as small creditors as compared to the proposal's estimates. Specifically, a small number of non-depository creditors who primarily report loans as not sold (e.g., several creditors that specialize in manufactured home lending) are now estimated to be ineligible to originate QMs as small creditors. These adjustments are all reflected in Table 7A. Table 7A also provides an estimate of the percentage of loans under the pricing thresholds of 1.5 percent above APOR (safe harbor) and 2.25 above APOR (rebuttable presumption) adopted in this final rule.

Table 7A—Final Rule's Share of 2018 Conventional First-Lien Purchase Loans Within Various Price-Based Safe Harbor (SH) QM and Rebuttable Presumption (RP) QM Definitions

[HMDA data]

Approach	Safe harbor QM (share of conventional purchase market)	QM overall (share of conventional purchase market)
Temporary GSE QM + DTI 43	89.6	94.7
Final Rule (SH 1.50, RP 2.25)	91.3	96.3
SH 0.75, RP 1.50	74.2	93.9
SH 1.00, RP 1.50	83.1	93.9
SH 1.25, RP 1.75	88.1	95.0
SH 1.35, RP 2.00	89.6	95.8
SH 1.40, RP 2.00	90.2	95.8
SH 1.50, RP 2.00	91.3	95.8
SH 1.75, RP 2.25	92.8	96.3
SH 2.00, RP 2.50	93.9	96.6

As discussed further below, the Bureau is maintaining the current safe harbor threshold for QMs, such that a loan is a safe harbor QM if its APR does not exceed APOR for a comparable transaction by 1.5 percentage points or more as of the date the interest rate is set (or by 3.5 percentage points or more for subordinate-lien transactions). As discussed in the section-by-section analysis of § 1026.43(e)(2)(vi)(A), the Bureau is adopting a threshold of 2.25 percentage points over APOR for transactions with a loan amount greater than or equal to $110,260 (indexed for inflation).^[220] As shown in Table 7A, under these thresholds and using the 2018 HMDA data, 91.3 percent of conventional purchase loans would have been safe harbor QMs and 96.3 percent would have been safe harbor QMs or rebuttable presumption QMs.

As discussed above in part V.B.3, the Bureau also analyzed the potential effects of a DTI-based approach on the size of the QM market, as reflected in Table 8 in the proposal and above. For comparison, the Bureau has also adjusted that analysis to reflect the revised methodology, discussed above, to identify creditors eligible to originate QMs as small creditors under § 1026.43(e)(5). These adjustments are reflected in Table 8A.

Table 8A—Final Rule's Share of 2018 Conventional Purchase Loans Within Various Safe Harbor QM and Rebuttable Presumption QM Definitions (HMDA Data) Under the Final Rule

Approach	Safe harbor QM (share of conventional market)	QM overall (share of conventional market)
Temporary GSE QM + DTI 43	89.6	94.7
Final Rule (Pricing at 2.25)	91.3	96.3
DTI limit 43	68.9	73.1
DTI limit 45	75.7	80.5
DTI limit 46	78.5	83.5
DTI limit 47	81.1	86.3
DTI limit 48	83.8	89.1
DTI limit 49	86.7	92.2
DTI limit 50	90.5	96.3

As noted above, some commenters stated that the proposed price-based approach would expand access to credit for certain underserved market segments, such as low-income and minority consumers. At the same time, some commenters, including a consumer advocate commenter, expressed concern that a price-based approach would curtail access to credit for some low-income and minority consumers because these consumers are more likely to have mortgages with higher rate spreads. The Bureau concludes that the thresholds in the final rule best balance considerations related to ability to repay while retaining access to responsible, affordable mortgage credit, including for minority consumers. In particular, using 2018 HMDA data that was used in the proposal to estimate the size of the QM market under various pricing thresholds, the Bureau estimates that 96.8 percent of conventional purchase loans to minority consumers would receive QM status under the final rule, compared to 94.9 percent under the current rule with the Temporary GSE QM loan definition and the General QM loan definition with a DTI limit of 43 percent, or 67.9 percent under only a General QM loan definition with a DTI limit of 43 percent. Under the proposed price-based thresholds, 95.5 percent of conventional purchase loans to minority consumers would have received QM status.

Finally, the Bureau concludes that a price-based General QM loan definition will provide compliance certainty to creditors because they will be able to readily determine whether a loan is a General QM. As described above, creditors have experience with APR calculations due to the existing price-based regulatory requirements and for various other disclosure and compliance reasons under Regulation Z. Creditors also have experience determining the appropriate APOR for use in calculating rate spreads. As such, the Bureau concludes that the price-based approach will provide certainty to creditors regarding a loan's status as a QM.

The Bureau acknowledges that a small percentage of loans eligible for General QM status under the current rule would be ineligible for General QM status under the final rule. Specifically, those are loans with DTI ratios below 43 percent and that otherwise satisfy the current General QM loan definition that are priced above the rate-spread thresholds established by the final rule (e.g., 2.25 percentage points or higher for a first lien transaction with a loan amount greater than or equal to $110,260 (indexed for inflation)). As described below in the Dodd-Frank Act section 1022(b) analysis, the Bureau expects that creditors may adjust the price of some of these loans to meet the General QM pricing thresholds under the final rule. For other loans, creditors may instead originate those loans as non-QM loans or under other QM definitions, including as FHA loans, although the Bureau acknowledges that consumers may pay higher costs for these loans. The Bureau further acknowledges that some consumers who would be eligible for a General QM under the current rule but not under the final rule's pricing thresholds may be unable to obtain a mortgage, although the Bureau expects that the number of such consumers will be small. As shown in Table 8A and discussed further below in the section-by-section analysis of § 1026.43(e)(2)(vi), the final rule represents an overall expansion of loans eligible for General QM status relative to the current definition. Further, and as the Bureau observed in the January 2013 rule, it is not possible to define by a bright-line rule a class of mortgages for which each consumer will have ability to repay.^[221] The Bureau's decision to adopt a price-based approach reflects an appropriate balance of credit access and ability-to-repay considerations, taking into account the most efficient and effective means to ensure compliance.

The Bureau also acknowledges comments suggesting that a test that combines rate spread and DTI may better predict early delinquency rates than either metric on its own. However, the Bureau's concerns about a DTI-based approach also apply to these hybrid approaches. The Bureau agrees with commenters asserting that hybrid approaches would be unduly complex and are not necessary given that price is also strongly correlated with loan performance, as described above. The Bureau also concludes that multi-factor approaches suggested by commenters are complex and unnecessary given that price is strongly correlated with loan performance.

One commenter criticized the price-based approach based on analysis showing that for loans with identical rate spreads, default occurrences vary, and for loans with similar default occurrences, the rate spreads vary greatly. The Bureau disagrees that such a finding shows that price is not an effective indicator of a consumer's ability to repay. The commenter's analysis shows that pricing and the commenter's preferred risk metric are both correlated with early delinquency, even when holding the other metric fixed. This only demonstrates that neither metric is perfectly correlated with early delinquency and that each metric is predictive of early delinquency independently of the other. The Bureau has concluded that pricing is an effective indicator of a consumer's ability to repay in part because it is strongly correlated with early delinquency, based on the Bureau's delinquency analysis and external analysis described above, recognizing that there is not a perfect correlation between price and early delinquency. However, there also is not a perfect correlation between early delinquency and DTI, nor between early delinquency and the alternative measures proposed by commenters. Because many different factors are correlated with early delinquency, the Bureau expects that, even at a fixed level of one potential measure of a consumer's ability to repay, early delinquency rates will still vary with other factors. While multi-factor approaches that incorporate additional variables may achieve higher correlations with early delinquency, such approaches are more complex and may involve greater prescriptiveness.

As noted above, a consumer advocate commenter expressed concern about the use of 60-day early delinquency rates in the first two years of a mortgage to measure ability to repay. That commenter raised concerns that mortgage payments may not be affordable but consumers may forgo paying other expenses so that they are able to continue making timely mortgage payments. The Bureau acknowledges that this may occur for some consumers, consistent with the Experian analysis cited by the consumer advocate commenter which showed that consumers with a mortgage and other credit obligations were less likely to be delinquent on their mortgage than on their other credit obligations.^[222] However, the Bureau believes that, as a general matter, 60-day early delinquencies in the first two years is an appropriate metric to measure ability to repay. Moreover, the Bureau notes that an analysis provided by a research center commenter, described above, measured loan performance by whether the consumer was ever 60 days or more delinquent, rather than by reference to the two-year period used in the Bureau's delinquency analysis. The commenter noted that its analysis also found delinquency to be highly correlated with rate spreads, when delinquency is measured over the life of the loan.

As noted above, some comments asserted that pricing is not an appropriate QM criterion because it reflects risk of loss to the creditor and not the consumer's ability to repay the loan. The proposal recognized that there is a distinction between credit risk, which largely determines pricing relative to APOR, and a particular consumer's ability to repay, which is one component of credit risk. While a consumer's ability to afford loan payments is an important component of pricing, the loan's price will reflect additional factors related to the loan that may not in all cases be probative of the consumer's repayment ability. While the Bureau recognizes these concerns about a price-based approach, the Bureau's delinquency analysis and the analyses by external parties discussed above provide evidence that rate spreads are correlated with delinquency. Further, the Bureau notes that the final rule includes a requirement to consider the consumer's DTI ratio or residual income as part of the General QM loan definition, and to verify the debt and income used to calculate DTI or residual income. These requirements are discussed further below in the section-by-section analysis of § 1026.43(e)(2)(v)(A) and are included in the General QM loan definition to further ensure that, consistent with the purposes of TILA, creditors appropriately consider consumers' financial capacity and that consumers are thus offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loan.

Similarly, some commenters raised concerns that factors unrelated to the consumer, or the individual loan, can influence the price of a loan and that a price-based approach may be more consistent with some business models than others. Some commenters also raised concerns that a price-based approach is variable and that whether a consumer receives a General QM under the price-based approach may vary by creditor. While the Bureau acknowledges these criticisms of a price-based approach, the Bureau's delinquency analysis and the analyses by external parties discussed above provide evidence that rate spreads are correlated with delinquency, across a range of datasets, time periods, loan types, measures of rate spread, and measures of delinquency.

The Bureau also recognizes concerns that a price-based approach may incentivize some creditors to price some loans just below the threshold so that the loans will receive the presumption of compliance that comes with QM status. The proposal acknowledged that creditors are likely to react to the final rule by adjusting the price of some loans they offer to fall just below the threshold separating QMs from non-QM loans. To the extent creditors offer loans at lower prices to obtain QM status under the final rule, consumers will pay less for those loans. Those loans would also be subject to the QM product-feature restrictions and limits on points and fees, which would provide a benefit to consumers who might have otherwise received a non-QM loan that included a more risky product feature or included points and fees above the QM limits. The Bureau does not expect significant changes in loan pricing as a result of the safe harbor threshold, which exists under the current ATR/QM Rule. The Bureau points to research cited by some commenters, which suggests that, while creditors reacted to the safe harbor pricing threshold in the January 2013 Final Rule by reducing the share of higher-priced mortgages that they originated, the economic significance of the response was minor and did not materially affect the mortgage market at the time the rule took effect.^[223]

The Bureau disagrees with the comment asserting that the price-based approach is inappropriate because LTV ratios are a component of pricing. Nothing in the statutory text of TILA prohibits the Bureau from adopting the price-based approach. Indeed, TILA provides the Bureau with considerable flexibility to determine the appropriate criteria to define QM and to adjust the statutory QM requirements as necessary or proper to achieve Congress's objectives. The Bureau's authority with respect to defining QMs is discussed above in part IV. TILA section 129C(b)(2)(A)(vi) provides the Bureau with authority to establish guidelines or regulations relating to ratios of total monthly debt to monthly income or alternative measures of ability to pay regular expenses after payment of total monthly debt, taking into account the income levels of the borrower and such other factors as the Bureau may determine relevant and consistent with the purposes described in TILA section 129C(b)(3)(B)(i). TILA section 129C(b)(3)(B)(i) authorizes the Bureau to prescribe regulations that revise, add to, or subtract from the criteria that define a QM upon a finding that such regulations are necessary or proper to ensure that responsible, affordable mortgage credit remains available to consumers in a manner consistent with the purposes of TILA section 129C; or are necessary and appropriate to effectuate the purposes of TILA sections 129B and 129C, to prevent circumvention or evasion thereof, or to facilitate compliance with such sections. In addition, TILA section 129C(b)(3)(A) directs the Bureau to prescribe regulations to carry out the purposes of section 129C.

The Bureau finds that the price-based approach is consistent with this authority and with the purposes of TILA and section 129C's presumption of compliance with the ATR requirements for QMs. TILA sections 129B and 129C do not suggest that, in prohibiting creditors from considering the consumers' equity in the property securing the transaction as a financial resource to repay the loan, Congress intended to limit the Bureau's authority to impose loan pricing restrictions that, if incorporated into the QM definition, would provide sufficient assurance of the consumer's ability to repay. The Dodd-Frank Act amendments to TILA rely on pricing thresholds to distinguish between and among categories of QM and non-QM loans that should receive heightened consumer protections.^[224] And, as described above, Dodd-Frank amendments to TILA in part codify and expand a pre-existing HOEPA regime that relied on pricing for similar purposes. Further, the Bureau notes that under this final rule creditors must consider the consumer's monthly DTI ratio or residual income; current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan; and debt obligations, alimony, and child support to satisfy the General QM loan definition.^[225] In light of this requirement, including the exclusion of the value of the dwelling that secures the loan from the assets the creditor may consider for purposes of this requirement,^[226] the Bureau concludes that the price-based approach is consistent with TILA section 129C(a)(3). For these reasons, and consistent with the statutory text, structure and purposes of the TILA, the Bureau concludes that it is an appropriate use of its authority to include a loan's price as one criterion to define General QMs.

With respect to commenters expressing concern about the sensitivity of a price-based General QM loan definition to macroeconomic cycles, the Bureau acknowledged this concern in the proposal. The proposal noted that periods of economic expansion, increasing house prices, and strong demand from consumers with weaker credit characteristics often lead to greater availability of credit. This is because as house prices increase, home equity also increases, and secondary market investors expect fewer losses accordingly. Even if a consumer were to default, increasing collateral values make it more likely that the investors would still recover the full amount of their investment. This increased likelihood of recovery may result in an underpricing of credit risk. To the extent such underpricing occurs, rate spreads over APOR would compress and additional higher-priced, higher-risk loans would fit within the proposed General QM loan definition. Further, the proposal recognized that, during periods of economic downturn, investors' demand for mortgage credit may fall as they seek safer investments to limit losses in the event of a broader economic decline. This may result in creditors reducing the availability of mortgage credit to riskier borrowers, through credit overlays and price increases, to protect against the risk that creditors may be unable to sell the loans profitably in the secondary markets, or even sell the loans at all. The proposal recognized that, while APOR would also increase during periods of economic stress and low secondary market liquidity, consumers with riskier credit characteristics may see disproportionate pricing increases relative to the increases in a more normal economic environment. These effects would likely make price-based QM standards pro-cyclical, with a more expansive QM market when the economy is expanding, and a more restrictive QM market when credit is tight. As a result, a rate spread-based QM threshold would likely be less effective than a binding DTI limit in deterring risky loans during periods of strong housing price growth or encouraging safe loans during periods of weak housing price growth. As described above, some commenters to the proposal highlighted these concerns and argued that the Bureau should not finalize the price-based approach due to potential systemic risks. However, the Bureau notes that a binding DTI limit risks restricting access to affordable credit relative to this final rule. The Bureau concludes that the advantages of the price-based approach in providing a flexible and holistic indicator of ability to repay outweigh the macroeconomic cycle concerns as considerations toward ensuring the availability of responsible, affordable mortgage credit. In addition, the Bureau believes that the QM product feature restrictions, the consider and verify requirements, and the final rule's special rule for ARMs mitigate some concerns regarding the pro-cyclical risks during economic expansions.

As noted, a commenter expressed concern that the Bureau's delinquency analysis does not reflect the full extent of rate compression. That commenter argued that the Bureau should analyze rate spreads and associated default risk by vintage year, citing analysis showing that rate spreads fell significantly between 2004 and 2006 and suggesting that the Bureau's analysis therefore may not capture potential declines in the correlation between price and early delinquency rates at the height of the subprime mortgage boom. With respect to this comment, the Bureau recognizes, as stated above, that there is not a perfect correlation between pricing and early delinquency rates. However, the Bureau has concluded that pricing is strongly correlated with early delinquency, based on the Bureau's delinquency analysis, external analysis described in the proposal, and analysis provided by commenters, which cover a wide range of years and economic conditions.^[227] With respect to other commenters that expressed concerns about fluctuations in rate spreads over time, the Bureau recognizes that overall market spreads expand and tighten over time, as described above.^[228] The Bureau concludes the pricing thresholds in the final rule provide the best balance between ability-to-repay considerations and ensuring access to responsible, affordable mortgage credit. The Bureau further notes that it monitors changing market and economic conditions and it could consider changes to the thresholds if circumstances warrant.

With respect to commenters that expressed concern about the connection between the price-based General QM loan definition and fair lending laws, including ECOA and the Fair Housing Act, the Bureau recognizes that some creditors may violate Federal fair lending laws by charging certain borrowers higher prices on the basis of race or national origin compared to non-Hispanic White borrowers with similar credit characteristics, and the Bureau reaffirms its commitment to consistent, efficient, and effective enforcement of Federal fair lending laws.^[229] The Bureau further emphasizes that the General QM loan definition, as amended by this final rule, does not create an inference or presumption that a loan satisfying the General QM loan definition is compliant with any Federal, State, or local anti-discrimination laws that pertain to lending. A creditor has an independent obligation to comply with ECOA and Regulation B, and an effective way for a creditor to minimize and evaluate fair lending risks under these laws is by monitoring their policies and practices and implementing effective compliance management systems. The Bureau declines to amend the ATR/QM Rule to provide that a loan loses its QM safe harbor status if there is a confirmed instance of discriminatory pricing on a prohibited basis that is not self-reported and remedied by the creditor.

The Bureau disagrees with commenters who assert that the price-based General QM loan definition does not advance fair lending. As noted above, the Bureau concludes that conditioning QM status on a specific DTI limit may impair access to responsible, affordable credit for some consumers for whom it might be appropriate to presume ability to repay their loans at consummation. Specifically, using a bright-line DTI ratio threshold may have an adverse impact on responsible access to credit, including for low-to-moderate-income and minority homeowners. As discussed above, a price-based General QM loan definition is better than the alternatives because a loan's price, as measured by comparing a loan's APR to APOR for a comparable transaction, is a strong indicator of a consumer's ability to repay and is a more holistic and flexible measure of a consumer's ability to repay than DTI alone. The Bureau therefore expects that this final rule will improve access to credit for low-to-moderate-income and minority homeowners, without the unnecessary complexity of hybrid or multi-factor alternatives urged by some commenters.

With respect to the comment that provided analysis of loan performance for loans secured by condominiums and urged the Bureau to study the relationship between high DTI ratios, property type, and delinquency prior to issuing the final rule or expand its delinquency analysis to include property type as a variable, the Bureau declines to undertake that further analysis at this time. As described above, the Bureau has concluded that pricing is strongly correlated with early delinquency and is concerned that a DTI limit may have an adverse impact on responsible access to credit. The Bureau also notes that fees and special assessments imposed by a condominium, cooperative, or homeowners association are mortgage-related obligations that must be included in the calculation of the consumer's debt-to-income or residual income for purposes of § 1026.43(e)(2)(v)(A) and therefore are incorporated into the General QM loan definition. Further, mortgage creditors often account for the property type when pricing a mortgage, and the rate-spread threshold would thus capture any differential risk for such loans that is reflected in their price. However, the Bureau will monitor the effects of the General QM final rule to determine if future changes are necessary to ensure continued access to responsible, affordable credit, including for particular property types such as condominiums.

The Bureau also declines to eliminate the conservatorship clause of the Temporary GSE QM loan definition. As explained in the Extension Final Rule, when the Bureau adopted the January 2013 Final Rule, the FHFA's conservatorship of the GSEs was central to its willingness to presume that loans that are eligible for purchase, guarantee, or insurance by the GSEs would be originated with appropriate consideration of consumers' ability to repay.^[230] If the GSEs are not under conservatorship, the Bureau is concerned about presuming that loans eligible for purchase or guarantee by either of the GSEs have been originated with appropriate consideration of the consumer's ability to repay.

With respect to the comment that expressed concern about the expiration of the Temporary GSE QM loan definition in light of the current GSE loan market, the Bureau anticipates that the final rule will preserve access to credit relative to the status quo. In particular, the Bureau concludes the General QM loan definition's pricing thresholds included in this final rule, in conjunction with the debt and income verification provisions in § 1026.43(e)(2)(v)(B), will ensure continued access to responsible, affordable mortgage credit, including for loans that have historically been eligible for purchase by the GSEs. With respect to the comment suggesting the Bureau consider evaluating changes to the General QM loan definition and the Seasoned QM Proposal at the same time, the Bureau has considered the expected effects of both proposals and is issuing rules on both of these topics at the same time.

C. The QM Presumption of Compliance Under a Price-Based General QM Loan Definition

To address potential uncertainty regarding the reasonableness of some ability-to-repay determinations, all QMs provide creditors with a presumption of compliance with the ATR requirements. Lower-priced QMs provide a conclusive presumption of compliance (i.e., a safe harbor) whereas higher-priced QMs provide a rebuttable presumption of compliance.^[231] The proposal would have preserved the current § 1026.43(b)(4) pricing threshold that generally separates safe harbor QMs from rebuttable presumption QMs, such that a loan is a safe harbor QM if its APR exceeds APOR for a comparable transaction by less than 1.5 percentage points as of the date the interest rate is set (or by less than 3.5 percentage points for subordinate-lien transactions).^[232]

1. Considerations Related to the Safe Harbor Threshold

As stated in the proposal, in developing the approach to the presumptions of compliance for QMs in the January 2013 Final Rule, the Bureau first considered whether the statute prescribes if QMs receive a conclusive or rebuttable presumption of compliance with the ATR provisions. As discussed above in part II.A, TILA section 129C(b) provides that loans that meet certain requirements are “qualified mortgages” and that creditors making QMs “may presume” that such loans have met the ATR requirements. However, the statute does not specify whether the presumption of compliance means that the creditor receives a conclusive presumption or a rebuttable presumption of compliance with the ATR provisions. The Bureau noted that its analysis of the statutory construction and policy implications demonstrates that there are sound reasons for adopting either interpretation.^[233] The Bureau concluded that the statutory language is ambiguous and does not mandate either interpretation and that the presumptions should be tailored to promote the policy goals of the statute.^[234] The Bureau interpreted the statute to provide for a rebuttable presumption of compliance with the ATR provisions but used its adjustment and exception authority to establish a conclusive presumption of compliance for loans that are not “higher-priced covered transactions.” ^[235]

In the January 2013 Final Rule, the Bureau identified several reasons why loans that are not higher-priced loans (generally prime loans) should receive a safe harbor. The Bureau noted that the fact that a consumer receives a prime rate is itself indicative of the absence of any indicia that would warrant a loan-level price adjustment, and thus is suggestive of the consumer's ability to repay.^[236] The Bureau noted that prime rate loans have performed significantly better historically than subprime loans and that the prime segment of the market has been subject to fewer abuses.^[237] The Bureau noted that the QM requirements will ensure that the loans do not contain certain risky product features and are underwritten with careful attention to consumers' DTI ratios.^[238] The Bureau also noted that a safe harbor provides greater legal certainty for creditors and secondary market participants and may promote enhanced competition and expand access to credit.^[239] The Bureau determined that if a loan met the product and underwriting requirements for QM and was not a higher-priced covered transaction, there are sufficient grounds for concluding that the creditor satisfied the ATR provisions.^[240]

The Bureau in the January 2013 Final Rule pointed to factors to support its decision to adopt a rebuttable presumption for QMs that are higher-priced covered transactions. The Bureau noted that QM requirements, including the restrictions on product features and the 43 percent DTI limit, would help prevent the return of the lax lending practices of some lenders in the years before the financial crisis, but that it is not possible to define by a bright-line rule a class of mortgages for which each consumer will have ability to repay, particularly for subprime loans.^[241] The Bureau noted that subprime pricing is often the result of loan-level price adjustments established by the secondary market and calibrated to default risk.^[242] The Bureau also noted that consumers in the subprime market tend to be less sophisticated and have fewer options and thus are more susceptible to predatory lending practices.^[243] The Bureau noted that subprime loans have performed considerably worse than prime loans.^[244] The Bureau therefore concluded that QMs that are higher-priced covered transactions would receive a rebuttable presumption of compliance with the ATR provisions. The Bureau recognized that this approach could increase by a modest amount the litigation risk for subprime QMs but did not expect that imposing a rebuttable presumption for higher-priced QMs would have a significant impact on access to credit.^[245]

2. The Bureau's Proposal

The safe harbor threshold. The Bureau did not propose to alter the approach in the current ATR/QM Rule, under current § 1026.43(b)(4) and (e)(1)(i), of providing a conclusive presumption of compliance with the ATR requirements (i.e., a safe harbor) to loans that meet the General QM requirements in § 1026.43(e)(2) and for which the APR exceeds the APOR by less than 1.5 percentage points (or by less than 3.5 percentage points for subordinate-lien loans).^[246] In the proposal, when discussing the safe harbor threshold, the Bureau restated its preliminary conclusion that pricing is strongly correlated with loan performance and that pricing thresholds should be included in the General QM loan definition in §  1026.43(e)(2). The Bureau also preliminarily concluded that for prime loans, the pricing, in conjunction with the revised QM requirements in proposed § 1026.43(e)(2), provides sufficient grounds for supporting a conclusive presumption that the creditor complied with the ATR requirements. The Bureau further noted that, under the proposed price-based approach, creditors would be required to consider DTI or residual income for a loan to satisfy the requirements of the General QM loan definition. The Bureau also stated that a safe harbor for prime QMs appears to be supported by the better performance of prime loans compared to subprime loans, and by the potential benefits of greater competition and access to credit from the greater certainty and reduced litigation risk arising from a safe harbor. The Bureau tentatively concluded that the current safe harbor threshold of 1.5 percentage points for first liens restricts safe harbor QMs to lower-priced, generally less risky, loans while ensuring that responsible, affordable credit remains available to consumers. The Bureau stated its general belief that these same considerations support not changing the current safe harbor threshold of 3.5 percentage points for subordinate-lien transactions, which generally perform better and have stronger credit characteristics than first-lien transactions. The Bureau's proposal to address subordinate-lien transactions is discussed further below in the section-by-section analysis of § 1026.43(e)(2)(vi). For the reasons discussed below, this final rule is maintaining the current safe harbor thresholds in current § 1026.43(b)(4) and (e)(1)(i).

Rebuttable Presumption QMs. The proposal generally would have maintained the current ATR/QM Rule's rebuttable presumption of compliance with the ATR requirements for loans that exceed the safe harbor threshold but that otherwise meet the General QM requirements in § 1026.43(e)(2).^[247] The Bureau did not propose to revise § 1026.43(e)(1)(ii)(B), which defines the grounds on which the presumption of compliance that applies to higher-priced QMs can be rebutted. Section 1026.43(e)(1)(ii)(B) provides that a consumer may rebut the presumption by showing that, at the time the loan was originated, the consumer's income and debt obligations left insufficient residual income or assets to meet living expenses. The analysis considers the consumer's monthly payments on the loan, mortgage-related obligations, and any simultaneous loans of which the creditor was aware, as well as any recurring, material living expenses of which the creditor was aware. The Bureau stated in the January 2013 Final Rule that this standard was sufficiently broad to provide consumers a reasonable opportunity to demonstrate that the creditor did not have a good faith and reasonable belief in the consumer's repayment ability, despite meeting the prerequisites of a QM. At the same time, the Bureau stated that it believed the standard was sufficiently clear to provide certainty to creditors, investors, and regulators about the standards by which the presumption can successfully be rebutted in cases in which creditors have met the QM requirements. The Bureau also noted that the standard was consistent with the standard in the 2008 HOEPA Final Rule.^[248] Commentary to that rule provides, as an example of how its presumption may be rebutted, that the consumer could show “a very high debt-to-income ratio and a very limited residual income . . . depending on all of the facts and circumstances.” ^[249] The Bureau noted that, under the definition of QM that the Bureau was adopting, the creditor was generally not entitled to a presumption if the consumer's DTI ratio was “very high.” As a result, the Bureau focused on the standard for rebutting the presumption in the January 2013 Final Rule on whether, despite meeting a DTI test, the consumer nonetheless had insufficient residual income to cover the consumer's living expenses.^[250]

The Bureau did not propose to change the standard for rebutting the presumption of compliance with the ATR requirements and stated its belief that the existing standard continues to balance consumer protection and access-to-credit considerations. For example, the Bureau did not propose amending the presumption of compliance to provide that the consumer may use the DTI ratio to rebut the presumption of compliance by establishing that the DTI ratio is very high, or by establishing that the DTI ratio is very high and that the residual income is not sufficient. First, the Bureau tentatively determined that permitting the consumer to rebut the presumption by establishing that the DTI ratio is very high is not necessary because the existing rebuttal standard already incorporates an examination of the consumer's actual income and debt obligations (i.e., the components of the DTI ratio) by providing the consumer the option to show that the consumer's residual income—which is calculated using the same components—was insufficient at consummation. Accordingly, the Bureau anticipated that the addition of a DTI ratio to the rebuttal standard would not add probative value beyond the current residual income test in § 1026.43(e)(1)(ii)(B). Second, the Bureau anticipated that the addition of a DTI ratio as a ground to rebut the presumption of compliance would undermine compliance certainty to creditors and the secondary market without providing any clear benefit to consumers. The Bureau tentatively determined that the rebuttable presumption standard would continue to be sufficiently broad to provide consumers a reasonable opportunity to demonstrate that the creditor did not have a good faith and reasonable belief in the consumer's repayment ability, despite meeting QM standards. The Bureau did not receive comments regarding the grounds on which the presumption of compliance can be rebutted.

3. Comments on the Safe Harbor Threshold

The Bureau received several comments concerning the proposed 1.5-percentage-point safe harbor threshold. A joint comment from consumer advocates stated that, if the Bureau finalizes a price-based approach, the proposed threshold should not be increased. A GSE commenter supported the 1.5-percentage-point threshold and stated it would be equally supportive if the Bureau increases the threshold. Various commenters, including a research center and several consumer advocate and industry commenters, specifically recommended increasing the safe harbor threshold to 2 percentage points. Commenters generally acknowledged that delinquency rates for safe harbor QMs would increase as the pricing threshold increases but expressed differing views on whether the proposed threshold should nonetheless be increased to expand access to credit.

A joint comment from consumer advocates generally objected to a price-based approach but specifically stated that increasing the safe harbor threshold would not significantly increase access to credit. The joint comment stated that the ATR/QM Rule's 1.5-percentage-point threshold is consistent with the Board's 2008 HOEPA Final Rule, which offered only a rebuttable presumption—not a safe harbor—for loans priced 1.5 percentage points or more above APOR. The joint comment stated that in markets with less competition, including minority communities, creditors routinely face no downward pressure on prices and will charge consumers more than they would in a more competitive market. The joint comment stated that, in less competitive markets, the current 1.5-percentage-point safe harbor threshold has benefited consumers by providing some downward pressure on prices. Notwithstanding such creditor reticence to price loans beyond the safe harbor threshold, the joint comment stated that there has not been an actual difference in litigation risk (i.e., for rebuttable presumption QMs versus safe harbor QMs) that would reasonably justify increasing the threshold. The joint comment further stated that increasing the safe harbor pricing threshold would not expand consumers' access to credit but instead would facilitate creditors raising prices to take advantage of less competitive markets and result in the same consumers obtaining the same loans but at higher prices.

A research center generally objected to a price-based approach but also stated that increasing the safe harbor threshold would not have a significant impact on access to credit. Based on 2018 loan data, the commenter stated that the current pricing threshold has relatively little impact on originating rebuttable presumption QMs priced 1.5 percentage points or more above APOR. Moreover, the commenter stated that even for rebuttable presumption QMs, litigation risk would be significantly reduced by the proposed rule's income and debt verification safe harbor, as discussed in the section-by-section analysis of § 1026.43(e)(2)(v)(B).

Various commenters, including a research center and multiple consumer advocate and industry commenters, specifically recommended increasing the safe harbor threshold to 2 percentage points, arguing that it would achieve a better balance of ability to repay with access to credit. Several of those commenters referenced the research center's analysis of Fannie Mae and Black Knight McDash data and stated that a 2-percentage-point threshold would increase the delinquency rate for safe harbor QMs. However, that subset of commenters argued that the analysis showed that the increased delinquency rate would nonetheless remain low relative to delinquency rates experienced in the past 20 years. Those commenters stated that addressing access-to-credit concerns with a 2-percentage-point threshold would therefore strike an appropriate balance with ability-to-repay concerns. One consumer advocate commenter stated that delinquency rate improvement, relative to the Great Recession, is largely due to the effects of the Dodd-Frank Act, which has helped ensure stronger product protections, better underwriting, and improved income, employment, and asset verification and documentation. Citing an FHFA working paper that was also cited in the General QM Proposal,^[251] a joint comment from consumer advocate and industry groups stated that loans with non-QM features—including interest-only loans, ARM loans that combined teaser rates with subsequent large jumps in payments, negative amortization loans, and loans made with limited or no documentation of the borrower's income or assets—accounted for about half of the rise in risk leading up to the 2008 financial crisis and subsequent passage of the Dodd-Frank Act. Given that the delinquency rate would be low on a relative basis, these commenters stated that addressing access-to-credit concerns with a 2-percentage-point threshold would strike an appropriate balance with ability-to-repay concerns.

Multiple consumer advocate and industry commenters stated that, in contrast to safe harbor QMs, creditors generally are less willing to make rebuttable presumption QMs. These commenters stated that their unwillingness to make rebuttable presumption QMs is evidenced by 2019 HMDA data showing that less than 5 percent of conventional, first-lien purchase loans were priced 1.5 percentage points or more above APOR.^[252] Citing Board economists' analysis of 2014 HMDA data,^[253] a joint comment from consumer advocate and industry groups stated that creditors reduced the share of higher-priced mortgages that they originated in response to the ATR/QM Rule. A research center stated that, based on 2019 HMDA data, increasing the safe harbor threshold to 2 percentage points would have replaced 75,265 rebuttable presumption QMs with safe harbor QMs instead. The research center stated that, because safe harbor QMs would provide those loans' creditors with greater protection from litigation than rebuttable presumption QMs, it suspects that the reduction in litigation risk would result in better pricing for consumers. The research center, as well as multiple consumer advocate and industry commenters, stated that increasing the safe harbor threshold to 2 percentage points would improve access to credit by reducing racial and ethnic disparities while helping increase lending volumes for every racial and ethnic group.

Several industry commenters elaborated on how rebuttable presumption QMs present more litigation risk to creditors than safe harbor QMs. One commenter stated that—even if a creditor has, in fact, made a reasonable and good faith determination of a consumer's repayment ability at the time of consummation—a creditor could still find itself in court providing evidentiary proof should a consumer challenge a rebuttable presumption QM. As a general matter, another commenter stated that—even if a defendant ultimately prevails in court—legal determinations regarding “reasonableness” are expensive to defend as they often require time-consuming litigation, extensive discovery, and possibly a trial. Another commenter stated that—even among creditors that would ultimately prevail in court—some creditors will choose the expense of settling with plaintiffs, rather than incurring the greater expense of paying a legal team to continue defending in court. The commenter stated that the safe harbor's conclusive presumption of compliance is necessary to stop meritless ability‐to‐repay litigation as early as possible in the legal process and to eliminate the settlement value of such litigation. These industry commenters each stated that increasing the safe harbor threshold to 2 percentage points would help address the negative effect that litigation risk has on access to credit.

Various commenters, including a research center and multiple consumer advocate and industry commenters, stated that increasing the safe harbor threshold in the Bureau's ATR/QM Rule to 2 percentage points would create a more level playing field between conventional and FHA lending. These commenters stated that FHA's own QM rule provides creditors with a safe harbor if the loan's APR is no more than APOR plus the FHA annual mortgage insurance premium plus 115 basis points. These commenters further stated that the current FHA annual mortgage insurance premium is 85 basis points, such that the FHA's QM rule effectively has a 2-percentage-point-over-APOR threshold. Some comments, including one from a consumer advocate commenter and a joint comment from consumer advocate and industry groups, stated that the Bureau's current 1.5-percentage-point safe harbor threshold has the effect of steering consumers, including minority consumers, to FHA loans rather than conventional loans and thus limits consumer choice among lenders and product offerings. Those comments further stated that a smaller pool of lenders originate FHA loans and that in 2019 there were approximately 3,200 HMDA reporting lenders for conventional purchase loans versus approximately 1,200 HMDA reporting lenders for FHA purchase loans.

Various commenters, including a research center and multiple consumer advocate and industry commenters, also stated that rate spreads fluctuate over time and recommended that this final rule increase pricing thresholds as a buffer to absorb the pricing impact of future market changes. In particular, regarding FHFA's GSE capital rule,^[254] these commenters stated that it would require GSEs to maintain more capital as a precaution against riskier loans in their portfolio (i.e., risk-based capital requirements). These commenters stated that they expect spreads over APOR will likely increase for riskier borrowers as a result of the FHFA's rule. The research center also stated that spreads for refinance loans could widen relative to APOR in response to the additional loan-level price adjustment of 50 basis points on most Fannie Mae and Freddie Mac refinances, effective December 1, 2020. However, an industry commenter stated that such changes also affect APOR itself, which adds further uncertainty regarding the actual magnitude of any future changes to spreads over APOR.

4. The Final Rule

For the reasons discussed below, as proposed, the Bureau is maintaining the current safe harbor threshold in § 1026.43(b)(4), such that a loan is a safe harbor QM under § 1026.43(e)(1) if its APR does not exceed APOR for a comparable transaction by 1.5 percentage points or more as of the date the interest rate is set (or by 3.5 percentage points or more for subordinate-lien transactions).^[255] The Bureau concludes that maintaining the current 1.5-percentage-point threshold, in conjunction with the revised General QM requirements in proposed § 1026.43(e)(2), addresses access-to-credit concerns while striking an appropriate balance with ability-to-repay concerns.

The Bureau declines to extend the safe harbor to loans priced 1.5 percentage points or more above APOR given that such loans have higher delinquency rates and have, since the January 2013 Final Rule took effect, received a rebuttable presumption of compliance with the Bureau's ATR/QM rule with no evidence to suggest that the 1.5-percentage-point line has caused a significant disruption of access to responsible, affordable mortgage credit. Further, since the Board's 2008 rule, loans priced above the current 1.5-percentage-point threshold have been subject to an ability-to-repay requirement that is substantially similar to the rebuttable presumption standard for QMs under the Bureau's ATR/QM Rule. Consistent with one of the research center comments discussed above, HMDA data analyzed by the Bureau in the Assessment Report suggest that the safe harbor threshold of 1.5 percentage points has not constrained creditors, as the share of originations above the safe harbor threshold remained steady after the implementation of the ATR/QM Rule.^[256] In response to various commenters above who stated that less than 5 percent of conventional, first-lien purchase loans were priced 1.5 percentage points or more above APOR, the Bureau is unaware of reliable data evidencing that the low lending levels at higher rate spreads are caused by the 1.5 percentage point safe harbor threshold as opposed to other factors. Regarding the Board economists' analysis of 2014 HMDA data cited by a joint comment from consumer advocate and industry groups, the Bureau notes that the researchers “provide evidence in this note that lenders responded to the ATR and QM rules, particularly by favoring loans priced to obtain safe harbor protections,” but “the estimated magnitudes indicate the rules did not materially affect the mortgage market in 2014.” ^[257] In response to commenters recommending that the Bureau increase the current 1.5-percentage-point safe harbor threshold to create a more level playing field between conventional and FHA lending, the Bureau reiterates that no evidence has been presented to suggest that the existing safe harbor threshold under the Bureau's ATR/QM Rule has caused any significant disruption of access to responsible, affordable mortgage credit. Moreover, the Bureau is balancing access-to-credit concerns with concerns about ability to repay as measured by early delinquency rates.

In declining to provide a conclusive (rather than a rebuttable) QM presumption of compliance for loans priced above the current 1.5-percentage-point threshold, the Bureau concludes that such loans have higher delinquency rates and that access-to-credit concerns do not outweigh those ability to repay concerns.^[258] For example, Table 1 shows for 2002-2008 loans a 12 percent early delinquency rate for loans priced 1.50 to 1.74 percentage points above APOR, as compared to a 10 percent early delinquency rate for loans priced 1.25 to 1.49 percentage points above APOR. The comparable early delinquency rates for 2018 loans from Table 2 also show a higher early delinquency rate for loans priced 1.50 to 1.99 percentage points above APOR compared to loans priced 1.00 to 1.49 percentage points above APOR: 2.7 percent versus 1.7 percent.

In response to comments recommending that the Bureau increase the safe harbor threshold to account for possible future rate spread widening in the market, including in response to FHFA's GSE capital rule that was recently finalized and the additional loan-level price adjustment of 50 basis points on most Fannie Mae and Freddie Mac refinances, effective December 1, 2020, the Bureau concludes that it would be premature to increase the safe harbor threshold based on possible future spread widening in the market. For example, as discussed by an industry commenter above, such changes may also affect APOR itself, which would cause uncertainty regarding the actual magnitude of any future changes to spreads over APOR. Moreover, while it is possible that future spread widening could result in some safe harbor QMs instead becoming rebuttable presumption QMs, the Bureau concludes there is insufficient evidence to suggest that shifts in QMs' status from safe harbor to rebuttable presumption due to future spread widening would have a significant impact on access to responsible, affordable mortgage credit.^[259] However, the Bureau will monitor the market and take action as needed to maintain the best balance between consumers' ability to repay and access to responsible, affordable mortgage credit.

As discussed above in part V.B.4, several commenters generally objected to a price-based approach, but the Bureau did not receive comments requesting a lower safe harbor threshold if the Bureau finalizes a price-based approach. In maintaining and not lowering the current 1.5 percentage point safe harbor threshold, the Bureau concludes that there is some uncertainty as to what the consequences would be for the market and consumers with loans that would be safe harbor QMs under the existing rule but rebuttable presumption QMs under a lower safe harbor threshold. Since it took effect, the Bureau's ATR/QM Rule has provided a safe harbor to loans priced below the 1.5-percentage-point threshold—and such loans were never subject to the ability-to-repay requirements in the Board's 2008 HOEPA Final Rule. The 1.5-percentage-point threshold in the Bureau's ATR/QM Rule is the same as that used in the Board's 2008 HOEPA Final Rule. When the Bureau established the safe harbor in the January 2013 Final Rule, the Bureau stated that the “line the Bureau is drawing is one that has long been recognized as a rule of thumb to separate prime loans from subprime loans” and, “under the existing regulations that were adopted by the Board in 2008, only higher-priced mortgage loans are subject to an ability-to-repay requirement. . . .” ^[260] Thus, the January 2013 Final Rule stated that “investors will likely require creditors to agree to . . . representations and warranties when assigning or selling loans under the [Bureau's] new rule” and, for loans with rate spreads less than 1.5 percentage points, “this may represent an incremental risk of put-back to creditors, given that such loans are not subject to the current [2008 HOEPA Final Rule] regime, but those loans are being provided a safe harbor if they are qualified mortgages.” ^[261] In contrast, for loans with rate spreads of 1.5 percentage points or more, the Bureau stated that “it is not clear that there is any incremental risk beyond that which exists today under the Board's rule.” ^[262] The Bureau's January 2013 Final Rule further stated that there is “a widespread fear about the litigation risks associated with the Dodd-Frank Act ability-to-repay requirements,” ^[263] and that the safe harbor for loans with rate spreads less than 1.5 percentage points helps ensure that “litigation and secondary market impacts do not jeopardize access to credit.” ^[264] As discussed above, there is also concern among some commenters on the General QM Proposal regarding rebuttable presumption QMs presenting more litigation risk to creditors than safe harbor QMs.

Based on the Bureau's analysis of the 2018 NMDB data, the Bureau expects that the early delinquency rate of loans obtaining safe harbor QM status under this final rule will be on par with loans obtaining safe harbor QM status under the current rule, which includes the Temporary GSE QM loan definition. Table 6 shows the early delinquency rate for 2018 NMDB first-lien purchase originations by rate spread and DTI ratio. For loans with rate spreads between 1 and 1.49 percentage points and DTI ratios above 43 percent, the early delinquency rate is 2.3 percent. These are loans that would not meet the current General QM loan definition due to the 43 percent DTI limit, but that would receive safe harbor General QM status under this final rule. If the 2018 data are restricted to only those loans purchased and guaranteed by the GSEs (i.e., loans made under the Temporary GSE QM loan definition), loans with DTI ratios above 43 percent and rate spreads between 1 and 1.49 percentage points had an early delinquency rate of 2.4 percent.

The Bureau acknowledges that removing the 43 percent DTI limit will lead to somewhat higher-risk loans obtaining safe harbor QM status relative to loans within the current General QM loan definition (not including the Temporary GSE QM loan definition). In Table 5, the Bureau compared projected early delinquency rates for 2002-2008 first-lien purchase originations under the General QM loan definition with and without a 43 percent DTI limit under a range of potential rate-spread based safe harbor thresholds. Under the current 43 percent DTI limit for first-lien General QMs, Table 5 indicates that early delinquency rates for loans with rate spreads just below 1.5 percentage points increase with DTI ratio, from 6 percent for loans with a DTI ratio of 20 percent or below to 11 percent for loans with DTI ratios from 41 to 43 percent. For loans with rate spreads just below 1.5 percentage points and DTI ratios above 43 percent, Table 5 indicates early delinquency rates between 12 percent (for loans with 44 to 45 percent DTI ratios) and 15 percent (for loans with DTI ratios of 61 to 70 percent). Therefore, the loans with DTI ratios above 43 percent that would be granted safe harbor status under the price-based approach at a safe harbor threshold of 1.5 percentage points are likely to have a somewhat higher early delinquency rate than those just at or below 43 percent DTI ratios, 12 to 15 percent versus 11 percent. The comparable early delinquency rates for 2018 loans from Table 6 also show a slightly higher early delinquency rate for loans with rate spreads just below 1.5 percentage points with DTI ratios above 43 percent compared to loans with DTI ratios of 36 to 43 percent: 2.3 percent versus 1.5 percent. However, as noted above, if the 2018 data are restricted to loans made under the Temporary GSE QM loan definition, such loans with DTI ratios above 43 percent and rate spreads between 1 and 1.49 percentage points had an early delinquency rate of 2.4 percent. Thus, the Bureau expects that the early delinquency rate of loans obtaining safe harbor QM status under this final rule will be on par with loans obtaining safe harbor QM status under the current rule, which includes the Temporary GSE QM loan definition.

The Bureau concludes that the safe harbor threshold under this final rule strikes the best balance between ability-to-repay risk and the access-to-credit benefits discussed above and the overall safety of the prime QM market relative to the subprime market. As discussed by commenters above, loans that meet the General QM loan definition are relatively low-risk compared to loans with non-QM features. In response to commenters and based on findings in the Assessment Report, the Bureau concludes that loans with non-QM features—including interest-only loans, negative amortization loans, and loans made with limited or no documentation of the borrower's income or assets—had a substantial negative effect on consumers' ability to repay leading up to the 2008 financial crisis and subsequent passage of the Dodd-Frank Act.

In maintaining and not lowering the current 1.5-percentage-point safe harbor threshold as part of this final rule, the Bureau also acknowledges that the January 2013 Final Rule relied in part on the 43 percent DTI limit to support its conclusion that a 1.5 percentage-point safe harbor threshold is appropriate. However, as discussed above, the 43 percent DTI limit was only one of several supporting factors listed in the January 2013 Final Rule.^[265] Moreover, the January 2013 Final Rule did not include a DTI limit for Temporary GSE QMs but nonetheless provided both those loans and General QMs with the same 1.5-percentage-point safe harbor threshold. The January 2013 Final Rule stated that, “even in today's credit-constrained market, approximately 22 percent of mortgage loans are made with a debt-to-income ratio that exceeds 43 percent” and “many of those loans will fall within the temporary exception that the Bureau is recognizing for qualified mortgages.” ^[266] Further, as discussed in the section-by-section-analysis of § 1026.43(e)(2)(v)(A), this final rule imposes requirements for the creditor to consider the consumer's DTI ratio or residual income, income or assets other than the value of the dwelling, and debts to satisfy the General QM loan definition, thus requiring that the creditor consider key aspects of the consumer's financial capacity.^[267]

With respect to General QM prime first-lien loans (General QM first-lien loans with an APR that does not exceed APOR by 1.5 or more percentage points), the Bureau concludes that it is appropriate to use its adjustment authority under TILA section 105(a) to retain a conclusive presumption (i.e., a safe harbor). The Bureau concludes this approach strikes the best balance between the competing consumer protection and access-to-credit considerations described above. The Bureau concludes these same considerations support not changing the current safe harbor threshold of 3.5 percentage points for subordinate-lien transactions, which generally perform better and have stronger credit characteristics than first-lien transactions.^[268] The Bureau also concludes that providing a safe harbor for prime first-lien and subordinate-lien loans is necessary and proper to facilitate compliance with and to effectuate the purposes of section 129C and TILA, including to assure that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loans.

In addition, the Bureau also is also relying on TILA section 129C(b)(3)(B)(i), which authorizes the Bureau to prescribe regulations that revise, add to, or subtract from the criteria that define a QM, as authority for retaining a conclusive presumption. For the same reasons outlined above, the Bureau concludes that this conclusive presumption is necessary or proper to ensure that responsible, affordable mortgage credit remains available to consumers in a manner consistent with the purposes of TILA section 129C, as well as necessary and appropriate to effectuate the purposes of TILA section 129C and facilitate compliance with section 129C.

The final rule generally maintains the current ATR/QM Rule's rebuttable presumption of compliance for loans that exceed the safe harbor threshold but that otherwise meet the General QM requirements in § 1026.43(e)(2).^[269] The Bureau is not revising § 1026.43(e)(1)(ii)(B), which defines the grounds on which the presumption of compliance that applies to higher-priced QMs can be rebutted. The Bureau did not receive comments regarding the grounds on which borrowers can rebut the presumption of compliance. The Bureau concludes that existing § 1026.43(e)(1)(ii)(B) continues to strike the best balance between consumer protection and access to credit considerations and is sufficiently broad to provide consumers a reasonable opportunity to demonstrate that the creditor did not have a good faith and reasonable belief in the consumer's repayment ability, despite meeting the prerequisites of a QM.

VI. Section-by-Section Analysis

1026.43 Minimum Standards for Transactions Secured by a Dwelling

43(b) Definitions

43(b)(4)

Section 1026.43(b)(4) provides the definition of a higher-priced covered transaction. It provides that a covered transaction is a higher-priced covered transaction if the APR exceeds APOR for a comparable transaction as of the date the interest rate is set by the applicable rate spread specified in the ATR/QM Rule. For General QMs under § 1026.43(e)(2), the applicable rate spreads are 1.5 or more percentage points for a first-lien covered transaction and 3.5 or more percentage points for a subordinate-lien covered transaction. Pursuant to § 1026.43(e)(1), a loan that satisfies the requirements of a QM and is a higher-priced covered transaction under § 1026.43(b)(4) is eligible for a rebuttable presumption of compliance with the ATR requirements. A QM that is not a higher-priced covered transaction is eligible for a conclusive presumption of compliance with the ATR requirements.

The Bureau's Proposal

The Bureau proposed to revise § 1026.43(b)(4) to create a special rule for purposes of determining whether certain types of General QMs under § 1026.43(e)(2) are higher-priced covered transactions. Under the proposal, this special rule would have applied to loans for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due. For such loans, the creditor would have been required to determine the APR, for purposes of determining whether a General QM under § 1026.43(e)(2) is a higher-priced covered transaction, by treating the maximum interest rate that may apply during that five-year period as the interest rate for the full term of the loan.

Under the proposed rule, an identical special rule would have applied to loans for which the interest rate may or will change under proposed § 1026.43(e)(2)(vi), which would have revised the definition of a General QM under § 1026.43(e)(2) to implement the price-based approach described in part V of this final rule. The proposed rule stated that the special rules in the proposed revisions to § 1026.43(b)(4) and § 1026.43(e)(2)(vi) would not modify other provisions in Regulation Z for determining the APR for other purposes, such as the disclosures addressed in or subject to the commentary to § 1026.17(c)(1).

Proposed comment 43(b)(4)-4 stated that provisions in subpart C, including commentary to § 1026.17(c)(1), address how to determine the APR disclosures for closed-end credit transactions and that provisions in § 1026.32(a)(3) address how to determine the APR to determine coverage under § 1026.32(a)(1)(i). It further provided that proposed § 1026.43(b)(4) required, only for purposes of a QM under paragraph (e)(2), a different determination of the APR for purposes of paragraph (b)(4) for a loan for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due. It also cross-referenced proposed comment 43(e)(2)(vi)-4 for how to determine the APR of such a loan for purposes of § 1026.43(b)(4) and (e)(2)(vi).

The Bureau sought comment on all aspects of the special rule it proposed in § 1026.43(b)(4).

The Final Rule

The Bureau is finalizing § 1026.43(b)(4) and comment 43(b)(4)-4 as proposed. The section-by-section analysis of § 1026.43(e)(2)(vi), which the Bureau also is finalizing as proposed, explains the Bureau's reasoning for adopting these provisions as proposed. That section-by-section analysis also summarizes comments received in response to the proposed special rule and provides the Bureau's response to those comments.

Legal authority. As discussed above in part IV, TILA section 105(a) directs the Bureau to prescribe regulations to carry out the purposes of TILA and provides that such regulations may contain additional requirements, classifications, differentiations, or other provisions, and may provide for such adjustments and exceptions for all or any class of transactions that the Bureau judges are necessary or proper to effectuate the purposes of TILA, to prevent circumvention or evasion thereof, or to facilitate compliance therewith. In particular, it is the purpose of TILA section 129C, as amended by the Dodd-Frank Act, to assure that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loans and that are understandable.

As also discussed above in part IV, TILA section 129C(b)(3)(B)(i) authorizes the Bureau to prescribe regulations that revise, add to, or subtract from the criteria that define a QM upon a finding that such regulations are necessary or proper to ensure that responsible, affordable mortgage credit remains available to consumers in a manner consistent with the purposes of section 129C, necessary and appropriate to effectuate the purposes of section 129C and section 129B, to prevent circumvention or evasion thereof, or to facilitate compliance with such section.

The Bureau is finalizing the special rule in § 1026.43(b)(4) regarding the APR determination of certain loans for which the interest rate may or will change pursuant to its authority under TILA section 105(a) to make such adjustments and exceptions as are necessary and proper to effectuate the purposes of TILA, including that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loans. The Bureau concludes that these provisions will ensure that General QM status will not be accorded to certain loans for which the interest rate may or will change that pose a heightened risk of becoming unaffordable relatively soon after consummation. The Bureau is also finalizing these provisions pursuant to its authority under TILA section 129C(b)(3)(B)(i) to revise and add to the statutory language. The Bureau concludes that the special rule's APR determination provisions in § 1026.43(b)(4) will ensure that responsible, affordable mortgage credit remains available to consumers in a manner consistent with the purpose of TILA section 129C, referenced above, as well as effectuate that purpose.

43(c) Repayment Ability

43(c)(4) Verification of Income or Assets

TILA section 129C(a)(4) states that a creditor making a residential mortgage loan shall verify amounts of income or assets that such creditor relies on to determine repayment ability, including expected income or assets, by reviewing the consumer's Internal Revenue Service (IRS) Form W-2, tax returns, payroll receipts, financial institution records, or other third-party documents that provide reasonably reliable evidence of the consumer's income or assets. In the January 2013 Final Rule, the Bureau implemented this requirement in § 1026.43(c)(4), which states that a creditor must verify the amounts of income or assets that the creditor relies on under § 1026.43(c)(2)(i) to determine a consumer's ability to repay a covered transaction using third-party records that provide reasonably reliable evidence of the consumer's income or assets. Section 1026.43(c)(4) further states that a creditor may verify the consumer's income using a tax-return transcript issued by the IRS and lists several examples of other records the creditor may use to verify the consumer's income or assets, including, among others, financial institution records. Additionally, current § 1026.43(e)(2)(v)(A) provides that a General QM is a covered transaction for which the creditor considers and verifies at or before consummation the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan in accordance with § 1026.43(c)(4), as well as § 1026.43(c)(2)(i) and appendix Q.

The Bureau proposed to add comment 43(c)(4)-4 to clarify that a creditor does not meet the requirements of § 1026.43(c)(4) if it observes an inflow of funds into the consumer's account without confirming that the funds qualify as a consumer's personal income. The proposed comment also stated that, for example, a creditor would not meet the requirements of § 1026.43(c)(4) where it observes an unidentified $5,000 deposit in the consumer's account but fails to take any measures to confirm or lacks any basis to conclude that the deposit represents the consumer's personal income and not, for example, proceeds from the disbursement of a loan. The Bureau did not propose to change the text of § 1026.43(c)(4).

Commenters to the proposal did not address proposed comment 43(c)(4)-4. Accordingly, the Bureau is adopting new comment 43(c)(4)-4 as proposed. The Bureau determines, based on outreach and on its experience supervising creditors, that this clarification would be useful to creditors because the ATR/QM Rule includes “financial institution records” as one of the examples of records that a creditor may use to verify a consumer's income or assets. As part of their underwriting process, creditors may seek to use transactions in electronic or paper financial records such as consumer account statements to examine inflows and outflows from consumers' accounts. In many cases, there may be a sufficient basis in transaction data alone, or in combination with other information, to determine that a deposit or other credit to a consumer's account is the consumer's personal income, such that a creditor's use of the data in an underwriting process is distinguishable from the example in the proposed comment, and, therefore, the creditor may use the data in verifying the consumer's income. The Bureau also concludes that this clarification would help creditors understand their verification requirements under the General QM loan definition. Under this final rule, § 1026.43(e)(2)(v)(B) provides that, to satisfy the General QM loan definition, the creditor must verify the consumer's current or reasonably expected income or assets using third-party records that provide reasonably reliable evidence of the consumer's income or assets, in accordance with § 1026.43(c)(4).

The Bureau is adding comment 43(c)(4)-4 pursuant to TILA section 129C(a)(4), which states that a creditor making a residential mortgage loan shall verify amounts of income or assets that such creditor relies on to determine repayment ability, including expected income or assets, by reviewing the consumer's IRS Form W-2, tax returns, payroll receipts, financial institution records, or other third-party documents that provide reasonably reliable evidence of the consumer's income or assets.

43(e) Qualified Mortgages

43(e)(2) Qualified Mortgage Defined—General

43(e)(2)(v)

As discussed above in part V, this final rule removes the specific DTI limit in § 1026.43(e)(2)(vi). Furthermore, as discussed below in this section-by-section analysis, this final rule requires that creditors consider the consumer's DTI ratio or residual income and removes the appendix Q requirements from § 1026.43(e)(2)(v). The Bureau concludes that these amendments necessitate additional revisions to the General QM loan definition to clarify a creditor's obligation to consider and verify certain information for purposes of the General QM loan definition. Consequently, this final rule amends the consider and verify requirements in § 1026.43(e)(2)(v) and its associated commentary.

TILA section 129C contains several requirements that creditors consider and verify various types of information. In the statute's general ATR provisions, TILA section 129C(a)(1) requires that a creditor make a reasonable and good faith determination, based on “verified and documented information,” that a consumer has a reasonable ability to repay the loan. TILA section 129C(a)(3) states that a creditor's ATR determination shall include “consideration” of the consumer's credit history, current income, expected income the consumer is reasonably assured of receiving, current obligations, DTI ratio or the residual income the consumer will have after paying non-mortgage debt and mortgage-related obligations, employment status, and other financial resources other than the consumer's equity in the dwelling or real property that secures repayment of the loan. TILA section 129C(a)(4) states that a creditor making a residential mortgage loan shall verify amounts of income or assets that such creditor relies on to determine repayment ability, including expected income or assets, by reviewing the consumer's IRS Form W-2, tax returns, payroll receipts, financial institution records, or other third-party documents that provide reasonably reliable evidence of the consumer's income or assets. Finally, in the statutory QM definition, TILA section 129C(b)(2)(A)(iii) provides that, for a loan to be a QM, the income and financial resources relied on to qualify the obligors on the loan must be “verified and documented.”

In the January 2013 Final Rule, the Bureau implemented the requirements to consider and verify various factors for the general ATR standard in § 1026.43(c)(2), (3), (4), and (7). Section 1026.43(c)(2) states that—except as provided in certain other provisions (including the General QM loan definition)—a creditor must consider several specified factors in making its ATR determination. These factors include, among others, the consumer's current or reasonably expected income or assets, other than the value of the dwelling, including any real property attached to the dwelling, that secures the loan (under § 1026.43(c)(2)(i)); the consumer's current debt obligations, alimony, and child support (§ 1026.43(c)(2)(vi)); and the consumer's monthly DTI ratio or residual income in accordance with § 1026.43(c)(7). Section 1026.43(c)(3) requires a creditor to verify the information the creditor relies on in determining a consumer's repayment ability using reasonably reliable third-party records, with a few specified exceptions. Section 1026.43(c)(3) further states that a creditor must verify a consumer's income and assets that the creditor relies on in accordance with § 1026.43(c)(4). Section 1026.43(c)(4) requires that a creditor verify the amounts of income or assets that the creditor relies on to determine a consumer's ability to repay a covered transaction using third-party records that provide reasonably reliable evidence of the consumer's income or assets. It also provides examples of records the creditor may use to verify the consumer's income or assets.

As noted in part V, the January 2013 Final Rule incorporated some aspects of the general ATR standards into the General QM loan definition, including the requirement to consider and verify income or assets and debt obligations, alimony, and child support. Section 1026.43(e)(2)(v) states that a General QM is a covered transaction for which the creditor considers and verifies at or before consummation: (A) The consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan, in accordance with appendix Q, § 1026.43(c)(2)(i), and (c)(4); and (B) the consumer's current debt obligations, alimony, and child support in accordance with appendix Q and § 1026.43(c)(2)(vi) and (c)(3). The Bureau used its adjustment and exception authority under TILA section 129C(b)(3)(B)(i) to require creditors to consider and verify the consumer's debt obligations, alimony, and child support pursuant to the General QM loan definition.

The Bureau proposed to revise § 1026.43(e)(2)(v) to separate and clarify the requirements to consider and verify certain information for purposes of the General QM loan definition. Proposed § 1026.43(e)(2)(v)(A) contained the “consider” requirements and proposed § 1026.43(e)(2)(v)(B) contained the “verify” requirements. Specifically, proposed § 1026.43(e)(2)(v) stated that a General QM is a covered transaction for which the creditor: (A) Considers the consumer's income or assets, debt obligations, alimony, child support, and monthly DTI ratio or residual income, using the amounts determined from § 1026.43(e)(2)(v)(B); and (B) verifies the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan using third-party records that provide reasonably reliable evidence of the consumer's income or assets, in accordance with § 1026.43(c)(4), and the consumer's current debt obligations, alimony, and child support using reasonably reliable third-party records in accordance with § 1026.43(c)(3). Proposed § 1026.43(e)(2)(v)(A) also stated that, for purposes of § 1026.43(e)(2)(v)(A), the consumer's monthly DTI ratio or residual income is determined in accordance with § 1026.43(c)(7), except that the consumer's monthly payment on the covered transaction, including the monthly payment for mortgage-related obligations, is calculated in accordance with § 1026.43(e)(2)(iv). To further clarify the requirements in § 1026.43(e)(2)(v), the Bureau also proposed to add comments 43(e)(2)(v)(A)-1 through -3 and comments 43(e)(2)(v)(B)-1 through -3.

As discussed below, this final rule adopts § 1026.43(e)(2)(v)(A) largely as proposed—with minor technical additions to the rule text—and adopts § 1026.43(e)(2)(v)(B) as proposed. The Bureau is also adopting the proposed commentary for § 1026.43(e)(2)(v)(A) and § 1026.43(e)(2)(v)(B) largely as proposed, with two substantive changes from the proposal. First, the Bureau has added language to comment 43(e)(2)(v)(A)-1 to clarify that, in order to meet the General QM consider requirement, a creditor must maintain written policies and procedures for how it takes into account income, debt, and DTI or residual income and document how it took into account these factors. Second, the Bureau has added a list of specific verification standards to comment 43(e)(2)(v)(B)-3.i, which provides a safe harbor for compliance with the verification requirement in § 1026.43(e)(2)(v)(B). These verification standards include relevant provisions in specified versions of the Fannie Mae Single Family Selling Guide, the Freddie Mac Single-Family Seller/Servicer Guide, the FHA's Single Family Housing Policy Handbook, the VA's Lenders Handbook, and the USDA's Field Office Handbook for the Direct Single Family Housing Program and Handbook for the Single Family Guaranteed Loan Program, current as of the date of the proposal's public release.

The Bureau also proposed to remove comments 43(e)(2)(v)-2 and -3. In general, these comments explain that a creditor must consider and verify any income and debt specified in appendix Q, and that while a creditor may consider and verify any other income and debt, such income and debt would not be included in the DTI ratio determination required by § 1026.43(e)(2)(vi). This final rule removes these comments. The Bureau concludes that these comments are no longer needed due to this final rule's revisions to § 1026.43(e)(2)(v). The first sentence of each of these comments merely restates language in the regulatory text. The second sentence of each of these comments is no longer needed because this final rule removes references to appendix Q from § 1026.43(e)(2)(v). And the third sentence of each of these comments is no longer needed because this final rule removes the DTI limit in § 1026.43(e)(2)(vi).

43(e)(2)(v)(A)

The Bureau's Proposal

Section 1026.43(e)(2)(v) currently provides that a General QM is a covered transaction for which the creditor, at or before consummation, considers and verifies the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan, debt obligations, alimony, and child support. In the General QM Proposal, the Bureau proposed to separate the consider and verify requirements in § 1026.43(e)(2)(v) into § 1026.43(e)(2)(v)(A) for the “consider” requirements and § 1026.43(e)(2)(v)(B) for the “verify” requirements. The Bureau proposed to revise § 1026.43(e)(2)(v)(A) to provide that a General QM is a covered transaction for which the creditor, at or before consummation, considers the consumer's income or assets, debt obligations, alimony, child support, and monthly DTI ratio or residual income, using the amounts determined from proposed § 1026.43(e)(2)(v)(B). Proposed § 1026.43(e)(2)(v)(A) also stated that, for purposes of § 1026.43(e)(2)(v)(A), the consumer's monthly DTI ratio or residual income is determined in accordance with § 1026.43(c)(7), except that the consumer's monthly payment on the covered transaction, including the monthly payment for mortgage-related obligations, is calculated in accordance with § 1026.43(e)(2)(iv).

To clarify the consider requirement in proposed § 1026.43(e)(2)(v)(A), the Bureau proposed to add comments 43(e)(2)(v)(A)-1 to -3. Proposed comment 43(e)(2)(v)(A)-1 provided that, in order to comply with the consider requirement, a creditor must take into account income or assets, debt obligations, alimony, child support, and monthly DTI ratio or residual income in its ability-to-repay determination. The proposed comment further stated that, pursuant to requirements in § 1026.25(a) to retain records showing compliance with the rule, a creditor must retain documentation showing how it took into account the required factors. The proposed comment provided examples of the types of documents that a creditor might use to show that it took into account the required factors.

The Bureau proposed comment 43(e)(2)(v)(A)-2 to clarify that creditors have flexibility in how they consider these factors and that the proposed rule would not have prescribed a specific monthly DTI or residual income threshold. The proposed comment also included two examples of how a creditor may comply with the requirement to consider DTI.

The Bureau proposed comment 43(e)(2)(v)(A)-3 to clarify that the requirement in § 1026.43(e)(2)(v)(A) to consider income or assets, debt obligations, alimony, child support, and monthly DTI or residual income would not preclude the creditor from taking into account additional factors that are relevant in making its ability-to-repay determination.

This final rule adopts § 1026.43(e)(2)(v)(A) largely as proposed, with minor technical additions to the rule text. This final rule also adopts comments 43(e)(2)(v)(A)-1 to -3 largely as proposed, with some adjustments in comment 43(e)(2)(v)(A)-1 to clarify that creditors must maintain certain policies and procedures and retain certain documentation to satisfy § 1026.43(e)(2)(v)(A).

Comments Received

The Bureau's general approach to the consider requirement. Both industry and consumer advocate commenters supported the proposal to retain a requirement to consider income or assets, debt obligations, alimony, child support, and monthly DTI or residual income for General QMs. Commenters generally stated that the consider requirement is an important consumer protection for QMs and that such a requirement is necessary to achieve the statutory intent of TILA. Both industry and consumer advocate commenters generally supported the retention of a requirement to consider a consumer's monthly DTI ratio and the option of considering residual income in lieu of DTI. These commenters explained that DTI is an important factor in assessing a consumer's ability to repay and that the residual income option creates space for flexibility and industry innovation. One industry commenter noted that creditors use DTI as part of their underwriting processes and will continue to do so even if the General QM loan definition no longer includes a specific DTI limit. Another industry commenter explained that it uses DTI as part of its underwriting process and makes responsible loans with DTI ratios above 43 percent. Another industry commenter stated that the VA loan program has successfully used residual income for underwriting purposes.

One industry commenter expressed concerns about the requirement to calculate DTI according to § 1026.43(c)(7), arguing that this cross-reference could be interpreted to import a requirement that creditors adopt an “appropriate” DTI threshold. The commenter suggested that the Bureau could avoid that interpretation by removing any requirement to calculate a DTI ratio. As explained in the proposed rule and below, the General QM Proposal incorporated the cross-reference only for purposes of calculating monthly DTI, residual income, and monthly payment on the covered loan.

Commentary provisions. Industry commenters generally supported the inclusion of proposed comments 43(e)(2)(v)(A)-1 through -3. These commenters generally stated that the proposed comments provide the clarity needed to facilitate industry compliance and assurance of QM status. Many industry commenters specifically encouraged the Bureau to adopt the proposed comments because they would provide creditors with flexibility in applying their own underwriting methodologies. One industry commenter stated that the examples in the proposed comments reflected the current underwriting practices of community banks.

Many industry commenters supported the proposed documentation approach to the consider requirement. One industry commenter explained that the proposed documentation approach would be an effective means for a creditor to meet the consider requirement and have assurance of QM status. A comment letter signed by 12 civil rights and consumer groups included a “term sheet” that provided a variety of suggested changes to the consider requirement (“joint consumer advocate term sheet”) and asked the Bureau to clearly state that in order to maintain QM status, the creditor must retain documentation of how it satisfied the consider requirement. A consumer advocate commenter that also signed the term sheet explained that, without documentation, examiners could not meaningfully assess whether the creditor had in fact considered the consumer's debts and income. An industry commenter asked the Bureau to adopt a cure provision for situations where a loan file is incomplete due to an alleged oversight.

Several commenters recommended that the Bureau expressly require creditors to develop and maintain procedures to consider debts and income. In its support for the documentation examples in the first proposed comment, one industry commenter suggested that the Bureau require creditors to provide underwriter spreadsheets or other documentation that showed the creditor followed procedures in its consideration of the required factors. Another industry commenter recommended that the Bureau require creditors to maintain an independently developed credit policy setting forth the manner in which they will consider and verify the required factors. The commenter stated that such a requirement would facilitate investor and regulator evaluation of a loan's QM status and would align with OCC guidance and appraiser guidance under the Financial Institutions Reform, Recovery, and Enforcement Act. Another industry commenter asked the Bureau to develop specific operational guidelines for the calculation of DTI and residual income, including minimum threshold values for residual income. Another industry commenter stated that the Bureau should require creditors to comply with a specific set of underwriting criteria that includes compensating factors for consumers with high DTI.

Similar to these industry commenters, consumer advocate commenters asked the Bureau to require creditors to develop and maintain procedures to consider debts and income. One consumer advocate commenter that signed the joint consumer advocate term sheet explained that, without a component requiring such procedures, the consider requirement would exist in name only and individual loan officers could make individual decisions about what meets the consider standard. This commenter explained that without procedures, creditors under pressure to make loans could use their discretion to make a pro forma note of consideration.

Some industry commenters specifically encouraged the Bureau to adopt the language in proposed comment 43(e)(2)(v)(A)-2 explaining that the proposed rule would not prescribe a particular DTI or residual income threshold. One industry commenter stated that it appreciated how the proposed comments provided creditors with flexibility as to how they considered monthly DTI and additional factors in their underwriting processes. One industry commenter asked the Bureau to refrain from enumerating appropriate compensating factors. In contrast, some industry commenters stated that the proposed consider requirement was still too vague and requested additional clarification. One of these commenters warned that risk-averse lenders would not originate loans under the proposed approach.

One industry commenter supported the consider requirement but requested that the Bureau require a creditor to show that it took into account the required factors, rather than how it took into account the required factors.

Several industry and consumer advocate commenters supported the Bureau's statement in the proposal that if creditors ignore income or assets, debt obligations, alimony, child support, and DTI or residual income, they do not consider these factors sufficiently for purposes of the General QM loan definition.

Both industry and consumer advocate commenters raised concerns that the proposed General QM consider standard, even with the proposed clarifying commentary, would not prevent loans from obtaining QM status if the consumer lacks the ability to repay. One consumer advocate commenter stated that the proposed General QM consider standard needs more specificity to ensure that creditors engage in a meaningful ability-to-repay analysis. The joint consumer advocate term sheet provided a variety of suggested changes to the consider requirement, such as adding extreme examples of non-compliance (100 percent DTI or zero or negative residual income loans); deeming LTV-based loans to be a per se violation of the consider requirement; clarifying that not retaining documentation of how the creditor considered the required factors would result in loss of QM status; and expanding the documentation requirement so that an examiner could confirm that a creditor followed its procedures. Another consumer advocate commenter that signed the joint consumer advocate term sheet stated that examples of non-compliant underwriting practices would provide some clarity to consumers and industry; establish an outer bound for responsible mortgage lending; and ensure that lenders adopt systems that would prevent behavior that falls outside the scope of a reasonable consideration of the required factors. This consumer advocate commenter stated that the joint consumer advocate term sheet's recommendation to clearly exclude loans where the creditor relied on LTV ratio in lieu of debt, income, and DTI or residual income would prevent loan flipping practices, which rely on the consumer's existing equity in the home to repeatedly refinance and strip equity in order to pay financed closing costs immediately to the creditor or broker. In contrast, one industry commenter stated that LTV-based lending should not be a concern given the fixed cost of foreclosure and how a creditor determines loan pricing. One industry commenter stated that a loan with 100 percent DTI could meet the proposed General QM consider standard.

The Final Rule

This final rule adopts § 1026.43(e)(2)(v)(A) and comments 43(e)(2)(v)(A)-1 to -3 largely as proposed, with minor technical additions to the rule text and some adjustments in comment 43(e)(2)(v)(A)-1 to clarify that creditors must maintain certain policies and procedures and retain certain documentation. As explained above, the Bureau is separating the consider and verify requirements in § 1026.43(e)(2)(v) into § 1026.43(e)(2)(v)(A) for the “consider” requirements and § 1026.43(e)(2)(v)(B) for the “verify” requirements. Final § 1026.43(e)(2)(v)(A) provides that a General QM is a covered transaction for which the creditor, at or before consummation, considers the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan, debt obligations, alimony, child support, and monthly DTI ratio or residual income, using the amounts determined from § 1026.43(e)(2)(v)(B). Although the proposed consider provision would have required creditors to consider current or reasonably expected income or assets other than the value of the dwelling through the requirement to use amounts determined from the § 1026.43(e)(2)(v)(B), the final rule makes this connection more clear by including the clauses “current or reasonably expected” and “other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan” in § 1026.43(e)(2)(v)(A). Final § 1026.43(e)(2)(v)(A) also states that, for purposes of § 1026.43(e)(2)(v)(A), the consumer's monthly DTI ratio or residual income is determined in accordance with § 1026.43(c)(7), except that the consumer's monthly payment on the covered transaction, including the monthly payment for mortgage-related obligations, is calculated in accordance with § 1026.43(e)(2)(iv).

The Bureau's general approach to the consider requirement. The Bureau concludes that requiring creditors to consider DTI as part of the General QM loan definition ensures that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loan. The Bureau determines that DTI continues to be an important factor in assessing a consumer's ability to repay. Comments on the General QM Proposal and on the ANPR indicate that creditors generally use DTI as part of their underwriting process. These comments indicate that requiring as part of the General QM loan definition that creditors consider DTI when determining a consumer's ability to repay—even if the General QM loan definition no longer includes a specific DTI limit—is consistent with current market practices.

As discussed in the June 2013 Final Rule, the Bureau created an exception from the DTI limit for certain small creditors that hold QMs on portfolio.^[270] The Bureau determined that, even though the DTI limit was not appropriate for a small creditor that holds loans on their portfolio, DTI (or residual income, as discussed below) was still a fundamental part of the creditor's ability-to-repay determination.^[271] The Bureau similarly concludes that DTI is a fundamental part of the creditor's ability-to-repay determination for General QMs.

Section 1026.43(e)(2)(v)(A) provides creditors with the option to consider either a consumer's monthly residual income or DTI. The Bureau concludes that residual income is an appropriate alternative to monthly DTI for creditors to consider under § 1026.43(e)(2)(v). The January 2013 Final Rule adopted a bright-line DTI limit for the General QM loan definition under § 1026.43(e)(2)(vi), but the Bureau concluded that it did not have enough information to establish a bright-line residual income limit as an alternative to the DTI limit.^[272] In comparison, consistent with TILA section 129C(a)(3), the January 2013 Final Rule allows creditors to consider either residual income or DTI as part of the general ATR requirements in § 1026.43(c)(2)(vii), and the June 2013 Final Rule allows small creditors originating QMs pursuant to § 1026.43(e)(5) to consider DTI or residual income. Given the elimination of the bright-line DTI limit in § 1026.43(e)(2)(vi), comments on the proposed rule, comments from stakeholders in the January 2013 Final Rule regarding the value of residual income in determining ability to repay,^[273] and the Bureau's determination in the June 2013 Final Rule that residual income can be a valuable measure of ability to repay, the Bureau concludes that allowing creditors the option to consider residual income in lieu of DTI would allow for creditor flexibility and innovation and is necessary and proper to preserve access to responsible, affordable mortgage credit.

The Bureau concludes that the amounts considered under § 1026.43(e)(2)(v)(A) should be consistent with the amounts verified according to § 1026.43(e)(2)(v)(B). For example, if the creditor seeks to comply with the consider requirement under § 1026.43(e)(2)(v)(A) using the consumer's assets, the creditor could consider assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan as those assets are calculated under § 1026.43(e)(2)(v)(B).

The final rule also adopts the proposed requirement in § 1026.43(e)(2)(v)(A) to calculate monthly DTI, monthly residual income, and monthly payment for mortgage-related obligations in a manner consistent with the method used in current § 1026.43(e)(2)(vi). As explained in the proposed rule, this calculation method was previously adopted in the January 2013 Final Rule and is being moved to the § 1026.43(e)(2)(v)(A) consider requirement given the Bureau's removal of the DTI limit in § 1026.43(e)(2)(vi) and appendix Q. To preserve the incorporation of alimony and child support that was previously facilitated by appendix Q, the calculation method in § 1026.43(e)(2)(v)(A) now cross-references § 1026.43(c)(7) for purposes of calculating monthly DTI or residual income. The Bureau concludes that incorporating the pre-existing reference to simultaneous loans is no longer necessary because the new cross-reference to § 1026.43(c)(7) requires creditors to consider simultaneous loans. Additionally, given that this final rule allows creditors to consider residual income in lieu of monthly DTI, the Bureau is expanding the calculation method requirement to include residual income. This calculation method also incorporates the pre-existing cross-reference to § 1026.43(e)(2)(iv) to determine the monthly payments for the covered loan.

As explained in the proposed rule, this calculation method was previously adopted in the January 2013 Final Rule. This calculation method does not appear to be unduly burdensome given that, as described further below, only one commenter addressed the proposed calculation provision, and the comment related not to the calculation method itself but to the commenter's concern that cross-referencing § 1026.43(c)(7) could be interpreted to import a requirement that creditors adopt an “appropriate” DTI threshold. The Bureau also believes that providing a calculation method will facilitate compliance and decrease creditor compliance costs by reducing ambiguity as to how DTI must be calculated. Accordingly, the Bureau concludes that the information in the rulemaking record does not support amending the rule to delete or change the calculation method. The Bureau also notes that the requirement merely provides the method for calculating DTI, residual income, and monthly mortgage payments. As detailed in comments 43(e)(2)(v)(A)-2 to -3, General QM creditors still retain the flexibility to determine how the required factors are taken into account in the consumer's ATR determination.

The Bureau declines to remove the requirement to calculate and consider DTI (or residual income) according to § 1026.43(c)(7) in order to address the industry commenter's concern that this could be interpreted to import a requirement that creditors adopt an “appropriate” DTI threshold. Instead, as explained in the proposed rule and above, the Bureau emphasizes that this final rule incorporates the cross-reference only for purposes of calculating monthly DTI, residual income, and monthly payment on the covered loan. As comment 43(e)(2)(v)(A)-2 makes clear, creditors have flexibility in how they consider income or assets, debt obligations, alimony, child support, and monthly DTI ratio or residual income and the final rule does not prescribe a specific monthly DTI or residual income threshold. More generally, the Bureau emphasizes that § 1026.43(e)(2)(v)(A) requires only that the creditor “consider” the specified factors. It does not permit a broader challenge that a loan is not a General QM because the creditor failed to make a reasonable and good-faith determination of the consumer's ability to repay under § 1026.43(c)(1), as this would undermine the certainty of whether a loan is a General QM.

Commentary provisions. For the reasons discussed below, the Bureau is finalizing comments 43(e)(2)(v)(A)-1 to -3 largely as proposed, with some adjustments in comment 43(e)(2)(v)(A)-1 to clarify that creditors must maintain certain policies and procedures and must retain certain documentation.

This final rule adds comments 43(e)(2)(v)(A)-1 to -3 because the Bureau concludes they are appropriate to ensure that the Rule's requirement to consider the consumer's income or assets, debt obligations, alimony, child support, and DTI or residual income is clear and detailed enough to provide creditors with sufficient certainty about whether a loan satisfies the General QM loan definition. Under the final rule, the General QM loan definition no longer includes a specific DTI limit in § 1026.43(e)(2)(vi) and instead requires in § 1026.43(e)(2)(v)(A) that creditors consider the consumer's income or assets, debt obligations, alimony, child support, and DTI or residual income . By requiring creditors to calculate DTI and compare that calculation to a DTI limit, the DTI limit from the January 2013 Final Rule provided creditors with a bright-line rule demonstrating how to consider the consumer's income or assets and debts for purposes of determining whether the General QM loan requirements are met. Without additional explanation of the requirement to consider DTI or residual income, along with the consumer's income or assets and debts, elimination of the DTI limit could create compliance uncertainty that could leave some creditors reluctant to originate QMs to consumers and could allow other creditors to originate risky loans without considering DTI or residual income and still receive QM status. In addition, without additional explanation, it may be difficult to enforce the requirement to consider. Commentary examples of compliance that reflect standard market practices also may help ensure that the consider requirement is not unduly burdensome. Many commenters supported the Bureau's proposal to maintain the consider requirement in the General QM loan definition, while also emphasizing the importance of clarity of QM safe harbor status and the utility of compliance examples. While commenters generally supported inclusion of the proposed comments, some commenters requested additions such as clarification of the documentation requirement and examples of non-compliance. Accordingly, the Bureau concludes that it is appropriate to provide additional explanation for the § 1026.43(e)(2)(v) consider requirement in comments 43(e)(2)(v)(A)-1 to -3, as discussed below.

Comment 43(e)(2)(v)(A)-1. Consistent with the proposal, comment 43(e)(2)(v)(A)-1 explains that, in order to comply with the requirement to consider, a creditor must take into account current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan, debt obligations, alimony, child support, and monthly DTI ratio or residual income in its ability-to-repay determination. As adopted by this final rule, comment 43(e)(2)(v)(A)-1 also provides that a creditor must maintain written policies and procedures for how it takes into account, pursuant to its underwriting standards, income or assets, debt obligations, alimony, child support, and monthly debt-to-income ratio or residual income in its ability-to-repay determination. The Bureau is also adding a clause to comment 43(e)(2)(v)(A)-1 to explain that the creditor must document how it applied its policies and procedures. The Bureau is also clarifying the documentation example to reflect how the creditor may also comply by providing the required documents in combination with any applicable exceptions used from the creditor's policies and procedures. Bureau experience in market outreach and regulation shows that it is standard practice for creditors to maintain written policies and procedures, including underwriting standards, for considering debt, income, and DTI or residual income, and commenters representing creditors explained that their members already have underwriting procedures to take into account DTI in the ability-to-repay determination. The creditor's policies and procedures typically refer to the creditor's underwriting standards and describe how to address exceptions to the creditor's underwriting standards.

The Bureau concludes that this policies and procedures clarification will facilitate confirmation by investors, auditors, consumers, regulators, and other stakeholders that a creditor has, in fact, taken into account the required factors. The Bureau determines that, as some commenters noted, it would be difficult for these stakeholders to identify how a creditor took into account the required factors if the creditor does not have written policies and procedures for how it takes them into account. Further, given the flexibility that this final rule provides to creditors by removing the DTI limit, the Bureau concludes that it is important for creditors to adopt and memorialize their institutional policies and procedures (including underwriting standards) for considering the consumer's income or assets, debt obligations, alimony, child support, and DTI or residual income, to help ensure that the consideration is sufficiently rigorous. The Bureau also concludes that this clarification will assist creditors in ensuring compliance with the General QM requirements by helping to prevent individual loan officers and underwriters from attempting to originate General QMs without having met the consider requirement. The Bureau additionally concludes that this clarification will impose a limited burden given that standard market practice is to maintain underwriting standards and policies and procedures.

Comment 43(e)(2)(v)(A)-1 also explains that to comply with § 1026.43(e)(2)(v)(A)—and thereby to qualify for General QM status—a creditor must retain documentation showing how it took into account the required factors in its ability-to-repay determination, including how it applied its policies and procedures. This reflects a modification from the proposal, which would have cross-referenced the creditor's obligation under § 1026.25(a) to retain documentation. The requirement continues to defer to creditors on how to consider the required factors, allowing creditors the flexibility to use their own underwriting standards as long as the loan file documents how the required factors were taken into account in the creditor's ability-to-repay determination.

The General QM loan definition currently contains a 43 percent DTI limit, so any third party can compare the consumer's DTI (as reflected in the loan file) to the limit to confirm that the requirement to consider income or assets and debts was met. In contrast, under this final rule, the General QM consider requirement allows the creditor to determine how debt, alimony, child support, income or assets, and DTI or residual income should be taken into account in its ability-to-repay determination. Although there is a general record retention requirement in the ATR/QM Rule, the Bureau agrees with the commenter that this revised consider requirement should include a documentation component because, absent a documentation requirement, only the creditor would know how and whether it took into account the required factors in its ability-to-repay determination. Documentation of how the creditor considered the required factors is necessary for any third party, such as consumers, investors, and regulators, to confirm that the creditor did, in fact, consider the required factors.

Given statements from commenters about the interaction between the documentation requirement and QM status, the Bureau concludes that adding clarifying language to this documentation retention requirement is necessary. The final rule's commentary explains that in order to meet the consider requirement and thereby meet the requirements for a QM under § 1026.43(e)(2)—whether the loan is a safe harbor QM under § 1026.43(e)(1)(i) or a rebuttable presumption QM under § 1026.43(e)(1)(ii)—a creditor must retain documentation showing how it took into account these factors in its ability-to-repay determination, including how it applied its policies and procedures. To clarify that a lack of documentation showing how the creditor took into account the required factors would result in loss of QM status, rather than constituting a mere violation of the record retention requirement in § 1026.25(a), the Bureau is removing the proposed cross-reference to the record retention requirement in § 1026.25(a). The Bureau is adopting the documentation examples in the last sentence, with new language to clarify that a creditor can also comply by relying on any applicable exceptions in the creditor's policies and procedures (in combination with the example underwriting documents) to show how the creditor took into account the required factors. As examples of the type of documents that a creditor might use to show that income or assets, debt obligations, alimony, child support, and DTI or residual income were taken into account, the comment cites an underwriter worksheet or a final automated underwriting system certification, in combination with the creditor's applicable underwriting standards and any applicable exceptions described in its policies and procedures, that shows how these required factors were taken into account in the creditor's ability-to-repay determination.

In summary, comment 43(e)(2)(v)(A)-1 explains that the § 1026.43(e)(2)(v)(A) consider requirement means to take into account income or assets, debt obligations, alimony, child support, and monthly debt-to-income ratio or residual income in the consumer's ability-to-repay determination, including maintaining written policies and procedures to take into account and retaining documentation of how the creditor took into account. As detailed in comments 43(e)(2)(v)(A)-2 and 43(e)(2)(v)(A)-3, a creditor has flexibility in how it considers income or assets, debt obligations, alimony, child support, and monthly debt-to-income ratio or residual income, as long as the creditor documents how it took into account these required factors in its ability-to-repay determination. For example, a creditor might originate a loan with a DTI that deviates from the standard DTI threshold in its underwriting guidelines because the consumer's significant savings meets an exception in those guidelines. Under this example, the internal thresholds and exceptions qualify as procedures for taking into account, and documentation of how the creditor applied this exception to the loan file shows how the required factors were taken into account under § 1026.43(e)(2)(v)(A).

The creditor's maintenance of written policies and procedures facilitates review of the loan file to confirm that the creditor did, in fact, document how it took into account income or assets, debt, alimony, child support, and DTI ratio or residual income. The documentation provision requires a creditor to retain documentation to show how it applied its written policies and procedures, and, to the extent it deviated from them, to further retain documentation of how the creditor nonetheless took into account the required factors. The documentation examples listed in the comment (an underwriter worksheet or a final automated underwriting system certification, in combination with the creditor's applicable underwriting standards and any applicable exceptions described in its policies and procedures, that show how these required factors were taken into account in the creditor's ability-to-repay determination) can be sufficient to show how the creditor applied its written policies and procedures. For example, a typical loan application may fall within the creditor's underwriting standards, so an underwriter worksheet could contain enough information to show how the creditor took into account the required factors under the creditor's underwriting standards. Another example is a loan application that triggers exceptions, where the underwriter worksheet might state that certain exceptions were applied, and referring to the creditor's policies and procedures would clarify how those exceptions took into account the required factors. In contrast to the discussion in the previous paragraph, a creditor would not meet the § 1026.43(e)(2)(v)(A) consider requirement if the creditor deviated from its policies and procedures and its documentation failed to show how the required factors were taken into account. For example, a creditor would not meet the § 1026.43(e)(2)(v)(A) consider requirement if the consumer did not meet its own underwriting standards and the creditor merely made a note that the loan was approved by management.

As the Bureau explained in the General QM Proposal, the § 1026.43(e)(2)(v)(A) consider requirement means that if a creditor ignores the required factors of income or assets, debt obligations, alimony, child support, and DTI or residual income—or otherwise did not take them into account as part of its ability-to-repay determination—the loan would not be eligible for QM status. Consumer advocate commenters asked the Bureau to add examples of non-compliance, such as loans with 100 percent DTI or zero residual income, and LTV-based loans, arguing that these examples would help prevent loans from receiving QM status when debts and income did not demonstrate a consumer's ability to repay.

The Bureau declines to codify extreme examples of non-compliance in the final rule. Although the Bureau concludes that loans for which a consumer has 100 percent DTI or zero or negative residual income—and no significant assets unrelated to the value of the dwelling that could support the mortgage loan payments—would not meet the General QM consider standard because the only reasonable conclusion would be that the creditor did not consider DTI or residual income, putting such extreme examples in the rule could be incorrectly interpreted to permit any less extreme practices. For example, a creditor might originate a loan to consumer in a family of four with $200 in monthly residual income and no significant assets unrelated to the value of the dwelling. Although the only reasonable conclusion is that the creditor ignored the consumer's residual income and did not meet the General QM consider requirement, creditors might perceive the extreme non-compliance example to mean that only zero or negative residual income loans could violate the rule.

The Bureau concludes that adding an LTV ratio or other home equity discussion to the General QM consider requirement would introduce too much confusion, thereby undermining the need for clarity of QM status, and declines to adopt this recommendation. For example, some creditors may determine that consumers with a higher DTI have an ability to repay according to their underwriting policy, but due to market risk tolerance will only originate that higher DTI loan if the consumer has a relatively low LTV ratio. Although that loan may meet the consider requirement because the creditor applied its underwriting guidelines and showed how that DTI met its established DTI underwriting thresholds, adding a discussion about LTV ratio to the General QM consider requirement could be misconstrued to undermine the loan's General QM status. In contrast, commenters raised concerns about industry practices when a creditor ignores consumer debt, income, and DTI or residual income and instead relies on LTV ratio, such as with loan flipping. As discussed in the General QM Proposal and the January 2013 Final Rule, the Bureau is aware of concerns about creditors relying on factors related to the value of the dwelling, like LTV ratio, and how such reliance may have contributed to the mortgage crisis.^[274] The Bureau agrees that reliance on LTV ratio or another measure of current or future home equity, in conjunction with a 100 percent DTI or no residual income and no other significant assets unrelated to the value of the dwelling, support a conclusion that a creditor did not meet the § 1026.43(e)(2)(v)(A) requirement to consider the consumer's current or reasonably expected income or assets other than the value of the dwelling securing the mortgage, debt obligations, alimony, child support, and monthly DTI ratio or residual income.

The Bureau declines to change the General QM consider requirement from a standard to show how the creditor took into account to a standard to show that the creditor took into account. The suggested language change would remove the requirement for creditors to connect their consideration of the required factors to the ability-to-repay determination, making the consider requirement a check-the-box exercise under which a file could merely state that the factors were considered even if the creditor ignored debts and income. Instead, the Bureau concludes that creditors must show how it took into account the required factors, including, for example, showing how it applied its underwriting procedures to the consumer's loan application.

Comment 43(e)(2)(v)(A)-2. The Bureau is finalizing comment 43(e)(2)(v)(A)-2 as proposed. To reinforce that the General QM loan definition no longer includes a specific DTI limit, comment 43(e)(2)(v)(A)-2 highlights that creditors have flexibility in how they consider these factors. Comment 43(e)(2)(v)(A)-2 clarifies that § 1026.43(e)(2)(v)(A) does not prescribe specifically how a creditor must consider monthly debt-to-income ratio or residual income and also does not prescribe a particular monthly debt-to-income ratio or residual income threshold with which a creditor must comply. To assist creditors in understanding their compliance obligations, the Bureau is finalizing two examples of how to comply with the requirement to consider DTI or residual income. Comment 43(e)(2)(v)(A)-2 provides an example in which a creditor considers monthly DTI or residual income by establishing monthly DTI or residual income thresholds for its own underwriting standards and documenting how those thresholds were applied to determine the consumer's ability to repay. Given that some creditors use several thresholds that depend on any relevant compensating factors, the Bureau is finalizing a second example. The second example provides that a creditor may also consider DTI or residual income by establishing monthly DTI or residual income thresholds and exceptions to those thresholds based on other compensating factors, and documenting application of the thresholds along with any applicable exceptions. The Bureau concludes that both examples are consistent with current market practices and therefore providing these examples would clarify a loan's QM status without imposing a significant burden on the market.

Comment 43(e)(2)(v)(A)-3. The Bureau is finalizing comment 43(e)(2)(v)(A)-3 as proposed. The Bureau is aware that some creditors look to factors in addition to income or assets, debt obligations, alimony, child support, and DTI or residual income in determining a consumer's ability to repay. For example, the Bureau is aware that some creditors may look to net cash flow into a consumer's deposit account as a method of residual income analysis. A net cash flow calculation typically consists of residual income, further reduced by consumer expenditures other than those already subtracted as part of the residual income calculation. Accordingly, the result of a net cash flow calculation may be useful in assessing the adequacy of a particular consumer's residual income. Comment 43(e)(2)(v)(A)-3 clarifies that the requirement in § 1026.43(e)(2)(v)(A) to consider income or assets, debt obligations, alimony, child support, and monthly DTI or residual income does not preclude the creditor from taking into account additional factors that are relevant in making its ability-to-repay determination.

The comment further provides that creditors may look to existing comment 43(c)(7)-3 for guidance on considering additional factors in determining the consumer's ability to repay. Comment 43(c)(7)-3 explains that creditors may consider additional factors when determining a consumer's ability to repay and provides an example of looking to consumer assets other than the value of the dwelling, such as a savings account.

Legal Authority

The Bureau is finalizing the requirement that the creditor consider the consumer's monthly DTI ratio or residual income, current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan, debt obligations, alimony, and child support under § 1026.43(e)(2)(A) pursuant to its adjustment and exception authority under TILA section 129C(b)(3)(B)(i). The Bureau finds that this addition to the General QM criteria is necessary and proper to ensure that responsible, affordable mortgage credit remains available to consumers in a manner that is consistent with the purposes of TILA section 129C and necessary and appropriate to effectuate the purposes of TILA section 129C, which includes assuring that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loan. The Bureau also incorporates this requirement pursuant to its authority under TILA section 105(a) to issue regulations that, among other things, contain such additional requirements or other provisions, or that provide for such adjustments for all or any class of transactions, that in the Bureau's judgment are necessary or proper to effectuate the purposes of TILA, which include the above purpose of section 129C. The Bureau finds that including consideration of DTI or residual income in the General QM loan criteria is necessary and proper to fulfill the purpose of assuring that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loan. The Bureau also finds that § 1026.43(e)(2)(A) is authorized by TILA section 129C(b)(2)(A)(vi), which permits, but does not require, the Bureau to adopt guidelines or regulations relating to DTI ratios or alternative measures of ability to pay regular expenses after payment of total monthly debt.

43(e)(2)(v)(B)

The Bureau's Proposal

The Bureau proposed to revise § 1026.43(e)(2)(v)(B) to provide that a General QM would be a covered transaction for which the creditor, at or before consummation, verifies the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan using third-party records that provide reasonably reliable evidence of the consumer's income or assets, in accordance with § 1026.43(c)(4) and verifies the consumer's current debt obligations, alimony, and child support using reasonably reliable third-party records in accordance with § 1026.43(c)(3). The proposal would have removed requirements that creditors verify this information in accordance with appendix Q and would have removed appendix Q from Regulation Z entirely.

To clarify the verification requirement in § 1026.43(e)(2)(v)(B), the Bureau proposed to add comments 43(e)(2)(v)(B)-1 through -3. Proposed comment 43(e)(2)(v)(B)-1 stated that § 1026.43(e)(2)(v)(B) does not prescribe specific methods of underwriting that creditors must use. This proposed comment further provided that, as long as a creditor complies with the provisions of § 1026.43(c)(3) with respect to verification of debt obligations, alimony, and child support and § 1026.43(c)(4) with respect to verification of income and assets, creditors would be permitted to use any reasonable verification methods and criteria.

The Bureau proposed comment 43(e)(2)(v)(B)-2 to clarify that “current and reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan” is determined in accordance with § 1026.43(c)(2)(i) and its commentary and that “current debt obligations, alimony, and child support” has the same meaning as under § 1026.43(c)(2)(vi) and its commentary. The proposed comment further stated that § 1026.43(c)(2)(i) and (vi) and the associated commentary apply to a creditor's determination with respect to what inflows and property it may classify and count as income or assets and what obligations it must classify and count as debt obligations, alimony, and child support, pursuant to its compliance with § 1026.43(e)(2)(v)(B).

Proposed comment 43(e)(2)(v)(B)-3.i provided that a creditor also complies with § 1026.43(e)(2)(v)(B) if the creditor satisfies specified verification standards (verification safe harbor). In the section-by-section analysis of proposed § 1026.43(e)(2)(v)(B), the Bureau stated that these verification standards may include relevant provisions in specified versions of the Fannie Mae Single Family Selling Guide, the Freddie Mac Single-Family Seller/Servicer Guide, the FHA's Single Family Housing Policy Handbook, the VA's Lenders Handbook, and the USDA's Field Office Handbook for the Direct Single Family Housing Program and the Handbook for the Single Family Guaranteed Loan Program (“manuals”), as of the date of the proposal's public release. The Bureau sought comment on whether these or other verification standards should be incorporated into proposed comment 43(e)(2)(v)(B)-3.i. In the section-by-section analysis of proposed § 1026.43(e)(2)(v)(B), the Bureau also encouraged stakeholders to develop additional verification standards and stated that it would review any such standards for potential inclusion in the safe harbor.

Proposed comment 43(e)(2)(v)(B)-3.ii provided that a creditor complies with § 1026.43(e)(2)(v)(B) if it complies with requirements in the verification standards listed in comment 43(e)(2)(v)(B)-3 for creditors to verify income or assets, debt obligations, alimony and child support using specified documents or to include or exclude particular inflows, property, and obligations as income, assets, debt obligations, alimony, and child support. Proposed comment 43(e)(2)(v)(B)-3.iii stated that, for purposes of compliance with § 1026.43(e)(2)(v)(B), a creditor need not comply with requirements in the verification standards listed in comment 43(e)(2)(v)(B)-3.i other than those that require creditors to verify income, assets, debt obligations, alimony, and child support using specified documents or to classify and count particular inflows, property, and obligations as income, assets, debt obligations, alimony, and child support.

Proposed comment 43(e)(2)(v)(B)-3.iv stated that a creditor also complies with § 1026.43(e)(2)(v)(B) where it complies with revised versions of verification standards listed in comment 43(e)(2)(v)(B)-3.i, provided that the two versions are substantially similar. Finally, proposed comment 43(e)(2)(v)(B)-3.v provided that a creditor complies with § 1026.43(e)(2)(v)(B) if it complies with the verification requirements in one or more of the verification standards specified in comment 43(e)(2)(v)(B)-3.i. The proposed comment stated that, accordingly, a creditor may, but need not, comply with § 1026.43(e)(2)(v)(B) by complying with the verification standards from more than one manual (in other words, by “mixing and matching” verification requirements).

For the reasons described below, the Bureau adopts § 1026.43(e)(2)(v)(B) and comments 43(e)(2)(v)(B)-1 through -3 as proposed, except that, in this final rule, § 1026.43(e)(2)(v)(B) lists the applicable verification standards for the verification safe harbor in comment 43(e)(2)(v)(B)-3.i and includes minor edits to provide clarity. The verification standards listed in comment 43(e)(2)(v)(B)-3.i are the same verification standards that the Bureau listed in the proposal and stated that it may include in the verification safe harbor.

Comments Received

Commenters generally supported the Bureau's overall approach of replacing appendix Q with a requirement to use third-party records that provide reasonably reliable evidence of the consumer's income, assets, debt obligations, alimony, and child support. Several commenters recommended modifications to the proposal, as described and organized below based on the topic of concern.^[275]

Verification safe harbor. Commenters generally supported including, in the list of specified external verification standards, the portions of the GSE, FHA, VA, and USDA manuals that the Bureau listed in the proposal. Both GSEs supported the safe harbor for the verification standards in their manuals resulting from proposed comment 43(e)(2)(v)(B)-3. Both GSEs stated that the commentary should reference not only the verification standards in their manuals but should also reference amendments, letters, and other creditor-specific waivers of provisions that are not included in their manuals. One GSE stated that the Bureau should require creditors to comply with its entire manual—not just with its verification standards—to receive the verification safe harbor. An industry commenter stated that automatic loan origination system reports, specifically Fannie Mae's Desktop Underwriter and Freddie Mac's Loan Prospector, should be conclusive proof of compliance with the verification requirements of § 1026.43(e)(2)(v)(B) and its related commentary. A research center commenter stated that, for rebuttable presumption General QM loans, income and debt verification is effectively the only issue a consumer might challenge, and therefore the verification safe harbor would result in creditors facing about the same legal exposure on a rebuttable presumption QM as on a safe harbor QM. The commenter asserted that this would provide less protection to consumers and more leverage for increased home prices.

The Bureau declines to extend the verification safe harbor for materials outside of the scope of the verification standards in the specified manuals. The Bureau is concerned that the automatic inclusion of any amendments or modifications to manuals could cause significant changes in the creditor obligations and consumer protections without review by the Bureau. The Bureau will monitor changes to the manuals and incorporate updated versions if necessary. The Bureau is also concerned about incorporating standards that are not publicly available. The Bureau also declines to extend the safe harbor for matters beyond the verification standards within the specified GSE manuals. The Bureau is not aware of a reason why a creditor's compliance with standards unrelated to verification should be required for the creditor to obtain the benefit of the safe harbor for compliance with the Bureau's verification requirement. In addition, referencing the rest of the GSE manuals could lead to confusion among creditors or secondary market participants, because those manuals also contain requirements not related to verification standards—for example, housing expense ratios, DTI limits, or LTV limits that may be inconsistent with the provisions on related issues in the General QM loan definition. The Bureau also declines to extend a verification safe harbor merely for the inclusion of an approval acknowledgment generated by an automated underwriting system maintained by the GSEs or other institution, because modifications to the automated underwriting system approval process may deviate from the specified manuals and the Bureau would not be able to evaluate the nature and extent of such deviations without prior review.

The Bureau additionally disagrees with the research center commenter's assertion that the verification safe harbor would result in creditors facing about the same legal exposure on a rebuttable presumption QM as on a safe harbor QM. The Bureau notes that the verification safe harbor provides creditors with a safe harbor only for compliance with the verification requirement in § 1026.43(e)(2)(v)(B). The verification safe harbor does not preclude consumers from asserting that the creditor did not comply with § 1026.43(e)(2)(v)(A), for example, by failing to take into account the consumer's DTI ratio or residual income in the creditor's ability-to-repay determination. Moreover, consumers could still rebut the presumption by demonstrating that they had insufficient residual income to cover their living expenses as explained in comment 43(e)(1)(ii)-1.

Use of revised manuals that are substantially similar. The Bureau requested comment on whether creditors that comply with verification standards in revised versions of the listed manuals that are substantially similar to the listed versions should also receive a verification safe harbor, as the Bureau proposed. The Bureau also requested comment on whether the Rule should include illustrations of revisions to the manuals that might qualify as substantially similar, and if so, what types of illustrations would provide helpful clarification to creditors and other stakeholders.

Commenters generally supported the inclusion of a verification safe harbor for verification standards in the listed manuals that have been revised but are substantially similar, but some commenters suggested alternative approaches. A GSE supported the substantially similar standard but requested that the Bureau clarify the meaning of substantially similar. In contrast, some industry commenters stated that creditors should receive a safe harbor for compliance with the revised version of the manuals whether or not they are substantially similar. Some industry commenters stated that the Bureau should adjust the commentary to presume the revised versions of manuals are valid unless they materially deviate from the prior version. Some industry commenters stated that the Bureau should adopt a mechanism by which the Bureau could review and determine if revised manuals are substantially similar to the versions referenced in comment 43(e)(2)(v)(B)-3.i. Some industry commenters stated that the Bureau should include a statement that affirms that verification standards adopted by a creditor that are materially similar to those in the manuals referenced in comment 43(e)(2)(v)(B)-3.i should also receive a verification safe harbor.

The Bureau is adopting comment 43(e)(2)(v)(B)-3.i as proposed. The Bureau determines that commenters' suggested clarifications of the substantially similar standard in fact would not provide greater clarity. For example, the Bureau determines that a standard providing that the revised manual receives a verification safe harbor provided that it does not “materially deviate” or is “materially similar” would not be appreciably clearer than a standard that the revised manual be “substantially similar.”

The Bureau additionally notes that, in proposing to extend the verification safe harbor to substantially similar versions of the verification standards in the manuals, the Bureau did not intend for creditors to always be responsible for determining on their own whether a revised version of a listed manual is substantially similar to a version adopted in this final rule. Rather, the Bureau intends to provide further clarity to creditors by releasing guidance, as appropriate, regarding whether future revisions of manuals qualify as “substantially similar” for purposes of the verification safe harbor. The following three illustrations show how the Bureau may evaluate future changes to the manuals. The Bureau believes these illustrations may help creditors anticipate if and when the Bureau may address whether future revisions of manuals are eligible for a safe harbor.

First, revisions only to provisions within the manuals that are not referenced in comment 43(e)(2)(v)(B)-3.i would result in a revised version that is substantially similar. For example, a revised version of the FHA's Single Family Housing Policy Handbook that makes changes only to Section III, Servicing and Loss Mitigation, would be substantially similar for purposes of comment 43(e)(2)(v)(B)-3.i because there are no changes to the verification standards contained in Sections II.A.1 and II.A.4-5 of that Handbook.

Second, the portions of the manuals referenced in comment 43(e)(2)(v)(B)-3.i contain not only verification standards, but also additional provisions related to the underwriting of the mortgage. Consistent with comment 43(e)(2)(v)(B)-3.iii, revisions only to these unrelated underwriting provisions would produce a revised version that would be substantially similar. As an illustration, the Freddie Mac Single-Family Seller/Servicer Guide chapter 5401.1 requires a review of the consumer's monthly housing expense-to-income ratio. Chapter 5401.1 is contained within the portions of the Freddie Mac Single-Family Seller/Servicer Guide listed in comment 43(e)(2)(v)(B)-3.i. However, revised versions of Chapter 5401.1 concerning a consumer's monthly housing expense-to-income ratio would be substantially similar to the manual in comment 43(e)(2)(v)(B)-3.i, since these provisions of chapter 5401.1 do not relate to the verification of income, assets, debt obligations, alimony, or child support by use of reasonably reliable third-party records.

Third, revisions to the manuals concerning verification standards may or may not be substantially similar. The Bureau may evaluate such revisions to determine if the revised manual is substantially similar to the version referenced in comment 43(e)(2)(v)(B)-3.i. As an illustration, Fannie Mae Selling Guide chapter B-3-3.2-01 generally requires two years of individual and business tax returns to verify a consumer's income. Business tax returns, however, are not required if the consumer is using personal funds to pay for down payment, closing, and escrow account amounts; the consumer has been in same business for five years; and the consumer's individual tax returns show an increase in self-employment income. A revised version of the Fannie Mae Selling Guide that amends chapter B-3-3.2-01 to change any of these requirements for verifying self-employed income may or may not make the revised Selling Guide substantially similar to the Fannie Mae Selling Guide issued on June 3, 2020. The Bureau may consider providing additional guidance to address any such revisions.

“Mixing and matching” of verification standards. The Bureau also sought comment on its proposal to allow creditors to “mix and match” verification standards from different manuals, including whether examples of such mixing and matching would be helpful and whether the Bureau should instead limit or prohibit such mixing and matching, and why. Some industry commenters supported the ability of creditors to mix and match the verification standards from the manuals because it would provide flexibility and would not restrict creditors from adopting wholesale verification standards from a single external party. Some consumer advocate commenters opposed permitting creditors to mix and match verification standards from the manuals because allowing mixing and matching would introduce unnecessary subjectivity into the rule, although the commenters did not explain how. These consumer advocate commenters also stated that allowing mixing and matching could enable creditors to exploit differences in approaches between manuals. These commenters did not explain or provide examples of how creditors might do so or of what harm could result.

The Bureau concludes that permitting creditors to mix and match standards for verifying income, assets, debt obligations, alimony, and child support from each of the manuals would provide creditors with greater flexibility without undermining consumer protection. The GSEs and Federal agencies that maintain the manuals have had considerable historical experience in determining which records and supplemental records are reasonably reliable third-party records for purposes of verifying income, assets, debt obligations, alimony, and child support, as well as determining the need for updated information over applicable timeframes. Each of the manuals has also been historically relied upon for those purposes by Congress, the Bureau, secondary market participants, and creditors. Congress included separate QM definitions for loans insured or guaranteed by FHA, VA, and USDA without establishing separate third-party verification standards other than those established by their respective agencies.^[276] The third-party verification standards of the GSEs also served as a basis for verification under the Temporary GSE QM loan definition under § 1026.43(e)(4), and the Bureau is not aware of resulting instances of harm caused by inadequately verified income or assets, debt obligations, alimony and child support.

The Bureau has analyzed the relevant provisions of the manuals and has not identified ways that creditors may exploit differences between them or how mixing and matching would add subjectivity to the ATR/QM Rule's verification requirements. As noted, commenters did not cite examples of how this might occur. Permitting creditors to mix and match verification standards may allow creditors to use different manuals, but the Bureau has not identified evidence that combinations of historically accepted third-party record verification standards will, by virtue of their combination, result in insufficient verification of income, assets, debt obligations, alimony, or child support because the creditor uses different manuals for the verification of the information provided. The Bureau also determines, based on its analysis of the relevant provisions of the manuals, that permitting creditors to “mix and match” would not add subjectivity to the Rule's verification requirements.

Adding standards created by a self-regulatory organization (SRO). In the General QM Proposal, the Bureau encouraged stakeholders to develop additional verification standards that the Bureau could incorporate into the verification safe harbor and stated that it would review any such standards for potential inclusion in the safe harbor. Commenters did not provide any stakeholder-developed verification standards for review. However, several industry commenters stated that the Bureau should use verification standards adopted by a self-regulatory organization (SRO), in addition to or as a replacement for the standards listed in the proposal. Commenters that suggested this approach generally discussed such adoption as a future objective, as such standards, or even such an SRO, do not appear to exist at this time. One of these commenters recommended that the Bureau include in the safe harbor the GSE and Federal agency manuals listed in the proposal only until an industry-developed standard is established and approved by the Bureau.

The Bureau notes that there is no evidence in the record that such an SRO, much less verification standards created by such an entity or other consortium of industry stakeholders, exists. Accordingly, the Bureau determines that it would be premature to include such standards in the verification safe harbor. However, the Bureau continues to encourage stakeholders, including groups of stakeholders, to develop verification standards.^[277] The Bureau is interested in reviewing any such standards that stakeholders develop for potential inclusion in the verification safe harbor. Stakeholder standards could incorporate, in whole or in part, any standards that the Bureau specifies as providing a verification safe harbor, including mixing and matching these standards.

Preventing use of fraudulent documentation. The joint consumer advocate term sheet requested that the Bureau affirm that documentation that is falsified or the subject of fraud by or with the knowledge and consent of the lender, broker, or their agents would not comply with the verification requirement in § 1026.43(e)(2)(v)(B). The Bureau agrees that falsified or fraudulent documentation is, by definition, not a “reasonably reliable” third party record. The Bureau further notes that creditors have legal obligations to protect against such instances of mortgage fraud.^[278] The Bureau also notes that the manuals listed in the verification safe harbor have embedded limitations and restrictions on what third-party documentation may be used for verification that address similar sources of law. Accordingly, the Bureau determines that the issues presented by commenters are already adequately addressed by this final rule and by existing legal requirements.

The Final Rule

The Bureau is adopting § 1026.43(e)(2)(v)(B) and comments 43(e)(2)(v)(B)-1 through -3 as proposed, except that, in this final rule, § 1026.43(e)(2)(v)(B) lists the applicable verification standards for the verification safe harbor in comment 43(e)(2)(v)(B)-3.i.^[279] These verification standards are: (1) Chapters B3-3 through B3-6 of the Fannie Mae Single Family Selling Guide, published June 3, 2020; (2) sections 5102 through 5500 of the Freddie Mac Single-Family Seller/Servicer Guide, published June 10, 2020; (3) sections II.A.1 and II.A.4-5 of the FHA's Single Family Housing Policy Handbook, issued October 24, 2019; (4) chapter 4 of the VA's Lenders Handbook, revised February 22, 2019; (5) chapter 4 of the USDA's Field Office Handbook for the Direct Single Family Housing Program, revised March 15, 2019; and (6) chapters 9 through 11 of the USDA's Handbook for the Single Family Guaranteed Loan Program, revised March 19, 2020. These verification standards are the same standards that the Bureau listed in the proposal and requested comment on. Based on its review of the standards and the comments received, Bureau concludes that each of the verification standards listed in comment 43(e)(2)(v)(B)-3.i is sufficient to satisfy the final rule's verification requirement.

The Bureau concludes that these amendments to § 1026.43(e)(2)(v)(B) will ensure that the ATR/QM Rule's verification requirements are clear and detailed enough to provide creditors with sufficient certainty about whether a loan satisfies the General QM loan definition. The Bureau concludes that, without such certainty, creditors may be less likely to provide General QMs to consumers, reducing the availability of responsible, affordable mortgage credit. The Bureau also finds that these verification requirements are flexible enough to adapt to emerging issues with respect to the treatment of certain types of income, assets, debt obligations, alimony, and child support, advancing the provision of responsible, affordable mortgage credit to consumers. The Bureau aims to ensure that the verification requirement provides substantial flexibility for creditors to adopt innovative verification methods, such as the use of bank account data that identifies the source of deposits to determine personal income, while also specifying examples of compliant verification standards to provide greater certainty that a loan has QM status.

As described above, this final rule provides that creditors must verify income, assets, debt obligations, alimony, and child support in accordance with the general ATR verification provisions in § 1026.43(c)(3) and (4). This final rule also provides a safe harbor for compliance with § 1026.43(e)(2)(v)(B) if a creditor complies with verification standards in the manuals listed in comment 43(e)(2)(v)(B)-3.i. These verification standards are available to the public for free online.^[280]

The Bureau determines, based on extensive public feedback and its own experience and review, that these external standards are reasonable and would provide creditors with substantially greater certainty about whether many loans satisfy the General QM loan definition—particularly with respect to verifying income for self-employed consumers, consumers with part-time employment, and consumers with irregular or unusual income streams. The Bureau determines that these types of income would be addressed more fully by these external standards than by appendix Q. The Bureau determines that, as a result, final § 1026.43(e)(2)(v)(B) would increase access to responsible, affordable credit for consumers.

The Bureau emphasizes that a creditor would not be required to comply with any of the verification standards listed in comment 43(e)(2)(v)(B)-3.i in order to comply with § 1026.43(e)(2)(v)(B). Rather, under this final rule, compliance with the listed verification standards constitutes compliance with the verification requirements of § 1026.43(c)(3) and (4) and their commentary, which generally require creditors to verify income, assets, debt obligations, alimony, and child support using reasonably reliable third-party records. The Bureau determines that this would help address the compliance concerns of many creditors and commenters associated with appendix Q's lack of clarity.

The Bureau also determines that this final rule would provide creditors with the flexibility to develop other methods of compliance with the verification requirements of § 1026.43(e)(2)(v)(B), consistent with § 1026.43(c)(3) and (4) and their commentary, an option that the Bureau intends to address the concerns of creditors and commenters that found appendix Q to be too rigid or prescriptive. As explained in comment 43(e)(2)(v)(B)-1, § 1026.43(e)(2)(v)(B) does not prescribe specific methods of underwriting, and as long as a creditor complies with § 1026.43(c)(3) and (4), the creditor is permitted to use any reasonable verification methods and criteria. Furthermore, as comment 43(e)(2)(v)(B)-3.v clarifies, creditors have the flexibility to mix and match the verification requirements in the standards specified in comment 43(e)(2)(v)(B)-3.i, and receive a safe harbor with respect to verification that is made consistent with those standards.

Comment 43(e)(2)(v)(B)-3.iv explains that a creditor complies with § 1026.43(e)(2)(v)(B) if it complies with revised versions of the verification standards specified in comment 43(e)(2)(v)(B)-3.i, provided that the two versions are substantially similar. The GSE and Federal agency standards listed in comment 43(e)(2)(V)(B)-3.i are regularly updated in response to emerging issues with respect to the treatment of certain types of debt or income. This comment explains that the safe harbor described in comment 43(e)(2)(v)(B)-3.i applies not only to verification standards in the specific listed versions, but also to revised versions of these verification standards, as long as the revised version is substantially similar.

As discussed above, the Bureau encourages stakeholders, including groups of stakeholders, to develop verification standards. The Bureau is interested in reviewing any such standards for potential inclusion in the verification safe harbor. Stakeholder standards could incorporate, in whole or in part, any standards that the Bureau specifies as providing a safe harbor, including mixing and matching these standards.

Legal Authority

The Bureau is incorporating the requirement that the creditor verify the consumer's current or reasonably expected income, assets other than the value of the dwelling (including any real property attached to the dwelling), debt obligations, alimony, and child support into the definition of a General QM in § 1026.43(e)(2) and revisions to its commentary pursuant to its authority under TILA section 129C(b)(3)(B)(i). The Bureau finds that these provisions are necessary and proper to ensure that responsible, affordable mortgage credit remains available to consumers in a manner that is consistent with the purposes of TILA section 129C and necessary and appropriate to effectuate the purposes of TILA section 129C, which includes assuring that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loan.

The Bureau also adopts these provisions pursuant to its authority under TILA section 105(a) to issue regulations that, among other things, contain such additional requirements or other provisions, or that provide for such adjustments for all or any class of transactions, that in the Bureau's judgment are necessary or proper to effectuate the purposes of TILA, which include the above purpose of section 129C, among other things. The Bureau finds that these provisions are necessary and proper to achieve this purpose. In particular, the Bureau finds that incorporating the requirement that a creditor verify a consumer's current debt obligations, alimony, and child support into the General QM criteria—as well as clarifying that a creditor complies with the General QM verification requirement where it complies with certain verification standards issued by third parties that the Bureau would specify—ensures that creditors verify whether a consumer has the ability to repay a General QM. Finally, the Bureau concludes that these regulatory amendments are authorized by TILA section 129C(b)(2)(A)(vi), which permits, but does not require, the Bureau to adopt guidelines or regulations relating to debt-to-income ratios or alternative measures of ability to pay regular expenses after payment of total monthly debt.

43(e)(2)(vi)

TILA section 129C(b)(2)(vi) states that the term “qualified mortgage” includes any mortgage loan that complies with any guidelines or regulations established by the Bureau relating to ratios of total monthly debt to monthly income or alternative measure of ability to pay regular expenses after payment of total monthly debt, taking into account the income levels of the consumer and such other factors as the Bureau may determine relevant and consistent with the purposes described in TILA section 129C(b)(3)(B)(i). TILA section 129C(b)(3)(B)(i) authorizes the Bureau to revise, add to, or subtract from the criteria that define a QM upon a finding that the changes are necessary or proper to ensure that responsible, affordable mortgage credit remains available to consumers in a manner consistent with the purposes of TILA section 129C, necessary and appropriate to effectuate the purposes of TILA sections 129C and 129B, to prevent circumvention or evasion thereof, or to facilitate compliance with TILA sections 129C and 129B. Current § 1026.43(e)(2)(vi) implements TILA section 129C(b)(2)(vi), consistent with TILA section 129C(b)(3)(B)(i), and provides that, as a condition to be a General QM under § 1026.43(e)(2), the consumer's total monthly DTI ratio may not exceed 43 percent. Section 1026.43(e)(2)(vi) further provides that the consumer's total monthly DTI ratio is generally determined in accordance with appendix Q.

For the reasons described in part V above, the Bureau proposed to remove the 43 percent DTI limit in current § 1026.43(e)(2)(vi) and replace it with a price-based approach. The proposal also would have required a creditor to consider the consumer's DTI ratio or residual income, income or assets other than the value of the dwelling, and debts and verify the consumer's income or assets other than the value of the dwelling and the consumer's debts. Specifically, the Bureau proposed to remove the text of current § 1026.43(e)(2)(vi) and to provide instead that, to be a General QM under § 1026.43(e)(2), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by the amounts specified in § 1026.43(e)(2)(vi)(A) through (E).^[281] Proposed § 1026.43(e)(2)(vi)(A) through (E) provided specific rate-spread thresholds for purposes of § 1026.43(e)(2), including higher thresholds for small loan amounts and subordinate-lien transactions. Proposed § 1026.43(e)(2)(vi)(A) provided that for a first-lien covered transaction with a loan amount greater than or equal to $109,898 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by two or more percentage points. Proposed § 1026.43(e)(2)(vi)(B) and (C) provided higher thresholds for smaller first-lien covered transactions. Proposed § 1026.43(e)(2)(vi)(D) and (E) provided higher thresholds for subordinate-lien covered transactions. Under the proposal, loans priced at or above the thresholds in proposed § 1026.43(e)(2)(vi)(A) through (E) would not have been eligible for QM status under § 1026.43(e)(2). The proposal also provided that the loan amounts specified in § 1026.43(e)(2)(vi)(A) through (E) would be adjusted annually for inflation based on changes in the Consumer Price Index for All Urban Consumers (CPI-U).

Proposed § 1026.43(e)(2)(vi) also provided a special rule for determining the APR for purposes of determining a loan's status as a General QM loan under § 1026.43(e)(2) for certain ARMs and other loans for which the interest rate may or will change in the first five years of the loan. Specifically, proposed § 1026.43(e)(2)(vi) provided that, for purposes of § 1026.43(e)(2)(vi), the creditor must determine the APR for a loan for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due by treating the maximum interest rate that may apply during that five-year period as the interest rate for the full term of the loan.

The Bureau proposed these revisions to § 1026.43(e)(2)(vi) for the reasons set forth above in part V.B. As explained above, the Bureau proposed to remove the 43 percent DTI limit in current § 1026.43(e)(2)(vi) and replace it with a price-based approach because the Bureau is concerned that retaining the existing General QM loan definition with the 43 percent DTI limit after the Temporary GSE QM loan definition expires would significantly reduce the size of the QM market and could significantly reduce access to responsible, affordable credit. The Bureau proposed a price-based approach to replace the specific DTI limit approach because it is concerned that imposing a DTI limit as a condition for QM status under the General QM loan definition may be overly burdensome and complex in practice and may unduly restrict access to credit because it provides an incomplete picture of the consumer's financial capacity. In the proposal, the Bureau preliminarily concluded that a price-based General QM loan definition is appropriate because a loan's price, as measured by comparing a loan's APR to APOR for a comparable transaction, is a strong indicator of a consumer's ability to repay and is a more holistic and flexible measure of a consumer's ability to repay than DTI alone.

The Bureau also proposed to remove current comment 43(e)(2)(vi)-1, which relates to the calculation of monthly payments on a covered transaction and for simultaneous loans for purposes of calculating the consumer's DTI ratio under current § 1026.43(e)(2)(vi). The Bureau did so because, under the proposal to move the text of current § 1026.43(e)(2)(vi)(B) and revise it to remove the references to appendix Q, current comment 43(e)(2)(vi)-1 would have been unnecessary. The Bureau proposed to replace current comment 43(e)(2)(vi)-1 with a cross-reference to comments 43(b)(4)-1 through -3 for guidance on determining APOR for a comparable transaction as of the date the interest rate is set. The Bureau also proposed new comment 43(e)(2)(vi)-2, which provided that a creditor must determine the applicable rate-spread threshold based on the face amount of the note, which is the “loan amount” as defined in § 1026.43(b)(5), and provided an example of a $75,000 loan amount that would fall into the proposed tier for loans greater than or equal to $65,939 (indexed for inflation) but less than $109,898 (indexed for inflation). In addition, the Bureau proposed comment 43(e)(2)(vi)-3 in which it would have published the annually adjusted loan amounts to reflect changes in the CPI-U. The Bureau also proposed new comment 43(e)(2)(vi)-4 to explain the proposed special rule that, for purposes of § 1026.43(e)(2)(vi), the creditor must determine the APR for a loan for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due by treating the maximum interest rate that may apply during that five-year period as the interest rate for the full term of the loan. The Bureau did not receive comments regarding comments 43(e)(2)(vi)-1 through -3 and is adopting them as proposed, except that the $65,939 and $109,898 loan amount thresholds in comment 43(e)(2)(vi)-2 have been revised to $66,156 and $110,260, respectively, for consistency with the Bureau's recently-issued final rule that adjusted for inflation the related thresholds in comment 43(e)(3)(ii)-1.^[282] The Bureau is also adopting comment 43(e)(2)(vi)-4 as proposed and that comment is discussed further below.

For the reasons discussed in part V and below, the Bureau is adopting a price-based approach to defining General QMs in § 1026.43(e)(2)(vi) pursuant to its authority under TILA section 129C(b)(3)(B)(i). The Bureau concludes that a price-based approach to the General QM loan definition is necessary and proper to ensure that responsible, affordable mortgage credit remains available to consumers in a manner that is consistent with the purposes of TILA section 129C and is necessary and appropriate to effectuate the purposes of TILA section 129C, which includes assuring that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loan.

As noted above in part V, the Bureau concludes that a price-based General QM loan definition best balances consumers' ability to repay with ensuring access to responsible, affordable mortgage credit. The Bureau is amending the General QM loan definition because retaining the existing 43 percent DTI limit would reduce the size of the QM market and likely would lead to a significant reduction in access to responsible, affordable credit when the Temporary GSE QM definition expires. The Bureau continues to believe that General QM status should be determined by a simple, bright-line rule to provide certainty of QM status, and the Bureau concludes that pricing achieves this objective. Furthermore, the Bureau concludes that pricing, rather than a DTI limit, is a more appropriate standard for the General QM loan definition. While not a direct measure of financial capacity, loan pricing is strongly correlated with early delinquency rates, which the Bureau uses as a proxy for repayment ability. The Bureau concludes that conditioning QM status on a specific DTI limit would likely impair access to credit for some consumers for whom it is appropriate to presume their ability to repay their loans at consummation. Although a pricing limit that is set too low could also have this effect, compared to DTI, loan pricing is a more flexible metric because it can incorporate other factors that may also be relevant to determining ability to repay, including credit scores, cash reserves, or residual income. The Bureau concludes that a price-based General QM loan definition is better than the alternatives because a loan's price, as measured by comparing a loan's APR to APOR for a comparable transaction, is a strong indicator of a consumer's ability to repay and is a more holistic and flexible measure of a consumer's ability to repay than DTI alone.

The Bureau concludes that a price-based approach to the General QM loan definition will both ensure that responsible, affordable mortgage credit remains available to consumers and assure that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loan. For these same reasons, the Bureau is adopting a price-based requirement in § 1026.43(e)(2)(vi) pursuant to its authority under TILA section 105(a) to issue regulations that, among other things, contain such additional requirements or other provisions, or that provide for such adjustments for all or any class of transactions, that in the Bureau's judgment are necessary or proper to effectuate the purposes of TILA, which include the above purpose of section 129C, among other things. The Bureau concludes that the price-based addition to the General QM criteria is necessary and proper to achieve this purpose, for the reasons described above in part V. Finally, the Bureau concludes a price-based approach is authorized by TILA section 129C(b)(2)(A)(vi), which permits, but does not require, the Bureau to adopt guidelines or regulations relating to DTI ratios or alternative measures of ability to pay regular expenses after payment of total monthly debt.

43(e)(2)(vi)(A)

The Bureau's Proposal

Proposed § 1026.43(e)(2)(vi)(A) provided that, for a first-lien covered transaction with a loan amount greater than or equal to $109,898 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 2 or more percentage points. Thus, under the proposal, loans priced at or above the proposed 2-percentage-point threshold would not have been eligible for QM status under § 1026.43(e)(2) (except that, as discussed below, the proposal provided higher thresholds for loans with smaller loan amounts and for subordinate-lien transactions).

In the proposal, the Bureau stated that the 2002-2008 time period corresponds to a market environment that, in general, demonstrates looser, higher-risk credit conditions and that ended with very high unemployment and falling home prices. The Bureau's analysis set forth in Table 5 found direct correlations between rate spreads and early delinquency rates across all DTI ranges reviewed. The proposal stated that loans with low rate spreads had relatively low early delinquency rates even at high DTI levels and the highest early delinquency rates corresponded to loans with both high rate spreads and high DTI ratios. For loans with DTI ratios of 41 to 43 percent—the category in Table 5 that includes the current DTI limit of 43 percent—the early delinquency rates reached 16 percent at rate spreads including and above 2.25 percentage points over APOR. At rate spreads inclusive of 1.75 through 1.99 percentage points over APOR—the category that is just below the proposed 2 percentage-point rate-spread threshold—the early delinquency rate reached 22 percent for DTI ratios of 61 to 70 percent. At DTI ratios of 41 to 43 percent and rate spreads inclusive of 1.75 through 1.99 percentage points over APOR, the early delinquency rate is 15 percent.

In the proposal, the Bureau stated that, in contrast to Table 5, the 2018 time period in Table 6 corresponds to a market environment that, in general, demonstrates tighter, lower-risk credit conditions and that featured very low unemployment and rising home prices. The proposal stated that this more recent sample of data provides insight into early delinquency rates under post-crisis lending standards for a dataset of loans that had not undergone an economic downturn. In the 2018 data in Table 6, early delinquency rates also increased as rate spreads increased across each range of DTI ratios analyzed, although the overall performance of loans in the Table 6 dataset was significantly better than those represented in Table 5. For loans with DTI ratios of 36 to 43 percent—the category in Table 6 that includes the current DTI limit of 43 percent—early delinquency rates reached 3.9 percent (at rate spreads of at least 2 percentage points). The highest early delinquency rate associated with the proposed rate-spread threshold (less than 2 percentage points over APOR) is 3.2 percent and corresponds to loans with the DTI ratios of 26 to 35 percent. At the same rate-spread threshold, the early delinquency rate for the loans with the highest DTI ratios is 2.3 percent. The Bureau stated that the apparent anomalies in the progression of the early delinquency rates across DTI ratios at the higher rate spread categories in Table 6 are likely because there are relatively few loans in the 2018 data with the indicated combinations of higher rate spreads and lower DTI ratios and some creditors require that consumers demonstrate more compensating factors on higher DTI loans.

In the proposal, the Bureau stated that, although in Tables 5 and 6 delinquency rates rise with rate spread, there is no clear point at which delinquency rates accelerate and comparisons between a high-risk credit market (Table 5) and a low-risk credit market (Table 6) show substantial expansion of early delinquency rates during an economic downturn across all rate spreads and DTI ratios. Data show that, for example, prime loans that experience a 0.2 percent early delinquency rate in a low-risk market might experience a 2 percent early delinquency rate in a higher-risk market, while subprime loans with a 4.2 percent early delinquency rate in a low-risk market might experience a 19 percent early delinquency rate in a higher-risk market.

The proposal referenced data and analyses provided by CoreLogic and the Urban Institute, as discussed in part V.B.2 above, which the Bureau stated also show a strong positive correlation of delinquency rates with interest rate spreads. The Bureau stated that this evidence collectively suggests that higher rate spreads—including the specific measure of APR over APOR—are strongly correlated with early delinquency rates. The proposal stated the Bureau's expectation that, for loans just below the respective thresholds, a pricing threshold of 2 percentage points over APOR would generally result in similar or somewhat higher early delinquency rates relative to the current DTI limit of 43 percent. However, the proposal stated that Bureau analysis shows the early delinquency rate for this set of loans is on par with loans that have received QM status under the Temporary GSE QM loan definition. Restricting the sample of 2018 NMDB-HMDA matched first-lien conventional purchase originations to only those purchased and guaranteed by the GSEs, the proposal stated that loans with rate spreads at or above 2 percentage points had an early delinquency rate of 4.2 percent, higher than the maximum early delinquency rates observed for loans with rate spreads below 2 percentage points in either Table 2 (2.7 percent) or Table 6 (3.2 percent). The proposal explained that this comparison uses 2018 data on GSE originations because such loans were originated while the Temporary GSE QM loan definition was in effect and the GSEs were in conservatorship. The proposal further explained that GSE loans from the 2002 to 2008 period were originated under a different regulatory regime and with different underwriting practices (e.g., GSE loans more commonly had DTI ratios over 50 percent during the 2002 to 2008 period), and thus may not be directly comparable to loans made under the Temporary GSE QM loan definition.

In the proposal, the Bureau used 2018 HMDA data to estimate that 95.8 percent of conventional purchase loans currently meet the criteria to be defined as QMs, including under the Temporary GSE QM loan definition. The Bureau also used 2018 HMDA data to project that the proposed 2 percentage-point-over-APOR threshold would result in a 96.1 percent market share for QMs with an adjustment for small loans, as discussed below. The Bureau stated that creditors may also respond to such a threshold by lowering pricing on some loans near the threshold, further increasing the QM market share. The proposal stated that, using the size of the QM market as an indicator of access to credit, the Bureau expects that a pricing threshold of 2 percentage points over APOR, in combination with the proposed adjustments for small loans, would result in an expansion of access to credit as compared to the current rule including the Temporary GSE QM loan definition, particularly as creditors are likely to adjust pricing in response to the rule, allowing additional loans to obtain QM status. The Bureau also acknowledged, however, that some loans that do not meet the current General QM loan definition, but that would be General QMs under the proposed price-based approach, would have been made under other QM definitions (e.g., FHA, small-creditor QM). Further, the Bureau stated that the proposal would result in a substantial expansion of access to credit as compared to the current rule without the Temporary GSE QM loan definition, under which only an estimated 73.6 percent of conventional purchase loans would be QMs.

In the proposal, the Bureau tentatively concluded that, in general, a 2 percentage-point-over-APOR threshold would appropriately balance ensuring consumers' ability to repay with maintaining access to responsible, affordable mortgage credit. The Bureau requested comment on the threshold amount, as well as comment on expected market changes and the possibility of adjusting the threshold in emergency situations. For the reasons discussed below, the Bureau is finalizing § 1026.43(e)(2)(vi)(A) with a threshold of 2.25 percentage points over APOR for transactions with a loan amount greater than or equal to $110,260 (indexed for inflation).

Comments Received

The Bureau received several comments concerning the proposed 2-percentage-point threshold for General QM eligibility under § 1026.43(e)(2)(vi)(A).^[283] Various commenters supported finalizing the proposed threshold or raising it by some unspecified amount. A GSE supported the proposed 2-percentage-point threshold to both continue access to affordable credit and ensure consumers' ability to repay. Another GSE supported the 2-percentage-point threshold and stated it was equally supportive of increasing the threshold by an unspecified amount. Similarly, an industry commenter stated that it does not oppose increasing the threshold by some unspecified amount.

Some comments, including one from an academic commenter and a joint comment from consumer advocates, generally opposed a price-based approach but also stated concerns specifically regarding the proposed 2-percentage-point threshold for QM eligibility under § 1026.43(e)(2)(vi)(A). Citing an Urban Institute analysis that was also cited in the proposal,^[284] the comments stated that, among loans with rate spreads of 1.51 to 2.00 percentage points originated from 1995 through 2008, even 30-year fixed-rate, fully documented and fully amortizing loans had high delinquency rates—especially those originated during periods of greater rate spread compression. Citing General QM Proposal Tables 1 and 3 regarding 2002-2008 first-lien purchase originations (i.e., reproduced as Tables 1 and 3 above), the comments also stated that the 13 percent early delinquency rate for loans priced 1.75 to 1.99 percentage points above APOR is more than double the 6 percent early delinquency rate for loans with DTI ratios of 41 to 43 percent—and is almost double the 7 percent early delinquency rate for loans with DTI ratios of 46 to 48 percent.

A research center specifically recommended increasing the General QM eligibility threshold to 2.5 percentage points to balance ability to repay with access to credit. The commenter stated that, based on Fannie Mae and Black Knight McDash data, a 2.5-percentage-point threshold would increase the delinquency rate ^[285] but nonetheless the delinquency rate would remain low relative to delinquency rates experienced in the past 20 years. The research center also stated that, based on 2019 HMDA data, a 2.5-percentage-point threshold would cause 32,044 more loans to be QM-eligible than a 2-percentage-point threshold. The commenter further stated that FHA's QM rule does not limit pricing for rebuttable presumption QMs and thus increasing the Bureau's threshold under § 1026.43(e)(2)(vi)(A) would create a more level playing field and increase consumer choice.

An individual commenter generally supported proposed § 1026.43(e)(2)(vi)(A) but suggested incrementally increasing the General QM eligibility threshold to as high as 2.75 percentage points for transactions with lower points and fees. The commenter stated that the approach would provide more flexibility and help consumers avoid paying upfront points and fees.

Several commenters recommended increasing the General QM eligibility threshold to 3 percentage points. A joint comment from consumer advocate and industry groups included some signatories recommending a 3-percentage-point threshold and no signatories opposing it. Another joint comment from consumer advocate and industry groups supported a 3-percentage-point threshold to balance ability to repay with access to credit. The latter joint comment stated that, based on Fannie Mae data and accounting for current risk-based mortgage insurance premiums, a 3-percentage-point threshold would increase the early delinquency rate but nonetheless the delinquency rate would be low relative to the Great Recession. Citing an FHFA working paper that was also cited by the General QM Proposal,^[286] the joint comment further stated that loans with non-QM features—including interest-only loans, ARM loans that combined teaser rates with subsequent large jumps in payments, negative amortization loans, and loans made with limited or no documentation of the borrower's income or assets—accounted for about half of the rise in risk leading up to the 2008 financial crisis and subsequent passage of the Dodd-Frank Act. The joint comment stated that the Bureau should promote more consumers receiving the important benefits of the Dodd-Frank Act's QM product restrictions—including lower-income and minority consumers that would otherwise be disproportionally excluded—by increasing the threshold for QM eligibility under § 1026.43(e)(2)(vi)(A).

The Bureau also received comments—including one from a research center and a joint comment from consumer advocate and industry groups—recommending an increase in the General QM pricing threshold to account for possible future rate spread widening in the market, as also discussed above in part V.C with respect to the safe harbor threshold. The Bureau also received a joint comment from consumer advocates that generally opposed a price-based approach but also stated that the Bureau should not increase the General QM pricing threshold in future emergency situations without notice-and-comment rulemaking.

The Final Rule

For the reasons discussed below, the Bureau is adopting § 1026.43(e)(2)(vi)(A) with a threshold of 2.25 percentage points over APOR for transactions with a loan amount greater than or equal to $110,260 (indexed for inflation). The Bureau concludes that, for most first-lien covered transactions, a 2.25 percentage point pricing threshold strikes the best balance between ensuring consumers' ability to repay and ensuring access to responsible, affordable mortgage credit. The Bureau is adopting § 1026.43(e)(2)(vi)(A) with a $110,260 loan amount threshold for consistency with the Bureau's recently-issued final rule that adjusted for inflation the related $109,898 threshold in comment 43(e)(3)(ii)-1.^[287] As discussed below, the final rule provides higher thresholds for loans with smaller loan amounts and for subordinate-lien transactions. The final rule provides an increase from the proposed thresholds for some small manufactured housing loans to ensure continued access to credit.

The Bureau concludes that a General QM eligibility threshold lower than 2.25 percentage points would unduly limit some consumers to non-QM or FHA loans, which generally have materially higher costs, or would unduly result in some consumers not being able to obtain a loan at all despite their ability to afford one, given the current lack of a robust non-QM market.^[288] As discussed in part V.B.5 above, Table 7A shows that 96.3 percent of 2018 conventional first-lien purchase originations would have been QMs under this revised ATR/QM Rule, as compared to a 94.7 percent share under the existing ATR/QM Rule, including the Temporary GSE QM loan definition. As discussed in the Bureau's Dodd-Frank Act section 1022(b) analysis below, among loans that fall outside the current General QM loan definition because they have a DTI ratio above 43 percent, the Bureau estimates that 959,000 of these conventional loans in 2018 would fall within this final rule's General QM loan definition. The Bureau concludes that some consumers with those conventional loans with DTI ratios above 43 percent could have instead obtained non-QM or FHA loans, which generally have materially higher costs, but others would not have obtained a loan at all. For example, based on application-level data obtained from nine large lenders, the Assessment Report found that the January 2013 Final Rule eliminated between 63 and 70 percent of non-GSE eligible home purchase loans with DTI ratios above 43 percent.^[289] The Bureau concludes that a 2.25 percentage point General QM eligibility threshold helps address those access-to-credit concerns—including concerns related to certain ARMs and manufactured housing loans discussed below—while striking an appropriate balance with ability-to-repay concerns.

A 2.25 percentage point pricing threshold for QM eligibility under § 1026.43(e)(2)(vi)(A) is also supported by the Bureau's conclusion that the Dodd-Frank Act QM product restrictions contribute to ensuring that consumers have the ability to repay their loans and are important for maintaining and expanding access to responsible, affordable mortgage credit. The Bureau concludes that loans with non-QM features—including interest-only loans, negative amortization loans, and loans made with limited or no documentation of the borrower's income or assets—had a substantial negative effect on consumers' ability to repay leading up to the 2008 financial crisis and subsequent passage of the Dodd-Frank Act. The Bureau concludes that promoting access to more QMs with the important benefits of the Act's QM product restrictions will help ensure consumers' ability to repay. Furthermore, for General QMs priced greater than or equal to 1.5 but less than 2.25 percentage points above APOR, consumers would also be afforded the opportunity to rebut the creditor's QM presumption of compliance.

In response to commenters who stated that the early delinquency rate for the proposed 2-percentage-point threshold would be too high to justify a QM presumption of compliance, the Bureau acknowledges that Table 1 for 2002-2008 first-lien purchase originations shows a 14 percent early delinquency rate for loans priced 2.00 to 2.24 percentage points above APOR, as compared to a 13 percent early delinquency rate for loans priced 1.75 to 1.99 percentage points above APOR and a 12 percent early delinquency rate for loans priced 1.50 to 1.74 percentage points above APOR.^[290] The comparable early delinquency rates for 2018 loans from Table 2 also show a higher early delinquency rate for loans priced 2.00 percentage points or more above APOR compared to loans priced 1.50 to 1.99 percentage points above APOR: 4.2 percent versus 2.7 percent.^[291] However, Bureau analysis shows the early delinquency rate for this set of loans is on par with loans that have received QM status under the Temporary GSE QM loan definition. Specifically, when restricting the sample of 2018 NMDB-HMDA matched first-lien conventional purchase originations to only those purchased and guaranteed by the GSEs, loans with rate spreads at or above 2 percentage points had an early delinquency rate of 4.2 percent. As explained above, this comparison uses 2018 data because such loans were originated while the Temporary GSE QM loan definition was in effect and the GSEs were in conservatorship, whereas GSE loans from the 2002 to 2008 period were originated under a different regulatory regime and with different underwriting practices that may not be directly comparable to loans made under the Temporary GSE QM loan definition.

In response to commenters, and as discussed above in part V.C.4, the Bureau concludes that it would be premature at this point to increase the QM safe harbor threshold based on possible future spread widening both because of uncertainty regarding effects on APOR itself as well as insufficient evidence of a significant access-to-credit difference between safe harbor and rebuttable presumption QMs. But for the General QM eligibility threshold under § 1026.43(e)(2)(vi)(A), notwithstanding the uncertainty regarding effects on APOR itself, the Bureau concludes that a robust non-QM market has not yet emerged and, thus, loans that exceed that threshold may not be available to some consumers, even though they would have been within the consumer's ability to repay. Thus, the Bureau concludes that (in addition to the reasons above) future spread widening also supports the 2.25 percentage point pricing threshold because future spread widening poses a greater potential access-to-credit concern for the General QM eligibility threshold under § 1026.43(e)(2)(vi)(A) than for the safe harbor threshold under § 1026.43(b)(4), if levels of non-QM lending remain low. This conclusion is consistent with the Bureau's findings in the Assessment Report, which suggest that, while the safe harbor threshold of 1.5 percentage points has not constrained lenders from originating rebuttable presumption QMs, only a modest amount of non-QM lending has occurred since the January 2013 Final Rule took effect.^[292] Moreover, the Bureau will monitor the market and take action as needed to maintain the best balance between consumers' ability to repay and access to responsible, affordable mortgage credit.

The Bureau concludes that it has insufficient evidence as to whether a threshold higher than 2.25 percentage points would strike the best balance with ability-to-repay concerns, particularly given the limited expected access to credit gains from increasing the threshold higher than 2.25 percentage points.^[293] While the 14 percent early delinquency rate in Table 1 for loans priced 2.00 to 2.24 percentage points above APOR is the same early delinquency rate as for loans priced 2.25 percentage points or more above APOR, all loans with rate spreads of 2.25 percentage points or more needed to be grouped to ensure sufficient sample size for reliable analysis of the 2002-2008 data.^[294]

43(e)(2)(vi)(B)-(F)

Thresholds for Smaller Loans and Subordinate-Lien Transactions

The Bureau proposed to establish higher pricing thresholds for smaller loans. Under the proposal, smaller loans priced at or above the proposed thresholds would not have been eligible for QM status under § 1026.43(e)(2). Specifically, proposed § 1026.43(e)(2)(vi)(B) provided that, for first-lien covered transactions with loan amounts greater than or equal to $65,939 but less than $109,898, the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 3.5 or more percentage points.^[295] Proposed § 1026.43(e)(2)(vi)(C) provided that, for first-lien covered transactions with loan amounts less than $65,939, the APR may not exceed the APOR for a comparable transaction as of the date the interest rate is set by 6.5 or more percentage points.

The Bureau also proposed to establish higher thresholds for subordinate-lien transactions. Under the proposal, subordinate-lien transactions priced at or above the proposed thresholds would not have been eligible for QM status under § 1026.43(e)(2). Specifically, proposed § 1026.43(e)(2)(vi)(D) provided that, for subordinate-lien covered transactions with loan amounts greater than or equal to $65,939, the APR may not exceed the APOR for a comparable transaction as of the date the interest rate is set by 3.5 or more percentage points. Proposed § 1026.43(e)(2)(vi)(E) provided that, for subordinate-lien covered transactions with loan amounts less than $65,939, the APR may not exceed the APOR for a comparable transaction as of the date the interest rate is set by 6.5 or more percentage points.

The proposal also provided that the loan amounts specified in § 1026.43(e)(2)(vi)(A) through (E) would be adjusted annually for inflation based on changes in CPI-U. Specifically, the Bureau proposed adjusting the loan amounts in § 1026.43(e)(2)(vi) annually on January 1 by the annual percentage change in the CPI-U that was reported on the preceding June 1. The Bureau proposed publishing adjustments in new comment 43(e)(2)(vi)-3 after the June figures became available each year.

For the reasons discussed below, the Bureau is finalizing § 1026.43(e)(2)(vi)(B) through (E) as proposed, except that proposed § 1026.43(e)(2)(vi)(D) has been redesignated as § 1026.43(e)(2)(vi)(E) and proposed § 1026.43(e)(2)(vi)(E) has been redesignated as § 1026.43(e)(2)(vi)(F) because the Bureau is finalizing a threshold for smaller manufactured housing loans in § 1026.43(e)(2)(vi)(D).^[296] The Bureau is also finalizing two additional comments to clarify terms and phrases used in § 1026.43(e)(2)(vi)(D). Specifically, comment 43(e)(2)(vi)-5 clarifies that the term “manufactured home,” as used in § 1026.43(e)(2)(vi)(D), means any residential structure as defined under HUD regulations establishing manufactured home construction and safety standards (24 CFR 3280.2). The comment further clarifies that modular or other factory-built homes that do not meet the HUD code standards are not manufactured homes for purposes of § 1026.43(e)(2)(vi)(D). Comment 43(e)(2)(vi)-6 provides that the threshold in § 1026.43(e)(2)(vi)(D) applies to first-lien covered transactions less than $110,260 (indexed for inflation) that are secured by a manufactured home and land, or by a manufactured home only.

Comments Received

The Bureau received several comments from consumer advocates, the mortgage industry, research centers, and others in response to the proposed pricing thresholds for smaller loans and subordinate-lien transactions. While some commenters supported the Bureau's proposed thresholds, others expressed various concerns, as described below.

Pricing thresholds for smaller loans. Consumer advocates and industry commenters offered differing viewpoints on whether the Bureau should consider the creditor's costs in developing the thresholds for smaller loans. Consumer advocate commenters noted that the statute requires the Bureau to consider the consumer's ability to repay when defining General QM; thus, in developing thresholds, the Bureau should not consider the creditor's costs or profit margins, which the commenter perceived was the Bureau's basis for developing higher thresholds for smaller loans, absent a showing that the available credit is responsible and affordable. Conversely, industry commenters suggested that the Bureau should consider the creditor's costs in developing the thresholds for smaller loans, given the impact these costs have on the price of these loans, specifically manufactured housing loans. For example, these commenters noted that, despite having smaller loan amounts, manufactured housing loans, including chattel loans, tend to have the same or similar origination and servicing costs as traditional mortgages. They also asserted that, unlike traditional mortgages, manufactured housing loans, including chattel loans, lack access to secondary market funding and to private mortgage insurance to offset credit risk and protect against potential losses. Overall, industry commenters stated that the thresholds for smaller loans should provide creditors with the ability to recover their costs for originating and servicing smaller loans, and still originate qualified mortgages.

The Bureau also received comments about the impact of the proposed thresholds on low- to moderate-income and minority consumers and on land installment contracts. With respect to the former, one large credit union expressed concern about the impact the proposed loan amount thresholds for smaller loans would have on these consumers given the rise in home prices. In addition, one State trade association observed that some loans greater than $65,939 exceeded the proposed pricing thresholds due to various risk factors, such as high LTV ratios or negative credit history, and that it was unclear whether these risk factors were more common among low- to moderate-income and minority consumers. With respect to land installment contracts, consumer advocate commenters asserted that under the Bureau's proposed thresholds for smaller loans, land installment contracts would newly be eligible for QM status, which would impede consumer lawsuits against creditors.

Data to support the thresholds for smaller loans. Consumer advocate commenters recommended that the Bureau further refine the data used to support the thresholds for smaller loans. Specifically, they recommended that the Bureau refine the data to include the volume of loans in each rate-spread range, loan performance data using incremental rate-spread ranges instead of cumulative rate-spread ranges, and an analysis that separates chattel loans from real estate-secured mortgages.

A few consumer advocate commenters underscored the need for refining the data by analyzing the early delinquency rates shown in General QM Proposal Table 5,^[297] which, according to these commenters, indicate that the proposed thresholds for smaller loans would harm vulnerable consumers. Specifically, these commenters noted that for loans priced 2.25 or more percentage points above APOR and with a DTI ratio greater than 26 percent, early delinquency rates were 10 percent or higher; and for similarly priced loans with DTI ratios between 40 and 50 percent, early delinquency rates were between 16 to 19 percent. These commenters also noted that General QM Proposal Table 5 did not show the early delinquency rate for 2002-2008 first-lien purchase originations in the NMDB at the proposed thresholds for smaller loans (3.5 or 6.5 percentage points above APOR). These commenters recommended that the Bureau make available for comment a revised version of General QM Proposal Table 5 that shows the historical early delinquency rates for first-lien purchase originations categorized by DTI and rate spreads greater than 2.25 percentage points above APOR, before it presumes ability to repay for consumers taking out loans with higher rate spreads.

Aside from noting issues with the Bureau's data, consumer advocate commenters also noted that the limited public data appears to suggest that smaller loans do not perform well, citing a newspaper article on manufactured housing loans, which described features unique to manufactured housing loans and reported that 28 percent of chattel loans fail to perform, as an example.^[298]

QM share of manufactured housing loans. A few industry commenters asserted that a substantial share of manufactured housing loans qualifying as General QMs under the current definition would fail to qualify as General QMs under the proposed thresholds. Some of these commenters surveyed their members to obtain information to estimate the decline in shares of manufactured housing loans that would meet the standards to be General QMs. For example, members of a national manufactured housing trade association stated that they expect up to 50 percent of their manufactured housing loans would lose General QM status under the proposed thresholds for smaller loans. Members of a trade group representing credit unions likewise stated that they expect up to 90 percent of their manufactured housing loans would lose General QM status. Other commenters used 2019 HMDA data to estimate the decline in shares of manufactured housing loans that would be eligible for General QM status. For instance, while comparing data from General QM Proposal Table 7 with 2019 HMDA data, a non-depository manufactured housing creditor asserted that, compared to first-lien manufactured housing loans, the Bureau's proposed thresholds would allow for far more first-lien conventional purchase loans for site-built housing to be eligible for General QM status.^[299]

To prevent a decline in the share of manufactured housing loans eligible for General QM status, commenters recommended the following adjustments or alternatives to the Bureau's proposed thresholds for smaller loans. One industry commenter recommended that the Bureau increase the pricing threshold for smaller loans but did not provide specific thresholds. Two other industry commenters recommended increasing the loan amount thresholds instead, from $65,939 to $110,000 and from $109,898 to $210,000. One of these commenters added that the Bureau should set these thresholds either for all loans or for only manufactured housing loans, while the other added that 91 percent of the first-lien manufactured housing loans originated in 2019 would have been eligible for General QM status if these higher loan amount thresholds were in place. One of these commenters also recommended a complementary DTI approach for manufactured housing loans. Under this approach, a manufactured housing loan would be eligible for General QM status by either satisfying the pricing thresholds or having a DTI ratio no higher than 45 percent, when determined in accordance with GSE or Federal agency underwriting guidelines. Lastly, a manufacturing housing creditor recommended incorporating HOEPA's APR thresholds for high-cost mortgages into a definition of General QM for manufactured housing loans. Specifically, the creditor recommended that a first-lien covered transaction secured by a manufactured home would have a conclusive presumption of compliance if the APR at consummation did not exceed the APOR by more than 1.5 percentage points; a rebuttable presumption of compliance if the APR at consummation did not exceed the APOR by 6.5 percentage points; and a rebuttable presumption of compliance if the transaction was a first-lien, personal property loan under $50,000 and the APR at consummation did not exceed the APOR by 8.5 percentage points. To underscore the importance of preventing an estimated decline in the share of manufactured housing loans that are General QMs, these commenters asserted that, without General QM status, creditors may either extend manufactured housing loans as more expensive non-QMs, or not extend these loans at all.^[300]

Consumer advocate commenters, however, asserted that creditors offering manufactured housing loans could adjust the price of these loans to fit within the Bureau's proposed thresholds, noting that creditors were able to price manufactured housing loans below HOEPA's APR thresholds for high-cost mortgages after those thresholds were adopted. Consumer advocate commenters also added that a high threshold would encourage exploitative lending right under the threshold.

QM share of subordinate-lien transactions. A few industry commenters noted that a sizable share of subordinate-lien transactions qualifying as General QMs under the current definition would fail to qualify as General QMs under the proposed thresholds.

To prevent the estimated decline in the share of subordinate-lien transactions that would obtain QM status under the proposed thresholds, one industry commenter recommended that the Bureau retain the current General QM loan definition for higher-priced mortgage loans, increase the pricing threshold for subordinate-lien transactions while using the same proposed loan amount thresholds used for first-lien transactions, or both. Under the commenter's second recommendation, a subordinate-lien transaction would qualify as a General QM if the APR at consummation does not exceed the APOR by 5 percentage points for transactions with a loan amount greater than or equal to $109,898; by 5.5 percentage points for transactions with a loan amount greater than or equal to $65,939 but less than $109,898; and by 8.5 percentage points for transactions with a loan amount less than $65,939. The commenter pointed to General QM Proposal Table 10 to demonstrate that delinquency rates did not materially differ under these recommended thresholds.^[301]

The Final Rule

The Bureau is adopting the proposed pricing thresholds for smaller loans and subordinate-lien transactions. However, as described below, the Bureau is finalizing an additional, higher pricing threshold for smaller loans secured by a manufactured home. In developing pricing thresholds under the General QM loan definition for smaller loans, smaller loans secured by a manufactured home, and subordinate-lien transactions, the Bureau balanced considerations related to ensuring consumers' ability to repay with maintaining access to responsible, affordable mortgage credit.^[302]

The final rule amends § 1026.43 by revising § 1026.43(e)(2)(vi) to provide higher pricing thresholds to define General QM for smaller loans, smaller loans secured by a manufactured home, and subordinate-lien transactions. The Bureau is also adjusting the loan amounts specified in § 1026.43(e)(2)(vi)(A) through (F). As discussed in the proposal, the Bureau proposed loan amount thresholds of $65,939 and $109,898, because those thresholds aligned with certain thresholds for the limits on points and fees, as updated for inflation, in § 1026.43(e)(3)(i) and the associated commentary.^[303] On August 19, 2020, the Bureau issued a final rule adjusting the loan amounts for the limits on points and fees under § 1026.43(e)(3)(i), based on the annual percentage change reflected in the CPI-U in effect on June 1, 2020.^[304] To ensure that the loan amounts for § 1026.43(e) remain synchronized, the Bureau is finalizing the loan amount thresholds specified in § 1026.43(e)(2)(vi)(A) through (F) with a threshold of $66,156, rather than a threshold of $65,939, and $110,260, rather than a threshold of $109,898. As clarified in comment 43(e)(2)(vi)-3, these amounts shall be adjusted annually on January 1 by the annual percentage change in the CPI-U that was reported on the preceding June 1.

Final § 1026.43(e)(2)(vi)(B) provides that, for first-lien covered transactions with loan amounts greater than or equal to $66,156 (indexed for inflation) but less than $110,260 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 3.5 or more percentage points. Section 1026.43(e)(2)(vi)(C) provides that, for first-lien covered transactions with loan amounts less than $66,156 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 6.5 or more percentage points. Section 1026.43(e)(2)(vi)(D) provides that, for first-lien covered transactions secured by a manufactured home with loan amounts less than $110,260 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 6.5 or more percentage points. Section 1026.43(e)(2)(vi)(E) provides that, for subordinate-lien covered transactions with loan amounts greater than or equal to $66,156 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 3.5 or more percentage points. Section 1026.43(e)(2)(vi)(F) provides that, for subordinate-lien covered transactions with loan amounts less than $66,156 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 6.5 or more percentage points.

The Bureau is also adding two comments to provide additional clarification on terms and phrases used in § 1026.43(e)(2)(vi)(D). Comment 43(e)(2)(vi)-5 clarifies that the term “manufactured home,” as used in § 1026.43(e)(2)(vi)(D), means any residential structure as defined under HUD regulations establishing manufactured home construction and safety standards (24 CFR 3280.2). Modular or other factory-built homes that do not meet the HUD code standards are not manufactured homes for purposes of § 1026.43(e)(2)(vi)(D). The Bureau is aligning the definition of “manufactured home” with the HUD standards to maintain consistency with the definition the Bureau uses elsewhere in Regulation Z.^[305] Comment 43(e)(2)(vi)-6 provides that the threshold in § 1026.43(e)(2)(vi)(D) applies to first-lien covered transactions less than $110,260 (indexed for inflation) that are secured by a manufactured home and land, or by a manufactured home only.

Smaller loans. The Bureau is adopting higher thresholds for smaller loans because it is concerned that loans with smaller loan amounts are typically priced higher than loans with larger loan amounts, even though a consumer with a smaller loan may have similar credit characteristics and likelihood of early delinquency, which the Bureau uses as a proxy for measuring whether a consumer had a reasonable ability to repay at the time the loan was consummated. As discussed in the General QM Proposal—and noted by commenters supporting the proposed higher thresholds for smaller loans—many of the creditors' costs for a transaction may be the same or similar between smaller loans and larger loans. For creditors to recover their costs for originating and servicing smaller loans, they may have to charge higher interest rates or higher points and fees as a percentage of the loan amount than they would for comparable larger loans. As a result, smaller loans tend to have higher APRs than larger loans to consumers with similar credit characteristics and who may have a similar ability to repay. The Bureau concludes that its observation of the components of creditors' costs, in this limited regard, is consistent with its statutory obligations. As stated above, TILA section 129C(b)(3)(B)(i) authorizes the Bureau to prescribe regulations that revise, add to, or subtract from the criteria that define a QM upon a finding that those regulations are necessary or proper to ensure that responsible, affordable credit remains available to consumers in a manner consistent with the purposes of TILA section 129C. Here, as further explained below, the Bureau's analysis indicates that consumers who take out smaller loans with APRs within higher thresholds may have similar credit characteristics as consumers who take out larger loans. The Bureau's analysis also indicates that smaller loans with APRs within higher thresholds may have comparable levels of early delinquencies as larger loans within lower thresholds. However, as explained further below, the Bureau's analysis of delinquency levels for smaller loans, compared to larger loans, does not appear to indicate a threshold at which delinquency levels significantly accelerate. Nevertheless, the Bureau concludes that the finalized thresholds for smaller loans best ensure that responsible, affordable credit remains available to consumers taking out smaller loans, while also helping to ensure that the risks are limited. The Bureau thus concludes that smaller loans that are higher-priced loans under § 1026.43(b)(4) but are priced below the applicable thresholds in § 1026.43(e)(2)(vi)(B) or (C) will receive a rebuttable presumption of compliance with the ATR requirements.

Moreover, adopting the same threshold of 2.25 percentage points above APOR for all loans could disproportionately prevent smaller loans with comparable levels of early delinquencies as larger loans, potentially including a disproportionate number of loans to minority consumers, from being originated as General QMs. The Bureau's analysis of 2018 HMDA data found that 3.7 percent of site-built loans to minority consumers are priced 2.25 percentage points or more over APOR, but 2.7 percent of site-built loans to non-Hispanic White consumers are priced 2.25 percentage points or more over APOR. While some loans may be originated under other QM definitions or as non-QM loans, those loans may cost materially more to consumers, and some loans may not be originated at all. As discussed in part V, the non-QM market has been slow to develop, and the negative impact on the non-QM market from the disruptions caused by the COVID-19 pandemic raises further concerns about the capacity of the non-QM market to provide consumers with access to credit through such loans.

The Bureau also notes that, in the Dodd-Frank Act, Congress provided for additional pricing flexibility for creditors making smaller loans, allowing smaller loans to include higher points and fees while still meeting the QM definition. TILA section 129C(b)(2)(A)(vi) defines a QM as a loan for which, among other things, the total points and fees payable in connection with the loan do not exceed 3 percent of the total loan amount. However, TILA section 129C(b)(2)(D) requires the Bureau to prescribe rules adjusting the points-and-fees limits for smaller loans. In the January 2013 Final Rule, the Bureau implemented this requirement in § 1026.43(e)(3), adopting higher points-and-fees thresholds for different tiers of loan amounts less than or equal to $100,000, adjusted for inflation.^[306] The Bureau's conclusion that creditors originating smaller loans typically impose higher points and fees or higher interest rates to recover their costs, regardless of the consumer's creditworthiness, and that higher thresholds for smaller loans in § 1026.43(e)(2)(vi) are therefore warranted, is generally consistent with the statutory directive to adopt higher points-and-fees thresholds for smaller loans.

To develop the thresholds for smaller loans in § 1026.43(e)(2)(vi)(B) and (C), the Bureau analyzed evidence related to credit characteristics and loan performance for first-lien purchase transactions at various rate spreads and loan amounts (adjusted for inflation) using HMDA and NMDB data, as shown in Table 9.^[307] To ensure a sufficient sample size was available for a reliable analysis, the Bureau used cumulative rate-spread ranges.

Table 9—Loan Characteristics and Performance for Different Sizes of First-Lien Transactions at Various Rate Spreads

Loan size group	Rate spread range (percentage points over APOR)	Mean CLTV, 2018 HMDA	Mean DTI, 2018 HMDA	Mean credit score, 2018 HMDA	Percent observed 60+ days delinquent within first 2 years, 2002-2008 NMDB (%)	Percent observed 60+ days delinquent within first 2 years, 2018 NMDB (%)
Under $66,156	1.5-2.0	81.9	32.3	717	6.1	2.8
Under $66,156	1.5-2.5	82.2	32.3	714	6.1	2.3
Under $66,156	1.5-3.0	82.1	32.2	714	6.2	2.3
Under $66,156	1.5-3.5	81.9	32.1	715	6.2	2.5
Under $66,156	1.5-4.0	81.7	32.3	714	6.3	2.5
Under $66,156	1.5-4.5	81.7	32.5	710	6.4	2.6
Under $66,156	1.5-5.0	81.7	32.6	706	6.4	2.5
Under $66,156	1.5-5.5	81.6	32.7	699	6.5	2.4
Under $66,156	1.5-6.0	81.7	32.9	694	6.5	2.5
Under $66,156	1.5-6.5	81.9	33.1	685	6.5	3.4
Under $66,156	1.5 and above	82.0	33.3	676	6.6	4.1
$66,156 to $110,259	1.5-2.0	89.9	35.5	704	11.1	3.4
$66,156 to $110,259	1.5-2.5	90.1	35.4	702	12.2	4.2
$66,156 to $110,259	1.5-3.0	90.0	35.5	702	12.9	4.2
$66,156 to $110,259	1.5-3.5	89.7	35.5	703	13.0	4.3
$66,156 to $110,259	1.5-4.0	89.4	35.6	703	13.1	4.0
$66,156 to $110,259	1.5-4.5	89.3	35.7	701	13.2	4.2
$66,156 to $110,259	1.5-5.0	89.1	35.8	699	13.3	4.1
$66,156 to $110,259	1.5-5.5	89.1	35.9	696	13.4	4.0
$66,156 to $110,259	1.5-6.0	89.2	36.0	692	13.4	4.2
$66,156 to $110,259	1.5-6.5	89.3	36.1	684	13.4	4.5
$66,156 to $110,259	1.5 and above	89.3	36.1	684	13.7	4.5
$110,260 and above, manufactured and site-built housing	1.5-2.25 (for comparison)	92.4	39.3	698	15.6	2.7

The Bureau's analysis indicates that consumers with smaller loans with APRs within higher thresholds, such as 6.5 or 3.5 percentage points above APOR, have similar credit characteristics as consumers with larger loans with APRs between 1.5 and 2.25 percentage points above APOR.^[308]

More specifically, the Bureau analyzed 2018 HMDA data on first-lien conventional purchase loans and found that loans less than $66,156 that are priced between 1.5 and 6.5 percentage points above APOR have a mean DTI ratio of 33.1 percent, a mean combined LTV ratio of 81.9 percent, and a mean credit score of 685. Loans greater than or equal to $66,156 but less than $110,260 that are priced between 1.5 and 3.5 percentage points above APOR have a mean DTI ratio of 35.5 percent, a mean combined LTV of 89.7 percent, and a mean credit score of 703. Loans greater than or equal to $110,260 that are priced between 1.5 and 2.25 percentage points above APOR have a mean DTI ratio of 39.3 percent, a mean combined LTV of 92.4 percent, and a mean credit score of 698. These data comparisons all suggest that the credit characteristics, and potentially the ability to repay, of consumers taking out smaller loans with higher APRs, may be at least comparable to those of consumers taking out larger loans with lower APRs.

With respect to early delinquencies, the evidence summarized in Table 9 generally provides support for higher thresholds for smaller loans. Loans less than $66,156 had lower delinquency rates than loans greater than or equal to $66,156 but less than $110,260 across all rate spread ranges and generally had delinquency rates lower than larger loans (greater than or equal to $110,260) priced between 1.5 and 2.25 percentage points above APOR, except as described below. Loans greater than or equal to $66,156 but less than $110,260 had lower delinquency rates than larger loans between 2002 and 2008, but higher delinquency rates in 2018.

More specifically, the Bureau analyzed NMDB data from 2002 through 2008 on first-lien conventional purchase loans and found that loans less than $66,156 that were priced between 1.5 and 6.5 percentage points above APOR had an early delinquency rate of 6.5 percent. Loans greater than or equal to $66,156 but less than $110,260 that were priced between 1.5 and 3.5 percentage points above APOR had an early delinquency rate of 13 percent. Loans greater than or equal to $110,260 that were priced between 1.5 and 2.25 percentage points above APOR had an early delinquency rate of 15.6 percent. These rates suggest that the historical loan performance of smaller loans with higher APRs may be comparable, if not better, than larger loans with lower APRs.

However, the Bureau's analysis found that early delinquency rates for 2018 loans are somewhat higher for smaller loans with higher APRs than larger loans with lower APRs. More specifically, NMDB data from 2018 on first-lien conventional purchase loans indicates that loans less than $66,156 that were priced between 1.5 and 6.5 percentage points above APOR had an early delinquency rate of 3.4 percent and those that were priced 1.5 percentage points over APOR and above had an early delinquency rate of 4.1 percent. Loans greater than or equal to $66,156 but less than $110,260 that were priced between 1.5 and 3.5 percentage points above APOR had an early delinquency rate of 4.3 percent. Loans greater than or equal to $110,260 that were priced between 1.5 and 2.25 percentage points above APOR had an early delinquency rate of 2.7 percent.

Although the data in the rulemaking record do not appear to indicate a particular threshold at which the credit characteristics or loan performance for smaller loans with higher APRs decline significantly, the Bureau concludes that the thresholds in § 1026.43(e)(2)(vi)(B) and (C) for smaller, first-lien covered transactions strike the best balance between ensuring consumers' ability to repay and ensuring access to responsible, affordable mortgage credit.

As described in more detail above, consumer advocate commenters recommended that the Bureau further refine the data before concluding that smaller loans with APRs within higher thresholds have similar credit characteristics and comparable levels of early delinquencies as larger loans. The commenters based their recommendation on specific concerns, including: (1) The absence of loan volume data and the use of cumulative rate-spread ranges, instead of incremental rate-spread ranges, in General QM Proposal Table 9; and (2) the absence of an analysis of chattel loans, separate from that of real-estate secured mortgages. The Bureau understands these concerns to suggest three issues: (1) That without loan volume data, it was not clear if there was a sufficient sample size for a reliable analysis; (2) that cumulative rate-spread ranges resulted in a skewed analysis of the early delinquency rates for smaller loans at or near the threshold; and (3) that differences between chattel loans and real-estate secured mortgages, with respect to pricing and performance, were not adequately considered.

However, the Bureau took all these issues into account when using HMDA and NMDB data to analyze the evidence related to the credit characteristics and loan performance of first-lien purchase transactions at various rate spread and loan amounts. As explained in the General QM Proposal, the Bureau grouped loans at higher rate spreads when a sufficient number of observations did not exist in the data for a reliable analysis. For example, the Bureau grouped loans with rate spreads of 2.25 percentage points or more to ensure a sufficient sample size for a reliable analysis of the 2002-2008 data in Tables 1 and 5 of the General QM Proposal.^[309] This grouping ensured that all cells shown in these tables contained at least 500 loans. For similar reasons, the Bureau grouped loans in General QM Proposal Table 9 (and Table 9 above).^[310] The Bureau determined that it was necessary to use a cumulative rate-spread range to ensure a sufficient sample size for a reliable analysis of 2018 NMDB data for higher-priced, smaller loans. More specifically, by grouping first-lien loans less than $65,939 ($66,156, when adjusted for inflation), priced between 1.5 and 6.5 percentage points above APOR, the Bureau was able to analyze the performance of 677 loans from 2018 NMDB data compared to only 87 loans if the Bureau looked at first-lien loans less than $65,939 that were priced between 6 and 6.5 percentage points above APOR.

Moreover, an analysis using incremental rate-spread ranges would have also supported higher thresholds for smaller loans. When using only 2002-2008 NMDB data, because of limitations in 2018 NMDB data, loans less than $66,156 and loans greater than or equal to $66,156 but less than $110,260 that were priced at or a half percentage point below the threshold had lower delinquency rates than larger loans (greater than or equal to $110,260) priced between 1.5 and 2.25 percentage points above APOR.

Specifically, loans less than $66,156 that were priced between 6 and 6.5 percentage points above APOR had an early delinquency rate of 7.7 percent. Loans greater than or equal to $66,156 but less than $110,260 that were priced between 3 and 3.5 percentage points above APOR had an early delinquency rate of 13.9 percent. Loans greater than or equal to $110,260 that were priced between 1.5 and 2.25 percentage points above APOR had an early delinquency rate of 15.6 percent. These early delinquency rates suggest that even under an approach using incremental rate-spread ranges, the historical performance of smaller loans with higher APRs remained comparable, if not better, than larger loans with lower APRs.

Some commenters recommended analyzing chattel loans separately from real-estate secured mortgages because of potential differences between the two with respect to pricing and performance. Consumer advocate commenters cited a newspaper article suggesting that chattel loans may not perform well. However, the Bureau is not aware of any data that sufficiently address how pricing at various thresholds correlates with performance or demonstrate how pricing varies with the performance of chattel loans relative to real-estate secured mortgages. Further, the Bureau's own data are not sufficient to separately analyze chattel loans from real-estate secured mortgages at various pricing thresholds. The Bureau's merged historical HMDA and NMDB data do not have reliable indicators for chattel loans. And although 2018 HMDA and NMDB data do have more reliable indicators, there are too few loans in 2018 data to reliably distinguish performance across different rate spread or loan size groupings. Accordingly, the Bureau lacks a reasoned basis for setting a different pricing threshold for chattel loans relative to real-estate secured mortgages, particularly given the access-to-credit concerns and other concerns described below. The Bureau will, however, continue to monitor the market and, if additional data become available and indicate that an adjustment to the thresholds for smaller loans and smaller manufactured housing loans is warranted, the Bureau will consider making an adjustment.

Lastly, as described above, some consumer advocate commenters suggested that land installment contracts would be newly eligible for General QM status under this final rule. The commenters, however, did not provide the Bureau with evidence or data indicating that land installment contracts that were previously ineligible for General QM status would become eligible for General QM status under the amended General QM loan definition in this final rule. As described above, the Bureau anticipates the price-based approach in this final rule will change the share of covered transactions that would be eligible for General QM status. Specifically, loans with DTI ratios over 43 percent priced under the thresholds will be eligible for General QM status, and loans with DTI ratios under 43 percent but priced over the thresholds will not be eligible for General QM status. However, the Bureau does not have data or other evidence indicating that the final rule will change the scope of transactions covered by the Rule so that certain land installment contracts will now be eligible for General QM status.

Smaller manufactured housing loans. As discussed above, commenters asserted that a substantial share of manufactured housing loans that qualify as General QMs under the current definition would fail to qualify under the proposed pricing thresholds. These commenters confirmed the Bureau's concerns, as discussed in the General QM Proposal, regarding the impact a price-based General QM definition, without higher thresholds, would have on the availability of responsible, affordable mortgage credit for manufactured homes. Specifically, the commenters confirmed the Bureau's concern that manufactured housing loans with smaller loan amounts are typically priced higher than loans with larger loan amounts, even though a consumer with a smaller manufactured housing loan may have similar ability to repay; and that while some smaller manufactured housing loans may be originated under other QM definitions or as non-QM loans, those loans may cost materially more to consumers, and some may not be originated at all. The Bureau also analyzed 2018 HMDA data to confirm its concerns on the potential effects on access to credit of a price-based approach to defining a General QM. The Bureau's analysis found that 55 percent of manufactured housing loans are priced 2.25 percentage points or more above APOR. Moreover, as indicated by the various combinations in Table 10 below,^[311] the Bureau estimates, based on 2018 HMDA data, that under the current rule—including the Temporary GSE QM loan definition, the General QM loan definition with a 43 percent DTI limit, and the Small Creditor QM loan definition in § 1026.43(e)(5)—83.6 percent of first-lien covered transactions secured by a manufactured home were General QMs. However, under the proposed General QM thresholds for larger loans and smaller loans, the Bureau estimates that 72.3 percent of first-lien covered transactions secured by a manufactured home would have been General QMs.

Table 10—Share of 2018 Manufactured Housing Conventional First-Lien Purchase Transactions Within Various QM Definitions

[HMDA data]

Approach	QM (share of manufactured housing loans)
Temporary GSE QM + DTI 43	83.6
Proposal	72.3
Final rule with small, manufactured housing loan pricing at 6.5	84.6

In view of commenter confirmation of the Bureau's concerns regarding the potential effects of the proposal on the availability of responsible, affordable mortgage credit for manufactured homes, the Bureau has reconsidered whether the proposed thresholds for smaller loans strike the best balance between ensuring consumers' repayment ability and maintaining access to responsible, affordable mortgage credit for manufactured homes. Specifically, the Bureau concludes that it achieves a better balance of these competing considerations by expanding the proposed rebuttable presumption of compliance with the ATR requirements to loans for manufactured housing less than $110,260 that are higher-priced loans under § 1026.43(b)(4) but are priced below the threshold in § 1026.43(e)(2)(vi)(D). In so concluding, the Bureau acknowledges that Table 9 suggests a higher risk of early delinquency among first-lien covered transactions secured by a manufactured home priced equal to or greater than $66,156. But the Bureau concludes that the degree of risk is acceptable in view of a potentially significant reduction of access to such mortgage credit and the fact that consumers obtaining such loans will retain the opportunity to rebut the presumption of compliance by showing that the creditor in fact lacked a good faith and reasonable belief in the consumer's reasonable ability to repay the loan.

Section 1026.43(e)(2)(vi)(D) as finalized thus provides that, for first-lien covered transactions secured by a manufactured home with a loan amount less than $110,260 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 6.5 or more percentage points. Smaller loans secured by a manufactured home and priced at or above the 6.5-percentage-point threshold are not eligible for QM status under § 1026.43(e)(2).^[312] Under the final rule with this threshold, the Bureau estimates that, based on 2018 HMDA data, 84.6 percent of first-lien covered transactions secured by a manufactured home would have been General QMs. This is consistent with the share of first-lien covered transactions secured by a manufactured home that were QMs under the current rule, which includes the Temporary GSE QM loan definition, the General QM loan definition with a 43 percent DTI limit, and the Small Creditor QM loan definition in § 1026.43(e)(5).

The access-to-credit concerns described above are sufficient by themselves to support the Bureau's decision to adopt a higher pricing threshold for smaller manufactured housing loans. This threshold also is independently supported by the credit characteristics of consumers with these loans. Specifically, the Bureau considered 2018 HMDA data to assess whether consumers who take out smaller manufactured housing loans with higher APRs have similar credit characteristics, and thus similar ability to repay, as consumers who take out larger loans with lower APRs. The Bureau would have also considered whether the consumer was ever 60 or more days past due within the first 2 years after origination, i.e., the early delinquency rate. However, as described above, the Bureau does not have sufficient loan performance data on manufactured housing loans for a reliable analysis of whether consumers who take out these smaller manufactured housing loans had early difficulties in making payments. Accordingly, the Bureau limited its ability-to-repay analysis to the credit characteristics of consumers taking out smaller manufactured housing loans with APRs within higher thresholds, as shown in Table 11.

Table 11—Loan Characteristics for Different Sizes of Manufactured Housing First-Lien Transactions at Various Rate Spreads

Loan size group	Rate spread range (percentage points over APOR)	Mean CLTV, 2018 HMDA	Mean DTI, 2018 HMDA	Mean credit score, 2018 HMDA
Under $66,156	1.5-2.0	74.2	31.8	733
Under $66,156	1.5-2.5	73.7	31.2	735
Under $66,156	1.5-3.0	74.6	31.5	737
Under $66,156	1.5-3.5	75.6	31.6	734
Under $66,156	1.5-4.0	76.3	32.1	728
Under $66,156	1.5-4.5	77.4	32.7	717
Under $66,156	1.5-5.0	77.8	32.8	709
Under $66,156	1.5-5.5	78.1	33.0	697
Under $66,156	1.5-6.0	78.6	33.2	689
Under $66,156	1.5-6.5	79.4	33.6	676
Under $66,156	1.5 and above	80.1	33.6	665
$66,156 to $110,259	1.5-2.0	85.4	23.3	732
$66,156 to $110,259	1.5-2.5	85.2	34.2	735
$66,156 to $110,259	1.5-3.0	85.5	34.6	731
$66,156 to $110,259	1.5-3.5	85.8	35.0	728
$66,156 to $110,259	1.5-4.0	85.9	35.5	723
$66,156 to $110,259	1.5-4.5	86.1	35.9	715
$66,156 to $110,259	1.5-5.0	86.5	36.1	707
$66,156 to $110,259	1.5-5.5	86.8	36.3	699
$66,156 to $110,259	1.5-6.0	87.6	36.5	690
$66,156 to $110,259	1.5-6.5	88.2	36.6	677
$66,156 to $110,259	1.5 and above	88.2	36.7	676
$110,260 and above, manufactured and site-built housing	1.5-2.25 (for comparison)	92.4	39.3	698

The Bureau's analysis indicates that consumers with smaller manufactured housing loans with APRs up to 6.5 percentage points above APOR have credit characteristics that are comparable to, if not better than, consumers with larger loans priced between 1.5 and 2.25 percentage points above APOR. More specifically, the Bureau found that smaller manufactured housing loans less than $66,156 that are priced between 1.5 and 6.5 percentage points above APOR have a mean DTI ratio of 33.6 percent, a mean combined LTV ratio of 79.4 percent, and a mean credit score of 676. Smaller manufactured housing loans greater than or equal to $66,156 but less than $110,260 that are priced between 1.5 and 6.5 percentage points above APOR have a mean DTI ratio of 36.6 percent, a mean combined LTV ratio of 88.2 percent, and a mean credit score of 677. Loans greater than or equal to $110,260 that are priced between 1.5 and 2.25 percentage points above APOR have a mean DTI ratio of 39.3 percent, a mean combined LTV ratio of 92.4 percent, and a mean credit score of 698. These all suggest that the credit characteristics of consumers taking out smaller manufactured housing loans with higher APRs appear to be at least comparable to, if not better than, those of consumers taking out larger loans with lower APRs. This suggests that consumers taking out smaller manufactured housing loans with higher APRs may have an ability to repay these loans at least comparable to the consumers who take out larger loans with lower APRs.

Although the current data appear to indicate some thresholds at which certain credit characteristics, in particular credit score, decline for smaller manufactured housing loans with higher APRs, the Bureau concludes that the adopted threshold in § 1026.43(e)(2)(vi)(D) for smaller, first-lien covered transactions secured by a manufactured home strikes the best balance between ensuring consumers' ability to repay and ensuring access to responsible, affordable mortgage credit for manufactured homes.

The Bureau is also adding two comments to provide additional clarification on the pricing threshold for smaller loans secured by a manufactured home. Comment 43(e)(2)(vi)-5 clarifies that the term “manufactured home,” as used in § 1026.43(e)(2)(vi)(D), means any residential structure as defined under HUD regulations establishing manufactured home construction and safety standards (24 CFR 3280.2). Modular or other factory-built homes that do not meet the HUD code standards are not manufactured homes for purposes of § 1026.43(e)(2)(vi)(D). Comment 43(e)(2)(vi)-6 provides that the threshold in § 1026.43(e)(2)(vi)(D) applies to first-lien covered transactions less than $110,260 (indexed for inflation) that are secured by a manufactured home and land, or by a manufactured home only.

The Bureau is aware that whether a manufactured home is titled as personal property or as real property factors into the cost of the loan and that the price may be relatively higher for a loan in which the manufactured home is titled as personal property (i.e., a chattel loan).^[313] However, the Bureau is not adopting a higher threshold for only smaller chattel loans. Doing so would incentivize manufactured home creditors to encourage consumers to title their manufactured homes as personal property to originate a QM-eligible loan. Generally, titling manufactured homes as personal property may have disadvantages for consumers because chattel loans tend to be more expensive,^[314] and have fewer consumer protections.^[315] Moreover, as explained above, the Bureau does not have sufficient performance data to analyze how chattel loans perform relative to real estate-secured mortgages at various pricing thresholds. Without this data and given the risks for consumers' titling their manufactured homes as personal property, the Bureau has decided to adopt a higher pricing threshold for smaller loans secured by either a manufactured home and land, or by a manufactured home only.

Moreover, the Bureau understands that creditors may either increase or decrease the price of these loans to just below the adopted threshold. To the extent creditors reduce the price of the loan, this would result in more affordable prices; for example, some consumers whose loans would have otherwise been priced above the threshold may now be eligible for loans below the threshold. These loans would also be subject to the QM prohibitions on certain loan features and limits on points and fees, which would provide protections for consumers. However, this development could also lead to an increase in the number of consumers with delinquent loans who would have to rebut the creditor's presumption of compliance to benefit from an ability-to-repay cause of action or defense against foreclosure. Regardless, the Bureau does not have sufficient data to determine whether these developments would occur and the impact these developments would have on the benefits and costs to consumers. However, as described above, the Bureau intends to monitor the market for additional data that might indicate the need for the Bureau to consider a future adjustment.

A few commenters recommended alternatives other than the one adopted here to address the access-to-credit concern for manufactured homes. However, the Bureau concludes that adopting a higher pricing threshold for smaller loans secured by a manufactured home addresses the access-to-credit concerns better than the recommended alternatives. The first recommendation to increase the dollar thresholds defining “smaller loans,” would result in a definition that is inconsistent with the meaning of “smaller loans” in the small loan exception to the QM points and fees cap, which could potentially lead to certain compliance challenges. The other recommendation to incorporate HOEPA's APR thresholds into the General QM loan definition does not properly acknowledge HOEPA's statutory objective, which was to identify transactions requiring creditors to provide additional disclosures and prohibiting creditors from engaging in certain practices. The Bureau does not believe that it should implement thresholds designed for those discrete uses here, in determining whether the transaction should be eligible for a rebuttable presumption of compliance with the ATR requirements. Lastly, the Bureau declines to adopt a complementary DTI alternative for manufactured housing loans. A complementary DTI alternative would be unduly complex and not necessary given that the Bureau expects the final pricing threshold to improve access to credit for manufactured homes. Moreover, the Bureau believes that a loan's price, as measured by comparing a loan's APR to APOR for a comparable transaction, is a strong indicator of a consumer's ability to repay and is a more holistic and flexible measure of a consumer's ability to repay than DTI alone. For these reasons, the Bureau concludes that adopting a higher pricing threshold for smaller loans secured by a manufactured home strikes a better balance between ensuring consumers' ability to repay and ensuring access to responsible, affordable mortgage credit for manufactured homes.

Subordinate-lien transactions. The Bureau is adopting higher thresholds in § 1026.43(e)(2)(vi)(E) and (F) for subordinate-lien transactions because subordinate-lien transactions may be priced higher than comparable first-lien transactions for reasons other than consumers' ability to repay. In general, the creditor of a subordinate lien will recover its principal, in the event of default and foreclosure, only to the extent funds remain after the first-lien creditor recovers its principal. Thus, to compensate for this risk, creditors typically price subordinate-lien transactions higher than first-lien transactions, even though the consumer in the subordinate-lien transaction may have similar credit characteristics and ability to repay. In addition, subordinate-lien transactions are often for smaller loan amounts, so the pricing factors discussed above for smaller loan amounts may further increase the price of subordinate-lien transactions, regardless of the consumer's ability to repay. To the extent the higher pricing for a subordinate-lien transaction is not related to consumers' ability to repay, applying the same pricing to them as first-lien transactions results in them being excluded from QM status under § 1026.43(e)(2).

In the January 2013 Final Rule, the Bureau adopted higher thresholds for determining if subordinate-lien QMs received a rebuttable presumption or a conclusive presumption of compliance with the ATR requirements.^[316] For subordinate-lien transactions, the definition of “higher-priced covered transaction” in § 1026.43(b)(4) is used in § 1026.43(e)(1) to set a threshold of 3.5 percentage points above APOR to determine which subordinate-lien QMs receive a safe harbor and which receive a rebuttable presumption of compliance. As discussed above in part V, the Bureau is not proposing to alter the threshold for subordinate-lien transactions in § 1026.43(b)(4). To avoid the odd result that a subordinate-lien transaction would otherwise be eligible to receive a safe harbor under § 1026.43(b)(4) and (e)(1) but would not be eligible for QM status under § 1026.43(e)(2)(vi), the Bureau considered which threshold or thresholds at or above 3.5 percentage points above APOR to propose for subordinate-lien transactions in § 1026.43(e)(2)(vi).

To develop the thresholds for subordinate-lien transactions in § 1026.43(e)(2)(vi)(E) and (F), the Bureau considered evidence related to credit characteristics and loan performance for subordinate-lien transactions at various rate spreads and loan amounts (adjusted for inflation) using HMDA and Y-14M data, as shown in Table 12.^[317] To ensure a sufficient sample size was available for a reliable analysis, the Bureau used cumulative rate-spread ranges.^[318]

Table 12—Loan Characteristics and Performance for Different Sizes of Subordinate-Lien Transactions at Various Rate Spreads

Loan size group	Rate spread range (percentage points over APOR)	Mean CLTV, 2018 HMDA	Mean DTI, 2018 HMDA	Mean credit score, 2018 HMDA	Percent observed 90+ days delinquent within first 2 years, 2013-2016 Y-14M data (subset) (%)
Under $66,156	2.0-2.5	76.9	36.1	728	2.1
Under $66,156	2.0-3.0	78.4	36.5	724	1.6
Under $66,156	2.0-3.5	79.7	36.8	721	1.4
Under $66,156	2.0-4.0	80.1	36.9	720	1.4
Under $66,156	2.0-4.5	80.2	36.9	719	1.3
Under $66,156	2.0-5.0	80.3	37.0	718	1.3
Under $66,156	2.0-5.5	80.3	37.1	718	1.3
Under $66,156	2.0-6.0	80.3	37.1	717	1.3
Under $66,156	2.0-6.5	80.4	37.2	717	1.3
Under $66,156	2.0 and above	80.7	37.3	715	1.4
$66,156 and above	2.0-2.5	79.5	37.2	738	1.9
$66,156 and above	2.0-3.0	80.5	37.3	735	1.7
$66,156 and above	2.0-3.5	81.0	37.4	732	1.6
$66,156 and above	2.0-4.0	81.3	37.5	732	1.7
$66,156 and above	2.0-4.5	81.3	37.6	731	1.7
$66,156 and above	2.0-5.0	81.5	37.7	731	1.8
$66,156 and above	2.0-5.5	81.6	37.7	730	1.8
$66,156 and above	2.0-6.0	81.6	37.8	729	1.8
$66,156 and above	2.0-6.5	81.7	37.9	729	1.8
$66,156 and above	2.0 and above	81.8	37.9	728	1.9

In general, the Bureau's analysis found strong credit characteristics and loan performance for subordinate-lien transactions at various thresholds greater than 2 percentage points above APOR. The current data do not appear to indicate a particular threshold at which the credit characteristics or loan performance decline significantly.

With respect to larger subordinate-lien transactions, the Bureau's analysis of 2018 HMDA data on subordinate-lien conventional loans found that, for consumers with subordinate-lien transactions greater than or equal to $66,156 that were priced up to 2 to 3.5 percentage points above APOR, the mean DTI ratio was 37.4 percent, the mean combined LTV was 81 percent, and the mean credit score was 732. The Bureau also analyzed Y-14M loan data for 2013 to 2016 and estimated that subordinate-lien transactions greater than or equal to $66,156 that were priced up to 2 to 3.5 percentage points above APOR had an early delinquency rate of approximately 1.6 percent.^[319] These factors appear to provide a strong indication of ability to repay, so the Bureau has decided to set the threshold at 3.5 percentage points above APOR for larger subordinate-lien transactions (greater than or equal to $66,156) to be eligible for QM status under § 1026.43(e)(2).

The Bureau recognizes that, because the price-based approach would leave the threshold in § 1026.43(b)(4) for higher-priced QMs at 3.5 percentage points above APOR for subordinate-lien transactions (and that such transactions that are not higher priced would, therefore, receive a safe harbor under § 1026.43(e)(1)(i)), this approach would result in subordinate-lien transactions for amounts over $66,156 either being a safe harbor QM or not being eligible for QM status under § 1026.43(e)(2). No such loans would be eligible to be a rebuttable presumption QM. Nevertheless, the Bureau concludes that the threshold best balances the relatively strong credit characteristics and loan performance of these transactions historically, which is indicative of ability to repay, against the concern that the supporting data are limited to recent years with strong economic performance and conservative underwriting.

For smaller subordinate-lien transactions, the Bureau's analysis of 2018 HMDA data on subordinate-lien conventional loans found that for consumers with subordinate-lien transactions less than $66,156 that were priced between 2 and 6.5 percentage points above APOR, the mean DTI ratio was 37.2 percent, the mean combined LTV was 80.4 percent, and the mean credit score was 717. The Bureau also analyzed Y-14M loan data for 2013 to 2016 and estimated that subordinate-lien transactions less than $66,156 that were priced between 2 and 6.5 percentage points above APOR, the early delinquency rate was approximately 1.3 percent. Based on these relatively strong credit characteristics and low delinquency rates, the Bureau has decided to set the threshold at 6.5 percentage points above APOR for subordinate-lien transactions less than $66,156 to be eligible for QM status under § 1026.43(e)(2). The Bureau notes that under this approach, these transactions would be eligible only for a rebuttable presumption of compliance under § 1026.43(e)(1)(ii) when higher-priced under § 1026.43(b)(4), and that consumers, therefore, would have the opportunity to rebut the presumption under § 1026.43(e)(1)(ii)(B).

Some subordinate-lien transactions currently meeting the General QM loan definition may fail to do so under the adopted thresholds. However, based on 2018 HMDA data, the Bureau estimates that the adopted thresholds will increase the overall share of subordinate-lien transactions that are eligible for QM status. Accordingly, the Bureau concludes that its approach strikes the best balance between ensuring consumers' ability to repay and access to responsible, affordable credit for subordinate-lien transactions.

Determining the APR for Certain Loans for Which the Interest Rate May or Will Change

The Bureau's Proposal

The Bureau also proposed to revise § 1026.43(e)(2)(vi) to include a special rule for determining the APR for certain types of loans for purposes of whether a loan meets the General QM loan definition under § 1026.43(e)(2). This proposed special rule would have applied to loans for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due. For such loans, for purposes of determining whether the loan is a General QM under § 1026.43(e)(2)(vi), the creditor would have been required under the proposal to determine the APR by treating the maximum interest rate that may apply during that five-year period as the interest rate for the full term of the loan.^[320] The proposed special rule would have applied principally to ARMs with initial fixed-rate periods of five years or less (referred to in the proposal as “short-reset ARMs”) but also would have applied to step-rate mortgages ^[321] that have an initial period of five years or less. The special rule in the proposed revisions to § 1026.43(e)(2)(vi) would not have modified other provisions in Regulation Z for determining the APR for other purposes, such as the disclosures addressed in or subject to the commentary to § 1026.17(c)(1).

In the proposed rule, the Bureau said that it anticipated that the proposed price-based approach to defining General QMs would in general be effective in identifying which loans consumers have the ability to repay and should therefore be eligible for QM status under § 1026.43(e)(2). However, the Bureau recognized that, absent the special rule, the proposed price-based approach may less effectively capture specific unaffordability risks of certain loans for which the interest rate may or will change relatively soon after consummation. Therefore, the Bureau stated that, for loans for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due, a modified approach to determining the APR for purposes of the rate-spread thresholds under proposed § 1026.43(e)(2) may be warranted.

Proposed comment 43(e)(2)(vi)-4.i stated that provisions in subpart C, including the existing commentary to § 1026.17(c)(1), address the determination of the APR disclosures for closed-end credit transactions and that provisions in § 1026.32(a)(3) address how to determine the APR to determine coverage under § 1026.32(a)(1)(i). It further stated that proposed § 1026.43(e)(2)(vi) requires, for the purposes of that paragraph, a different determination of the APR for a QM under proposed § 1026.43(e)(2) for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due. In addition, proposed comment 43(e)(2)(vi)-4.i stated that an identical special rule for determining the APR for such a loan also applies for purposes of proposed § 1026.43(b)(4).

The Bureau proposed comment 43(e)(2)(vi)-4.ii to explain the application of the special rule in proposed § 1026.43(e)(2)(vi) for determining the APR for a loan for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due. Specifically, it stated that the special rule applies to ARMs that have a fixed-rate period of five years or less and to step-rate mortgages for which the interest rate changes within that five-year period.

Proposed comment 43(e)(2)(vi)-4.iii provided that, to determine the APR for purposes of proposed § 1026.43(e)(2)(vi), a creditor must treat the maximum interest rate that could apply at any time during the five-year period after the date on which the first regular periodic payment will be due as the interest rate for the full term of the loan, regardless of whether the maximum interest rate is reached at the first or subsequent adjustment during the five-year period. Further, the proposed comment cross-referenced existing comments 43(e)(2)(iv)-3 and -4 for additional instruction on how to determine the maximum interest rate during the first five years after the date on which the first regular periodic payment will be due.

The Bureau proposed comment 43(e)(2)(vi)-4.iv to explain how to use the maximum interest rate to determine the APR for purposes of proposed § 1026.43(e)(2)(vi). Specifically, the proposed comment provided that the creditor must determine the APR by treating the maximum interest rate described in proposed § 1026.43(e)(2)(vi) as the interest rate for the full term of the loan. It further provided an example of how to determine the APR by treating the maximum interest rate as the interest rate for the full term of the loan.

The Bureau requested comment on all aspects of the proposed special rule in proposed § 1026.43(e)(2)(vi). In particular, the Bureau requested data regarding short-reset ARMs and those step-rate mortgages that would be subject to the proposed special rule, including default and delinquency rates and the relationship of those rates to price. The Bureau also requested comment on alternative approaches for such loans, including the ones discussed in the proposed rule, such as imposing specific limits on annual rate adjustments for short-reset ARMs, applying a different rate spread, and excluding such loans from General QM eligibility altogether.

Comments Received

Of the approximately 75 comments the Bureau received in response to its General QM Proposal, approximately 25 comments addressed the ARM special rule proposed in § 1026.43(b)(4) and (e)(2)(vi). Nearly all of these ARM commenters represented industry entities—mostly trade associations and a few individual companies. Two commenters represented a coalition of industry and consumer advocates. One individual consumer advocate submitted a comment.

Most ARM commenters acknowledged that short-reset ARMs pose a heightened risk to consumers, with many commenters acknowledging the risks of payment shock. Some commenters agreed that it is appropriate for the Bureau to adopt more stringent requirements for these loans to obtain QM status. Whether or not they acknowledged the need for more stringent requirements, nearly all commenters urged the Bureau to adopt some alternative instead of the proposed special rule.

Commenter criticism generally fell into two categories: (1) That the special rule would be overly burdensome; and (2) that, because some ARMs allow up to a 2 percentage point increase at the first reset,^[322] the special rule would limit or eliminate QM eligibility for some or all short-reset ARMs as they are currently structured—with some commenters predicting that, as a result, some or all short-reset ARMs would cease to be offered in the marketplace. Based on one or both of these criticisms, most ARM commenters recommended that the Bureau either (1) narrow the scope of the special rule to exclude some subset of short-reset ARMs from its coverage or (2) adopt an alternative special rule. One commenter stated that ARMs should no longer be eligible for the QM safe harbor at all, and should instead be designated as rebuttable presumption loans if they are eligible under the General QM loan definition, or non-QM loans if not.

Several commenters criticized the special rule as burdensome. These commenters asserted that the new APR calculation required under the special rule would be “operationally difficult” and would require “significant systems adjustment.” One commenter specifically stated that the APR calculation would add compliance risk and uncertainty to the mortgage market for creditors offering ARM products by adding to the “costs of system updates, staff training, and compliance monitoring; costs that would likely be passed on in one form or another to consumers.” One commenter asserted the adjustments would be “operationally difficult, if not impossible.” Three commenters (including two of the aforementioned commenters asserting burden) requested a longer implementation period due to the added complications of the COVID pandemic and the upcoming replacement of the London InterBank Offered Rate (LIBOR) index with the Secured Overnight Financing Rate (SOFR) index.

Several other commenters stated that the special rule would adversely affect the market for GSE short-reset ARMs that have been developed specifically for the new SOFR index and that such ARMs likely would be unable to achieve QM status under the special rule.

In addition to these SOFR-related market concerns, many other commenters more generally asserted that the special rule would limit or eliminate QM eligibility for some or all short-reset ARMs. Of these commenters, seven predicted that the special rule would eliminate or at least reduce short-reset ARM originations. Three industry commenters predicted that the special rule would result in total elimination of short-reset ARM originations. Four other commenters predicted that the special rule would prevent origination of at least some short-reset ARMs, with two asserting that five-year ARMs would be eliminated and one specifying that three-year ARMs would be eliminated.

Several commenters recommended that the Bureau restrict the scope of the special rule—either to exclude five-year ARMs from coverage or to restrict the scope to short-reset ARMs with an initial fixed-rate period of less than five years, three years, or two years. Some of these commenters urged the Bureau to exclude five-year ARMs from coverage and others recommended narrowing the scope of the special rule to three-year ARMs (or shorter). Some commenters recommended excluding from coverage ARMs that reset after exactly five years or, in the alternative, excluding from coverage ARMs with initial terms of three years or less. One commenter recommended narrowing the special rule to apply to ARMs with an initial period of two years or less.

Several commenters recommended that the Bureau adopt an alternative to the proposed special rule. One industry commenter recommended setting the QM rate-spread threshold for ARMs in a manner that references the maximum interest rate possible in the first five years. The commenter suggested, as an example, requiring that the maximum interest rate possible in the first five years be within a given rate spread of APOR. Similarly, another industry commenter recommended that the Bureau adopt a separate qualification test that compares the maximum interest rate possible in the first five years to the APOR plus an appropriate threshold.

Three commenters, including a consumer advocate and a coalition of industry and consumer advocates, recommended adopting a different special rule that uses the Average Initial Interest Rate (AIIR) instead of APOR as the comparison rate. The Bureau understands that commenters are using AIIR to refer to the mean initial interest rate for a particular ARM product, which is one input into the APOR calculation for ARMs. Another commenter recommended removing QM eligibility for most short-reset ARMs but, in the alternative, supported the special rule using AIIR. These commenters generally maintained that a special rule employing AIIR would ease implementation and preserve the availability of short-reset ARMs for certain consumers while still protecting them from payment shock. As described by commenters, the AIIR special rule would be one part of a two-part test. First, creditors would be required to compare the maximum interest rate in the first five years with the AIIR for a comparable ARM product, plus 2.5 percent, regardless of loan size. If the maximum possible rate is less than or equal to the AIIR plus 2.5 percent, the loan potentially would be eligible for QM status. Second, loans satisfying the initial test would then be subject to the same APR-to-APOR rate-spread tests as other loans under the General QM rule for purposes of determining whether the loans are safe harbor QMs, rebuttable presumption QMs, or non-QM loans under the applicable thresholds.

Three industry commenters recommended a different special rule for short-reset ARMs. They recommended that the Bureau establish “reasonable secondary caps for rate changes allowed within the short-reset period” such that short-reset ARMs meeting those caps would be eligible for QM status. These commenters did not specify their views on what caps on interest rate resets would be reasonable. In the alternative, these commenters, plus a GSE, recommended that the Bureau require creditors to use the fully indexed rate for the remaining loan term after the first five years (rather than the highest possible interest rate in the first five years) to calculate the APR for short-reset ARMs. Although these commenters did not specify which interest rate to use for the first five years, the Bureau understands this approach to be similar to the APR calculation for ARMs in § 1026.17(c)(1), which requires the creditor to disclose a composite APR based on the initial rate for as long as it is charged and, for the remainder of the term, on the fully indexed rate.^[323] In a variation of this approach, another GSE recommended that the Bureau adopt that GSE's own requirements for short-reset ARMs in lieu of the special rule. Specifically, the GSE recommended that the Bureau require creditors to calculate the APR using the greater of the fully indexed rate or 2 percent over the initial note rate for the full term of the loan.

The Final Rule

For the reasons set forth herein, the Bureau is finalizing as proposed the revisions to § 1026.43(e)(2)(vi) and comment 43(e)(2)(vi)-4 regarding the special rule for determining the APR for certain types of loans for purposes of whether a loan meets the General QM loan definition under § 1026.43(e)(2). This special rule applies to loans for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due. For such loans, for purposes of determining whether the loan is a General QM under § 1026.43(e)(2)(vi), the creditor is required to determine the APR by treating the maximum interest rate that may apply during that five-year period as the interest rate for the full term of the loan.^[324] The special rule applies principally to ARMs with initial fixed-rate periods of five years or less (referred to herein as “short-reset ARMs”).^[325] The Bureau concludes that the risks associated with short-reset ARMs can be effectively addressed without prohibiting them from receiving General QM status altogether. This conclusion is consistent with the fact that the Dodd-Frank Act expressly states that short-reset ARMs are eligible for General QM status and includes a specific provision for addressing the potential for payment shock from such loans.

Careful consideration of its data and rationale, and of comments received, leads the Bureau to conclude that while the price-based approach to defining General QMs is generally effective in identifying which loans consumers have the ability to repay and should therefore be eligible for QM status under § 1026.43(e)(2), the special rule is necessary to effectively capture specific unaffordability risks of certain short-reset ARMs. The Bureau further concludes that the burden of implementing the special rule is not unreasonable, as discussed further below, given that all of the inputs needed to calculate the special rule's APR—including the five year maximum interest rate—are already required under existing provisions in Regulation Z and that creditors can offer short-reset ARMs that satisfy the new General QM pricing requirements under the special rule.

As a general matter, as discussed above, the Bureau is adopting in this final rule a non-QM threshold for loans greater than or equal to $110,260 that is higher than the threshold that it proposed. Specifically, § 1026.43(e)(2) provides that loans greater than or equal to $110,260 may be eligible for QM status if the APR does not exceed APOR 2.25 or more percentage points. The Bureau notes that this change will increase the pool of loans that achieve QM status under the ATR/QM Rule, including short-reset ARMs subject to the special rule. Thus, the 2.25-percentage-point threshold under this final rule will result in more short-reset ARMs achieving QM status than would have under the 2-percentage-point threshold in the proposal. While short-reset ARMs offer consumers who can afford them an important alternative to fixed-rate mortgage loans, the Bureau estimates that the special rule will apply to a relatively small percentage of the mortgage market. Based on 2018 HMDA data, the Bureau estimates that approximately 36,000 conventional purchase loans, or approximately 1.3 percent of conventional purchase loans in the U.S. mortgage market, would have been subject to the special rule had it been in effect that year.

Structure and pricing particular to ARMs. As explained in the proposal, absent special treatment, short-reset ARMs may present particular concerns under an approach that uses APR as an indicator of ability to repay. Short-reset ARMs may be affordable for the initial fixed-rate period but may become unaffordable relatively soon after consummation if the payments increase appreciably after reset, causing payment shock. The APR for short-reset ARMs is not as predictive of ability to repay as for fixed-rate mortgages because of how ARMs are structured and priced and how the APR for ARMs is determined under various provisions in Regulation Z. Several different provisions in Regulation Z address the calculation of the APR for ARMs. For disclosure purposes, if the initial interest rate is determined by the index or formula to make later interest rate adjustments, Regulation Z generally requires the creditor to base the APR disclosure on the initial interest rate at consummation and to not assume that the rate will increase during the remainder of the loan.^[326] In some transactions, including many ARMs, the creditor may set an initial interest rate that is lower (or, less commonly, higher) than the rate would be if it were determined by the index or formula used to make later interest rate adjustments. For these ARMs, Regulation Z requires the creditor to disclose a composite APR based on the initial rate for as long as it is charged and, for the remainder of the term, on the fully indexed rate.^[327] The fully indexed rate at consummation is the sum of the value of the index at the time of consummation plus the margin, based on the contract. The Dodd-Frank Act requires a different APR calculation for ARMs for the purpose of determining whether ARMs are subject to certain HOEPA requirements.^[328] As implemented in § 1026.32(a)(3)(ii), the creditor is required to determine the APR for HOEPA coverage for transactions in which the interest rate may vary during the term of the loan in accordance with an index, such as with an ARM, by using the fully indexed rate or the introductory rate, whichever is greater.^[329]

The requirements in Regulation Z for determining the APR for disclosure purposes and for HOEPA coverage purposes do not account for any potential increase or decrease in interest rates based on changes to the underlying index. If interest rates rise after consummation, and therefore the value of the index rises to a higher level, the loan can reset to a higher interest rate than the fully indexed rate at the time of consummation. The result would be a higher payment than the one calculated based on the rates used in determining the APR, and a higher effective rate spread (and increased likelihood of delinquency) than the spread that would be taken into account for determining General QM status at consummation under the price-based approach in the absence of a special rule.

ARMs can present more risk for consumers than fixed-rate mortgages, depending on the direction and magnitude of changes in interest rates. In the case of a 30-year fixed-rate loan, creditors or mortgage investors assume both the credit risk and the interest-rate risk (i.e., the risk that interest rates rise above the fixed rate the consumer is obligated to pay), and the price of the loan, which is fully captured by the APR, reflects both risks. In the case of an ARM, the creditor or investor assumes the credit risk of the loan, but the consumer assumes most of the interest-rate risk, as the interest rate will adjust along with the market. The extent to which the consumer assumes the interest-rate risk is established by caps in the note on how high the interest rate charged to the consumer may rise. To compensate for the added interest-rate risk assumed by the consumer (as opposed to the creditor or investor), ARMs are generally priced lower—in absolute terms—than a 30-year fixed-rate mortgage with comparable credit risk.^[330] Yet with rising interest rates, the risks that ARMs could become unaffordable, and therefore lead to delinquency or default, are more pronounced. As noted above, the requirements for determining the APR for ARMs in Regulation Z do not reflect this risk because they do not take into account potential increases in the interest rate over the term of the loan based on changes to the underlying index. This APR may therefore understate the risk that the loan may become unaffordable to the consumer if interest rates increase.

Unaffordability risk more acute for short-reset ARMs. As the Bureau noted in the proposal, short-reset ARMs may present greater risks of unaffordability than other ARMs. While all ARMs run the risk of increases in interest rates and payments over time, longer-reset ARMs (i.e., ARMs with initial fixed-rate periods of longer than five years) present a less acute risk of unaffordability than short-reset ARMs. Longer-reset ARMs permit consumers to take advantage of lower interest rates for more than five years and thus, akin to fixed-rate mortgages, provide consumers significant time to pay off or refinance, or to otherwise adjust to anticipated changes in payment during the relatively long period during which the interest rate is fixed and before payments may increase.

Short-reset ARMs can also contribute to speculative lending because they permit creditors to originate loans that could be affordable in the short term, with the expectation that property values will increase and thereby permit consumers to refinance before payments may become unaffordable. Further, creditors can minimize their credit risk on such ARMs by, for example, requiring lower LTV ratios, as was common in the run-up to the 2008 financial crisis.^[331] Additionally, creditors may be more willing to market these ARMs in areas of strong housing-price appreciation, irrespective of a consumer's ability to absorb the potentially higher payments after reset, because creditors may expect that consumers will have the equity in their homes to refinance if necessary.

In the Dodd-Frank Act, Congress addressed affordability concerns specific to short-reset ARMs and their eligibility for QM status by providing in TILA section 129C(b)(2)(A)(v) that, to receive QM status, ARMs must be underwritten using the maximum interest rate that may apply during the first five years.^[332] The ATR/QM Rule implemented this requirement in Regulation Z at § 1026.43(e)(2)(iv). For many short-reset ARMs, this requirement resulted in a higher DTI that would have to be compared to the Rule's 43 percent DTI limit to determine whether the loans were eligible to receive General QM status. Particularly in a higher-rate environment in which short-reset ARMs could become more attractive, the five-year maximum interest-rate requirement, combined with the Rule's 43 percent DTI limit, would have likely prevented some of the riskiest short-reset ARMs (i.e., those that adjust sharply upward in the first five years and cause payment shock) from obtaining General QM status. As discussed above, the Bureau is finalizing a price-based approach that removes the DTI limit from the General QM loan definition in § 1026.43(e)(2)(vi). As a result, the Bureau finds that, without the special rule, a price-based approach would not adequately address the risk that consumers taking out short-reset ARMs may not have the ability to repay those loans but that such loans would nonetheless be eligible for General QM status under § 1026.43(e)(2).^[333]

The price-based approach to addressing affordability concerns. As noted in the proposal, the Bureau's analysis of historical ARM pricing and performance indicates that the General QM product restrictions combined with the price-based approach would have effectively excluded many—but not all—of the riskiest short-reset ARMs from obtaining General QM status. As a result, the Bureau concludes that an additional mechanism is merited to exclude from the General QM loan definition these short-reset ARMs for which the pricing and structure indicate a risk of delinquency that is inconsistent with the presumption of compliance with the ATR requirements that comes with QM status.

The Bureau's analysis of NMDB data shows that short-reset ARMs originated from 2002 through 2008 had, on average, substantially higher early delinquency rates (14.9 percent) than other ARMs (10.1 percent) or than fixed-rate mortgages (5.4 percent). Many of these short-reset ARMs were also substantially higher-priced relative to APOR and more likely to have product features that TILA and the ATR/QM Rule now prohibit for QMs, such as interest-only payments or negative amortization. In considering only loans without such restricted features and with rate spreads within 2 percentage points of APOR (the proposed non-QM threshold), short-reset ARMs still have the highest average early delinquency rate (5.5 percent), but the difference relative to other ARMs (4.3 percent) and fixed-rate mortgages (4.2 percent) is smaller. While not a factor in the Bureau's decision to finalize the special rule as proposed, the Bureau's analysis of early delinquency rates of loans without restricted features and with rate spreads within 2.25 percentage point of APOR (the non-QM threshold under the Final Rule) yields similar results, though the delinquency rates for short-reset ARMs as compared to all other loans are slightly higher. Under that analysis, the early delinquency rate for short-reset ARMs is 6.2 percent as compared to 4.4 percent for all other ARMs and 4.3 percent for fixed-rate mortgages.^[334]

In the proposal, the Bureau requested additional data or evidence comparing loan performance of short-reset ARMs, other ARMs, and fixed-rate mortgages, as well as data comparing the performance of such loans during periods of rising interest rates. In response, a few commenters stated that their internal data for loans originated post-crisis—in an environment of relatively low interest rates—showed generally comparable delinquency rates between certain ARMs and fixed-rate mortgages. Those delinquency rates are generally consistent with those reflected in the data on which the Bureau relied, in part, to propose the special rule. No commenters, however, provided data on comparative loan performance during periods of rising interest rates—which, as discussed herein, is the interest-rate environment for which the special rule's additional safeguards are primarily designed. The Bureau recognizes that rising interest rates may also pose some risk of unaffordability for longer-reset ARMs later in the loan term. However, as also discussed herein, the Bureau is finalizing the special rule to address the specific concern that short-reset ARMs pose a higher risk than other ARMs of becoming unaffordable in the first five years, before consumers have sufficient time to refinance or adjust to the larger payments—a concern Congress also identified in the Dodd-Frank Act. Short-reset ARMs have the potential for a significant interest rate increase early in the loan term and present concerns that the payments may therefore become unaffordable. Commenters did not present evidence controverting that short-reset ARMs may present particular risks. Indeed, most commenters acknowledged that short-reset ARMs do in fact present additional concerns about affordability.

A combination of factors post-financial crisis—including a sharp drop in ARM originations and the restriction of such originations to highly creditworthy borrowers, as well as the prevalence of low interest rates—likely has muted the overall risks of short-reset ARMs. During the peak of the mid-2000s housing boom, ARMs accounted for as much as 52 percent of all new originations. In contrast, the current market share of ARMs is relatively small. Post-crisis, the ARM share had declined to 12 percent by December 2013 and to 1.4 percent by July 2020, only slightly above the historical low of 1 percent in 2009.^[335] One major factor contributing to the overall decline in ARM volume is the low-interest-rate environment since the end of the financial crisis. Typically, ARMs are more popular when conventional interest rates are high, since the rate (and monthly payment) during the initial fixed period is typically lower than the rate of a comparable conventional fixed-rate mortgage.

Consistent with TILA section 129C(b)(2)(A), the January 2013 Final Rule prohibited ARMs with higher-risk features such as interest-only payments or negative amortization from receiving General QM status. According to the Assessment Report, short-reset ARMs comprised 17 percent of ARMs in 2012, prior to the January 2013 Final Rule, and fell to 12.3 percent in 2015, after the effective date of the Rule.^[336] The Assessment Report also found that short-reset ARMs originated after the effective date of the Rule were restricted to highly creditworthy borrowers.^[337] The Assessment Report further found that conventional, non-GSE short-reset ARMs originated after the effective date of the Rule had early delinquency rates of only 0.2 percent.^[338] Due to the post-crisis low interest rate environment and restriction of ARM originations to highly creditworthy borrowers, these recent originations may not accurately reflect the potential unaffordability of short-reset ARMs under different market conditions than those that currently prevail.

Special rule for APR determination for short-reset ARMs.^[339] Given the potential that rising interest rates could cause short-reset ARMs to become unaffordable for consumers following consummation and the fact that the price-based approach does not account for some of those risks because of how APRs are determined for ARMs, the Bureau is finalizing the proposed special rule to determine the APR for short-reset ARMs for purposes of defining General QM under § 1026.43(e)(2). As noted above, in defining QM in TILA, Congress adopted a special requirement to address affordability concerns for short-reset ARMs. Specifically, TILA provides that, for an ARM to be a QM, the underwriting must be based on the maximum interest rate permitted under the terms of the loan during the first five years. With the 43 percent DTI limit in the current ATR/QM Rule, implementing the five-year underwriting requirement is straightforward: The Rule requires a creditor to calculate DTI using the mortgage payment that results from the maximum possible interest rate that could apply during the first five years.^[340] This ensured that the creditor calculates the DTI using the highest interest rate that the consumer may experience in the first five years, and the loan is not eligible for QM status under § 1026.43(e)(2) if the DTI calculated using that interest rate exceeds 43 percent. The Bureau concludes that using the fully indexed rate to determine the APR for purposes of the rate-spread thresholds in § 1026.43(e)(2)(vi) as finalized would not provide a sufficiently meaningful safeguard against the elevated likelihood of delinquency for short-reset ARMs. For that reason, the Bureau is finalizing the proposed special rule for determining the APR for such loans.

The Bureau concludes the statutory five-year underwriting requirement provides a basis for the special rule for determining the APR for short-reset ARMs for purposes of General QM rate-spread thresholds under § 1026.43(e)(2). Specifically, under the special rule, the creditor must determine the APR by treating the maximum interest rate that may apply during the first five years, as described in § 1026.43(e)(2)(vi), as the interest rate for the full term of the loan. That APR determination is then compared to the APOR ^[341] to determine General QM status. This approach addresses in a targeted manner the primary concern about short-reset ARMs—payment shock—by accounting for the risk of delinquency and default associated with payment increases under these loans. And it would do so in a manner that is consistent with the five-year framework embedded in TILA for such ARMs and implemented in the current ATR/QM Rule.

In sum, the Bureau finds that the special rule is consistent with both TILA's statutory mandate for short-reset ARMs and the proposed price-based approach. As discussed above in part V, the rate spread of APR over APOR is strongly correlated with early delinquency rates. As a result, such rate spreads may generally serve as an effective indicator for a consumer's ability to repay. However, the structure and pricing of ARMs can result in early interest rate increases that are not fully accounted for in Regulation Z provisions for determining the APR for ARMs. Such increases could diminish the effectiveness of the rate spread as an indicator and lead to heightened risk of early delinquency for short-reset ARMs relative to other loans with comparable APR over APOR rate spreads. The special rule, by requiring creditors to more fully incorporate this interest-rate risk in determining the APR for short-reset ARMs, will more fully ensure that the resulting pricing accounts for that risk for such loans.

The special rule requires that the maximum interest rate in the first five years be treated as the interest rate for the full term of the loan to determine the APR. The Bureau concludes that a composite APR determination based on the maximum interest rate in the first five years and the fully indexed rate for the remaining loan term could understate the APR for short-reset ARMs by failing to sufficiently account for the risk that consumers with such loans could face payment shock early in the loan term. Accordingly, to account for that risk, and to ensure that the QM presumption of compliance is accorded to short-reset ARMs for which the consumer has the ability to repay, the Bureau is requiring that the APR for short-reset ARMs be based on the maximum interest rate during the first five years.

Commenter criticism of the special rule: Burden and market effects. As noted above, commenter criticism of the proposed special rule generally fell into two categories: (1) The special rule would be overly burdensome; and (2) because some ARMs allow up to a 2 percentage point increase at the first reset, the special rule would limit or eliminate QM eligibility for some or all short-reset ARMs—with some commenters predicting that, as a result, some or all short-reset ARMs would cease to be offered in the marketplace.

With regard to the first criticism, some commenters asserted that the special rule would increase burden by adding operational complexity and compliance uncertainty. These commenters provided no further explanation or data to justify their claims. The Bureau recognizes that the special rule's APR calculation is a new regulatory requirement. However, the Bureau concludes that its special rule addresses the risk posed by short-reset ARMs without adding unreasonable burden. Cognizant of reducing burden resulting from calculating a new APR, the Bureau proposed the special rule, in part, because it parallels the underwriting requirement in existing § 1026.43(e)(2)(iv), which already requires creditors to calculate the five-year maximum interest rate for short-reset ARMs. As such, the special rule's APR calculation is based on an input already required for short-reset ARMs under the underwriting calculation. Moreover, creditors already have all of the other inputs required for the special rule's APR calculation from existing APR regulatory requirements. The Bureau expects that these factors will mitigate the burden of implementing systems changes to comply with the special rule. The Bureau also notes that the different APR calculation required under § 1026.32(a)(3)(ii) for purposes of determining whether ARMs are subject to certain HOEPA requirements has not resulted in compliance uncertainty.

Three commenters raised concerns that adapting to the special rule would be burdensome because it would overlap with the transition from the LIBOR index to the SOFR index (and because of the pandemic) and therefore requested a longer implementation period. The implementation period of the Final Rule is addressed in part VII below.

A few other commenters stated that the special rule would adversely affect the market for GSE short-reset ARMs that have been developed specifically for the new SOFR index, and that such ARMs likely would be unable to obtain QM status under the special rule. The Bureau notes that the special rule does not depend on which index a creditor uses to determine the interest rate of a short-reset ARM. Thus, the transfer from LIBOR ARMs to SOFR ARMs has no effect on the application of the special rule, as it is the structure of the rate resets permitted under the contract within the first five years that will determine the maximum interest rate for the purposes of calculating the APR under the special rule. Creditors offering ARM products, including short-reset ARMs, will have to complete the work to transition from LIBOR to SOFR regardless of the parameters of the Bureau's special rule. Moreover, the Bureau understands that the 2 percentage point cap on the initial reset of most GSE short-reset ARMs is the same for both GSE LIBOR ARMs and GSE SOFR ARMs. While the current ATR/QM Rule's GSE Patch granted QM status to all GSE-eligible ARMs, under this final rule, GSE ARMs will require similar adjustments due to their rate reset caps in order to qualify for QM status—regardless of which index is used. Further, the Bureau notes that only approximately 5 percent of 2018 conventional purchase ARMs that would have been subject to the special rule were GSE loans. In sum, the Bureau recognizes the operational challenges posed by the transition from LIBOR to SOFR, but the Bureau finds that the special rule would not exacerbate these challenges and that these challenges are unrelated to the types of ARMs that qualify for a QM presumption of compliance under the special rule.

With respect to the remaining criticisms of the special rule's projected market effects, commenters claimed that, because some short-reset ARMs allow up to a 2 percentage point increase at the first reset, the special rule would limit or eliminate QM eligibility for some or all short-reset ARMs. A few of these commenters further predicted that, as a result, some or all short-reset ARMs would cease to be offered in the marketplace. These commenters did not provide additional data or evidence to support their projections. As discussed above, the Bureau is increasing the rate-spread threshold for eligibility under the General QM loan definition from the proposed 2-percentage-point threshold to 2.25 percentage points for loans less than or equal to $110,260. As a result of this increased threshold, more short-reset ARMs will achieve QM status than would have under the proposal. This is especially true for many five-year ARMs, including existing GSE five-year ARMs, which under the proposed special rule might have required modifications to the current interest rate cap to obtain QM status. Under the 2.25-percentage-point threshold, many of these loans may qualify as QMs as currently structured. Because most GSE five-year ARMs (both LIBOR and SOFR) provide for a 2 percentage point cap on the first reset, many of these short-reset ARMs will fall within the new QM threshold. Due to this increased threshold, any five-year ARM with an initial APR within 0.25 percentage points of the APOR at origination can have an initial adjustment of up to 2 percent and still qualify as a QM under the special rule.

The Bureau recognizes that, because the QM safe harbor threshold remains unchanged, many of the short-reset ARMs that achieve QM status under the Final Rule's expanded spread will receive a rebuttable presumption of compliance rather than a conclusive presumption. Due to the risk that these short-reset ARMs (i.e., those with relatively high interest rate caps) may become unaffordable after early resets, the Bureau concludes that rebuttable presumption status, as opposed to safe harbor status, is appropriate for such loans. Furthermore, according to the Bureau's evidence, as discussed in the proposal and above, the fact that many of these loans may qualify only for a rebuttable presumption and not a safe harbor is not likely to have a significant impact on access to responsible, affordable mortgage credit. As discussed in more detail above, creditors readily make rebuttable presumption QMs, thus indicating that the non-QM threshold is the more relevant threshold in determining access to responsible, affordable mortgage credit under the General QM amendments.

The Bureau is aware that the increase in the rate-spread threshold will have a greater impact on QM eligibility of five-year ARMs as compared to three-year ARMs. For example, GSE three-year ARMs permit interest rate increases as high as 6 percentage points in the first five years and as such likely will not qualify for General QM status. The Bureau notes that the purpose of the special rule is to ensure that General QM status will not be accorded to certain loans for which the interest rate may sharply increase in the first five years, resulting in pricing that exceeds the non-QM threshold in this final rule and in potentially unaffordable payments. Consistent with this purpose, the special rule would preclude such loans from obtaining General QM status, including many three-year ARMs with interest rates that may increase by as much as 6 percentage points in the first five years. Loans for which the interest rate may increase so sharply early in the term of the loan do not warrant the General QM presumption of compliance with the ATR requirements.

To the extent the increased rate-spread threshold does not address commenter concerns with regard to access to credit, the Bureau notes that creditors can and do market QM-eligible ARMs that either satisfy the requirements of the special rule by not permitting resets above 2.25 percentage points within the first five years or that fall outside the purview of the special rule by resetting later than five years (60 months) after the first payment is due. Market participants currently originate some five-year ARMs with sufficiently low initial reset caps or with an initial reset that occurs shortly after 60 months. For example, the definition of a GSE five-year ARM allows an initial fixed-rate period of up to 66 months.^[342] Thus, GSEs and creditors can offer ARMs that satisfy the General QM pricing requirements under the special rule or that fall outside the scope of the special rule. Also, while interest rate reset data for privately-held non-agency loans is not reliably available, the Bureau notes that both FHA and VA ARMs, although subject to their own agency QM rules, contain interest rate reset caps that would fall within the parameters of the special rule as finalized.^[343]

A few commenters asked for clarification of certain aspects of the special rule. One commenter requested that the Bureau clarify whether the special rule applies to five-year ARMs. Specifically, the commenter asked for clarification as to whether the first interest rate reset of a five-year ARM is included in the special rule's APR calculation, given the special rule's applicability to loans for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due. To the extent that the first interest rate reset of a five-year ARM occurs on the five-year anniversary of the due date of the first periodic payment, such ARMs are subject to the special rule. As noted in the proposal, the special rule is identical in this regard to the existing underwriting requirement for short-reset ARMs in § 1026.43(e)(2)(iv). Also, comment 43(e)(2)(vi)-4.ii, which the Bureau is finalizing as proposed, clarifies that the special rule applies to five-year ARMs.

One commenter posed several questions concerning how the special rule applies to certain loan products or in various factual scenarios. To the extent that the commenter's questions are not addressed in the final rule, the Bureau notes that it has a variety of tools for answering such questions once a final rule is issued, including external guidance materials and an informal guidance function.

Commenter recommendations. Commenters that criticized the special rule generally recommended one of two ways to address their criticisms: Narrow the scope of the special rule or substitute an alternative special rule.

Some commenters recommended narrowing the scope of the special rule to expand the number of short-reset ARMs that obtain QM status—either to exclude five-year ARMs from coverage or to restrict the scope to short-reset ARMs with an initial fixed-rate period of less than five years, three years, or two years. The Bureau declines to adopt these recommendations and is finalizing the special rule as proposed to cover short-reset ARMs with initial fixed-rate periods of five years or less, for the following reasons and those discussed above.

The majority of these commenters specifically recommended excluding five-year ARMs from coverage. The Bureau notes that coverage of the special rule is already relatively narrow. Including five-year ARMs, the Bureau estimates that the special rule would apply to 36,000 conventional purchase loans annually, according to 2018 HMDA data. Excluding five-year ARMs from the scope of the special rule would reduce that number to 3,500 loans. Further, as discussed above, because the Bureau is increasing the rate-spread threshold from 2 percentage points to 2.25 percentage points for loans greater than or equal to $110,260, more five-year ARMs will obtain QM status under the special rule as finalized.

The Bureau reiterates that the purpose of the special rule is to prevent certain short-reset ARMs from obtaining QM status if there may be a sharp rise in interest rates soon after origination. This rise may occur with three-year ARMs, which may have contracts that permit the interest rate to increase by as much as 6 percentage points in the first five years. Because consumers may lack sufficient time to adjust to larger payments early in the loan term or to build enough equity to refinance, such ARMs pose a higher risk of early delinquency. For these additional reasons, the Bureau declines to narrow coverage to short-reset ARMs with initial fixed-rate periods of three years or less.

Some commenters recommended the Bureau implement alternative special rules to address the risks presented by short-reset ARMs. The Bureau declines to adopt the alternative special rules recommended by these commenters. To the extent that commenters are advocating for alternative special rules to increase the number of short-reset ARMs that could obtain QM status, the Bureau notes that the increase of the rate-spread threshold in the Final Rule will expand the pool of QM-eligible short-reset ARMs compared to the proposal.

As noted above, a few commenters recommended adopting a special rule that uses the maximum interest rate in the first five years of the loan (as opposed to using the APR required by the special rule) to compare with the AIIR (instead of APOR), plus the additional cushion of 2.5 percentage points (“AIIR special rule”). As the Bureau understands this recommendation, short-reset ARMs satisfying the initial test would then be subject to the same APR-to-APOR rate-spread tests as other loans under the General QM loan definition for purposes of determining whether the loans receive a safe harbor or a rebuttable presumption or are non-QM under the applicable thresholds.

The Bureau recognizes that adopting this AIIR special rule would expand the number of short-reset ARMs that would achieve QM status, as interest rate increases of up to 2.5 percentage points early in the life of the loan would meet that special rule's pricing threshold. The Bureau also recognizes that using the five-year maximum interest rate in this special rule could be a burden-reduction measure, since creditors will already have calculated that input, as it is currently required for underwriting loans pursuant to § 1026.43(e)(2)(iv).

The AIIR special rule would expand the pool of QM-eligible short-reset ARMs to those whose interest rates increase by as much as 2.5 percentage points. However, commenters provided no evidence that this threshold would appropriately identify which loans are likely affordable and should receive a presumption of compliance. Moreover, the Bureau concludes that any potential burden-reduction benefits are outweighed by the complexity of introducing into the General QM loan definition a new measure—the AIIR—and a new formula that requires, as the first step in a two-step process, comparing the maximum five-year interest rate to the AIIR and then adding 2.5 percentage points. (Then, if the short-reset ARM meets the threshold of the first test, it is still subject to the price-based APR-APOR rate-spread test.) In addition, because “AIIR” is not a commonly used term, the Bureau is concerned that creditors may not understand AIIR to mean what the Bureau believes the commenters intended, i.e., the mean initial interest rate for a particular ARM product. As such, a requirement to use the AIIR could necessitate significant regulatory explanation, likely adding implementation and compliance burden. Additionally, the AIIR special rule would deviate from the final rule's straight-forward APR-to-APOR comparison, requiring an additional comparison of interest rates. For these reasons, the Bureau declines to adopt the AIIR special rule.

Two commenters recommended a special rule using the maximum interest rate in the first five years for short-reset ARMs instead of the APR calculation required by the special rule (“five-year maximum interest rate special rule”). These commenters advocated this alternative special rule as way to expand QM eligibility for short-reset ARMs and to ease burden, as this calculation of the five-year maximum interest rate is already required for underwriting short-reset ARMs in the current ATR/QM Rule ^[344] and therefore would not require an additional calculation. One commenter recommended setting the General QM rate-spread threshold for short-reset ARMs in a manner that compares the maximum interest rate possible in the first five years with a given rate spread of APOR. The other commenter similarly recommended adopting a separate qualification test that compares the highest interest rate within five years to the APOR plus an appropriate threshold.

The Bureau recognizes that the five-year maximum interest rate special rule suggested by the commenter would expand the pool of QM-eligible short-reset ARMs. However, this would be accomplished in part by excluding from the APR calculation non-interest finance charges, which are included for other types of loans subject to the Rule. Such finance charges are key components of a loan's pricing and therefore contribute to making pricing an effective indicator of a consumer's ability to repay. As such, the Bureau declines to exclude non-interest finance charges from the APR calculation for short-reset ARMs.

The Bureau further notes that the interest-rate-to-APOR comparison would allow riskier loans—that is, loans that may reset to a significantly higher interest rate in the first five years—to obtain QM status. As discussed above, the intended effect of the Bureau's special rule is to guard against certain short-reset ARMs with early, potentially unaffordable, sharp increases in interest rates from obtaining QM status. For these reasons, the Bureau declines to adopt the five-year maximum interest rate special rule.

As noted above, a few commenters recommended replacing the special rule with reasonable secondary interest rate caps during the first five years for short-reset ARMs (“rate cap special rule”). While this alternative special rule would directly address the threat of payment shock, these commenters did not specify what rate caps would be reasonable or how such caps would operate in relation to the contractual rate caps under the ARM note. In the proposed rule, for these same reasons, the Bureau considered and declined to propose interest rate caps that commenters had suggested in response to the ANPR and noted that the special rule would address the problem in a more streamlined manner. Additionally, the rate cap special rule would deviate from the pricing approach that would apply to other ARMs and fixed-rate mortgages subject to this final rule. Moreover, commenters provided no evidence indicating that rate caps in general or that specific rate caps would identify more accurately than the Bureau's special rule those short-reset ARMs likely to be affordable and thus meriting a presumption of compliance.

The commenters that recommended secondary rate caps alternatively recommended that the Bureau require creditors to use the fully indexed rate for the remaining loan term after the first five years to calculate the APR for short-reset ARMs (without specifying which interest rate to use for the first five years). The Bureau understands this approach to be similar to the general APR requirements for ARMs in § 1026.17(c)(1), which require the creditor to disclose a composite APR based on the initial rate for as long as it is charged and, for the remainder of the term, on the fully indexed rate. Absent the Bureau's special rule, this would be the applicable APR formula for short-reset ARMs under the price-based approach. Another GSE recommended the Bureau adopt that GSE's own requirements for short-reset ARMs, which the GSE described as using the greater of the fully indexed rate or 2 percent over the initial note rate for the full term of the loan.

The Bureau declines to adopt either of these approaches. Using the fully indexed rate to calculate the APR for short-reset ARMs—for some or all of the loan term—would not adequately address the risk that such ARMs can become unaffordable. As noted above, if interest rates rise after consummation, and therefore the value of the index rises to a higher level, the loan can reset to a higher interest rate than the fully indexed rate at the time of consummation. The result would be a higher payment than the one that would be calculated based on the rates used in determining the APR. Requiring the use of 2 percent over the initial note rate (if greater than the fully indexed rate) also would not adequately address this risk. As noted above, many short-reset ARMs are permitted to adjust substantially more than 2 percent early in the life of the loan, particularly those structured to have multiple adjustments within the first five years. The interest rate of such ARMs can adjust upward 6 percentage points in the first five years of the loan. By requiring that the APR for short-reset ARMs be determined by treating the maximum interest rate during the first five years as the interest rate for the full term of the loan, the Bureau's special rule is designed to account for that risk, and to ensure that General QM status is accorded to short-reset ARMs that merit a presumption of compliance.

Legal authority. As discussed above in part IV, TILA section 105(a), directs the Bureau to prescribe regulations to carry out the purposes of TILA, and provides that such regulations may contain additional requirements, classifications, differentiations, or other provisions, and may provide for such adjustments and exceptions for all or any class of transactions that the Bureau judges are necessary or proper to effectuate the purposes of TILA, to prevent circumvention or evasion thereof, or to facilitate compliance therewith. In particular, a purpose of TILA section 129C, as amended by the Dodd-Frank Act, is to assure that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loans.

As also discussed above in part IV, TILA section 129C(b)(3)(B)(i) authorizes the Bureau to prescribe regulations that revise, add to, or subtract from the criteria that define a QM upon a finding that such regulations are necessary or proper to ensure that responsible, affordable mortgage credit remains available to consumers in a manner consistent with the purposes of section 129C, necessary and appropriate to effectuate the purposes of section 129C and section 129B, to prevent circumvention or evasion thereof, or to facilitate compliance with such section.

The Bureau is finalizing the special rule in § 1026.43(e)(2)(vi) regarding the APR determination of certain loans for which the interest rate may or will change pursuant to its authority under TILA section 105(a) to make such adjustments and exceptions as are necessary and proper to effectuate the purposes of TILA, including that consumers are offered and receive residential mortgage loans on terms that reasonably reflect their ability to repay the loans. The Bureau concludes that these provisions will ensure that General QM status would not be accorded to short-reset ARMs and certain other loans that pose a heightened risk of becoming unaffordable relatively soon after consummation. The Bureau is also finalizing these provisions pursuant to its authority under TILA section 129C(b)(3)(B)(i) to revise and add to the criteria that define a QM. The Bureau believes that the special rule's APR determination provisions in § 1026.43(e)(2)(vi) will ensure that responsible, affordable mortgage credit remains available to consumers in a manner consistent with the purpose of TILA section 129C, referenced above, as well as effectuate that purpose.

43(e)(4)

TILA section 129C(b)(3)(B)(ii) directs HUD, VA, USDA, and RHS to prescribe rules defining the types of loans they insure, guarantee, or administer, as the case may be, that are QMs. Pending the other agencies' implementation of this provision, the Bureau included in the ATR/QM Rule a temporary category of QMs in the special rules in § 1026.43(e)(4)(ii)(B) through (E) consisting of mortgages eligible to be insured or guaranteed (as applicable) by HUD, VA, USDA, and RHS. The Bureau also created the Temporary GSE QM loan definition in § 1026.43(e)(4)(ii)(A).

Section 1026.43(e)(4)(i) currently states that, notwithstanding § 1026.43(e)(2), a QM is a covered transaction that satisfies the requirements of § 1026.43(e)(2)(i) through (iii)—the General QM loan-feature prohibitions and points-and-fees limits—as well as one or more of the criteria in § 1026.43(e)(4)(ii). Section 1026.43(e)(4)(ii) currently states that a QM under § 1026.43(e)(4) must be a loan that is eligible under enumerated “special rules” to be (A) purchased or guaranteed by the GSEs while under the conservatorship of the FHFA (the Temporary GSE QM loan definition), (B) insured by HUD under the National Housing Act, (C) guaranteed by VA, (D) guaranteed by USDA pursuant to 42 U.S.C. 1472(h), or (E) insured by RHS. Section 1026.43(e)(4)(iii)(A) currently states that § 1026.43(e)(4)(ii)(B) through (E) shall expire on the effective date of a rule issued by each respective agency pursuant to its authority under TILA section 129C(b)(3)(ii) to define a QM. Section 1026.43(e)(4)(iii)(B) currently states that, unless otherwise expired under § 1026.43(e)(4)(iii)(A), the special rules in § 1026.43(e)(4) are available only for covered transactions consummated on or before January 10, 2021.

In the General QM Proposal, the Bureau proposed to replace current § 1026.43(e)(4) with a provision stating that, notwithstanding § 1026.43(e)(2), a QM is a covered transaction that is defined as a QM by HUD under 24 CFR 201.7 or 24 CFR 203.19, VA under 38 CFR 36.4300 or 38 CFR 36.4500, or USDA under 7 CFR 3555.109. The Bureau proposed these amendments because, in the Extension Proposal, the Bureau proposed to revise § 1026.43(e)(4)(iii)(B) to state that, unless otherwise expired under § 1026.43(e)(4)(iii)(A), the special rules in § 1026.43(e)(4) would be available only for covered transactions consummated on or before the effective date of a final rule issued by the Bureau amending the General QM loan definition.^[345] In the General QM Proposal, the Bureau also noted that, after the promulgation of the January 2013 Final Rule, each of the agencies described in § 1026.43(e)(4)(ii)(B) through (E) adopted separate definitions of qualified mortgages.^[346] The Bureau noted that, as a result, the special rules in § 1026.43(e)(4)(ii)(B) through (E) are already superseded by the actions of HUD, VA, and USDA. The Bureau's proposed amendments to § 1026.43(e)(4) provided cross-references to each of these other agencies' definitions so that creditors and practitioners have a single point of reference for all QM definitions.

The Bureau proposed to amend comment 43(e)(4)-1 to reflect the cross-references to the QM definitions of other agencies and to clarify that a covered transaction that meets another agency's definition is a QM for purposes of § 1026.43(e). The Bureau proposed to amend Comment 43(e)(4)-2 to clarify that covered transactions that met the requirements of § 1026.43(e)(2)(i) through (iii), were eligible for purchase or guarantee by Fannie Mae or Freddie Mac, and were consummated prior to the effective date of any final rule promulgated as a result of the proposal would still be considered a QM for purposes of § 1026.43(e) after the adoption of such potential final rule. Comments 43(e)(4)-3, -4, and -5 would have been removed. The Bureau requested comment on the proposed amendments to § 1026.43(e)(4) and related commentary. Comments on the proposal did not discuss the proposed amendments to § 1026.43(e)(4) and its related commentary.

In this final rule, the Bureau amends § 1026.43(e)(4) as proposed, with modifications to the commentary to clarify the application of this final rule's effective date to the availability of the Temporary GSE QM loan definition.

As noted above, on October 20, 2020, the Bureau issued the Extension Final Rule to replace the January 10, 2021 sunset date of the Temporary GSE QM loan definition with a provision stating that the Temporary GSE QM loan definition will be available only for covered transactions for which the creditor receives the consumer's application before the mandatory compliance date of this final rule.^[347] As noted in part VII below, this final rule will have an effective date of March 1, 2021, and a mandatory compliance date of July 1, 2021. As a result, the Temporary GSE QM loan definition will still be used by creditors after the effective date of March 1, 2021 and will not expire until July 1, 2021. In this final rule, the Bureau is making changes to proposed comment 43(e)(4)-2 to reflect this final rule's effective date and mandatory compliance date.

As noted above, the Bureau proposed to remove 43(e)(4)-3. In this final rule, the Bureau is instead revising comment 43(e)(4)-3 to cross-reference new comment 43-2. As discussed further in part VII below, new comment 43-2 clarifies that, for transactions for which a creditor received an application on or after March 1, 2021, but prior to July 1, 2021, a creditor has the option of complying either with 12 CFR part 1026 as it is in effect or with 12 CFR part 1026 as it was in effect on February 26, 2021. The Bureau believes this comment will assist creditors and secondary market participants with compliance with the final rule because it will clarify that, even though the Temporary GSE QM loan definition will not appear in § 1026.43(e)(4) after this final rule's effective date of March 1, 2021, creditors may continue to use it for transactions for which they received the consumer's application prior to July 1, 2021.

The Bureau is amending § 1026.43(e)(4) and related commentary pursuant to TILA section 129C(b)(3)(B)(ii), since the respective agencies directed to create their own definitions of qualified mortgages have done so and the Temporary GSE patch provisions will cease to be applicable on July 1, 2021.

Conforming Changes

As discussed above, the Bureau proposed, among other things, to revise the requirements in § 1026.43(e)(2)(v) that creditors consider and verify certain information; to remove the DTI limit in § 1026.43(e)(2)(vi); to remove references to appendix Q from § 1026.43; and to remove appendix Q from Regulation Z entirely. Accordingly, the Bureau proposed non-substantive conforming changes in certain provisions to reflect these proposed changes.

Specifically, the Bureau proposed to update comment 43(c)(7)-1 by removing the reference to the DTI limit in § 1026.43(e). The Bureau also proposed conforming changes to provisions related to small creditor QMs in § 1026.43(e)(5)(i) and to balloon-payment QMs in § 1026.43(f)(1). Both § 1026.43(e)(5) and (f)(1) currently provide that as part of the respective QM definitions, loans must comply with the requirements in existing § 1026.43(e)(2)(v) to consider and verify certain information. As discussed above, the Bureau proposed to reorganize and revise § 1026.43(e)(2)(v) in order to provide that creditors must consider DTI or residual income and to clarify the requirements for creditors to consider and verify income or assets, debts, and other information. The proposed conforming changes to § 1026.43(e)(5) and (f)(1) would generally have inserted the substantive requirements of existing § 1026.43(e)(2)(v) into § 1026.43(e)(5)(i) and (f)(1), respectively, and would have provided that loans under § 1026.43(e)(5) and (f) do not have to comply with revised § 1026.43(e)(2)(v) or (vi). However, the proposed conforming changes would not have inserted the requirement that creditors consider and verify income or assets, debts, and other information in accordance with appendix Q because the Bureau proposed to remove appendix Q from Regulation Z. The Bureau also proposed conforming changes to the related commentary.

The Bureau did not receive any comments on the proposed conforming changes. While the Bureau, in this final rule, is making some modifications to the proposal, none of these modifications affects the proposed conforming changes. Therefore, this final rule adopts the conforming changes to comment 43(c)(7)-1 and to the provisions related to small creditor QMs in § 1026.43(e)(5)(i) and balloon-payment QMs in § 1026.43(f)(1) as proposed.

Appendix Q to Part 1026—Standards for Determining Monthly Debt and Income

Appendix Q to part 1026 contains standards for calculating and verifying debt and income for purposes of determining whether a mortgage satisfies the 43 percent DTI limit for General QMs. As explained in the section-by-section analysis of § 1026.43(e)(2)(v)(B) above, the Bureau proposed to remove appendix Q from Regulation Z entirely in light of concerns from creditors and investors that its rigidity, ambiguity, and static nature result in standards that are both confusing and outdated. The Bureau sought comment on whether, instead of removing appendix Q entirely, it should retain appendix Q as an option for complying with the ATR/QM Rule's verification requirements.

Commenters generally supported removing appendix Q. Commenters stated that appendix Q's requirements to consider and verify income and debt are outdated, ambiguous, and inflexible. Commenters also stated that appendix Q is difficult for creditors to use for self-employed and gig economy consumers and in some cases has resulted in reduced access to credit. A consumer advocate, for example, stated that appendix Q consisted of “ossified and complex detail” and supported the Bureau's proposal to amend § 1026.43(e)(2)(v). These commenters generally supported replacing appendix Q with the provisions of § 1026.43(e)(2)(v) discussed above. In contrast, two industry commenters supported retaining appendix Q and suggested detailed edits to its provisions. However, both comment letters discussed such edits to appendix Q in the context of retaining a DTI limit within the General QM loan definition, which is not being adopted for the reasons discussed in part V above.

This final rule removes the appendix Q requirements from § 1026.43(e)(2)(v) and removes appendix Q from Regulation Z entirely, as the Bureau proposed. The Bureau determines that, due to the well-founded and consistent concerns articulated by stakeholders and described in detail in the General QM Proposal,^[348] appendix Q does not provide sufficient compliance certainty to creditors and does not provide flexibility to adapt to emerging issues with respect to the treatment of certain types of debt or income categories. The Bureau does not anticipate that removing appendix Q and using the new requirements of 1026.43(e)(2)(v) to consider and verify income, assets, debts, alimony, and child support will lead to higher risk loans obtaining QM status beyond loans that will receive such status from the removal of DTI limits as discussed in part C.4 above.

The Bureau recognizes that some findings in the Assessment Report suggest that the issues raised by creditors with respect to appendix Q do not appear to have had a substantial impact for certain loans. For example, although creditors have stated that it may be difficult to comply with certain appendix Q requirements for self-employed borrowers, the Assessment Report noted that application data indicated that the approval rates for non-high DTI, non-GSE eligible self-employed borrowers have decreased by only 2 percentage points since the January 2013 Final Rule became effective.^[349] The Bureau concludes, however, that this limited decrease in approvals for such applications does not undermine creditors' concerns that appendix Q's definitions of debt and income are rigid and difficult to apply and do not provide the level of compliance certainty that the Bureau anticipated in the January 2013 Final Rule. Additionally, the Assessment Report showed that about 40 percent of respondents to a lender survey indicated that they “often” or “sometimes” originate non-QM loans if the borrower cannot provide documentation required by appendix Q. The Bureau concluded in the Assessment Report that these results left open the possibility that appendix Q requirements may have had an impact on access to credit.^[350]

The Bureau declines to retain and revise appendix Q. As noted above, the Bureau concludes that appendix Q is inflexible, ambiguous and static, which results in standards that are both confusing and outdated. The Bureau concludes that it would be time- and resource-intensive to revise appendix Q in a manner to try to resolve these concerns. The Bureau therefore concludes that removing appendix Q entirely would be more efficient and practicable than retaining and revising it. The Bureau also does not anticipate a decrease in consumer protection as a result of removing appendix Q and adopting the provisions of 1026.43(e)(2)(v).

VII. Effective Date

A. The Bureau's Proposal

The Bureau proposed an effective date for a revised General QM loan definition of six months after publication in the Federal Register of a final rule. The Bureau further proposed that the revised regulations would apply to covered transactions for which creditors receive an application on or after that effective date. In the proposal, the Bureau tentatively determined that a six-month period between Federal Register publication of a final rule and the final rule's effective date would give creditors enough time to bring their systems into compliance with the revised regulations. The Bureau also stated it did not intend to issue a final rule amending the General QM loan definition early enough for it to take effect before April 1, 2021.

For the reasons described below, this final rule adopts an effective date of March 1, 2021, and a mandatory compliance date of July 1, 2021, resulting in an optional early compliance period between March 1, 2021 and July 1, 2021.^[351] This final rule adds new comment 43-2, which explains that, for transactions for which a creditor received the consumer's application on or after March 1, 2021 and prior to July 1, 2021, creditors have the option of using either the current General QM loan definition (i.e., the version in effect on February 26, 2021) or the revised General QM loan definition. Comment 43-2 also explains that, for transactions for which a creditor received the consumer's application on or after July 1, 2021, creditors seeking to originate General QMs are required to use the revised General QM loan definition. Comment 43-2 also specifies the meaning of “application” for these purposes.

B. Comments Received

The Bureau received several comments concerning the effective date and implementation period.^[352] Several industry commenters supported the proposal to link the effective date to the date the creditor received the consumer's application. One of these commenters stated that using the application date is preferable to using the consummation date because, while a loan is being processed and underwritten, the consummation date remains unknown, making it difficult for the creditor to anticipate which General QM loan definition to apply. Another commenter recommended clarifying that “application” has the same definition as under the Bureau's TILA-RESPA Integrated Disclosure Rule (TRID) because that definition is commonly used by the secondary market.

As discussed below, this final rule adds new comment 43-2 to clarify the operation of the final rule's effective date and mandatory compliance date, including clarifying that the effective date and mandatory compliance date are linked to the date the creditor received the consumer's application. Comment 43-2 also clarifies that, for transactions subject to TRID, creditors determine the date the creditor received the consumer's application, for purposes of this final rule's effective date and mandatory compliance date, in accordance with the TRID definition of application in § 1026.2(a)(3)(ii). This new comment also clarifies that, for transactions that are not subject to TRID, creditors can determine the date the creditor received the consumer's application, for purposes of this final rule's effective date and mandatory compliance date, in accordance with either § 1026.2(a)(3)(i) or (ii). The Bureau concludes that the clarifications provided in comment 43-2 will reduce uncertainty throughout the origination process.

Several industry commenters addressed the length of the implementation period. One industry commenter supported the Bureau's proposed effective date of six months after the final rule's publication in the Federal Register. Another industry commenter requested an implementation period extending to June 2021 and a 90-day grace period during which loans would still be reviewed for compliance with the revised definition but the Bureau would take no action to penalize simple mistakes and interpretation differences. The commenter stated that it took many months for small-to-mid-size creditors and investor channels to adjust to TRID.

Several industry commenters stated that an implementation period longer than six months is needed for creditors to work with vendors to develop and install software updates, conduct testing, update training policies, complete staff training, and educate consumers on product offerings. These commenters' recommendations for the length of the implementation period ranged from 12 months to 24 months. One of these industry commenters did not recommend a specific timeframe but stated that implementation would, on average, take from six months to 12 months depending on the size and complexity of both the vendor and creditor—or even up to 18 months depending on the overall complexity of the final rule, the timing of its effective date, and its impact on key operations such as underwriting. Another of these industry commenters requested at least one year for implementation while also stating that: Many creditors needed more than a year to implement the January 2013 Final Rule; a longer implementation period might avoid wasted time and expense if the regulation is changed again as a result of the 2020 elections; and small-to-mid-size creditors need more implementation time than larger creditors. Several industry commenters—including the commenter that generally supported the proposed effective date—stated that, in particular, the APR calculation for certain ARMs under proposed § 1026.43(e)(2)(vi) would require a significant (but unspecified) amount of implementation time.

As noted above, this final rule adopts a mandatory compliance date of July 1, 2021. This date is approximately six months after the date the Bureau expects this final rule to be published in the Federal Register. Therefore, this final rule adopts an implementation period similar to the six-month implementation period the Bureau proposed. The Bureau declines to adopt a longer implementation period because the Bureau concludes that a six-month period gives creditors and the secondary market enough time to prepare to comply with the amendments in this final rule. For example, with respect to the price-based thresholds in revised § 1026.43(e)(2)(vi), the Bureau understands that creditors currently calculate the APR and APOR for mortgage loans. With respect to the consider and verify requirements in revised § 1026.43(e)(2)(v), the Bureau understands that the revised consider requirements generally reflect existing market practices and that creditors currently use and are familiar with the verification standards in the verification safe harbor. The Bureau also concludes that this final rule is less complex to implement relative to other rules the Bureau has issued, such as the January 2013 Final Rule or TRID. The Bureau further concludes that it would be imprudent to provide a longer than necessary implementation period based on mere speculation that the Bureau might propose additional changes in the future. The Bureau declines to adopt a 90-day grace period or allow more implementation time for small-to-mid-size creditors because the Bureau concludes, for the reasons described above, that a six-month period gives all creditors and secondary market participants enough time to prepare to comply with the amendments in this final rule. The Bureau also concludes that establishing an optional early compliance period will facilitate implementation for all creditors, including small-to-mid-size creditors, for the reasons described below in the discussion of the final rule.

Several industry commenters also stated that this final rule's implementation period should generally account for other simultaneous challenges for creditors, including responding to the COVID-19 pandemic and its economic effects; transitioning indices away from LIBOR; ^[353] and implementing the GSEs' redesigned Uniform Residential Loan Application (URLA).^[354] One of those commenters specified that this final rule's implementation period should extend at least six months after the URLA's March 2021 mandatory compliance date. The Bureau concludes that a six-month implementation period gives creditors and secondary market participants enough time to prepare for the amendments in this final rule, even in light of these other commitments. As stated above, the Bureau concludes that this final rule is less complex to implement relative to other rules, such as the January 2013 Final Rule or TRID, and will not require significant changes to creditors' existing practices. Moreover, the Bureau concludes that current market conditions do not require a longer implementation period.

Several industry commenters responded to the General QM Proposal by requesting that the Bureau establish a period during which the Temporary GSE QM loan definition would remain in effect after the date creditors are required to transition from the current General QM loan definition to the revised General QM loan definition (Overlap Period). With respect to the length of the Overlap Period, commenters suggested periods between six months and one year. The Bureau also received several requests for an Overlap Period in response to the Extension Proposal, with commenters suggesting that the period last between four months and one year. The Bureau declines to adopt an Overlap Period in this final rule for the same reasons it declined to adopt an Overlap Period in the Extension Final Rule. In that final rule, the Bureau concluded that establishing an Overlap Period would keep the Temporary GSE QM loan definition in place longer than necessary to facilitate a smooth and orderly transition to a revised General QM loan definition and would prolong the negative effects of the Temporary GSE QM loan definition on the mortgage market.^[355]

In contrast with an Overlap Period, one group of industry commenters requested an optional early compliance period during which the revised General QM loan definition would become available, on an optional basis, before the date creditors are required to transition from the current General QM loan definition to the revised General QM loan definition. The group did not specify how much earlier, in its view, the Bureau should make the revised General QM loan definition available. As discussed below, the Bureau concludes that establishing an optional early compliance period will facilitate a smooth and orderly transition to a revised General QM loan definition without prolonging the negative effects of the Temporary GSE QM loan definition.

C. The Final Rule

For the reasons discussed below (and above in response to commenters), this final rule adopts an effective date of March 1, 2021, and a mandatory compliance date of July 1, 2021, resulting in an optional early compliance period between March 1, 2021 and July 1, 2021.^[356] This final rule adds new comment 43-2, which explains that, for transactions for which a creditor received the consumer's application on or after March 1, 2021, and prior to July 1, 2021, creditors seeking to originate General QMs have the option of complying with either the current General QM loan definition (i.e., the version in effect on February 26, 2021) or the revised General QM loan definition. This comment also explains that, for transactions for which a creditor received the consumer's application on or after July 1, 2021, creditors seeking to originate General QMs must use the revised General QM loan definition. Comment 43-2 also specifies the meaning of “application” for these purposes.

The Bureau also notes that the Temporary GSE QM loan definition will be available for transactions for which the creditor receives the consumer's application before July 1, 2021, unless the applicable GSEs ceases to operate under conservatorship before July 1, 2021.^[357] As noted above, the Extension Final Rule amended Regulation Z to replace the January 10, 2021 sunset date of the Temporary GSE QM loan definition with a provision stating that the Temporary GSE QM loan definition will be available only for covered transactions for which the creditor receives the consumer's application before the mandatory compliance date of final amendments to the General QM loan definition in Regulation Z. Under this final rule, which amends the General QM loan definition, that mandatory compliance date is July 1, 2021. The Extension Final Rule did not amend the conservatorship clause in § 1026.43(e)(4)(ii)(A). As a result, the Temporary GSE QM loan definition will be available for transactions for which the creditor receives the consumer's application before July 1, 2021, unless the applicable GSE ceases to operate under conservatorship before July 1, 2021.

Consistent with the practice of other agencies in similar contexts, the revised General QM loan definition will be incorporated into the Code of Federal Regulations on the March 1, 2021 effective date. Comment 43-2 clarifies that for transactions for which the creditor receives the application on or after March 1, 2021, but prior to July 1, 2021, the creditor has the option of complying either with Regulation Z (as interpreted by the commentary) as it is in effect (including the amendments set forth in this final rule) or as it was in effect on February 26, 2021, together with any amendments that become effective other than the amendments set forth in this final rule.^[358] The Bureau concludes that this final rule will reduce uncertainty throughout the origination process by linking the effective date and mandatory compliance date to the date the creditor received the consumer's application.

The applicability of this final rule's effective date and mandatory compliance date, as well as compliance with this final rule's revisions to Regulation Z, is determined on a loan-by-loan basis. For example, if a creditor receives an application for a given loan on March 1, 2021 March 1, 2021, and that loan satisfies the current General QM loan definition (including satisfying the 43 percent DTI limit), then the loan is eligible for General QM status—even if the loan does not satisfy the revised General QM loan definition (e.g., exceeds the applicable § 1026.43(e)(2)(vi) pricing threshold). Similarly, if a creditor receives an application for a different loan on March 1, 2021, and that loan satisfies the revised General QM loan definition (including satisfying the applicable § 1026.43(e)(2)(vi) pricing threshold), then the loan is eligible for General QM status—even if the loan does not satisfy the current General QM loan definition (e.g., exceeds the 43 percent DTI limit).

As discussed above, the Bureau concludes that a mandatory compliance date of July 1, 2021, will provide stakeholders with a sufficient amount of time—approximately six months—to prepare to implement the revised General QM loan definition. While the Bureau proposed an effective date that would vary based on the date of publication in the Federal Register, the Bureau concludes that using a date certain for the mandatory compliance date (July 1, 2021) will facilitate implementation of this final rule by allowing stakeholders to begin preparing to implement by a particular date (i.e., no later than July 1, 2021) as soon as the Bureau issues this final rule, rather than when the Federal Register publishes the final rule some days later.

The Bureau has decided to adopt an optional early compliance period starting on March 1, 2021 (i.e., to allow creditors to begin using the revised General QM loan definition for applications received on or after the March 1, 2021 effective date). In the General QM Proposal, the Bureau stated that it did not intend to issue a final rule early enough for it to take effect before April 1, 2021. With this statement, the Bureau sought to reassure creditors and other market participants that creditors seeking to originate General QMs would not be required to discontinue using the existing General QM loan definition or to implement the revised General QM loan definition before April 1, 2021.^[359] In the proposal, the Bureau expected that this would occur on the final rule's effective date, because the proposal did not provide for an optional early compliance period with a separate mandatory compliance date. In contrast, under this final rule, creditors may continue using the existing General QM loan definition or wait to implement the revised General QM loan definition, should they wish to do so, until the rule's mandatory compliance date, which is July 1, 2021. This mandatory compliance date of July 1, 2021 is consistent with the Bureau's expectation, at the proposal stage, that creditors seeking to originate General QMs would not be required to implement the revised General QM loan definition before April 1, 2021 (as creditors have the option of waiting until July 1, 2021).

The Bureau further concludes that the flexibility afforded under the optional early compliance period may help creditors implement the provisions of the final rule more quickly and easily. To the extent that large creditors are more likely to avail themselves of optional early compliance, the Bureau notes that small-to-mid-size correspondent lenders will also benefit, as they often wait for larger wholesale creditors to implement a rule before finalizing their own implementation strategy to ensure their systems are compatible with the wholesale creditors.

New comment 43-2 clarifies that, for transactions subject to TRID, creditors determine the date the creditor received the consumer's application, for purposes of this comment, in accordance with the TRID definition of application in § 1026.2(a)(3)(ii). This new comment also clarifies that, for transactions that are not subject to TRID, creditors can determine the date the creditor received the consumer's application, for purposes of this comment, in accordance with either § 1026.2(a)(3)(i) or (ii).

As discussed in the Extension Final Rule,^[360] Regulation Z contains two definitions of “application.” Section 1026.2(a)(3)(i) defines “application” as the submission of a consumer's financial information for the purposes of obtaining an extension of credit. This definition applies to all transactions covered by Regulation Z. Section 1026.2(a)(3)(ii) also contains a more specific definition of “application.” Under this definition, for transactions subject to § 1026.19(e), (f), or (g)—i.e., transactions subject to TRID—an application consists of the submission of the consumer's name, the consumer's income, the consumer's social security number to obtain a credit report, the property address, an estimate of the value of the property, and the mortgage loan amount sought. The more specific definition of application in § 1026.2(a)(3)(ii) applies not just for purposes of TRID, but extends to all transactions subject to TRID. Therefore, for transactions that are subject to the ATR/QM Rule and that are also subject to TRID, the Bureau concludes that the more specific definition applies for purposes of the ATR/QM Rule as well. However, for transactions that are subject to the ATR/QM Rule but that are not subject to TRID, the Bureau finds that there may be ambiguity as to when the creditor received the consumer's application for purposes of the effective date of the revised General QM loan definition, optional compliance provision, and mandatory compliance date.^[361] This potential ambiguity arises because the general definition of application in § 1026.2(a)(3)(i) is less precise than the TRID definition.

To address this potential ambiguity, new comment 43-2 clarifies that, for transactions that are not subject to TRID, creditors can determine the date the creditor received the consumer's application, for purposes of this comment, in accordance with either § 1026.2(a)(3)(i) or (ii). The Bureau concludes that this clarification is appropriate because it will facilitate compliance with this final rule by reducing uncertainty throughout the origination process.

VIII. Dodd-Frank Act Section 1022(b) Analysis

A. Overview

As discussed above, this final rule amends the General QM loan definition to, among other things, remove the specific DTI limit and add pricing thresholds. In developing this final rule, the Bureau has considered the potential benefits, costs, and impacts as required by section 1022(b)(2)(A) of the Dodd-Frank Act. Specifically, section 1022(b)(2)(A) of the Dodd-Frank Act requires the Bureau to consider the potential benefits and costs of a regulation to consumers and covered persons, including the potential reduction of access by consumers to consumer financial products or services, the impact on depository institutions and credit unions with $10 billion or less in total assets as described in section 1026 of the Dodd-Frank Act, and the impact on consumers in rural areas. The Bureau consulted with appropriate prudential regulators and other Federal agencies regarding the consistency of this final rule with prudential, market, or systemic objectives administered by such agencies as required by section 1022(b)(2)(B) of the Dodd-Frank Act.

1. Data and Evidence

The discussion in these impact analyses relies on data from a range of sources. These include data collected or developed by the Bureau, including HMDA ^[362] and NMDB ^[363] data, as well as data obtained from industry, other regulatory agencies, and other publicly available sources. The Bureau also conducted the Assessment and issued the Assessment Report as required under section 1022(d) of the Dodd-Frank Act. The Assessment Report provides quantitative and qualitative information on questions relevant to this final rule, including the extent to which DTI ratios are probative of a consumer's ability to repay, the effect of rebuttable presumption status relative to safe harbor status on access to credit, and the effect of QM status relative to non-QM status on access to credit. Consultations with other regulatory agencies, industry, and research organizations inform the Bureau's impact analyses.

The data the Bureau relied upon provided detailed information on the number, characteristics, pricing, and performance of mortgage loans originated in recent years. As discussed above, commenters provided some supplemental data and estimates with more information relevant to pricing and APR calculations (particularly PMI costs) for originations before 2018. PMI costs are an important component of APRs, particularly for loans with smaller down payments, and thus should be included or estimated in calculations of rate spreads relative to APOR. The data provided by commenters show a strong positive relationship between rate spread over APOR and delinquency rates, similar to the relationship shown in the Bureau's analyses of 2002-2008 data and 2018 data.

The data do not provide information on creditor costs. As a result, analyses of any impacts of this final rule on creditor costs, particularly realized costs of implementing underwriting criteria or potential costs from legal liability, are based on more qualitative information. Similarly, estimates of any changes in burden on consumers resulting from increased or decreased verification requirements are based on qualitative information.

Finally, a group of consumer advocate commenters submitted a joint letter arguing that because the mortgage finance market is in flux, any assumptions made regarding the impact of pricing as an adequate substitute for more direct measures of ability to repay are rendered uncertain by the current economic conditions, and thus the Bureau should refrain from revising the General QM definition. In the proposal, the Bureau acknowledged the importance of economic disruptions and mortgage market changes due to the COVID-19 pandemic. However, the Bureau did not receive data or evidence from commenters that would lead it to anticipate that market changes or other circumstances will significantly alter its estimates of the benefits and costs of this final rule. These commenters also stated that the Bureau must fulfill its statutory obligation “to study ability-to-repay” before amending the General QM definition. However, the Bureau has already done so by completing the Assessment Report and through its monitoring of the performance of mortgage loans and the availability of mortgage credit.

Description of the Baseline

The Bureau considers the benefits, costs, and impacts of this final rule against the baseline in which the Bureau takes no action and the Temporary GSE QM loan definition expires when the GSEs cease to operate under conservatorship. Under this final rule, creditors that wish to originate General QMs will be required to comply with the amended General QM loan definition either at the time or after the Temporary GSE QM loan definition expires, depending on whether the GSEs remain in conservatorship on the mandatory compliance date of this final rule. As a result, this final rule's direct market impacts are considered relative to a baseline in which the Temporary GSE QM has expired and no changes have been made to the General QM loan definition. While there is not a fixed date on which the Temporary GSE QM loan definition will expire in the absence of this final rule, the Bureau anticipates that the GSEs will cease to operate under conservatorship in the foreseeable future and the baseline will occur at that time. Unless described otherwise, estimated loan counts under the baseline, final rule, and alternatives are annual estimates.

Under the baseline, conventional loans could receive QM status under the Bureau's rules only by underwriting according to the General QM requirements, Small Creditor QM requirements, Balloon Payment QM requirements, or the expanded portfolio QM amendments created by the 2018 Economic Growth, Regulatory Relief, and Consumer Protection Act. The General QM loan definition, which would be the only type of QM available to larger creditors for conventional loans, requires that consumers' DTI ratio not exceed 43 percent and requires creditors to determine debt and income in accordance with the standards in appendix Q.

The Bureau anticipates that, under the baseline in which the Temporary GSE QM loan definition expires, there are two main types of conventional loans that would be affected: Over-43-Percent-DTI ^[364] GSE loans and GSE-eligible loans without appendix Q-required documentation. These loans are currently originated as QMs due to the Temporary GSE QM loan definition but would not be originated as General QMs, and may not be originated at all, if the Temporary GSE QM loan definition were to expire without this final rule's amendments to the General QM loan definition. This section 1022 analysis refers to these loans as potentially displaced loans.

The proposal's analysis of the potential market impact of the Temporary GSE QM loan definition's expiration cited data and analysis from the Bureau's ANPR, as described below. None of the comments on the proposal challenged the data or analysis from the ANPR or the proposal related to the potential market impacts of the Temporary GSE QM loan definition's expiration. The Bureau concludes that the data and analysis in the proposal and ANPR provide a well-supported estimate of the potential impact of the Temporary GSE QM loan definition's expiration for this final rule.

Over-43-Percent-DTI GSE Loans. The ANPR provided an estimate of the number of loans potentially affected by the expiration of the Temporary GSE QM loan definition.^[365] In providing the estimate, the ANPR focused on loans that fall within the Temporary GSE QM loan definition but not the General QM loan definition because they have DTI ratios above 43 percent. This final rule refers to these loans as Over-43-Percent-DTI GSE loans. Based on NMDB data, the Bureau estimated that there were approximately 6.01 million closed-end first-lien residential mortgage originations in the United States in 2018.^[366] Based on supplemental data provided by the FHFA, the Bureau estimated that the GSEs purchased or guaranteed 52 percent—roughly 3.12 million—of those loans.^[367] Of those 3.12 million loans, the Bureau estimated that 31 percent—approximately 957,000 loans—had DTI ratios greater than 43 percent.^[368] Thus, the Bureau estimated that, as a result of the General QM loan definition's 43 percent DTI limit, approximately 957,000 loans—16 percent of all closed-end first-lien residential mortgage originations in 2018—were Over-43-Percent-DTI GSE loans.^[369] This estimate does not include Temporary GSE QMs that were eligible for purchase by the GSEs but were not sold to the GSEs.

Loans Without Appendix Q-Required Documentation That Are Otherwise GSE-Eligible. In addition to Over-43-Percent-DTI GSE loans, the Bureau noted that an additional, smaller number of Temporary GSE QMs with DTI ratios of 43 percent or less, when calculated using GSE underwriting guides, may not fall within the General QM loan definition because their method of verifying income or debt is incompatible with appendix Q.^[370] These loans would also likely be affected once the Temporary GSE QM loan definition expires. The Bureau understands, from extensive public feedback and its own experience, that appendix Q does not specifically address whether and how to verify certain forms of income. The Bureau understands these concerns are particularly acute for self-employed consumers, consumers with part-time employment, and consumers with irregular or unusual income streams.^[371] As a result, these consumers' access to credit may be affected if the Temporary GSE QM loan definition were to expire without amendments to the General QM loan definition.

The Bureau's analysis of the market under the baseline focuses on Over-43-Percent-DTI GSE loans because the Bureau estimates that most potentially displaced loans are Over-43-Percent-DTI GSE loans. The Bureau also lacks the loan-level documentation and underwriting data necessary to estimate with precision the number of potentially displaced loans that do not fall within the other General QM requirements and are not Over-43-Percent-DTI GSE loans. However, the Assessment did not find evidence of substantial numbers of loans in the non-GSE-eligible jumbo market being displaced when appendix Q verification requirements became effective in 2014.^[372] Further, the Assessment Report found evidence of only a limited reduction in the approval rate of self-employed applicants for non-GSE eligible mortgages.^[373] Based on this evidence, along with qualitative comparisons of GSE and appendix Q verification requirements and available data on the prevalence of borrowers with non-traditional or difficult-to-document income (e.g., self-employed borrowers, retired borrowers, those with irregular income streams), the Bureau estimates this second category of potentially displaced loans is considerably less numerous than the category of Over-43-Percent-DTI GSE loans.

Additional Effects on Loans Not Displaced. While the most significant market effects under the baseline are displaced loans, loans that continue to be originated as QMs after the expiration of the Temporary GSE QM loan definition would also be affected. After the expiration date, all loans with DTI ratios at or below 43 percent which are or would have been purchased and guaranteed as GSE loans under the Temporary GSE QM loan definition—approximately 2.16 million loans in 2018—and that continue to be originated as General QMs after the provision expires would be required to verify income and debts according to appendix Q, rather than only according to GSE guidelines. Given the concerns raised about appendix Q's ambiguity and lack of flexibility, this would likely entail both increased documentation burden for some consumers as well as increased costs or time-to-origination for creditors on some loans.^[374]

B. Benefits and Costs to Covered Persons and Consumers

1. Benefits to Consumers

The primary benefit to consumers of this final rule is increased access to credit, largely through the expanded availability of Over-43-Percent-DTI conventional QMs. Given the large number of consumers who obtain Over-43-Percent-DTI GSE loans rather than available alternatives, including loans from the private non-QM market and FHA loans, such Over-43-Percent-DTI conventional QMs may be preferred due to their pricing, underwriting requirements, or other features. Based on HMDA data, the Bureau estimates that 959,000 Over-43-Percent-DTI conventional loans in 2018 would fall outside the QM definitions under the baseline, but fall within this final rule's amended General QM loan definition.^[375] In addition, some consumers who would have been limited in the amount they could borrow due to the DTI limit under the baseline will likely be able to obtain larger mortgages at higher DTI levels.

Under the baseline, a sizeable share of potentially displaced Over-43-Percent-DTI GSE loans may instead be originated as FHA loans. Thus, under this final rule, any price advantage of GSE or other conventional QMs over FHA loans will be a realized benefit to consumers. Based on the Bureau's analysis of 2018 HMDA data, FHA loans comparable to the loans received by Over-43-Percent-DTI GSE borrowers, based on loan purpose, credit score, and combined LTV ratio, on average have $3,000 to $5,000 higher upfront total loan costs at origination. APRs provide an alternative, annualized measure of costs over the life of a loan. FHA borrowers typically pay different APRs, which can be higher or lower than APRs for GSE loans depending on a borrower's credit score and LTV ratio. Borrowers with credit scores at or above 720 pay an APR 30 to 60 basis points higher than borrowers of comparable GSE loans, leading to higher monthly payments over the life of the loan. However, FHA borrowers with credit scores below 680 and combined LTV ratios exceeding 85 percent pay an APR 20 to 40 basis points lower than borrowers of comparable GSE loans, leading to lower monthly payments over the life of the loan.^[376] For a loan size of $250,000, these APR differences amount to $2,800 to $5,600 in additional total monthly payments over the first five years of mortgage payments for borrowers with credit scores above 720, and $1,900 to $3,800 in reduced total monthly payments over five years for borrowers with credit scores below 680 and LTV ratios exceeding 85 percent.^[377] Thus, all FHA borrowers are likely to pay higher costs at origination, while some pay higher monthly mortgage payments, and others pay lower monthly mortgage payments. Assuming for comparison that all 959,000 additional loans falling within the amended General QM loan definition would be made as FHA loans in the absence of this final rule, the average of the upfront pricing estimates results in total savings for consumers of roughly $4 billion per year on upfront costs.^[378] The total savings or costs over the life of the loan based on APR differences would vary substantially across borrowers depending on credit scores, LTV ratios, and length of time holding the mortgage. While this comparison assumed all potentially displaced loans would be made as FHA loans, higher costs (either upfront or in monthly payments) are likely to prevent some borrowers from obtaining loans at all.

In the absence of this final rule, some of these potentially displaced consumers, particularly those with higher credit scores and the resources to make larger down payments, likely would be able to obtain credit in the non-GSE private market at a cost comparable to or slightly higher than the costs for GSE loans, but below the cost of an FHA loan. As a result, the above cost comparisons between GSE and FHA loans provide an estimated upper bound on pricing benefits to consumers of this final rule. However, under the baseline, some potentially displaced consumers may not obtain loans, and thus will experience benefits of credit access under this final rule. As discussed above, the Assessment Report found that the January 2013 Final Rule eliminated between 63 and 70 percent of home purchase loans with DTI ratios above 43 percent that were not Temporary GSE QMs.^[379]

This final rule will also benefit those consumers with incomes difficult to verify using appendix Q to obtain General QM status, as this final rule's General QM amendments will no longer require the use of appendix Q for verification of income. Under this final rule—as under the current rule—creditors will be required to verify income and assets in accordance with § 1026.43(c)(4) and debt obligations, alimony, and child support in accordance with § 1026.43(c)(3). This final rule also states that a creditor complies with the General QM requirement to verify income, assets, debt obligations, alimony, and child support if it complies with verification requirements in standards the Bureau specifies. The greater flexibility of verification standards allowed under this final rule is likely to reduce effort and costs for these consumers, and in the most difficult cases in which consumers' documentation cannot satisfy appendix Q, this final rule will allow consumers to obtain General QMs rather than potential FHA or non-QM alternatives. These consumers—likely including self-employed borrowers and those with non-traditional forms of income—will likely benefit from cost savings under this final rule, similar to those for High-DTI consumers discussed above.

Finally, as noted below under “Costs to consumers,” the Bureau estimates that 25,000 low-DTI conventional loans which are QM under the baseline will fall outside the amended QM definition under this final rule, due to exceeding the pricing thresholds in § 1026.43(e)(2)(vi). If consumers of such loans are able to obtain non-QM loans with the amended General QM loan definition in place, they will gain the benefit of the ability-to-repay causes of action and defenses against foreclosure. However, some of these consumers may instead obtain FHA loans with QM status.

2. Benefits to Covered Persons

This final rule's primary benefit to covered persons, specifically mortgage creditors, is the expanded profits from originating Over-43-Percent DTI conventional QMs. Under the baseline, creditors would be unable to originate such loans under the Temporary GSE QM loan definition and would instead have to originate loans with comparable DTI ratios as FHA, Small Creditor QM, or non-QM loans, or originate at lower DTI ratios as conventional General QMs. Creditors' current preference for originating large numbers of Over-43-Percent-DTI Temporary GSE QMs likely reflects advantages in a combination of costs or guarantee fees (particularly relative to FHA loans), liquidity (particularly relative to Small Creditor QM), or litigation and credit risk (particularly relative to non-QM loans). Moreover, QMs—including Temporary GSE QMs—are exempt from the Dodd-Frank Act risk retention requirement whereby creditors that securitize mortgage loans are required to retain at least 5 percent of the credit risk of the security, which adds significant cost. As a result, this final rule conveys benefits to mortgage creditors originating Over-43-Percent-DTI conventional QMs on each of these dimensions.

In addition, for those lower-DTI GSE loans that could satisfy General QM requirements, creditors may realize cost savings from underwriting loans using the more flexible verification standards allowed under this final rule compared with using appendix Q. Under this final rule, creditors will be required to consider DTI or residual income in addition to income or assets other than the value of the dwelling and debts but will not need to comply with the appendix Q standards required for General QMs under the baseline. For conventional consumers unable to provide documentation compatible with appendix Q, this final rule allows such loans to continue receiving QM status, providing comparable benefits to creditors as described for Over-43-Percent-DTI GSE loans above.

Finally, creditors with business models that rely most heavily on originating Over-43-Percent-DTI GSE loans will likely see a competitive benefit from the continued ability to originate such loans as General QMs. Under the baseline, creditors that primarily originate FHA or private non-QM loans likely would have gained market share at the expense of creditors originating many Over-43-Percent-DTI GSE loans. The final rule will prevent this shift from occurring, which is effectively a transfer in market share to the creditors originating many Over-43-Percent-DTI GSE loans.

3. Costs to Consumers

As discussed above, relative to the baseline, the Bureau estimates that 959,000 additional Over-43-Percent-DTI loans could be originated as General QMs under this final rule. Some of these loans would have been non-QM loans (if originated) under the baseline. As a result, this final rule is likely to increase the number of consumers who become delinquent on QMs, meaning an increase in consumers with delinquent loans who do not have the benefit of the ability-to-repay causes of action and defenses against foreclosure.

Tables 5 and 6 in part V provide historical early delinquency rates for loans under different combinations of DTI ratio and rate spread. Under this final rule, conventional loans originated with rate spreads below 2.25 percentage points and DTI above 43 percent will newly fall within the amended General QM loan definition relative to the baseline. Based on the number and characteristics of 2018 HMDA originations, the Bureau estimates that between 8,000 and 58,000 additional General QMs annually could become delinquent within two years of origination, based on the observed early delinquencies from Table 6 (2018) and Table 5 (2002-2008), respectively.^[380] Further, consumers who would have been limited in the amount they could borrow due to the DTI limit under the baseline may obtain larger mortgages at higher DTI levels, further increasing the expected number of delinquencies. However, given that many of these loans may have been originated as FHA (or other non-General QM) loans under the baseline, the increase in delinquent loans held by consumers without the ability-to-repay causes of action and defenses against foreclosure is likely smaller than the upper bound estimates cited above.

For the estimated 25,000 consumers obtaining low-DTI General QM or Temporary GSE QMs priced 2.25 percentage points or more above APOR under the baseline, the amended General QM loan definition may restrict access to conventional QM credit. There are several possible outcomes for these consumers. Many may instead obtain FHA loans, likely paying higher total loan costs, as discussed above. Others may be able to obtain General QMs priced below 2.25 percentage points over APOR due to creditor responses to this final rule or obtain loans under the Small Creditor QM definition. However, some consumers may not be able to obtain a mortgage at all.

In addition, this final rule reduces the scope of the non-QM market relative to the baseline, which could slow the development of new non-QM loan products which may have become available under the baseline. To the extent that some consumers would prefer some of these products to conventional QMs due to pricing, verification flexibility, or other advantages, the delay of their development will be a cost to consumers of this final rule.

4. Costs to Covered Persons

For creditors retaining the credit risk of their General QM mortgages (e.g., portfolio loans and private securitizations), an increase in Over-43-Percent-DTI General QM originations may lead to increased risk of credit losses. However, some of this increased risk may be offset by lender pricing responses. Further, on average the effects on portfolio lenders may be small. Creditors that hold loans on portfolio have an incentive to verify ability to repay regardless of liability under the ATR provisions, because they hold the credit risk. While portfolio lenders (or those that manage the portfolios) may recognize and respond to this incentive to different degrees, this final rule is likely on average to cause a small increase in the willingness of these creditors to originate loans with a greater risk of default and credit losses, such as certain loans with high DTI ratios. The credit losses to investors in private securitizations are harder to predict. In general, these losses will depend on the scrutiny that investors are willing and able to give to the non-QM loans under the baseline that become QMs (with high DTI ratios) under this final rule. It is possible, however, that the reduction in liability under the ATR provisions will lead to securitizations with more loans that have a greater risk of default and credit losses.

In addition, creditors will generally no longer be able to originate low-DTI conventional loans priced 2.25 percentage points or higher above APOR as General QMs under this final rule.^[381] Creditors may be able to originate some of these loans at prices below 2.25 percentage points above APOR or as non-QM loans or other types of QMs, but in these cases may pay higher costs or receive lower revenues relative to under the baseline. If creditors are unable to originate such loans at all, they will see a larger reduction in revenue.

This final rule also generates what are effectively transfers between creditors relative to the baseline, reflecting reduced loan origination volume for creditors that primarily originate FHA or private non-QM loans and increased origination volume for creditors that primarily originate conventional QMs. Business models vary substantially within market segments, with portfolio lenders and lenders originating non-QM loans most likely to forgo market share gains possible under the baseline, while GSE-focused bank and non-bank creditors are likely to maintain market share that might be lost in the absence of this final rule.

5. Other Benefits and Costs

This final rule may limit the development of the secondary market for non-QM mortgage loan securities. Under the baseline, loans that do not fit within General QM requirements represent a potential new market for non-QM loan securitizations. Thus, this final rule will reduce the scope of the potential non-QM loan market, likely lowering total profits and revenues for participants in the private secondary market. This will effectively be a transfer from these non-QM loan secondary market participants to participants in the agency or other QM secondary markets.

6. Consideration of Alternatives

The Bureau considered potential alternatives to this final rule, including maintaining the General QM loan definition's DTI limit but at a higher level, for example, 45 or 50 percent. The Bureau estimates the effects of such alternatives relative to this final rule, assuming no change in consumer or creditor behavior. For an alternative General QM loan definition with a DTI limit of 45 percent, the Bureau estimates that 673,000 fewer loans would have been General QM due to DTI ratios over 45 percent, while 28,000 additional loans with rate spreads above the final rule's QM pricing thresholds would have newly fit within the General QM loan definition due to DTI ratios at or below 45 percent. For an alternative DTI limit of 50 percent, the Bureau estimates 51,000 fewer loans would have fit within the General QM loan definition due to DTI ratios over 50 percent, while 35,000 additional loans with rate spreads above the final rule's QM pricing thresholds would have newly fit within the General QM loan definition due to DTI ratios at or below 50 percent.

In addition to these effects on the composition of loans within the General QM loan definition, the Bureau uses the historical delinquency rates from Tables 5 and 6 in part V to estimate the number of loans that would have been expected to become delinquent within the General QM loan definition relative to this final rule. The Bureau estimates that under an alternative DTI limit of 45 percent, 4,000 to 37,000 fewer General QMs would have become delinquent relative to this final rule, based on delinquency rates for 2018 and 2002-2008 originations respectively. Under an alternative DTI limit of 50 percent, the Bureau estimates approximately 1,000 additional General QMs would have become delinquent relative to this final rule, due to loans priced 2.25 percentage points or more above APOR gaining QM status.

For an alternative DTI limit of 45 percent, these estimates collectively indicate that substantially fewer loans would have fit within the General QM loan definition relative to this final rule, which would also have reduced the number of General QMs becoming delinquent. By contrast, the estimates indicate that an alternative DTI limit of 50 percent would have led to a comparable number of General QMs relative to this final rule, both overall and among those that would have become delinquent. However, consumer and creditor responses to such alternatives, such as reducing loan amounts to lower DTI ratios, could have increased the number of loans that would have fit within the alternative General QM loan definitions relative to this final rule.

The Bureau considered other potential alternatives to the proposed rule, including imposing a DTI limit only for loans above a certain pricing threshold, for example a DTI limit of 50 percent for loans with rate spreads at or above 1 percentage point. Such an alternative would have functioned as a hybrid of this final rule and an alternative which maintains a DTI limit at a higher level, 50 percent in the case of this example. As a result, the number of loans fitting within the General QM loan definition would have generally been between the Bureau's estimates for this final rule and its estimates for the corresponding alternative which would have maintained the higher DTI limit. Thus, this hybrid approach would have brought fewer loans within the General QM loan definition compared to this final rule but more loans within the General QM loan definition compared to the alternative DTI limit of 50 percent, both overall and among loans that would have become delinquent.

C. Potential Impact on Depository Institutions and Credit Unions With $10 Billion or Less in Total Assets, as Described in Section 1026

This final rule's expected impact on depository institutions and credit unions that are also creditors making covered loans (depository creditors) with $10 billion or less in total assets is similar to the expected impact on larger depository creditors and on non-depository creditors. As discussed in part VIII.B.4 (Costs to Covered Persons), depository creditors originating portfolio loans may forgo potential market share gains that would occur under the baseline. In addition, depository creditors with $10 billion or less in total assets that originate portfolio loans can originate Over-43-Percent-DTI Small Creditor QMs under the rule. These depository creditors may currently rely less on the Temporary GSE QM loan definition for originating Over-43-Percent-DTI loans. If the expiration of the Temporary GSE QM loan definition in the absence of this final rule would confer a competitive advantage to these small creditors in their origination of Over-43-Percent-DTI loans, this final rule will offset this outcome.

Conversely, those small depository creditors that primarily rely on the GSEs as a secondary market outlet because they do not have the capacity to hold numerous loans on portfolio or the infrastructure or scale to securitize loans may continue to benefit from the ability to make Over-43-Percent-DTI GSE loans as QMs. Under the baseline, these creditors would be limited to originating GSE loans as QMs only with DTI ratios at or below 43 percent under the current General QM loan definition. These creditors may also originate FHA, VA, or USDA loans or non-QM loans for private securitizations, likely at a higher cost relative to originating Temporary GSE QMs. This final rule will allow these creditors to originate more GSE loans under the General QM loan definition and have a lower cost of origination relative to the baseline.^[382]

D. Potential Impact on Rural Areas

This final rule's expected impact on rural areas is similar to the expected impact on non-rural areas. Based on 2018 HMDA data, the Bureau estimates that Over-43-Percent-DTI conventional purchase mortgages originated for homes in rural areas are approximately as likely to be reported as initially sold to the GSEs (52.5 percent) as loans in non-rural areas (52 percent).^[383] In addition, the Bureau estimates that in 2018, 94.6 percent of conventional purchase loans originated for homes in rural areas would have been QMs under this final rule, similar to the Bureau's estimate for all conventional purchase loans in rural and non-rural areas (96.3 percent).^[384]

IX. Regulatory Flexibility Act Analysis

The Regulatory Flexibility Act (RFA), as amended by the Small Business Regulatory Enforcement Fairness Act of 1996, requires each agency to consider the potential impact of its regulations on small entities, including small businesses, small governmental units, and small not-for-profit organizations. The RFA defines a “small business” as a business that meets the size standard developed by the Small Business Administration pursuant to the Small Business Act.^[385]

The RFA generally requires an agency to conduct an initial regulatory flexibility analysis (IRFA) and a final regulatory flexibility analysis (FRFA) of any rule subject to notice-and-comment rulemaking requirements, unless the agency certifies that the rule would not have a significant economic impact on a substantial number of small entities (SISNOSE).^[386] The Bureau also is subject to certain additional procedures under the RFA involving the convening of a panel to consult with small business representatives before proposing a rule for which an IRFA is required.^[387]

In the proposal, the Bureau certified that an IRFA was not required because the proposal, if adopted, would not have a SISNOSE. The Bureau did not receive comments on its analysis of the impact of the proposal on small entities. As the below analysis makes clear, relative to the baseline, this final rule has only one sizeable adverse effect. Certain loans with DTI ratios under 43 percent that would otherwise be originated as rebuttable presumption QMs under the baseline will be non-QM loans under this final rule. This final rule will also have a number of more minor effects on small entities which are not quantified in this analysis, including adjustments to the APR calculation used for certain ARMs when determining QM status and amendments to the Rule's requirements to consider and verify income, assets, debt obligations, alimony, and child support. The Bureau expects only small increases or decreases in burden from these more minor effects.

The analysis divides potential originations into different categories and considers whether this final rule has any adverse impact on originations relative to the baseline. Note that under the baseline, the category of Temporary GSE QMs no longer exists. The Bureau has identified five categories of small entities that may be subject to this final rule: Commercial banks, savings institutions and credit unions (NAICS 522110, 522120, and 522130) with assets at or below $600 million; mortgage brokers (NAICS 522310) with average annual receipts at or below $8 million; and mortgage companies (NAICS 522292 and 522298) with average annual receipts at or below $41.5 million. As discussed further below, the Bureau relies primarily on 2018 HMDA data for the analysis.^[388]

Type I: First Liens That Are Not Small Loans, DTI Is Over 43 Percent

Under the baseline, small entities cannot originate Type I loans as safe harbor or rebuttable presumption QMs unless they are also small creditors and comply with the additional requirements of the small creditor QM category. Neither the removal of DTI requirements nor the addition of the pricing conditions has an adverse impact on the ability of small entities to originate these loans.

Type II: First Liens That Are Not Small Loans, DTI Is 43 Percent or Under

Under the baseline, small entities can originate these loans as either safe harbor QMs or rebuttable presumption QMs, depending on pricing. The removal of DTI requirements has no adverse impact on the ability of small entities to originate these loans. The addition of the pricing conditions has no adverse impact on the ability of small creditors to originate these loans as safe harbor QMs: A loan with APR within 1.5 percentage points of APOR that can be originated as a safe harbor QM under the baseline can be originated as a safe harbor QM under the pricing conditions of this final rule. Similarly, the addition of the pricing conditions has no adverse impact on the ability of small creditors to originate rebuttable presumption QMs with APR between 1.5 percentage points and 2.25 percentage points over APOR. The addition of the pricing conditions will, however, prevent small creditors from originating rebuttable presumption QMs with APR 2.25 percentage points or more over APOR. In the SISNOSE analysis below, the Bureau conservatively assumes that none of these loans will be originated.

Type III: First-Liens That Are Small Loans

Under the baseline, small entities can originate these loans as General QMs if they have DTI ratios at or below the DTI limit of 43 percent. This final rule's amended General QM loan definition preserves QM status for some smaller, low-DTI loans priced 2.25 percentage points or more over APOR. Specifically, loans under $66,156 with APR less than 6.5 percentage points over APOR and loans under $110,260 with APR less than 3.5 percentage points over APOR can be originated as General QMs, assuming they meet all other General QM requirements.^[389] This final rule will prevent small creditors from originating smaller, low-DTI loans with APR at or above these higher thresholds as General QMs. For the SISNOSE analysis below, the Bureau conservatively assumes that none of these loans will be originated.

Type IV: Closed-End Subordinate-Liens

Under the baseline, small entities can originate these loans as General QMs if they have DTI ratios at or below the DTI limit of 43 percent. This final rule's amended General QM loan definition creates new pricing thresholds for subordinate-lien originations. Subordinate-lien loans under $66,156 with APR less than 6.5 percentage points over APOR and larger subordinate-lien loans with APR less than 3.5 percentage points over APOR can be originated as General QMs, assuming they meet all other General QM requirements. The final rule will prevent small creditors from originating low-DTI, subordinate-lien loans with APR at or above these thresholds as General QMs. For the SISNOSE analysis below, the Bureau conservatively assumes that none of these loans will be originated.

Analysis

For purposes of this analysis, the Bureau assumes that average annual receipts for small entities is proportional to mortgage loan origination volume. The Bureau further assumes that a small entity experiences a significant negative effect from this final rule if it will cause a reduction in origination volume of over 2 percent. Using the 2018 HMDA data, the Bureau estimates that if none of the Type II, III, or IV loans adversely affected were originated, 97 small entities would experience a loss of over 2 percent in mortgage loan origination volume. Thus, there are at most 97 small entities that experience a significant adverse economic impact. The Bureau estimates that there are 2,027 small entities in the HMDA data. Ninety-seven is not a substantial number relative to 2,027.

The Bureau recognizes that there are small entities that originate mortgage credit that do not report HMDA data. The Bureau has no reason to expect, however, that small entities that originate mortgage credit that do not report HMDA data would be affected differently than small HMDA reporters by the final rule. In other words, the Bureau expects that including HMDA non-reporters in the analysis would increase the number of small entities that will experience a loss of over 2 percent in mortgage loan origination volume and the number of relevant small entities by the same proportion. Thus, the overall number of small entities that will experience a significant adverse economic impact will not be a substantial number of the overall number of small entities that originate mortgage credit.

Accordingly, the Director certifies that this final rule will not have a significant economic impact on a substantial number of small entities.

X. Paperwork Reduction Act

Under the Paperwork Reduction Act of 1995 (PRA),^[390] Federal agencies are generally required to seek, prior to implementation, approval from the Office of Management and Budget (OMB) for information collection requirements. Under the PRA, the Bureau may not conduct or sponsor, and, notwithstanding any other provision of law, a person is not required to respond to, an information collection unless the information collection displays a valid control number assigned by OMB.

The Bureau has determined that this final rule does not contain any new or substantively revised information collection requirements other than those previously approved by OMB under OMB control number 3170-0015. This final rule amends 12 CFR part 1026 (Regulation Z), which implements TILA. OMB control number 3170-0015 is the Bureau's OMB control number for Regulation Z.

XI. Congressional Review Act

Pursuant to the Congressional Review Act,^[391] the Bureau will submit a report containing this rule and other required information to the U.S. Senate, the U.S. House of Representatives, and the Comptroller General of the United States at least 60 days prior to the rule's published effective date. The Office of Information and Regulatory Affairs has designated this rule as a “major rule” as defined by 5 U.S.C. 804(2).

XII. Signing Authority

The Director of the Bureau, Kathleen L. Kraninger, having reviewed and approved this document, is delegating the authority to electronically sign this document to Grace Feola, a Bureau Federal Register Liaison, for purposes of publication in the Federal Register.

List of Subjects in 12 CFR Part 1026

• Advertising
• Banks
• Banking
• Consumer protection
• Credit
• Credit unions
• Mortgages
• National banks
• Reporting and recordkeeping requirements
• Savings associations
• Truth-in-lending

Authority and Issuance

For the reasons set forth above, the Bureau amends Regulation Z, 12 CFR part 1026, as set forth below:

PART 1026—TRUTH IN LENDING (REGULATION Z)

1. The authority citation for part 1026 continues to read as follows:

Authority: 12 U.S.C. 2601, 2603-2605, 2607, 2609, 2617, 3353, 5511, 5512, 5532, 5581; 15 U.S.C. 1601 et seq.

Subpart E—Special Rules for Certain Home Mortgage Transactions

2. Amend § 1026.43 by revising paragraphs (b)(4), (e)(2)(v) and (vi), (e)(4), (e)(5)(i)(A) and (B), and (f)(1)(i) and (iii) to read as follows:

§ 1026.43

Minimum standards for transactions secured by a dwelling.

* * * * *

(b) * * *

(4) Higher-priced covered transaction means a covered transaction with an annual percentage rate that exceeds the average prime offer rate for a comparable transaction as of the date the interest rate is set by 1.5 or more percentage points for a first-lien covered transaction, other than a qualified mortgage under paragraph (e)(5), (e)(6), or (f) of this section; by 3.5 or more percentage points for a first-lien covered transaction that is a qualified mortgage under paragraph (e)(5), (e)(6), or (f) of this section; or by 3.5 or more percentage points for a subordinate-lien covered transaction. For purposes of a qualified mortgage under paragraph (e)(2) of this section, for a loan for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due, the creditor must determine the annual percentage rate for purposes of this paragraph (b)(4) by treating the maximum interest rate that may apply during that five-year period as the interest rate for the full term of the loan.

* * * * *

(e) * * *

(2) * * *

(v) For which the creditor, at or before consummation:

(A) Considers the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan, debt obligations, alimony, child support, and monthly debt-to-income ratio or residual income, using the amounts determined from paragraph (e)(2)(v)(B) of this section. For purposes of this paragraph (e)(2)(v)(A), the consumer's monthly debt-to-income ratio or residual income is determined in accordance with paragraph (c)(7) of this section, except that the consumer's monthly payment on the covered transaction, including the monthly payment for mortgage-related obligations, is calculated in accordance with paragraph (e)(2)(iv) of this section.

(B)(1) Verifies the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan using third-party records that provide reasonably reliable evidence of the consumer's income or assets, in accordance with paragraph (c)(4) of this section; and

(2) Verifies the consumer's current debt obligations, alimony, and child support using reasonably reliable third-party records in accordance with paragraph (c)(3) of this section.

(vi) For which the annual percentage rate does not exceed the average prime offer rate for a comparable transaction as of the date the interest rate is set by the amounts specified in paragraphs (e)(2)(vi)(A) through (F) of this section. The amounts specified here shall be adjusted annually on January 1 by the annual percentage change in the Consumer Price Index for All Urban Consumers (CPI-U) that was reported on the preceding June 1. For purposes of this paragraph (e)(2)(vi), the creditor must determine the annual percentage rate for a loan for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due by treating the maximum interest rate that may apply during that five-year period as the interest rate for the full term of the loan.

(A) For a first-lien covered transaction with a loan amount greater than or equal to $110,260 (indexed for inflation), 2.25 or more percentage points;

(B) For a first-lien covered transaction with a loan amount greater than or equal to $66,156 (indexed for inflation) but less than $110,260 (indexed for inflation), 3.5 or more percentage points;

(C) For a first-lien covered transaction with a loan amount less than $66,156 (indexed for inflation), 6.5 or more percentage points;

(D) For a first-lien covered transaction secured by a manufactured home with a loan amount less than $110,260 (indexed for inflation), 6.5 or more percentage points;

(E) For a subordinate-lien covered transaction with a loan amount greater than or equal to $66,156 (indexed for inflation), 3.5 or more percentage points;

(F) For a subordinate-lien covered transaction with a loan amount less than $66,156 (indexed for inflation), 6.5 or more percentage points.

* * * * *

(4) Qualified mortgage defined—other agencies. Notwithstanding paragraph (e)(2) of this section, a qualified mortgage is a covered transaction that is defined as a qualified mortgage by the U.S. Department of Housing and Urban Development under 24 CFR 201.7 and 24 CFR 203.19, the U.S. Department of Veterans Affairs under 38 CFR 36.4300 and 38 CFR 36.4500, or the U.S. Department of Agriculture under 7 CFR 3555.109.

(5) * * *

(i) * * *

(A) That satisfies the requirements of paragraph (e)(2) of this section other than the requirements of paragraphs (e)(2)(v) and (vi) of this section;

(B) For which the creditor:

(1) Considers and verifies at or before consummation the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan, in accordance with paragraphs (c)(2)(i) and (c)(4) of this section;

(2) Considers and verifies at or before consummation the consumer's current debt obligations, alimony, and child support in accordance with paragraphs (c)(2)(vi) and (c)(3) of this section;

(3) Considers at or before consummation the consumer's monthly debt-to-income ratio or residual income and verifies the debt obligations and income used to determine that ratio in accordance with paragraph (c)(7) of this section, except that the calculation of the payment on the covered transaction for purposes of determining the consumer's total monthly debt obligations in paragraph (c)(7)(i)(A) shall be determined in accordance with paragraph (e)(2)(iv) of this section instead of paragraph (c)(5) of this section;

* * * * *

(f) * * *

(1) * * *

(i) The loan satisfies the requirements for a qualified mortgage in paragraphs (e)(2)(i)(A) and (e)(2)(ii) and (iii) of this section;

* * * * *

(iii) The creditor:

(A) Considers and verifies at or before consummation the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan, in accordance with paragraphs (c)(2)(i) and (c)(4) of this section;

(B) Considers and verifies at or before consummation the consumer's current debt obligations, alimony, and child support in accordance with paragraphs (c)(2)(vi) and (c)(3) of this section;

(C) Considers at or before consummation the consumer's monthly debt-to-income ratio or residual income and verifies the debt obligations and income used to determine that ratio in accordance with paragraph (c)(7) of this section, except that the calculation of the payment on the covered transaction for purposes of determining the consumer's total monthly debt obligations in (c)(7)(i)(A) shall be determined in accordance with paragraph (f)(1)(iv)(A) of this section, together with the consumer's monthly payments for all mortgage-related obligations and excluding the balloon payment;

* * * * *

Appendix Q to Part 1026 [Removed]

3. Remove appendix Q to part 1026.

4. In supplement I to part 1026, under Section 1026.43—Minimum Standards for Transactions Secured by a Dwelling:

a. Under introductory paragraph 1, add introductory paragraph 2;

b. Revise sections 43(b)(4) Higher-priced covered transaction, 43(c)(4) Verification of income or assets, and 43(c)(7) Monthly debt-to-income ratio or residual income;

c. Revise Paragraph 43(e)(2)(v);

d. Add Paragraphs 43(e)(2)(v)(A) and 43(e)(2)(v)(B) after Paragraph 43(e)(2)(v);

e. Revise Paragraph 43(e)(2)(vi);

f. Revise section 43(e)(4); and

g. Revise Paragraph 43(e)(5) and Paragraphs 43(f)(1)(i), 43(f)(1)(ii), 43(f)(1)(iii), 43(f)(1)(iv), 43(f)(1)(v), and 43(f)(1)(vi),.

The additions and revisions read as follows:

Supplement I to Part 1026—Official Interpretations

* * * * *

Section 1026.43—Minimum Standards for Transactions Secured by a Dwelling

* * * * *

2. General QM Amendments Effective on March 1, 2021. The Bureau's revisions to Regulation Z contained in Qualified Mortgage Definition Under the Truth in Lending Act (Regulation Z): General QM Loan Definition published on December 29, 2020 (2021 General QM Amendments) apply with respect to transactions for which a creditor received an application on or after March 1, 2021 (effective date). Compliance with the 2021 General QM Amendments is mandatory with respect to transactions for which a creditor received an application on or after July 1, 2021 (mandatory compliance date). For a given transaction for which a creditor received an application on or after March 1, 2021 but prior to July 1, 2021, a person has the option of complying either: With 12 CFR part 1026 as it is in effect; or with 12 CFR part 1026 as it was in effect on February 26, 2021, together with any amendments to 12 CFR part 1026 that become effective after February 26, 2021, other than the 2021 General QM Amendments. For transactions subject to § 1026.19(e), (f), or (g), creditors determine the date the creditor received the consumer's application, for purposes of this comment, in accordance with § 1026.2(a)(3)(ii). For transactions that are not subject to § 1026.19(e), (f), or (g), creditors can determine the date the creditor received the consumer's application, for purposes of this comment, in accordance with either § 1026.2(a)(3)(i) or (ii).

* * * * *

43(b)(4) Higher-Priced Covered Transaction

1. Average prime offer rate. The average prime offer rate is defined in § 1026.35(a)(2). For further explanation of the meaning of “average prime offer rate,” and additional guidance on determining the average prime offer rate, see comments 35(a)(2)-1 through -4.

2. Comparable transaction. A higher-priced covered transaction is a consumer credit transaction that is secured by the consumer's dwelling with an annual percentage rate that exceeds by the specified amount the average prime offer rate for a comparable transaction as of the date the interest rate is set. The published tables of average prime offer rates indicate how to identify a comparable transaction. See comment 35(a)(2)-2.

3. Rate set. A transaction's annual percentage rate is compared to the average prime offer rate as of the date the transaction's interest rate is set (or “locked”) before consummation. Sometimes a creditor sets the interest rate initially and then re-sets it at a different level before consummation. The creditor should use the last date the interest rate is set before consummation.

4. Determining the annual percentage rate for certain loans for which the interest rate may or will change. Provisions in subpart C of this part, including the commentary to § 1026.17(c)(1), address how to determine the annual percentage rate disclosures for closed-end credit transactions. Provisions in § 1026.32(a)(3) address how to determine the annual percentage rate to determine coverage under § 1026.32(a)(1)(i). Section 1026.43(b)(4) requires, only for the purposes of a qualified mortgage under § 1026.43(e)(2), a different determination of the annual percentage rate for purposes of § 1026.43(b)(4) for a loan for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due. See comment 43(e)(2)(vi)-4 for how to determine the annual percentage rate of such a loan.

* * * * *

43(c)(4) Verification of Income or Assets

1. Income or assets relied on. A creditor need consider, and therefore need verify, only the income or assets the creditor relies on to evaluate the consumer's repayment ability. See comment 43(c)(2)(i)-2. For example, if a consumer's application states that the consumer earns a salary and is paid an annual bonus and the creditor relies on only the consumer's salary to evaluate the consumer's repayment ability, the creditor need verify only the salary. See also comments 43(c)(3)-1 and -2.

2. Multiple applicants. If multiple consumers jointly apply for a loan and each lists income or assets on the application, the creditor need verify only the income or assets the creditor relies on in determining repayment ability. See comment 43(c)(2)(i)-5.

3. Tax-return transcript. Under § 1026.43(c)(4), a creditor may verify a consumer's income using an Internal Revenue Service (IRS) tax-return transcript, which summarizes the information in a consumer's filed tax return, another record that provides reasonably reliable evidence of the consumer's income, or both. A creditor may obtain a copy of a tax-return transcript or a filed tax return directly from the consumer or from a service provider. A creditor need not obtain the copy directly from the IRS or other taxing authority. See comment 43(c)(3)-2.

4. Unidentified funds. A creditor does not meet the requirements of § 1026.43(c)(4) if it observes an inflow of funds into the consumer's account without confirming that the funds are income. For example, a creditor would not meet the requirements of § 1026.43(c)(4) where it observes an unidentified $5,000 deposit in the consumer's account but fails to take any measures to confirm or lacks any basis to conclude that the deposit represents the consumer's personal income and not, for example, proceeds from the disbursement of a loan.

* * * * *

43(c)(7) Monthly Debt-to-Income Ratio or Residual Income

1. Monthly debt-to-income ratio or monthly residual income. Under § 1026.43(c)(2)(vii), the creditor must consider the consumer's monthly debt-to-income ratio, or the consumer's monthly residual income, in accordance with the requirements in § 1026.43(c)(7). Section 1026.43(c) does not prescribe a specific monthly debt-to-income ratio with which creditors must comply. Instead, an appropriate threshold for a consumer's monthly debt-to-income ratio or monthly residual income is for the creditor to determine in making a reasonable and good faith determination of a consumer's ability to repay.

2. Use of both monthly debt-to-income ratio and monthly residual income. If a creditor considers the consumer's monthly debt-to-income ratio, the creditor may also consider the consumer's residual income as further validation of the assessment made using the consumer's monthly debt-to-income ratio.

3. Compensating factors. The creditor may consider factors in addition to the monthly debt-to-income ratio or residual income in assessing a consumer's repayment ability. For example, the creditor may reasonably and in good faith determine that a consumer has the ability to repay despite a higher debt-to-income ratio or lower residual income in light of the consumer's assets other than the dwelling, including any real property attached to the dwelling, securing the covered transaction, such as a savings account. The creditor may also reasonably and in good faith determine that a consumer has the ability to repay despite a higher debt-to-income ratio in light of the consumer's residual income.

* * * * *

Paragraph 43(e)(2)(v)

1. General. For guidance on satisfying § 1026.43(e)(2)(v), a creditor may rely on commentary to § 1026.43(c)(2)(i) and (vi), (c)(3), and (c)(4).

Paragraph 43(e)(2)(v)(A)

Consider. In order to comply with the requirement to consider under § 1026.43(e)(2)(v)(A), a creditor must take into account current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan, debt obligations, alimony, child support, and monthly debt-to-income ratio or residual income in its ability-to-repay determination. A creditor must maintain written policies and procedures for how it takes into account, pursuant to its underwriting standards, income or assets, debt obligations, alimony, child support, and monthly debt-to-income ratio or residual income in its ability-to-repay determination. A creditor must also retain documentation showing how it took into account income or assets, debt obligations, alimony, child support, and monthly debt-to-income ratio or residual income in its ability-to-repay determination, including how it applied its policies and procedures, in order to meet this requirement to consider and thereby meet the requirements for a qualified mortgage under § 1026.43(e)(2). This documentation may include, for example, an underwriter worksheet or a final automated underwriting system certification, in combination with the creditor's applicable underwriting standards and any applicable exceptions described in its policies and procedures, that shows how these required factors were taken into account in the creditor's ability-to-repay determination.

2. Requirement to consider monthly debt-to-income ratio or residual income. Section 1026.43(e)(2)(v)(A) does not prescribe specifically how a creditor must consider monthly debt-to-income ratio or residual income. Section 1026.43(e)(2)(v)(A) also does not prescribe a particular monthly debt-to-income ratio or residual income threshold with which a creditor must comply. A creditor may, for example, consider monthly debt-to-income ratio or residual income by establishing monthly debt-to-income or residual income thresholds for its own underwriting standards and documenting how it applied those thresholds to determine the consumer's ability to repay. A creditor may also consider these factors by establishing monthly debt-to-income or residual income thresholds and exceptions to those thresholds based on other compensating factors, and documenting application of the thresholds along with any applicable exceptions.

3. Flexibility to consider additional factors related to a consumer's ability to repay. The requirement to consider income or assets, debt obligations, alimony, child support, and monthly debt-to-income ratio or residual income does not preclude the creditor from taking into account additional factors that are relevant in determining a consumer's ability to repay the loan. For guidance on considering additional factors in determining the consumer's ability to repay, see comment 43(c)(7)-3.

Paragraph 43(e)(2)(v)(B)

1. Verification of income, assets, debt obligations, alimony, and child support. Section 1026.43(e)(2)(v)(B) does not prescribe specific methods of underwriting that creditors must use. Section 1026.43(e)(2)(v)(B)(1) requires a creditor to verify the consumer's current or reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan in accordance with § 1026.43(c)(4), which states that a creditor must verify such amounts using third-party records that provide reasonably reliable evidence of the consumer's income or assets. Section 1026.43(e)(2)(v)(B)(2) requires a creditor to verify the consumer's current debt obligations, alimony, and child support in accordance with § 1026.43(c)(3), which states that a creditor must verify such amounts using reasonably reliable third-party records. So long as a creditor complies with the provisions of § 1026.43(c)(3) with respect to debt obligations, alimony, and child support and § 1026.43(c)(4) with respect to income and assets, the creditor is permitted to use any reasonable verification methods and criteria.

2. Classifying and counting income, assets, debt obligations, alimony, and child support. “Current and reasonably expected income or assets other than the value of the dwelling (including any real property attached to the dwelling) that secures the loan” is determined in accordance with § 1026.43(c)(2)(i) and its commentary. “Current debt obligations, alimony, and child support” has the same meaning as under § 1026.43(c)(2)(vi) and its commentary. Section 1026.43(c)(2)(i) and (vi) and the associated commentary apply to a creditor's determination with respect to what inflows and property it may classify and count as income or assets and what obligations it must classify and count as debt obligations, alimony, and child support, pursuant to its compliance with § 1026.43(e)(2)(v)(B).

3. Safe harbor for compliance with specified external standards.

i. Meeting the standards in the following manuals for verifying current or reasonably expected income or assets using third-party records provides a creditor with reasonably reliable evidence of the consumer's income or assets. Meeting the standards in the following manuals for verifying current debt obligations, alimony, and child support using third-party records provides a creditor with reasonably reliable evidence of the consumer's debt obligations, alimony, and child support obligations. Accordingly, a creditor complies with § 1026.43(e)(2)(v)(B) if it complies with verification standards in one or more of the following manuals:

A. Chapters B3-3 through B3-6 of the Fannie Mae Single Family Selling Guide, published June 3, 2020;

B. Sections 5102 through 5500 of the Freddie Mac Single-Family Seller/Servicer Guide, published June 10, 2020;

C. Sections II.A.1 and II.A.4-5 of the Federal Housing Administration's Single Family Housing Policy Handbook, issued October 24, 2019;

D. Chapter 4 of the U.S. Department of Veterans Affairs' Lenders Handbook, revised February 22, 2019;

E. Chapter 4 of the U.S. Department of Agriculture's Field Office Handbook for the Direct Single Family Housing Program, revised March 15, 2019; and

F. Chapters 9 through 11 of the U.S. Department of Agriculture's Handbook for the Single Family Guaranteed Loan Program, revised March 19, 2020.

ii. Applicable provisions in manuals. A creditor complies with § 1026.43(e)(2)(v)(B) if it complies with requirements in the manuals listed in comment 43(e)(2)(v)(B)-3 for creditors to verify income, assets, debt obligations, alimony and child support using specified reasonably reliable third-party documents or to include or exclude particular inflows, property, and obligations as income, assets, debt obligations, alimony, and child support.

iii. Inapplicable provisions in manuals. For purposes of compliance with § 1026.43(e)(2)(v)(B), a creditor need not comply with requirements in the manuals listed in comment 43(e)(2)(v)(B)-3 other than those that require creditors to verify income, assets, debt obligations, alimony and child support using specified documents or to classify and count particular inflows, property, and obligations as income, assets, debt obligations, alimony, and child support.

iv. Revised versions of manuals. A creditor also complies with § 1026.43(e)(2)(v)(B) where it complies with revised versions of the manuals listed in comment 43(e)(2)(v)(B)-3.i, provided that the two versions are substantially similar.

v. Use of standards from more than one manual. A creditor complies with § 1026.43(e)(2)(v)(B) if it complies with the verification standards in one or more of the manuals specified in comment 43(e)(2)(v)(B)-3.i. Accordingly, a creditor may, but need not, comply with § 1026.43(e)(2)(v)(B) by complying with the verification standards from more than one manual (in other words, by “mixing and matching” verification standards).

Paragraph 43(e)(2)(vi)

1. Determining the average prime offer rate for a comparable transaction as of the date the interest rate is set. For guidance on determining the average prime offer rate for a comparable transaction as of the date the interest rate is set, see comments 43(b)(4)-1 through -3.

2. Determination of applicable threshold. A creditor must determine the applicable threshold by determining which category the loan falls into based on the face amount of the note (the “loan amount” as defined in § 1026.43(b)(5)). For example, for a first-lien covered transaction with a loan amount of $75,000, the loan would fall into the tier for loans greater than or equal to $66,156 (indexed for inflation) but less than $110,260 (indexed for inflation), for which the applicable threshold is 3.5 or more percentage points.

3. Annual adjustment for inflation. The dollar amounts in § 1026.43(e)(2)(vi) will be adjusted annually on January 1 by the annual percentage change in the CPI-U that was in effect on the preceding June 1. The Bureau will publish adjustments after the June figures become available each year.

4. Determining the annual percentage rate for certain loans for which the interest rate may or will change.

i. In general. The commentary to § 1026.17(c)(1) and other provisions in subpart C address how to determine the annual percentage rate disclosures for closed-end credit transactions. Provisions in § 1026.32(a)(3) address how to determine the annual percentage rate to determine coverage under § 1026.32(a)(1)(i). Section 1026.43(e)(2)(vi) requires, for the purposes of § 1026.43(e)(2)(vi), a different determination of the annual percentage rate for a qualified mortgage under § 1026.43(e)(2) for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due. An identical special rule for determining the annual percentage rate for such a loan also applies for purposes of § 1026.43(b)(4).

ii. Loans for which the interest rate may or will change. Section 1026.43(e)(2)(vi) includes a special rule for determining the annual percentage rate for a loan for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due. This rule applies to adjustable-rate mortgages that have a fixed-rate period of five years or less and to step-rate mortgages for which the interest rate changes within that five-year period.

iii. Maximum interest rate during the first five years. For a loan for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due, a creditor must treat the maximum interest rate that could apply at any time during that five-year period as the interest rate for the full term of the loan to determine the annual percentage rate for purposes of § 1026.43(e)(2)(vi), regardless of whether the maximum interest rate is reached at the first or subsequent adjustment during the five-year period. For additional instruction on how to determine the maximum interest rate during the first five years after the date on which the first regular periodic payment will be due, see comments 43(e)(2)(iv)-3 and -4.

iv. Treatment of the maximum interest rate in determining the annual percentage rate. For a loan for which the interest rate may or will change within the first five years after the date on which the first regular periodic payment will be due, the creditor must determine the annual percentage rate for purposes of § 1026.43(e)(2)(vi) by treating the maximum interest rate that may apply within the first five years as the interest rate for the full term of the loan. For example, assume an adjustable-rate mortgage with a loan term of 30 years and an initial discounted rate of 5.0 percent that is fixed for the first three years. Assume that the maximum interest rate during the first five years after the date on which the first regular periodic payment will be due is 7.0 percent. Pursuant to § 1026.43(e)(2)(vi), the creditor must determine the annual percentage rate based on an interest rate of 7.0 percent applied for the full 30-year loan term.

5. Meaning of a manufactured home. For purposes of § 1026.43(e)(2)(vi)(D), manufactured home means any residential structure as defined under regulations of the U.S. Department of Housing and Urban Development (HUD) establishing manufactured home construction and safety standards (24 CFR 3280.2). Modular or other factory-built homes that do not meet the HUD code standards are not manufactured homes for purposes of § 1026.43(e)(2)(vi)(D).

6. Scope of threshold for transactions secured by a manufactured home. The threshold in § 1026.43(e)(2)(vi)(D) applies to first-lien covered transactions less than $110,260 (indexed for inflation) that are secured by a manufactured home and land, or by a manufactured home only.

* * * * *

43(e)(4) Qualified Mortgage Defined—Other Agencies

1. General. The Department of Housing and Urban Development, Department of Veterans Affairs, and the Department of Agriculture have promulgated definitions for qualified mortgages under mortgage programs they insure, guarantee, or provide under applicable law. Cross-references to those definitions are listed in § 1026.43(e)(4) to acknowledge the covered transactions covered by those definitions are qualified mortgages for purposes of this section.

2. Mortgages for which the creditor received the consumer's application prior to July 1, 2021. Covered transactions that met the requirements of § 1026.43(e)(2)(i) thorough (iii), were eligible for purchase or guarantee by the Federal National Mortgage Association (Fannie Mae) or the Federal Home Loan Mortgage Corporation (Freddie Mac) (or any limited-life regulatory entity succeeding the charter of either) operating under the conservatorship or receivership of the Federal Housing Finance Agency pursuant to section 1367 of the Federal Housing Enterprises Financial Safety and Soundness Act of 1992 (12 U.S.C. 4617), and for which the creditor received the consumer's application prior to the mandatory compliance date of July 1, 2021 continue to be qualified mortgages for the purposes of this section, including those covered transactions that were consummated on or after July 1, 2021.

3. Mortgages for which the creditor received the consumer's application on or after March 1, 2021 and prior to July 1, 2021. For a discussion of the optional early compliance period for the 2021 General QM Amendments, please see comment 43-2.

4. [Reserved].

5. [Reserved].

* * * * *

Paragraph 43(e)(5)

1. Satisfaction of qualified mortgage requirements. For a covered transaction to be a qualified mortgage under § 1026.43(e)(5), the mortgage must satisfy the requirements for a qualified mortgage under § 1026.43(e)(2), other than the requirements in § 1026.43(e)(2)(v) and (vi). For example, a qualified mortgage under § 1026.43(e)(5) may not have a loan term in excess of 30 years because longer terms are prohibited for qualified mortgages under § 1026.43(e)(2)(ii). Similarly, a qualified mortgage under § 1026.43(e)(5) may not result in a balloon payment because § 1026.43(e)(2)(i)(C) provides that qualified mortgages may not have balloon payments except as provided under § 1026.43(f). However, a covered transaction need not comply with § 1026.43(e)(2)(v) and (vi).

2. Debt-to-income ratio or residual income. Section 1026.43(e)(5) does not prescribe a specific monthly debt-to-income ratio with which creditors must comply. Instead, creditors must consider a consumer's debt-to-income ratio or residual income calculated generally in accordance with § 1026.43(c)(7) and verify the information used to calculate the debt-to-income ratio or residual income in accordance with § 1026.43(c)(3) and (4). However, § 1026.43(c)(7) refers creditors to § 1026.43(c)(5) for instructions on calculating the payment on the covered transaction. Section 1026.43(c)(5) requires creditors to calculate the payment differently than § 1026.43(e)(2)(iv). For purposes of the qualified mortgage definition in § 1026.43(e)(5), creditors must base their calculation of the consumer's debt-to-income ratio or residual income on the payment on the covered transaction calculated according to § 1026.43(e)(2)(iv) instead of according to § 1026.43(c)(5).

3. Forward commitments. A creditor may make a mortgage loan that will be transferred or sold to a purchaser pursuant to an agreement that has been entered into at or before the time the transaction is consummated. Such an agreement is sometimes known as a “forward commitment.” A mortgage that will be acquired by a purchaser pursuant to a forward commitment does not satisfy the requirements of § 1026.43(e)(5), whether the forward commitment provides for the purchase and sale of the specific transaction or for the purchase and sale of transactions with certain prescribed criteria that the transaction meets. However, a forward commitment to another person that also meets the requirements of § 1026.43(e)(5)(i)(D) is permitted. For example, assume a creditor that is eligible to make qualified mortgages under § 1026.43(e)(5) makes a mortgage. If that mortgage meets the purchase criteria of an investor with which the creditor has an agreement to sell loans after consummation, then the loan does not meet the definition of a qualified mortgage under § 1026.43(e)(5). However, if the investor meets the requirements of § 1026.43(e)(5)(i)(D), the mortgage will be a qualified mortgage if all other applicable criteria also are satisfied.

4. Creditor qualifications. To be eligible to make qualified mortgages under § 1026.43(e)(5), a creditor must satisfy the requirements stated in § 1026.35(b)(2)(iii)(B) and (C). Section 1026.35(b)(2)(iii)(B) requires that, during the preceding calendar year, or, if the application for the transaction was received before April 1 of the current calendar year, during either of the two preceding calendar years, the creditor and its affiliates together extended no more than 2,000 covered transactions, as defined by § 1026.43(b)(1), secured by first liens, that were sold, assigned, or otherwise transferred to another person, or that were subject at the time of consummation to a commitment to be acquired by another person. Section 1026.35(b)(2)(iii)(C) requires that, as of the preceding December 31st, or, if the application for the transaction was received before April 1 of the current calendar year, as of either of the two preceding December 31sts, the creditor and its affiliates that regularly extended, during the applicable period, covered transactions, as defined by § 1026.43(b)(1), secured by first liens, together, had total assets of less than $2 billion, adjusted annually by the Bureau for inflation.

5. Requirement to hold in portfolio. Creditors generally must hold a loan in portfolio to maintain the transaction's status as a qualified mortgage under § 1026.43(e)(5), subject to four exceptions. Unless one of these exceptions applies, a loan is no longer a qualified mortgage under § 1026.43(e)(5) once legal title to the debt obligation is sold, assigned, or otherwise transferred to another person. Accordingly, unless one of the exceptions applies, the transferee could not benefit from the presumption of compliance for qualified mortgages under § 1026.43(e)(1) unless the loan also met the requirements of another qualified mortgage definition.

6. Application to subsequent transferees. The exceptions contained in § 1026.43(e)(5)(ii) apply not only to an initial sale, assignment, or other transfer by the originating creditor but to subsequent sales, assignments, and other transfers as well. For example, assume Creditor A originates a qualified mortgage under § 1026.43(e)(5). Six months after consummation, Creditor A sells the qualified mortgage to Creditor B pursuant to § 1026.43(e)(5)(ii)(B) and the loan retains its qualified mortgage status because Creditor B complies with the limits on asset size and number of transactions. If Creditor B sells the qualified mortgage, it will lose its qualified mortgage status under § 1026.43(e)(5) unless the sale qualifies for one of the § 1026.43(e)(5)(ii) exceptions for sales three or more years after consummation, to another qualifying institution, as required by supervisory action, or pursuant to a merger or acquisition.

7. Transfer three years after consummation. Under § 1026.43(e)(5)(ii)(A), if a qualified mortgage under § 1026.43(e)(5) is sold, assigned, or otherwise transferred three years or more after consummation, the loan retains its status as a qualified mortgage under § 1026.43(e)(5) following the transfer. The transferee need not be eligible to originate qualified mortgages under § 1026.43(e)(5). The loan will continue to be a qualified mortgage throughout its life, and the transferee, and any subsequent transferees, may invoke the presumption of compliance for qualified mortgages under § 1026.43(e)(1).

8. Transfer to another qualifying creditor. Under § 1026.43(e)(5)(ii)(B), a qualified mortgage under § 1026.43(e)(5) may be sold, assigned, or otherwise transferred at any time to another creditor that meets the requirements of § 1026.43(e)(5)(i)(D). That section requires that a creditor together with all its affiliates, extended no more than 2,000 first-lien covered transactions that were sold, assigned, or otherwise transferred by the creditor or its affiliates to another person, or that were subject at the time of consummation to a commitment to be acquired by another person; and have, together with its affiliates that regularly extended covered transactions secured by first liens, total assets less than $2 billion (as adjusted for inflation). These tests are assessed based on transactions and assets from the calendar year preceding the current calendar year or from either of the two calendar years preceding the current calendar year if the application for the transaction was received before April 1 of the current calendar year. A qualified mortgage under § 1026.43(e)(5) transferred to a creditor that meets these criteria would retain its qualified mortgage status even if it is transferred less than three years after consummation.

9. Supervisory sales. Section 1026.43(e)(5)(ii)(C) facilitates sales that are deemed necessary by supervisory agencies to revive troubled creditors and resolve failed creditors. A qualified mortgage under § 1026.43(e)(5) retains its qualified mortgage status if it is sold, assigned, or otherwise transferred to another person pursuant to: A capital restoration plan or other action under 12 U.S.C. 1831o; the actions or instructions of any person acting as conservator, receiver or bankruptcy trustee; an order of a State or Federal government agency with jurisdiction to examine the creditor pursuant to State or Federal law; or an agreement between the creditor and such an agency. A qualified mortgage under § 1026.43(e)(5) that is sold, assigned, or otherwise transferred under these circumstances retains its qualified mortgage status regardless of how long after consummation it is sold and regardless of the size or other characteristics of the transferee. Section 1026.43(e)(5)(ii)(C) does not apply to transfers done to comply with a generally applicable regulation with future effect designed to implement, interpret, or prescribe law or policy in the absence of a specific order by or a specific agreement with a governmental agency described in § 1026.43(e)(5)(ii)(C) directing the sale of one or more qualified mortgages under § 1026.43(e)(5) held by the creditor or one of the other circumstances listed in § 1026.43(e)(5)(ii)(C). For example, a qualified mortgage under § 1026.43(e)(5) that is sold pursuant to a capital restoration plan under 12 U.S.C. 1831o would retain its status as a qualified mortgage following the sale. However, if the creditor simply chose to sell the same qualified mortgage as one way to comply with general regulatory capital requirements in the absence of supervisory action or agreement it would lose its status as a qualified mortgage following the sale unless it qualifies under another definition of qualified mortgage.

10. Mergers and acquisitions. A qualified mortgage under § 1026.43(e)(5) retains its qualified mortgage status if a creditor merges with, is acquired by, or acquires another person regardless of whether the creditor or its successor is eligible to originate new qualified mortgages under § 1026.43(e)(5) after the merger or acquisition. However, the creditor or its successor can originate new qualified mortgages under § 1026.43(e)(5) only if it complies with all of the requirements of § 1026.43(e)(5) after the merger or acquisition. For example, assume a creditor that originates 250 covered transactions each year and originates qualified mortgages under § 1026.43(e)(5) is acquired by a larger creditor that originates 10,000 covered transactions each year. Following the acquisition, the small creditor would no longer be able to originate § 1026.43(e)(5) qualified mortgages because, together with its affiliates, it would originate more than 500 covered transactions each year. However, the § 1026.43(e)(5) qualified mortgages originated by the small creditor before the acquisition would retain their qualified mortgage status.

* * * * *

43(f)(1) Exemption

Paragraph 43(f)(1)(i)

1. Satisfaction of qualified mortgage requirements. Under § 1026.43(f)(1)(i), for a mortgage that provides for a balloon payment to be a qualified mortgage, the mortgage must satisfy the requirements for a qualified mortgage in paragraphs (e)(2)(i)(A), (e)(2)(ii), and (e)(2)(iii). Therefore, a covered transaction with balloon payment terms must provide for regular periodic payments that do not result in an increase of the principal balance, pursuant to § 1026.43(e)(2)(i)(A); must have a loan term that does not exceed 30 years, pursuant to § 1026.43(e)(2)(ii); and must have total points and fees that do not exceed specified thresholds pursuant to § 1026.43(e)(2)(iii).

Paragraph 43(f)(1)(ii)

1. Example. Under § 1026.43(f)(1)(ii), if a qualified mortgage provides for a balloon payment, the creditor must determine that the consumer is able to make all scheduled payments under the legal obligation other than the balloon payment. For example, assume a loan in an amount of $200,000 that has a five-year loan term, but is amortized over 30 years. The loan agreement provides for a fixed interest rate of 6 percent. The loan consummates on March 3, 2014, and the monthly payment of principal and interest scheduled for the first five years is $1,199, with the first monthly payment due on April 1, 2014. The balloon payment of $187,308 is required on the due date of the 60th monthly payment, which is April 1, 2019. The loan can be a qualified mortgage if the creditor underwrites the loan using the scheduled principal and interest payment of $1,199, plus the consumer's monthly payment for all mortgage-related obligations, and satisfies the other criteria set forth in § 1026.43(f).

2. Creditor's determination. A creditor must determine that the consumer is able to make all scheduled payments other than the balloon payment to satisfy § 1026.43(f)(1)(ii), in accordance with the legal obligation, together with the consumer's monthly payments for all mortgage-related obligations and excluding the balloon payment, to meet the repayment ability requirements of § 1026.43(f)(1)(ii). A creditor satisfies § 1026.43(f)(1)(ii) if it uses the maximum payment in the payment schedule, excluding any balloon payment, to determine if the consumer has the ability to make the scheduled payments.

Paragraph 43(f)(1)(iii)

1. Debt-to-income or residual income. A creditor must consider and verify the consumer's monthly debt-to-income ratio or residual income to meet the requirements of § 1026.43(f)(1)(iii)(C). To calculate the consumer's monthly debt-to-income or residual income for purposes of § 1026.43(f)(1)(iii)(C), the creditor may rely on the definitions and calculation rules in § 1026.43(c)(7) and its accompanying commentary, except for the calculation rules for a consumer's total monthly debt obligations (which is a component of debt-to-income and residual income under § 1026.43(c)(7)). For purposes of calculating the consumer's total monthly debt obligations under § 1026.43(f)(1)(iii), the creditor must calculate the monthly payment on the covered transaction using the payment calculation rules in § 1026.43(f)(1)(iv)(A), together with all mortgage-related obligations and excluding the balloon payment.

Paragraph 43(f)(1)(iv)

1. Scheduled payments. Under § 1026.43(f)(1)(iv)(A), the legal obligation must provide that scheduled payments must be substantially equal and determined using an amortization period that does not exceed 30 years. Balloon payments often result when the periodic payment would fully repay the loan amount only if made over some period that is longer than the loan term. For example, a loan term of 10 years with periodic payments based on an amortization period of 20 years would result in a balloon payment being due at the end of the loan term. Whatever the loan term, the amortization period used to determine the scheduled periodic payments that the consumer must pay under the terms of the legal obligation may not exceed 30 years.

2. Substantially equal. The calculation of payments scheduled by the legal obligation under § 1026.43(f)(1)(iv)(A) are required to result in substantially equal amounts. This means that the scheduled payments need to be similar, but need not be equal. For further guidance on substantially equal payments, see comment 43(c)(5)(i)-4.

3. Interest-only payments. A mortgage that only requires the payment of accrued interest each month does not meet the requirements of § 1026.43(f)(1)(iv)(A).

Paragraph 43(f)(1)(v)

1. Forward commitments. A creditor may make a mortgage loan that will be transferred or sold to a purchaser pursuant to an agreement that has been entered into at or before the time the transaction is consummated. Such an agreement is sometimes known as a “forward commitment.” A balloon-payment mortgage that will be acquired by a purchaser pursuant to a forward commitment does not satisfy the requirements of § 1026.43(f)(1)(v), whether the forward commitment provides for the purchase and sale of the specific transaction or for the purchase and sale of transactions with certain prescribed criteria that the transaction meets. However, a purchase and sale of a balloon-payment qualified mortgage to another person that separately meets the requirements of § 1026.43(f)(1)(vi) is permitted. For example: Assume a creditor that meets the requirements of § 1026.43(f)(1)(vi) makes a balloon-payment mortgage that meets the requirements of § 1026.43(f)(1)(i) through (iv); if the balloon-payment mortgage meets the purchase criteria of an investor with which the creditor has an agreement to sell such loans after consummation, then the balloon-payment mortgage does not meet the definition of a qualified mortgage in accordance with § 1026.43(f)(1)(v). However, if the investor meets the requirement of § 1026.43(f)(1)(vi), the balloon-payment qualified mortgage retains its qualified mortgage status.

Paragraph 43(f)(1)(vi)

1. Creditor qualifications. Under § 1026.43(f)(1)(vi), to make a qualified mortgage that provides for a balloon payment, the creditor must satisfy three criteria that are also required under § 1026.35(b)(2)(iii)(A), (B) and (C), which require:

i. During the preceding calendar year or during either of the two preceding calendar years if the application for the transaction was received before April 1 of the current calendar year, the creditor extended a first-lien covered transaction, as defined in § 1026.43(b)(1), on a property that is located in an area that is designated either “rural” or “underserved,” as defined in § 1026.35(b)(2)(iv), to satisfy the requirement of § 1026.35(b)(2)(iii)(A) (the rural-or-underserved test). Pursuant to § 1026.35(b)(2)(iv), an area is considered to be rural if it is: A county that is neither in a metropolitan statistical area, nor a micropolitan statistical area adjacent to a metropolitan statistical area, as those terms are defined by the U.S. Office of Management and Budget; a census block that is not in an urban area, as defined by the U.S. Census Bureau using the latest decennial census of the United States; or a county or a census block that has been designated as “rural” by the Bureau pursuant to the application process established in 2016. See Application Process for Designation of Rural Area under Federal Consumer Financial Law; Procedural Rule, 81 FR 11099 (Mar. 3, 2016). An area is considered to be underserved during a calendar year if, according to HMDA data for the preceding calendar year, it is a county in which no more than two creditors extended covered transactions secured by first liens on properties in the county five or more times.

A. The Bureau determines annually which counties in the United States are rural or underserved as defined by § 1026.35(b)(2)(iv)(A)(1) or § 1026.35(b)(2)(iv)(B) and publishes on its public website lists of those counties to assist creditors in determining whether they meet the criterion at § 1026.35(b)(2)(iii)(A). Creditors may also use an automated tool provided on the Bureau's public website to determine whether specific properties are located in areas that qualify as “rural” or “underserved” according to the definitions in § 1026.35(b)(2)(iv) for a particular calendar year. In addition, the U.S. Census Bureau may also provide on its public website an automated address search tool that specifically indicates if a property address is located in an urban area for purposes of the Census Bureau's most recent delineation of urban areas. For any calendar year that begins after the date on which the Census Bureau announced its most recent delineation of urban areas, a property is located in an area that qualifies as “rural” according to the definitions in § 1026.35(b)(2)(iv) if the search results provided for the property by any such automated address search tool available on the Census Bureau's public website do not identify the property as being in an urban area. A property is also located in an area that qualifies as “rural,” if the Bureau has designated that area as rural under § 1026.35(b)(2)(iv)(A)(3) and published that determination in the Federal Register. See Application Process for Designation of Rural Area under Federal Consumer Financial Law; Procedural Rule, 81 FR 11099 (Mar. 3, 2016).

B. For example, if a creditor extended during 2017 a first-lien covered transaction that is secured by a property that is located in an area that meets the definition of rural or underserved under § 1026.35(b)(2)(iv), the creditor meets this element of the exception for any transaction consummated during 2018.

C. Alternatively, if the creditor did not extend in 2017 a transaction that meets the definition of rural or underserved test under § 1026.35(b)(2)(iv), the creditor satisfies this criterion for any transaction consummated during 2018 for which it received the application before April 1, 2018, if it extended during 2016 a first-lien covered transaction that is secured by a property that is located in an area that meets the definition of rural or underserved under § 1026.35(b)(2)(iv).

ii. During the preceding calendar year, or, if the application for the transaction was received before April 1 of the current calendar year, during either of the two preceding calendar years, the creditor together with its affiliates extended no more than 2,000 covered transactions, as defined by § 1026.43(b)(1), secured by first liens, that were sold, assigned, or otherwise transferred to another person, or that were subject at the time of consummation to a commitment to be acquired by another person, to satisfy the requirement of § 1026.35(b)(2)(iii)(B).

iii. As of the preceding December 31st, or, if the application for the transaction was received before April 1 of the current calendar year, as of either of the two preceding December 31sts, the creditor and its affiliates that regularly extended covered transactions secured by first liens, together, had total assets that do not exceed the applicable asset threshold established by the Bureau, to satisfy the requirement of § 1026.35(b)(2)(iii)(C). The Bureau publishes notice of the asset threshold each year by amending comment 35(b)(2)(iii)-1.iii.

Dated: December 10, 2020.

Grace Feola,

Federal Register Liaison, Bureau of Consumer Financial Protection.

Footnotes

1.  85 FR 41448 (July 10, 2020).

2.  85 FR 41716 (July 10, 2020).

3.  85 FR 53568 (Aug. 28, 2020).

4.  Dodd-Frank Wall Street Reform and Consumer Protection Act, Public Law 111-203, 124 Stat. 1376 (2010).

5.  15 U.S.C. 1601 et seq.

6.  Dodd-Frank Act sections 1411-12, 1414, 124 Stat. 2142-48, 2149; 15 U.S.C. 1639c.

7.  15 U.S.C. 1639b(a)(2).

8.  15 U.S.C. 1639c(a)(1). TILA section 103 defines “residential mortgage loan” to mean, with some exceptions including open-end credit plans, “any consumer credit transaction that is secured by a mortgage, deed of trust, or other equivalent consensual security interest on a dwelling or on residential real property that includes a dwelling.” 15 U.S.C. 1602(dd)(5). TILA section 129C also exempts certain residential mortgage loans from the ATR requirements. See, e.g., 15 U.S.C. 1639c(a)(8) (exempting reverse mortgages and temporary or bridge loans with a term of 12 months or less).

9.  15 U.S.C. 1639c(a)(3).

10.  15 U.S.C. 1639c(b)(1).

11.  15 U.S.C. 1639c(b)(2)(A).

12.  78 FR 6408 (Jan. 30, 2013).

13.  See 78 FR 35429 (June 12, 2013); 78 FR 44686 (July 24, 2013); 78 FR 60382 (Oct. 1, 2013); 79 FR 65300 (Nov. 3, 2014); 80 FR 59944 (Oct. 2, 2015); 81 FR 16074 (Mar. 25, 2016); 85 FR 67938 (Oct. 26, 2020).

14.  12 CFR 1026.43(c), (e).

15.  The QM definition is related to the definition of Qualified Residential Mortgage (QRM). Section 15G of the Securities Exchange Act of 1934, added by section 941(b) of the Dodd-Frank Act, generally requires the securitizer of asset-backed securities (ABS) to retain not less than 5 percent of the credit risk of the assets collateralizing the ABS. 15 U.S.C. 78o-11. Six Federal agencies (not including the Bureau) are tasked with implementing this requirement. Those agencies are the Board of Governors of the Federal Reserve System (Board), the Office of the Comptroller of the Currency (OCC), the Federal Deposit Insurance Corporation (FDIC), the Securities and Exchange Commission, the FHFA, and the U.S. Department of Housing and Urban Development (HUD) (collectively, the QRM agencies). Section 15G of the Securities Exchange Act of 1934 provides that the credit risk retention requirements shall not apply to an issuance of ABS if all of the assets that collateralize the ABS are QRMs. See 15 U.S.C. 78o-11(c)(1)(C)(iii), (4)(A) and (B). Section 15G requires the QRM agencies to jointly define what constitutes a QRM, taking into consideration underwriting and product features that historical loan performance data indicate result in a lower risk of default. See 15 U.S.C. 78o-11(e)(4). Section 15G also provides that the definition of a QRM shall be “no broader than” the definition of a “qualified mortgage,” as the term is defined under TILA section 129C(b)(2), as amended by the Dodd-Frank Act, and regulations adopted thereunder. 15 U.S.C. 78o-11(e)(4)(C). In 2014, the QRM agencies issued a final rule adopting the risk retention requirements. 79 FR 77601 (Dec. 24, 2014). That final rule aligns the QRM definition with the QM definition defined by the Bureau in the ATR/QM Rule, effectively exempting securities comprised of loans that meet the QM definition from the risk retention requirement. That final rule also requires the agencies to review the definition of QRM no later than four years after the effective date of the final risk retention rules. In 2019, the QRM agencies initiated a review of certain provisions of the risk retention rule, including the QRM definition. 84 FR 70073 (Dec. 20, 2019). Among other things, the review allows the QRM agencies to consider the QRM definition in light of any changes to the QM definition adopted by the Bureau.

16.  12 CFR 1026.43(e)(2)(i) through (iii).

17.  12 CFR 1026.43(e)(2)(iv).

18.  12 CFR 1026.43(e)(2)(v).

19.  12 CFR 1026.43(e)(2)(vi).

20.  78 FR 6408, 6527-28 (Jan. 30, 2013) (noting that appendix Q incorporates, with certain modifications, the definitions and standards in HUD Handbook 4155.1, Mortgage Credit Analysis for Mortgage Insurance on One-to-Four-Unit Mortgage Loans).

21.  12 CFR 1026, appendix Q.

22.  12 CFR 1026.43(e)(2)(i) through (iii).

23.  12 CFR 1026.43(e)(4).

24.  12 CFR 1026.43(e)(4)(iii)(B). The ATR/QM Rule created several additional categories of QMs. The first additional category consisted of mortgages eligible to be insured or guaranteed (as applicable) by HUD (FHA loans), the U.S. Department of Veterans Affairs (VA loans), the U.S. Department of Agriculture (USDA loans), and the Rural Housing Service (RHS loans). 12 CFR 1026.43(e)(4)(ii)(B) through (E). This temporary category of QMs no longer exists because the relevant Federal agencies have since issued their own QM rules. See, e.g., 24 CFR 203.19 (HUD rule). Other categories of QMs provide more flexible standards for certain loans originated by certain small creditors. 12 CFR 1026.43(e)(5), (f); cf. 12 CFR 1026.43(e)(6) (applicable only to covered transactions for which the application was received before Apr. 1, 2016).

25.  78 FR 6408, 6527 (Jan. 30, 2013).

26.  Id. at 6527-28.

27.  Id. at 6533-34.

28.  Id. at 6534.

29.  Id. at 6533.

30.  Id. at 6534.

31.  Id. at 6536.

32.  Id. at 6534.

33.  Id.

34.  Id.

35.  Id. at 6511.

36.  Id. at 6507.

37.  Id. at 6511.

38.  Id. at 6514.

39.  12 CFR 1026.43(b)(4).

40.  12 CFR 1026.43(e)(1)(i).

41.  12 CFR 1026.43(e)(1)(ii).

42.  78 FR 6408 at 6506, 6510-14 (Jan. 30, 2013).

43.  Id. at 6408.

44.  Id. at 6511.

45.  Id.

46.  Id.

47.  Id.

48.  Id.

49.  Id.

50.  Id. at 6511-12.

51.  Id. at 6413-14, 6510-11.

52.  Riegle Community Development and Regulatory Improvement Act of 1994, Public Law 103-325, 108 Stat. 2160 (1994).

53.  As originally enacted, HOEPA defined a class of “high-cost mortgages,” which were generally closed-end home-equity loans (excluding home-purchase loans) with APRs or total points and fees exceeding prescribed thresholds. Mortgages covered by HOEPA have been referred to as “HOEPA loans,” “Section 32 loans,” or “high-cost mortgages.”

54.  The Dodd-Frank Act adjusted the baseline for the APR comparison, lowered the points-and-fees threshold, and added a prepayment trigger.

55.  TILA section 129(h); 15 U.S.C. 1639(h). In addition to the disclosures and limitations specified in the statute, HOEPA expanded the Board's rulemaking authority, among other things, to prohibit acts or practices the Board found to be unfair and deceptive in connection with mortgage loans.

56.  Subsequently renumbered as sections 1026.31, 1026.32, and 1026.33 of Regulation Z.

57.  See 60 FR 15463 (Mar. 24, 1995).

58.  Under the Board's 2008 HOEPA Final Rule, an HPML is a consumer credit transaction secured by the consumer's principal dwelling with an APR that exceeds APOR for a comparable transaction, as of the date the interest rate is set, by 1.5 or more percentage points for loans secured by a first lien on the dwelling, or by 3.5 or more percentage points for loans secured by a subordinate lien on the dwelling. 73 FR 44522 (July 30, 2008) (2008 HOEPA Final Rule). The definition of an HPML includes practically all “high-cost mortgages” because the latter transactions are determined by higher loan pricing threshold tests. See 12 CFR 226.35(a)(1).

59.  73 FR 44522 (July 30, 2008).

60.  See 12 CFR 1026.34(a)(4)(iii), (iv).

61.  12 U.S.C. 5512(d).

62.  82 FR 25246 (June 1, 2017).

63.  See generally Bureau of Consumer Fin. Prot., Ability to Repay and Qualified Mortgage Assessment Report (Jan. 2019), https://files.consumerfinance.gov/​f/​documents/​cfpb_​ability-to-repay-qualified-mortgage_​assessment-report.pdf (Assessment Report).

64.  See, e.g., id. at 83-84, 100-05.

65.  See, e.g., id. at 10, 194-96.

66.  See, e.g., id. at 10-11, 117, 131-47.

67.  Id. at 188. Because the Temporary GSE QM loan definition generally affects only loans that conform to the GSEs' guidelines, the Assessment Report's discussion of the Temporary GSE QM loan definition focused on the conforming segment of the market, not on non-conforming (e.g., jumbo) loans.

68.  Id. at 191.

69.  Id. at 192.

70.  Id. at 13, 190, 238.

71.  Id. at 193.

72.  Id. at 193-94.

73.  Id. at 194.

74.  Id.

75.  Id. at 194-95.

76.  Id. at 119-20.

77.  Id. at 153.

78.  Id. at 196.

79.  Id.

80.  Id.

81.  Id.

82.  Id. at 197.

83.  Id. at 196.

84.  Id. at 205.

85.  Id.

86.  The Quarterly CARES Act Report to Congress: Hearing Before the S. Comm. on Banking, Housing, and Urban Affairs, 116th Cong. 2-3 (2020) (statement of Jerome H. Powell, Chairman, Board of Governors of the Federal Reserve System).

87.  Agency MBS are backed by loans guaranteed by Fannie Mae, Freddie Mac, and the Government National Mortgage Association (Ginnie Mae).

88.  Laurie Goodman et al., Urban Inst., Housing Finance at a Glance, Monthly Chartbook (Mar. 26, 2020), https://www.urban.org/​sites/​default/​files/​publication/​101926/​housing-finance-at-a-glance-a-monthly-chartbook-march-2020.pdf.

89.  Press Release, Bd. of Governors of the Fed. Reserve Sys., Federal Reserve announces extensive new measures to support the economy (Mar. 23, 2020), https://www.federalreserve.gov/​newsevents/​pressreleases/​monetary20200323b.htm.

90.  The Quarterly CARES Act Report to Congress: Hearing Before the S. Comm. on Banking, Housing, and Urban Affairs, 116th Cong. 3 (2020) (statement of Jerome H. Powell, Chairman, Board of Governors of the Federal Reserve System).

91.  Non-agency MBS are not backed by loans guaranteed by Fannie Mae, Freddie Mac or Ginnie Mae. This includes securities collateralized by non-QM loans.

92.  Brandon Ivey, Non-Agency MBS Issuance Slowed in First Quarter, Inside Mortg. Fin. (Apr. 3, 2020), https://www.insidemortgagefinance.com/​articles/​217623-non-agency-mbs-issuance-slowed-in-first-quarter.

93.  Brandon Ivey, Non-Agency MBS Issuance Slow in Mid-August, Inside Mortg. Fin. (Aug. 21, 2020), https://www.insidemortgagefinance.com/​articles/​218973-non-agency-mbs-issuance-slow-in-mid-august.

94.  Brandon Ivey, Expanded-Credit Lending Inches Up in Third Quarter, Inside Mortg. Fin. (Nov. 25, 2020), https://www.insidemortgagefinance.com/​articles/​219861-expanded-credit-lending-ticks-up-in-3q-amid-slow-recovery.

95.  Brandon Ivey, Jumbo Originations Drop Nearly 22% in First Quarter, Inside Mortg. Fin. (May 15, 2020) https://www.insidemortgagefinance.com/​articles/​218028-jumbo-originations-drop-nearly-22-in-first-quarter.

96.  Brandon Ivey, Jumbo Lending Flat in 3Q, Wide Variation Among Lenders, Inside Mortg. Fin. (Nov. 13, 2020) https://www.insidemortgagefinance.com/​articles/​219738-jumbo-lending-level-in-3q-wide-variation-among-lenders.

97.  Laurie Goodman et al., Urban Inst., Housing Finance at a Glance, Monthly Chartbook, Inside Mortg. Fin. (Nov. 24, 2020), https://www.urban.org/​sites/​default/​files/​publication/​103273/​housing-finance-at-a-glance-a-monthly-chartbook-november-2020_​0.pdf.

98.  Public Law 116-136, 134 Stat. 281 (2020) (includes loans backed by HUD, USDA, VA, Fannie Mae, and Freddie Mac).

99.  See, e.g., Fed. Hous. Fin. Agency, FHFA Extends Foreclosure and REO Eviction Moratoriums (Dec. 2, 2020), https://www.fhfa.gov/​Media/​PublicAffairs/​Pages/​FHFA-Extends-Foreclosure-and-REO-Eviction-Moratoriums-12022020.aspx;​; Press Release, U.S. Dep't of Hous. & Urban Dev., FHA Extends Foreclosure And Eviction Moratorium For Homeowners Through Year End (Aug. 27, 2020), https://www.hud.gov/​press/​press_​releases_​media_​advisories/​HUD_​No_​20_​134;​; Veterans Benefits Admin., Extended Foreclosure Moratorium for Borrowers Affected by COVID-19 (Aug. 24, 2020), https://www.benefits.va.gov/​HOMELOANS/​documents/​circulars/​26-20-30.pdf;​; Rural Dev., U.S. Dep't of Agric., Extension of Foreclosure and Eviction Moratorium for Single Family Housing Direct Loans (Aug. 28, 2020), https://content.govdelivery.com/​accounts/​USDARD/​bulletins/​29c3a9e.

100.  The GSEs typically repurchase loans out of the trust after they fall 120 days delinquent, after which the servicer is no longer required to advance principal and interest, but Ginnie Mae requires servicers to advance principal and interest until the default is resolved. On April 21, 2020, the FHFA confirmed that servicers of GSE loans will only be required to advance four months of mortgage payments, regardless of whether the GSEs repurchase the loans from the trust after 120 days of delinquency. Fed. Hous. Fin. Agency, FHFA Addresses Servicer Liquidity Concerns, Announces Four Month Advance Obligation Limit for Loans in Forbearance (Apr. 21, 2020), https://www.fhfa.gov/​Media/​PublicAffairs/​Pages/​FHFA-Addresses-Servicer-Liquidity-Concerns-Announces-Four-Month-Advance-Obligation-Limit-for-Loans-in-Forbearance.aspx.

101.  Press Release, Mortg. Bankers Ass'n, Share of Mortgage Loans in Forbearance Increases to 5.54% (Dec. 1, 2020), https://www.mba.org/​2020-press-releases/​december/​share-of-mortgage-loans-in-forbearance-increases-to-554-percent.

102.  Warehouse providers are creditors that provide financing to mortgage originators and servicers to fund and service loans.

103.  Maria Volkova, FHA/VA Lenders Raise Credit Score Requirements, Inside Mortg. Fin. (Apr. 3, 2020), https://www.insidemortgagefinance.com/​articles/​217636-fhava-lenders-raise-fico-credit-score-requirements.

104.  Press Release, Fed. Hous. Fin. Agency, Adverse Market Refinance Fee Implementation now December 1 (Aug. 25, 2020), https://www.fhfa.gov/​Media/​PublicAffairs/​Pages/​Adverse-Market-Refinance-Fee-Implementation-Now-December-1.aspx.

105.  On April 10, 2020, Ginnie Mae released guidance on a Pass-Through Assistance Program whereby Ginnie Mae will provide financial assistance at a fixed interest rate to servicers facing a principal and interest shortfall as a last resort. Ginnie Mae, All Participant Memorandum (APM) 20-03: Availability of Pass-Through Assistance Program for Participants in Ginnie Mae's Single-Family MBS Program (Apr. 10, 2020), https://www.ginniemae.gov/​issuers/​program_​guidelines/​Pages/​mbsguideapmslibdisppage.aspx?​ParamID=​105. On April 7, 2020, Ginnie Mae also announced approval of a servicing advance financing facility, whereby mortgage servicing rights are securitized and sold to private investors. Press Release, Ginnie Mae, Ginnie Mae approves private market servicer liquidity facility (Apr. 7, 2020), https://www.ginniemae.gov/​newsroom/​Pages/​PressReleaseDispPage.aspx?​ParamID=​194.

106.  Brandon Ivey, Non-QM Lenders Regaining Footing, Inside Mortg. Fin. (July 24, 2020), https://www.insidemortgagefinance.com/​articles/​218696-non-qm-lenders-regaining-footing-with-a-positive-outlook (on file).

107.  The Bureau has consulted with agencies including the FHFA, the Board, FHA, the FDIC, the OCC, the Federal Trade Commission, the National Credit Union Administration, HUD, and the Department of the Treasury.

108.  82 FR 25246 (June 1, 2017).

109.  See Assessment Report, supra note 63, appendix B (summarizing comments received in response to the Assessment RFI).

110.  See Bureau of Consumer Fin. Prot., Call for Evidence, https://www.consumerfinance.gov/​policy-compliance/​notice-opportunities-comment/​archive-closed/​call-for-evidence (last updated Apr. 17, 2018).

111.  83 FR 10437 (Mar. 9, 2018).

112.  83 FR 12286 (Mar. 21, 2018).

113.  83 FR 12881 (Mar. 26, 2018).

114.  84 FR 37155, 37160-62 (July 31, 2019).

115.  The Bureau stated that if the amount of time industry would need to change its practices in response to the rule depends on how the Bureau revises the General QM loan definition, the Bureau requested time estimates based on alternative possible definitions.

116.  Based on 2018 data, the Bureau estimated in the General QM Proposal that 943,000 High-DTI conventional loans would fall outside the QM definitions if there are no changes to the General QM loan definition prior to the expiration of the Temporary GSE QM loan definition but would fall within the General QM loan definition if amended as the Bureau proposed.

117.  85 FR 53568 (Aug. 28, 2020).

118.  85 FR 60096 (Sept. 24, 2020).

119.  The Extension Final Rule also did not affect the QM definitions that apply to FHA, VA, USDA, or RHS loans.

120.  12 U.S.C. 5581(a)(1)(A).

121.  Dodd-Frank Act section 1002(14), 12 U.S.C. 5481(14) (defining “Federal consumer financial law” to include the “enumerated consumer laws” and the provisions of title X of the Dodd-Frank Act), Dodd-Frank Act section 1002(12)(O), 12 U.S.C. 5481(12)(O) (defining “enumerated consumer laws” to include TILA).

122.  15 U.S.C. 1604(a).

123.  15 U.S.C. 1601(a).

124.  15 U.S.C. 1639b(a)(2).

125.  15 U.S.C. 1639c(b)(2)(A).

126.  15 U.S.C. 1639c(b)(3)(B)(i).

127.  15 U.S.C. 1639c(b)(3)(A).

128.  12 U.S.C. 5512(b)(1).

129.  76 FR 27390, 27453 (May 11, 2011) (2011 ATR/QM Proposal).

130.  Id. at 27453.

131.  Id.

132.  78 FR 6408, 6516 (Jan. 30, 2013).

133.  Id. at 6526-27.

134.  Id. at 6526.

135.  Id. at 6526-27.

136.  Id.

137.  Id.

138.  Id.

139.  Id.

140.  Id.

141.  Id.

142.  Id. at 6528.

143.  Id.

144.  Id.

145.  Id.

146.  Id. at 6527.

147.  Id.

148.  Id.

149.  Id. at 6528.

150.  Id.

151.  Id.

152.  Id. at 6533.

153.  See Assessment Report, supra note 63, at 83-84, 100-05.

154.  Id. at 104-05.

155.  Id. at 105.

156.  See, e.g., id. at 10, 194-96.

157.  See, e.g., id. at 10-11, 117, 131-47.

158.  See, e.g., Assessment Report supra note 63, at 150, 153, Table 20. Table 20 illustrates how the pool of denied non-GSE eligible applicants with DTI ratios above 43 percent has changed between 2013 and 2014. After the introduction of the Rule, the pool of denied applicants contains more consumers with higher incomes, higher FICO scores, and higher down payments.

159.  78 FR 6408, 6527 (Jan. 30, 2013).

160.  Assessment Report, supra note 63, at 198.

161.  Id. at 198.

162.  Proposed Rule's Dodd-Frank Act section 1022(b) analysis (citing the Bureau's prior estimate of affected loans in the ANPR); see 84 FR 37155, 37159 (July 31, 2019).

163.  Id. at 37159 n.58.

164.  In fiscal year 2019, approximately 57 percent of FHA-insured purchase mortgages had a DTI ratio above 43 percent. U.S. Dep't of Hous. & Urban Dev., Annual Report to Congress Regarding the Financial Status of the FHA Mutual Mortgage Insurance Fund, Fiscal Year 2019, at 33 (using data from App. B Tbl. B9) (Nov. 14, 2018), https://www.hud.gov/​sites/​dfiles/​Housing/​documents/​2019FHAAnnualReportMMIFund.pdf.

165.  84 FR 37155, 37159 (July 31, 2019).

166.  Id. In 2018, FHA's county-level maximum loan limits ranged from $294,515 to $679,650 in the continental United States. See U.S. Dep't of Hous. & Urban Dev., FHA Mortgage Limits, https://entp.hud.gov/​idapp/​html/​hicostlook.cfm (last visited Dec. 8, 2020).

167.  84 FR 37155, 37159 (July 31, 2019).

168.  Interest rates and insurance premiums on FHA loans generally feature less risk-based pricing than conventional loans, charging more similar rates and premiums to all consumers. As a result, they are likely to cost more than conventional loans for consumers with stronger credit scores and larger down payments. Consistent with this pricing differential, consumers with higher credit scores and larger down payments chose FHA loans relatively rarely in 2018 HMDA data on mortgage originations. See Bureau of Consumer Fin. Prot., Introducing New and Revised Data Points in HMDA (Aug. 2019), https://files.consumerfinance.gov/​f/​documents/​cfpb_​new-revised-data-points-in-hmda_​report.pdf.

169.  84 FR 37155, 37159 (July 31, 2019).

170.  See 12 CFR 1026.43(e)(5) (extending QM status to certain portfolio loans originated by certain small creditors). In addition, section 101 of the Economic Growth, Regulatory Relief, and Consumer Protection Act, Public Law 115-174, 132 Stat. 1296 (2018), amended TILA to add a safe harbor for small creditor portfolio loans. See 15 U.S.C. 1639c(b)(2)(F).

171.  84 FR 37155, 37159 (July 31, 2019).

172.  Id.

173.  Id.

174.  Id.

175.  Id.

176.  See Assessment Report, supra note 63, at 10-11, 117, 131-47.

177.  Id. at 193.

178.  Id. at 193-94.

179.  See, e.g., Norbert Michel, The Best Housing Finance Reform Options for the Trump Administration, Forbes (July 15, 2019), https://www.forbes.com/​sites/​norbertmichel/​2019/​07/​15/​the-best-housing-finance-reform-options-for-the-trump-administration/​#4f5640de7d3f;​; Eric Kaplan et al., Milken Institute, A Blueprint for Administrative Reform of the Housing Finance System, at 17 (2019), https://assets1b.milkeninstitute.org/​assets/​Publication/​Viewpoint/​PDF/​Blueprint-Admin-Reform-HF-System-1.7.2019-v2.pdf (suggesting that the Bureau both (1) expand the 43 percent DTI limit to 45 percent to move market share of higher-DTI loans from the GSEs and FHA to the non-agency market, and (2) establish a residual income test to protect against the risk of higher DTI loans); Morris Davis et al., A Quarter Century of Mortgage Risk (FHFA, Working Paper 19-02, 2019), https://www.fhfa.gov/​PolicyProgramsResearch/​Research/​Pages/​wp1902.aspx (examining various loan characteristics and a summary measure of risk—the stressed default rate—for predictiveness of loan performance).

180.  See Assessment Report, supra note 63, at 83.

181.  The NMDB, jointly developed by the FHFA and the Bureau, provides de-identified loan characteristics and performance information for a 5 percent sample of all mortgage originations from 1998 to the present, supplemented by de-identified loan and borrower characteristics from Federal administrative sources and credit reporting data. See Bureau of Consumer Fin. Prot., Sources and Uses of Data at the Bureau of Consumer Financial Protection, at 55-56 (Sept. 2018), https://www.consumerfinance.gov/​documents/​6850/​bcfp_​sources-uses-of-data.pdf.

182.  HMDA was originally enacted by Congress in 1975 and is implemented by Regulation C, 12 CFR part 1003. See Bureau of Consumer Fin. Prot., Mortgage data (HMDA), https://www.consumerfinance.gov/​data-research/​hmda/​ (last visited Dec. 8, 2020). HMDA requires many financial institutions to maintain, report, and publicly disclose loan-level information about mortgages. These data are housed here to help show whether lenders are serving the housing needs of their communities; they give public officials information that helps them make decisions and policies; and they shed light on lending patterns that could be discriminatory. The public data are modified to protect applicant and borrower privacy.

183.  See Neil Bhutta and Benjamin J. Keys, Eyes Wide Shut? The Moral Hazard of Mortgage Insurers during the Housing Boom, (NBER Working Paper No. 24844, 2018), https://www.nber.org/​papers/​w24844.pdf. APOR is approximated with weekly Freddie Mac Primary Mortgage Market Survey (PMMS) data, retrieved from Fed. Reserve Bank of St. Louis, Fed. Reserve Econ. Data, https://fred.stlouisfed.org/​ (Mar. 4, 2020). Each loan's APR is approximated by the sum of the interest rate in the NMDB data and an assumed PMI payment of 0.32, 0.52, or 0.78 percentage points for loans with LTVs above 80 but at or below 85, above 85 but at or below 90, and above 90, respectively. These PMI are based on standard industry rates during this time period. The 30-year Fixed Rate PMMS average is used for fixed-rate loans with terms over 15 years, and 15-year Fixed Rate PMMS is used for loans with terms of 15 years or less. The 5/1-year Adjustable-Rate PMMS average is used (for available years) for ARMs with a first interest rate reset occurring 5 or more years after origination, while the 1-year adjustable-rate PMMS average is used for all other ARMs.

184.  Loans with rate spreads of 2.25 percentage points or more are grouped in Tables 1 and 5 to ensure sufficient sample size for reliable analysis of the 2002-2008 data. This grouping ensures that all cells shown in Table 5 contain at least 500 loans.

185.  Freddie Mac's PMMS is the source of data underlying APOR for most mortgages. See supra note 183 for additional details.

186.  Where possible, the FHFA provided an anonymized match of HMDA loan identifiers for 2018 NMDB originations, allowing the Bureau to analyze more detailed HMDA loan characteristics (e.g., rate spread over APOR) for approximately half of 2018 NMDB originations.

187.  Loans with rate spreads of 2 percentage points or more are grouped in Tables 2 and 6 to ensure sufficient sample size for reliable analysis of the 2018 data. This grouping ensures that all cells shown in Table 6 contain at least 500 loans.

188.  Fewer than 0.7 percent of loans have reported DTI ratios over 70 percent in the 2002-2008 data. These loans are excluded from Tables 3 and 5 due to reliability concerns (including outliers which may reflect reporting errors) and to ensure that all cells shown in Table 5 contain at least 500 loans.

189.  Fewer than 0.5 percent of loans have reported DTI ratios over 50 percent in the 2018 data. These loans are excluded from Tables 4 and 6 due to reliability concerns (including outliers which may reflect reporting errors) and to ensure that all cells shown in Table 6 contain at least 500 loans.

190.  As in Tables 2 and 4, above, the 2018 data are divided into larger bins to ensure enough loans per bin. Loans with a DTI ratio greater than 50 percent are excluded, as well as loans with a DTI ratio at or below 25 percent and rate spreads of 1.5 percentage points and above, because these bins contained fewer than 500 loans in the matched 2018 NMDB-HMDA sample.

191.  See Archana Pradhan & Pete Carroll, Expiration of the CFPB's Qualified Mortgage (QM) GSE Patch—Part V, CoreLogic Insights Blog (Jan. 13, 2020), https://www.corelogic.com/​blog/​2020/​1/​expiration-of-the-cfpbs-qualified-mortgage-qm-gse-patch-part-v.aspx. Delinquency was measured as of October 2019, so loans do not have two full years of payment history.

192.  The Bureau analyzes the performance and pricing for smaller loans in the section-by-section analysis for § 1026.43(e)(2)(vi).

193.  See Archana Pradhan & Pete Carroll, Expiration of the CFPB's Qualified Mortgage (QM) GSE Patch—Part IV, CoreLogic Insights Blog (Jan. 11, 2020), https://www.corelogic.com/​blog/​2020/​1/​expiration-of-the-cfpbs-qualified-mortgage-qm-gse-patch-part-iv.aspx. (Delinquency measured as of October 2019.)

194.  See Karan Kaul & Laurie Goodman, Urban Inst., Updated: What, If Anything, Should Replace QM GSE Patch, at 9 (Oct. 2020), https://www.urban.org/​sites/​default/​files/​publication/​99268/​2018_​10_​30_​qualified_​mortgage_​rule_​update_​finalized_​4.pdf.

195.  See Karan Kaul et al., Urban Inst., Comment Letter to the Consumer Financial Protection Bureau on the Qualified Mortgage Rule, at 9-10 (Sept. 2019), https://www.urban.org/​sites/​default/​files/​publication/​101048/​comment_​letter_​to_​the_​consumer_​financial_​protection_​bureau_​0.pdf.

196.  See TILA section 103(aa)(i); Regulation Z § 1026.32(a)(1)(i). TILA and Regulation Z also provide a separate price-based coverage trigger based on the points and fees charged on a loan. See TILA section 130(aa)(ii); Regulation Z § 1026.32(a)(1)(ii).

197.  See generally 73 FR 44522 (July 30, 2008).

198.  The Board's 2008 rule was superseded by the January 2013 Final Rule, which imposed ability-to-repay requirements on a broader range of closed-end consumer credit transactions secured by a dwelling. See generally 78 FR 6408 (Jan. 30, 2013).

199.  See § 1026.35(b) and (c).

200.  The Bureau understands from feedback that creditors are concerned about errors in DTI calculations and have previously requested that the Bureau permit a cure of DTI overages that are discovered after consummation. See 79 FR 25730, 25743-45 (May 6, 2014) (requesting comment on potential cure or correction provisions for DTI overages).

201.  All estimates in Table 7 included loans that meet the Small Creditor QM loan definition in § 1026.43(e)(5). In particular, loans originated by small creditors that meet the criteria in § 1026.43(e)(5) are safe harbor QMs if priced below 3.5 percentage points over APOR or are rebuttable presumption QMs if priced 3.5 percentage points or more over APOR. The Bureau has provided revised analysis in part V.B.5 to reflect a revised methodology to identify creditors eligible to originate loans as small creditors under § 1026.43(e)(5).

202.  85 FR 41716, 41736-37 (July 10, 2020).

203.  Id. at 41737. See also 78 FR 6408, 6534 (Jan. 13, 2013) (stating that the Bureau believed it was appropriate to presume that loans that are eligible to be purchased or guaranteed by the GSEs “while under conservatorship” have been originated with appropriate consideration of consumers' ability to repay “in light of this significant Federal role and the government's focus on affordability in the wake of the mortgage crisis”).

204.  The Bureau acknowledged that some loans currently originated as Temporary GSE QMs have higher DTI ratios. However, the proposal expressed concern about adopting a DTI limit above a range of 45 to 48 percent without a requirement to consider compensating factors.

205.  Robert Bartlett et al., Haas School of Business UC Berkeley, Consumer Lending Discrimination in the FinTech Era (2019), https://faculty.haas.berkeley.edu/​morse/​research/​papers/​discrim.pdf.

206.  See supra note 194.

207.  Among other things, the commenter cited a recent Experian consumer “payment hierarchy” study, which used samples of consumers at various points in time and with various combinations of credit obligations and observed the relative performance of the credit obligations for two years. The commenter pointed out that, with respect to the consumers observed from February 2018 to February 2020—the most recent cohort in the study—Experian found that among those consumers with a mortgage, auto loan, retail card, and general purpose credit card, 0.81 percent became 90 days delinquent on their mortgage, whereas 4.26 percent became 90 days delinquent on their bank card. The disparities were roughly the same for consumers with a mortgage, bank card, and personal loan. See Experian, Consumer payment hierarchy by trade type: Time-series analysis (July 2020), http://images.go.experian.com/​Web/​ExperianInformationSolutionsInc/​%7Ba6ad2c78-e1da-46eb-b97b-bf2d953ce38d%7D_​Payment_​Hierarchy_​Report.pdf. The commenter stated that this suggests that originating a mortgage where the consumer lacks a reasonable ability to repay may manifest in delinquencies on credit obligations other than the mortgage itself.

208.  The commenter collectively referred to homeowners associations, condominium associations, and housing cooperatives as “community associations.”

209.  15 U.S.C. 1691 et seq.

210.  42 U.S.C. 3601 et seq.

211.  These thresholds are discussed below in the section-by-section analysis of § 1026.43(e)(2)(vi)(B)-(F). Final § 1026.43(e)(2)(vi)(B) provides that, for first-lien covered transactions with loan amounts greater than or equal to $66,156 (indexed for inflation) but less than $110,260 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 3.5 or more percentage points. Section 1026.43(e)(2)(vi)(C) provides that, for first-lien covered transactions with loan amounts less than $66,156 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 6.5 or more percentage points. Section 1026.43(e)(2)(vi)(E) provides that, for subordinate-lien covered transactions with loan amounts greater than or equal to $66,156 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 3.5 or more percentage points. Section 1026.43(e)(2)(vi)(F) provides that, for subordinate-lien covered transactions with loan amounts less than $66,156 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 6.5 or more percentage points.

212.  Final § 1026.43(e)(2)(vi)(D) provides that, for first-lien covered transactions secured by a manufactured home with loan amounts less than $110,260 (indexed for inflation), the APR may not exceed APOR for a comparable transaction as of the date the interest rate is set by 6.5 or more percentage points.

213.  In 2018, FHA's county-level maximum loan limits ranged from $294,515 to $679,650 in the continental United States. See U.S. Dep't of Hous. & Urban Dev., FHA Mortgage Limits, https://entp.hud.gov/​idapp/​html/​hicostlook.cfm (last visited Dec. 8, 2020).

214.  84 FR 37155, 37159 (July 31, 2019).

215.  See 12 CFR 1026.43(e)(5) (extending QM status to certain portfolio loans originated by certain small creditors). In addition, section 101 of the Economic Growth, Regulatory Relief, and Consumer Protection Act, Public Law 115-174, 132 Stat. 1296 (2018), amended TILA to add a safe harbor for small creditor portfolio loans. See 15 U.S.C. 1639c(b)(2)(F).

216.  84 FR 37155, 37159 (July 31, 2019).

217.  Id.

218.  Id.

219.  See Assessment Report, supra note 63, at 10-11, 117, 131-47.

220.  As discussed in the section-by-section analysis of § 1026.43(e)(2)(vi)(B)-(F), the Bureau proposed a loan amount threshold of $109,898 to align with the threshold for the limits on points and fees, as updated for inflation, in § 1026.43(e)(3)(i) and the associated commentary. On August 19, 2020, the Bureau issued a final rule adjusting the loan amounts for the limits on points and fees under § 1026.43(e)(3)(i), based on the annual percentage change reflected in the CPI-U in effect on June 1, 2020. 85 FR 50944 (Aug. 19, 2020). To ensure consistency, the Bureau is finalizing a loan amount threshold of $110,260 rather than a threshold of $109,898.

221.  See 78 FR 6408, 6511 (Jan. 30, 2013).

222.  See supra note 207.

223.  Neil Bhutta & Daniel Ringo, Effects of the Ability to Repay and Qualified Mortgage Rules on the Mortgage Market, FEDS Notes, Bd. of Governors of the Fed. Reserve Sys. (2015), https://www.federalreserve.gov/​econresdata/​notes/​feds-notes/​2015/​effects-of-the-ability-to-repay-and-qualified-mortgage-rules-on-the-mortgage-market-20151229.html.

224.  See, e.g., TILA section 129C(b)(2)(C) (establishing distinct points-and-fees thresholds for QMs based on loan pricing); section 129C(c)(ii) (establishing price-based restrictions on QMs permitted to impose prepayment penalties).

225.  See section-by-section analysis of § 1026.43(e)(2)(v)(A).

226.  In the January 2013 Final Rule, the Bureau exercised its authority under TILA section 105(a) to provide, in the context of the ATR provisions in § 1026.43(c)(2)(i), that a creditor may not look to the value of the dwelling that secures the covered transaction in assessing the consumer's repayment ability, instead of providing that a creditor may not look to the consumer's equity in the dwelling, as provided in TILA section 129C(a). The Bureau adopted this approach to provide broader protections to consumers. See 78 FR 6408, 6463-64 (Jan. 30, 2013).

227.  While the Bureau's conclusion on the strong correlation between pricing and early delinquency is based on its own delinquency analysis in this final rule, an Urban Institute analysis cited by a commenter also showed a positive correlation between pricing and rate spread during the years 2005 to 2008, largely covering the market conditions present during the subprime mortgage boom. See supra note 194.

228.  With respect to the commenter who presented analysis of subprime loans from the 2000s housing boom and asserted that the data show that pricing as a measure of ability to repay fails when there is a credit boom due to rate spread compression, the Bureau notes that it is unclear from the analysis whether these loans would have also satisfied the QM product feature restrictions and limits on points and fees, or how the performance of the loans varied with rate spreads.

229.  See, e.g., Consent Order, U.S. v. Bancorpsouth Bank, No. 1:16-cv-00118, ECF No. 8 (N.D. Miss.) (July 25, 2016), https://files.consumerfinance.gov/​f/​documents/​201606_​cfpb_​bancorpSouth-consent-order.pdf (joint action for discriminatory mortgage lending practices including charging African-American customers for certain mortgage loans more than non-Hispanic White borrowers with similar loan qualifications).

230.  78 FR 6408, 6534 (Jan. 13, 2013) (stating that the Bureau believed it was appropriate to presume that loans that are eligible to be purchased or guaranteed by the GSEs “while under conservatorship” have been originated with appropriate consideration of consumers' ability to repay “in light of this significant Federal role and the government's focus on affordability in the wake of the mortgage crisis”).

231.  As discussed in the section-by-section analysis of § 1026.43(e)(2)(vi) below, this final rule provides that loans with an APR exceeding the APOR by 2.25 percentage points or more (or exceeding higher thresholds for certain small or subordinate-lien loans) are not eligible for General QM status under § 1026.43(e)(2). Unless otherwise eligible for QM status (such as under § 1026.43(e)(5) or § 1026.43(f)), for non-QM loans a creditor must make a reasonable and good faith determination of the consumer's ability to repay and does not receive a presumption of compliance.

232.  Subordinate-lien transactions are discussed below in the section-by-section analysis of § 1026.43(e)(2)(vi).

233.  78 FR 6408, 6507 (Jan. 30, 2013).

234.  Id. at 6511.

235.  Id. at 6514.

236.  Id. at 6511.

237.  Id.

238.  Id.

239.  Id.

240.  Id.

241.  Id.

242.  Id.

243.  Id.

244.  Id. at 6511.

245.  Id. at 6511-13.

246.  Subordinate-lien transactions are discussed below in the section-by-section analysis of § 1026.43(e)(2)(vi).

247.  However, as discussed in the section-by-section analysis of § 1026.43(e)(2)(vi) below, under the proposal a loan would not have been eligible for QM status (i.e., would not receive any presumption of compliance with the ATR requirements) under § 1026.43(e)(2) if the loan exceeded the separate pricing thresholds in proposed § 1026.43(e)(2)(vi).

248.  78 FR 6408, 6512 (Jan. 30, 2013).

249.  See Regulation Z comment 34(a)(4)(iii)-1.

250.  78 FR 6408, 6511-12 (Jan. 30, 2013). The Bureau in the January 2013 Final Rule stated that it interpreted TILA section 129C(b)(1) to create a rebuttable presumption of compliance with the ATR requirements, but exercised its adjustment authority under TILA section 105(a) to limit the ability to rebut the presumption because the Bureau found that an open-ended rebuttable presumption would unduly restrict access to credit without a corresponding benefit to consumers. Id. at 6514.

251.  Davis et al., supra note 179.

252.  Bureau of Consumer Fin. Prot., Data Point: 2019 Mortgage Market Activity and Trends (June 2020), https://files.consumerfinance.gov/​f/​documents/​cfpb_​2019-mortgage-market-activity-trends_​report.pdf (4.6 percent of conventional, first-lien loans for purchasing one-to-four-family, owner-occupied, site-built homes). As explained in the Assessment Report, because of their nearly identical definitions, HMDA data regarding higher-priced mortgage loans (HPMLs) may serve as a proxy for higher-priced covered transactions under the ATR/QM Rule.

253.  See Bhutta & Ringo, supra note 223.

254.  Fed. Hous. Fin. Agency, Enterprise Regulatory Capital Framework Final Rule (2020), https://www.fhfa.gov/​SupervisionRegulation/​Rules/​Pages/​Enterprise-Regulatory-Capital-Framework-Final-Rule.aspx (Final Rule currently available on the FHFA website and awaiting Federal Register publication).

255.  Subordinate-lien transactions are discussed below in the section-by-section analysis of § 1026.43(e)(2)(vi).

256.  Assessment Report, supra note 63, section 5.5, at 187.

257.  See Bhutta & Ringo, supra note 223.

258.  As discussed in the section-by-section analysis of § 1026.43(e)(2)(vi)(A) below, this final rule generally provides that, for transactions that are covered by § 1026.43(e)(2)(vi)(A) and priced greater than or equal to 1.5 but less than 2.25 percentage points above APOR, the transaction receives a rebuttable QM (rather than a conclusive QM) presumption of compliance with the ATR requirements.

259.  As discussed in the section-by-section analysis of § 1026.43(e)(2)(vi)(A) below, this final rule generally provides that, for transactions that are covered by § 1026.43(e)(2)(vi)(A) and priced greater than or equal to 1.5 but less than 2.25 percentage points above APOR, the transaction receives a rebuttable QM (rather than a conclusive QM) presumption of compliance with the ATR requirements. The Bureau concludes that a General QM eligibility threshold lower than 2.25 percentage points could unduly limit some consumers to non-QM or FHA loans with materially higher costs, or no responsible, affordable loan at all, given the current lack of a robust non-QM market.

260.  78 FR 6408, 6513 (Jan. 30, 2013).

261.  Id. at 6512-13.

262.  Id. at 6513.

263.  Id. at 6505.

264.  Id. at 6513.

265.  78 FR 6408, 6511 (Jan. 30, 2013).

266.  Id. at 6528. The January 2013 Final Rule also did not include a DTI limit for balloon-payment QMs under § 1026.43(f). Id. at 6539.

267.  See id. at 6511 (“Moreover, requiring creditors to prove that they have satisfied the qualified mortgage requirements in order to invoke the presumption of compliance will itself ensure that the loans in question do not contain certain risky features and are underwritten with careful attention to consumers' debt-to-income ratios.”).

268.  Subordinate-lien transactions are discussed below in the section-by-section analysis of § 1026.43(e)(2)(vi).

269.  However, as discussed in the section-by-section analysis of § 1026.43(e)(2)(vi) below, under the final rule a loan is not eligible for QM status (i.e., will not receive any presumption of compliance with the ATR requirements) under § 1026.43(e)(2) if the loan exceeds the separate pricing thresholds in § 1026.43(e)(2)(vi), as finalized.

270.  78 FR 35430 (June 12, 2013).

271.  Id. at 35487 (“The Bureau continues to believe that consideration of debt-to-income ratio or residual income is fundamental to any determination of ability to repay. A consumer is able to repay a loan if he or she has sufficient funds to pay his or her other obligations and expenses and still make the payments required by the terms of the loan. Arithmetically comparing the funds to which a consumer has recourse with the amount of those funds the consumer has already committed to spend or is committing to spend in the future is necessary to determine whether sufficient funds exist.”).

272.  78 FR 6408, 6528 (Jan. 30, 2013) (“Unfortunately, however, the Bureau lacks sufficient data, among other considerations, to mandate a bright-line rule based on residual income at this time.”).

273.  Id. at 6527 (“Another consumer group commenter argued that residual income should be incorporated into the definition of QM. Several commenters suggested that the Bureau use the general residual income standards of the VA as a model for a residual income test, and one of these commenters recommended that the Bureau coordinate with FHFA to evaluate the experiences of the GSEs in using residual income in determining a consumer's ability to repay.”); id. at 6528 (“Finally, the Bureau acknowledges arguments that residual income may be a better measure of repayment ability in the long run. A consumer with a relatively low household income may not be able to afford a 43 percent debt-to-income ratio because the remaining income, in absolute dollar terms, is too small to enable the consumer to cover his or her living expenses. Conversely, a consumer with a relatively high household income may be able to afford a higher debt ratio and still live comfortably on what is left over.”).

274.  78 FR 6408, 6561 (Jan. 30, 2013) (“In some cases, lenders and borrowers entered into loan contracts on the misplaced belief that the home's value would provide sufficient protection. These cases included subprime borrowers who were offered loans because the lender believed that the house value either at the time of origination or in the near future could cover any default. Some of these borrowers were also counting on increased housing values and a future opportunity to refinance; others likely understood less about the transaction and were at an informational disadvantage relative to the lender.”); id. at 6564 (“During those periods there were likely some lenders, as evidenced by the existence of no-income, no-asset (NINA) loans, that used underwriting systems that did not look at or verify income, debts, or assets, but rather relied primarily on credit score and LTV.”); id. at 6559 (“If the lender is assured (or believes he is assured) of recovering the value of the loan by gaining possession of the asset, the lender may not pay sufficient attention to the ability of the borrower to repay the loan or to the impact of default on third parties. For very low LTV mortgages, i.e., those where the value of the property more than covers the value of the loan, the lender may not care at all if the borrower can afford the payments. Even for higher LTV mortgages, if prices are rising sharply, borrowers with even limited equity in the home may be able to gain financing since lenders can expect a profitable sale or refinancing of the property as long as prices continue to rise . . . . In all these cases, the common problem is the failure of the originator or creditor to internalize particular costs, often magnified by information failures and systematic biases that lead to underestimation of the risks involved. The first such costs are simply the pecuniary costs from a defaulted loan—if the loan originator or the creditor does not bear the ultimate credit risk, he or she will not invest sufficiently in verifying the consumer's ability to repay.”).

275.  The Bureau addresses comments on the Bureau's proposal regarding appendix Q in the section-by-section analysis for appendix Q, below.

276.  TILA section 129C(b)(3)(ii); 15 U.S.C. 1639c(b)(3)(ii).

277.  See, e.g., OMB Circular A-119: Federal Participation in the Development and Use of Voluntary Consensus Standards and in Conformity Assessment Activities (Jan. 27, 2016), https://www.whitehouse.gov/​wp-content/​uploads/​2020/​07/​revised_​circular_​a-119_​as_​of_​1_​22.pdf.

278.  See, e.g., 18 U.S.C. 1001, 1010, 1014, 1028, 1341 through 1344.

279.  The Bureau has also made some non-substantive changes to terminology in final comments 43(e)(2)(v)(B)-1 through -3 to ensure consistent usage of terms throughout the commentary.

280.  The referenced versions of the guides, or relevant sections thereof, are publicly available on the internet. The Fannie Mae Single Family Selling Guide, published June 3, 2020 can be found at http://www.allregs.com/​tpl/​public/​fnma_​freesiteconv_​tll.aspx. The Freddie Mac Single-Family Seller/Servicer Guide, published June 10, 2020 can be found at https://www.allregs.com/​tpl/​public/​fhlmc_​freesite_​tll.aspx. The FHA's Single Family Housing Policy Handbook, issued October 24, 2019 can be found at https://www.regulations.gov/​document?​D=​CFPB-2020-0020-0002. The chapter 4 of the VA's Lenders Handbook revised February 22, 2019 can be found at https://www.regulations.gov/​document?​D=​CFPB-2020-0020-0003. The USDA's Field Office Handbook for the Direct Single Family Housing Program, revised March 15, 2019 can be found at https://www.regulations.gov/​document?​D=​CFPB-2020-0020-0005. The USDA's Handbook for the Single Family Guaranteed Loan Program, revised March 19, 2020 can be found at https://www.regulations.gov/​document?​D=​CFPB-2020-0020-0004.

281.  As explained above in the section-by-section discussion of § 1026.43(e)(2)(v)(A), the Bureau proposed to move to § 1026.43(e)(2)(v)(A) the provisions in existing § 1026.43(e)(2)(vi)(B), which specify that the consumer's monthly DTI ratio is determined using the consumer's monthly payment on the covered transaction and any simultaneous loan that the creditor knows or has reason to know will be made.

282.  85 FR 50944, 50948 (Aug. 19, 2020).

283.  As discussed above in part V.C, the Bureau also received comments both for and against increasing the § 1026.43(b)(4) safe harbor threshold spread from 1.5 percentage points to 2 percentage points.

284.  See Kaul & Goodman, supra note 194.

285.  The analysis provided by the commenter looked at loans that had ever been 60 days or more delinquent, rather than 60 or more days delinquent during the first two years, which is the standard used in the Bureau's analysis.

286.  Davis et al., supra note 179.

287.  85 FR 50944, 50948 (Aug. 19, 2020).

288.  The Bureau stated in the January 2013 Final Rule that it believed a significant share of mortgages would be made under the general ATR standard. 78 FR 6408, 6527 (Jan. 30, 2013). However, the Assessment Report found that a robust market for non-QM loans above the 43 percent DTI limit has not materialized as the Bureau had predicted and, therefore, there is limited capacity in the non-QM market to provide access to credit after the expiration of the Temporary GSE QM loan definition. Assessment Report, supra note 63, at 198. As described above, the non-QM market has been further reduced by the recent economic disruptions associated with the COVID-19 pandemic, with most mortgage credit now available in the QM lending space. The Bureau acknowledges that the slow development of the non-QM market and the recent economic disruptions associated with the COVID-19 pandemic may significantly hinder its development in the near term.

289.  Assessment Report, supra note 63, at 10-11, 117, 131-47.

290.  The Bureau also acknowledges that Table 5 shows that for loans with DTI ratios of 61-70 in the 2002-2008 data, the early delinquency rates were 26 percent for loans priced 2.00 to 2.24 percentage points above APOR, relative to 22 percent for loans priced 1.75 to 2.00 percentage points above APOR.

291.  Similarly, Table 6 shows that for the DTI ratios with the highest early delinquency rates (DTI ratios of 26-35), the early delinquency rates were 4.4 percent for loans priced 2.00 or more percentage points over APOR, compared to 3.2 percent for loans priced 1.50 to 1.99 percentage points over APOR.

292.  Assessment Report, supra note 63, section 5.5, at 187.

293.  As discussed in part V.B.5 above, Table 7A shows that 96.3 percent of 2018 conventional first-lien purchase originations would have been QMs under this revised ATR/QM Rule including § 1026.43(e)(2)(vi)(A) with a threshold of 2.25 percentage points over APOR. Table 7A shows a 96.6 percent share if the threshold were instead increased to 2.5 percentage points over APOR.

294.  85 FR 41716, 41732 n.190 (July 10, 2020).

295.  On August 19, 2020, the Bureau issued a final rule adjusting the loan amounts for the limits on points and fees under § 1026.43(e)(3)(i), based on the annual percentage change reflected in the CPI-U in effect on June 1, 2020. 85 FR 50944 (Aug. 19, 2020). To ensure that the loan amounts for § 1026.43(e) remain synchronized, the Bureau is finalizing this rule with a threshold of $66,156, rather than a threshold of $65,939, and $110,260, rather than a threshold of $109,898.

296.  As noted above, and discussed in more detail below, the Bureau is increasing the loan amounts specified in § 1026.43(e)(2)(vi)(A) through (F) because the new adjustments for 2021 have been published. See 85 FR 50944 (Aug. 19, 2020).

297.  85 FR 41716, 41733 (July 10, 2020) (showing early delinquency rates for 2002-2008 first-lien purchase originations in NMDB data categorized according to both their DTI ratios and their approximate rate spreads).

298.  Mike Baker & Daniel Wagner, The mobile-home trap: How a Warren Buffet empire preys on the poor, The Seattle Times (Apr. 2, 2015), https://www.seattletimes.com/​business/​real-estate/​the-mobile-home-trap-how-a-warren-buffett-empire-preys-on-the-poor/​#:~:text=​Special%20Reports-,The%20mobile%20home%20trap%3A%20How%20a%20Warren%20Buffett,empire%20preys%20on%20the%20poor&​text=​Billionaire%20philanthropist%20Warren%20Buffett%20controls,loans%20and%20rapidly%20depreciating%20homes.

299.  85 FR 41716, 41736 (July 10, 2020) (showing the share of 2018 first-lien conventional purchase loans under various General QM loan definitions).

300.  The non-depository manufactured housing creditor specifically discussed the impact of a manufactured housing loan being subject to TILA's appraisal requirements for higher-priced mortgages because, without QM status, these loans would not be eligible for the exemption from these requirements under 12 CFR 1026.35(c)(2)(i).

301.  85 FR 41716, 41760 (July 10, 2020) (analyzing credit characteristics and loan performance for subordinate-lien transactions at various rate spreads and loan amounts (adjusted for inflation) using HMDA and Y-14M data).

302.  The Bureau's decisions to adopt basic pricing thresholds of 1.5 and 2.25 percentage points above APOR and to supplement them with higher pricing thresholds for smaller loans, for smaller loans secured by a manufactured home, and for subordinate-lien transactions are each independent of one another.

303.  Id. at 41757 n.270.

304.  85 FR 50944 (Aug. 19, 2020).

305.  See, e.g., 12 CFR 1026.35(c)(1)(iii).

306.  See 78 FR 6408, 6528 (Jan. 30, 2013).

307.  See Bureau of Labor and Statistics, Historical Consumer Price Index for All Urban Consumers (CPI-U), (Apr. 2020), https://www.bls.gov/​cpi/​tables/​supplemental-files/​historical-cpi-u-202004.pdf. (Using the CPI-U price index, nominal loan amounts are inflated to June 2020 dollars from the price level in June of the year prior to origination. This effectively categorizes loans according to the inflation-adjusted thresholds for smaller loans that would have been in effect on the origination date. The set of loans categorized within a given threshold remains the same as in the proposal, in which nominal loan amounts were inflated to June 2019 dollars and compared against the corresponding threshold levels of $65,939 and $109,898.)

308.  Portfolio loans made by small creditors, as defined in § 1026.35(b)(2)(iii)(B) and (C), are excluded, as such loans are likely Small Creditor QMs pursuant to § 1026.43(e)(5) regardless of pricing.

309.  85 FR 41716, 41732 n.190 (July 10, 2020). The Bureau also grouped loans with rate spreads of 2 percentage points or more to ensure a sufficient sample size for a reliable analysis of 2018 data in Tables 2 and 6 of the General QM Proposal. Id. at 41732 n.193.

310.  The Bureau grouped loans in General QM Proposal Table 10 for the same reasons. This grouping ensured a sufficient sample size for a reliable analysis of Y-14M data for subordinate-lien transactions.

311.  All estimates in Table 10 includes loans that meet the Small Creditor QM loan definition in § 1026.43(e)(5).

312.  The Bureau notes that one consequence of this 6.5 percent threshold and the other pricing thresholds in the final rule, like the pricing thresholds in the proposal, is that high-cost mortgages under HOEPA cannot qualify for General QM status. See 12 CFR 1026.32(a), 1026.34(a)(4), 1026.43(e)(3), (g)(1). Thus, for the reasons discussed in this final rule for adopting these pricing thresholds, the Bureau is no longer exercising authority under HOEPA to permit certain lower-DTI high-cost mortgages to qualify as General QMs. Cf. 78 FR 6855, 6861-62, 6924-25 (Jan. 31, 2013).

313.  Bureau of Consumer Fin. Prot., Introducing New and Revised Data Points in HMDA, at 207 (Aug. 2019), https://files.consumerfinance.gov/​f/​documents/​cfpb_​new-revised-data-points-in-hmda_​report.pdf.

314.  Id.

315.  For example, chattel loans are not subject to the TILA-RESPA Integrated Disclosure Rule. See 12 CFR 1026.19(e) and (f).

316.  78 FR 6408, 6506 (Jan. 30, 2013).

317.  See Bureau of Labor and Statistics, Historical Consumer Price Index for All Urban Consumers (CPI-U), (Apr. 2020), https://www.bls.gov/​cpi/​tables/​supplemental-files/​historical-cpi-u-202004.pdf. (Using the CPI-U price index, nominal loan amounts are inflated to June 2020 dollars from the price level in June of the year prior to origination. This effectively categorizes loans according to the inflation-adjusted thresholds for smaller loans that would have been in effect on the origination date. The set of loans categorized within a given threshold remains the same as in the proposal, in which nominal loan amounts were inflated to June 2019 dollars and compared against the corresponding threshold levels of $65,939 and $109,898.)

318.  As with its analysis of higher-priced, smaller loans above, the Bureau determined that it was necessary to use cumulative rate-spread ranges to ensure sufficient sample sizes for a reliable analysis of Y-14M data for subordinate lien loans. Without this cumulative grouping, the sample sizes for some rate-spread ranges would be insufficient for reliable analysis.

319.  The loan data were a subset of the supervisory loan-level data collected as part of the Federal Reserve's Comprehensive Capital Analysis and Review, known as Y-14M data. The early delinquency rate measured the percentage of loans that were 90 or more days late in the first two years. The Bureau used loans with payments that were 90 or more days late to measure delinquency, rather than the 60 or more days used with the data discussed above for first-lien transactions, because the Y-14M data do not include a measure for payments 60 or more days late. Data from a small number of lenders were not included due to incompatible formatting.

320.  The Bureau also stated that, under proposed § 1026.43(b)(4), an identical special rule for determining the APR for certain loans for which the interest rate may or will change also applies under that paragraph for purposes of determining whether a QM under § 1026.43(e)(2) is a higher-priced covered transaction and whether it is therefore subject to a rebuttable presumption as opposed to a conclusive presumption of compliance with the with the ATR requirements.

321.  A step-rate mortgage is a transaction secured by real property or a dwelling for which the interest rate will change after consummation and the rates that will apply and the periods for which they will apply are known at consummation. See 12 CFR 1026.18(s)(7)(ii).

322.  For example, many GSE ARM products provide for a 2 percentage point cap on the first reset.

323.  Regulation Z requirements for calculating the APR for ARMs are summarized below in the discussion of the structure and pricing particular to ARMs.

324.  As discussed above, the Bureau is also finalizing § 1026.43(b)(4), an identical special rule for determining the APR for certain loans for which the interest rate may or will change, which applies under that paragraph for purposes of determining whether a QM under § 1026.43(e)(2) is a higher-priced covered transaction.

325.  In addition to short-reset ARMs, the special rule applies to step-rate mortgages that have an initial fixed-rate period of five years or less. The Bureau recognizes that the interest rates of step-rate mortgages are known at consummation. However, unlike fixed-rate mortgages and akin to ARMs, the interest rate of step-rate mortgages changes, thereby raising the concern that interest-rate increases relatively soon after consummation may present affordability risks due to higher loan payments. Moreover, applying the APR determination requirement to such loans is consistent with the treatment of step-rate mortgages pursuant to the requirement in the General QM loan definition to underwrite loans using the maximum interest rate during the first five years after the date on which the first regular periodic payment will be due. See comment 43(e)(2)(iv)-3.iii.

326.  See 12 CFR 1026.17(c)(1) through (8).

327.  See 12 CFR 1026.17(c)(1) through (10).

328.  See TILA section 103(bb)(1)(B)(ii).

329.  See 12 CFR 1026.32(a)(3)-3.

330.  The lower absolute pricing of ARMs with comparable credit risk is reflected in the lower ARM APOR, which is typically 50 to 150 basis points lower than the fixed-rate APOR.

331.  Bureau analysis of NMDB data shows crisis-era short-reset ARMs had lower LTV ratios at consummation relative to comparably priced fixed-rate loans.

332.  This approach for ARMs is different from the approach in § 1026.43(c)(5) for underwriting ARMs under the ATR requirements, which, like the APR determination for HOEPA coverage for ARMs under § 1026.32(a)(3), is based on the greater of the fully indexed rate or the initial rate.

333.  As discussed below in the Legal Authority section, the Bureau is exercising its adjustment and revision authorities to amend § 1026.43(e)(2)(vi) to provide that, to determine the APR for short-reset ARMs for purposes of General QM status, the creditor must treat the maximum interest rate that may apply during that five-year period as the interest rate for the full term of the loan. The Bureau observes that the requirement in TILA section 129C(b)(2)(A)(v) to underwrite ARMs for QM purposes using the maximum interest rate that may apply during the first five years is at least ambiguous with respect to whether it independently obligates the creditor to determine the APR for short-reset ARMs in the same manner as the special rule, at least when the Bureau relies on pricing thresholds as the primary indicator of likely repayment ability in the General QM loan definition. Furthermore, the Bureau concludes that it would be reasonable, in light of the definition of a General QM and in light of the policy concerns already described, to construe TILA section 129C(b)(2)(A)(v) as imposing the same obligations as the special rule in § 1026.43(e)(2)(vi). Thus, in addition to relying on its adjustment and revision authorities to amend § 1026.43(e)(2)(vi), the Bureau concludes that it may do so under its general authority to interpret TILA in the course of prescribing regulations under TILA section 105(a) to carry out the purposes of TILA.

334.  Many ARMs in the data during this period do not report the time between consummation and the first interest-rate reset, and so are excluded from this analysis.

335.  Laurie Goodman et al., Urban Inst., Housing Finance at a Glance, at 9 (Sept. 2020), https://www.urban.org/​sites/​default/​files/​publication/​102979/​september-chartbook-2020.pdf.

336.  Assessment Report, supra note 63, at 94 (fig. 25).

337.  Id. at 93-95.

338.  Id. at 95 (fig. 26).

339.  As noted above, the special rule also applies to step-rate mortgages for which the interest rate changes in the first five years.

340.  12 CFR 1026.43(e)(2)(iv).

341.  This refers to the standard APOR for ARMs. The requirement modifies the determination for the APR of ARMs but does not affect the determination of the APOR. The Bureau notes that the APOR used for step-rate mortgages is the ARM APOR because, as with ARMs, the interest rate in step-rate mortgages adjusts and is not fixed. Thus, the APOR for fixed-rate mortgages would be inapt.

342.  Fannie Mae, Standard ARM Plan Matrix for 2020 (Apr. 2020), https://singlefamily.fanniemae.com/​media/​6951/​display.

343.  VA caps all interest rate increases at 1 percent a year for all VA ARMS. FHA caps interest rate increases at 1 percent for one-year and three-year ARMs. FHA caps annual interest rate increases at 1 percent for a lifetime cap of 5 percent or 2 percent increases for a lifetime cap of 6 percent.

344.  12 CFR 1026.43(e)(2)(iv).

345.  85 FR 41448 (July 10, 2020).

346.  78 FR 75215 (Dec. 11, 2013) (HUD); 79 FR 26620 (May 9, 2014) and 83 FR 50506 (Oct. 9, 2018) (VA); and 81 FR 26461 (May 3, 2016) (USDA).

347.  85 FR 67938 (Oct. 26, 2020).

348.  85 FR 41448, 41752 (July 10, 2020).

349.  See Assessment Report, supra note 63, at 11.

350.  See id. at 155.

351.  The Bureau's use of the term “mandatory compliance date” does not imply that creditors are required to use the General QM loan definition to comply with the ATR/QM Rule's ability-to-repay requirements. Unless a loan is eligible for QM status—such as under § 1026.43(e)(2), § 1026.43(e)(5), or § 1026.43(f)—a creditor must make a reasonable and good faith determination of the consumer's ability to repay and does not receive a presumption of compliance.

352.  This final rule uses the term “implementation period” to refer to the period between the date the Bureau issues this final rule and the date that creditors seeking to originate General QMs must comply with the General QM loan definition as amended by this final rule. Under the General QM Proposal, this implementation period would have ended on the effective date, while under this final rule the implementation period will end on the mandatory compliance date.

353.  The Bureau has separately proposed to amend Regulation Z to facilitate creditors' transition away from using LIBOR as an index for variable-rate consumer credit products. 85 FR 36938 (June 18, 2020).

354.  See Fannie Mae & Freddie Mac, Extended URLA Implementation Timeline (Apr. 14, 2020), https://singlefamily.fanniemae.com/​media/​22661/​display.

355.  85 FR 67938, 67951 (Oct. 26, 2020).

356.  The Bureau's use of the term “mandatory compliance date” does not imply that creditors are required to use the General QM loan definition to comply with the ATR/QM Rule's ability-to-repay requirements. Unless a loan is eligible for QM status—such as under § 1026.43(e)(2), § 1026.43(e)(5) or § 1026.43(f)—a creditor must make a reasonable and good faith determination of the consumer's ability to repay and does not receive a presumption of compliance.

357.  In that case, pursuant to the conservatorship clause, the Temporary GSE QM loan definition would expire with respect to that GSE on the date that GSE ceases to operate under conservatorship.

358.  The Seasoned QM Final Rule, which the Bureau is releasing simultaneously with this final rule, has an effective date of 60 days after publication of the final rule in the Federal Register. Unlike this final rule, there is no optional early compliance period for the Seasoned QM Final Rule.

359.  In the Extension Proposal, which the Bureau released concurrently with the General QM Proposal, the Bureau proposed to extend the Temporary GSE QM loan definition until the effective date of a final rule amending the General QM loan definition. See supra part III.C. Thus, when the Bureau issued the General QM Proposal, it expected that the Temporary GSE QM loan definition would expire on the effective date of this final rule, along with the current General QM loan definition (unless one or both of the GSEs were to cease to operate under conservatorship prior to the effective date). However, the Extension Final Rule extended the Temporary GSE QM loan definition until the mandatory compliance date, not the effective date, of a final rule amending the General QM loan definition. As a result, the Temporary GSE QM loan definition will be available until the mandatory compliance date of this final rule (July 1, 2021), unless one or both of the GSEs cease to operate under conservatorship prior to July 1, 2021. See supra part III.D.

360.  85 FR 67938, 67952 (Oct. 26, 2020).

361.  The ATR/QM Rule generally applies to closed-end consumer credit transactions that are secured by a dwelling, as defined in 12 CFR 1026.2(a)(19), including any real property attached to a dwelling. 12 CFR 1026.43(a). Therefore, the Rule applies to a dwelling, as defined in § 1026.19(a), whether or not it is attached to real property. In contrast, TRID generally applies to closed-end consumer credit transactions secured by real property or a cooperative unit. 12 CFR 1026.19(e)(1)(i). Therefore, some transactions that are a secured by a dwelling that is not considered real property under State or other applicable law will be subject to the ATR/QM Rule but not TRID.

362.  HMDA requires many financial institutions to maintain, report, and publicly disclose loan-level information about mortgages. These data help show whether creditors are serving the housing needs of their communities; they give public officials information that helps them make decisions and policies; and they shed light on lending patterns that could be discriminatory. HMDA was originally enacted by Congress in 1975 and is implemented by Regulation C. See Bureau of Consumer Fin. Prot., Mortgage data (HMDA), https://www.consumerfinance.gov/​data-research/​hmda/​.

363.  The NMDB, jointly developed by the FHFA and the Bureau, provides de-identified loan characteristics and performance information for a 5 percent sample of all mortgage originations from 1998 to the present, supplemented by de-identified loan and borrower characteristics from Federal administrative sources and credit reporting data. See Bureau of Consumer Fin. Prot., Sources and Uses of Data at the Bureau of Consumer Financial Protection, at 55-56 (Sept. 2018), https://www.consumerfinance.gov/​documents/​6850/​bcfp_​sources-uses-of-data.pdf. (Differences in total market size estimates between NMDB data and HMDA data are attributable to differences in coverage and data construction methodology.)

364.  The Assessment Report, the ANPR, the Extension Proposal, the General QM Proposal, and the Extension Final Rule used the term “High-DTI loans” to refer to loans with DTI ratios over 43 percent. For greater precision and because this final rule is eliminating the 43 percent DTI limit, this final rule instead uses the term “Over-43-Percent-DTI loans” to refer to such loans.

365.  84 FR 37155, 37158-59 (July 31, 2019).

366.  Id. at 37158-59.

367.  Id. at 37159.

368.  Id. The Bureau estimates that 616,000 of these loans were for home purchases, and 341,000 were refinance loans. In addition, the Bureau estimates that the share of these loans with DTI ratios over 45 percent has varied over time due to changes in market conditions and GSE underwriting standards, rising from 47 percent in 2016 to 56 percent in 2017, and further to 69 percent in 2018.

369.  Id. at 37159.

370.  Id. at 37159 n.58. Where these types of loans have DTI ratios above 43 percent, they would be captured in the estimate above relating to Over-43-Percent-DTI GSE loans.

371.  For example, in qualitative responses to the Bureau's Lender Survey conducted as part of the Assessment, underwriting for self-employed borrowers was one of the most frequently reported sources of difficulty in originating mortgages using appendix Q. These concerns were also raised in comments submitted in response to the Assessment RFI, noting that appendix Q is ambiguous with respect to how to treat income for consumers who are self-employed, have irregular income, or want to use asset depletion as income. See Assessment Report, supra note 63, at 200.

372.  Id. at 107 (“For context, total jumbo purchase originations increased from an estimated 108,700 to 130,200 between 2013 and 2014, based on nationally representative NMDB data.”).

373.  Id. at 118 (“The Application Data indicates that, notwithstanding concerns that have been expressed about the challenge of documenting and verifying income for self-employed borrowers under the General QM standard and the documentation requirements contained in appendix Q to the Rule, approval rates for non-High DTI, non-GSE eligible self-employed borrowers have decreased only slightly, by 2 percentage points . . . .”).

374.  See part V.B for additional discussion of concerns raised about appendix Q.

375.  This estimate includes only HMDA loans which have a reported DTI and rate spread over APOR, and thus may underestimate the true number of loans gaining QM status under the proposal.

376.  The Bureau expects consumers could continue to obtain FHA loans where such loans were cheaper or preferred for other reasons.

377.  Based on NMDB data, the Bureau estimates that the average loan amount among High-DTI GSE borrowers in 2018 was $250,000. While the time to repayment for mortgages varies with economic conditions, the Bureau estimates that half of mortgages are typically closed or paid off five to seven years into repayment. Payment comparisons based on typical 2018 HMDA APRs for GSE loans, 5 percent for borrowers with credit scores over 720, and 6 percent for borrowers with credit scores below 680 and LTVs exceeding 85 percent.

378.  This approximation assumes $4,000 in savings from total loan costs for all 959,000 consumers. Actual expected savings would vary substantially based on loan and credit characteristics, consumer choices, and market conditions.

379.  See Assessment Report supra note 63, at 10-11, 117, 131-47.

380.  In the proposal, the Bureau stated that 8,000 to 59,000 additional loans annually would become delinquent within two years of origination under the proposal. The Bureau's has revised its range of estimates under the proposal to 8,000 to 56,000.

381.  The comparable thresholds are 6.5 percentage points over APOR for loans priced under $66,156, 3.5 percentage points over APOR for loans priced under $110,260 but at or above $66,156, and 6.5 percentage points over APOR for loans for manufactured housing priced under $110,260.

382.  Alternative approaches, such as retaining a DTI limit of 45 or 50 percent, would have had similar effects of allowing small depository creditors to originate more GSE loans under an expanded General QM loan definition relative to the baseline, while offsetting potential competitive advantages for small depository creditors that originate Small Creditor QMs.

383.  These statistics are estimated based on originations from the first nine months of the year, to allow time for loans to be sold before HMDA reporting deadlines. In addition, a higher share of Over-43-Percent-DTI conventional purchase non-rural loans (33.3 percent) report being sold to other non-GSE purchasers compared to rural loans (22.3 percent).

384.  For alternative approaches, the Bureau estimates 83.3 percent of conventional purchase loans for homes in rural areas would have been QMs under a DTI limit of 45 percent, and 95.1 percent of conventional purchase loans for homes in rural areas would have been QMs under a DTI limit of 50 percent.

385.  5 U.S.C. 601(3) (the Bureau may establish an alternative definition after consultation with the Small Business Administration and an opportunity for public comment).

386.  5 U.S.C. 603 through 605.

387.  5 U.S.C. 609.

388.  Non-depositories are classified as small entities if they had fewer than 5,188 total originations in 2018. The classification for non-depositories is based on the SBA small entity definition for mortgage companies (less than $41.5 million in annual revenues) and an estimate of $8,000 for revenue-per-origination from the Assessment Report, supra note 63, at 78. The HMDA data do not directly distinguish mortgage brokers from mortgage companies, so the more inclusive revenue threshold is used.

389.  In addition, all loans for manufactured housing under $110,260 with APR less than 6.5 percentage points over APOR can be originated as General QMs, assuming they meet all other General QM requirements.

390.  44 U.S.C. 3501 et seq.

391.  5 U.S.C. 801 et seq.