2018-02649. 2018-2020 Enterprise Housing Goals

AGENCY:

Federal Housing Finance Agency.

ACTION:

Final rule.

SUMMARY:

The Federal Housing Finance Agency (FHFA) is issuing a final rule on the housing goals for Fannie Mae and Freddie Mac (the Enterprises) for 2018 through 2020. The Federal Housing Enterprises Financial Safety and Soundness Act of 1992 (the Safety and Soundness Act) requires FHFA to establish annual housing goals for mortgages purchased by the Enterprises. The housing goals include separate categories for single-family and multifamily mortgages on housing that is affordable to low-income and very low-income families, among other categories.

The final rule establishes the benchmark levels for each of the housing goals and subgoals for 2018 through 2020. In addition, the final rule makes a number of clarifying and conforming changes, including revisions to the requirements for the housing plan that an Enterprise may be required to submit to FHFA in response to a failure to achieve one or more of the housing goals or subgoals.

DATES:

The final rule is effective on March 14, 2018.

FOR FURTHER INFORMATION CONTACT:

Ted Wartell, Manager, Housing & Community Investment, Division of Housing Mission and Goals, at [phone]. This is not a toll-free number. The mailing address is: Federal Housing Finance Agency, 400 Seventh Street SW, Washington, DC 20219. The telephone number for the Telecommunications Device for the Deaf is [phone].

SUPPLEMENTARY INFORMATION:

I. Background

A. Statutory and Regulatory Background for the Existing Housing Goals

The Safety and Soundness Act requires FHFA to establish annual housing goals for several categories of both single-family and multifamily mortgages purchased by Fannie Mae and Freddie Mac.^[1] The annual housing goals are one measure of the extent to which the Enterprises are meeting their public purposes, which include “an affirmative obligation to facilitate the financing of affordable housing for low- and moderate-income families in a manner consistent with their overall public purposes, while maintaining a strong financial condition and a reasonable economic return.” ^[2]

The housing goals provisions of the Safety and Soundness Act were substantially revised in 2008 with the enactment of the Housing and Economic Recovery Act, which amended the Safety and Soundness Act.^[3] Under this revised structure, FHFA established housing goals for the Enterprises for 2010 and 2011 in a final rule published on September 14, 2010.^[4] FHFA established housing goals levels for the Enterprises for 2012 through 2014 in a final rule published on November 13, 2012.^[5] In a final rule published on September 3, 2015, FHFA announced the housing goals for the Enterprises for 2015 through 2017, including a new small multifamily low-income housing subgoal.^[6]

Single-family goals. The single-family goals defined under the Safety and Soundness Act include separate categories for home purchase mortgages for low-income families, very low-income families, and families that reside in low-income areas. Performance on the single-family home purchase goals is measured as the percentage of the total home purchase mortgages purchased by an Enterprise each year that qualify for each goal or subgoal. There is also a separate goal for refinancing mortgages for low-income families, and performance on the refinancing goal is determined in a similar way.

Under the Safety and Soundness Act, the single-family housing goals are limited to mortgages on owner-occupied housing with one to four units total. The single-family goals cover conventional, conforming mortgages, defined as mortgages that are not insured or guaranteed by the Federal Housing Administration (FHA) or another government agency and with principal balances that do not exceed the loan limits for Enterprise mortgages.

Market measurement. The performance of the Enterprises on the single-family housing goals is evaluated using a two-part approach, which compares the goal-qualifying share of the Enterprise's mortgage purchases to two separate measures: A benchmark level and a market level. FHFA considered alternatives to this method in the 2015-2017 housing goals rulemaking and determined that the two-part approach continued to be the most appropriate method for evaluating performance on the single-family goals. FHFA is continuing that approach in this final rule.

In order to meet a single-family housing goal or subgoal, the percentage of mortgage purchases by an Enterprise that meet each goal or subgoal must meet or exceed either the benchmark level or the market level for that year. The benchmark level is set prospectively by rulemaking based on various factors, including FHFA's forecast of the goal-qualifying share of the overall market for each year. The market level is determined retrospectively each year, based on the actual goal-qualifying share of the overall market as measured by FHFA based on Home Mortgage Disclosure Act (HMDA) data for that year. The overall mortgage market that FHFA uses for both the prospective market forecasts and the retrospective market measurement consists of all single-family owner-occupied conventional conforming mortgages that would be eligible for purchase by either Enterprise. It includes loans reported in HMDA as sold to the Enterprises as well as comparable loans reported to HMDA as held in a lender's portfolio. It also includes comparable loans that are reported in HMDA as “sold to others.” This category includes loans reported as sold to Farmer Mac, private securitization, commercial banks, savings banks, life insurance companies, credit unions, mortgage bank and finance companies and their affiliates. Because HMDA data is reported as of a single point in time, the same loan could be reported in any of these categories in a particular calendar year, regardless of the ultimate disposition of the loan.

The market as measured based on HMDA data is different from the “actual market” of loans that an Enterprise may purchase for purposes of meeting the goals. Both the benchmark level and the retrospective market level measure the goal-qualifying share of the overall market for the year in question and exclude “seasoned loans.” Seasoned loans are loans that were originated in prior years and acquired by the Enterprise in the current year. While both the benchmark and the retrospective market measure are designed to measure the current year's mortgage originations, the performance of the Enterprises on the housing goals includes all Enterprise purchases in that year, regardless of the year in which the loan was originated. This provides housing goals credit when the Enterprises acquire qualified seasoned loans. The Enterprises' acquisition of seasoned loans provides an important source of liquidity for this market segment.

The market as measured based on HMDA data is also different from the “actual market” because the “actual market” includes loans from institutions that are not required to report under HMDA.^[7] For instance, Bhutta, Laufer, and Ringo (2017) estimate that loans in HMDA data for 2016 represented 90% of the first-lien, home purchase and refinance loans found in Equifax's consumer credit files.^[8]

The differences between the market as measured based on HMDA data and the “actual market” of loans available for purchase by the Enterprises may help explain why Enterprise performance on the income-based home purchase goals generally do not coincide with the market as measured by HMDA. As noted by commenters on the proposed rule, between 2010-2015, each Enterprise met the retrospective HMDA market level for the low-income home purchase goal in only one year (2014 for Fannie Mae and 2010 for Freddie Mac), and only one Enterprise met the retrospective HMDA market level for the very low-income home purchase goal in one year (2014 for Fannie Mae). While the performance of the Enterprises has generally lagged the retrospective HMDA market levels, particularly for the income-based home purchase goals, FHFA continues to believe that the HMDA market levels represent feasible targets for the Enterprises. FHFA expects the Enterprises to continue to make efforts to meet the retrospective HMDA market levels, consistent with maintaining safe and sound credit quality standards, regardless of whether the market levels exceed or fall below the benchmark levels.

Recent changes to the HMDA regulations will likely result in the HMDA data covering an even greater portion of the single-family mortgage market.^[9] The changes will also provide more detailed information about the loans included in the HMDA data. The changes to the HMDA regulations generally took effect at the start of 2018, so the new, more detailed information will not be available until after the 2018 performance year.

FHFA has considered the possible impact that certain changes to the HMDA regulations may have on the Enterprise housing goals. However, at this time the impact that such changes might have on the retrospective measure of the market is uncertain. FHFA is not making any changes to the Enterprise housing goals in anticipation of the revised HMDA data. FHFA will assess the impact of the changes and, if necessary, may propose changes to the housing goals regulation at a later date.

Multifamily goals. The multifamily goals defined under the Safety and Soundness Act include separate categories for mortgages on multifamily properties (properties with five or more units) with rental units affordable to low-income families and for mortgages on multifamily properties with rental units affordable to very low-income families. FHFA has also established by regulation a small multifamily low-income subgoal for properties with 5-50 units. The multifamily goals evaluate the performance of the Enterprises based on numeric targets, not percentages, for the number of affordable units in properties backed by mortgages purchased by an Enterprise. The regulation does not include a retrospective market level measure for the multifamily goals and subgoals, due in part to a lack of comprehensive data about the multifamily market such as that provided by HMDA for single-family mortgages. As a result, FHFA currently measures Enterprise multifamily goals performance against the benchmark levels only. The expanded HMDA fields that will be available for the 2018 performance year are expected to include information on the number of units in the properties securing each multifamily loan and should be helpful in evaluating performance for this market segment.

B. Adjusting the Housing Goals

Under the housing goals regulation first established by FHFA in 2010, as well as under this final rule, FHFA may reduce the benchmark levels for any of the single-family or multifamily housing goals in a particular year without going through notice and comment rulemaking based on a determination by FHFA that (1) market and economic conditions or the financial condition of the Enterprise require a reduction, or (2) “efforts to meet the goal or subgoal would result in the constraint of liquidity, over-investment in certain market segments, or other consequences contrary to the intent of the Safety and Soundness Act or the purposes of the Charter Acts.” ^[10] The housing goals regulation also takes into account the possibility that achievement of a particular housing goal may or may not have been feasible for the Enterprise. If FHFA determines that a housing goal was not feasible for the Enterprise to achieve, then the regulation provides for no further enforcement of that housing goal for that year.^[11]

If after publication of this final rule FHFA determines that any of the single-family or multifamily housing goals should be adjusted in light of market conditions, to ensure the safety and soundness of the Enterprises, or for any other reason, FHFA will take steps as necessary and appropriate to adjust that goal. Such steps could include adjusting the benchmark levels through the processes in the existing regulation or establishing revised housing goal levels through notice and comment rulemaking.

C. Housing Goals Under Conservatorship

On September 6, 2008, FHFA placed each Enterprise into conservatorship. Although the Enterprises remain in conservatorship at this time, they continue to have the mission of supporting a stable and liquid national market for residential mortgage financing. FHFA has continued to establish annual housing goals for the Enterprises and to assess their performance under the housing goals each year during conservatorship.

II. Proposed Rule and Comments

FHFA published a proposed rule in the Federal Register on July 7, 2017 that proposed benchmark levels for each of the single-family and multifamily housing goals and technical changes to the regulations.^[12] The comment period ended on September 5, 2017.

FHFA received 24 comment letters on the proposed rule, representing the views of more than 40 organizations and individuals. Comments were submitted by seven individuals; eight policy advocacy organizations; seven trade associations representing lenders, home builders, credit unions, and other mortgage market participants; and Fannie Mae and Freddie Mac. FHFA has reviewed and considered all of the comments. A number of comment letters raised issues unrelated to the housing goals or beyond the scope of the proposed rule, and those comments are not addressed in this final rule. Specific provisions of the proposed rule, and the comments received on those provisions, are discussed below and throughout this final rule.

Qualitative Measures. Four commenters—a trade organization, an advocacy organization, and Fannie Mae and Freddie Mac—suggested that FHFA consider qualitative efforts when evaluating the performance of the Enterprises under the housing goals, including building partnerships with community-based organizations and developing new or innovative products. Freddie Mac highlighted efforts like outreach, education, and relationship building with organizations, which take time and energy to create and maintain, but noted that these activities are not technically counted until they result in actual loan purchases. Fannie Mae stated that qualitative measures should be taken into account when determining whether the goals were met. Fannie Mae also suggested that qualitative measures should be considered by FHFA in determining whether an Enterprise should be required to submit a housing plan. FHFA recognizes that the quantitative performance outcomes of the Enterprises may not fully reflect the efforts that the Enterprises have made in seeking to improve their performance. In particular, quantitative measures will not always reflect the impact of market developments outside the control of the Enterprises that may have a significant impact on the ability of the Enterprises to meet the housing goals. On the other hand, quantitative benchmarks provide a bright line for measuring performance that qualitative measures do not. In addition, FHFA does take into account the qualitative efforts of the Enterprises in attempting to meet the housing goals when FHFA assesses the feasibility of any housing goals that an Enterprise fails to achieve, as well as whether to require an Enterprise to submit a housing plan if the Enterprise fails to achieve a goal that was feasible. On balance, FHFA remains unconvinced about the value of adding qualitative factors to the benchmarks or of replacing quantitative benchmarks against which progress can be objectively measured.

Single-Family Rental. Two commenters discussed the treatment of single-family rental housing under the goals, recognizing that this is still an emerging segment. One comment letter (representing multiple consumer advocacy groups) noted that “while our organizations have significant concerns about the Enterprises financing investors in the single-family rental market, if this financing becomes more firmly established as part of the Enterprise multifamily channel, it is critical that FHFA develop a goal that addresses affordability in this context.” Further, noting that “the Enterprises have always played a part in single-family rental by financing 2-4 unit properties owned by an owner-occupant,” the letter recommended that FHFA offer “bonus credit for owner-occupied 2-4 unit properties . . . when the owner has participated in a certified counseling program that includes landlord training.”

The other comment letter (from a trade organization) encouraged FHFA to develop an approach to single-family rental as a part of the multifamily goals and to provide clarity on whether single-family rental will be counted for multifamily housing goals, and if so, how it will be categorized and measured.

FHFA is actively monitoring this market segment and developing an overall regulatory approach to single-family rental. The housing goals regulation permits FHFA to “determine whether and how any transaction or class of transactions shall be counted for purposes of the housing goals.” ^[13] FHFA may provide specific guidance to the Enterprises under this provision that may allow the Enterprises to count some single-family rental properties that are financed as multifamily transactions toward the performance of the Enterprises on the multifamily housing goals. Any such guidance would be subject to appropriate limits to ensure that the overall multifamily housing goals continue to provide meaningful incentives for the Enterprises in the categories targeted by the housing goals. FHFA may also consider options to address single-family rental properties more systematically through future notice and comment rulemaking.

Manufactured Housing—Chattel. One commenter stated that “loans for owner-occupied real property and chattel manufactured home have always counted toward single-family housing goals, provided they meet the appropriate income threshold for the goal.” This statement is not an accurate description of the housing goals regulation. Prior to 2010, the regulation defined the term “mortgage” to include a loan secured by “a manufactured home that is personal property under the laws of the State in which the manufactured home is located.” FHFA revised the definition in 2010 to remove this language and thus to exclude chattel loans on manufactured housing from coverage under the housing goals regulation. The Supplementary Information for the 2010 final rule recognized that the role of the Enterprises with respect to chattel loans on manufactured housing was subject to change, and also stated that “FHFA may revise the definition of `mortgage' in future rulemaking to ensure conformance with the final regulation on duty to serve.” ^[14]

In December 2016, FHFA published a final rule implementing the statutory requirements for the Fannie Mae and Freddie Mac Duty to Serve underserved markets. The Duty to Serve final rule does not require the Enterprises to purchase chattel loans on manufactured housing, but the final rule does permit the Enterprises to receive Duty to Serve credit for such purchases to the extent that the Enterprises choose to pursue a pilot initiative for chattel loans on manufactured housing and any required FHFA approvals are received.^[15]

While both Enterprises have adopted Duty to Serve plans to pursue pilot initiatives for chattel loans on manufactured housing, those plans are still in the early stages. In addition, because neither Enterprise has purchased chattel loans on manufactured housing in recent years, there is limited data available on the market for such loans or their performance. As a result, FHFA would be unable to set benchmark levels for this market segment or assess the impact of any Enterprise purchases on their housing goals performance. Due to the limited information available at this time, the final rule does not make any change to the housing goals treatment of chattel loans on manufactured housing. FHFA may propose changes in a future rulemaking based on its assessment of additional information that may become available, especially from Enterprise chattel pilot activities. If FHFA does propose a change in the definition of the term “mortgage” to include chattel loans on manufactured housing, FHFA will also need to size this market segment and appropriately adjust the benchmark levels upwards to reflect the new definition.

Blanket loans on Manufactured Housing Communities (MHCs). The housing goals regulation does not explicitly address blanket loans on MHCs, but FHFA has interpreted the regulation to exclude blanket loans on MHCs from counting toward the performance of the Enterprises under the multifamily housing goals. In the 2015-2017 Enterprise housing goals proposed rule, FHFA requested comment on whether such loans should be counted. FHFA received a number of comments at that time supporting housing goals credit for blanket loans on MHCs, but the final rule did not adopt that change due to the difficulty of accurately determining “a manufactured housing unit's affordability under the housing goals, because bedroom count information on individual manufactured housing units in the communities is not collected by the Enterprises, and the pad rent alone does not include the full cost of housing for the residents, which includes paying for their unit financing.” ^[16]

One commenter on the July 17, 2017 proposed rule stated that blanket loans on MHCs should be included for counting toward the housing goals, arguing that it would be inconsistent to include them in the Duty to Serve regulation but not in the housing goals regulation. The commenter stated that goals eligibility should include investor-owned rental communities as well as resident-owned communities, arguing that the former are the dominant segment of the MHC segment. The commenter further argued that housing goals credit should be limited to occupied units located in the community rather than the total number of rental spaces available. Given the large volume of the segment, the commenter asserted the proposed multifamily goals should be increased to “reflect the expanded scope of the housing goals.”

Both Enterprises renewed their requests for FHFA to provide housing goals credit for blanket loans on manufactured housing communities (MHCs). Freddie Mac also suggested a different affordability standard than either of the two affordability methods defined in the Duty to Serve regulation. The Duty to Serve regulation includes two methods for estimating the number of units that could be counted as “affordable” for purposes of receiving Duty to Serve credit. For an MHC owned by a government unit or instrumentality, a nonprofit organization, or the residents, units in the MHC may be treated as affordable for Duty to Serve purposes if they are subject to affordability restrictions under laws or regulations governing the affordability of the community, or the community's or ownership entity's founding, chartering, governing, or financing documents. The Duty to Serve regulation also allows affordability for blanket loans on MHCs to be determined by estimating the affordability of units in the community based on the median income of the census tract in which the MHC is located. Freddie Mac proposed instead that FHFA determine affordability under the housing goals for blanket loans on MHCs based on an estimated “MHC Adjustment Factor” that would estimate the total housing cost for manufactured housing units based on the actual site rent plus an estimated utility allowance and an estimated additional amount to reflect the cost of the unit itself (including insurance and taxes).^[17]

FHFA does not believe that it would be inconsistent to allow credit for blanket loans on MHCs under Duty to Serve while not allowing credit for such loans under the housing goals. The scope of activities included under the Duty to Serve regulation differs from the scope of activities covered by the housing goals. The Duty to Serve regulation addresses certain specific market segments identified by Congress in the Safety and Soundness Act, one of which is manufactured housing, and appropriately includes credit related to blanket loans on MHCs. In contrast, the housing goals are directed at the full range of Enterprise loan purchase activities and are designed to evaluate the performance of the Enterprises particularly in serving low- and very low-income borrowers and renters. While FHFA has determined not to include credit for blanket loans on MHCs in this final rule, FHFA will continue to monitor this market segment. Moreover, as discussed in more detail below, FHFA exempts blanket loans on MHCs from the annual Conservatorship Scorecard cap on multifamily mortgage purchases, to avoid discouraging the flow of capital to the MHC sector.

III. Summary of Final Rule

A. Benchmark Levels for the Single-Family Housing Goals

The final rule establishes the benchmark levels for the single-family housing goals and subgoal for 2018-2020 as follows:

Goal	Criteria	Benchmark level for 2015-2017 (percent)	Benchmark level for 2018-2020 (percent)
Low-Income Home Purchase Goal	Home purchase mortgages on single-family, owner-occupied properties with borrowers with incomes no greater than 80 percent of area median income	24	24
Very Low-Income Home Purchase Goal	Home purchase mortgages on single-family, owner-occupied properties with borrowers with incomes no greater than 50 percent of area median income	6	6
Low-Income Areas Home Purchase Subgoal 18	Home purchase mortgages on single-family, owner-occupied properties with:
	• Borrowers in census tracts with tract median income no greater than 80 percent of area median income; or	14	14
	• Borrowers with incomes no greater than 100 percent of area median income in census tracts where (i) tract income is less than 100 percent of area median income, and (ii) minorities comprise at least 30 percent of the tract population.
Low-Income Refinancing Goal	Refinancing mortgages on single-family, owner-occupied properties with borrowers with incomes no greater than 80 percent of area median income	21	21

B. Multifamily Housing Goal Levels

The final rule establishes the levels for the multifamily goal and subgoals for 2018-2020 as follows:

Goal	Criteria	Goal level for 2017 (units)	Goal level for 2018-2020 (units)
Low-Income Multifamily Goal	Units affordable to families with incomes no greater than 80 percent of area median income in multifamily rental properties with mortgages purchased by an Enterprise	300,000	315,000
Very Low-Income Multifamily Subgoal	Units affordable to families with incomes no greater than 50 percent of area median income in multifamily rental properties with mortgages purchased by an Enterprise	60,000	60,000
Small Multifamily Low-Income Subgoal	Units affordable to families with incomes no greater than 80 percent of area median income in small multifamily rental properties (5 to 50 units) with mortgages purchased by an Enterprise	10,000	10,000

C. Other Changes

The final rule makes changes and clarifications to the existing regulation, including minor technical changes to some regulatory definitions. The final rule also revises the requirements applicable to the housing plan an Enterprise may be required to submit based on a failure to achieve one or more of the housing goals.

IV. Single-Family Housing Goals

This final rule establishes the single-family housing goals for 2018-2020. FHFA considered the required statutory factors described below in setting the benchmark levels for the single-family housing goals. FHFA's analysis and goal setting process includes developing market forecast models for each of the single-family housing goals, as well as considering a number of other variables that impact affordable homeownership. Many of these variables indicate that low-income and very low-income households are facing, and will continue to face, difficulties in achieving homeownership or in refinancing an existing mortgage. These factors, such as rising property values and stagnant household incomes, also impact the Enterprises' ability to meet their mission and facilitate affordable homeownership for low-income and very low-income households. Nevertheless, FHFA expects and encourages the Enterprises to work toward meeting their housing goal requirements in a safe and sound manner. This may include steps the Enterprises take to fulfill FHFA's expectations for supporting access to credit expressed in the Conservatorship Scorecard, which requires the Enterprises to undertake a number of research and related efforts including the development of pilots and initiatives.^[19]

A. Setting the Single-Family Housing Goal Levels

FHFA Process for Setting the Single-Family Benchmark Levels

Section 1332(e)(2) of the Safety and Soundness Act requires FHFA to consider the following seven factors in setting the single-family housing goals:

1. National housing needs;

2. Economic, housing, and demographic conditions, including expected market developments;

3. The performance and effort of the Enterprises toward achieving the housing goals in previous years;

4. The ability of the Enterprises to lead the industry in making mortgage credit available;

5. Such other reliable mortgage data as may be available;

6. The size of the purchase money conventional mortgage market, or refinance conventional mortgage market, as applicable, serving each of the types of families described, relative to the size of the overall purchase money mortgage market or the overall refinance mortgage market, respectively; and

7. The need to maintain the sound financial condition of the Enterprises.^[20]

FHFA has considered each of these seven statutory factors in setting the benchmark levels for each of the single-family housing goals and subgoal.

Recognizing that some of the factors required by statute to be considered can be readily captured using reliable data series while others cannot, FHFA implemented the following approach. FHFA's statistical market models considered factors that are captured through well-known and established data series, and these are then used to generate a point forecast for each goal, as well as a confidence interval for the point forecast. FHFA then considered the remaining statutory factors, as well as other relevant policy factors, in selecting the specific point forecast within the confidence interval as the benchmark level. FHFA's market forecast models incorporate four of the seven statutory factors: National housing needs; economic, housing, and demographic conditions; other reliable mortgage data; and the size of the purchase money conventional mortgage market or refinance conventional mortgage market for each single-family housing goal. The market forecast models generated a point estimate, as well as a confidence interval. FHFA then considered the remaining three statutory factors (historical performance and effort of the Enterprises toward achieving the housing goal; ability of the Enterprises to lead the industry in making mortgage credit available; and need to maintain the sound financial condition of the Enterprises), as well as other relevant policy factors, in selecting the specific point forecast within the confidence interval as the benchmark level for the goal period.

Market forecast models. The purpose of FHFA's market forecast models is to forecast the market share of the goal-qualifying mortgage originations in the market for the 2018-2020 period. The models are intended to generate reliable forecasts rather than to test various economic hypotheses about the housing market or to explain the relationship between variables. Following standard practice among forecasters and economists at other federal agencies, FHFA estimated a reduced-form equation for each of the housing goals and fit an Autoregressive Integrated Moving Average (or ARIMA) model to each goal share. The models look at the statistical relationship between (a) the historical market share for each single-family housing goal or subgoal, as calculated from monthly HMDA data, and (b) the historical values for various factors that may influence the market shares, e.g., interest rates, inflation, house prices, home sales, the unemployment rate, and other factors. The models then project the future value of the affordable market share using forecast values of the model inputs. FHFA developed separate models for each of the single-family housing goals and subgoal.

FHFA has employed similar models in past housing goals rulemakings to generate market forecasts. The models were developed using monthly series generated from HMDA and other data sources, and the resulting monthly forecasts were then averaged into an annual forecast for each of the three years in the goal period. The models rely on 13 years of HMDA data, from 2004 to 2016, the latest year for which HMDA data are available. Additional discussion of the market forecast models can be found in an updated research paper, available at http://www.fhfa.gov/​PolicyProgramsResearch/​Research/​.^[21]

In the final rule establishing the housing goals for 2015-2017, FHFA stated that it would engage directly with commenters to obtain detailed feedback on FHFA's econometric models for the housing goals. Throughout 2016, FHFA met with industry modeling experts about potential improvements to the econometric models. Considering input received, FHFA has revised the market forecast models to include better specifications and new variables for all goal-qualifying shares, while still following generally accepted practices and standards adopted by economists, including those at other federal agencies. During the model development process, FHFA grouped factors that are expected by housing market economists to have an impact on the market share of affordable housing into seven broad categories. For each category of variables, many variables were tested but only retained when they exhibited predictive power. The new set of models includes new driver variables that reflect factors that impact the affordable housing market—for example, household debt service ratio, labor force participation rate, and underwriting standards.

As is the case with any forecasting model, the accuracy of the forecast will vary depending on the accuracy of the inputs to the model and the length of the forecast period. FHFA has attempted to minimize the first source of variability by using third party forecasts published by Moody's and other accredited mortgage market forecasters. The second source of variability is harder to address. The models underlying this final rule rely on the most up-to-date data available as of November 2017, and use forecasted input values for the rest of 2017 (depending on the data series) to produce the forecasts for 2018-2020. The confidence intervals for the benchmark levels become wider as the forecast period lengthens. In other words, it becomes more likely that the actual market levels will be different from the forecasts the farther into the future the forecasts attempt to make predictions. Predicting three years out is not the usual practice in forecasting. A number of industry forecasters, including the Mortgage Bankers Association (MBA), Fannie Mae and Freddie Mac, do not publish forecasts beyond two years because accuracy of forecasts decreases substantially beyond a two-year period.

Market outlook. There are many factors that impact the affordable housing market as a whole, and changes to any one of them may significantly impact the ability of the Enterprises to meet the goals. In developing the market models, FHFA used Moody's forecasts, where available, as the source for macroeconomic variables.^[22] In cases where Moody's forecasts were not available (for example, the share of government-guaranteed/insured home purchases and the share of government-guaranteed/insured refinances), FHFA generated and tested its own forecasts.^[23] Elements that impact the models and the determination of benchmark levels are discussed below.

Interest rates are arguably one of the most important variables in determining the trajectory of the mortgage market. The Federal Reserve launched its “interest rate normalization” process in December 2015 with a 0.25 percentage point increase. In the September 2017 meeting of the Federal Open Market Committee (FOMC), the FOMC indicated a commitment to a low federal funds rate policy for the time being. Storm-related disruptions and rebuilding, resulting from hurricanes Harvey, Irma, and Maria, are expected to affect economic activity in the near term.^[24] However, there is some consensus among economists that the Federal Reserve will resume rate hikes if the economic signals indicate a need for it. Mortgage interest rates—in particular the 30-year fixed rate, which is closely tied to the federal funds rate and the 10-year Treasury note yield—are expected (in Moody's forecasts) to rise gradually from the historic low of 3.4 percent in August 2016 to 4.8 percent by 2020.

The unemployment rate has fallen steadily over the last few years to 4.1 percent in November 2017.^[25] Moody's forecasts expect it to remain around the same levels, between 4.1 and 4.5 percent over the next three years, given the expected growth of the economy at the modest range of 2.0 to 2.4 percent per year. Per capita disposable nominal income growth is forecast by Moody's to be modest as well: From $45,500 in 2018 to $48,400 in 2020. While household incomes are increasing slowly, the inflation rate is forecast to remain flat at 1.9 to 2.3 percent throughout the period, although that depends in the near term on the recovery from the recent hurricane devastation and Federal Reserve policy in the near and medium term.

Industry analysts generally expect the overall housing market to continue its recovery, although the growth of house prices is not expected to be as large as in the last few years given the interest rate environment. As forecast by Moody's, FHFA's purchase-only House Price Index (HPI) is forecast to increase at the annual rates of 3.8, 4.8, and 2.9 percent in 2018, 2019, and 2020, respectively.

The expected increase in mortgage interest rates and house prices will likely impact the ability of low- and very low-income households to purchase homes. Housing affordability, as measured by Moody's forecast of the National Association of Realtors' (NAR) Housing Affordability Index (HAI), is expected to decline from an index value of 156.5 in 2017 to 148.3 in 2020.^[26]

Over the past few years, low interest rates coupled with rising house prices have created an incentive for many homeowners to refinance. The refinance share has increased from 39.9 percent of overall mortgage originations in 2014 to 47.4 percent in 2016. However, assuming that interest rates are going to rise over the next few years, Moody's forecasts that the refinance rate is expected to fall as low as 27 percent during the 2018 to 2020 period.

Additional factors reflecting affordability challenges in the single-family market. While FHFA's models can address and forecast many of the statutory factors that can make affordability for single-family homeownership more challenging for low-income and very low-income households, including increasing interest rates and rising property values, some factors are not captured in the models. FHFA, therefore, considers additional factors when selecting the benchmark level within the model-generated confidence interval for each of the single-family housing goals. Some of these additional factors may affect a subset of the market rather than the market as a whole. These factors include an uneven economic recovery, stagnant wages even where unemployment is decreasing, demographic trends, and the Enterprises' share of the mortgage market. Variability in these factors can also have a substantial impact on the ability of the Enterprises to meet the housing goals. Consequently, as discussed further below, FHFA will carefully monitor these factors and consider the potential impact of market shifts or larger trends on the ability of the Enterprises to achieve the housing goals.

Throughout 2016 and 2017, the economy and the housing market continued to recover from the financial crisis, but the recovery has been uneven across the country.^[27] In some areas, economic growth, job gains, and demand are outpacing housing supply, sparking rapidly rising property values, while other areas of the country have not regained pre-crisis home values and are not projected to do so in the near future.

Income trends. Trends in factors such as area median income (AMI) point to a recovery in most areas in 2017. FHFA uses AMIs published by the U.S. Department of Housing and Urban Development (HUD) to determine affordability for Enterprise single-family and multifamily mortgage acquisitions. AMI is a measure of median family income derived from the Census Bureau's American Community Survey (ACS). Since the 1990s, AMIs have been used widely by HUD, state housing finance agencies, the Federal Deposit Insurance Corporation (FDIC), the U.S. Department of Treasury, and local governments across the nation to determine eligibility for various affordable housing and public assistance programs. The HUD-published AMIs are considered the standard benchmark in the affordable housing industry. HUD changed the methodology for determining AMIs in 2015 because of changes in the Census Bureau's data collection methodology and changes in the reporting schedules of the ACS data.

AMI shifts reflect changes in borrower income levels at the census tract level. In general, a decrease in an area's AMI represents a decline in housing affordability in the area because the households will have relatively less income with which to purchase a home where property values have either remained the same or increased during the same time period.^[28] This can make it more challenging for the Enterprises to meet the housing goals. Conversely, increases in AMIs would make it easier for the Enterprises to meet the housing goals. While there are annual fluctuations in AMI, the trends over a longer period (for instance, over two years or more) indicate that the economy is recovering, albeit in an uneven manner. Over the five-year period from 2012 to 2017, AMIs increased in approximately 80 percent of counties nationwide, indicating a geographically wide-spread recovery. However, some areas experienced AMI decreases in some years. For example, from 2015 to 2016 there were AMI decreases concentrated in South Dakota, Arizona, Pennsylvania, West Virginia, North Carolina, and the coast of South Carolina.

Overall, there are multiple trends in the single-family market that indicate that lower income households that are seeking to buy a home are likely to continue to face difficulty affording homes. While mortgage rates and home prices are projected to rise, the backdrop remains one of slow increases in average household income (as indicated by the AMI), and it is likely that the resources for lower income households seeking to buy a home will remain stretched. The current high house price appreciation, which is projected to continue even at the lower end of the house price spectrum, coupled with a limited supply of lower priced homes (largely due to the lack of construction of lower priced homes) suggests that it will be more challenging for the Enterprises to meet the single-family home purchase goals.^[29]

Additionally, many households have experienced stagnant wages or limited wage growth even though unemployment levels have decreased significantly since the peak of the financial crisis. Data released by the U.S. Census Bureau show that while median household income increased by 3.2 percent from 2015 to 2016, it was only the second year since 2007 that median household income increased.^[30] Further, real median earnings were not statistically different in 2016 compared to 2015. Constrained wages, in addition to rising interest rates and increasing property values, could make it difficult for many low-income and very low-income households to achieve homeownership.

Demographic factors. Demographic changes, such as the housing patterns of millennials or the growth of minority households, also reflect challenges in the affordable homeownership market. The homeownership rate among millennials is lower than other demographic groups, but household formation will likely increase as this group ages. However, many millennials will face multiple challenges, including difficulty finding affordable homes to buy and building enough wealth for a down payment and closing costs, particularly in light of student loan and other debt burdens. Another continuing demographic trend is the growth of minority households, which is projected to be over 70 percent of net household growth through 2025.^[31] Because the median net worth of minority households historically has been low, building the necessary wealth to meet down payment and closing costs will likely also continue to be a challenge for many of these new households.

FHFA is committed to identifying new market conditions and challenges and working with the Enterprises to identify solutions to help meet these challenges. The effectiveness of these solutions, however, cannot be accounted for in a model.

Enterprise market share. Another factor that can affect the Enterprises' ability to support affordable homeownership for low-income and very low-income households is the Enterprises' overall share of the mortgage market, which has fluctuated over time. Graph 1 shows the distribution of conforming mortgage originations by market segment from 2011-2016. The Enterprises' share of the market was at its lowest immediately before and directly after the housing crisis in 2008, at around 45 percent. After that period, the Enterprises' share rose steadily for many years, but began to decline from a peak of 67 percent in 2013, accounting for about 53 percent of the market in 2016. Similarly, the total government share of the mortgage market remained stable for many years after the housing crisis, but expanded to 29 percent in 2015 and 28 percent in 2016, up from 25 percent in 2014.

As discussed in the proposed rule, FHFA's analysis of the mortgage insurance market indicates that a substantial share of the conforming market could switch from private mortgage insurance to FHA insurance if FHA premiums are reduced by similar magnitudes as in the past. FHFA will continue to pay close attention to any changes in the mortgage insurance market.

As discussed above, multiple factors impact the Enterprises' ability to meet their mission and support affordable homeownership through the housing finance market. Nevertheless, FHFA expects the Enterprises to continue efforts in a safe and sound manner to support affordable homeownership under the single-family housing goals categories.

B. Single-Family Benchmark Levels

1. Low-Income Home Purchase Goal

The low-income home purchase goal is based on the percentage of all single-family, owner-occupied home purchase mortgages purchased by an Enterprise that are for low-income families, defined as families with incomes less than or equal to 80 percent of AMI. The final rule sets the annual low-income home purchase housing goal benchmark level for 2018-2020 at 24 percent, the same as the 2015-2017 benchmark level. FHFA has determined that, despite the various challenges to affordability highlighted above, the Enterprises will be able to take steps to maintain or increase their performance on this goal. The 24 percent benchmark level will serve as an appropriate target that will channel Enterprise efforts in this segment.

Table 1—Enterprise Low-Income Home Purchase Goal

	Historical performance (year)				Projected performance (year)
	2013	2014	2015	2016	2017	2018	2019	2020
Benchmark (%)	23	23	24	24	24	24	24	24
Actual Market * (%)	24.0	22.8	23.6	22.9	21.9 +/−2.5	22.7 +/− 4.3	24.4 +/− 5.5	24.3 +/−6.5
Fannie Mae:
Low-Income Purchase	193,712	177,846	188,891	221,628
Total Home Purchase	814,137	757,870	802,432	966,800
% Low-Income	23.8	23.5	23.5	22.9
Freddie Mac:
Low-Income Purchase	93,478	108,948	129,455	153,434
Total Home Purchase	429,158	519,731	579,340	644,988
% Low-Income	21.8	21.0	22.3	23.8
* Market forecast shown for 2017-2020.

Recent performance and forecasts. As shown in Table 1, performance at both Enterprises fell short of the benchmark level for the low-income home purchase goal in 2015 and 2016, and both Enterprises missed both the benchmark level and the market level for the low-income home purchase goal in 2015. Both Enterprises met this goal in 2016 by exceeding the market level. Recent past performance of the Enterprises indicates that it has been difficult for the Enterprises to consistently exceed the benchmark level and lead this market segment in making credit available.

From 2013 to 2014, the low-income home purchase market decreased from 24.0 percent to 22.8 percent. In 2015, the market rebounded to 23.6 percent and then decreased to 22.9 percent in 2016. FHFA's current model forecasts that the market for this goal will continue to decrease to 21.9 percent in 2017 before increasing to 22.7 percent in 2018, 24.4 percent in 2019 and 24.3 in 2020. The actual market for each of these years will be calculated by FHFA using HMDA data for the year when it becomes available.

Although the Enterprises have been challenged in meeting the single-family housing goal levels in recent years, each Enterprise has increased the number of single-family home purchase loans it has purchased that were made to low-income households. Fannie Mae's eligible single-family loan purchases increased from 193,712 loans in 2013 to 221,628 in 2016. Freddie Mac's eligible single-family loan purchases increased from 93,478 in 2013 to 153,434 in 2016.

Proposed rule and comments. In the proposed rule, FHFA proposed maintaining the benchmark level for 2018-2020 at the 2015-2017 level of 24 percent. At that time, using data through December 2016, the average market level forecast for 2018-2020 was 24.2 percent. Since the publication of the proposed rule, FHFA updated the model using data through November 2017 and additional 2016 data from HMDA and Moody's. The updated FHFA model forecasts that the market for this goal will be slightly lower, with the average forecast at 23.8 percent.

Five comment letters expressed support for the proposed benchmark levels for the single-family goals, including the low-income home purchase goal. Commenters commended FHFA for appropriately challenging the Enterprises while taking into account safety and soundness and the realities of the mortgage market. Four comments endorsed a higher benchmark level for the low-income home purchase goal. These commenters recommended setting the low-income goal benchmark at levels between 27 and 30 percent, arguing that more aggressive targets will encourage focus on this income segment, which will benefit consumers and improve access to credit. Only one commenter (Fannie Mae) asserted that the proposed benchmark level for the low-income home purchase goal was too high, and should be lowered to 21 percent. The letter cited ongoing market challenges that make it difficult to meet the benchmark level, including the lack of supply of moderately-priced homes and limited job growth.

FHFA determination. Consistent with the proposed rule, the final rule sets the benchmark level for the low-income home purchase housing goal at 24 percent. This is slightly above the average market forecast for the three years, to encourage the Enterprises to continue to find ways to support lower income borrowers while not compromising safe and sound lending standards. Even though the benchmark is slightly higher than the average market forecast for this goal, due to the two-part nature of the goals, the level that will be used to judge the Enterprises' year-end performance will be the lower of the market level or the benchmark. Therefore, the 24 percent benchmark level is appropriate, reasonable, and supported by the current market forecast. FHFA recognizes that there may be challenges to meeting this goal, including uneven growth in AMI and the relative affordability of private mortgage insurance, which may be beyond the control of the Enterprises and impact their ability to achieve these goals. FHFA will continue to monitor the performance of the Enterprises on this goal and, if FHFA determines in later years that the benchmark level for the low-income home purchase housing goal is no longer feasible for the Enterprises to achieve in light of market conditions or for any other reason, FHFA may take appropriate steps to adjust the benchmark level.

2. Very Low-Income Home Purchase Goal

The very low-income home purchase goal is based on the percentage of all single-family, owner-occupied home purchase mortgages purchased by an Enterprise that are for very low-income families, defined as families with incomes less than or equal to 50 percent of the area median income. The final rule sets the annual very low-income home purchase housing goal benchmark level for 2018 through 2020 at 6 percent. FHFA has determined that, despite the various challenges to affordability highlighted above, the Enterprises will be able to take steps to maintain or increase their performance on this goal. The 6 percent benchmark level will serve as an appropriate target that will channel Enterprise efforts in this segment.

Table 2—Very Low-Income Home Purchase Goal

	Historical performance (year)				Projected performance (year)
	2013	2014	2015	2016	2017	2018	2019	2020
Benchmark (%)	7	7	6	6	6	6	6	6
Actual Market * (%)	6.3	5.7	5.8	5.4	5.1 +/−0.9	5.3 +/−1.5	5.9 +/−1.9	5.9 +/−2.2
Fannie Mae:
Very Low-Income Purchase	48,810	42,872	45,022	49,932
Total Home Purchase	814,137	757,870	802,432	966,800
% Very Low-Income	6.0	5.7	5.6	5.2
Freddie Mac:
Very Low-Income Purchase	23,705	25,232	31,146	36,837
Total Home Purchase	429,158	519,731	579,340	644,988
% Very Low-Income	5.5	4.9	5.4	5.7
* Market forecast shown for 2017-2020.

Recent performance and forecasts. As shown in Table 2, the market for very low-income home purchase loans has been declining since 2013, as reflected in HMDA data, although there was a slight uptick in 2015. FHFA has gradually lowered the benchmark level for this goal from 8 percent in 2010 to 6 percent in 2015. Despite this reduction, the performance of both Enterprises has continued to fall below the benchmark level in each year since 2013. In 2016, Freddie Mac achieved the very low-income goal by meeting the market level, but Fannie Mae failed to meet the goal.

FHFA's models forecast this segment to remain between 5.1 percent and 5.9 percent for 2017-2020. For the 2018-2020 goal period, FHFA's forecast indicates an increase from 5.1 percent in 2017 to 5.3 percent in 2018 and to 5.9 percent in 2019 and 2020. As noted earlier, the confidence intervals widen as the forecast period lengthens.

Proposed rule and comments. In the proposed rule, FHFA proposed maintaining the benchmark level for 2018-2020 at the 2015-2017 level of 6 percent. At that time, using data through December 2016, the average market level forecast for 2018-2020 was 6.4 percent. FHFA adjusted the model using data through November 2017 and additional 2016 data from HMDA and Moody's, and the current model forecasts that the average market level for 2018-2020 for this goal will be lower, at 5.7 percent.

As highlighted in the low-income goal discussion above, there were five comment letters that expressed support for the proposed benchmark levels for the single-family goals, including the very low-income home purchase goal at 6 percent. Commenters commended FHFA for appropriately challenging the Enterprises while taking into account safety and soundness and the realities of the mortgage market. Four comments endorsed a higher benchmark level for the very low-income home purchase goal. Commenters recommended setting the very low-income goal benchmark at levels between 7 and 10 percent. These commenters argued that more aggressive targets will encourage the Enterprises to focus on this income segment, which will benefit consumers and improve access to credit. Only one commenter (Fannie Mae) asserted that the proposed benchmark level for the very low-income home purchase goal was too high, and should be lowered to 5 percent. Fannie Mae cited ongoing market challenges that make it difficult to meet the benchmark level, including lack of supply of moderately-priced homes and limited job growth.

FHFA determination. Consistent with the proposed rule, the final rule sets the very low-income home purchase housing goal benchmark level at 6 percent, slightly higher than the current 5.7 percent forecast average. FHFA considered lowering the benchmark level for the very low-income home purchase goal to 5.5 percent but decided to keep the benchmark level at 6 percent for multiple reasons. This level is near but slightly higher than the market forecast average. This level should serve as a “stretch goal” to encourage the Enterprises to continue their efforts to promote safe and sustainable lending to very low-income families. As noted in the low-income home purchase goal discussion above, there are significant challenges to housing affordability that may be beyond the control of the Enterprises that could make the benchmark level a challenge for the Enterprises. However, given the two-part nature of the goals, the level that will be likely to constrain the Enterprises will be the lower of the market level or the benchmark. Thus, FHFA is persuaded that setting the benchmark level at 6 percent is appropriate, reasonable, and supported by the current market forecast.

FHFA will continue to monitor the Enterprises' performance on this goal and, if FHFA determines in later years that the benchmark level for the very low-income areas home purchase housing goal is no longer feasible for the Enterprises to achieve in light of market conditions or for any other reason, FHFA may take appropriate steps to adjust the benchmark level.

3. Low-Income Areas Home Purchase Subgoal

The low-income areas home purchase subgoal is based on the percentage of all single-family, owner-occupied home purchase mortgages purchased by an Enterprise that are either: (1) For families in low-income areas, defined to include census tracts with median income less than or equal to 80 percent of AMI; or (2) for families with incomes less than or equal to AMI who reside in minority census tracts (defined as census tracts with a minority population of at least 30 percent and a tract median income of less than 100 percent of AMI). Mortgage loans may qualify under either or both conditions. As discussed in the proposed rule, mortgages satisfying condition (1) above, or borrowers in low-income areas, are typically almost double the share of mortgages satisfying condition (2), or moderate-income borrowers in minority census tracts. The share of mortgages that satisfy both conditions is generally small (for example, 4.6 percent of low-income areas subgoal mortgages in 2015).

The final rule sets the annual low-income areas home purchase subgoal benchmark level for 2018 through 2020 at 14 percent, which is lower than the 15 percent in the proposed rule, based on comments received by FHFA. FHFA has determined that this benchmark level will serve as an appropriate target for the Enterprises. FHFA will continue to evaluate the impact and efficacy of this subgoal.

Table 3—Low-Income Areas Home Purchase Subgoal

	Historical performance (year)				Projected performance (year)
	2013	2014	2015	2016	2017	2018	2019	2020
Benchmark (%)	11	11	14	14	14	14	14	14
Actual Market * (%)	14.2	15.2	15.2	15.9	16.5	16.6	16.8	16.4
					+/−1.2	+/−2.0	+/−2.5	+/−3.0
Fannie Mae Performance:
Low-Income Area Home Purchase Mortgages	86,430	91,691	99,723	125,956
High-Minority Area Home Purchase Mortgages	27,425	25,650	25,349	30,535
Subgoal-Qualifying Total Home Purchase Mortgages	113,855	117,341	125,072	156,491
Total Home Purchase Mortgages	814,137	757,870	802,432	966,800
Low-Income Area % of Home Purchase Mortgages	14.0	15.5	15.6	16.2
Freddie Mac Performance:
Low-Income Area Home Purchase Mortgages	40,444	55,987	67,172	80,805
High-Minority Area Home Purchase Mortgages	12,177	14,808	16,601	19,788
Subgoal-Qualifying Total Home Purchase Mortgages	52,621	70,795	83,773	100,593
Total Home Purchase Mortgages	429,158	519,731	579,340	644,988
Low-Income Area % of Home Purchase Mortgages	12.3	13.6	14.5	15.6
* Market forecast shown for 2017-2020.

Recent performance and forecasts. As shown in Table 3, both Enterprises have met this subgoal every year since 2013, regularly exceeding both the market and the benchmark levels. Fannie Mae's performance exceeded both the market and the benchmark level in 2014 through 2016, although its performance was below the market level in 2013. From 2013 through 2016, Freddie Mac's performance exceeded the benchmark level but was below the market level.

The forecast for this subgoal was obtained by generating separate forecasts for the two sub-populations (the low-income areas component and the high-minority component). FHFA has tested alternate model specifications for this subgoal and determined that aligning the overlapping portion with the low-income areas component yields forecast estimates that are more precise (in terms of a narrower confidence interval).^[32] FHFA's forecast indicates that the market will increase slightly in the coming years, reaching a maximum level of 16.8 percent in 2019.

Proposed rule and comments. In the proposed rule, FHFA proposed raising the benchmark level to 15 percent for 2018-2020 from the 2015-2017 level of 14 percent. FHFA has adjusted the model using data through November 2017 and additional 2016 data from HMDA and Moody's, and the current model forecasts that the average market for 2018-2020 for this goal will be approximately 16.6 percent, slightly higher than the 15.9 percent average from the proposed rule forecast. As noted in the proposed rule, FHFA's analysis found that the mortgage market (as measured by HMDA data) in both low-income areas and the high-minority areas had increasing shares of borrowers with incomes at or above 100 percent of AMI.^[33] This trend lies at the heart of the public policy dilemma that FHFA is addressing: While the presence of higher income borrowers in lower income and high minority areas may be a sign of economic diversity in those areas and may be related to the possibility of improved economic indicators for the community, there is nevertheless some concern that such a trend could displace existing residents in those areas, especially lower income households. FHFA is aware that this particular subgoal may encourage the Enterprises to focus on purchasing loans for higher income households in low-income and high-minority areas while at the same time fueling concerns about the impact of rising housing costs on existing or displaced households in lower-income or higher-minority areas.

FHFA sought comment on this issue in its proposed rule and received two comment letters that addressed this issue. Both commenters agreed with FHFA's concerns. One encouraged FHFA to continue to carefully monitor the policy objectives and efficacy of this goal. The other commenter opposed raising the benchmark levels for this goal. After considering these and other comments, FHFA is setting the very low-income areas home purchase housing subgoal benchmark level at 14 percent, which is lower than the current 16.6 percent average market forecast.

FHFA determination. The final rule sets the benchmark level for the low-income areas home purchase subgoal at 14 percent. This level reflects a balance between the market and recent performance levels of the Enterprises while FHFA continues to evaluate whether the goal meets all policy objectives. FHFA will continue to monitor the Enterprises' performance on this subgoal and, if FHFA determines in later years that the benchmark level for the low-income areas home purchase subgoal is no longer feasible for the Enterprises to achieve in light of market conditions or for other reasons, FHFA may take appropriate steps to adjust the benchmark level.

4. Low-Income Areas Home Purchase Goal

The low-income areas home purchase goal covers the same categories as the low-income areas home purchase subgoal, but it also includes moderate income families in designated disaster areas. As a result, the low-income areas home purchase goal is based on the percentage of all single-family, owner-occupied home purchase mortgages purchased by an Enterprise that are: (1) For families in low-income areas, defined to include census tracts with median income less than or equal to 80 percent of AMI; (2) for families with incomes less than or equal to AMI who reside in minority census tracts (defined as census tracts with a minority population of at least 30 percent and a tract median income of less than 100 percent of AMI); or (3) for families with incomes less than or equal to 100 percent of AMI who reside in designated disaster areas.

The low-income areas goal benchmark level is established by a two-step process. The first step is setting the benchmark level for the low-income areas subgoal, as established by this final rule. The second step is establishing an additional increment for mortgages to families with incomes less than or equal to AMI located in federally-declared disaster areas.^[34] Each year, FHFA sets the disaster area increment separately from this rule and notifies the Enterprises by letter of the benchmark level for the low-income areas home purchase goal that year. The final rule sets the annual low-income areas home purchase goal benchmark level for 2018 through 2020 at the subgoal benchmark level of 14 percent plus a disaster areas increment that FHFA will set separately each year.

Table 4—Low-Income Areas Home Purchase Goal

	Historical performance (year)
	2010	2011	2012	2013	2014	2015	2016
Benchmark (%)	24	24	20	21	18	19	17
Actual Market * (%)	24.0	22.0	23.2	22.1	22.1	19.8	19.7
Fannie Mae Performance:
Subgoal-Qualifying Home Purchase Mortgages	59,281	54,285	83,202	113,855	117,341	125,072	156,441
Disaster Areas Home Purchase Mortgages	56,076	50,209	58,085	62,314	54,548	38,885	38,545
Goal-Qualifying Total Home Purchase Mortgages	115,357	104,494	141,287	176,169	171,889	163,957	194,986
Total Home Purchase Mortgages	479,200	467,066	633,627	814,137	757,870	802,432	964,847
Goal Performance (%)	24.1	22.4	22.3	21.6	22.7	20.4	20.2
Freddie Mac Performance:
Subgoal-Qualifying Home Purchase Mortgages	32,089	23,902	32,750	52,621	70,795	83,773	100,608
Disaster Areas Home Purchase Mortgages	38,898	26,232	26,486	33,123	33,923	26,411	27,709
Goal-Qualifying Total Home Purchase Mortgages	70,987	50,134	59,236	85,744	104,718	110,184	128,317
Total Home Purchase Mortgages	307,555	260,796	288,007	429,158	519,731	579,340	644,991
Goal Performance (%)	23.1	19.2	20.6	20.0	20.1	19.0	19.9

5. Low-Income Refinancing Goal

The low-income refinancing goal is based on the percentage of all single-family, owner-occupied refinance mortgages purchased by an Enterprise that are for low-income families, defined as families with incomes less than or equal to 80 percent of AMI. The final rule sets the annual low-income refinancing housing goal benchmark level for 2018 through 2020 at 21 percent. FHFA has determined that this benchmark level will serve as an appropriate target for the Enterprises. While this benchmark level is unchanged from the current 2015 to 2017 benchmark level, it will nevertheless be challenging for the Enterprises given the current level of interest rates (which are at historic low levels) and the likelihood of interest rate hikes. Because of the significant impact interest rate changes have on this market, Enterprise and market performance on this goal are particularly susceptible to fluctuation. Moderation in the setting of this goal is also supported by the fact that many borrowers have already refinanced during the recent extended period of historically low interest rates.

Table 5—Low-Income Refinancing Goal

	Historical performance (year)				Projected performance (year)
	2013	2014	2015	2016	2017	2018	2019	2020
Benchmark (%)	20	20	21	21	21	21	21	21
Actual Market * (%)	24.3	25.0	22.5	19.8	23.4	23.4	20.6	18.0
					+/−3.0	+/−5.1	+/−6.5	+/−7.7
Fannie Mae Performance:
Low-Income Refinance Mortgages	519,753	215,826	227,817	246,571
Total Refinance Mortgages	2,170,063	831,218	1,038,663	1,270,542
Low-Income % of Refinance Mortgages	24.0	26.0	21.9	19.4
Low-Income HAMP Modification Mortgages	11,858	6,503	3,563	2,127
Total HAMP Modification Mortgages	16,478	9,288	6,595	3,800
Low-Income % of HAMP Modification Mortgages	72.0	70.0	54.0	56.0
Low-Income Refinance & HAMP Modification Mortgages	531,611	222,329	231,380	248,698
Total Refinance & HAMP Modification Mortgages	2,186,541	840,506	1,045,258	1,274,342
Low-Income % of Refinance & HAMP Modification Mortgages	24.3	26.5	22.1	19.5
Freddie Mac Performance:
Low-Income Refinance Mortgages	306,205	131,921	179,530	172,987
Total Refinance Mortgages	1,309,435	514,936	795,936	828,553
Low-Income % of Refinance Mortgages	23.4	25.6	22.6	20.9
Low-Income HAMP Modification Mortgages	14,757	6,795	3,064	1,721
Total HAMP Modification Mortgages	21,599	10,335	4,433	2,335
Low-Income % of HAMP Modification Mortgages	68.3	65.7	69.1	73.7
Low-Income Refinance & HAMP Modification Mortgages	320,962	138,716	182,594	174,708
Total Refinance & HAMP Modification Mortgages	1,331,034	525,271	800,369	830,888
Low-Income % of Refinance & HAMP Modification Mortgages	24.1	26.4	22.8	21.0
* Market forecast shown for 2017-2020.

Recent performance and forecasts. As shown in Table 5, the performance of the Enterprises on the low-income refinancing housing goal has historically been very close to the actual market levels. In 2014, when the market level was at its highest point, both Enterprises met the goal by exceeding the market level. In 2015, Freddie Mac surpassed the market and the benchmark levels, and Fannie Mae exceeded the benchmark level. In 2016, Freddie Mac met the benchmark level and exceeded the market level with its performance at 21.0 percent, but Fannie Mae missed the benchmark and the market levels, with its performance reaching only 19.5 percent.

The low-income share of the refinance market as measured by HMDA data has changed dramatically in recent years, increasing from 20.2 percent in 2010 to a peak of 25 percent in 2014, and dropping from 22.5 percent in 2015 to 19.8 percent in 2016. FHFA's model predicts that this share will increase to 23.4 percent in 2017 and 2018, and then decline to 20.6 percent in 2019 and 18.0 percent in 2020. The confidence intervals for this model are fairly wide because of the considerable uncertainty around interest rates. Recent macroeconomic forecasts have predicted interest rate hikes that have yet to materialize in any substantive way.

Since 2010, the low-income refinancing housing goal has included modifications under the Home Affordable Modification Program (HAMP).^[35] HAMP modifications, however, are not included in the data used to calculate the market levels. Including HAMP modifications in the Enterprise performance numbers increases the measured performance of the Enterprises on the low-income refinancing housing goal because lower income borrowers make up a greater proportion of the borrowers receiving HAMP modifications than the low-income share of the overall refinancing mortgage market. However, HAMP modifications have been declining over time, and the program stopped taking applications at the end of 2016.^[36] The expiration of the HAMP program may make it slightly more difficult for the Enterprises to meet the low-income refinancing goal.

Proposed rule and comments. In the proposed rule, FHFA proposed maintaining the benchmark level for 2018-2020 at the 2015-2017 level of 21 percent. FHFA received one comment stating generally that all single-family goals should be increased. The comment noted the importance of the low-income refinance goal in preserving homeownership.

FHFA determination. Consistent with the proposed rule, the final rule sets the low-income refinance benchmark level at 21 percent, slightly higher than the current 20.7 percent average market forecast. FHFA is setting this benchmark at a relatively low level compared to the 23.4 percent market forecast for 2018, based in part on the forecast decreasing significantly over the three year period covered by the forecast. FHFA is also mindful of the higher level of uncertainty about the forecasts for this goal given the unpredictability of future interest rate changes. The 21 percent benchmark level reflects a balance between the market and recent performance levels of the Enterprises. FHFA will continue to monitor the performance of the Enterprises on this goal and, if FHFA determines in later years that the benchmark level for the low-income refinancing housing goal is no longer feasible for the Enterprises to achieve in light of market conditions or for other reasons, FHFA may take appropriate steps to adjust the benchmark level.

V. Multifamily Housing Goals

This final rule also establishes the multifamily housing goals for 2018-2020. FHFA considered the required statutory factors described below in setting the benchmark levels for the multifamily housing goals. Two divergent trends underlie FHFA's analysis: a strong multifamily mortgage market for units that are affordable to higher-income households but a continued gap in the supply of units affordable to lower-income households. There are some forecasts that support a softening of the first trend but all forecasts uniformly expect the second trend to continue during the goal period. FHFA expects and encourages the Enterprises to fully support affordable multifamily housing, in part by fulfilling the multifamily housing goals in a safe and sound manner.

A. Factors Considered in Setting the Multifamily Housing Goal Levels

In setting the benchmark levels for the multifamily housing goals, FHFA considered the statutory factors outlined in section 1333(a)(4) of the Safety and Soundness Act. These factors include:

1. National multifamily mortgage credit needs and the ability of the Enterprises to provide additional liquidity and stability for the multifamily mortgage market;

2. The performance and effort of the Enterprises in making mortgage credit available for multifamily housing in previous years;

3. The size of the multifamily mortgage market for housing affordable to low-income and very low-income families, including the size of the multifamily markets for housing of a smaller or limited size;

4. The ability of the Enterprises to lead the market in making multifamily mortgage credit available, especially for multifamily housing affordable to low-income and very low-income families;

5. The availability of public subsidies; and

6. The need to maintain the sound financial condition of the Enterprises.^[37]

Unlike the single-family housing goals, performance on the multifamily housing goals is measured solely against a benchmark level, without any retrospective market measure. The absence of a retrospective market measure for the multifamily housing goals results, in part, from the lack of comprehensive data about the multifamily mortgage market. Unlike the single-family market, for which HMDA provides a reasonably comprehensive dataset about single-family mortgage originations each year, the multifamily market (including the affordable multifamily market segment) has no comparable source of data. Consequently, it can be difficult to correlate different datasets on the multifamily market because they usually rely on different reporting formats. For example, some data are available by dollar volume of mortgages while other data are available by unit production.^[38]

Another difference between the single-family and multifamily goals is that there are separate single-family housing goals for home purchase and refinance mortgages, while the multifamily goals include all Enterprise multifamily mortgage purchases, regardless of the purpose of the loan. In addition, unlike the single-family housing goals, the multifamily housing goals are measured based on the total volume of affordable multifamily mortgage purchases rather than on a percentage of multifamily mortgage purchases. The use of total volume, which FHFA measures by the number of eligible units, rather than percentages of each Enterprises' overall multifamily purchases, requires that FHFA take into account the expected size of the overall multifamily mortgage market and the affordable share of the market, as well as the expected volume of the Enterprises' overall multifamily purchases and the affordable share of those purchases.

The lack of comprehensive data for the multifamily mortgage market is even more acute with respect to the segments of the market that are targeted to low-income families, defined as families with incomes at or below 80 percent of AMI, and very low-income families, defined as families with incomes at or below 50 percent of AMI. As required by the Safety and Soundness Act, FHFA determines affordability of multifamily units based on maximum rent levels not exceeding 30 percent of the area median income standard for low- and very low-income families.^[39] This affordability definition is sometimes referred to as the “Brooke Amendment,” and states that to be considered a low-income multifamily unit (i.e., affordable at the 80 percent AMI level), the rent levels must be less than or equal to 30 percent of the maximum income at 80 percent of the AMI, with appropriate adjustments for unit size as measured by the number of bedrooms.^[40] Similarly, to be considered a very low-income multifamily unit (i.e., affordable at the 50 percent AMI level), the rent levels must be less than or equal to 30 percent of the maximum income at 50 percent of the AMI, with appropriate adjustments for unit size as measured by the number of bedrooms. While much of the analysis that follows discusses trends in the overall multifamily mortgage market, FHFA recognizes that these trends may not apply to the same extent to all segments of the multifamily market. Notwithstanding these challenges, FHFA has considered each of the required statutory factors (a number of which are related) as discussed below.

Multifamily mortgage market. FHFA's consideration of the multifamily mortgage market addressed the size of and competition within the multifamily mortgage market, as well as the subset of the multifamily market affordable to low-income and very low-income families. In 2016, the multifamily mortgage origination market experienced continued growth: Year-over-year origination volume grew 8 percent from about $250 billion to $269 billion, fueled largely by a recovery in multifamily construction. Forecasts from various industry experts indicate that overall multifamily mortgage market volumes and mortgage originations are expected to increase only modestly in 2017, both for refinancing activity and for financing new multifamily units, and will likely decrease modestly in 2018. FHFA's internal forecasts are consistent with this view.

The total number of renter households grew from 35 million in 2005 to 44 million in 2015, an increase of about one quarter.^[41] According to the National Multifamily Housing Council's tabulation of 2016 American Community Survey (ACS) data, about 43 percent of renter households (18.9 million households or 38.8 million residents) lived in structures with five or more rental units.^[42] This growth led to an increase in demand for rental units that has only partially been met by expansions in supply. Vacancy rates hit a 30-year low in 2016, and are especially low in lower-priced segments of the market, while climbing in the higher-priced segments of the market.^[43] Rents also continued to rise nationally and outpaced inflation in 2016.^[44]

Affordability in the multifamily market. There are several factors that make it difficult to accurately forecast the affordable share of the multifamily mortgage market. First, the portion of the overall multifamily mortgage market that provides housing units affordable to low-income and very low-income families varies from year to year. Second, competition between purchasers of mortgages within the multifamily market overall may differ from the competition within the affordable multifamily market segment. Finally, the volume for the affordable multifamily market segment depends on the availability of affordable housing subsidies. Thus in some ways, the multifamily market is segmented into the affordable and non-affordable segments with loose linkages between the two segments. Despite strength in the non-affordable multifamily market in recent years, there has been little increase in the affordable segment. Using the standard measure of affordability, where rent and utilities do not exceed 30 percent of AMI (required by the Brooke Amendment), families living in rental units have faced decreasing affordability in recent years.

The Joint Center for Housing Studies (JCHS) has released two reports noting concerning trends in the supply of affordable multifamily units. The overall inventory of affordable multifamily units is low, and rent on most newly built units are out of reach for lower-income families. As the JCHS's 2017 Report on America's Rental Housing notes:

“Soaring demand sparked a sharp expansion of the rental stock over the past decade. Initially, most of the additions to supply came from conversions of formerly owner-occupied units, particularly single family homes, which provided housing for the increasing number of families with children in the rental market. Between 2006 and 2016, the number of single-family homes available for rent increased by nearly 4 million, lifting the total to 18.2 million. While single-family homes have always accounted for a large share of rental housing, they now make up 39 percent of the stock. More recently, though, growth in the single-family supply has slowed. The American Community Survey shows that the number of single-family rentals (including detached, attached, and mobile homes) increased by only 74,000 units between 2015 and 2016, substantially below the 400,000 annual increase averaged in 2005-2015. With this slowdown in single-family conversions and a boom in multifamily construction, new multifamily units have come to account for a growing share of new rentals.” ^[45]

The Report on America's Rental Housing goes on to note that much of this new multifamily construction is aimed at higher income households and located primarily in high-rise buildings in downtown neighborhoods while the supply of moderate and lower cost units has only grown modestly. The Report on America's Rental Housing notes that the share of new units renting for less than $850 a month has actually declined from two-fifths to one-fifth between 2001 and 2016.

The JCHS's 2017 State of the Nation's Housing Report indicates that the majority of growth in rental housing stock in recent years was primarily the result of new multifamily construction. Moreover, most of this new construction consists of apartments with fewer bedrooms and has been concentrated in urban areas with higher median rents. According to the State of the Nation's Housing Report, there have been significant declines in the supply of low-cost rental housing. Using ACS data from 2005 and 2015, the report notes that gains in the supply of high-end units and losses of low- and modest-priced units over the past decade have shifted the entire rental stock toward the high end. The State of the Nation's Housing Report notes, “bolstered by new, high-end construction and rising rents for existing apartments, the number of units renting for $2,000 or more per month increased 97 percent in real terms between 2005 and 2015.” At the same time, “the number of units renting for below $800 fell by 2 percent.” ^[46]

The State of the Nation's Housing Report also notes the significant prevalence of cost-burdened renters. In 2015, nearly one-third of all tenants paid more than 30 percent of their household income for rental housing, especially in high-cost urban markets where most renters reside and where a majority of the multifamily loans purchased by Fannie Mae and Freddie Mac have been located. Among lower-income households, cost burdens are especially severe.^[47] The same report notes that while housing affordability is a growing concern for communities nationwide, the cost-burdened shares in 11 of the country's largest metropolitan areas were above 40 percent. In addition, a recent study showed that the median incomes of renter households have experienced slight declines in some large metropolitan areas in recent years, leading to increased cost burdens for these households.^[48]

One source of growth in the stock of lower-rent apartments is “filtering,” a process by which existing units become more affordable as they age. However, in recent years, this downward filtering of rental units has occurred at a slow pace in most markets. Coupled with the permanent loss of affordable units, as these units fall into disrepair or units are demolished to create new higher-rent or higher-valued ownership units, this trend has severely limited the supply of lower rent units. As a result, there is an acute shortfall of affordable units for extremely low-income renters (earning up to 30 percent of AMI) and very low-income renters (earning up to 50 percent of AMI). This supply gap is especially wide in certain metropolitan areas in the southern and western United States.^[49]

The combination of the supply gap in affordable units, which has resulted in significant increases in rental rates, and the prevalence of cost-burdened renters resulting from largely flat real incomes has led to an erosion of affordability, with fewer units qualifying for the housing goals.^[50] This challenge of affordability is also reflected in the falling share of low-income multifamily units financed by loans purchased by the Enterprises. While 77 percent of the multifamily units financed by Fannie Mae in 2011 were low-income, that ratio dropped steadily in the intervening years to 64 percent in 2016. At Freddie Mac, the low-income share also peaked in 2011 and 2012 at 79 percent, and decreased gradually to 68 percent in 2016. For the very low-income goal, the share at Fannie Mae peaked in 2012 at 22 percent before falling to 12 percent in 2016, and at Freddie Mac the share peaked at 17 percent in 2013 before falling to 12 percent in 2016.

Small multifamily properties with 5 to 50 units are also an important source of affordable rental housing and represent approximately one-third of the affordable rental market. Because they have different operating and ownership characteristics than larger properties, small multifamily properties often have different financing needs. For example, small multifamily properties are more likely to be owned by an individual or small investor and less likely to be managed by a third party property management firm.^[51] Likewise, the affordability of small multifamily units means they generate less revenue per unit than larger properties. These factors can make financing more difficult to obtain for small multifamily property owners. While the volume of Enterprise-supported loans on small multifamily properties has been inconsistent in recent years, each Enterprise continues to refine its approach to serving this market.

Availability of public subsidies. Multifamily housing subsidy assistance is primarily available in two forms—demand-side subsidies that either assist low-income tenants directly (e.g., Section 8 vouchers) or provide project-based rental assistance (Section 8 contracts), and supply-side subsidies that support the creation and preservation of affordable housing (e.g., public housing and Low-Income Housing Tax Credit (LIHTC)). The availability of public subsidies impacts the overall affordable multifamily housing market, and changes to longstanding housing subsidy programs could significantly impact the ability of the Enterprises to meet the goals.

Financing for affordable multifamily buildings—particularly those affordable to very low-income families—often uses an array of state and federal supply-side housing subsidies, such as LIHTC, tax-exempt bonds, project-based rental assistance, or soft subordinate financing.^[52] In recent years, competition for affordable housing subsidies has been intense and investor interest in tax credit equity projects of all types and in all markets has been strong, especially in markets in which bank investors are seeking to meet Community Reinvestment Act (CRA) goals. By contrast, in recent months, the subsidy provided by the LIHTC program has been volatile and uncertain due to potential impacts of recent changes in tax laws. Projections carried out by housing industry groups suggest that the level of LIHTC production will decrease because of the reduction in corporate tax rates.^[53]

Subject to the continuing availability of these subsidies, there should continue to be opportunities in the multifamily market to provide permanent financing for properties with LIHTC during the 2018-2020 period. There should also be opportunities for market participants, including the Enterprises, to purchase mortgages that finance the preservation of existing affordable housing units, especially for restructurings of older properties that reach the end of their initial 15-year LIHTC compliance periods and for refinancing properties with expiring Section 8 rental assistance contracts.

In recent years, demand-side public subsidies and the availability of public housing have not kept pace with the growing number of low-income and very low-income households in need of federal housing assistance. As a result, the number of renter households with “worst case needs” has grown to 8.3 million, an increase of more than one-third since 2005.^[54]

Role of the Enterprises. In setting the multifamily housing goals, FHFA considered the ability of the Enterprises to lead the market in making multifamily mortgage credit available. The share of the overall multifamily market purchased by the Enterprises increased in the years immediately following the financial crisis but has declined more recently in response to growing private sector participation. The Enterprise share (in dollar volume terms) of the multifamily origination market was approximately 70 percent of the market in 2008 and 2009 compared to 38 percent in 2015 and 39 percent in 2016.^{[55, 56]} The total share is expected to remain at around these lower levels in 2017 and 2018, particularly in light of the Scorecard cap imposed by FHFA in its role as conservator, which is discussed below.

Despite the Enterprises' reduced market share in the overall multifamily market and due to the segmented nature of the multifamily market noted earlier, FHFA expects the Enterprises to continue to demonstrate leadership in multifamily affordable housing by providing liquidity and supporting housing for tenants at different income levels in various geographic markets and in various market segments.

Conservatorship limits on multifamily mortgage purchases (Conservatorship Scorecard cap). As conservator of the Enterprises, FHFA has established a yearly cap in the Conservatorship Scorecard that limits the amount of conventional, market-rate multifamily loans that each Enterprise can purchase. The multifamily cap is intended to further FHFA's conservatorship goals of maintaining the presence of the Enterprises as a backstop for the multifamily finance market, while not impeding the participation of private capital. This target for the Enterprise share of the multifamily origination market reflects what FHFA considers an appropriate market share for the Enterprises during normal market conditions. The cap prevents the Enterprises from crowding out other capital sources and restrains the rapid growth of the Enterprises' multifamily businesses that started in 2011.^[57] FHFA has designed the cap so that most loans eligible for housing goals credit, as well as certain other categories of transactions for underserved market segments, are excluded from the cap. As a result, increases and decreases in the cap itself should not impact the ability of the Enterprises to meet these goals.

In 2015, FHFA established a cap of $30 billion on new conventional multifamily loan purchases for each Enterprise in response to increased participation in the market from private sector capital. In 2016, the cap increased from $30 billion to $36.5 billion in response to growth of the overall multifamily origination market throughout the year. This increase maintained the Enterprises' current market share at about 40 percent. In 2017, FHFA kept the cap at $36.5 billion. In 2018, the cap has been reduced to $35 billion.

FHFA reviews the market size estimates quarterly, using current market data provided by the MBA, the National Multifamily Housing Council, and Fannie Mae and Freddie Mac. FHFA also produces an internal forecast. If FHFA determines during the year that the actual market size is greater than was projected, FHFA will consider an increase to the capped (conventional market-rate) category of the Conservatorship Scorecard for each Enterprise. In light of the need for market participants to be able to plan sales of mortgages during long origination processes, if FHFA determines that the actual market size is smaller than projected, there will be no reduction to the capped volume for the current year from the amount initially established under the Conservatorship Scorecard.

As noted earlier, in order to encourage affordable lending activities, FHFA excludes many types of loans in underserved markets from the Conservatorship Scorecard cap on conventional multifamily loans. The Conservatorship Scorecard has no volume targets in the market segments excluded from the cap. There is significant overlap between the types of multifamily mortgages that are excluded from the Conservatorship Scorecard cap and the multifamily mortgages that contribute to the performance of the Enterprises under the affordable housing goals. The 2018 Conservatorship Scorecard excludes either the entirety of the loan amount or a pro rata share of the loan for the following categories: (1) Targeted affordable housing (such as loans on properties subsidized by LIHTC, properties developed under state or local inclusionary zoning, real estate tax abatement, loan or similar programs, and properties covered by a Section 8 Housing Assistance Payment contract limiting tenant incomes to 80 percent of AMI or below); (2) small multifamily properties; (3) blanket loans on manufactured housing communities; (4) blanket loans on senior housing and assisted living communities; (5) loans in rural areas; (6) loans to finance energy or water efficiency improvements; and (7) market rate affordable units in standard (60 percent of AMI), high cost (80 percent of AMI), very high cost (100 percent of AMI), and extremely high cost (120 percent of AMI) markets. By excluding these categories from the cap, the Conservatorship Scorecard continues to encourage the Enterprises to support affordable housing in their purchases of multifamily mortgages.^[58]

B. Multifamily Housing Goal Benchmark Levels

The final rule sets the multifamily housing goals at benchmark levels intended to encourage the Enterprises to provide liquidity and to support various multifamily finance market segments in a safe and sound manner. The Enterprises have served as a stabilizing force in the multifamily market in the years since the financial crisis. During the conservatorship period, the Enterprise portfolios of loans on multifamily affordable housing properties have experienced low levels of delinquency and default, similar to the performance of Enterprise loans on market rate properties. In light of this performance, the Enterprises should be able to sustain or increase their volume of purchases of loans on affordable multifamily housing properties without adversely impacting the Enterprises' safety and soundness or negatively affecting the performance of their total loan portfolios.

FHFA continues to monitor the activities of the Enterprises, both in FHFA's capacity as regulator and as conservator. If necessary, FHFA will make appropriate changes in the benchmark levels for the multifamily housing goals to ensure the Enterprises' continued safety and soundness.

1. Multifamily Low-Income Housing Goal.

The multifamily low-income housing goal is based on the total number of rental units in multifamily properties financed by mortgages purchased by the Enterprises that are affordable to low-income families, defined as families with incomes less than or equal to 80 percent of AMI. The final rule sets the annual benchmark level for the low-income multifamily housing goal for each Enterprise at 315,000 units in each year from 2018 through 2020.

Table 6—Multifamily Low-Income Housing Goal

Year	Historical Performance					2017	2018-2020
	2012	2013	2014	2015	2016
Fannie Mae Goal	285,000	265,000	250,000	300,000	300,000	300,000	315,000
Freddie Mac Goal	225,000	215,000	200,000	300,000	300,000	300,000	315,000
Fannie Mae Performance:
Low-Income Multifamily Units	375,924	326,597	260,124	307,510	352,368
Total Multifamily Units	501,256	430,751	372,089	468,798	552,785
Low-Income % Total	75.0%	75.8%	69.9%	65.6%	63.7%
Freddie Mac Performance:
Low-Income Multifamily Units	298,529	254,628	273,807	379,042	406,958
Total Multifamily Units	377,522	341,921	366,377	514,275	597,399
Low-Income % of Total Units	79.1%	74.5%	74.7%	73.7%	68.1%

Recent performance and forecasts. As shown in Table 6, from 2012 through 2016, both Enterprises exceeded the low-income multifamily goal. Prior to 2015, Fannie Mae had higher goals than Freddie Mac. For the 2015-2017 goal period, FHFA set the same benchmark levels for both Enterprises for the first time, reflecting parity between Freddie Mac and Fannie Mae multifamily market share in terms of unit counts.

In 2016, the goal for each Enterprise was 300,000 units. Fannie Mae purchased mortgages financing 352,368 low-income units, and Freddie Mac purchased mortgages financing 406,958 low-income units. While total volumes have increased, the share of low-income units financed at each Enterprise has been declining from peak levels in 2012.

Industry forecasts and FHFA internal forecasts for the overall multifamily originations market indicate a modest increase in 2017 over 2016 and a decrease in 2018.

Proposed rule and comments. In the proposed rule, FHFA proposed setting the benchmark for 2018-2020 at 315,000 units. Three commenters supported the proposed benchmark levels for the multifamily goals. One commenter stated, “the goals are only meaningful if they are achievable.” The three commenters that argued for higher goals did not suggest a specific number. One commenter (Fannie Mae) suggested lowering the low-income multifamily goal to 300,000 units, which was the 2015-2017 benchmark level. Regardless of whether they supported the proposed benchmark levels or supported different benchmark levels, commenters pointed out the particular difficulty for renters in finding affordable units and paying for them, given decreasing affordable rental housing stock, stagnant wages, and rapid rent increases in recent years. Several commenters pointed out that the overall multifamily market had been strong and growing, and the demand for rental housing is projected to continue to increase in coming years.

FHFA determination. As discussed above, the Conservatorship Scorecard cap has been lowered to $35 billion for 2018. Because the Scorecard cap has been designed to exclude affordable housing goal categories, lowering the cap should not significantly impact the ability of the Enterprises to meet the multifamily housing goals. However, FHFA expects that availability of housing subsidies will likely continue to be challenging for renter households. As a result, the gap between the supply of low-income and very low-income units and the needs of low-income households, as described in the affordability discussion above, is expected to continue in the next goal period. These trends, along with industry forecasts and FHFA internal forecasts, support a cautious approach in considering any increase in the benchmark levels for the multifamily housing goals.

Given recent Enterprise performance and balancing these considerations, the final rule sets the annual benchmark level for the low-income multifamily housing goal for each Enterprise at 315,000 units in each year from 2018 through 2020, a modest increase from the 300,000 unit goal for each Enterprise in 2015-2017.

2. Multifamily Very Low-Income Housing Subgoal

The multifamily very low-income housing subgoal is based on the total number of rental units in multifamily properties financed by mortgages purchased by the Enterprises that are affordable to very low-income families, defined as families with incomes no greater than 50 percent of AMI. The final rule sets the benchmark level for the very low-income multifamily housing subgoal for each Enterprise at 60,000 units for each year from 2018 through 2020.

Table 7—Multifamily Very Low-Income Subgoal

Year	Historical performance
	2012	2013	2014	2015	2016	2017	2018-2020
Fannie Mae Goal	80,000	70,000	60,000	60,000	60,000	60,000	60,000
Freddie Mac Goal	59,000	50,000	40,000	60,000	60,000	60,000	60,000
Fannie Mae Performance:
Very Low-Income Multifamily Units	108,878	78,071	60,542	69,078	65,910
Total Multifamily Units	501,256	430,751	372,089	468,798	552,785
Very Low-Income % of Total Units	21.7%	18.1%	16.3%	14.7%	11.9%
Freddie Mac Performance:
Very Low-Income Multifamily Units	60,084	56,752	48,689	76,935	73,030
Total Home Purchase Mortgages	377,522	341,921	366,377	514,275	597,399
Very Low-Income % of Total Units	15.9%	16.6%	13.3%	15.0%	12.2%

Recent performance and forecasts. As shown in Table 7, from 2012 through 2016, both Enterprises exceeded the very low-income multifamily subgoal. In 2016, the subgoal for each Enterprise was 60,000 units. Fannie Mae purchased mortgages financing 65,910 very low-income units, while Freddie Mac purchased mortgages financing 73,030 very low-income units. Similar to the low-income multifamily goal, the share of very low-income units financed at each Enterprise has been declining in recent years.

As discussed above, industry forecasts and FHFA internal forecasts for the overall multifamily originations market indicate a modest increase in 2017 over 2016 and a decrease in 2018.

Proposed rule and comments. In the proposed rule, FHFA proposed setting the very low-income multifamily subgoal at 60,000 units. Three commenters supported the proposed benchmark levels for the multifamily goals. The three commenters that argued for higher goals did not suggest a specific number. One commenter (Fannie Mae) suggested lowering the very low-income goal to 55,000 units. Regardless of whether they supported the proposed benchmarks or supported different benchmarks, commenters pointed out the particular difficulty for renters in finding affordable units and paying for them, given decreasing affordable stock, stagnant wages, and rapid rent increases in recent years. Several comments pointed out the fact that the overall multifamily market had been strong and growing, and the demand for rental housing is projected to continue to increase in coming years.

FHFA determination. The very low-income multifamily market faces many of the same constraints as the low-income multifamily market. However, very low-income multifamily housing is inherently even more difficult to build, finance, and maintain, and a larger element of public subsidy is required to make such projects viable. The availability of public subsidies has been severely diminished in recent years, and FHFA expects the availability of subsidies to remain at historically low levels or decline further. The recent disruption in the tax credit market, described above, will pose an additional challenge to the very low-income multifamily market. These factors suggest moderation in setting the benchmark level for the very low-income multifamily subgoal for the Enterprises.

Given the challenges associated with the Enterprises meeting this housing goal and the trends described, the final rule sets the benchmark level for the very low-income multifamily housing subgoal for each Enterprise at 60,000 units for each year from 2018 through 2020, the same as the 60,000 unit goal for each Enterprise in 2015-2017.

3. Small Multifamily Low-Income Housing Subgoal

A small multifamily property is defined for purposes of the housing goals as a property with 5 to 50 units. The small multifamily low-income housing subgoal is based on the total number of units in small multifamily properties financed by mortgages purchased by the Enterprises that are affordable to low-income families, defined as families with incomes less than or equal to 80 percent of AMI. The final rule sets the benchmark level for the small multifamily subgoal for each Enterprise at 10,000 units for each year from 2018 through 2020.

Table 8—Small Multifamily Low-Income Subgoal

Year	Historical performance
	2012	2013	2014	2015	2016	2017	2018-2020
Small Low-Income Multifamily Goal				6,000	8,000	10,000	10,000
Fannie Mae Performance:
Small Low-Income Multifamily Units	16,801	13,827	6,732	6,731	9,312
Total Small Multifamily Units	26,479	21,764	11,880	11,198	15,211
Low-Income % of Total Small Multifamily Units	63.5%	63.5%	56.7%	60.1%	61.2%
Freddie Mac Performance:
Small Low-Income Multifamily Units	829	1,128	2,076	12,801	22,101
Total Small Multifamily Units	2,194	2,375	4,659	21,246	33,984
Low-Income % of Total Small Multifamily Units	37.8%	47.5%	44.6%	60.3%	65.0%

Recent performance and forecasts. The small multifamily low-income housing subgoal was a new subgoal established by regulation for the 2015-2017 goal period. The subgoal was set at 6,000 units in 2015, 8,000 units in 2016, and 10,000 units in 2017. As shown in Table 8, both Enterprises exceeded the subgoal of 8,000 units in 2016. Fannie Mae purchased mortgages financing 9,312 units, and Freddie Mac purchased mortgages financing 22,101 units. As discussed above, industry forecasts and FHFA internal forecasts for the overall multifamily originations market indicate a modest increase in 2017 over 2016 and a decrease in 2018.

Proposed rule and comments. In the proposed rule, FHFA proposed setting the small multifamily subgoal at 10,000 units for each year. FHFA received five comments specifically on the small multifamily goal, and those comments were generally positive. For example, one commenter stressed the importance of small multifamily properties and the lack of “consistent access to secondary market liquidity,” and stated that the proposed benchmark levels for 2018-2020 were appropriate. Further, the commenter stated, “keeping these goals at an achievable level keeps them as meaningful incentives.” Other commenters also supported the benchmark levels and maintaining the small multifamily low-income subgoal. There were two commenters that recommended that FHFA increase the benchmark level for the small multifamily low-income subgoal, but neither commenter specified a number.

FHFA determination. The final rule sets the annual small multifamily subgoal for each Enterprise at 10,000 units for each year from 2018 through 2020, the same as the 2017 goal. The Enterprises continue to innovate in their approaches to serving this market. FHFA is still monitoring the trends in this market segment as well as Enterprise performance for this new subgoal. Maintaining the current goal should continue to encourage the Enterprises' participation in this market and ensure the Enterprises have the expertise necessary to serve this market should private sources of financing become unable or unwilling to lend on small multifamily properties.

Given the importance of this market segment, the final rule sets the benchmark level for the small multifamily subgoal for each Enterprise at 10,000 units for each year from 2018 through 2020, the same as the 10,000 unit subgoal for each Enterprise in 2017.

VI. Section-by-Section Analysis of Other Changes

The final rule also revises other provisions of the housing goals regulation, as discussed below.

A. Changes to Definitions—Proposed § 1282.1

Consistent with the proposed rule, the final rule includes changes to definitions used in the current housing goals regulation. Specifically, the final rule revises the definitions of “median income,” “metropolitan area,” and “non-metropolitan area” and removes the definition of “AHS.”

1. Definition of “Median Income”

The current regulation defines “median income” as the unadjusted median family income estimates for an area as most recently determined by HUD. While this definition accurately identifies the source that FHFA uses to determine median incomes each year, the definition does not reflect the longstanding practice FHFA has followed in providing the Enterprises with the median incomes that the Enterprises must use each year. The final rule revises the definition to be clear that the Enterprises are required to use the median incomes provided by FHFA each year in determining affordability for purposes of the housing goals.

The final rule also makes two additional technical changes to the definition of “median income.” First, the final rule adds a reference to “non-metropolitan areas” in the definition because FHFA determines median incomes for both metropolitan areas and non-metropolitan areas each year. Second, the final rule removes the word “family” in one place so that the term “median income” is used consistently throughout the regulation.

The revised definition reads: “Median income means, with respect to an area, the unadjusted median family income for the area as determined by FHFA. FHFA will provide the Enterprises annually with information specifying how the median family income estimates for metropolitan and non-metropolitan areas are to be applied for purposes of determining median income.”

Comments on Proposed Rule and FHFA determination. FHFA did not receive any comments on these technical revisions, and the final rule adopts the changes as proposed.

2. Definitions of “Metropolitan Area” and “Non-Metropolitan Area”

The current regulation defines both “metropolitan area” and “non-metropolitan area” based on the areas for which HUD defines median family incomes. The definition of “metropolitan area” refers to median family income estimates “determined by HUD,” while the definition of “non-metropolitan area” refers to median family income estimates “published annually by HUD.”

To be consistent with the changes to the definition of “median income,” the final rule revises the definition of “metropolitan area” by replacing the phrase “for which median family income estimates are determined by HUD” with the phrase “for which median incomes are determined by FHFA.” For the same reason, the final rule revises the definition of “non-metropolitan area” by replacing the phrase “for which median family income estimates are published annually by HUD” with the phrase “, for which median incomes are determined by FHFA.”

Comments on Proposed Rule and FHFA determination. FHFA did not receive any comments on these technical revisions, and the final rule adopts the changes as proposed.

3. Definition of “AHS” (American Housing Survey)

Consistent with the proposed rule, the final rule removes the definition of “AHS” from § 1282.1 because the term is no longer used in the Enterprise housing goals regulation.

Prior to the 2015 amendments to the Enterprise housing goals regulation, the term “AHS” was used to specify the data source from which FHFA derives the utility allowances used to determine the total rent for a rental unit which, in turn, is used to determine the affordability of the unit when actual utility costs are not available. The 2015 amendments consolidated and simplified the definitions applicable to determining the total rent and eliminated the reference to AHS in the part of the definition related to utility allowances, providing FHFA with flexibility in how it determines the nationwide average utility allowances. The current nationwide average utility allowances are still fixed numbers based on AHS data, but the regulation does not require FHFA to rely solely on AHS data to determine those utility allowances. The term “AHS” is not used anywhere else in the regulation, so the final rule removes the definition from § 1282.1.

Comments on Proposed Rule and FHFA determination. FHFA did not receive any comments on this technical revision, and the final rule adopts the change as proposed.

B. Data Source for Estimating Affordability of Multifamily Rental Units—Proposed § 1282.15(e)(2)

The final rule revises § 1282.15(e)(2) to update the data source used by FHFA to estimate affordability where actual information about rental units in a multifamily property is not available.

Section 1282.15(e)(3) permits the Enterprises to use estimated affordability information to determine the affordability of multifamily rental units for up to 5 percent of the total multifamily rental units in properties securing mortgages purchased by the Enterprise each year when actual rental information about the units is not available. The estimations are based on the affordable percentage of all rental units in the census tract in which the property for which the Enterprise is estimating affordability is located.

The current regulation provides that the affordable percentage of all rental units in the census tract will be determined by FHFA based on the most recent decennial census. However, the 2000 decennial census was the last decennial census that collected this information. The U.S. Census Bureau now collects this information through the ACS. Since 2011, FHFA has used the most recent data available from the ACS to determine the affordable percentage of rental units in a census tract for purposes of estimating affordability. The final rule revises § 1282.15(e)(2) to reflect this change. To take into account possible future changes in how rental affordability data is collected, the revised sentence does not refer specifically to data derived from the ACS. The final rule revises § 1282.15(e)(2) to replace the phrase “as determined by FHFA based on the most recent decennial census” with the phrase “as determined by FHFA.”

Comments on Proposed Rule and FHFA determination. FHFA did not receive any comments on this change, and the final rule adopts the change as proposed.

C. Determination of Median Income for Certain Census Tracts—Proposed § 1282.15(g)(2)

Consistent with the proposed rule, the final rule revises § 1282.15(g) to remove paragraph (g)(2), an obsolete provision describing the method that the Enterprises were required to use to determine the median income for a census tract where the census tract was split between two areas with different median incomes.

Current § 1282.15(g)(2) requires the Enterprises to use the method prescribed by the Federal Financial Institutions Examination Council to determine the median income for certain census tracts that were split between two areas with different median incomes. This provision was put in place by the 1995 final rule published by HUD establishing Enterprise housing goals under the Safety and Soundness Act.^[59]

As discussed above regarding the definition of “median income,” the process of determining median incomes has changed over the years, so that the Enterprises are now required to use median incomes provided by FHFA each year when determining affordability for purposes of the housing goals. Because FHFA provides median incomes for every location in the United States, it is no longer necessary for the regulation to set forth a process for the Enterprises to use when it is not certain what the applicable median income would be for a particular location. Consequently, the final rule removes § 1282.15(g)(2) from the regulation and renumbers § 1282.15(g)(1).

Comments on Proposed Rule and FHFA determination. FHFA did not receive any comments on this change, and the final rule adopts the change as proposed.

D. Housing Plan Timing—Proposed § 1282.21(b)(3)

Consistent with the proposed rule, the final rule revises § 1282.21(b)(3) to make clear that the Director has discretion to determine the appropriate period of time that an Enterprise may be subject to a housing plan to address a failure to meet a housing goal.

The final rule revises § 1282.21(b)(3) to state explicitly that a housing plan that is required based on an Enterprise's failure to achieve a housing goal will be required to address a time period determined by the Director. If FHFA requires an Enterprise to submit a housing plan, FHFA will notify the Enterprise of the applicable time period in FHFA's final determination on the housing goals performance of the Enterprise for a particular year. This change is based on (1) FHFA's experience in overseeing the housing goals, in particular the experience in requiring Freddie Mac to submit a housing plan based on its failure to achieve certain housing goals in 2014 and 2015, (2) the inherent conflict in the timeframes set out in the Safety and Soundness Act, and (3) the importance of ensuring that any housing plans are focused on sustainable improvements in Enterprise goals performance.

Comments on Proposed Rule. FHFA received four comments on this proposed revision. One commenter supported the revision and FHFA's efforts to provide “a clear and transparent process by which [the Enterprise] is expected to carry out the housing plan.” One commenter was supportive but recommended that the housing plan timing be “time bound and defined,” rather than left to the discretion of the Director. Two commenters recommended a tougher approach to enforcement of the goals and encouraged FHFA to impose civil and monetary penalties for failure to meet the goals. One commenter also requested that FHFA publish the housing plans and progress reports, and provide an opportunity for the public to review and comment on the housing plans.

FHFA determination. The final rule amends § 1282.21(b)(3) to provide that a housing plan will be required to address a time period determined by the Director. This change is consistent with the proposed rule. The final rule does not define the applicable time period, which will allow FHFA to establish an appropriate time period based on the facts and circumstances in each case.

FHFA is committed to enforcing the housing goals as provided in the Safety and Soundness Act. FHFA required that an Enterprise submit a housing plan for the first time in 2015. FHFA required Freddie Mac to submit a housing plan for 2016-2017 based on Freddie Mac's failure to meet the low-income and very low-income housing goals in 2013 and 2014. FHFA extended the housing plan through 2018 after Freddie Mac failed to meet the same goals in 2015. Freddie Mac submitted detailed proposals for improving its performance on those housing goals in the housing plan, and Freddie Mac continues to provide regular updates to FHFA. The Safety and Soundness Act provides for enforcement through civil money penalties and cease and desist orders if an Enterprise refuses to submit a housing plan when required, submits an unacceptable plan, or fails to comply with a housing plan.^[60] FHFA may take such action in appropriate circumstances.

When FHFA has required an Enterprise to submit a housing plan based on a failure to meet one or more housing goals, FHFA has required that the housing plan include detailed plans for future business initiatives and other actions that the Enterprise will take to improve its performance on the housing goals. For example, the Freddie Mac housing plan included proprietary forecasts for specific initiatives and programs that Freddie Mac is undertaking to improve its performance on the applicable housing goals. The level of detail required means that almost all of the information in the housing plan will be competitively sensitive. For that reason, the final rule does not provide for publication of any housing plan that an Enterprise may be required to submit. FHFA values the input of external entities on this process and recognizes commenters' desires for more information. FHFA will continue to review policies and procedures related to housing goals enforcement and may consider options to increase transparency related to Enterprise housing plans, either by future rulemaking or other changes to FHFA's processes.

VII. Paperwork Reduction Act

This final rule does not contain any information collection requirement that would require the approval of the Office of Management and Budget (OMB) under the Paperwork Reduction Act (44 U.S.C. 3501 et seq.). Therefore, FHFA has not submitted any information to OMB for review.

VIII. Regulatory Flexibility Act

The Regulatory Flexibility Act (5 U.S.C. 601 et seq.) requires that a regulation that has a significant economic impact on a substantial number of small entities, small businesses, or small organizations must include an initial regulatory flexibility analysis describing the regulation's impact on small entities. Such an analysis need not be undertaken if the agency has certified that the regulation will not have a significant economic impact on a substantial number of small entities. 5 U.S.C. 605(b). FHFA has considered the impact of this final rule under the Regulatory Flexibility Act. The General Counsel of FHFA certifies that the rule is not likely to have a significant economic impact on a substantial number of small entities because the rule applies to Fannie Mae and Freddie Mac, which are not small entities for purposes of the Regulatory Flexibility Act.

List of Subjects in 12 CFR Part 1282

• Mortgages
• Reporting and recordkeeping requirements

Authority and Issuance

For the reasons stated in the SUPPLEMENTARY INFORMATION, under the authority of 12 U.S.C. 4511, 4513 and 4526, FHFA amends part 1282 of Title 12 of the Code of Federal Regulations as follows:

PART 1282—ENTERPRISE HOUSING GOALS AND MISSION

1. The authority citation for part 1282 continues to read as follows:

Authority: 12 U.S.C. 4501, 4502, 4511, 4513, 4526, 4561-4566.

2. Amend § 1282.1 as follows:

a. Remove the definition of “AHS”; and

b. Revise the definitions of “Median income,” “Metropolitan area,” and “Non-metropolitan area.”

The revisions read as follows:

§ 1282.1

Definitions.

* * * * *

Median income means, with respect to an area, the unadjusted median family income for the area as determined by FHFA. FHFA will provide the Enterprises annually with information specifying how the median family income estimates for metropolitan and non-metropolitan areas are to be applied for purposes of determining median income.

Metropolitan area means a metropolitan statistical area (MSA), or a portion of such an area, including Metropolitan Divisions, for which median incomes are determined by FHFA.

* * * * *

Non-metropolitan area means a county, or a portion of a county, including those counties that comprise Micropolitan Statistical Areas, located outside any metropolitan area, for which median incomes are determined by FHFA.

* * * * *

3. Revise paragraphs (c)(2), (d)(2), (f)(2), and (g)(2) of § 1282.12 to read as follows:

§ 1282.12

Single-family housing goals.

* * * * *

(c) * * *

(2) The benchmark level, which for 2018, 2019 and 2020 shall be 24 percent of the total number of purchase money mortgages purchased by that Enterprise in each year that finance owner-occupied single-family properties.

(d) * * *

(2) The benchmark level, which for 2018, 2019 and 2020 shall be 6 percent of the total number of purchase money mortgages purchased by that Enterprise in each year that finance owner-occupied single-family properties.

* * * * *

(f) * * *

(2) The benchmark level, which for 2018, 2019 and 2020 shall be 14 percent of the total number of purchase money mortgages purchased by that Enterprise in each year that finance owner-occupied single-family properties.

(g) * * *

(2) The benchmark level, which for 2018, 2019 and 2020 shall be 21 percent of the total number of refinancing mortgages purchased by that Enterprise in each year that finance owner-occupied single-family properties.

4. Revise § 1282.13 to read as follows:

§ 1282.13

Multifamily special affordable housing goal and subgoals.

(a) Multifamily housing goal and subgoals. An Enterprise shall be in compliance with a multifamily housing goal or subgoal if its performance under the housing goal or subgoal meets or exceeds the benchmark level for the goal or subgoal, respectively.

(b) Multifamily low-income housing goal. The benchmark level for each Enterprise's purchases of mortgages on multifamily residential housing affordable to low-income families shall be at least 315,000 dwelling units affordable to low-income families in multifamily residential housing financed by mortgages purchased by the Enterprise in each year for 2018, 2019, and 2020.

(c) Multifamily very low-income housing subgoal. The benchmark level for each Enterprise's purchases of mortgages on multifamily residential housing affordable to very low-income families shall be at least 60,000 dwelling units affordable to very low-income families in multifamily residential housing financed by mortgages purchased by the Enterprise in each year for 2018, 2019, and 2020.

(d) Small multifamily low-income housing subgoal. The benchmark level for each Enterprise's purchases of mortgages on small multifamily properties affordable to low-income families shall be at least 10,000 dwelling units affordable to low-income families in small multifamily properties financed by mortgages purchased by the Enterprise in each year for 2018, 2019, and 2020.

§ 1282.15

[Amended]

5. Amend § 1282.15 as follows:

a. In paragraph (e)(2) remove the phrase “based on the most recent decennial census”; and

b. Revise paragraph (g).

The revisions read as follows:

§ 1282.15

General counting requirements.

* * * * *

(g) Application of median income. For purposes of determining an area's median income under §§ 1282.17 through 1282.19 and the definitions in § 1282.1, the area is:

(1) The metropolitan area, if the property which is the subject of the mortgage is in a metropolitan area; and

(2) In all other areas, the county in which the property is located, except that where the State non-metropolitan median income is higher than the county's median income, the area is the State non-metropolitan area.

* * * * *

6. Amend § 1282.21 by revising paragraph (b)(3) to read as follows:

§ 1282.21

Housing plans.

* * * * *

(b) * * *

(3) Describe the specific actions that the Enterprise will take in a time period determined by the Director to improve the Enterprise's performance under the housing goal; and

* * * * *

Dated: February 5, 2018.

Melvin L. Watt,

Director, Federal Housing Finance Agency.

Footnotes

1.  See 12 U.S.C. 4561(a).

2.  See 12 U.S.C. 4501(7).

3.  Housing and Economic Recovery Act of 2008, Public Law 110-289, 122 Stat. 2654 (July 30, 2008).

4.  See 75 FR 55892.

5.  See 77 FR 67535.

6.  See 80 FR 53392.

7.  See https://www.ffiec.gov/​hmda/​pdf/​2017letter.pdf for complete list of institutions required to report under HMDA. For 2016, this included depositories with greater than $44 million in assets and non-depositories with greater than $10 million in assets that originated more than 100 home purchase and refinance loans.

8.  See https://www.federalreserve.gov/​publications/​files/​2016_​HMDA.pdf.

9.  See Home Mortgage Disclosure Act final rule, 80 FR 66128 (Oct. 28, 2015).

10.  12 CFR 1282.14(d).

11.  12 CFR 1282.21(a).

12.  82 FR 31514 (July 7, 2017).

13.  12 CFR 1282.16(e).

14.  75 FR 55892, 55895 (Sept. 14, 2010).

15.  81 FR 96242, 96251 (Dec. 29, 2016).

16.  80 FR at 53429.

17.  FHFA found insufficient data supporting the Freddie Mac suggested “MHC Adjustment Factor” for determining affordability. The $450/unit estimate suggested by Freddie Mac was based on a very small and non-national sample, provided by an appraiser and is not suitable for a nationwide proxy.

18.  The Enterprise housing goals also include a low-income areas home purchase goal. The low-income areas goal benchmark level is established by a two-step process. The first step is setting the benchmark level for the low-income areas subgoal, as established by this final rule. The second step is establishing an additional increment for mortgages to families located in federally-declared disaster areas with incomes less than or equal to AMI. Each year, FHFA sets the disaster area increment separately from this rule and notifies the Enterprises by letter of the benchmark level for that year. The final rule sets the annual low-income areas home purchase goal benchmark level for 2018 through 2020 at the subgoal benchmark level of 14 percent plus a disaster areas increment that FHFA will set separately each year.

19.  See 2017 Scorecard for Fannie Mae, Freddie Mac, and Common Securitization Solutions, December 2016, available at https://www.fhfa.gov/​AboutUs/​Reports/​ReportDocuments/​2017-Scorecard-for-Fannie-Mae-Freddie-Mac-and-CSS.pdf.

20.  12 U.S.C. 4562(e)(2).

21.  Details on FHFA's single-family market models are available in the most recent technical paper, “The Size of the Affordable Mortgage Market: 2018-2020 Enterprise Single-Family Housing Goals.”

22.  The macroeconomic outlook described here is based on Moody's and other forecasts as of August 2017.

23.  This refers to the mortgages insured/guaranteed by government agencies such as FHA, Department of Veterans Affairs (VA), and Rural Housing Service (RHS).

24.  Board of Governors of the Federal Reserve System. Federal Open Market Committee Press Release, September 20, 2017.

25.  Bureau of Labor Statistics, The Employment Situation—November 2017, published on December 22, 2017.

26.  NAR's housing affordability index is a national index. It does not capture regional differences. It measures, nationally, whether an average family could qualify for a mortgage on a typical home. A typical home is defined as the national median-priced, existing single-family home as reported by NAR. An average family is defined as one earning the median family income. The calculation assumes a down payment of 20 percent of the home price and a monthly payment that does not exceed 25 percent of the median family income. An index value of 100 means that a family earning the median family income has exactly enough income to qualify for a mortgage on a median-priced home. An index value above 100 signifies that a family earning the median family income has more than enough income to qualify for a mortgage on a median-priced home. A decrease in the index value over time means that housing is becoming less affordable.

27.  National Association of Counties, “County Economies 2016: Widespread Recovery, Slower Growth,” February 2017: available at http://www.naco.org/​sites/​default/​files/​documents/​County-Economies-2016.pdf.

28.  The supply of single-family homes at the more affordable end of the market also impacts a low-income or very low-income household's ability to purchase a home. See The State of the Nation's Housing 2017, Joint Center for Housing Studies of Harvard University, June 2017.

29.  For example, according to the State of the Nation's Housing 2017 Report, the construction of single-family homes has shifted toward larger, more expensive homes in recent years. The share of small-size single-family homes (under 1,800 square feet) dropped from 37 percent of all construction completions to 21 percent in 2015, while the share of large-size homes (over 3,000 square feet) almost doubled from 17 percent to 31 percent.

30.  See Income and Poverty in the United States: 2016, United States Census Bureau, September 2017: https://www.census.gov/​content/​dam/​Census/​library/​publications/​2017/​demo/​P60-259.pdf.

31.  Daniel McCue, Christopher Herbert, Working Paper: Updated Household Projections, 2015-2035: Methodology and Results, Joint Center for Housing Studies of Harvard University, December 2016.

32.  Details are available in the market model paper, “The Size of the Affordable Mortgage Market: 2018-2020 Enterprise Single-Family Housing Goals,” available at http://www.fhfa.gov/​PolicyProgramsResearch/​Research/​PaperDocuments/​Market-Estimates_​2018-2020.pdf.

33.  82 FR 31514 (July 7, 2017).

34.  Disaster declarations are listed on the Federal Emergency Management Agency (FEMA) website at https://www.fema.gov/​disasters.

35.  The goal has included permanent HAMP modifications to low-income borrowers in the numerator and all HAMP permanent modifications in the denominator.

36.  The HAMP program expired at the end of 2016. There will be some HAMP modifications that will count toward the Enterprise housing goals in 2017 as applications that were initiated before the end of the program are converted to permanent modifications.

37.  12 U.S.C. 4563(a)(4).

38.  The Consumer Financial Protection Bureau (CFPB) will collect additional data fields (including the number of units in the properties securing each multifamily loan that is reported) beginning in 2018 that may be useful in the future in considering whether to create a retrospective market measure for the multifamily housing goals.

39.  12 U.S.C. 4563(c).

40.  See https://www.huduser.gov/​portal/​pdredge/​pdr_​edge_​featd_​article_​092214.html for description of the Brooke Amendment and background on the definition of affordability embedded in the housing goals.

41.  “America's Rental Housing: Expanding Options for Diverse and Growing Demand,” Joint Center for Housing Studies of Harvard University, December 2015.

42.  Source: http://www.nmhc.org/​Content.aspx?​id=​4708#Type_​of_​Structure. Accessed 10/30/2017.

43.  “State of the Nation's Housing 2017,” Joint Center for Housing Studies of Harvard University,” June 2017.

44.  Id.

45.  “America's Rental Housing,” Joint Center for Housing Studies of Harvard University, January 2018.

46.  “State of the Nation's Housing 2017,” Joint Center for Housing Studies of Harvard University, June 2017. Available at www.jchs.harvard.edu/​research/​state_​nations_​housing.

47.  Id.

48.  “Renting in America's Largest Metropolitan Areas,” NYU Furman Center, March 2016.

49.  “The Gap: The Affordable Housing Gap Analysis 2017,” National Low Income Housing Coalition, March 2017.

50.  “State of the Nation's Housing 2017,” Joint Center for Housing Studies of Harvard University, June 2017.

51.  “2012 Rental Housing Finance Survey,” U.S. Census Bureau and U.S. Department of Housing and Urban Development, Tables 2b, 2c, 2d and 3.

52.  LIHTC is a supply-side subsidy created under the Tax Reform Act of 1986 and is the main source of new affordable rental housing construction in the United States today. Tax credits are used for the acquisition, rehabilitation, and/or new construction of rental housing for low-income and very low-income households. LIHTC has facilitated the creation or rehabilitation of approximately 2.4 million affordable rental units since 1986.

53.  Novogradac & Company, “Final Tax Reform Bill Would Reduce Affordable Rental Housing Production by Nearly 235,000.” December 19, 2017.

54.  “Worst Case Housing Needs: 2017 Report to Congress,” U.S. Department of Housing and Urban Development, August 2017. Renters with worst case needs have very low incomes, lack housing assistance, and have either severe rent burdens or severely inadequate housing (or both).

55.  Urban Institute, “The GSEs' Shrinking Role in the Multifamily Market,” April 2015.

56.  MBA, 2016 Annual Report on Multifamily Lending, October 2017.

57.  MBA, 2015 Annual Report on Multifamily Lending, October 2016.

58.  For more information on the Conservatorship Scorecard, see https://www.fhfa.gov/​AboutUs/​Reports/​ReportDocuments/​2017-Scorecard-for-Fannie-Mae-Freddie-Mac-and-CSS.pdf.

59.  See 60 FR 61846 (Dec. 1, 1995).

60.  12 U.S.C. 4566(c)(1).