2017-14039. 2018-2020 Enterprise Housing Goals

AGENCY:

Federal Housing Finance Agency.

ACTION:

Proposed rule.

SUMMARY:

The Federal Housing Finance Agency (FHFA) is issuing a proposed rule with request for comments 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 existing housing goals for the Enterprises include benchmark levels for each housing goal through the end of 2017. This proposed rule would establish benchmark levels for each of the housing goals and subgoals for 2018 through 2020. In addition, the proposed rule would make a number of clarifying and conforming changes, including revisions to the requirements for the housing plan that an Enterprise may be required to submit in response to a failure to achieve one or more of the housing goals.

DATES:

FHFA will accept written comments on the proposed rule on or before September 5, 2017.

ADDRESSES:

You may submit your comments on the proposed rule, identified by regulatory information number (RIN) 2590-AA81, by any one of the following methods:

• Agency Web site: www.fhfa.gov/​open-for-comment-or-input.
• Federal eRulemaking Portal: http://www.regulations.gov. Follow the instructions for submitting comments. If you submit your comment to the Federal eRulemaking Portal, please also send it by email to FHFA at [e-mail] to ensure timely receipt by FHFA. Include the following information in the subject line of your submission: Comments/RIN 2590-AA81.
• Hand Delivered/Courier: The hand delivery address is: Alfred M. Pollard, General Counsel, Attention: Comments/RIN 2590-AA81, Federal Housing Finance Agency, Eighth Floor, 400 Seventh Street SW., Washington, DC 20219. Deliver the package at the Seventh Street entrance Guard Desk, First Floor, on business days between 9 a.m. and 5 p.m.
• U.S. Mail, United Parcel Service, Federal Express, or Other Mail Service: The mailing address for comments is: Alfred M. Pollard, General Counsel, Attention: Comments/RIN 2590-AA81, Federal Housing Finance Agency, Eighth Floor, 400 Seventh Street SW., Washington, DC 20219. Please note that all mail sent to FHFA via U.S. Mail is routed through a national irradiation facility, a process that may delay delivery by approximately two weeks.

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. Comments

FHFA invites comments on all aspects of the proposed rule and will take all comments into consideration before issuing the final rule. Copies of all comments will be posted without change, including any personal information you provide such as your name, address, email address, and telephone number, on the FHFA Web site at http://www.fhfa.gov. In addition, copies of all comments received will be available for examination by the public on business days between the hours of 10 a.m. and 3 p.m., at the Federal Housing Finance Agency, 400 Seventh Street SW., Washington, DC 20219. To make an appointment to inspect comments, please call the Office of General Counsel at [phone].

Commenters are encouraged to review and comment on all aspects of the proposed rule, including the single-family benchmark levels, the multifamily benchmark levels, and other changes to the regulation.

II. 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.

Two-part approach. The performance of the Enterprises on the 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 proposing to continue that approach in this 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 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. 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 actually purchased by the Enterprises as well as comparable loans held in a lender's portfolio. It also includes comparable loans that are part of a private label security (PLS), although very few such securities have been issued for conventional conforming mortgages since 2008.

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. (Seasoned loans are loans that were originated in prior years and acquired by the Enterprise in the current year.) The Enterprises' acquisition of seasoned loans provides an important source of liquidity for this market segment.

Recent changes to the HMDA regulations will result in the HMDA data covering a greater portion of the single-family mortgage market.^[7] The changes will also provide more detailed information about the loans included in the HMDA data. The changes to the HMDA regulations generally take effect at the start of 2018, so the new, more detailed information will not be available until after the 2018 performance year.

For example, the Enterprise housing goals currently count all loans purchased by an Enterprise with original principal balances that are within the conforming loan limits. The conforming loan limits are different for single-family properties depending on the number of units in the property. However, the definition of the retrospective market excludes all loans with original principal balances above the conforming loan limits for single unit properties because the current HMDA data do not identify the number of units for each loan. Starting with the new HMDA data reported, it will be possible to identify the number of units for each loan. This may allow FHFA to revise the definition of the retrospective market to exclude only those loans above the conforming loan limits applicable to the size of the property, instead of excluding all loans above the conforming loan limit applicable to a single unit property.

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 proposing to make any changes to the Enterprise housing goals in anticipation of the upcoming changes to the 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 on multifamily properties with rental units affordable to very low-income families, as well as a small multifamily low-income subgoal for properties with 5-50 units. The multifamily goals established by FHFA in 2010, 2012, and 2015 evaluated 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. FHFA has not established 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 for 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 proposed 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.” ^[8] The proposal 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.^[9]

If, after publication of a final rule establishing the housing goals for 2018 through 2020, 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.

III. Summary of Proposed Rule

A. Benchmark Levels for the Single-Family Housing Goals

This proposed rule would establish the benchmark levels for the single-family housing goals and subgoal for 2018-2020 as follows:

Goal	Criteria	Current benchmark level for 2015-2017	Proposed benchmark level for 2018-2020
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 percent	24 percent.
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 percent	6 percent.
Low-Income Areas Home Purchase Subgoal	Home purchase mortgages on single-family, owner-occupied properties with:
	• Borrowers in census tracts with tract median income of no greater than 80 percent of area median income; or	14 percent	15 percent.
	• Borrowers with income 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 percent	21 percent.

B. Multifamily Housing Goal Levels

The proposed rule would establish the levels for the multifamily goal and subgoals for 2018-2020 as follows:

Goal	Criteria	Current goal level for 2017	Proposed goal level for 2018-2020
Low-Income 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 units	315,000 units.
Very Low-Income 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 units	60,000 units.
Low-Income Small Multifamily 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 units	10,000 units.

C. Other Proposed Changes

The proposed rule would make changes and clarifications to the existing rules, including minor technical changes to some regulatory definitions. The proposed rule also would revise 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 proposed rule sets out FHFA's views about benchmark levels for the single-family housing goals from 2018-2020. In making this proposal, FHFA has considered the required statutory factors described below. 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 access to credit expectations expressed in the most recent Conservatorship Scorecard, which requires the Enterprises to undertake a number of research and related efforts including the development of pilots and initiatives.^[10]

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.^[11]

FHFA has considered each of these seven statutory factors in setting the proposed 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 proposed 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 generate 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 proposed 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. Separate models were developed for each of the single-family housing goals and subgoals.

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 12 years of HMDA data, from 2004 to 2015, the latest year for which HMDA data are available. Additional discussion of the market forecast models can be found in a research paper, available at http://www.fhfa.gov/​PolicyProgramsResearch/​Research/​.^[12]

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 and adhering to 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 variable by using third party forecasts published by Moody's and other accredited mortgage market forecasters. The second variable is harder to address. The proposed rule relies on the most up-to-date data available as of December 2016, and uses forecasted input values for 2017 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 four years out is not the usual practice in forecasting. A number of industry forecasters, including Fannie Mae, Freddie Mac, and the Mortgage Bankers Association (MBA), 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 our market models, FHFA used Moody's forecasts, where available, as the source for macroeconomic variables.^[13] In cases where Moody's forecasts were not available (for example, the share of government-guaranteed home purchases and the share of government-guaranteed refinances), FHFA generated and tested its own forecasts.^[14] 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. At the July 2016 meeting of the Federal Open Market Committee (FOMC), policymakers indicated their commitment to a low federal funds rate for the time being, signaling a pause in the interest rate normalization path. However, there is broad consensus among economists that the Federal Reserve will resume rate hikes if the economy performs as expected. Based on Moody's January 2017 forecast, 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 projected to rise gradually from the current historic low of 3.6 percent in 2016 to 5.5 percent by 2020.

The unemployment rate has steadily fallen over the last few years and according to Moody's is expected to remain at 4.7 percent over the next four years, given expected growth of the economy at the modest range of 1.5 to 2.9 percent per year (January 2017 forecast). Moody's also forecasts a modest increase in per capita disposable nominal income growth—from $43,100 in 2016 to $50,300 in 2020. Moody's estimates that the inflation rate will remain flat at 2.0 percent throughout the same period, although this also depends on Federal Reserve policy.

Industry analysts generally expect the overall housing market to continue its recovery, although the growth of house prices may slow down, assuming continued increases in interest rates. According to Moody's forecast (as of January 2017) based on FHFA's purchase-only House Price Index (HPI), house prices are expected to increase at the annual rates of 3.9, 1.8, and 2.0 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' Housing Affordability Index, is projected to decline from an index value of 162.2 in 2016 to 152.5 in 2020. Low interest rates coupled with rising house prices usually create incentives for homeowners to refinance, and the refinance share of overall mortgage originations increased from 39.9 percent in 2014 to 50 percent in 2016. However, assuming that interest rates rise in the near future, the refinance rate is expected to fall below 21.4 percent by 2019, according to the Moody's forecast.

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 point within the model-generated confidence interval for each of the single-family housing goals. Some of these factors may affect a subset of the market rather than the market as a whole. Some of these additional 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 substantial impacts on the ability of the Enterprises to meet 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, the economy and the housing market continued to recover from the financial crisis, but the recovery has been uneven across the country. 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.

Trends in factors such as area median income (AMI) point to an uneven recovery. FHFA uses census-tract level AMIs published by the U.S. Department of Housing and Urban Development (HUD) to determine affordability for the 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.^[15] 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. Overall, while there are annual fluctuations in AMI, the trends over a longer period (for instance, over four years) indicate that the economy is recovering, albeit in an uneven manner. For instance, from 2014 to 2016, over 80 percent of census tracts experienced an AMI increase. Over the four-year period from 2012 to 2016, AMI increased in about 51 percent of census tracts. This unevenness of the economic recovery is particularly evident geographically. For instance, the census tracts that experienced more than a 10 percent decline in AMIs in 2016 are concentrated in the southern and midwestern regions of the country.

In addition to the uneven recovery reflected in changing AMI levels, 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 last year for the most recent year available reflected that while median household income increased by 5.2 percent in 2015, the first annual increase in median household income since 2007, median wages remained 1.6 percent lower than the median in 2007, the year before the most recent recession, and 2.4 percent lower than the median household income peak that occurred in 1999.^[16] 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 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. In addition, another continuing demographic trend is the growth of minority households, which is projected to be over 70 percent of net household growth through 2025.^[17] In light of the fact that the median net worth of minority households has been historically low, building the necessary wealth to meet down payment and closing costs will likely also 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.

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. The Enterprises' share of the market is continually subject to fluctuation. During the mortgage market bubble, the Enterprises' share of the market dropped to about 46 percent in the last quarter of 2005. The other significant low point occurred in 2008, when the Enterprises' acquisitions accounted for less than 45 percent of the mortgage market. Since then, the Enterprises' share has risen overall but declined slightly in recent years, accounting for about 52 percent of the market in 2015. As shown in Graph 1, over the same time period, the total government share of the mortgage market (including FHA, VA, and RHS) has been expanding. In 2015, the total government share accounted for 28 percent of overall mortgage originations, up from 24 percent in 2014. This is likely an impact of the FHA mortgage insurance premium reduction announced in January 2015. As seen in Graph 1, the increase in government share came from decreases in the other two segments.

Both Enterprises' charter acts require that all mortgages the Enterprises acquire have mortgage insurance (or one of the other forms of credit enhancement specified in the charter acts) if the loan-to-value (LTV) ratio for the loan at acquisition is greater than 80 percent. Private mortgage insurance rates are dependent on characteristics of the mortgage such as loan term, type of mortgage (purchase, type of refinance), LTV ratio, and credit score of the borrower. Lenders may also be able to negotiate and obtain lower private mortgage insurance directly from the mortgage insurer. Therefore, for certain market segments, the choice between government mortgage insurance or private mortgage insurance depends on the net impact of these factors.

In recent years private mortgage insurance rates have increased relative to government mortgage insurance rates, but the increase has not been uniform across the credit score and LTV spectrum. Changes in the mortgage insurance market can impact the cost of mortgage insurance and, consequently, may influence whether the mortgage is originated with private mortgage insurance or with FHA insurance. For example, FHA decreased its rates for mortgage insurance from 1.35 percent to 0.85 percent in January 2015. If FHA decreased or increased its mortgage insurance premiums, it would be reasonable to expect further shifts in the market that would not be uniform across the credit score and LTV spectrum. Reductions in the FHA insurance premium are likely to have two impacts on the conforming segment of the market: (1) The substitution effect—some borrowers will switch from private mortgage insurance to FHA insurance due to the lower premium rate; and (2) the expanded homeownership effect—new borrowers, especially those with lower credit scores seeking higher LTV loans, will enter the mortgage market because they are now able to meet the debt-to-income threshold due to the lower monthly mortgage payment. Analysis conducted by Federal Reserve Board staff indicates that both effects existed after the last FHA reduction.^[18] Increases in FHA premiums would likely result in reverse shifts.

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. Proposed 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 proposed rule would set the annual low-income home purchase housing goal benchmark level for 2018-2020 at 24 percent, the same as the current 2015-2017 benchmark level. FHFA believes 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.

Table 1—Enterprise Low-Income Home Purchase Goal

Year	Historical performance				Projected performance
	2013	2014	2015	2016	2017	2018	2019	2020
Actual Market	24.0%	22.8%	23.6%
Benchmark	23%	23%	24%	24%	24%
Current Market Forecast				23.9%	24.9%	25.5%	24.0%	23.0%
				+/−2.5%	+/−4.3%	+/−5.6%	+/−6.6%	+/−7.4%
Fannie Mae Performance:
Low-Income Home Purchase Mortgages	193,712	177,846	188,891	221,249
Total Home Purchase Mortgages	814,137	757,870	802,432	964,847
Low-Income % of Home Purchase Mortgages	23.8%	23.5%	23.5%	22.9%
Freddie Mac Performance:
Low-Income Home Purchase Mortgages	93,478	108,948	129,455	153,435
Total Home Purchase Mortgages	429,158	519,731	579,340	644,991
Low-Income % of Home Purchase Mortgages	21.8%	21.0%	22.3%	23.8%

As shown in Table 1, performance at both Enterprises has fallen short of the market in the low-income purchase goal almost every year since 2013 (with the exception of Fannie Mae in 2014), although the Enterprises have sometimes missed the market look-back goal only by one- or two-tenths of a percentage point. Performance at both Enterprises fell short of both the benchmark and the market level in 2015. The past performance of the Enterprises indicates that it has been difficult for the Enterprises to consistently 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 actual market rebounded to 23.6 percent. FHFA's current model forecasts that the market for this goal will increase slightly to 23.9 percent in 2016 and then to 24.9 percent in 2017. (Actual market levels for 2016 will not be available until HMDA data are published in September 2017.) Although the Enterprises have been challenged in meeting the percentage single-family housing goal levels in recent years, FHFA notes that each Enterprise has increased the number of single-family home purchase loans made to low-income households. Fannie Mae's eligible single-family loan purchases increased from 193,712 loans in 2013 to 221,249 in 2016. Freddie Mac's eligible single-family loan purchases increased from 93,478 in 2013 to 153,435 in 2016.

From 2018 to 2020, the proposed goals period, the current forecast peaks at 25.5 percent in 2018, before decreasing to 24.0 percent in 2019 and 23.0 percent in 2020. The average of these projections is 24.1 percent. This forecast is based on the latest data available and will be updated before the release of the final housing goals rule. The confidence intervals for the 2018-2020 goal period are wide, but they will narrow before the final rule is published.

FHFA is proposing a benchmark level for the low-income home purchase housing goal that is close to the market forecast, to encourage the Enterprises to continue to find ways to support lower income borrowers while not compromising safe and sound lending standards. FHFA notes that the proposed benchmark is close to the average of its market forecast for this goal. FHFA recognizes that there may be challenges to meeting this goal, including uneven growth in AMI and the relative affordability of private mortgage insurance, that may be beyond the control of the Enterprises and impact their ability to achieve these goals. FHFA will continue to monitor the Enterprises, both as regulator and as conservator, 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 can 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 proposed rule would set the annual very low-income home purchase housing goal benchmark level for 2018 through 2020 at 6 percent, also unchanged from the current 2015 to 2017 benchmark level.

Table 2—Very Low-Income Home Purchase Goal

Year	Historical performance				Projected performance
	2013	2014	2015	2016	2017	2018	2019	2020
Actual Market	6.3%	5.7%	5.8%
Benchmark	7%	7%	6%	6%	6%
Current Market Forecast				5.9%	6.4%	6.7%	6.3%	6.2%
				+/−0.8%	+/−1.4%	+/−1.8%	+/−2.1%	+/−2.4%
Fannie Mae Performance:
Very Low-Income Home Purchase Mortgages	48,810	42,872	45,022	49,852
Total Home Purchase Mortgages	814,137	757,870	802,432	964,847
Very Low-Income % of Home Purchase Mortgages	6.0%	5.7%	5.6%	5.2%
Freddie Mac Performance:
Very Low-Income Home Purchase Mortgages	23,705	25,232	31,146	36,838
Total Home Purchase Mortgages	429,158	519,731	579,340	644,991
Very Low-Income % of Home Purchase Mortgages	5.5%	4.9%	5.4%	5.7%

Since 2013, the market for very low-income home purchase loans has also been declining, as reflected in HMDA data, although there was a slight uptick in 2015. FHFA has gradually lowered the benchmark for this goal from 8 percent in 2010 to 6 percent in 2015. Despite this reduction, the performance of both Enterprises has fallen below the benchmark and the market levels in each year since 2013. In addition, both Enterprises are projected to fall below the 6 percent benchmark level in 2016.

FHFA market analysis reflects a relatively flat trend for this segment, at 5.7 percent in 2014 and 5.8 percent in 2015. FHFA's current model forecasted the market to increase slightly to 5.9 percent in 2016 and then to 6.4 percent in 2017. For the 2018-2020 goal period, FHFA's forecast indicates an increase to 6.7 percent in 2018, followed by declines to 6.3 percent and 6.2 percent in 2019 and 2020, respectively. As noted earlier, the confidence intervals widen as the forecast period lengthens, and will reduce somewhat as FHFA incorporates more information before publishing the final rule.

Similar to the low-income home purchase goal, FHFA is proposing a benchmark level that is near the market forecast 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 purchase goal discussion, FHFA believes that there are significant challenges to housing affordability that may be beyond the control of the Enterprises that could make the proposed benchmark a challenge for the Enterprises. As each Enterprise has been struggling to meet the current benchmark and market levels, the proposed benchmark will continue to encourage the Enterprise to safely and soundly innovate in this area. FHFA, as regulator and as conservator, will continue to monitor the Enterprises' performance, 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

Background. 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). Borrowers could qualify under either or both conditions. As noted in Table 3, mortgages satisfying condition (1) above (borrowers in low-income areas) are almost typically double the share of mortgages satisfying condition (2) (moderate-income borrowers in minority census tracts). For example, in 2015, 12.2 percent of mortgages met only condition (1), 7.6 percent met only condition (2), and 4.6 percent of mortgages met both conditions.

Table 3—Composition of Low-Income Areas Home Purchase Subgoal Based on HMDA Data

Year	Low-income area goal (%)	All low- income areas (%)	Low-income census tracts that are not high minority areas (%)	High minority areas that are also low-income census tracts (%)	High minority areas that are not low-income census tracts (%)	All high minority areas (%)
	(A) Grand Total	(B) LI	(C) LI, not HM	(D) HM and LI	(E) HM, not LI	(F) HM
Distribution of HMDA Borrowers by Census Tract Location:
2004	16.8	13.3	8.1	5.3	3.5	8.7
2005	15.3	12.5	8.3	4.2	2.8	7.0
2006	15.8	13.1	8.9	4.3	2.7	6.9
2007	16.2	13.3	8.5	4.8	3.0	7.7
2008	14.3	11.6	7.4	4.2	2.7	6.9
2009	13.1	10.0	5.9	4.1	3.0	7.2
2010	12.1	9.2	5.6	3.6	2.9	6.5
2011	11.4	8.8	5.5	3.3	2.6	5.9
2012	13.5	10.3	6.0	4.3	3.2	7.5
2013	14.1	10.9	6.6	4.3	3.1	7.4
2014	15.0	12.0	7.5	4.6	3.0	7.5
2015	15.1	12.2	7.6	4.6	2.9	7.5
Enterprises' Performance:
2010	11.6	8.7	5.2	3.5	2.9	6.4
2011	10.7	8.1	5.1	3.1	2.6	5.7
2012	12.6	9.3	5.4	3.9	3.3	7.2
2013	13.4	10.2	6.2	4.0	3.2	7.2
2014	14.7	11.6	7.0	4.5	3.2	7.7
2015	15.1	12.1	7.4	4.6	3.0	7.7
Source: FHFA's tabulation of Home Mortgage Disclosure Act (HMDA) and Enterprises' data. Conventional conforming single-family owner-occupied 1st lien non-HOEPA originations.

The forecast for this subgoal is obtained by generating separate forecasts for the two sub-populations (the low-income areas component and the high-minority income component). For this proposed rulemaking, FHFA has tested alternate model specifications for this subgoal and determined that aligning the overlapping portion with the low-income area component yields forecast estimates that are more precise (in terms of a narrower confidence interval).^[19]

FHFA sought to understand how the markets in low-income areas and high minority census tracts have evolved in recent years and who was being served by the Enterprises' efforts in these areas. FHFA's analysis found that the mortgage market in both low-income areas and in high-minority census tracts has been moving towards borrowers with higher incomes in recent years. As noted in Table 4, HMDA data show that both the low-income areas and the high-minority areas have increasing shares of borrowers with incomes at or above 100 percent of AMI, although loans to borrowers with incomes over 100 percent of AMI do not qualify for the minority areas component of the goal. For instance, the share of loans made to borrowers with incomes less than 50 percent of AMI and residing in low-income areas decreased from 17.8 percent in 2010 to 14.1 percent in 2015, after peaking at 19 percent in 2012. Over the same period, the share of loans made to borrowers with incomes greater than 100 percent of AMI and residing in these low-income census tracts increased from 38.8 percent in 2010 to 42.1 percent in 2015, after dipping to 36.5 percent in 2012. Thus, borrowers with higher incomes have made up an increasing share of the mortgage market in the low-income areas. A similar trend exists among borrowers residing in high minority census tracts. While borrowers with incomes greater than 100 percent of AMI represented 42.5 percent of borrowers in these census tracts in 2010, the share increased to 49.2 percent in 2015.

Table 4—Borrower Income Relative to AMI for Low-Income Areas Subgoal

[HMDA]

	2010 (%)	2011 (%)	2012 (%)	2013 (%)	2014 (%)	2015 (%)
Borrowers Residing in Low-Income Census Tracts:
Borrower Income ≤50% AMI	17.8	17.7	19.0	15.4	14.1	14.1
Borrower Income >50% and ≤60% AMI	9.6	9.0	10.5	9.8	9.3	9.3
Borrower Income >60% and ≤80% AMI	18.4	17.6	18.8	18.6	18.6	18.6
Borrower Income >80% and ≤100% AMI	14.3	13.9	13.9	14.7	14.9	14.9
Borrower Income >100% and ≤120% AMI	10.1	10.0	10.0	10.8	11.3	11.3
Borrower Income >120% AMI	28.7	30.5	26.5	29.3	30.9	30.8
Income Missing	1.0	1.4	1.3	1.3	0.9	1.0
Total	100.0	100.0	100.0	100.0	100.0	100.0
Borrowers Residing in High-Minority Census Tracts:
Borrower Income ≤50% AMI	14.9	15.0	14.6	11.3	10.1	10.3
Borrower Income >50% and ≤60% AMI	9.0	8.7	9.1	8.1	7.6	7.6
Borrower Income >60% and ≤80% AMI	18.0	17.7	17.7	16.9	16.8	17.0
Borrower Income >80% and ≤100% AMI	14.6	14.3	14.1	14.7	14.8	14.9
Borrower Income >100% and ≤120% AMI	10.9	10.6	11.0	11.7	12.0	12.2
Borrower Income >120% AMI	31.6	32.4	32.3	36.0	37.8	37.0
Income Missing	1.0	1.3	1.3	1.4	0.9	1.0
Total	100.0	100.0	100.0	100.0	100.0	100.0
Definitions:
Low-income census tracts = Census tracts with median income ≤80% Area Median Income (AMI).
High-minority census tracts = Census tracts where (i) tract median income ≤100% Area Median Income (AMI); and (ii) minorities comprise at least 30 percent of the tract population.
Source: FHFA's tabulation of HMDA data.

The presence of higher income borrowers in lower income and higher 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, but there is nevertheless some concern that such a trend could displace lower income households in these areas. Change in the mix of renters to owner-occupied households often precedes and accompanies these trends. 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, and FHFA is also aware of concerns about the impact of rising housing costs on existing households in lower-income or higher-minority areas. FHFA welcomes input on all aspects of the low-income areas goal and subgoal, and in particular how best to satisfy the policy objectives of the various components of the goal and subgoal.

Table 5 shows similar trends in Enterprise acquisitions of mortgages in low-income areas and high-minority areas. In 2015, 42.5 percent of Enterprise acquisitions were of loans made to borrowers with incomes greater than or equal to 100 percent of the AMI, up from 40.7 percent in 2010. Also in 2015, 48.3 percent of Enterprise acquisitions in high-minority census tracts were acquisitions of loans made to borrowers with incomes greater than or equal to 100 percent of AMI, up from 45.4 percent in 2010.

Table 5—Borrower Income Relative to AMI for Low-Income Areas Subgoal

[Enterprise Loans Only]

	2010 (%)	2011 (%)	2012 (%)	2013 (%)	2014 (%)	2015 (%)
Borrowers Residing in Low-Income Census Tracts:
Borrower Income ≤50% AMI	16.7	16.3	18.2	14.5	13.4	13.4
Borrower Income >50% and ≤60% AMI	9.2	8.8	10.0	9.6	9.4	9.4
Borrower Income >60% and ≤80% AMI	18.4	17.5	18.6	18.6	19.0	19.1
Borrower Income >80% and ≤100% AMI	14.8	14.4	14.6	15.3	15.5	15.6
Borrower Income >100% and ≤120% AMI	10.8	10.9	10.8	11.5	11.7	11.8
Borrower Income >120% AMI	29.9	32.0	27.7	30.5	31.0	30.7
Income Missing	0.2	0.0	0.0	0.0	0.0	0.0
Total	100.0	100.0	100.0	100.0	100.0	100.0
Borrowers Residing in High-Minority Census Tracts:
Borrower Income ≤50% AMI	13.3	12.9	15.2	11.5	10.3	10.3
Borrower Income >50% and ≤60% AMI	8.4	8.0	9.0	8.3	8.0	7.9
Borrower Income >60% and ≤80% AMI	17.7	16.9	18.0	17.7	17.7	17.7
Borrower Income >80% and ≤100% AMI	15.1	14.7	14.9	15.5	15.7	15.9
Borrower Income >100% and ≤120% AMI	11.6	11.4	11.5	12.4	12.6	12.8
Borrower Income >120% AMI	33.8	36.2	31.3	34.6	35.7	35.5
Income Missing	0.2	0.1	0.0	0.0	0.0	0.0
Total	100.0	100.0	100.0	100.0	100.0	100.0
Definitions:
Low-income census tracts = Census tracts with median income ≤80% Area Median Income (AMI).
High-minority census tracts = Census tracts where (i) tract median income ≤100% Area Median Income (AMI); and (ii) minorities comprise at least 30 percent of the tract population.
Source: FHFA's tabulation of Enterprises' data.

Proposed rule. The proposed rule would raise the annual low-income areas home purchase subgoal benchmark level for 2018 through 2020 to 15 percent from the 14 percent level set for the current goal period (2015-2017).

Table 6—Low-Income Areas Home Purchase Subgoal

Year	Historical performance				Projected performance
	2013	2014	2015	2016	2017	2018	2019	2020
Actual Market	14.2%	15.2%	15.2%
Benchmark	11%	11%	14%	14%	14%
Current Market Forecast				14.7%	15.6%	15.8%	16.1%	15.7%
				+/−1.2%	+/−2.0%	+/−2.6%	+/−3.1%	+/−3.5%
Fannie Mae Performance:
Low-Income Area Home Purchase Mortgages	86,430	91,691	99,723	n/a
High-Minority Area Home Purchase Mortgages	27,425	25,650	25,349	n/a
Subgoal-Qualifying Total Home Purchase Mortgages	113,855	117,341	125,072	156,441
Total Home Purchase Mortgages	814,137	757,870	802,432	964,847
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	n/a
High-Minority Area Home Purchase Mortgages	12,177	14,808	16,601	n/a
Subgoal-Qualifying Total Home Purchase Mortgages	52,621	70,795	83,773	100,608
Total Home Purchase Mortgages	429,158	519,731	579,340	644,991
Low-Income Area % of Home Purchase Mortgages	12.3%	13.6%	14.5%	15.6%

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 in 2014 and 2015, although its performance was lower than that of the market in 2013. From 2013 through 2015, Freddie Mac's performance exceeded the benchmark but was below the market level. FHFA's forecast indicates that the market will increase slightly in the coming years, reaching a maximum level of 16.1 in 2019.

FHFA is proposing only a modest increase in the benchmark level that reflects the recent performance levels of the Enterprises while FHFA continues to evaluate whether the measure meets policy objectives. FHFA, as regulator and as conservator, will continue to monitor the Enterprises' performance, and if FHFA determines in later years that the benchmark level for the low-income areas home purchase housing 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 proposed 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.^[20] 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 proposed rule would set the annual low-income areas home purchase goal benchmark level for 2018 through 2020 at the subgoal benchmark level of 15 percent plus a disaster areas increment that FHFA will set separately each year.

Table 7—Low-Income Areas Home Purchase Goal

Year	Historical performance
	2010	2011	2012	2013	2014	2015	2016
Actual Market	24.0%	22.0%	23.2%	22.1%	22.1%	19.8%	n/a
Benchmark	24%	24%	20%	21%	18%	19%	17%
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 proposed rule would set the annual low-income refinancing housing goal benchmark level for 2018 through 2020 at 21 percent. While this proposed benchmark level is unchanged from the current 2015 to 2017 benchmark level, FHFA believes that this level 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 impacts 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 8—Low-Income Refinancing Goal

Year	Historical performance				Projected performance
	2013	2014	2015	2016	2017	2018	2019	2020
Actual Market	24.3%	25.0%	22.5%
Benchmark	20%	20%	21%	21%	21%
Current Market Forecast				21.1%	23.4%	24.3%	25.5%	24.8%
				+/−2.9%	+/−4.9%	+/−6.2%	+/−7.3%	+/−8.3%
Fannie Mae Performance:
Low-Income Refinance Mortgages	519,753	215,826	227,817	247,663
Total Refinance Mortgages	2,170,063	831,218	1,038,663	1,268,648
Low-Income % of Refinance Mortgages	24.0%	26.0%	21.9%	19.5%
Low-Income HAMP Modification Mortgages	11,858	6,503	3,563	n/a
Total HAMP Modification Mortgages	16,478	9,288	6,595	n/a
Low-Income % of HAMP Modification Mortgages	72.0%	70.0%	54.0%	n/a
Low-Income Refinance & HAMP Modification Mortgages	531,611	222,329	231,380	n/a
Total Refinance & HAMP Modification Mortgages	2,186,541	840,506	1,045,258	n/a
Low-Income % of Refinance & HAMP Modification Mortgages	24.3%	26.5%	22.1%	n/a
Freddie Mac Performance:
Low-Income Refinance Mortgages	306,205	131,921	179,530	174,664
Total Refinance Mortgages	1,309,435	514,936	795,936	830,824
Low-Income % of Refinance Mortgages	23.4%	25.6%	22.6%	21.0%
Low-Income HAMP Modification Mortgages	14,757	6,795	3,064	n/a
Total HAMP Modification Mortgages	21,599	10,335	4,433	n/a
Low-Income % of HAMP Modification Mortgages	68.3%	65.7%	69.1%	n/a
Low-Income Refinance & HAMP Modification Mortgages	320,962	138,716	182,594	n/a
Total Refinance & HAMP Modification Mortgages	1,331,034	525,271	800,369	n/a
Low-Income % of Refinance & HAMP Modification Mortgages	24.1%	26.4%	22.8%	n/a

Both Enterprises have met this goal since 2013. The performance of the Enterprises on this goal has historically been very close to actual market levels. In 2014, when the market figure was at its highest point, both Enterprises met the goal and exceeded the market. In 2015, Freddie Mac exceeded the market and the benchmark level, and Fannie Mae exceeded the benchmark level.

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 25.0 percent in 2014. FHFA's model for this goal forecasts that this market will decrease in 2016, with a sharp rise in 2017-2019, followed by slight moderation in 2020. However, the confidence intervals around the forecasts are very wide, reflecting the uncertainty about interest rates. Recent macroeconomic forecasts have predicted interest rate hikes that have not materialized.

Since 2010 the low-income refinancing housing goal has included modifications under the Home Affordable Modification Program (HAMP).^[21] 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.^[22] The expiration of the HAMP program may make it slightly more difficult for the Enterprises to meet the low-income refinancing goal.

FHFA, as regulator and conservator, will continue to monitor the Enterprises and if FHFA determines in later years that the benchmark level for the low-income refinancing housing goal needs to be revised, FHFA may take appropriate steps to adjust the benchmark level.

V. Multifamily Housing Goals

This proposed rule also sets out FHFA's views about benchmark levels for the multifamily housing goals from 2018-2020. FHFA has considered the required statutory factors described below. Despite the strength of the multifamily mortgage market, data indicates a continued supply gap of units affordable to lower-income households. However, 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 Proposed Multifamily Housing Goal Levels

In setting the proposed benchmark levels for the multifamily housing goals, FHFA has 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.^[23]

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. Consequently, it can be difficult to correlate different datasets that usually rely on different reporting formats. For example, some data are available by dollar volume while other data are available by unit production.^[24]

Another difference between the single-family and multifamily goals is that there are separate single-family housing goals for home purchase and refinancing 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 volumes, 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 a unit's rent and utility expenses not exceeding 30 percent of the area median income standard for low- and very low-income families.^[25] While much of the analysis that follows discusses trends in the overall multifamily mortgage market, FHFA recognizes that these general 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 addresses 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 2015, the multifamily mortgage origination market experienced remarkable growth—year-over-year origination volume grew 28 percent over the prior year to nearly $250 billion, fueled largely by a recovery in multifamily construction. The overall market grew modestly in 2016. Forecasts from various industry experts indicate that overall multifamily growth in 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 remain level in 2018.

According to the National Multifamily Housing Council's tabulation of American Community Survey microdata, in 2015 about 43 percent of renter households (18.7 million households) lived in multifamily properties, defined as structures with five or more rental units.^[26] More generally, the population of renters continued to grow from 35 million in 2005 to 44 million in 2015, an increase of about one quarter.^[27] 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 high-end segment of many markets.^[28] As a result of these factors, rents continued to rise nationally and outpaced inflation in 2016.^[29]

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 will depend on the availability of affordable housing subsidies.

Using the measure under which affordable rent and utilities do not exceed 30 percent of AMI, affordability for families living in rental units has decreased for many households in recent years. The Joint Center for Housing Studies (JCHS) 2016 State of the Nation's Housing Report notes some concerning trends in the supply of affordable multifamily units. For example, the report found that the majority of growth in the multifamily housing stock has been the result of new construction. Moreover, most of the new construction consists of apartments with fewer bedrooms and has been concentrated in urban areas with higher median rents. In the same report, JCHS also noted, “the steep rent for new units reflect rising land and development costs, which push multifamily construction to the high end of the market.” ^[30]

JCHS has also noted the significant prevalence of cost-burdened renters. In 2015, nearly half of all tenants paid more than 30 percent of household income for rental housing, especially in high-cost urban markets where most renters reside and where Fannie Mae and Freddie Mac have focused their multifamily lending. Among lower-income households, cost burdens are especially severe.^[31] 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.^[32]

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.^[33]

The combination of the supply gap in affordable units which 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.^[34] 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.^[35] 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 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 historic 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.^[36] In recent years, competition for affordable housing subsidy 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 much more uncertain, as policymakers consider a broader range of potential tax reform legislation that could adversely impact the LIHTC program.

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 8.19 million, an increase of one-third since 2005.^[37]

Role of the Enterprises. In setting the proposed multifamily housing goals, FHFA has 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 of the multifamily origination market was approximately 70 percent of the market in 2008 and 2009 compared to 38 percent in 2015.^[38] The total share is expected to remain at around the 2015 level in 2016, 2017, and 2018 in light of the Scorecard cap imposed by FHFA in its role as conservator (discussed below).

Despite the Enterprises' reduced market share in the overall multifamily market, 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 lending cap is intended to further FHFA's conservatorship goal: 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 reflect what is generally considered by the industry as 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.^[39]

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 was initially set at $30 billion, raised in May 2016 to $35 billion, and further increased to $36.5 billion in August, in response to growth of the overall multifamily origination market throughout the year. These increases maintained the Enterprises' current market share at about 40 percent. FHFA has announced that for 2017, the cap will remain at $36.5 billion.

FHFA reviews the market size estimates quarterly, using current market data provided by Fannie Mae, Freddie Mac, the MBA, and the National Multifamily Housing Council. If FHFA determines that the actual market size is greater than was projected, the agency will consider an approximate increase to the capped (conventional market-rate) category of the Conservatorship Scorecard for each Enterprise. In light of the need for market participants 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.

In order to encourage affordable lending activities, FHFA excludes many types of loans in underserved markets from the Conservatorship Scorecard cap on conventional 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 2017 Conservatorship Scorecard excludes either the entirety of the loan amount or a pro rata share of the loan on the following categories: (1) Targeted affordable housing; (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 AMI), high cost (80 percent AMI), and very high cost (100 percent AMI) markets. By excluding the underserved market categories from the cap, the Conservatorship Scorecard continues to encourage the Enterprises to support affordable housing in their purchases of multifamily mortgages.^[40]

B. Proposed Multifamily Housing Goal Benchmark Levels

In setting the proposed multifamily housing goals, FHFA encourages the Enterprises to provide liquidity and to support various multifamily finance market segments while doing so 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 multifamily housing goals to ensure the Enterprises' continued safety and soundness.

The proposed rule establishes benchmark levels for the multifamily housing goals for the Enterprises. Before finalizing the benchmark levels for the low-income and very low-income multifamily goals in the final rule, FHFA will review any additional data that become available about the multifamily performance of the Enterprises in 2016, updated projections of the size of the multifamily market and affordable market share, and any public comments received on the proposed multifamily housing goals.

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.

Table 9—Multifamily Low-Income Housing Goal

Year	Historical performance
	2012	2013	2014	2015	2016
Fannie Mae Goal	285,000	265,000	250,000	300,000	300,000
Freddie Mac Goal	225,000	215,000	200,000	300,000	300,000
Fannie Mae Performance:
Low-Income Multifamily Units	375,924	326,597	260,124	307,510	351,235
Total Multifamily Units	501,256	430,751	372,089	468,798	551,666
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	379,043	407,340
Total Multifamily Units	377,522	341,921	366,377	514,275	597,033
Low-Income % of Total Units	79.1%	74.5%	74.7%	73.7%	68.2%

From 2012 through 2016, both Enterprises exceeded their low-income multifamily goals. Prior to 2015, Fannie Mae had higher goals than Freddie Mac. For the 2015-2017 goal period, FHFA set the same goal level 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 351,235 low-income units, and Freddie Mac purchased mortgages financing 407,340 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.

As noted above, the forecast for the multifamily originations market increases slightly and then levels off after 2017. The Conservatorship Scorecard cap for each Enterprise was raised from an initial $30 billion cap to $36.5 billion in August 2016 in response to growth of the multifamily origination market throughout the year. This change allowed the Enterprises to pursue purchases of a greater volume of multifamily originations and support the overall market and may seem to support an increase in the proposed goal levels for both Enterprises. However, 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. Moreover, the forecast for the multifamily originations market for 2017 and 2018 is relatively flat, and securing housing subsidies will likely continue to be challenging. These trends suggest moderation in any increase in the proposed goal levels. Therefore, balancing these considerations, the proposed rule sets the annual 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 includes units affordable to very low-income families, defined as families with incomes no greater than 50 percent of AMI.

Table 10—Multifamily Very Low-Income Subgoal

Year	Historical performance
	2012	2013	2014	2015	2016
Fannie Mae Goal	80,000	70,000	60,000	60,000	60,000
Freddie Mac Goal	59,000	50,000	40,000	60,000	60,000
Fannie Mae Performance:
Very Low-Income Multifamily Units	108,878	78,071	60,542	69,078	65,445
Total Multifamily Units	501,256	430,751	372,089	468,798	551,666
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	73,032
Total Home Purchase Mortgages	377,522	341,921	366,377	514,275	597,033
Very Low-Income % of Total Units	15.9%	16.6%	13.3%	15.0%	12.2%

From 2012 through 2016, both Enterprises met and exceeded their very low-income multifamily goals. In 2016, the goal for each Enterprise was 60,000 units. Fannie Mae purchased mortgages financing 65,445 very low-income units, while Freddie Mac purchased mortgages financing 73,032 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.

The market for very low-income multifamily housing faces even larger challenges than the market for low-income multifamily housing, given the need for lower rents—often requiring deeper subsidies—to make units affordable to these households. These factors suggest moderation in the setting of the very low-income multifamily subgoal for the Enterprises. Therefore, the proposed rule maintains the annual very low-income multifamily subgoal for each Enterprise at 60,000 units each year from 2018 through 2020.

3. Small Multifamily Low-Income Housing Subgoal

A small multifamily property is defined 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.

Table 11—Small Multifamily Low-Income Subgoal

Year	Historical performance
	2012	2013	2014	2015	2016
Small Low-Income Multifamily Goal				6,000	8,000
Fannie Mae Performance:
Small Low-Income Multifamily Units	16,801	13,827	6,732	6,731	9,310
Total Small Multifamily Units	26,479	21,764	11,880	11,198	15,230
Low-Income % of Total Small Multifamily Units	63.5%	63.5%	56.7%	60.1%	61.1%
Freddie Mac Performance:
Small Low-Income Multifamily Units	829	1,128	2,076	12,802	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%

This was a new subgoal created in the 2015-2017 goal period. The goal was set at 6,000 units in 2015, 8,000 units in 2016, and 10,000 units in 2017. In 2016, both Enterprises exceeded the goal of 8,000 units. Fannie Mae purchased mortgages financing 9,310 units, and Freddie Mac purchased mortgages financing 22,101 units.

The proposed rule would set 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, and will consider all input in preparation of the final rule. However, FHFA is proposing to maintain the same benchmark level for 2018 through 2020 as the 2017 benchmark level for both Enterprises. 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.

VI. Section-by-Section Analysis of Other Proposed Changes

The proposed rule would also revise other provisions of the housing goals regulation, as discussed below.

A. Changes to Definitions—Proposed § 1282.1

The proposed rule includes changes to definitions used in the current housing goals regulation. The proposed rule would revise the definitions of “median income,” “metropolitan area,” and “non-metropolitan area” and would remove the definition of “AHS.”

1. Definition of “Median Income”

The current regulation defines “median income” as the unadjusted median family income 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 proposed rule would revise 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 proposed rule would also make two additional technical changes to the definition of “median income.” First, the proposed rule would add 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 proposed rule would remove the word “family” in one place so that the term “median income” is used consistently throughout the regulation.

The revised definition would read: “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.”

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

The proposed rule would revise the definitions of “metropolitan area” and “non-metropolitan area” to be consistent with each other and to reflect the proposed changes to the definition of “median income” discussed above.

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 incomes “determined by HUD,” while the definition of “non-metropolitan area” refers to median family incomes “published annually by HUD.”

To be consistent with the proposed changes to the definition of “median income,” the proposed rule would revise 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 proposed rule would revise 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.”

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

The proposed rule would remove 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 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 proposed rule would remove the definition from § 1282.1.

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

The proposed rule would revise § 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) 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 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 proposed rule would revise § 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 would not refer specifically to data derived from the ACS. Section 1282.15(e)(2) would be revised to replace the phrase “as determined by FHFA based on the most recent decennial census” with the phrase “as determined by FHFA.”

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

The proposed rule would revise § 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 to establish the Enterprise housing goals under the Safety and Soundness Act.^[41]

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 proposed rule would remove § 1282.15(g)(2) from the regulation.

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

The proposed rule would revise § 1282.21(b)(3) to provide the Director with 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.

Section 1336 of the Safety and Soundness Act provides for the enforcement of the Enterprise housing goals. If FHFA determines that an Enterprise has failed to meet a housing goal and that achievement of the goal was feasible, FHFA may require the Enterprise to submit a housing plan describing the actions it will take “to achieve the goal for the next calendar year.” ^[42] The Safety and Soundness Act has similar provisions for requiring a housing plan if FHFA determines, during the year in question, that there is a substantial probability that an Enterprise will fail to meet a housing goal and that achievement of the goal is feasible. In such cases, the housing plan would describe the actions the Enterprise will take “to make such improvements and changes in its operations as are reasonable in the remainder of such year.” The current regulation generally mirrors the statutory language on the requirements for a housing plan, except that the regulation makes clear that the housing plan must also “[a]ddress any additional matters relevant to the plan as required, in writing, by the Director.” ^[43]

FHFA required an Enterprise to submit a housing plan for the first time in late 2015 in response to Freddie Mac's failure to achieve the single-family low-income and very low-income home purchase goals in 2014. FHFA required Freddie Mac to submit a housing plan setting out the steps Freddie Mac would take in 2016 and 2017 to achieve the two goals that it failed to achieve in 2013 and 2014. The requirement for the plan to address actions taken in both 2016 and 2017 was based on FHFA's authority under § 1282.21(b) to require a housing plan to address any additional matters required by the Director and was intended to address an issue of timing.

FHFA's final determination on Freddie Mac's performance on the housing goals for 2014 was issued on December 17, 2015. As described in more detail below, that timing was driven by procedural steps required by the Safety and Soundness Act and FHFA's own regulation. If FHFA interpreted narrowly the statutory and regulatory provisions stating that the housing plan should address the steps the Enterprise would take in the following year, the housing plan itself would become irrelevant because the year it would cover would have ended before the housing plan was even submitted to FHFA.

The extended time required to reach a final determination housing goals performance will occur every year as a result of the procedural steps required by the Safety and Soundness Act. Under those procedures, if FHFA determines that an Enterprise has failed to achieve a housing goal in a particular year, FHFA is first required to issue a preliminary determination that generally provides at least 30 days for the Enterprise to respond. FHFA must then consider any information submitted by the Enterprise before making a final determination on whether the Enterprise failed to meet the goal and whether achievement of the goal was feasible. If FHFA determines that the Enterprise should be required to submit a housing plan, the statute provides for up to 45 days for the Enterprise to submit its housing plan.^[44] FHFA must then evaluate the housing plan, generally within 30 days. The time necessary for FHFA's review and determination at each step of this procedural process is generally four to six months.

These procedural steps cannot begin until FHFA has the information necessary to make a determination on whether the Enterprise has met the housing goals. The Enterprises are required to submit their official performance numbers to FHFA within 75 days after the end of the year, usually March 15 of the following year. Therefore, the earliest that FHFA would be able to approve a housing plan from an Enterprise would be mid-July of the year following the performance year. For the single-family housing goals, this time period is extended even further because the HMDA data necessary to determine if an Enterprise met the retrospective market measurement portion of the single-family housing goals are not available until September of the year following the performance year.

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, (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, FHFA is proposing to amend § 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 performance of the Enterprise for a particular year.

VII. Paperwork Reduction Act

The proposed rule would 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 the proposed rule under the Regulatory Flexibility Act. The General Counsel of FHFA certifies that the proposed rule, if adopted as a final rule, is not likely to have a significant economic impact on a substantial number of small entities because the regulation 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 proposes to amend part 1282 of Title 12 of the Code of Federal Regulations as follows:

CHAPTER XII—FEDERAL HOUSING FINANCE AGENCY

Subchapter E—Housing Goals and Mission

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.

§ 1282.1

[Amended]

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 § 1282.12 to read as follows:

§ 1282.12

Single-family housing goals.

(a) Single-family housing goals. An Enterprise shall be in compliance with a single-family housing goal if its performance under the housing goal meets or exceeds either:

(1) The share of the market that qualifies for the goal; or

(2) The benchmark level for the goal.

(b) Size of market. The size of the market for each goal shall be established annually by FHFA based on data reported pursuant to the Home Mortgage Disclosure Act for a given year. Unless otherwise adjusted by FHFA, the size of the market shall be determined based on the following criteria:

(1) Only owner-occupied, conventional loans shall be considered;

(2) Purchase money mortgages and refinancing mortgages shall only be counted for the applicable goal or goals;

(3) All mortgages flagged as HOEPA loans or subordinate lien loans shall be excluded;

(4) All mortgages with original principal balances above the conforming loan limits for single unit properties for the year being evaluated (rounded to the nearest $1,000) shall be excluded;

(5) All mortgages with rate spreads of 150 basis points or more above the applicable average prime offer rate as reported in the Home Mortgage Disclosure Act data shall be excluded; and

(6) All mortgages that are missing information necessary to determine appropriate counting under the housing goals shall be excluded.

(c) Low-income families housing goal. The percentage share of each Enterprise's total purchases of purchase money mortgages on owner-occupied single-family housing that consists of mortgages for low-income families shall meet or exceed either:

(1) The share of such mortgages in the market as defined in paragraph (b) of this section in each year; or

(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) Very low-income families housing goal. The percentage share of each Enterprise's total purchases of purchase money mortgages on owner-occupied single-family housing that consists of mortgages for very low-income families shall meet or exceed either:

(1) The share of such mortgages in the market as defined in paragraph (b) of this section in each year; or

(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.

(e) Low-income areas housing goal. The percentage share of each Enterprise's total purchases of purchase money mortgages on owner-occupied single-family housing that consists of mortgages for families in low-income areas shall meet or exceed either:

(1) The share of such mortgages in the market as defined in paragraph (b) of this section in each year; or

(2) A benchmark level which shall be set annually by FHFA notice based on the benchmark level for the low-income areas housing subgoal, plus an adjustment factor reflecting the additional incremental share of mortgages for moderate-income families in designated disaster areas in the most recent year for which such data is available.

(f) Low-income areas housing subgoal. The percentage share of each Enterprise's total purchases of purchase money mortgages on owner-occupied single-family housing that consists of mortgages for families in low-income census tracts or for moderate-income families in minority census tracts shall meet or exceed either:

(1) The share of such mortgages in the market as defined in paragraph (b) of this section in each year; or

(2) The benchmark level, which for 2018, 2019 and 2020 shall be 15 percent of the total number of purchase money mortgages purchased by that Enterprise in each year that finance owner-occupied single-family properties.

(g) Refinancing housing goal. The percentage share of each Enterprise's total purchases of refinancing mortgages on owner-occupied single-family housing that consists of refinancing mortgages for low-income families shall meet or exceed either:

(1) The share of such mortgages in the market as defined in paragraph (b) of this section in each year; or

(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: June 28, 2017.

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

8.  12 CFR 1282.14(d).

9.  12 CFR 1282.21(a).

10.  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.

11.  12 U.S.C. 4562(e)(2).

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

13.  The macroeconomic outlook described here is based on Moody's and other forecasts as of September 2016.

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

15.  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 on Housing Studies, June 2017.

16.  See Income and Poverty in the United States: 2015, United States Census Bureau, September 2016 https://www.census.gov/​content/​dam/​Census/​library/​publications/​2016/​demo/​p60-256.pdf.

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

18.  Bhutta, Neil and Ringo, Daniel (2016). “Changing FHA Mortgage Insurance Premiums and the Effects on Lending,” FEDS Notes. Washington: Board of Governors of the Federal Reserve System, September 29, 2016, http://dx.doi.org/​10.17016/​2380-7172.1843.

19.  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.

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

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

22.  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.

23.  12 U.S.C. 4563(a)(4).

24.  CFPB is planning to collect and release additional data fields (including the number of units for each multifamily loan that is reported) beginning in 2018 that likely will be useful in creating a retrospective market measure for the multifamily market.

25.  12 U.S.C. 4563(c).

26.  Accessed on 9/22/2016 at http://www.nmhc.org/​Content.aspx?​id=​4708#Type_​of_​Structure.

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

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

29.  Id.

30.  “The State of the Nation's Housing 2016,” Joint Center for Housing Studies of Harvard University, June 2016, available at http://www.jchs.harvard.edu/​sites/​jchs.harvard.edu/​files/​jchs_​2016_​state_​of_​the_​nations_​housing_​lowres.pdf.

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

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

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

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

35.  “2012 Rental Housing Finance Survey,” U.S. Census Bureau and U.S. Department of Housing and Urban Development, Tables 2b, 2c, 2d and 3.

36.  LIHTC is a supply-side subsidy created under the Tax Reform Act of 1986 and is the main source of new affordable 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 households. LIHTC has facilitated the creation or rehabilitation of approximately 2.4 million affordable units since inception in 1986.

37.  “Preview of 2015 Worst Case Housing Needs,” U.S. Department of Housing and Urban Development, January 2017. Renters with worse case needs have very low incomes, lack housing assistance, and have either severe rent burdens or severely inadequate housing (or both).

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

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

40.  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.

41.  See 60 FR 61846 (Dec. 1, 1995).

42.  See 12 U.S.C. 4566(c)(2).

43.  See 12 CFR 1282.21(b).

44.  See 12 U.S.C. 4566(c)(3).