2015-20880. 2015-2017 Enterprise Housing Goals

AGENCY:

Federal Housing Finance Agency.

ACTION:

Final rule.

SUMMARY:

The Federal Housing Finance Agency (FHFA) is issuing a final rule regarding the housing goals for Fannie Mae and Freddie Mac (the Enterprises) for 2015 through 2017. The Federal Housing Enterprises Financial Safety and Soundness Act of 1992, as amended (the Safety and Soundness Act), requires FHFA to establish annual housing goals for mortgages purchased by the Enterprises. The housing goals include separate categories for single-family and multifamily mortgages on housing that is affordable to low-income and very low-income families, among other categories.

The final rule establishes the benchmark levels for each of the housing goals and subgoals for 2015 through 2017. In addition, the final rule establishes a new housing subgoal for small multifamily properties affordable to low-income families.

The final rule also adds or revises a number of other provisions in the housing goals regulation in order to provide greater clarity about the mortgages that will qualify for the goals or subgoals. In addition, the final rule makes a number of clarifying and conforming changes, including revisions to the definitions of “rent” and “utilities” and to the rules for determining affordability of both single-family and multifamily units. The final rule also establishes more transparent agency procedures for FHFA guidance on the housing goals.

FHFA also discusses here its plans to require more detailed Enterprise reporting to FHFA on the Enterprises' purchases of mortgages on single-family rental housing.

DATES:

The final rule is effective on October 5, 2015.

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 20024. The telephone number for the Telecommunications Device for the Hearing Impaired is [phone].

SUPPLEMENTARY INFORMATION:

I. Description of the Enterprise Affordable Housing Goals

The Safety and Soundness Act requires FHFA to establish several annual housing goals for single-family and multifamily mortgages purchased by Fannie Mae and Freddie Mac.^[1] The housing goals provisions were substantially revised in 2008 with the enactment of the Housing and Economic Recovery Act.^[2] Under the revised structure, FHFA established housing goals for the Enterprises for 2010 and 2011 in a final rule published on September 14, 2010.^[3] FHFA established new housing goals benchmark levels for the Enterprises for 2012 through 2014 in a final rule published on November 13, 2012.^[4] The housing goals established by FHFA in these two prior rulemakings include four goals and one subgoal for single-family owner-occupied housing and one goal and one subgoal for multifamily housing.

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 acquired by an Enterprise each year that qualifies 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 a total of one to four units, at least one of which must be owner-occupied. The single-family goals cover “conventional, conforming mortgages,” with “conventional” meaning not insured or guaranteed by the Federal Housing Administration (FHA) or other government agency, and “conforming” meaning those mortgages with a principal balance that does not exceed the loan limits for Enterprise mortgages.

The single-family goals established by FHFA in 2010 and 2012 compare the goal-qualifying share of an Enterprise's mortgage purchases to two separate measures: A “benchmark level” and a “market level.” The benchmark level is set prospectively by rulemaking, based on various factors, including FHFA's forecast of the goal-qualifying share of the overall conventional conforming mortgage market using FHFA's market estimation models. The market level is determined retrospectively each year based on the actual goal-qualifying share of the overall conventional conforming mortgage market as measured by FHFA based on Home Mortgage Disclosure Act (HMDA) data for that year. The overall mortgage market that FHFA uses for purposes of both the prospective market forecasts and the retrospective market measurement consists of all conventional conforming mortgages on single-family, owner-occupied properties 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 or sold to another mortgage conduit, some of which may be securitized into a private label security (PLS), although very few such securities have been issued for conventional conforming mortgages since 2008.

Under this two-part approach, determining whether an Enterprise has met the single-family goals and subgoals for a specific year requires looking at both the benchmark level and the market level for each goal and subgoal. In order to meet a single-family housing goal or subgoal during 2012-2014, the actual percentage of mortgage purchases by an Enterprise that met each goal or subgoal had to meet or exceed either the benchmark level or the market level for that goal or subgoal for that year.

Multifamily goals. The multifamily goals defined under the Safety and Soundness Act include separate categories for mortgages on multifamily properties (i.e., properties with five or more units) with rental units affordable to low-income families and very low-income families. The multifamily goals established by FHFA in 2010 and 2012 are based on numeric targets, not percentages of mortgage purchases, 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 because of the lack of comprehensive data about the multifamily mortgage market such as that provided by HMDA for single-family mortgages. As a result, FHFA measures Enterprise multifamily goals performance only against the benchmark levels, which are set prospectively by rulemaking based on various statutorily-prescribed factors, including FHFA's forecast of the goal-qualifying share of the overall conventional multifamily mortgage market.

II. Proposed Rule and Comments

FHFA published a proposed rule in the Federal Register on September 11, 2014 regarding the establishment of affordable housing goals for Fannie Mae and Freddie Mac for 2015-2017.^[5] The proposed rule would have established benchmark levels for each of the single-family and multifamily housing goals. The proposed rule also would have established a new multifamily housing subgoal for small multifamily properties with units that are affordable to low-income families and would have revised the rules for determining whether some types of transactions could be counted for purposes of the housing goals.

In addition, the proposed rule requested comment on three options for determining compliance with the single-family housing goals. Specifically, the proposed rule requested comment on whether the current two-part approach should be maintained (alternative #1), whether housing goals performance should be measured against a prospective benchmark level only (alternative #2), or whether it should be measured against a retrospective market level measure only (alternative #3).

FHFA received 144 comment letters on the proposed rule.^[6] Comments were submitted by policy advocacy groups, many of which have a specific focus on affordable housing; trade associations representing lenders, home builders, realtors, and other mortgage market participants; individuals, including many with personal or professional experience in housing or mortgage finance; members of Congress; a trade association representing government entities; businesses and non-profit organizations with an interest in housing, including mission-oriented housing developers and housing counseling groups; investors and groups representing investors; Fannie Mae; and Freddie Mac. FHFA has reviewed and considered all of the comments. Specific provisions of the proposed rule, and the comments received on those provisions, are discussed below. A significant number of comment letters discussed whether the conservatorships of the Enterprises should be ended or raised other issues unrelated to the housing goals. Those comments are beyond the scope of this rulemaking and are not addressed in the final rule.

III. Summary of the Final Rule

A. Single-Family Housing Goals

The final rule maintains the current two-part approach for determining Enterprise compliance with the single-family housing goals, under which FHFA compares Enterprise performance to both a benchmark level and a market level. The final rule establishes the benchmark levels for the single-family housing goals and subgoal for 2015-2017 as follows:

Goal	Criteria	Benchmark level for 2012-2014	Final Rule benchmark level for 2015-2017
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	23 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	7 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 no greater than 80 percent of area median income; and
	• 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	11 percent	14 percent.
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	20 percent	21 percent.

In addition to the low-income areas subgoal described in the above chart, the Enterprises are subject to a low-income areas home purchase goal, which includes the subgoal and mortgages to families with incomes no greater than area median income that live in counties that have been declared disaster areas within the previous three years. This goal is set at the beginning of each year and can vary from year to year, depending on the pattern of disaster areas. The Enterprises are notified by letter about the level of this goal, and these letters are posted on FHFA's public Web site.^[7]

B. Multifamily Housing Goals

The final rule establishes the benchmark levels for the multifamily goals and subgoals for 2015-2017 as shown below. The low-income multifamily goals are higher than the levels in the proposed rule for Fannie Mae and Freddie Mac, consistent with the larger multifamily finance market size in 2015 and the expanded number of exclusions from the cap on the dollar volume of multifamily financing established by FHFA in the 2015 Scorecard for Fannie Mae, Freddie Mac, and Common Securitization Solutions (2015 Conservatorship Scorecard). The agency announced expanded multifamily exclusions under the 2015 Conservatorship Scorecard cap on May 7, 2015. The expanded exclusions from the cap permit both Enterprises to purchase unlimited amounts of loans on multifamily properties that provide affordable rental units in the categories identified by the exclusions. Most of these units can be credited towards the Enterprises' annual multifamily housing goals benchmark levels. Under the final rule, the multifamily benchmark levels are now the same for both Enterprises.

The very low-income multifamily subgoal benchmark levels in the final rule are the same for Fannie Mae and higher than those in the proposed rule for Freddie Mac, consistent with the equal treatment of the two Enterprises in the 2015 Conservatorship Scorecard.

Consistent with the proposed rule, the final rule establishes for the first time a new subgoal for rental units that are affordable to low-income families, (i.e., families with incomes no greater than 80 percent of area median income) in small (5- to 50-unit) multifamily properties financed by mortgages purchased by an Enterprise.

Goal	Criteria	Goal levels for 2014	Final rule goal levels for 2015	Final rule goal levels for 2016	Final rule goal levels for 2017
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	Fannie Mae: 250,000 units Freddie Mac: 200,000 units	Fannie Mae: 300,000 units Freddie Mac: 300,000 units	Fannie Mae: 300,000 units Freddie Mac: 300,000 units	Fannie Mae: 300,000 units. Freddie Mac: 300,000 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	Fannie Mae: 60,000 units Freddie Mac: 40,000 units	Fannie Mae: 60,000 units Freddie Mac: 60,000 units	Fannie Mae: 60,000 units Freddie Mac: 60,000 units	Fannie Mae: 60,000 units. Freddie Mac: 60,000 units.
Low-Income Subgoal for Small Multifamily Rental Properties	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	None	Fannie Mae: 6,000 units Freddie Mac: 6,000 units	Fannie Mae: 8,000 units Freddie Mac: 8,000 units	Fannie Mae: 10,000 units. Freddie Mac: 10,000 units.

C. Changes to Counting Rules

The final rule makes a number of changes and clarifications to the existing rules concerning whether a particular Enterprise mortgage purchase may be counted toward the single-family and multifamily housing goals. These changes include updating and clarifying definitions and other provisions to reflect current Enterprise lending programs and market practices. The final rule also adds transparency to FHFA guidance on issues that arise under the housing goals by indicating that guidance will be placed on FHFA's public Web site.

IV. Affordability

The annual housing goals help measure 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.” ^[8] The Enterprise Charter Acts state that one of their purposes is to “provide ongoing assistance to the secondary market for residential mortgages (including activities relating to mortgages on housing for low- and moderate-income families involving a reasonable economic return that may be less than the return earned on other activities). . . .. .” ^[9]

FHFA received numerous comments on the proposed rule that emphasized the importance of affordable housing for families, including both options for ownership and rental, whether in single-family homes or multifamily housing. FHFA shares this understanding of the importance of affordable housing, and the approach to setting the levels for each of the housing goals is informed by it. While the housing goals target particular segments of the overall housing market, FHFA recognizes that the Enterprises have an important role to play in supporting liquidity for all parts of the housing market, not just those covered by the housing goals.

For households with credit sufficient to qualify for mortgages, homes remain relatively affordable, despite recent increases in home prices. The interest rate on 30-year fixed rate mortgages—the primary financing option for most homebuyers—was below 4.5 percent for most of 2014 and below 4.0 percent for most of the first six months of 2015. This rate is extraordinarily low by historical standards.^[10]

Increases in home prices have eroded affordability over the last several years, however. While interest rates have remained low, the recovery in home prices has been robust, with U.S. home prices rising by roughly five percent between the fourth quarters of 2013 and 2014. In the preceding four quarters, home price growth was almost eight percent. In some areas, home prices are now at levels that were prevalent prior to the recent housing collapse.^[11]

In addition to rising home prices, other challenges affect affordability. The quality and quantity of jobs in the U.S. economy play key roles in determining affordability and, while labor markets have improved since the onset of this recession, a full recovery remains elusive. Unemployment rates are still elevated in many areas, and the labor force participation rate is relatively low. Importantly, household incomes, which fell during the recession, have exhibited very little real growth since then. Although estimates may vary across data sources, the Census Bureau has determined the annual inflation-adjusted household income growth rate to be below one percent for 2011-2013 (the latest years available). Household income growth is important to affordability because it provides prospective homebuyers confidence that future mortgage payments can be made even as the cost of living rises.^[12]

Another challenge to affordability is the relatively limited resources that many prospective households have available for making down payments on a home purchase. For many households, the extent of household savings is extremely limited. For example, using data from the Federal Reserve's 2013 Survey of Consumer Finances, Harvard's Joint Center for Housing Studies estimated that the median household net worth for households that rented in 2013 was $5,400. For younger renting households—those with household heads under the age of 25 or between the ages of 25 and 34—median household net worth was even lower; the median net worth for renting households headed by individuals under 25 was $2,000, while the median net worth for households headed by 25-34 year-olds was $4,850.^[13] In a November 2014 speech, FHFA Director Watt noted that the problem of low wealth is particularly acute for communities of color. In his speech, he stated that:

“[such communities] . . . generally have significantly lower average household wealth and experienced record loss of wealth during the financial crisis as a result of abusive mortgage products, the economic downturn and other factors . . . . [T]his wealth disparity is likely to have a growing impact on the future housing market since people of color are projected to account for approximately 70 percent of the increase in number of households over the next decade.” ^[14]

For some households—particularly households headed by younger individuals—household debt is an impediment to home buying. Student loan and automobile debt are burdening household budgets, often making it difficult for prospective borrowers to afford to purchase a home. Outstanding balances for these types of non-mortgage debt have been growing in recent years. According to data recently published by the New York Federal Reserve Bank, between the fourth quarters of 2013 and 2014, the amount of automobile loan debt grew by more than ten percent and the amount of student loan debt grew by more than seven percent.^[15]

Increasing rents and nearly stagnant wages, particularly for low- and very low- income renters, have resulted in a significant decline in rental housing affordability over the past three years. A recent Harvard study shows that more than half of all tenants pay more than 30 percent of household income for rental housing, especially in the high-cost urban markets where most renters reside and where much of Fannie Mae and Freddie Mac lending is focused. Tenants in the lower income brackets, such as those at 50 or 80 percent of area median income, pay the highest percentage of income for rental housing. These are the income groups targeted by the very low-income and low-income goals, respectively.^[16]

V. Single-Family Housing Goals

A. Approach for Determining Enterprise Compliance With the Single-Family Housing Goals—§ 1282.12(a)

Since 2010, under the housing goals regulation, FHFA has determined Enterprise compliance with the single-family housing goals using a two-part approach under which FHFA compares each Enterprise's housing goals performance to both: (1) A benchmark level that is set in advance in the housing goals regulation; and (2) the actual market level, as measured retrospectively by FHFA based on HMDA data. An Enterprise is determined to have met the goal if it meets or exceeds either the benchmark level or the actual market level for the goal.

The proposed rule presented three alternatives for determining Enterprise compliance with the single-family housing goals. The first alternative would have maintained the current two-part approach. The second alternative would have measured Enterprise performance by comparing it only to a benchmark level set in advance in the regulation. The third alternative would have measured Enterprise performance by comparing it only to the actual market level, as measured retrospectively based on HMDA data.

After considering the comments on the three alternatives, which are discussed below, FHFA has decided to retain in the final rule the current two-part approach for determining Enterprise compliance with the single-family housing goals. This approach balances the risks of its two component tests. Under a benchmark level only approach, since benchmark levels are based on multi-year mortgage market forecasts, the Enterprises would know their goals in advance, thereby enabling more certainty in their planning for how they will meet the goals each year. FHFA recognizes, however, that the market forecasts could result in setting the levels too high relative to the actual market for the year as the market forecasts include factors such as prior market performance that do not necessarily reflect current or future market conditions. The market forecasts also depend on current forecasts of other economic indicators such as interest rates, economic growth, and unemployment.

The retrospective market measure is based on the actual performance of the market in the year being evaluated. The retrospective market measure helps to address the inherent difficulty of accurately forecasting, years in advance, the housing goals' shares of the overall market for purposes of establishing benchmark levels, and thereby help to ensure that the goals are feasible. The retrospective market measure is much more adaptive than a fixed benchmark level by itself, although the HMDA data used for the retrospective market measure do not become available until September of the following year. However, a retrospective market measure-only approach could make it more difficult for the Enterprises to plan their operations and calibrate their performance in the absence of prospectively set benchmark levels.

Even with the inclusion of retrospective market levels under the two-part approach, if FHFA determines in the future that the benchmark levels need to be adjusted in light of changes in the market, either to ensure the safety and soundness of the Enterprises or for any other reason, FHFA will take steps, including adjusting the benchmark levels, as appropriate.

Comments on Proposed Rule

Comments recommending current two-part approach. Several trade associations, housing advocacy groups, and both Enterprises commented that the current two-part approach for determining Enterprise compliance with the single-family housing goals should be retained in the final rule. The commenters stated that neither the benchmark level nor the market level alone is a perfect tool for measuring compliance with the goals. They stated, however, that together the two measures balance the need for predictable prospective targets which encourage the Enterprises to purchase more affordable loans with the need to ensure that the goals are feasible for the Enterprises.

Fannie Mae supported the current two-part approach, stating that prospective benchmark levels provide forecasted targets against which the Enterprises can calibrate and manage their resources, while relying solely on benchmark levels, which are based on multi-year mortgage market forecasts, risks setting levels that will be out of step with actual market conditions and may raise safety and soundness concerns. Fannie Mae noted that if it becomes apparent that an Enterprise is falling short of the benchmark levels, it may become increasingly inefficient economically for the Enterprise to acquire the last loans needed to achieve the benchmarks. Fannie Mae stated that the “price pay-up” needed to acquire those “last” loans could have the effect of “bidding up” the price to the Enterprises for other loans that would have come to the Enterprises anyway, which would be an inefficient use of Enterprise funds. Fannie Mae stated that the retrospective market measure diminishes the likelihood of such distortions and makes it less likely that additional FHFA regulatory action will be needed to address changing market conditions. Fannie Mae noted the concern raised in the proposed rule that the two-part approach may provide less of an incentive for the Enterprises to achieve the benchmark levels in years when the Enterprises anticipate that market levels will end up lower than the benchmark levels, but stated that the Enterprises will always strive to meet the benchmark levels rather than wager on HMDA data that is not available until months after the rating period closes to meet the market levels instead. Fannie Mae also recognized the concern raised in the proposed rule that the retrospective market measure may be less meaningful in years when the Enterprises purchase a large percentage of the overall mortgage market because it would effectively compare the performance of the Enterprises to their own activities, but noted that steps such as increasing guarantee fees have already been taken to reduce the role of the Enterprises and encourage other financial institutions to re-enter the market. Fannie Mae also noted that the Enterprises compete against each other, even in conservatorship, and neither has a controlling share of the market.

Freddie Mac also recommended that FHFA maintain the current two-part approach, stating that projecting market size and composition in setting the benchmark levels is a challenging task and that a changing economic environment can have a significant effect on the volume and goals-qualifying composition of the mortgage market. Freddie Mac stated that the current two-part approach strikes the right balance in providing the Enterprises with known targets, while recognizing that actual market performance may make meeting such targets infeasible.

Comment recommending modified two-part approach (meet retrospective market level only during downturns). One trade association commenter recommended modifying the current two-part approach by retaining both the benchmark level and retrospective market measure but applying the latter only during unexpected market downturns when the total goal-qualifying market share for the loans differs substantially from the benchmark level. The commenter noted that relying solely on the benchmark standard could spur the Enterprises to increase their support for affordable homeowner lending, but would also leave them vulnerable to unexpected market swings. The commenter also noted that relying solely on the retrospective market measure would make it impossible for the Enterprises to plan ahead, and the lack of a benchmark standard might lower the Enterprises' incentive to support affordable homeowner lending. The commenter stated that the benefit the Enterprises receive from their quasi-governmental status should come with a responsibility to be an affordable housing lending leader.

Comments recommending modified two-part approach (meet both levels). Several housing advocacy groups recommended modifying the current two-part approach by requiring that an Enterprise meet both the benchmark level and the retrospective market measure. The commenters stated that, by itself, the retrospective market measure is inherently circular because the Enterprises continue to purchase a high percentage of the loans originated in the conventional market, i.e., the market level is generally set—and largely guaranteed to be met—by the Enterprises regardless of their progress or failure to provide reasonable access to affordable home loans. The commenters stated that the benchmark level is an essential part of setting meaningful goals but would not alone be sufficient. The commenters stated that the Enterprises should be required to meet both the benchmark and market level tests. The commenters also suggested that exceeding the market level by some margin should be a significant factor in evaluating performance on a housing goal, to ensure that the Enterprises are making substantial progress in returning reasonable accessibility to the market.

FHFA Response

The housing goals are designed to motivate the Enterprises to help make financing available to more borrowers who are creditworthy and well positioned for homeownership. Both Enterprises have taken important steps to help provide access to credit for the populations the goal is intended to serve. However, if the goal is too high, the Enterprises may not be able to meet the goal due to the lack of qualifying loans available for purchase, and a goal set too high could lead them to make inappropriate business decisions to meet the goal that are not consistent with safety and soundness.

Comments recommending modified two-part approach (meet retrospective market level only for enforcement). Two policy advocacy groups recommended that FHFA maintain the current two-part approach but use the retrospective market measure only for enforcement purposes for determining whether to impose penalties on an Enterprise for failure to meet a benchmark level. The commenters noted that relying solely on a benchmark level can be problematic if the benchmark level is either too high or too low, but relying solely on the retrospective market measure would undermine the Enterprises' incentive to promote affordable lending products. The commenters' recommendation is similar to the current two-part approach in that under the commenters' recommendation, if an Enterprise fails to meet the benchmark level, FHFA would look at the market level for enforcement purposes and if the Enterprise met the market level, FHFA presumably would take no enforcement action against the Enterprise. Under the current approach, if an Enterprise fails to meet the benchmark level but meets the market level, it has met the goal and no enforcement action is taken against the Enterprise.

FHFA Response

Both tests—the benchmark and the retrospective market—serve important purposes. The benchmarks, which are prospective, provide targets against which the Enterprises can plan for and calibrate their performance. However, benchmarks, which predict market performance years out, are inevitably imperfect. Applying prospective benchmark levels only could result in some years where an Enterprise would be judged against a level that does not reflect what is reasonably feasible given market conditions. The retrospective market measure provides an important safety valve in years when the goal-qualifying share of the overall market turns out to be lower than anticipated. This situation may be expected when prospective benchmark levels are set several years in advance, especially if the benchmark levels are set to encourage the Enterprises to lead the market in supporting affordable housing. Applying the retrospective market measure only if there has been a “substantial” market downturn would be too uncertain due to the difficulties of defining whether there has been a substantial downturn triggering the use of the retrospective measure. Such an approach would introduce greater uncertainty to the process of evaluating Enterprise performance and would make it more difficult for the Enterprises to plan.

Comments recommending prospective benchmark test only. Several housing advocacy groups stated that FHFA lacks the legal authority under the Safety and Soundness Act to adopt the retrospective market measure as a stand-alone measure or as a component of the two-part approach for determining Enterprise compliance with the single-family goals. The commenters stated that the prospective benchmark standard is the most appropriate standard both legally and as a policy matter to encourage the Enterprises to lead the market.

FHFA Response

The inclusion of the retrospective market measure in the two-part approach is fully consistent with the Safety and Soundness Act and Congressional intent in delegating responsibility for setting the housing goals to FHFA. The statute provides that the single-family goals “shall be established as a percentage of the total number of conventional, conforming, single-family, owner-occupied, purchase money [or refinance] mortgages purchased by the [E]nterprise. . ..” ^[17] This language is consistent with setting goals prospectively as a fixed percentage of mortgages purchased, but it is also consistent with the retrospective market measure of FHFA's two-part approach. The retrospective market measure uses actual market performance, measured as the percentage of total market production that consists of goals-eligible mortgages, and that percentage is established as the goal for Enterprise purchases. The various provisions in the statute enabling the goals to be adjusted based on market conditions are evidence of Congressional intent that the goals generally be related to and even based on the market for loans in the various goal categories, and that the goals should be set in light of market conditions. Those provisions include: (i) The requirement that FHFA calculate the preceding three-year average percentages of goal-eligible originations for each goal category, and take that information into account in setting the single-family goals; ^[18] (ii) the authority to adjust goals, when they have been set for more than one year, based on market conditions; ^[19] (iii) the discretionary authority to reduce a goal in response to a petition from an Enterprise, either in response to market conditions or if efforts to meet the goal could potentially constrain liquidity; ^[20] and (iv) the provisions for relief from enforcement if goals are determined not to have been feasible.^[21]

Comments recommending enforcement and adjustment of the housing goals. A comment from housing advocacy groups recommended that FHFA more fully enforce the housing goals through detailed examination of failed or infeasible goals and by requiring a detailed housing plan, where appropriate. A comment from policy advocacy groups recommended that FHFA adjust the benchmark levels upwards for future years if the market level for a goal is consistently above the benchmark level.

FHFA Response

FHFA places a high priority on the housing goals and uses a range of tools, both formal and informal, to monitor, analyze and enforce the goals. As discussed above, FHFA has authority to adjust a benchmark level upward or downward through notice-and-comment rulemaking.^[22] If, after publication of this final rule, FHFA determines that any of the single-family or multifamily benchmark levels should be adjusted upward or downward in light of market conditions, to ensure the safety and soundness of the Enterprises, or for any other reason, FHFA will take any steps that are necessary and appropriate to adjust the benchmark levels.

B. Factors Considered in Setting the Single-Family Housing Goal 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 goal levels:

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 under this section in previous years;

4. The ability of the Enterprise 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.^[23]

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

The Safety and Soundness Act requires FHFA to consider the percentage of goal-qualifying mortgages under each housing goal, as calculated based on HMDA data for the three most recent years for which data are available, when setting the prospective benchmark levels for the single-family goals.^[24] FHFA has incorporated HMDA data in the goals process, by comparing actual goal performance with market performance through the retrospective approach. The HMDA performance numbers are provided in the tables in subsequent sections for each of the single-family housing goals.

1. FHFA's Market Estimation Models

In setting the benchmark levels for the single-family goals, FHFA relies extensively on projections of the estimated shares of home purchase or refinance mortgages originated in the single-family primary conventional conforming market that will qualify for each goal or subgoal. These projections are based on FHFA's market estimation models. The confidence intervals around the forecasted point estimates in the models for the final rule narrowed from those for the proposed rule due mainly to inclusion in the models of the additional year of HMDA data for 2013 and refining the models' equations to obtain statistically better fitting models. The addition of the 2013 HMDA data provided 12 additional data points (months) from which the parameters were estimated, resulting in one less year of forecasting, i.e., the forecasting started at January 2014 instead of January 2013. With the inclusion of the 2013 HMDA data, FHFA re-estimated all four housing goal/subgoal estimation models. This re-estimation resulted in a slightly different mix of explanatory variables, as some variables in the previous models no longer provided statistically significant impacts, while other variables that were not significant in past models proved to be significant in the current models. The specific market estimation model projections for each housing goal are discussed below.

The market estimation models incorporate four of the seven statutory factors that FHFA is required to consider in setting the benchmark levels. The models are designed to measure the size of the single-family mortgage market (Factor 6), and in so doing, they consider aspects of three of the other factors: Factor 1: National Housing Needs; Factor 2: Economic, Housing, and Demographic Conditions; and Factor 5: Other Mortgage Data. Information about economic and housing conditions, such as the unemployment rate, inflation, housing starts, home sales, and home prices, which are produced by the U.S. Bureau of Labor Statistics, U.S. Census Bureau, and FHFA, are included in the market estimation models. FHFA also considers various other mortgage data sources, including the Mortgage Bankers Association's mortgage default survey, the National Association of Realtors' Housing Affordability Index, and Freddie Mac's Primary Mortgage Market Survey.

FHFA's market estimation models ^[25] are econometric time-series models that examine the relationship between (a) the historical market performance for each single-family housing goal, as calculated from HMDA data, and (b) the historical values for various factors that may influence the market performance, such as interest rates, inflation, house prices, home sales, and the unemployment rate. The models use all available relevant historical information based on statistically significant correlations among economic, housing and mortgage data, and the mortgage affordability measures over time. The models' parameters are re-estimated annually as HMDA data become available in September of each year.

The market estimation models then use available updated government and industry forecasts for each of the variables influencing market performance—most significantly interest rates and inflation—to project an estimated goal-qualifying share of the market for each goal or subgoal. Specifically, the models yield a point estimate for each goal that represents the best estimate of goal-qualifying shares for each year (i.e., 2015, 2016, and 2017), as well as a range around that point estimate representing the confidence that the range includes the actual future market affordability measure for the goal (referred to as the “confidence interval”). The wider the confidence interval, the less exact the point estimate, and vice versa. For example, the estimate for the low-income home purchase goal for 2015 is 22.4 percent, with a 95 percent confidence interval of plus or minus 3.2 percent. In other words, the model prediction is that there is a 95 percent chance that the actual market share in 2015 will be between 19.2 percent and 25.6 percent. The same forecast for 2017 is 22.0 percent, with a 95 percent confidence interval of plus or minus 5.0 percent. Thus, the model prediction range for 2017 is between 17.0 percent and 27.0 percent. The same pattern holds for each of the forecasts: The confidence intervals widen for each successive year in the forecast, reflecting greater uncertainty about the market shares for the later years in the forecast.

The market estimation models are limited by two factors. First, to specify the market accurately, as defined in the regulation, affordability is measured using HMDA data going back to 2004; pre-2004 data are not used in the parameter estimation, because it was missing important variables that make comparisons to post-2004 originations problematic. Second, some explanatory variables, such as inventory, vacancy rates, rents and completions, which are known to be correlated with mortgage affordability, are not available in the government- and industry-produced forecasts and, therefore, those variables are not able to be included in the parameter estimation.

In response to the comments discussed below, FHFA plans to engage in additional discussions with interested parties regarding its market estimation models, and may make adjustments to the models as warranted. If changes are made to the models, FHFA may engage in additional rulemaking, if necessary, to adjust the benchmark levels for the goals.

Comments on Proposed Rule

Several commenters, including housing advocacy groups, policy advocacy groups and a trade association, provided similar comments on the market estimation models used by FHFA in setting the benchmark levels for the single-family housing goals. These comments are discussed below. Neither Fannie Mae nor Freddie Mac commented on the market estimation models.

(1) Confidence Intervals

Commenters stated that the confidence intervals for the market size estimates in the models showed wide ranges of possible affordable housing, limiting the usefulness of the estimates.

FHFA Response

Changes in the models since the proposed rule have narrowed these confidence intervals, in some cases considerably. In response to comments, FHFA tested additional explanatory variables and, in some cases, incorporated them into the revised models. In addition, FHFA had the benefit of an additional year of actual economic data that became available since the proposed rule was posted for comment in August, 2014. In addition, the updated forecasts incorporate changes in the economic outlook by government and industry observers. Most significantly, they reflect changes in the outlook for interest rates and inflation. As a result, the models' confidence intervals in the final rule are much narrower than in the proposed rule. For example, in the proposed rule, the point estimate for the 2015 low-income home purchase goal was 20.9 percent, with a 95 percent confidence interval of 14.2 percent to 27.6 percent. In the final rule, the point estimate for this goal is 22.4 percent, with a 95 percent confidence interval of 19.2 percent to 25.6 percent. This is about half the width of the confidence interval in the proposed rule.

In addition, FHFA notes that under its models, the mean forecast is FHFA's best estimate of what the goal-qualifying share of the market will be at any particular month between January 2015 and December 2017. FHFA has considered all applicable factors in setting the goals, which are generally not identical to the forecasted mean values for the goal-qualifying market shares. In particular, FHFA gives weight to past Enterprise performance on each goal. The inclusion of the retrospective market measure in the housing goals determination takes into account the uncertainty with the benchmark level forecasts.

(2) Certain Variables Not Included

Some commenters stated that certain important variables were omitted from the models for specific goals in order to keep the confidence intervals from becoming even wider. Commenters recommended that any variable that improves the fit of the models should be included, even if it is not statistically significant.

FHFA Response

FHFA notes that it followed common econometric practice by testing and evaluating many explanatory variables but publishing for the proposed rule only statistically significant explanatory variables that provided the best fit model. In the process of re-estimating the market models for the final rule and in response to the comments, FHFA has added and tested additional explanatory variables including: Monthly binary variables for the 2004-2007 period to capture structural shifts in the market; loan-to-value (LTV) share variables; an owner-occupied share variable; and an adjustable rate mortgage share variable. The additional variables in the models did not materially change the results in the forecast point estimates for the final rule. Four model specifications are presented for each single-family goal in FHFA's research paper published on its Web site in order to compare the impact of including or excluding explanatory variables. The paper is available at: http://www.fhfa.gov/​PolicyProgramsResearch/​Research/​.

(3) Data Period Used

Some commenters stated that FHFA's model forecasts give too much weight to recent years, which reflect more limited credit availability. The commenters recommended that FHFA consider market data from periods that may reflect more normal levels of credit availability. The commenters noted that FHFA based its best fit model forecasts on market data from 2004-2014 and stated that those years reflect atypical market conditions. From 2004-2007, the market was characterized by historically low interest rates, with home prices rising and falling dramatically and liberal extensions of credit. In contrast, from 2008-2013, the market was characterized by significant tightening of credit availability. The commenters stated that excluding market data from periods prior to 2004 resulted in benchmark estimates that are too low. The commenters pointed out that even if interest rates and home prices increase over the next three years, they will still be at very favorable levels historically and will be at least as favorable as the numerous years prior to the mortgage boom when affordable housing lending levels by the Enterprises were much higher.

FHFA Response

FHFA agrees that additional data points, including prior to the market boom, should improve forecast accuracy, i.e., better fit models. FHFA's forecasts do not use HMDA data prior to 2004 for several reasons. Explanatory variables that were found to be predictive in one or more of the models are not available prior to 2004. Pre-2004 HMDA data did not identify property type, lien status, Home Ownership Equity Protection Act (HOEPA) status, and the Average Prime Offer Rate (APOR) rate spread. It was also less precise in identifying manufactured loans and subprime loans. All of these factors make it difficult to define the market using pre-2004 data as specified in the regulation.

In response to comments, FHFA did test model specifications that included monthly data going back to January 1996. A detailed description of that analysis is included as an appendix to the FHFA research paper that was discussed earlier and that is posted on FHFA's Web site. The results using pre-2004 data may be less reliable because either the confidence intervals are wider using the 1996-2013 data (as in the case of the single-family, low-income borrower home purchase goal and low-income areas subgoal), or the predicted trends do not coincide with what we have observed in recent months (in the case of the single-family, very low-income home purchase and low-income refinance goals). FHFA determined that the predicted trends resulting from the models using the shorter 2004-2013 time series are preferable.

(4) Impact of Enterprises' Dominant Share of Market

Some commenters stated that the models do not capture the Enterprises' dominant share of the conventional mortgage market, which enables the Enterprises to greatly impact the mix of loans that lenders produce. The commenters stated that the models do not take into account factors that explain the impact of Enterprise policies on the market that are likely to significantly affect the market for affordable loans. These commenters cited as an example the changes in the representations and warranties policies that will reduce Enterprise mortgage buyback risk, which may result in elimination of lenders' credit overlays and, therefore, an increase in affordability of loans.

FHFA Response

FHFA considers these factors in its judgment involved in setting the final levels of the goals after it estimates its models. FHFA recognizes the significant impact that the Enterprises have on the market. While FHFA supports Enterprise efforts to expand credit availability for borrowers at different income levels and in different areas, those efforts must be consistent with the safe and sound operation of the Enterprises.

(5) Impact of Share of Government-Insured Mortgages

Some commenters stated that the models do not appropriately take into account the FHA, Department of Veterans Affairs, and other government agency market shares. These commenters stated that a large FHA market share raises questions about why the Enterprises cannot compete with FHA for the same segments of the market.

FHFA Response

FHFA recognizes that the FHA market share will have some impact on the affordable portion of the conventional mortgage market. In fact, FHA share was tested as an explanatory variable in the market models for both the home purchase and refinance goals. It proved to be statistically significant only in the low-income areas subgoal and refinance goal models.

(6) Frequency of Market Assessments

Several commenters raised the possibility of FHFA conducting more frequent reassessments of the single-family mortgage market if the models are not changed, including the use of a transparent metric for recalculating the benchmark levels based on changes in the forecasts. A policy advocacy group commenter noted that while this would create greater uncertainty that would make it more difficult for the Enterprises to plan to meet the benchmark levels, a tolerance for shortfall could be built into any goals increased through the reassessments. A trade association commenter recommended annual updates of the market projections and adjustments of the benchmark levels accordingly. A housing advocacy group commenter recommended that FHFA set the benchmark levels for a two-year period, as a means of addressing the uncertainty in the models about the size of the market in the third year of the forecast. Another housing advocacy group commenter stated that in light of the uncertainty of the models, FHFA could monitor market trends and revise the benchmark levels as needed, or set higher benchmark levels.

FHFA Response

After consideration of the comments, FHFA has decided to continue to set the benchmark levels in the final rule for a three-year period, as permitted by the Safety and Soundness Act.^[26] FHFA recognizes the limitations of forecasting the market for a three-year period. However, the inclusion of the retrospective market measure helps to ensure feasibility of the goals, especially during the later years of the three-year period. In addition, if FHFA determines that the benchmark levels need to be adjusted in light of changes in the market at any point in the future, FHFA will take all appropriate steps, including possibly adjusting the benchmark levels for the goals.

(7) Transparency of the Models

For the proposed rule, FHFA tested several specifications of the market estimation models but published only the best fit model on FHFA's Web site, since it was the model relevant to the market affordability forecasts. A number of commenters requested that FHFA make more information available about its market estimation models to enable more meaningful comments on the methodology used. A policy advocacy group commenter stated that a sensitivity analysis that shows how the models respond to changes in the values of variables, both for those used and those omitted, would be useful. The commenter stated that without more information about the models, it is difficult to suggest how the models could be improved or compensated for by setting different benchmark levels. A housing advocacy group commenter stated that the monthly nationwide time series provided by the Federal Financial Institutions Examination Council (FFIEC), which serves as the basis for FHFA's market estimation models, should be made publicly available. The commenter stated that disclosure of the data, which is aggregated, would not create privacy or confidentiality problems, and would allow researchers to reproduce, and possibly modify, FHFA's results, with the aim of improving their predictive accuracy.

FHFA Response

In response to the comments, FHFA is publishing on its Web site four models that capture different model specifications, as well as the model specification used in the proposed rule and re-estimated for the final rule using updated data. The models are contained in FHFA's research paper available at http://www.fhfa.gov/​PolicyProgramsResearch/​Research/​.

As noted above, FHFA welcomes input on how the market estimation models could be enhanced to improve market forecasts. FHFA plans to engage in additional discussions with interested parties on the models and may make adjustments to the models as warranted. If changes are made to the models, FHFA may engage in additional rulemaking, if necessary, to adjust the benchmark levels.

2. Past Performance of the Enterprises

The past performance of the Enterprises on each of the single-family housing goals and subgoal, Factor 3 above, is also an important factor in setting the benchmark levels. FHFA has reviewed the actual performance of the Enterprises on each housing goal in previous years and compared that performance to the performance of the overall single-family mortgage market to help FHFA ensure that the benchmark levels are set at levels that are feasible. For example, the market estimation models may not capture all of the factors that contribute to Enterprise performance, such as changes in lender underwriting standards and the resulting impact on credit availability. FHFA's measurements of the mortgage market using HMDA data may not reflect the exact portion of the market that is eligible for purchase by the Enterprises, for example, because not all lenders are required to report data under HMDA. FHFA may rely more heavily on past Enterprise performance if the market estimation models yield results that are far above, or far below, the past performance of either Enterprise on a housing goal. The Enterprises' past performance on the housing goals is discussed under each of the housing goals below.

3. Other Factors

FHFA has also considered the remaining two statutory factors in setting the single-family benchmark levels: Factor 4: Ability to Lead the Industry and Factor 7: Need to Maintain Sound Financial Condition. FHFA's consideration of these factors takes into account the financial condition of the Enterprises, the importance of maintaining the Enterprises in sound and solvent financial condition, and the appropriate role of the Enterprises in relation to the overall single-family mortgage market. The recent performance of the Enterprises and the past and expected performance of the overall single-family market also contribute to FHFA's consideration of these statutory factors.^[27] Factors 4 and 7 are discussed under each of the housing goals below.

FHFA continues to monitor the activities of the Enterprises, both in FHFA's capacity as safety and soundness regulator and as conservator. If necessary, FHFA will make any appropriate changes to the single-family benchmark levels to ensure the continued safety and soundness of the Enterprises.

C. Single-Family Benchmark Levels

1. Low-Income Home Purchase Goal—§ 1282.12(c)

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 area median income. After consideration of the statutory factors, including updated forecasts from FHFA's market estimation models, preliminary figures on goal performance in 2014, as reported by the Enterprises, and the comments received on the proposed benchmark level for this goal, which are discussed below, § 1282.12(c) of the final rule sets the annual benchmark level for this goal for 2015 through 2017 at 24 percent. The 24 percent level is one percentage point above the benchmark level for 2014 and the proposed benchmark level for 2015-2017.

Because this final rule is being published well into 2015, FHFA will consider that timing as part of the evaluation of the Enterprises' actual 2015 housing goals performance.

Market Size

FHFA's consideration of the size of the single-family mortgage market takes into account both the actual size of the market in previous years, as measured using the most recent HMDA data available, and FHFA's forecast for the size of the market based on its market estimation models.

As indicated in Table 1, FHFA's forecasts for the low-income share of the overall market for home purchase mortgages for 2015 through 2017, which are the result of updating the market estimation models used by FHFA to forecast the market size for the proposed rule through May 2015, are significantly lower than the actual low-income shares of the overall market for home purchase mortgages in 2010 through 2013. The proposed rule estimated the low-income shares of the market as 20.9 percent for 2015, 20.2 percent for 2016, and 19.8 percent for 2017. FHFA's updated market estimation models project that the low-income borrower shares of the overall home purchase mortgage market will be 22.4 percent for 2015, 22.9 percent for 2016, and 22.0 percent for 2017. The forecast ranges are 19.2 percent-25.6 percent for 2015, 18.7 percent-27.1 percent for 2016, and 17.0 percent-27.0 percent for 2017. As can be seen, the updated estimates for 2015 and 2016 are higher than the estimates that were used for the proposed rule, and this was taken into account in setting the goal level at 24 percent for 2015-2017, an increase of one percentage point from the 2014 benchmark level and from the level in the proposed rule.

Past Performance of the Enterprises

As indicated in Table 1, the performance of the Enterprises on the low-income home purchase goal has followed a pattern similar to that for the overall market performance on the goal since 2010—steady performance in 2010 through 2012, followed by lower levels in 2013 and 2014. However, while the low-income share of the market was lower in 2013 and 2014, the total volume of single-family home purchase loans in those years was significantly higher than in 2010 through 2012. Fannie Mae's performance in 2010 was 25.1 percent, which increased to 25.8 percent in 2011, before falling slightly to 25.6 percent in 2012 and 23.8 percent in 2013. Freddie Mac's performance in 2010 was 26.8 percent, before declining to 23.3 percent in 2011, increasing to 24.4 percent in 2012, and declining to 21.8 percent in 2013. Preliminary performance figures as reported by the Enterprises for 2014 indicate that Fannie Mae's performance on this goal was approximately 23.5 percent and Freddie Mac's performance was approximately 21.0 percent. Official 2014 performance figures as determined by FHFA, as well as the retrospective HMDA market performance numbers, will be available later in 2015. The market share shown in Table 1 for 2014 is a forecast based on FHFA's market model. With the exception of Fannie Mae's reported performance in 2014, the performance level of each Enterprise on the low-income home purchase goal was below the retrospective HMDA share for each year from 2010 through 2014.

Table 1—Enterprise Low-Income Home Purchase Goal

Year	Type of Home Purchase (HP) mortgages	Benchmark	Performance		Market share estimate
			Fannie Mae	Freddie Mac
2010	Low-Income HP Mortgages		120,430	82,443
	Total HP Mortgages		479,200	307,555
	Low-Inc. % of HP Mortgages	27%	25.1%	26.8%	27.2%
2011	Low-Income HP Mortgages		120,597	60,682
	Total HP Mortgages		467,066	260,796
	Low-Inc. % of HP Mortgages	27%	25.8%	23.3%	26.5%
2012	Low-Income HP Mortgages		162,486	70,393
	Total HP Mortgages		633,627	288,007
	Low-Inc. % of HP Mortgages	23%	25.6%	24.4%	26.6%
2013	Low-Income HP Mortgages		193,712	93,478
	Total HP Mortgages		814,137	429,158
	Low-Inc. % of HP Mortgages	23%	23.8%	21.8%	24.0%
2014	Low-Income HP Mortgages		177,846	108,948
	Total HP Mortgages		757,870	519,731
	Low-Inc. % of HP Mortgages	23%	23.5%	21.0%	22.0%
	95% Confidence Interval				+/−2.0%
2015	Final Rule Benchmark	24%			22.4%
	95% Confidence Interval				+/−3.2%
2016	Final Rule Benchmark	24%			22.9%
	95% Confidence Interval				+/−4.2%
2017	Final Rule Benchmark	24%			22.0%
	95% Confidence Interval				+/−5.0%
Source: Official performance as determined by FHFA for 2010-13; preliminary performance figures for 2014 as reported by the Enterprises. Actual goal-qualifyiing market shares, based on FHFA analysis of HMDA data, for 2010-13. FHFA estimates of goal-qualifying market shares for 2014-17.

Analysis

The final rule sets the annual benchmark level for the low-income home purchase goal at 24 percent for 2015 through 2017, which is one percentage above both the actual benchmark level for 2014 and the level in the proposed rule for 2015-2017. As shown in Table 1, the market estimation models forecast a range of possible market levels. The benchmark level of 24 percent is above the point estimates for 2015-2017 but within the confidence intervals for all three years. Although FHFA's market estimation models forecast declines in the low-income share of the overall home purchase mortgage market between 2015 and 2017, the point estimate of 22.4 percent for 2015 is subject to less uncertainty than the point estimate of 22.0 percent for 2017. Recent data also show a decline in the Enterprises' performances from 2012 to 2013 on this goal, and a further decline in market performance with a revised market size estimate of 22.0 percent for 2014. However, a benchmark level of 24 percent will encourage the Enterprises to continue their efforts to promote safe and sustainable lending to low-income families if the market share turns out to be smaller than 24 percent. This may include any steps the Enterprises take to bring greater certainty to origination and servicing representation and warranty standards for lenders, any additional outreach to small and rural lenders and to state and local housing finance agencies, and any other efforts by the Enterprises to reach underserved creditworthy borrowers. The above factors, taken together, support setting the benchmark level somewhat above the market estimate for 2015, but still well within the confidence interval.

FHFA will continue to monitor the Enterprises in its capacities as regulator and as conservator, and FHFA will take any steps appropriate to address changes in market conditions.

Comments on Proposed Rule

Several commenters supported the proposed benchmark level of 23 percent for the low-income home purchase goal.

A housing advocacy group commenter recommended that the benchmark levels be set at the upper ranges of the market estimates, or the Enterprises otherwise would have little incentive to increase their purchases of goal-qualifying loans. The commenter noted that the retrospective market measure will serve as a fallback if the levels turn out to be too high.

A number of housing advocacy and policy advocacy group commenters recommended setting a higher benchmark level of 27 percent. A housing advocacy group commenter cited limitations of the market estimation models, the fact that 27 percent was the level in effect in 2010 and 2011, and the fact that the Enterprises exceeded 23 percent in almost every year since 2001. Another housing advocacy group commenter also recommended 27 percent based on its concerns about the market estimation models and the Enterprises' “tight credit box,” which the commenter stated has driven many low-income borrowers and borrowers of color out of the home purchase market. A policy advocacy group commenter recommended setting an “aggressive” benchmark level of 27 percent given the uncertainty in the market estimation models and other data strongly indicating a lack of access to conventional conforming mortgage credit by lower-income and minority borrowers.

A housing advocacy group commenter recommended that the benchmark level be set higher than 27 percent, based on historical market size data from years pre-dating the housing crisis and on the Enterprises' goal performance during that period. The commenter stated that the period between 2000 and 2004 reflected economic conditions and a market environment that more closely align with 2015-2017 and, therefore, the 2000-2004 period would provide a more useful comparison for purposes of setting the benchmark levels for the single-family housing goals. The commenter stated that while the proposed 23 percent level might be higher than FHFA's point estimates for the overall market share projected for low-income home mortgage purchases for 2015-2017, the benchmark level should be set as a “stretch” goal of at least 28 percent. The commenter based its recommendation on the Enterprises' past performance during the 2000-2004 period, their current dominant position in the secondary mortgage market, and improved market performance expectations.

Fannie Mae commented that the proposed 23 percent level reflected an appropriate analysis and application of the statutory factors. Freddie Mac did not comment on the proposed benchmark level.

FHFA Response

As discussed above, the final rule sets the annual benchmark level for 2015-2017 at 24 percent, which is slightly higher (1.6 percentage points) than the point estimate for 2015 but well within the confidence intervals for all three years. FHFA believes this is an appropriate benchmark level based on the market estimation models' forecasts for 2015-2017, the Enterprises' recent performance, the updated market size estimate for 2014, and the goal to encourage the Enterprises to continue their efforts to promote safe and sustainable lending to low-income families.

2. Very Low-Income Home Purchase Goal—§ 1282.12(d)

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. After consideration of the statutory factors, including updated forecasts from FHFA's market estimation models, and the comments received on the proposed benchmark level for this goal, which are discussed below, § 1282.12(d) of the final rule sets the annual benchmark level for this goal for 2015 through 2017 at 6 percent. The 6 percent level is one percentage point below both the benchmark level for 2014 and the proposed benchmark level.

Market Size

As discussed above, FHFA's consideration of the size of the single-family market takes into account both the actual size of the market in previous years, as measured using the most recent HMDA data available and FHFA's forecast for the size of the market based on its market estimation model.

As shown in Table 2, FHFA's forecasts for the very low-income share of the overall market for home purchase mortgages for 2015 through 2017 are lower than the actual very low-income share of the overall market for home purchase mortgages in 2010 through 2013, and are similar to the estimated very low-income share for 2014. These estimates are the result of updating the market estimation models used by FHFA to forecast the market size for the proposed rule. The proposed rule estimated the very low-income shares of the market at 5.8 percent for 2015, 5.7 percent for 2016, and 5.6 percent for 2017. FHFA's updated market estimation models project through May 2015 that the very low-income shares of the overall market for home purchase mortgages will be almost the same for each year: 5.9 percent for 2015, 6.0 percent for 2016, and 5.7 percent for 2017. The forecast ranges at a 95 percent confidence level are 3.4 percent-8.4 percent for 2015, 2.8 percent-9.2 percent for 2016, and 1.9 percent-9.5 percent for 2017.

Past Performance of the Enterprises

As indicated in Table 2, the performance of the Enterprises on the very low-income home purchase goal was relatively stable between 2010 and 2012, before declining in 2013 and further in 2014. As discussed above for the low-income home purchase goal, while the very low-income share of the market was lower in 2013 and 2014, the total volume of single-family home purchase loans in those years was significantly higher than in 2010 through 2012. Fannie Mae's performance was 7.2 percent in 2010, 7.6 percent in 2011 and 7.3 percent in 2012, while Freddie Mac's performance was 7.9 percent in 2010, 6.6 percent in 2011 and 7.1 percent in 2012. Preliminary performance figures as reported by the Enterprises for 2014 indicate that Fannie Mae's performance on this goal was 5.7 percent, and Freddie Mac's performance was 4.9 percent. Official 2014 performance figures as determined by FHFA, as well as the retrospective HMDA market performance numbers, will be available later in 2015. The market share shown in Table 2 for 2014 is a forecast based on FHFA's market model. With the exception of Fannie Mae's reported performance in 2014, the performance level of each Enterprise on the very low-income home purchase goal was below the retrospective HMDA share each year from 2010 through 2014.

Table 2—Enterprise Very Low-Income Home Purchase Goal

Year	Type of Home Purchase (HP) mortgages	Benchmark	Performance		Market share/ estimate
			Fannie Mae	Freddie Mac
2010	Very Low-Income HP Mortgages		34,673	24,276
	Total HP Mortgages		479,200	307,555
	Very Low-Inc. % of HP Mortgages	8%	7.2%	7.9%	8.1%
2011	Very Low-Income HP Mortgages		35,443	17,303
	Total HP Mortgages		467,066	260,796
	Very Low-Inc. % of HP Mortgages	8%	7.6%	6.6%	8.0%
2012	Very Low-Income HP Mortgages		46,519	20,469
	Total HP Mortgages		633,627	288,007
	Very Low-Inc. % of HP Mortgages	7%	7.3%	7.1%	7.7%
2013	Very Low-Income HP Mortgages		48,810	23,705
	Total HP Mortgages		814,137	429,158
	Very Low-Inc. % of HP Mortgages	7%	6.0%	5.5%	6.3%
2014	Very Low-Income HP Mortgages		42,872	25,232
	Total HP Mortgages		757,870	519,731
	Very Low-Inc. % of HP Mortgages	7%	5.7%	4.9%	5.7%
	95% Confidence Interval				+/−1.4%
2015	Final Rule Benchmark	6%			5.9%
	95% Confidence Interval				+/−2.5%
2016	Final Rule Benchmark	6%			6.0%
	95% Confidence Interval				+/−3.2%
2017	Final Rule Benchmark	6%			5.7%
	95% Confidence Interval				+/−3.8%
Source: Official performance as determined by FHFA for 2010-13; preliminary performance figures for 2014 as reported by the Enterprises. Actual goal-qualifying market shares, based on FHFA analysis of HMDA data, for 2010-13. FHFA estimates of goal-qualifying market shares for 2014-17.

While the recovery in the home purchase market between 2012 and 2013 resulted in significantly higher volumes of home purchase mortgages acquired by the Enterprises, the volume of very low-income home purchase mortgages did not increase by nearly as much as the rest of the market. Between 2012 and 2013, the volume of Fannie Mae's purchases of very low-income home purchase mortgages increased by 5 percent, while its overall volume of home purchase mortgages increased by 28 percent. As a result, Fannie Mae's very low-income home purchase goal performance fell from 7.3 percent in 2012 to 6.0 percent in 2013. Similarly, the volume of Freddie Mac's purchases of very low-income home purchase mortgages increased by 16 percent, while its overall volume of home purchase mortgages increased by 49 percent. As a result, Freddie Mac's very low-income home purchase goal performance fell from 7.1 percent in 2012 to 5.5 percent in 2013.

Analysis

The final rule sets the annual benchmark level for the very low-income home purchase goal for 2015 through 2017 at 6 percent, which is one percentage point below both the actual benchmark level for 2014 and the level in the proposed rule for 2015-2017. It is more difficult for the Enterprises to manage their percentage of very low-income mortgage purchases because of the small number of such loans available to them. Further, given the Enterprises' preliminary performance figures for 2014 (Fannie Mae at 5.7 percent and Freddie Mac at 4.9 percent), FHFA believes the proposed 7 percent target would have been difficult for either Enterprise to achieve in 2014. The 6 percent benchmark level will still encourage the Enterprises to continue their efforts to promote safe and sustainable lending to very low-income families.

As shown in Table 2, the market estimation models forecast point estimates for this goal of 5.9 percent, 6.0 percent and 5.7 percent in 2015, 2016 and 2017, respectively. Recent data show a decline in the Enterprises' performances in 2012-2014, relative to previous years, on this goal. The 6 percent benchmark level is set essentially at the forecast midpoint to encourage the Enterprises to continue their efforts to promote safe and sustainable lending to very low-income families. As discussed above, this may include any steps the Enterprises take to bring greater certainty to origination and servicing standards for lenders, any additional outreach to small and rural lenders and to state and local housing finance agencies, and any other efforts by the Enterprises to reach underserved creditworthy borrowers. FHFA recognizes that this benchmark level may be challenging to meet, though less so than the 7 percent level in the proposed rule, as the Enterprises may not purchase loans inconsistent with safety and soundness. If an Enterprise fails to meet the benchmark level, it may still meet the goal if its performance equals or exceeds the retrospective market level.

HMDA data suggest that banks are keeping an increasingly higher share of mortgages to low-income and very low-income borrowers in their portfolios, meaning that they are not sold to any entity on the secondary market, making it more difficult for either Enterprise to reach the market level. Possible explanations are that: Lenders are originating these loans to comply with the Community Reinvestment Act but prefer to hold them in portfolio to protect against the risk that the Enterprises require the lenders to repurchase the loans, which they may consider somewhat more likely to default, because of violations to representations and warranties, or the loans are originated without private mortgage insurance and/or below market interest rates, meaning the lenders would need to sell the loans to the Enterprises at a loss and/or take recourse on the loans. In addition, FHA's mortgage insurance premium reduction of 50 basis points has the result that its execution is cheaper for many low-income borrowers with less than perfect credit scores.

FHFA will continue to monitor the Enterprises in its capacities as regulator and as conservator, and FHFA will take any steps appropriate to address changes in market conditions.

Comments on Proposed Rule

A housing advocacy group commenter recommended that the benchmark level be set at the upper range of the market estimates, or the Enterprises otherwise would have little incentive to increase their purchases of very low-income loans. A comment from policy advocacy groups recommended setting an “aggressive” benchmark level given the uncertainty in the market estimation models and other data strongly indicating a lack of access to conventional conforming mortgage credit by lower-income borrowers and minority borrowers. A comment from housing advocacy groups also recommended setting a higher benchmark level due to the uncertainty in the market estimation models. A non-profit housing developer suggested that the very low-income share of the market is expected to be around 7 to 8 percent, but did not provide a source for that forecast.

Fannie Mae commented that it opposed the proposed benchmark level of 7 percent for this goal, recommending a 6 percent level instead. Fannie Mae noted that FHFA's market size forecasts for this goal in the proposed rule were 5.8 percent for 2015, 5.7 percent for 2016, and 5.6 percent for 2017 and, thus, were lower than the proposed benchmark level of 7 percent. Fannie Mae stated that setting the benchmark level significantly higher than the market size forecasts in order to encourage the Enterprises to continue their efforts to promote safe and sustainable lending to very low-income families could have the unintended negative consequence of suggesting that the Enterprises should undertake efforts that may not contribute to a safe and sustainable market. In addition, Fannie Mae stated that it is already committed to a variety of efforts to support financing for very low-income borrowers, including its standard product eligibility criteria for 95 percent LTV loans, targeted products such as MyCommunityMortgage, acquiring loans through its partnerships with housing finance agencies, reintroducing acquisitions of loans from HUD's Section 184 program and the U.S. Department of Agriculture's Rural Development 502 program that serve Native American and rural communities, and changing requirements for loans to borrowers with derogatory credit events, such as foreclosures, short sales, deed-in-lieu transfers and bankruptcy, to facilitate earlier borrower requalification.

Freddie Mac did not comment on the proposed benchmark level for the very low-income home purchase goal.

FHFA Response

As discussed above, the final rule sets the annual benchmark level for 2015-2017 at 6 percent, which is above the point estimates but within the confidence intervals for all three years. FHFA believes this is an appropriate benchmark level based on the market estimation models' forecasts for 2015-2017, the Enterprises' recent performance, the updated market size estimate for 2014, and the goal to encourage the Enterprises to continue their efforts to promote safe and sustainable lending to very low-income families.

3. Low-Income Areas Home Purchase Subgoal—§ 1282.12(f)

The low-income areas home purchase subgoal is based on the percentage of all single-family, owner-occupied home purchase mortgages acquired 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 area median income; or (2) for families with incomes less than or equal to area median income 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 the area median income). After consideration of the statutory factors, including updated forecasts from FHFA's market estimation models, and the comments received on the proposed benchmark level for this subgoal, which are discussed below, § 1282.12(f) of the final rule sets the annual benchmark level for this subgoal for 2015 through 2017 at 14 percent. The 14 percent level is higher than the 11 percent level for 2014 and the same as the proposed benchmark level.

Market Size

As discussed above, FHFA's consideration of the size of the single-family market takes into account both the actual size of the market in previous years, as measured using the most recent HMDA data available, and FHFA's forecast for the size of the market based on its market estimation model.

As shown in Table 3, FHFA's forecasts for the low-income areas shares of the overall market for home purchase mortgages for 2015 and 2016 are lower than the actual low-income areas share of the overall market for home purchase mortgages in 2013 and the current estimate for 2014. The proposed rule estimated the low-income areas shares of the market as 14.7 percent for 2015, 14.7 percent for 2016, and 14.2 percent for 2017. FHFA's updated market estimation models project that the low-income areas shares of the overall home purchase market will be somewhat lower, with point estimates of 13.2 percent for 2015, 13.6 percent for 2016, and 14.2 percent for 2017. The forecast ranges are 11.7 percent-14.7 percent for 2015, 10.8 percent-16.4 percent for 2016, and 10.6 percent-17.8 percent for 2017.

Past Performance of the Enterprises

As indicated in Table 3, Fannie Mae's performance on the low-income areas home purchase subgoal was 12.4 percent in 2010, declined to 11.6 percent in 2011, and increased to 13.1 percent in 2012 and 14.0 percent in 2013. Freddie Mac's performance followed the same basic pattern—its performance was 10.4 percent in 2010, declined to 9.2 percent in 2011, and increased to 11.4 percent in 2012 and 12.3 percent in 2013. Preliminary performance figures as reported by the Enterprises for 2014 indicate that Fannie Mae's performance on this subgoal was 15.5 percent, and Freddie Mac's performance was 13.6 percent. Official 2014 performance figures, as well as the retrospective HMDA market performance numbers, will be determined by FHFA later in 2015. The market share shown in Table 3 for 2014 is a forecast based on FHFA's market model. While Freddie Mac's performance on the low-income areas home purchase subgoal was below the retrospective HMDA share each year from 2010 through 2014, Fannie Mae's performance exceeded the retrospective HMDA share in several of those years.

Table 3—Enterprise Low-Income Areas Home Purchase Subgoal

Year	Type of Home Purchase (HP) mortgages	Benchmark	Performance		Market share/ estimate
			Fannie Mae	Freddie Mac
2010	Low-Income Area HP Mortgages		44,467	23,928
	High-Minority Area HP Mortgages		14,814	8,161
	Subgoal-Qualifying Total		59,281	32,089
	Total HP Mortgages		479,200	307,555
	Subgoal Benchmark/Performance	13%	12.4%	10.4%	12.1%
2011	Low-Income Area HP Mortgages		40,736	18,270
	High-Minority Area HP Mortgages		13,549	5,632
	Subgoal-Qualifying Total		54,285	23,902
	Total HP Mortgages		467,066	260,796
	Subgoal Benchmark/Performance	13%	11.6%	9.2%	11.4%
2012	Low-Income Area HP Mortgages		60,927	24,586
	High-Minority Area HP Mortgages		22,275	8,164
	Subgoal-Qualifying Total		83,202	32,750
	Total HP Mortgages		633,627	288,007
	Subgoal Benchmark/Performance	11%	13.1%	11.4%	13.6%
2013	Low-Income Area HP Mortgages		86,430	40,444
	High-Minority Area HP Mortgages		27,425	12,177
	Subgoal-Qualifying Total		113,855	52,621
	Total HP Mortgages		814,137	429,158
	Subgoal Benchmark/Performance	11%	14.0%	12.3%	14.2%
2014	Low-Income Area HP Mortgages		91,691	55,987
	High-Minority Area HP Mortgages		25,650	14,808
	Subgoal-Qualifying Total		117,341	70,795
	Total HP Mortgages		757,870	519,731
	Subgoal Benchmark/Performance	11%	15.5%	13.6%	14.0%
	95% Confidence Interval				+/−0.6%
2015	Final Rule Benchmark	14%			13.2%
	95% Confidence Interval				+/−1.5%
2016	Final Rule Benchmark	14%			13.6%
	95% Confidence Interval				+/−2.8%
2017	Final Rule Benchmark	14%			14.2%
	95% Confidence Interval				+/−3.6%
Source: Official performance as determined by FHFA for 2010-13; preliminary performance figures for 2014 as reported by the Enterprises. Actual subgoal-qualifying market shares, based on FHFA analysis of HMDA data, for 2010-13. FHFA estimates of subgoal-qualifying market shares for 2014-17.

Analysis

The final rule sets the annual benchmark for this subgoal at 14 percent for 2015-2017, which is higher than the actual benchmark level of 11 percent for 2014 and the same as the level in the proposed rule for 2015-2017. As shown in Table 2, the market estimation models forecast a range of possible market levels. The benchmark level of 14 percent is above the point estimates of 13.2 percent and 13.6 percent for 2015 and 2016, respectively, and just below the point estimate of 14.2 percent for 2017, but well within the confidence intervals for all three years. It is the same as or higher than both Enterprises' performance on this subgoal in 2012 and 2013. Recent data also show an increase in the Enterprises' performances in 2012 through 2014, relative to previous years, on this subgoal. The benchmark level is not being raised to 15 percent as this would rely too heavily on Fannie Mae's reported performance of 15.5 percent for 2014. While Freddie Mac's performance has increased, reaching a reported 13.6 percent for 2014, it would be less likely to reach 15 percent in 2015-2017.

FHFA will continue to monitor the Enterprises in its capacities as regulator and as conservator, and FHFA will take any steps appropriate to address changes in market conditions.

Comments on Proposed Rule

Several policy advocacy group commenters and Fannie Mae supported the proposed 14 percent benchmark level. One commenter stated that, “[h]aving subgoals for . . . households in low-income areas will encourage credit to flow to these households and communities suffering from lack of access to credit.” The commenters supported the increase from the 11 percent benchmark level for 2014, noting that the Enterprises' past performance demonstrates their ability to meet an increased level without increasing risk, and an increase in the level will further meet the needs of geographically underserved areas. Fannie Mae stated that the proposed 14 percent level reflected an appropriate analysis and application of the statutory factors.

A housing advocacy group commenter recommended setting the benchmark level at the upper range of the market estimates because it believes that the Enterprises would otherwise have little incentive to increase their purchases of goal-qualifying loans. A comment from policy advocacy groups recommended setting an “aggressive” benchmark level, given the uncertainty in the market estimation models and other data strongly indicating a lack of access to conventional conforming mortgage credit by lower-income borrowers and minority borrowers. A comment from housing advocacy groups also recommended setting a higher benchmark level due to the uncertainty in the market estimation models.

Freddie Mac did not comment on the proposed benchmark level.

FHFA Response

As discussed above, the final rule sets the annual benchmark level for 2015-2017 for this subgoal at 14 percent, which is above the point estimates for 2015 and 2016 and just below the point estimate for 2017, but within the confidence intervals for all three years. FHFA believes this is an appropriate benchmark level based on the market estimation models' forecasts for 2015-2017, the Enterprises' recent performance, and the updated market size estimate for 2014.

4. Low-Income Areas Home Purchase Goal—§ 1282.12(e)

Section 1282.12(e) provides that the low-income areas home purchase goal includes all mortgages that are counted for purposes of the low-income areas home purchase subgoal discussed above (families in low-income areas and moderate-income families who reside in high-minority census tracts), as well as home purchase mortgages for families with incomes no greater than 100 percent of area median income who reside in Federally-declared disaster areas (regardless of the minority share of the population in the tract or the ratio of tract median family income to area median income).

FHFA does not separately forecast the size of the market for the low-income areas home purchase goal and does not establish a benchmark level for the goal in advance in the housing goals regulation. The benchmark level for this goal is determined each year based on the benchmark level for the low-income areas home purchase subgoal, plus an additional amount determined each year by FHFA separately from rulemaking to reflect the disaster areas covered for that year.

Designated disaster areas include counties declared by the Federal Emergency Management Agency to be disaster areas eligible for individual assistance during the previous three years. This is referred to as the “disaster areas increment.” It is established through an FHFA analysis of HMDA data for the most recent three-year period available. Given the lag in the release of HMDA data, the disaster areas increment for 2013 was based on disaster areas declared between 2010 and 2012, but the increment was calculated using HMDA data for 2009 through 2011, because 2012 HMDA data were not available until later in 2013. The disaster areas increment used in setting the benchmark level of the goal for 2014 was based on disaster areas declared between 2011 and 2013, but the increment was calculated using HMDA data for 2010 through 2012. Thus, the disaster areas increment, and the resulting low-income areas home purchase goal, can vary from one year to the next.

For 2012, the disaster areas increment was 9 percent; thus, the overall low-income areas home purchase goal for that year was 20 percent (11 percent + 9 percent). For 2013 and 2014, the disaster areas increment was 10 percent; thus, the overall low-income areas goal for those years was 21 percent (11 percent + 10 percent). For 2015-2017, the disaster areas increment will be provided by letter to the Enterprises each year based on updated disaster area information.

Past performance on the low-income areas home purchase goal is shown below in Table 4.

Table 4—Enterprise Low-Income Areas Home Purchase Goal

Year	Type of Home Purchase (HP) mortgages	Benchmark	Performance		Market share/ estimate
			Fannie Mae	Freddie Mac
2010	Subgoal-Qualifying HP Mortgages		59,281	32,089
	Disaster Areas HP Mortgages		56,076	38,898
	Goal-Qualifying Total		115,357	70,987
	Total HP Mortgages		479,200	307,555
	Goal Benchmark/Performance	24%	24.1%	23.1%	24.0%
2011	Subgoal-Qualifying HP Mortgages		54,285	23,902
	Disaster Areas HP Mortgages		50,209	26,232
	Goal-Qualifying Total		104,494	50,134
	Total HP Mortgages		467,066	260,796
	Goal Benchmark/Performance	24%	22.4%	19.2%	22.0%
2012	Subgoal-Qualifying HP Mortgages		83,202	32,750
	Disaster Areas HP Mortgages		58,085	26,486
	Goal-Qualifying Total		141,287	59,236
	Total HP Mortgages		633,627	288,007
	Goal Benchmark/Performance	20%	22.3%	20.6%	23.2%
2013	Subgoal-Qualifying HP Mortgages		113,855	52,621
	Disaster Areas HP Mortgages		62,314	33,123
	Goal-Qualifying Total		176,169	85,744
	Total HP Mortgages		814,137	429,158
	Goal Benchmark/Performance	21%	21.6%	20.0%	22.1%
2014	Subgoal-Qualifying HP Mortgages		117,341	70,795
	Disaster Areas HP Mortgages		54,548	33,923
	Goal-Qualifying Total		171,889	104,718
	Total HP Mortgages		757,870	519,731
	Goal Benchmark/Performance	18%	22.7%	20.1%	NA
Source: Official performance as determined by FHFA for 2010-13; preliminary performance figures for 2014 as reported by the Enterprises. Actual goal-qualifyiing market shares, based on FHFA analysis of HMDA data, for 2010-13. Goal-qualifying market share for 2014 will be available after FHFA analysis of HMDA data for 2014.

5. Low-Income Refinancing Goal—§ 1282.12(g)

The low-income refinancing goal is based on the percentage of all refinancing mortgages on owner-occupied single-family housing purchased by an Enterprise that are for low-income families, defined as families with incomes less than or equal to 80 percent of the area median income. After consideration of the statutory factors, including updated forecasts from FHFA's market estimation models and the comments received on the proposed benchmark level for this goal, which are discussed below, § 1282.12(g) of the final rule sets the annual benchmark level for this goal for 2015 through 2017 at 21 percent. The 21 percent level is higher than the 20 percent level for 2014, but lower than the proposed benchmark level of 27 percent. FHFA's updated forecasts project a significantly smaller low-income share of the overall refinancing mortgage market compared to the forecasts FHFA used to set the benchmark level in the proposed rule.

Market Size

FHFA's consideration of the size of the single-family market takes into account both the actual size of the market in previous years, as measured using the most recent HMDA data available, and FHFA's forecast for the size of the market based on its market estimation model.

The low-income share of the overall market for refinancing mortgages is strongly affected by the overall volume of refinancings. The size of the entire refinancing mortgage market has an impact on the share of affordable refinancing mortgages (defined as refinancing mortgages for borrowers with incomes of 80 percent or less of area median income) and, thus, on the development of the benchmark level for the Enterprises for the low-income refinancing goal. Refinancing mortgage volume has historically increased when the refinancing of mortgages is motivated by low interest rates, i.e., “rate-and-term refinances,” and this increased volume is typically dominated by higher-income borrowers. Consequently, in periods of low interest rates, the share of lower-income borrowers refinancing often decreases. The opposite is true when interest rates rise—there are usually fewer refinancings overall, but a greater percentage of those are cash-out refinancings by low-income borrowers. Because interest rates and mortgage rates are currently continuing at relatively low levels, the low-income share of borrowers who are refinancing has continued at relatively low levels.^[28]

The proposed rule estimated the low-income refinancing shares of the market as 31.0 percent for 2015, 33.5 percent for 2016, and 34.2 percent for 2017. As shown in Table 5, FHFA's updated market estimation models project that the low-income refinancing shares of the market will be much lower—21.8 percent for 2015, 22.4 percent for 2016 and 22.8 percent for 2017. The forecast ranges are 19.1 percent-24.5 percent for 2015; 17.7 percent-27.1 percent for 2016; and 16.2 percent-29.0 percent for 2017. FHFA's updated forecasts for 2015 through 2017 are significantly lower than the estimates used in the proposed rule, but still higher than the 2014 benchmark level of 20 percent.

Past Performance of the Enterprises

As indicated in Table 5, the performance of the Enterprises on the low-income refinancing goal has followed a similar pattern as the overall market performance on this goal since 2010, although the performance of the Enterprises varied more over the period than the overall market performance. Fannie Mae's performance on the low-income refinancing goal in 2010 was 20.9 percent, and increased to 24.3 percent in 2013 and a reported 26.5 percent in 2014. Freddie Mac's performance on the low-income refinancing goal in 2010 was 22.0 percent, and increased to 24.1 percent in 2013 and a reported 26.4 percent in 2014.

Table 5—Enterprise Low-Income Refinancing Goal

Year	Type of Home Purchase (HP) mortgages	Benchmark	Performance		Market share/ estimate
			Fannie Mae	Freddie Mac
2010	Low-Income % of Refinance Mortgages	NA	19.3%	20.8%	20.2%
	Low-Income % of HAMP Modifications	NA	69.9%	67.5%	NA
	Goal Benchmark & Performance	21%	20.9%	22.0%	NA
2011	Low-Income % of Refinance Mortgages	NA	21.3%	21.2%	21.5%
	Low-Income % of HAMP Modifications	NA	71.2%	67.3%	NA
	Goal Benchmark & Performance	21%	23.1%	23.4%	NA
2012	Low-Income % of Refinance Mortgages	NA	21.2%	21.5%	22.3%
	Low-Income % of HAMP Modifications	NA	72.9%	69.3%	NA
	Goal Benchmark & Performance	20%	21.8%	22.4%	NA
2013	Low-Income Refinance Mortgages		519,753	306,205
	Total Refinance Mortgages		2,170,063	1,309,435
	Low-Income % of Refinance Mortgages	NA	24.0%	23.4%	24.3%
	Low-Income HAMP Modifications		11,858	14,757
	Total HAMP Modifications		16,478	21,599
	Low-Income % of HAMP Mods	NA	72.0%	68.3%	NA
	Low-Income Refis/HAMP Mods		531,611	320,962
	Total Refis/HAMP Mods		2,186,541	1,331,034
	Goal Benchmark & Performance	20%	24.3%	24.1%	NA
2014	Low-Income Refinance Mortgages		215,826	131,921
	Total Refinance Mortgages		831,218	514,936
	Low-Income % of Refinance Mortgages	NA	26.0%	25.6%	26.2%
	95% Confidence Interval				±1.5%
	Low-Income HAMP Modifications		6,503	6,795
	Total HAMP Modifications		9,288	10,335
	Low-Income % of HAMP Mods	NA	70.0%	65.7%	NA
	Low-Income Refis/HAMP Mods		222,329	138,716
	Total Refis/HAMP Mods		840,506	525,271
	Goal Benchmark & Performance	20%	26.5%	26.4%	NA
2015	Final Rule Benchmark (incl. HAMP mods)	21%			21.8%
	95% Confidence Interval				±2.7%
2016	Final Rule Benchmark (incl. HAMP mods)	21%			22.4%
	95% Confidence Interval				±4.7%
2017	Final Rule Benchmark (incl. HAMP mods)	21%			22.8%
	95% Confidence Interval				±6.2%
Source: Official performance as determined by FHFA for 2010-13; preliminary performance figures for 2014 as reported by the Enterprises. Actual goal-qualifying market shares, based on FHFA analysis of HMDA data, for 2010-13. FHFA estimates of goal-qualifying market shares for 2014-17. Note that market results/estimates do not take into account HAMP modifications due to lack of data (See discussion below.)
Detailed data on the total and goal-qualifying volumes of refinancing mortgages and HAMP modifications for 2010-12 are presented in FHFA's proposed housing goals rule, Federal Register, September 11, 2014, Table 8, p. 54515.

Analysis

The final rule sets the annual benchmark level for this goal at 21 percent for 2015 through 2017, which is higher than the actual benchmark level of 20 percent for 2014, but below the level in the proposed rule for 2015-2017 of 27 percent. As shown in Table 5, the market estimation models forecast a range of possible market levels. The benchmark level of 21 percent is slightly lower than the point estimate of 21.8 percent for 2015, and lower than the point estimates of 22.4 percent for 2016 and 22.8 percent for 2017, and within the confidence intervals for all three years.

FHFA's current market forecast has moderated considerably for this goal, down by nine percentage points in 2015, and just over 11 percentage points in 2016 and 2017. This calls into question the magnitude of the increase in the proposed rule. FHFA has also reviewed the Enterprises' month-by-month performance for the second half of 2014 and observed a steady decline in the low-income share of refinance mortgages over this period.

The final rule, therefore, sets the benchmark level for this goal at 21 percent for 2015-2017, which is 1 percentage points higher than the 2014 level, but 6 percentage points lower than the level in the proposed rule. This is consistent with FHFA's updated forecasts for 2015-2017.

FHFA will continue to monitor the Enterprises in its capacities as regulator and as conservator, and FHFA will take any steps appropriate to address changes in market conditions.

Comments on Proposed Rule

Several commenters supported the proposed benchmark level of 27 percent for this goal. Fannie Mae commented that the proposed 27 percent level reflected an appropriate analysis and application of the statutory factors. Freddie Mac did not comment on the proposed benchmark level.

A comment from a housing counseling group suggested raising the benchmark level to 35 percent to help “reduce unnecessary displacement and loss of potential wealth building of homeowners with Enterprises' guaranteed mortgages.” A housing advocacy group commenter recommended that the benchmark level be set at the upper range of the market estimates because it believes that the Enterprises would otherwise have little incentive to increase their purchases of low-income refinancing loans. A comment from policy advocacy groups recommended setting an “aggressive” benchmark level, given the uncertainty in the market estimation models and other data strongly indicating a lack of access to conventional conforming mortgage credit by lower-income borrowers and minority borrowers. A comment from housing advocacy groups also recommended setting a higher benchmark level due to the uncertainty in the market estimation models.

FHFA Response

As described above, FHFA believes that given current conditions in the refinance market, a larger increase from the 2014 benchmark level of 20 percent would be too substantial an increase in the goal. As discussed above, the final rule sets the annual benchmark level for 2015-2017 for this goal at 21 percent, which is slightly lower than the point estimate of 21.8 percent for 2015, lower than the point estimates of 22.4 percent for 2016 and 22.8 percent for 2017, and within the confidence intervals for all three years. FHFA believes this is an appropriate benchmark level based on the market estimation models' forecasts for 2015-2017, the Enterprises' recent performance, and the updated market size estimate for 2014.

Counting Loan Modifications—§ 1282.16(c)(10)

Under § 1282.16(c)(10) of the housing goals regulation, Enterprise financings of qualifying permanent modifications of loans for low-income families under the Home Affordable Modification Program (HAMP) are counted toward the low-income refinancing goal. These HAMP permanent loan modifications are the only type of loan modification eligible for counting for purposes of the housing goals. The intent in permitting HAMP permanent loan modifications to count toward the low-income refinancing goal was to encourage support for the HAMP program. In every year from 2010 through 2014, low-income families received at least 67 percent of HAMP loan modifications at each Enterprise. Because the low-income share of all HAMP loan modifications is much higher than the low-income share of all refinancing transactions, including HAMP loan modifications, the low-income refinancing goal increases the performance of the Enterprises on the low-income refinancing goal. This was especially true for 2011, when Fannie Mae's performance was 21.3 percent without HAMP loan modifications, but 23.1 percent with HAMP loan modifications. The impact was even larger for Freddie Mac, whose performance in 2011 was 21.2 percent without HAMP loan modifications, but 23.4 percent with HAMP loan modifications.

However, HAMP loan modifications have had a smaller impact on low-income refinancing goal performance in recent years as HAMP loan modification volume has fallen—for Fannie Mae, from a high of 64,124 loan modifications in 2011 to 9,288 loan modifications in 2014, and for Freddie Mac, from 52,910 loan modifications in 2011 to 10,355 loan modifications in 2014.

Comments on Proposed Rule

Freddie Mac recommended that loan modifications other than HAMP loan modifications also be permitted to count for purposes of the low-income refinancing goal. Freddie Mac stated that its own non-HAMP loan modification programs are largely consistent with HAMP, serving similar goals.

FHFA Response

Because loan modifications are not considered new originations, they are not reported in HMDA data. As a result, it is difficult to adjust the market estimates based on expected modification volumes.

VI. Reporting Requirements for Single-Family Rental Units

In the Notice accompanying the proposed rule, FHFA noted that it plans to require the Enterprises to include more detailed information about their purchases of mortgages on single-family rental housing in the Annual Mortgage Reports (AMRs) that the Enterprises are required to submit under § 1282.62(b) of the current regulation. This additional information will be included in the Enterprise AMRs covering 2015 and years following.

The AMRs currently provide information on Enterprise purchases of all mortgages on owner-occupied and rental properties, regardless of whether the mortgage may be counted for purposes of the housing goals. The additional requirements will provide detailed affordability information on rental units in all single-family properties, whether owner-occupied (with one or more rental units in addition to the owner-occupied unit) or investor-owned.

Comments

FHFA received several comments from policy advocacy groups and housing advocacy groups supporting more detailed reporting in the AMRs. The same commenters also recommended that FHFA establish specific requirements in the regulation for Enterprise support of single-family rental properties.

FHFA Response

The final rule does not revise the regulation to specifically address single-family rental properties. This is consistent with the proposed rule. The additional AMR reporting requirements fall within the scope of the existing regulation, so no changes to the text of the regulation are necessary. FHFA is requiring the Enterprises to provide additional information regarding their purchases of mortgages on single-family rental properties as described in the Notice accompanying the proposed rule. This additional information will be publicly available as part of the housing goals tables submitted as part of the Enterprise AMRs. These housing goals tables are available on FHFA's Web site.^[29]

VII. Multifamily Housing Goals

A. Multifamily Housing Goals Benchmark Levels in Final Rule

1. Multifamily Low-Income Housing Goal—§ 1282.13(b)

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 area median income. FHFA has considered each of the statutory factors, including updated forecasts of the multifamily market and the comments received on the proposed benchmark levels for this goal, which are discussed below. Section 1282.13(b) of the final rule sets the same annual benchmark level for each Enterprise at 300,000 low-income units for each year from 2015 through 2017. This is higher than the 2014 benchmark levels (250,000 units for Fannie Mae and 200,000 units for Freddie Mac) and higher than the proposed benchmark levels (250,000 units for Fannie Mae and 210,000 to 230,000 units for Freddie Mac), to account for the overall size of the multifamily finance market, which has expanded substantially since the proposed rule was issued. Each Enterprise has exceeded 250,000 low-income units in each of the past three years, and given the larger size of the current multifamily mortgage market and the expanded exclusions from the 2015 Conservatorship Scorecard multifamily cap, FHFA believes that an annual 300,000 low-income unit goal for 2015-2017 is achievable and appropriate.

2. Multifamily Very Low-Income Housing Subgoal—§ 1282.13(c)

The multifamily very low-income housing subgoal is based on the total number of rental units in multifamily properties financed by mortgages purchased by the Enterprises that are affordable to very low-income families, defined as families with incomes less than or equal to 50 percent of area median income. FHFA has considered each of the statutory factors, including updated forecasts of the size of the multifamily market and the comments received on the proposed benchmark levels for this subgoal, which are discussed below. Freddie Mac has traditionally lagged Fannie Mae under this subgoal, but the gap narrowed considerably in 2013 and 2014. Section 1282.13(c) of the final rule sets Fannie Mae's very low-income subgoal benchmark level at 60,000 units for each year of the three-year goals period, as in the proposed rule. The final rule also sets Freddie Mac's very low-income subgoal benchmark level at 60,000 units for each year of the three-year goals period, which is an increase from the proposed annual benchmark level of 43,000 to 50,000 units. This is consistent with the 2015 Conservatorship Scorecard multifamily cap that permits the same volume cap and exclusions for each Enterprise.

The applicable statutory factors, comments received and analyses supporting these benchmark levels are discussed below.

B. Factors Considered in Setting the Multifamily Housing Goal Benchmark Levels

Section 1333(a)(4) of the Safety and Soundness Act requires FHFA to consider the following six factors in setting the multifamily housing goals:

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

2. The performance and effort of the Enterprise 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 Enterprise 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 Enterprise.^[30]

In setting the benchmark levels for the multifamily housing goals, FHFA has considered each of the six statutory factors. The statutory factors for the multifamily goals are very similar, but not identical, to the statutory factors that were considered in setting the benchmark levels for the single-family housing goals. There are several important distinctions between the single-family housing goals and the multifamily housing goals. While there are separate single-family goals for home purchase and refinancing mortgages, the multifamily goals include all Enterprise multifamily mortgage purchases, regardless of the purpose of the loan. In addition, unlike the single-family goals, the multifamily goals are set based on the total volume of multifamily mortgage purchases, not on a percentage of overall multifamily mortgage purchases.

Another difference between the single-family and multifamily goals is that performance on the multifamily goals is measured based solely on meeting a benchmark level, without any retrospective market measure. The absence of a retrospective market measure for the multifamily goals is due, in part, to the lack of reliable, comprehensive data about new loan origination activity in the multifamily mortgage market. Unlike the single-family mortgage market, where HMDA provides a reasonably comprehensive data set about single-family mortgage originations each year, the multifamily mortgage market (and the market segment that supports properties with affordable market rents) has no such comparable data set. As a result, it can be difficult to correlate different data sets that may rely on different reporting formats—for example, some data are available by dollar volume while other data are available by unit production. The lack of comprehensive data about the multifamily mortgage market is even more apparent with respect to the segments of the market that are targeted to low-income and very low-income renters. Much of the analysis that follows discusses general trends in the overall multifamily mortgage market, although FHFA recognizes that these general trends may not apply to the same extent to all segments of the market.

FHFA has considered each of the required statutory factors, which are discussed below, a number of which are related or overlap.

C. Analysis of Considerations in Setting the Multifamily Benchmark Levels

1. The Multifamily Mortgage Market: Market Size, Competition and the Affordable Multifamily Market (Factors 1 and 3)

FHFA's consideration of the multifamily mortgage market addresses the size of and competition within the market, as well as the subset of the market that finances units affordable to low-income and very low-income families. Recent trends in the multifamily mortgage market indicate that overall loan volumes have increased substantially from the volumes in 2014, both in terms of total refinancing activity and total financing for property acquisitions and for new multifamily units being completed. FHFA has also considered the importance of Enterprise support of the multifamily mortgage market in light of recent decreases in rental housing affordability.

(i) 2015 Conservatorship Scorecard—Multifamily Limits

Given the increasing participation in the market from private sector capital, FHFA's 2015 Conservatorship Scorecard established a cap of $30 billion on new multifamily loan purchases for each Enterprise. However, consistent with the recent expansion of the market and in order to facilitate market liquidity, especially in the segment of the market that supports properties with affordable rents, FHFA recently revised and expanded the types of affordable housing lending activities that are excluded from the Scorecard cap, as was discussed above.

(ii) Multifamily Mortgage Market Size

The total number of units in multifamily properties in the United States, defined as all units in structures with five or more rental units, was over 18 million in 2013, according to data from the U.S. Census Bureau in the 2013 American Community Survey.^[31] Multifamily mortgage origination volume varies significantly from year to year based on a variety of market conditions. During the financial crisis, the size of the multifamily mortgage market decreased significantly before rebounding in 2013 and beyond. Overall, multifamily mortgage originations fell from $147.7 billion in 2007 to $87.9 billion in 2008 to $52.5 billion in 2009, as shown in Table 6. The declines were even more pronounced in the private sector segment of the market, which decreased from almost $112 billion in 2007 to $46.4 billion in 2008 to $18.4 billion in 2009. The Enterprises' mortgage purchases provided a countercyclical source of liquidity during this same period. While the size of the overall multifamily market was declining, the volume of Enterprise purchases was relatively steady. The combined volume of Enterprise multifamily mortgage purchases in 2007, excluding purchases of commercial mortgage-backed securities (CMBS), was $34.6 billion, and rose to $40 billion in 2008 before declining to $31 billion in 2009.

Table 6—Government and Private Sector Market Shares of Multifamily Originations

Year	Total volume ($Bil.)	Fannie Mae (%)	Freddie Mac (%)	Enterprise total (%)	FHA (%)	Private sector (%)
2005	$133.1	11.7%	6.7%	18.4%	2.2%	79.3%
2006	$138.0	11.7%	7.1%	18.8%	1.0%	80.2%
2007	$147.7	13.1%	10.4%	23.4%	0.8%	75.8%
2008	$87.9	25.4%	20.1%	45.5%	1.7%	52.8%
2009	$52.5	30.2%	28.9%	59.2%	5.6%	35.2%
2010	$68.8	24.5%	20.3%	44.8%	15.3%	40.0%
2011	$110.1	20.9%	18.9%	39.8%	10.6%	49.6%
2012	$146.1	21.7%	18.3%	39.9%	10.2%	49.8%
2013	$170.0	16.6%	14.8%	31.4%	10.4%	58.3%
2014	$209.9	16.1%	14.9%	31.0%	6.0%	63.0%
Note: FHA data is for fiscal years 2005 to 2014.
Sources: “MBA Commercial Real Estate Finance Survey.”
Sources for 2014 data: Fannie Mae, Freddie Mac, and FHA. Total 2014 volume derived from “MBA Commercial Real Estate Finance Survey” data.
Note: All multifamily loans in CMBS issuances are included under “Private Sector,” regardless of the investor.

Since the financial crisis, the total multifamily origination market has rebounded and has shown increased private capital participation, with private capital defined to include CMBS and insurance company and bank/credit union portfolio purchases. The multifamily new origination market has increased from a low of $52.5 billion in 2009 to $176 billion in 2014.^[32] As the size of the overall market has increased, the Enterprise share of the market has decreased, from a high of almost 60 percent in 2009 to just over one-third in 2014.

Volumes in the overall multifamily new origination market are expected to continue to increase between 2015 and 2017, including refinancing activity, financing for newly constructed multifamily units, and financing for property acquisitions. However, the Enterprises are expected to roughly maintain or slightly increase their current percentage share of the overall market due to increased private capital participation and competition.

Comments on Proposed Rule

A comment from public advocacy groups suggested that, in evaluating the size of the multifamily mortgage market, FHFA should include all rental units, cooperative units and condominiums. The comment pointed to data from the American Community Survey suggesting that a more inclusive definition of the market would result in a significantly larger overall market size and, therefore, increased multifamily goal benchmark levels.

FHFA Response

Although certain cooperative housing blanket loans are eligible for goals credit under the housing goals, FHFA considers cooperative and condominium units to be primarily intended to be owner-occupied and, therefore, including them in the overall multifamily market size would overstate the number of rental units and properties available for financing.

(iii) Affordable Multifamily Mortgage Market Segment

FHFA's consideration of the multifamily mortgage market is limited by the lack of comprehensive data about the size of the market for financing properties affordable to low-income and very low-income families. The challenge of identifying goals-qualifying units is made more difficult because utility allowances must be added to the market rent on all individually metered rental units before calculating a unit's affordability.

FHFA recognizes that the portion of the overall multifamily rental market that is affordable to low-income and very low-income families may vary from year to year, that the competition among capital sources within the market as a whole may differ from the competition within the affordable segment of the market, and that the financing volume for the segment of the market that is affordable to very low-income renters is also related to the limited availability of affordable housing subsidies.

Increasing rents and nearly stagnant wages, particularly for low- and very low- income renters, has resulted in a significant decline in rental housing affordability over the past three years. The Safety and Soundness Act requires FHFA to determine affordability based on rents, which FHFA has defined by regulation to include utilities, not exceeding 30 percent of the relevant percentage of household income.^[33] However, as mentioned above, a recent Harvard study shows that more than half of all tenants pay more than 30 percent of household income for rental housing, especially in the high-cost urban markets where most renters reside and where much of Fannie Mae and Freddie Mac lending is focused. Tenants in the lowest income brackets, such as at the low-income and very low-income goal levels, pay the highest percentage of their income for rental housing. As a result, there are a declining number of low-income and very low-income units that qualify as affordable under the 30 percent test for the Enterprises to finance, and even fewer in the high-cost urban markets where their lenders are most active but where tenant rent burden is the greatest.^[34]

(iv) Factors Impacting the Multifamily Mortgage Market

FHFA has considered a variety of economic indicators and measures related to the size and affordability of the multifamily market, including the market fundamentals and the ongoing need for affordable rental units. This section examines the following such factors: Interest rates, property values, rents, vacancy rates, and housing permits, starts and completions. The trends in each of these factors in recent years have tended to show strong demand for multifamily housing relative to the overall supply, which is reflected in higher property values and rents, lower vacancy rates, and increasing multifamily construction. All of these factors indicate that multifamily mortgage origination volumes can be expected to continue at a relatively high rate.

Interest Rates

The volume of multifamily mortgage originations is heavily influenced by interest rates, with lower rates generating higher loan volumes. Multifamily properties benefit from lower interest rates because reduced borrowing costs increase net property cash flow and, thus, an owner's return on equity. Although interest rates rose in 2013, they decreased in 2014 and have remained low compared to historical levels. Continued low rates in 2015 have contributed to increased mortgage origination volumes for both refinancing and acquisition financing.

Property Values

As of the first quarter of 2015, multifamily property values were up over 16 percent from the first quarter of 2014 and more than 34 percent since the first quarter of 2013, and are now above the valuation peak reached in 2007.^[35] Rising multifamily property values usually spur increases in refinancings, property sales, and new construction activity. Multifamily property values continued to increase through the first quarter of 2015, with more modest increases expected to continue during the remainder of 2015 through 2017.

Multifamily Vacancy Rates and Rents

During the housing crisis, vacancy rates for multifamily properties increased significantly and median asking rents declined. Since then, vacancy rates have dropped while rents have increased. Rental vacancy rates peaked at over 13 percent in the third quarter of 2009, but have declined each year since then to less than 7.1 percent nationwide in the first quarter of 2015.^[36] Median asking rents nationwide have increased steadily since 2011, reaching $734 in 2013 and $756 in the third quarter of 2014.^[37] Both the low vacancy rates and higher asking rents indicate that the demand for multifamily housing will remain strong during the three-year goals period.

Multifamily Building Permits, Starts and Completions

Multifamily building permits and construction starts have recovered in recent years, after falling significantly after the housing market crisis. Multifamily building permits averaged 357,000 units annually between 2005 and 2008 but fell dramatically in 2009 and 2010, to approximately 130,000 units per year. The volume of permits has increased since 2010, exceeding 340,000 units in 2013 and almost reaching the same level in 2014.^[38] Actual multifamily housing starts have followed the same pattern, averaging approximately 287,000 units annually between 2005 and 2008, decreasing to just under 100,000 units annually in 2009 and 2010, but increasing since then to 338,000 units in 2013 and 339,000 units in 2014.^[39]

Multifamily completions have followed a similar pattern. Completions exceeded 250,000 units each year from 2005 through 2009 until declining in 2009 and 2010, when the number of units completed dropped below 150,000 units each year. Multifamily completions have since recovered to pre-2009 levels, reaching 254,000 units in 2014.^[40] Given the recent increases in the volume of multifamily building permits and starts, completions are expected to increase in the coming years, which will generate increased demand for permanent mortgage financing.

2. Past Performance of the Enterprises (Factor 2)

The Enterprises have served a consistent and critical role in the multifamily mortgage market in the years before, during, and since the financial crisis. The 2012 housing goals rule increased the overall multifamily goals for 2012 through 2014 compared to previous years, reflecting the Enterprises' increased market share since 2008. However, the 2012 rule also anticipated the increase in private market activity during 2012 through 2014, and as a result set goal levels that declined in each of those years, with 2012 the highest and 2014 the lowest.

As required by the Safety and Soundness Act, in setting the multifamily goals for 2015 through 2017, FHFA has considered the mortgage purchase performance of the Enterprises in previous years. Previously, FHFA had established higher multifamily goals for Fannie Mae than for Freddie Mac, reflecting the more established history and higher overall loan volumes of Fannie Mae's multifamily business. Moreover, because of its delegated underwriting platform, Fannie Mae, through its lenders, was seen to have a greater origination capacity than Freddie Mac, which underwrites each multifamily loan it purchases. Freddie Mac has also typically financed fewer total units than Fannie Mae on the same dollar volume of loan originations. This was because Freddie Mac usually financed fewer properties that had higher leverage, which were located in high-cost, urban core markets. Freddie Mac has also financed fewer small multifamily properties with 50 or fewer units and fewer properties in secondary, tertiary, or rural markets.

However, that changed in 2014 with Freddie Mac's increased loan production of $28.3 billion, which was a new record and only $500 million less than Fannie Mae. It is expected that both Enterprises will sustain similar high levels of loan production during the three-year goals period of the final rule.

Enterprise Performance on Multifamily Low-Income Housing Goal

The multifamily low-income housing goal includes units affordable to low-income families. Enterprise purchases of mortgages that finance properties with units affordable to low-income families over the 2010-2014 period, are shown in Table 7. From 2010 to 2014, Fannie Mae financed an average of 296,000 such units each year, peaking at 375,924 units in 2012, and Freddie Mac financed an average of 244,000 such units each year, peaking at 298,529 units in 2012. Since 2010, Fannie Mae and Freddie Mac financings have yielded a relatively stable percentage of mortgages financing low-income units relative to their total mortgage purchases, as is shown in Table 7. The share of low-income units financed by Fannie Mae compared to its total multifamily mortgage purchases rose from 68 percent in 2009 to a range of 75 to 77 percent between 2010 and 2014. Similarly, the share of low-income units financed by Freddie Mac rose from 65 percent in 2009 to a range of 75 to 79 percent between 2010 and 2014.^[41]

Until 2014, Fannie Mae had consistently financed more low-income units than Freddie Mac, by a relatively stable amount. However, in 2014, due to its increased loan volume, Freddie Mac surpassed Fannie Mae's low-income unit production. In that year, Freddie Mac financed 273,582 low-income units (above its goal of 200,000), compared to Fannie Mae's 262,050 units (above its goal of 250,000).

Table 7—Enterprise Past Performance on Low-Income Multifamily Goal, 2006-14

[Goals and performance measured in low-income multifamily units financed]

Year	Fannie Mae				Freddie Mac
	Total multifamily				Total multifamily
	Goal	Performance	Units financed	Low income (%)	Goal	Performance	Units financed	Low income (%)
2014	250,000	262,050	372,072	70%	200,000	273,582	366,377	75%
2013	265,000	326,597	430,751	76%	215,000	254,628	341,490	75%
2012	285,000	375,924	501,256	75%	225,000	298,529	377,522	79%
2011	177,750	301,224	390,526	77%	161,250	229,001	290,116	79%
2010	177,750	214,997	286,504	75%	161,250	161,500	216,042	75%
2009	NA	235,199	344,989	68%	NA	167,026	256,346	65%
2008	NA	450,850	653,060	69%	NA	268,036	375,760	71%
2007	NA	392,666	668,963	59%	NA	298,746	388,072	77%
2006	NA	313,620	427,130	73%	NA	174,377	224,608	78%
Source: Performance as reported by the Enterprises for 2014; official performance as determined by FHFA for 2010-13; performance if the goal had been in effect for 2006-09 as calculated by FHFA. “Low-income” refers to units affordable to renters with incomes no greater than 80 percent of Area Median Income (AMI), based on rental proxy.
Note: Figures do not include units financed by the purchase of commercial mortgage-backed securities (CMBS).

Enterprise Performance on Multifamily Very Low-Income Subgoal

The multifamily very low-income housing subgoal includes units affordable to very low-income families. Enterprise-financed properties with units affordable to very low-income families from 2010-2013 are shown in Table 8. From 2010 to 2013, Fannie Mae financed an average of 81,000 very low-income units each year, peaking at 108,878 units in 2012, whereas Freddie Mac financed an average of 46,000 such units each year, peaking at 60,084 units in 2012.

In recent years, Fannie Mae has financed a higher percentage of very low-income units than has Freddie Mac, although the difference was very small in 2013, as shown in Table 8. The share of very low-income units financed by Fannie Mae was 18 percent of its overall purchases in 2009, rising to 22 percent in 2011 and 2012, and then falling to 18 percent in 2013 and 16 percent in 2014. Freddie Mac financing of very low-income units was unusually low in 2009, at 8 percent of its overall purchases, but returned to a more typical level of 14 percent in 2010. It has fluctuated since then, increasing to 17 percent in 2013 and decreasing to 13 percent in 2014.^[42]

In 2014, both Enterprises reported that they exceeded their very low-income subgoals. As shown in Table 8, Fannie Mae financed 60,542 such units compared to its 2014 goal of 60,000 units, and Freddie Mac financed 48,689 such units compared to its 2014 goal of 40,000 units.

Table 8—Enterprise Past Performance on Very Low-Income Multifamily Goal, 2006-14

[Goals and performance measured in low-income multifamily units financed]

Year	Fannie Mae				Freddie Mac
	Total multifamily				Total multifamily
	Goal	Performance	Units financed	Very low income (%)	Goal	Performance	Units financed	Very low income (%)
2014	60,000	60,542	372,072	16%	40,000	48,689	366,377	13%
2013	70,000	78,071	430,751	18%	50,000	56,752	341,490	17%
2012	80,000	108,878	501,256	22%	59,000	60,084	377,522	16%
2011	42,750	84,244	390,526	22%	21,000	35,471	290,116	12%
2010	42,750	53,908	286,504	19%	21,000	29,656	216,042	14%
2009	NA	60,765	344,989	18%	NA	20,302	256,346	8%
2008	NA	96,242	653,060	15%	NA	45,154	375,760	12%
2007	NA	88,901	668,963	13%	NA	59,821	388,072	15%
2006	NA	88,521	427,130	21%	NA	34,638	224,608	15%
Source: Performance as reported by the Enterprises for 2014; official performance as determined by FHFA for 2010-13; performance if the goal had been in effect for 2006-09, as calculated by FHFA. “Very low-income” refers to units affordable to renters with incomes no greater than 50 percent of Area Median Income (AMI), based on rental proxy.
Note: Figures do not include units financed by the purchase of commercial mortgage-backed securities (CMBS).

3. Ability of the Enterprises to Lead the Market in Making Multifamily Mortgage Credit Available (Factor 4)

In setting the multifamily housing goals benchmark levels, FHFA has considered the ability of the Enterprises to lead the market in making multifamily mortgage credit available. As discussed, the Enterprises' share of the overall mortgage market increased in the years immediately following the financial crisis and decreased in subsequent years in response to growing private sector participation. Despite the Enterprises' reduced market share in the overall multifamily mortgage market, FHFA expects them to demonstrate leadership in multifamily affordable housing lending, which includes supporting housing for tenants at different income levels in various geographic markets and in various market segments.

4. Availability of Public Subsidies (Factor 5)

The broad decline in rental housing affordability has particularly affected very low-income renters (households with incomes at or below 50 percent of area median income), so the number of market rate units qualifying as affordable for the very low-income goal that are available for the Enterprises to finance is limited and will likely decline in each year of the three-year goals period. Thus, the ability of either Enterprise to meet the very low-income subgoal is largely dependent on the availability of rental housing subsidies to make units affordable to very low-income households (known as targeted affordable housing), because in many rental markets there are few, if any, units with market rents that are affordable to very low-income households using the required 30 percent of income test for rent plus tenant paid utilities.^[43]

The number of subsidized projects available to finance is finite due to the limited amount of subsidies available and the limited number of subsidized properties. Thus, it would be difficult for the Enterprises to increase their share of the subsidized housing finance market and to finance greater numbers of such units beyond their current levels of activity.

These factors have less impact on the low-income goal because that goal targets households with incomes at or below 80 percent of area median income, while housing subsidy programs generally target households with incomes at or below 60 percent of area median income.^[44] The low-income goal, thus, is usually met through financing properties that contain non-subsidized, market rate units, which have rents that are affordable to low-income households.

5. Need To Maintain Sound Financial Condition of the Enterprises (Factor 6)

In setting the multifamily goal benchmark levels, FHFA also considered the importance of maintaining the Enterprises in sound and solvent financial condition. During the conservatorships, under both stressed and normal market conditions, the delinquency and default performance of Enterprise loans on affordable housing properties has not been significantly different from loans on market rate properties, which have experienced extremely low delinquency and foreclosure rates. The Enterprises should, therefore, be able to sustain or increase their purchases of loans on affordable properties without impacting the Enterprises' safety and soundness or negatively affecting the performance of their portfolios. FHFA continues to monitor the activities of the Enterprises, both in FHFA's capacity as safety and soundness regulator and as conservator. If necessary, FHFA could make appropriate changes to the multifamily goal benchmark levels to ensure the Enterprises' continued safety and soundness.

Analysis

Based on FHFA's analysis of the factors discussed above, the final rule sets the multifamily goals generally higher than the Enterprises' reported actual low-income and very low-income goals performance in 2014, reflecting the substantially increased size of the multifamily finance market in 2015 and the revised 2015 Conservatorship Scorecard.

Beginning with their actual 2014 loan production totals and continuing in 2015, FHFA expects both Enterprises to have substantially equivalent total multifamily loan volumes for each year of the three-year goals period, with their combined volume representing between one-third to 40 percent of the estimated new origination market size during those years. Given the significant expansion of the multifamily market in 2015, the final rule revises the proposed benchmark level for the multifamily low-income goal by setting the same annual level for each Enterprise at 300,000 low-income units for each year of the three-year goals period. The fact that both Enterprises exceeded 250,000 low-income units in each of the past three years, when they had considerably lower annual loan origination volume than in 2015, demonstrates that the low-income goal of 300,000 units is achievable, given the larger size of the current market.

The final rule also revises the proposed benchmark level for the multifamily very low-income goal by setting both Fannie Mae's and Freddie Mac's goals at 60,000 units for each year of the three-year goals period. Fannie Mae's performance was above the 60,000 unit level in 2014, and Freddie Mac's performance fell below the 60,000 unit level in 2014. Nonetheless, in light of the significant expansion of the multifamily financing market in 2015 and the revised 2015 Conservatorship Scorecard, FHFA believes that the very low-income goals in the final rule are achievable.

However, in light of the declining number of qualifying affordable low-income units available to finance in the current rental market due to the market forces discussed in previous sections, FHFA expects both Enterprises will require increasing efforts to meet the low-income unit goal during the three-year goals period as compared to previous years. Those efforts will likely include adjustments to existing loan products, expanded specialized lender networks, and increased marketing efforts. FHFA does not expect either Enterprise to engage in any transaction that does not involve a reasonable rate of return. A reasonable rate of return on mortgages for properties with rents affordable to very low- and low-income families may be less than the return earned on other activities, in order to meet the goals. FHFA will take market conditions and other appropriate factors into account in assessing Enterprise performance on the multifamily goals. FHFA could also adjust the levels of the multifamily goals in future years if necessary.

Comments on Proposed Rule

FHFA specifically requested comment in the proposed rule on whether the benchmark levels would be achievable and appropriate for the Enterprises. A number of commenters stated that the benchmark levels should be set at higher levels than proposed. Commenters noted that while the Enterprises' role in the multifamily mortgage market is expected to be maintained or possibly decrease over the coming years as private capital becomes increasingly prevalent, the overall market is expected to continue to grow. A comment from policy advocacy groups stated that, even if the overall volume of Enterprise multifamily mortgage purchases declines, the number of affordable units they support should remain higher than proposed. The comment stated that increased market competition has come from life insurance companies that tend to invest in properties geared toward higher income earners. The comment also noted that both Enterprises easily exceeded both multifamily goals over the past several years. A trade association commenter recommended that the proposed benchmark levels be increased to encourage the Enterprises to expand their relationships with housing finance agencies, noting that the Enterprises have been strong partners in supporting housing finance agencies in the development and rehabilitation of affordable rental properties. Several commenters stated that there is a severe shortage of affordable rental housing and that both Enterprises could do more to support such housing. The commenters, thus, encouraged FHFA to set “stretch” benchmark levels as an incentive to the Enterprises to increase their affordable mortgage purchase volumes.

Another trade association commenter stated that in setting the benchmark levels, FHFA should consider market trends such as increased competition from the private market, as well as the interplay with regulatory directives such as the portfolio dollar volume limits for the Enterprises under conservatorship and FHFA's proposed rule on the Enterprise duty to serve underserved markets. The commenter stated that the housing goals should be aligned with the priorities set by FHFA for the Enterprises in conservatorship, whether in the Conservatorship Scorecard or through other means. The commenter recommended that FHFA monitor multifamily market conditions closely to determine whether any of the multifamily goals should be adjusted.

Fannie Mae commented that it was committed to meeting the benchmark levels in the proposed rule, but stated that the multifamily mortgage market has changed and will continue to change, including a decline in the Enterprises' multifamily mortgage market share and an overall trend of increased competition from the private sector. Fannie Mae also stated that while there have been recent increases in the volume of multifamily building permits and housing starts, very little of this new construction is targeting class B and C properties, which in general are older and smaller properties with fewer amenities and which generally provide more affordable units than class A properties. Fannie Mae provided data showing that class B and C properties made up 65 percent of all multifamily properties in 2000, but dropped to 58 percent by 2013. Fannie Mae stated that the market changes will make the proposed benchmark levels difficult to meet, and in the absence of a retrospective market measure for the multifamily goals, indicated that it may request that FHFA reduce the benchmark levels if circumstances warrant in the future.

FHFA also specifically requested comment in the proposed rule on whether the goals should be set at different levels for each Enterprise or if the levels should be the same. Several trade association commenters stated that the benchmark levels of both Enterprises should be the same, while others supported the proposal to raise Freddie Mac's goals levels, which have lagged behind Fannie Mae's goals levels for many years. A trade association commenter recommended that over time, both Enterprises should be subject to the same benchmark levels.

Freddie Mac commented that it welcomes the challenge of gradually increasing its multifamily loan purchases from 2015-2017, but stated that the historical difference in the volume of multifamily business at each Enterprise warrants maintaining the difference in the goal levels between the two Enterprises. Freddie Mac stated that every loan it finances supports affordable rental housing, and historically, approximately 90 percent of the total financing it provides in any given year supports moderate-income households, defined as households with incomes at or below 100 percent of area median income.

A comment from policy advocacy groups suggested that FHFA revisit pre-conservatorship initiatives such as those providing lines of credit to mission-based entities that build or preserve affordable housing. The comment also recommended that FHFA consider providing bonus goals credit for Enterprise purchases of mortgages financing multifamily properties located outside of areas with high concentrations of minority and low-income residents. The comment stated that housing located in communities with better schools, transportation, and employment potential can lead to significant improvements in resident outcomes.

FHFA Response

FHFA has taken into consideration the views of the commenters and has adjusted the goals in the final rule consistent with the expanded size of the market, the revised 2015 Conservatorship Scorecard and to reinforce FHFA's emphasis on providing financing for affordable rental housing. However, there is currently no shortage of private capital serving multifamily lending beyond the Enterprises' established market share, nor does FHFA expect there to be any shortage during the new three-year goals period, including from depository institutions. Mortgage Bankers Association data show the Enterprises' market share falling from over 60 percent during the height of the recession in 2009, to approximately one-third in 2014, or close to historical norms, with increased volumes in 2015. The Enterprises fulfilled their countercyclical function when most lenders withdrew from the market in 2008 and 2009 and remained the market leaders until commercial mortgage markets stabilized over the past several years. Furthermore, setting goals for the Enterprises that are too high could be disruptive to the multifamily mortgage market by compelling them to compete for lending business already adequately served by private capital sources and potentially making the multifamily mortgage industry more dependent on the Enterprises than is necessary.

FHFA has also concluded that, at this point in the growth of the Enterprises' multifamily businesses, the low-income housing goals should not be set at different levels for each Enterprise for the three-year goals period, because each Enterprise is expected to produce substantially the same loan volumes and unit counts and to have the same share of the multifamily market for mortgage purchases. The final rule sets the low-income goals at the same level for both Enterprises, based in part on FHFA's expectation that the Enterprises combined will comprise one-third to 40 percent of the estimated multifamily market for mortgage purchases for the three-year goals period.

Similarly, the final rule sets the very low-income goals at the same level for both Enterprises, under the assumption that the Enterprises will have similar shares of very low-income units and, thus, should have the same goal levels.

The policy advocacy groups' suggestion to re-establish lines of credit is not addressed in the final rule because that issue is beyond the scope of this specific rulemaking.

Regarding the recommendation on financing properties in certain geographic areas, FHFA will monitor the geographic distribution of the financing provided by the Enterprises to such properties.

As further discussed below, the final rule also changes several definitions to ensure that any rental unit claimed as goals-eligible is, in fact, a unit with affordable rents. These changes are expected, however, to have only a limited impact on the ability of the Enterprises to meet the 2015-2017 multifamily housing goals because they make up only a small percentage of very low- and low-income units financed by the Enterprises.

VIII. New Low-Income Housing Subgoal for Small Multifamily Properties

A. Small Multifamily Housing Subgoal Benchmark Levels in Final Rule—§ 1282.13(d)

The Enterprises have played a relatively limited role in supporting financing for small multifamily properties with 5 to 50 units. The proposed rule included establishment of a new subgoal for Enterprise purchases of mortgages on small multifamily properties with units affordable to low-income families. Based on FHFA's consideration of each of the applicable statutory factors, as well as the comments received on the proposed subgoal, § 1282.13(d) of the final rule establishes a new subgoal for Enterprise purchases of mortgages on small multifamily properties for low-income families. For both Fannie Mae and Freddie Mac, the benchmark levels in the final rule for this subgoal are 6,000 low-income units for 2015; 8,000 such units for 2016; and 10,000 such units for 2017. The benchmark levels in the final rule are generally lower than the levels in the proposed rule for Freddie Mac and substantially lower for Fannie Mae. Recent surveys indicate that there is currently ample liquidity available to small property owners, mainly through local banks and thrifts.^[45] Increasing the small multifamily goals to the levels in the proposed rule risks the Enterprises “crowding out” smaller lenders. Nonetheless, market conditions can change and both Enterprises must have the capability to serve the small multifamily market during stressed market conditions. The small multifamily goals are modest, but are intended to keep the Enterprises active in this market segment. Consistent with the proposed rule, the final rule defines “small multifamily property” as a property with 5 to 50 units.^[46] The new small multifamily properties subgoal will provide an additional incentive for the Enterprises to support these properties, which are an important source of affordable rental housing.

The applicable statutory factors, comments received, and analysis supporting these benchmark levels are discussed below.

B. Factors Considered in Setting the Small Multifamily Housing Subgoal Benchmark Levels

The Safety and Soundness Act provides that the Enterprises must report to FHFA on their purchases of mortgages on small multifamily properties with units affordable to low-income families, which may be defined as multifamily properties with 5 to 50 units (as such numbers may be adjusted by FHFA), or as mortgages of up to $5 million (as such amount may be adjusted by FHFA).^[47] These purchases (based on units) are included in the quarterly and annual activities reports published by the Enterprises. The Safety and Soundness Act further provides that FHFA may, by regulation, establish additional requirements related to such units.^[48] The statutory language, thus, provides FHFA with discretion to define small multifamily properties as those containing 5 to 50 units and to include in the rule a low-income families subgoal for small multifamily properties. FHFA has not established a subgoal for affordable small multifamily properties in previous rules.

The Safety and Soundness Act requires FHFA to consider the same six factors in setting a low-income housing subgoal for small multifamily properties as are considered in setting the multifamily low-income and very low-income housing goals:

1. National multifamily mortgage credit needs;

2. Past performance of the Enterprises;

3. Multifamily mortgage market size;

4. Ability to lead the market;

5. Availability of public subsidies; and

6. The need to maintain the sound financial condition of the Enterprises.^[49]

FHFA has considered each of these six factors in setting the benchmark levels for the low-income housing subgoal for small multifamily properties, as further discussed below.

C. Analysis of Considerations in Setting the Small Multifamily Housing Subgoal Benchmark Levels

1. Size of the Small Multifamily Mortgage Market (Factor 3)

Limited data is available on the overall size of the market for mortgages on small multifamily properties. Market data is generally reported based on loan balances rather than by property size, which necessitates using loan balances to estimate the size of the market for smaller properties with 5 to 50 units. Although using loan balances between $1 million and $3 million will include some smaller balance loans on larger properties and will exclude some larger loans on smaller properties, it can provide a reasonable estimate of the size of the mortgage market for properties with 5 to 50 units.

According to data from the Mortgage Bankers Association, the volume of multifamily loans with balances from $1 million to $3 million originated in 2006 and 2007 was just over $34 billion each year. These volumes declined significantly in 2008 through 2010, to as low as $8 billion in 2009, but have increased steadily since 2010, reaching $34 billion again in 2012, representing almost 25 percent of all multifamily loans by loan volume originated in 2012.

These trends in origination volumes have followed a similar pattern to those for the overall multifamily mortgage market, where volumes increased starting in 2014 and are expected to continue to increase through 2017 for both the overall market and for the segment consisting of loans with balances between $1 million and $3 million.

2. National Multifamily Mortgage Credit Needs (Factor 1)

Small multifamily properties have different operating and ownership characteristics than larger properties and as a result have different financing needs.^[50] 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. As a result, these properties are more likely to have informal documentation of the property's financial and other operating records, which can make it more difficult for property owners to obtain financing from some sources, including from the Enterprises.

Small multifamily properties also are often older than larger properties, have fewer, if any, amenities, and tend to have more affordable rents. As a result, small multifamily properties are likely to generate less revenue per unit than larger properties and support less leverage. While these factors make small multifamily properties an important source of affordable rental housing, they can also make financing more difficult to obtain. However, FHFA does not have any data showing that small multifamily property owners' financing needs are not currently being met or that there are liquidity gaps in this segment of the market.

3. Past Performance of Enterprises (Factor 2)

The Enterprises have played a relatively limited role in supporting financing for small multifamily properties, a role that is significantly smaller than their role in the multifamily market overall. In fact, small multifamily properties accounted for less than three percent of the total multifamily units financed by Fannie Mae in 2013, and less than one percent of the total multifamily units financed by Freddie Mac, even though the total small multifamily market comprises approximately 25 percent to one-third of the overall multifamily market.

While it appears that, currently, the small multifamily property finance sector has ample liquidity, primarily from community and larger banks, and that property owners' financing needs are largely being met, the Enterprises' loan products provide borrowers the option of longer, fixed rate loan terms and lower financing costs than other sources of financing, which are important features to some small property owners. Fixed rate financing provides borrowers with a predictable monthly mortgage payment for a longer period, as compared to alternatives such as adjustable rate mortgages or short-term loans with balloon payments, and can lock in lower, predictable mortgage costs that may result in less pressure to raise rents for low-income tenants.

Fannie Mae's purchases of mortgages financing low-income units in small multifamily properties were significantly greater in the years before the mortgage crisis than in subsequent years. Fannie Mae financed at least 40,000 low-income units in small multifamily properties each year between 2006 and 2008, peaking at 59,015 units in 2007, with much of this volume generated through loan pool purchases. Fannie Mae financed 12,552 low-income units in small multifamily properties in 2010, 13,480 such units in 2011, 16,801 such units in 2012, 13,827 such units in 2013, but only 6,732 such units in 2014.

Freddie Mac has played a much smaller role than Fannie Mae in this market, financing 459 low-income units in small multifamily properties in 2010, 691 such units in 2011, 829 such units in 2012, 1,128 such units in 2013, and 2,076 such units in 2014. Table 9 shows the number of low-income units in small multifamily properties financed by mortgages purchased by the Enterprises in 2006-2014.

Table 9—Enterprise Funding of Low-Income Units in Small Multifamily Properties, 2006-14

[“Small multifamily properties” are those with 5-50 units]

Year	Fannie Mae			Freddie Mac
	LI Units	Total units	Low-income (%)	LI Units	Total units	Low-income (%)
2014	6,732	11,880	56.7%	2,076	4,659	44.6%
2013	13,827	21,764	63.5%	1,128	2,375	47.5%
2012	16,801	26,479	63.5%	829	2,194	37.8%
2011	13,480	22,382	60.2%	691	2,173	31.8%
2010	12,552	20,810	60.3%	459	1,978	23.2%
2009	13,466	21,934	61.4%	528	1,619	32.6%
2008	43,782	82,706	52.9%	1,879	3,391	55.4%
2007	59,015	111,221	53.1%	2,147	3,522	61.0%
2006	40,631	60,174	67.5%	773	1,467	52.7%
Source: Funding as reported by the Enterprises for 2014; as calculated by FHFA for 2006-13.“Low-income” refers to units affordable to renters with incomes no greater than 80 percent of Area Median Income (AMI), based on rental proxy.
Note: Figures do not include units financed by the purchase of commercial mortgage-backed securities (CMBS).

4. Ability of the Enterprises To Lead the Market in Making Small Multifamily Mortgage Credit Available (Factor 4)

In setting the benchmark level for the low-income housing subgoal for small multifamily properties, FHFA considered the ability of the Enterprises to lead the market in making mortgage credit available. As discussed above, the Enterprises have played a smaller role in the small multifamily property mortgage market than in the overall market. The low-income housing subgoal for small multifamily properties will encourage the Enterprises to increase their participation in this market segment. It will also assure that the Enterprises and their lenders maintain an ongoing presence in the small multifamily property mortgage market so their role could be increased if there is a future financial crisis and other participating lenders withdraw from the market. FHFA will continue to assess the impact of Enterprise participation in the small multifamily property mortgage market and could adjust the benchmark levels for this subgoal as necessary.

5. Availability of Public Subsidies (Factor 5)

According to Rental Housing Finance Survey data, the availability of public subsidies for small multifamily properties is primarily through Section 8 rental assistance vouchers, although the data also show that small multifamily properties are less likely to contain subsidized rental units than larger multifamily properties.^[51] As discussed above, this is at least in part due to the fact that market rents in small multifamily properties are more likely to be affordable to low- and moderate-income families without needing to use rental subsidies.

6. Need To Maintain Sound Financial Condition of the Enterprises (Factor 6)

In setting the benchmark level for the low-income housing subgoal for small multifamily properties, FHFA also considered the importance of maintaining the Enterprises in sound and solvent financial condition. The delinquency rates for Fannie Mae's overall multifamily loan purchases are very low, as are the delinquency rates for the subset of those loans financing small properties. There is less data available on the performance of loans on small multifamily properties held by banks and thrifts, since detailed reporting data is not available or is combined with reporting on other income-producing properties. However, there is no evidence to suggest that increasing the Enterprises' purchases of loans on small multifamily properties will affect the Enterprises' financial conditions or negatively impact the performance of their loan portfolios as long as prudent underwriting judgments about such loans continue to be made.

FHFA will continue to monitor the activities of the Enterprises, both in FHFA's capacities as safety and soundness regulator and as conservator. If necessary, FHFA could make appropriate changes in the benchmark levels for this subgoal to ensure their continued safety and soundness.

Analysis

The primary benefit of increased purchases of loans on small multifamily properties by the Enterprises is to provide borrowers the opportunity to obtain longer-term, fixed rate financing at relatively low interest rates. Owners of small multifamily properties are more likely to have an adjustable rate mortgage or short-term loans with balloon payments than are owners of large properties.^[52] Adjustable rate mortgages usually have terms ranging from 1 to 5 years, with frequent rate adjustments based on changes to the LIBOR index, while balloon mortgages must be paid off or refinanced after a specific time period, often after five years. Further, during periods of financial instability, small property owners may be left with few, if any, sources of mortgage credit. By further addressing this financing need, the Enterprises would bring to small multifamily property owners the same benefits they provide to large multifamily property owners: Lower fixed interest rates, longer loan terms, and continued liquidity during periods of financial instability.

In setting the benchmark levels for the small multifamily property subgoal, FHFA considered the limited role the Enterprises have played in this market and the challenges of financing small multifamily properties, including a lack of standardization in this asset class, which can make the credit risk of small loans more difficult and time-consuming to assess. The mortgage origination process can be more costly, and it may be difficult to include small loans in securitizations for sale to investors. While small multifamily properties tend to have more affordable rents than larger properties, it is less profitable for the Enterprises' lenders to originate and service small loans. As a result, many small property lenders are banks that maintain a retail presence in the communities where properties are located and that can originate small loans for portfolio without securitizing them.^[53]

The challenges of supporting mortgage lending for small multifamily properties are even greater for properties with 24 or fewer units than for properties with 25 to 50 units. While the subgoal includes all properties with 5 to 50 units, FHFA expects that most Enterprise purchases of mortgages on small multifamily properties will be for properties with 25 to 50 units. The 2012 Rental Housing Finance Survey provides information on the characteristics of multifamily properties that have 5 to 24 units and properties that have 25 to 49 units.^[54] Multifamily properties with 25 to 49 units, unlike 5 to 24 unit properties, have operating characteristics that are similar to those of 50+ unit properties. For example, 25 to 49 unit properties and 50+ unit properties are more likely to be operated by a third party property management firm, have a mortgage, and be newer than 5 to 24 unit properties. The Enterprises should be able to provide additional liquidity to the 25 to 50 unit properties in light of the similarities of this property group to larger multifamily properties. In fact, data provided by Fannie Mae show that about 73 percent of all small multifamily units it financed in 2013 were in 25 to 50 unit properties.

For both Fannie Mae and Freddie Mac, the benchmark levels in the final rule for this subgoal are 6,000 low-income units for 2015, 8,000 such units for 2016, and 10,000 such units for 2017. These benchmark levels are generally lower than the levels in the proposed rule for Freddie Mac and substantially lower than the proposed benchmark levels for Fannie Mae.

By setting relatively low benchmark levels initially in the final rule, FHFA will have an opportunity to assess the impact of the new subgoal. For example, if there is unmet demand for alternative lending products, it is possible that additional support from the Enterprises could result in a wider array of long-term, fixed rate financing options for small multifamily property borrowers, with better mortgage terms (such as 10-year fixed rate loans) and lower financing costs than other sources of financing. These savings would lock in lower borrowing expenses for a multi-year period and may result in lower and more stable rents for low-income tenants. On the other hand, if the current market for lending to small multifamily properties is providing adequate long-term, fixed rate financing options for small multifamily property owners and investors, it is possible that the Enterprises would simply be competing on the same terms with existing sources of liquidity for small multifamily properties.

In addition, the Enterprises will be poised to quickly expand their financing activities in the event of a future financial crisis and a withdrawal from this market by other lending sources, such as commercial banks. Without having already established an ongoing market presence in this segment, including engaging the Enterprises' lender base in offering this financing, the Enterprises' programs would be unable to expand quickly when needed.

Comments on Proposed Rule

Most commenters on the proposed new small multifamily subgoal supported establishment of the subgoal. A trade association commenter noted that small multifamily properties play a key role in efforts to provide affordable housing in rural and other less densely populated areas, but that it is often difficult for developers to secure financing for such properties. Comments from a trade association and from policy advocacy groups urged FHFA to monitor developments in the small multifamily market and consider increasing the benchmark levels if market dynamics and the Enterprises' activities and capabilities justify such an increase. The commenters stated that the new subgoal will push the Enterprises to further innovate their approaches to the small multifamily market. The trade association commenter stated that the new subgoal would be an important step toward improving access to affordable, fixed rate financing, which the commenter stated is an urgent need for small multifamily units. Freddie Mac also supported establishment of the subgoal.

A trade association commenter stated that the proposed benchmark levels for the subgoal are high relative to the recent activity of the Enterprises in the small multifamily property market and other capital sources active in the market. The commenter cautioned that if the benchmark levels pressure the Enterprises to be overly aggressive in competing in the small multifamily market, it could result in a shift toward greater government-sponsored financing in this market, rather than promoting liquidity in other markets with substantial scarcity of capital.

A number of commenters suggested that the benchmark levels should increase more gradually from year to year. A trade association commenter noted that the Enterprises, especially Freddie Mac, may need more time to ramp up their small multifamily mortgage programs and that FHFA should consider this in setting the benchmark levels.

Another trade association commenter recommended increasing the proposed benchmark levels in order to promote readily available, consistently-priced, long-term credit. The commenter noted that the proposed levels are a relatively small percentage of the Enterprises' total low-income units. The commenter cited the lack of a functioning secondary market for 5 to 50 unit properties and that nearly three-fourths of small rental properties are affordable to very low-income households without government assistance.

A comment from an academic stated that more research is needed before FHFA makes a decision on establishing a small multifamily low-income subgoal. The comment noted that mortgages on small multifamily properties have significantly higher origination costs compared to large properties, since fixed origination costs are spread over fewer units. The comment stated that it is more efficient for the Enterprises to finance large properties than small properties.

Fannie Mae recommended that FHFA either delay implementation of the small multifamily subgoal to conduct further inquiry and analysis or significantly reduce the proposed benchmark level. Fannie Mae stated that existing data and information are insufficient to establish appropriate benchmark levels. Fannie Mae stated that it has been a leader in financing 5 to 50 unit small properties, notwithstanding the challenges inherent in such financings. Fannie Mae noted, however, that given the challenges with the data, it is difficult for it to fully evaluate the proposed subgoal benchmark levels, stating that the proposed level of 20,000 units for 2015 is likely to be 40 to 50 percent higher than Fannie Mae's own projections for 2015 based on current production.

Fannie Mae commented that it did not believe it would be able to meet the proposed benchmark levels solely through its Delegated Underwriting and Servicing (DUS) flow business. In addition, Fannie Mae stated that it is unclear whether the proposed benchmark levels could be met without re-entering the pools purchase business, which involves acquisition of an aggregation of seasoned permanent mortgages on multifamily rental properties from another lender. Fannie Mae stated that it made such pool acquisitions most recently in 2006-2008, but has not engaged in these transactions since then.

A trade association commenter expressed concerns over the impact of more Enterprise competition in the small multifamily market on smaller lenders. The commenter stated that small lenders may not be able to compete on price given the lower borrowing costs for the Enterprises. In addition, the Enterprises only make non-recourse loans, while small lenders almost always require recourse.

FHFA Response

Regardless of the level of support for this market segment from the secondary market, FHFA does not have any recent evidence of illiquidity or a lack of financing availability in the small multifamily property segment. Further, in spite of the limited empirical data that is currently available about the small multifamily property market, FHFA has determined that the data is sufficient for it to assess the statutory factors used to determine the benchmark levels and has set the benchmark levels in the final rule primarily based on the Enterprises' past and current histories of serving this market segment.

FHFA realizes that both Enterprises, and especially Freddie Mac, have limited experience in purchasing loans on small multifamily properties. The final rule establishes lower benchmark levels for Fannie Mae than the levels in the proposed rule due to the significant drop in small multifamily units Fannie Mae financed in 2014 compared to the levels it financed over previous years, and due to an apparent abundance of capital sources serving this segment of the multifamily market. These final lower benchmark levels should be achievable by Fannie Mae without needing to re-enter the pool purchase business. Consistent with the other multifamily benchmark levels set in this final rule, since Fannie Mae and Freddie Mac are expected to have the same loan volume during the three-year goals period, Fannie Mae will be expected to purchase the same volume of loans on small multifamily properties as does Freddie Mac, with both Enterprises being held to the same benchmark levels during that time.

The benchmark levels for Freddie Mac in the final rule are modest in volume due to Freddie Mac's limited experience in purchasing loans on small multifamily properties, but increase each year of the three-year goals period commensurate with Freddie Mac's projected increase in loan volume to this market segment.

As discussed above, while it appears that currently the small multifamily property finance sector has ample liquidity, primarily from community and larger banks, and that small multifamily property owners' financing needs are largely being met, the Enterprises' loan products could provide small multifamily property borrowers the option of longer, fixed rate loan terms and lower financing costs than other sources of financing.

FHFA also believes that,_ in light of the subgoal's relatively low benchmark levels in the final rule, the Enterprises will not take significant business away from local banks and thrifts.

A trade association commenter cited challenges facing implementation of a small multifamily mortgage program. Another trade association commenter noted high costs and credit risks of small multifamily lending. Comments from policy advocacy groups and a mission-oriented housing developer noted some of the risks of small multifamily lending including: Disparate borrowers; lack of standardization in underwriting, originating, and servicing, which makes financing more expensive and limits secondary market participation; and large fluctuations in property financial performance. Commenters recommended consideration of these factors in setting the benchmark levels and close monitoring by FHFA of the Enterprises' small multifamily mortgage purchases due to these challenges.

FHFA Response

FHFA has considered the factors pointed out by the commenters but believes that the Enterprises will be able to effectively manage the risks and any additional fixed costs associated with purchasing loans on small multifamily properties, and FHFA will closely monitor the Enterprises' participation in this market segment.

A trade association commenter expressed concern that the Enterprises would concentrate their loan purchases on 25 to 49 unit properties rather than the more numerous and more affordable 5 to 24 unit properties. The commenter noted that existing sources of liquidity for small multifamily properties, especially properties with fewer than 25 units, are not sufficient to meet the needs of the market and that the Enterprises could play a much larger role in supporting those segments of the market. The commenter stated that the Enterprises have not provided sufficient support for small multifamily properties, instead focusing on buildings with more than 50 units. The commenter noted that Fannie Mae has stated that nearly half of its small loan book of business is concentrated in just two MSAs, New York and Los Angeles and recommended that FHFA require the Enterprises to issue annual reports detailing the composition of the Enterprises' multifamily lending portfolios to show how the Enterprises are meeting the goals.

FHFA Response

As noted previously, no evidence has been presented of illiquidity or a lack of financing availability in the small multifamily property segment for properties with fewer than 25 units.

With respect to the proposed definition of “small multifamily property,” the Safety and Soundness Act provides FHFA with discretion to define “small multifamily property” either in terms of the number of units in the property or in terms of the size of the loan.^[55] Both Enterprises commented that the proposed definition of 5-50 units is different from the definitions used and reported by the Enterprises for their respective small loan products, both of which are based on the unpaid principal balance of the loan. Fannie Mae noted that the Mortgage Bankers Association also uses loan balances. Freddie Mac recommended that FHFA define “small multifamily property” as either properties with 5 to 50 units or a loan balance of up to $5 million. Freddie Mac stated that this definition would be consistent with the Safety and Soundness Act language and would facilitate the use of more accurate data in market size estimations for purposes of evaluating the appropriate levels for the small multifamily housing subgoal.

FHFA Response

FHFA has decided in the final rule not to define “small multifamily property” using loan amount, because some larger multifamily properties with more than 50 units may obtain low-leverage financing, meaning the Enterprise loan is small but the property securing the loan is not. Including smaller loans on larger properties would tend to overstate the level of support that the Enterprises provide for small multifamily properties.

Modifications of Multifamily Mortgages

Freddie Mac also recommended that modifications of multifamily mortgages be treated as mortgage purchases for purposes of the housing goals. Freddie Mac stated that such modifications mitigate risk and the adverse impacts of foreclosure, thereby benefiting tenants by preventing disinvestment, maintaining building services, and helping avoid destabilizing the surrounding community.

FHFA Response

FHFA agrees that for troubled multifamily properties at risk of default, loan modifications, which may split a loan into supportable and cash flow only payments and/or reduce the loan interest rate, are effective means of avoiding foreclosure and the potentially negative effects on tenants and communities. Indeed, these risk mitigation tools are already in wide use by the Enterprises and are their primary tools to address, stabilize, and resolve troubled multifamily assets and avoid foreclosure and further losses. However, Freddie Mac did not offer any reasons why loan modifications should be counted the same as new loan acquisitions for purposes of providing housing goals credit. Because simply modifying an existing loan on an existing Enterprise-financed property that has already been counted towards the housing goals does not represent a new loan on a property that was not previously financed, FHFA has determined that there is no reason to provide housing goals credit for such loan modifications. Although FHFA counts income-eligible single-family HAMP modifications as refinancing mortgages for purposes of the single-family housing goals, it began doing so to encourage the Enterprises to engage fully in that program. The same rationale is not applicable to modifications of multifamily mortgages.

IX. Section-by-Section Analysis of Other Changes in Final Rule

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

A. Changes to Definitions—§ 1282.1

The final rule makes changes to definitions used in the current housing goals regulation, including: (1) Definitions related to rent and utilities; (2) the definition of “dwelling unit;” (3) technical definition changes; and (4) other changes to definitions. The changes are discussed below.

1. Definitions Related to Rent and Utilities

Rents are used to determine the affordability of a unit for purposes of counting under the housing goals. Consistent with the proposed rule, the final rule consolidates and simplifies several terms related to rents that are defined separately in the current regulation. Specifically, the final rule deletes the separate definitions of “contract rent” and “utility allowance,” with the substance of those definitions included in a revised definition of “rent.”

“Rent” is defined generally in the final rule as the actual rent, or the average rent by unit size, for a dwelling unit. The rent is to be determined by the Enterprises based on the total combined rent for all bedrooms in the dwelling unit including fees or charges for management and maintenance services and any included utility charges. Where the rent does not also include all utilities provided to the unit, then “rent” also includes either the actual cost of utilities not included in the rent or a utility allowance, which is further discussed below.

Comments on Proposed Rule

Two policy advocacy groups supported clarification of the definition of “rent” as proposed.

Freddie Mac recommended that the definition of “rent” be revised to delete the proposed requirement that rent reflect the total combined rent for all bedrooms in the dwelling unit because in certain circumstances, such as student housing, there is a separate lease for each room in a unit and the combined rent of each room may not be equal to the rent if all four bedrooms were rented out under one lease. This aspect of the definition of “rent” relates to the more general issue regarding the definition of “dwelling unit,” which is discussed in more detail below in the context of the definition of “dwelling unit.”

FHFA Response

The final rule maintains the proposed requirement that rents for individual bedrooms in a dwelling unit be combined for purposes of determining the affordability of the dwelling unit in shared living arrangements. This requirement mirrors the revised definition of “dwelling unit” under the final rule, which generally does not permit individual bedrooms in a single living space to be treated as separate units for purposes of the housing goals.

Sources of Information for Determining Utility Allowances

The final rule expands the sources of information that may be used by an Enterprise for determining the utility allowance. Specifically, consistent with the proposed rule, the final rule allows an Enterprise to use the utility allowance established by a state or local housing finance agency that is used in determining the affordability of low-income housing tax credit (LIHTC) properties for the area where the property is located.

The current regulation requires the Enterprises to take into account the cost of utilities for rental units in determining affordability for purposes of the housing goals. The definition of “rent” provides that if the rent includes all utilities, the Enterprises must use that rent to determine affordability. If the rent does not include all utilities, then the Enterprises may use either: (a) Data on the actual cost of utilities paid by individual tenants but not included in the rent; or (b) a “utility allowance.” The current definition of “utility allowance” allows the use of either a nationwide average utility allowance provided by FHFA or the utility allowances issued by the U.S. Department of Housing and Urban Development (HUD), the Enterprises' former mission regulator, under the Section 8 Program for the area where the property is located. The expanded definition of “utility allowance” in the final rule will allow the Enterprises to use the same utility allowance data that is used in the administration of the LIHTC program and will facilitate alignment in determining affordability for such units.

Comments on Proposed Rule

A comment, signed by several members of Congress, stated that the proposed new source for calculating the utility allowance is acceptable and appropriate.

Freddie Mac recommended that the Enterprises also be permitted to use a fixed 8 percent of the rent as a proxy for utility costs. Freddie Mac stated that while the alternatives in the proposed rule for calculating utility allowances would more accurately reflect actual utility costs, it would be an administrative burden to implement. Freddie Mac also provided data on average operating expenses and utilities from the “2013 Survey of Operating Income and Expense in Rental Apartment Communities.” Based on that data, Freddie Mac suggested that the Enterprises be permitted to calculate the utility allowance as a fixed 8 percent of the rent.

FHFA Response

In order to provide additional flexibility in determining accurate rent levels that better reflect local and regional differences in utility costs, the final rule expands the permitted ways to determine the utility allowance as discussed above. The Enterprises will continue to have the option to use the nationwide average utility allowance provided by FHFA or the utility allowance established under the HUD Section 8 Program.

While the final rule does not adopt the alternative measure for determining utility allowances proposed by Freddie Mac, FHFA notes that the proposed and final rule language regarding the nationwide average utility allowances does not specify the sole method by which FHFA will determine such allowances. The current nationwide average utility allowances are fixed numbers based on data from the American Housing Survey, but the regulation is sufficiently broad to allow FHFA to adopt the measure proposed by Freddie Mac at a future date, without changing the regulation itself, if it chooses to do so.

Nationwide Average Utility Allowances

In the Notice accompanying the proposed rule, FHFA noted that it planned to issue updated figures for the nationwide average utility allowances as more recent American Housing Survey data becomes available. FHFA is providing updated figures to the Enterprises by letters, which will be posted on FHFA's Web site. These revised nationwide average utility allowances are based on the most recent American Housing Survey data available, as follows:

Type of property	Number of Bedrooms
	Efficiency	1	2	3 or more
Multifamily	$50	$77	$110	$149
Single-family	$70	$111	$161	$219

Definition of “Rental Unit”

Consistent with the proposed rule, the final rule streamlines the current regulation by deleting the term “rental housing” in § 1282.1, and replacing this term in § 1282.17 with the term “rental units,” the only other place in the regulation where the term “rental housing” appears.

Definition of “Utilities”

Consistent with the proposed rule, the final rule revises the existing definition of “utilities” to expand the list of excluded services. The current regulation excludes charges for cable and telephone services from the definition of “utilities.” The revised definition also excludes all subscription-based television, telephone and internet services (regardless of whether provided by a cable provider or other provider).

2. Definition of “Dwelling Unit”—Shared Living Arrangements

The final rule revises the current definition of “dwelling unit” by limiting the definition to include only units with plumbing and kitchen facilities. Section 1282.1 of the current regulation defines “dwelling unit” as “a room or unified combination of rooms intended for use, in whole or in part, as a dwelling by one or more persons, and includes a dwelling unit in a single-family property, multifamily property, or other residential or mixed-use property.” The proposed rule would have added a provision limiting the definition to units with complete plumbing and kitchen facilities. After considering the comments on the proposed change, the final rule adopts this limitation but omits the word “complete,” to ensure that FHFA retains flexibility, if necessary, to provide more specific guidance on specific classes of transactions in the future.

Limiting the definition of “dwelling unit” to units with plumbing and kitchen facilities is intended to address shared living arrangements where separate individuals rent separate bedrooms but share common areas and cooking and sanitary facilities. The final rule does this by providing that a unified combination of rooms will be treated as a single dwelling unit, regardless of whether there are individual leases for the separate bedrooms in the unit, if the rooms share plumbing and kitchen facilities. FHFA may provide additional guidance regarding whether particular types of housing should be counted as separate dwelling units despite the limitation added by this final rule.

Comments on Proposed Rule

One comment letter, signed by several members of Congress, supported the proposed change to the definition of “dwelling unit,” stating that it makes sense to count a unit as a single unit no matter how many bedrooms it has.

Fannie Mae agreed with the new definition but recommended that seniors housing units that lack a full kitchen (e.g., kitchenettes) or have no cooking facilities in the units due to safety concerns, such as in seniors housing Alzheimer's units, be considered “dwelling units” for housing goals purposes.

Freddie Mac opposed the proposed revision to the definition of “dwelling unit,” stating that the change may restrict the availability of safe, affordable housing for seniors and students, and could impact single-room occupancy (SRO) living space. Freddie Mac noted that shared living arrangements represent an important segment of the affordable housing market and are often used by unrelated persons who live together due to a lack of affordable housing alternatives. Freddie Mac also noted that the availability of affordable housing for students is becoming increasingly important as the costs of higher education continue to rise. Freddie Mac recommended that a bedroom rented to a tenant pursuant to a separate and independent lease be counted as a separate dwelling unit for purposes of the housing goals. Freddie Mac also suggested alternative criteria that could be used to limit potential “over-counting” of individual rooms in a single dwelling: Whether there are separate and independent leases; whether a separate rent amount is identifiable and reported; and/or whether each bedroom has a separate entrance and lock.

FHFA Response

FHFA has decided to adopt the revised definition of “dwelling unit” as proposed, with one change as described above. Under the final rule, bedrooms sharing the same plumbing and kitchen facilities will be treated as a single dwelling unit for housing goals counting purposes. For example, four individuals living in a shared living arrangement with separate bedrooms but with shared bathrooms and kitchen would be considered a single dwelling unit with four bedrooms rather than four efficiency units. For purposes of determining affordability under the housing goals, the rent for the dwelling unit would be the aggregate of all rent payments made by all of the individuals residing in the dwelling unit, even if each individual who resides in a bedroom has entered into a separate lease agreement or if the bedrooms have separate locks.

This change will also clarify the appropriate calculation of rent for dwelling units in student housing or other shared living arrangements in a single dwelling unit. Potential over-counting of such shared units under the housing goals can occur when the rent for each bedroom is calculated as if it were a separate unit. Thus, four bedrooms renting for $500 each could be considered affordable for housing goals purposes if they were considered efficiency units, but may not be affordable if they were considered a single four-bedroom unit renting for $2,000. To avoid potential over-counting of the Enterprises' housing goals performance, FHFA has decided to adopt the revised definition as proposed, except that the final rule omits the word “complete.”

FHFA recognizes that the Enterprises purchase mortgages secured by multifamily properties with a variety of different purposes and configurations. While the definition of “dwelling unit” will generally prevent an Enterprise from receiving credit under the housing goals for individual bedrooms that share the same plumbing and kitchen facilities, FHFA retains authority under § 1282.16(e) to determine how any class of transactions will be treated for purposes of the housing goals. FHFA may exercise this authority in the future to permit housing goals credit for particular types of housing, such as certain types of seniors housing or group housing for people with special needs, which may lack separate plumbing or kitchen facilities but that otherwise meet the criteria to be considered a separate dwelling unit. FHFA will provide any such guidance to the Enterprises, and post such guidance on FHFA's public Web site, in writing in accordance with the procedures in § 1282.16(e).

3. Technical Definition Changes

Consistent with the proposed rule, the final rule makes a number of technical changes to the existing definitions in § 1282.1. Specifically, the final rule removes two definitions that are not used anywhere in the current regulation, other than the definitions themselves: “HMDA” and “working day.” The final rule also revises the definition of “families in low-income areas” to remove the reference to “block numbering areas,” which conforms the words used in the definition to the terminology currently used by the U.S. Census Bureau. In addition, the final rule revises the existing definition of “HOEPA mortgage” to reflect renumbering in the statute cited in the definition.

Comments on Proposed Rule

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

4. Other Changes to Definitions

Other definitional changes in § 1282.1 are discussed below in the corresponding section dealing with the substantive provisions to which the definitions relate. These changes include: (i) Deleting the definitions of “mortgage with unacceptable terms or conditions” and “rental housing;” and (ii) adding a definition for “efficiency.” The definition of “small multifamily property” was discussed above under the section on the new small multifamily property subgoal.

B. General Counting Requirements—§ 1282.15

The final rule revises a number of provisions related to counting single-family owner-occupied units and rental units under the housing goals. Some provisions are being revised or eliminated because they are no longer necessary based on the affordability information that is available to the Enterprises. Other provisions are being amended or added in order to provide greater clarity and to minimize cases where a unit may be treated as affordable when it is not in fact affordable.

1. Use of Area Median Income at Single-Family Mortgage Loan Origination Date

Consistent with the proposed rule, the final rule revises current § 1282.15(b)(1) to provide that, for purposes of determining whether single-family mortgage loan purchases may be counted under a housing goal, the income of the mortgagors shall be determined based on the area median income as of the date the mortgage loan was originated, rather than as of the date of the mortgage loan application.

The data that is reported to the Enterprises typically includes an origination date, which is used by the Enterprises for purposes of determining affordability. This change conforms the regulatory language to the existing practice of the Enterprises.

Comments on Proposed Rule

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

2. Removal of Affordability Estimation Provision for Mortgages on Single-Family Owner-Occupied Units—§ 1282.15(b)

Consistent with the proposed rule, the final rule revises current § 1282.15(b) by removing the affordability estimation provisions in current paragraphs (b)(2) and (b)(3) for mortgages on single-family owner-occupied units where the borrower's income information is not available, and provides in § 1282.15(b)(2) that such mortgages may not be counted in the numerator but will still be included in the denominator for any of the housing goals.^[56] This change in the treatment of single-family mortgages with missing borrower income information is similar to the treatment of HOEPA loans under § 1282.16(d) and will continue to provide an incentive for the Enterprises to maintain their high rate of income data collection.

The current regulation allows the Enterprises to estimate affordability for single-family owner-occupied mortgages by multiplying the number of mortgage purchases with missing borrower income information in each census tract by the percentage of all single-family owner-occupied mortgage originations in the respective tracts that would count toward achievement of each housing goal, as determined by FHFA based on the most recent HMDA data available. The current regulation further provides that the estimation methodology may be used up to a nationwide maximum calculated by multiplying, for each census tract, the percentage of all single-family owner-occupied mortgage originations with missing borrower incomes (as determined by FHFA based on the most recent HMDA data available for home purchase and refinancing mortgages, respectively) by the number of Enterprise mortgage purchases secured by single-family owner-occupied properties for each census tract, summed up over all census tracts.

Comments on Proposed Rule

A housing advocacy group commenter agreed that mortgages with missing income data should not be included in the numerator for housing goals counting purposes. The final rule adopts the change as proposed.

3. Determination of Affordability of Rental Units Based on Rents, Not Incomes—§ 1282.15(d)(1)

Consistent with the proposed rule, the final rule revises current § 1282.15(d) to provide that, in determining whether rental units count under the housing goals, the affordability of a unit shall be determined based solely on the rent for the unit.

The current regulation provides that the affordability of rental units is to be determined based on the tenant's actual income, if available, and based on rents if the tenant's income is not available. Because lenders generally do not collect income information on tenants, the Enterprises use rents in all cases (except for certain seniors housing units) to determine affordability for purposes of the housing goals. The revision in the final rule to use rents, thus, conforms the counting rule to the Enterprises' actual practices and recognizes the general unavailability of actual tenant income data. The revision also more closely aligns the regulation's language with section 1333(c) of the Safety and Soundness Act, which provides that FHFA shall evaluate the performance of the Enterprises under the multifamily housing goals “based on whether the rent levels are affordable.” ^[57]

Section 1333(c) provides that to be counted as an affordable rent for purposes of the housing goals, a unit's rent may not exceed 30 percent of the maximum income level of very low- or low-income families, adjusted for the number of bedrooms in a unit.^[58] Section 1282.19 of the current regulation sets forth tables containing the applicable affordable amounts for each of the income categories targeted under the housing goals, adjusted for the number of bedrooms in a unit.

Comments on Proposed Rule

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

4. Reliance on Other Housing Program Affordability Restrictions for Determining Affordability of Rental Units—§ 1282.15(d)(2)

Consistent with the proposed rule, § 1282.15(d)(2) of the final rule adopts a new counting rule for rental units that are subject to affordability restrictions of local, state, or federal affordable housing programs, with a clarification regarding the applicable affordability restrictions. This provision is intended to ease the Enterprises' operational compliance requirements for determining affordability of units that are already required to be affordable under a separate governmental housing program.

The final rule permits an Enterprise to determine the affordability of rental units for housing goals purposes using the housing program's maximum permitted income level for a renter household or the maximum permitted rent for the units. Although affordability for a multifamily property is generally determined based solely on rent levels for each unit, the final rule permits rental units that are subject to affordability restrictions of local, state, or federal affordable housing programs to be counted assuming that the program restricts affordability based on tenant income or rent levels. The final rule clarifies that in order for a unit to be counted as affordable for purposes of the housing goals under a housing program with eligibility limits on income, the maximum income level for the unit under the program must be no greater than the maximum income level for the applicable family or unit size under each goal as set forth in § 1282.17 or § 1282.18, as appropriate. For a housing program with eligibility limits on rent, the maximum rent level for the unit under the program must be no greater than the maximum rent level for each goal, adjusted for unit size as set forth in § 1282.19.

If a property includes both units with affordability restrictions and units that are not restricted but that would nonetheless qualify as affordable, an Enterprise may only rely on the program restriction for purposes of determining affordability for the actual units that are restricted, with the affordability of the remainder of the units determined based on rent data.

An example of an applicable affordable housing program is the LIHTC program. LIHTC units restricted for occupancy by tenants with incomes at 50 percent of area median income and rents not exceeding 30 percent of tenant income, adjusted for bedroom count and household size, will receive credit toward the multifamily very low-income housing subgoal, and the Enterprises will not have to separately determine affordability for such units.

The Notice accompanying the proposed rule stated that the Enterprises must also confirm that the LIHTC or other monitoring entity that exercises compliance oversight over the property has determined that the units are in compliance with the program's affordability restrictions as to maximum tenant incomes or maximum permitted rents charged. FHFA expects the Enterprises to have appropriate procedures in place to ensure the accuracy and reliability of the information they report to FHFA regarding whether units meet the necessary criteria for counting under the housing goals. Therefore, the final rule does not include a specific requirement for the Enterprises to document compliance with the housing programs' affordability restrictions on maximum tenant incomes or rents. Confirming compliance with the affordability restrictions is a standard due diligence requirement imposed on lenders who are authorized to participate in the Enterprises' loan programs. In addition, LIHTC properties rarely go out of compliance with their affordability restrictions because of the potentially adverse tax consequences to investors. LIHTC properties are also subject to ongoing compliance monitoring by designated oversight agencies and other participants in the transaction.

Comments on Proposed Rule

Several housing advocacy groups, Fannie Mae and Freddie Mac supported the proposed new counting rule for properties with affordability restrictions on the basis that compliance with the restrictions is already monitored by a designated public agency and it would be redundant for the Enterprises to independently conduct such compliance monitoring themselves.

Fannie Mae recommended expanding the proposal to include limited equity cooperatives (where unit affordability is tied to limitations on the amount of equity shareholders may retain when they sell their cooperative shares) when the cooperative units are subject to rent and income restrictions that meet the affordability targets for low-income and very low-income families if the units are rented out. Fannie Mae noted that such properties are generally valued and the blanket loan is sized using unrestricted market rents. As a result, limited equity cooperatives that are subject to rent restrictions are generally not counted as affordable for housing goals purposes.

Freddie Mac recommended that the proposal be revised to allow an Enterprise to rely on a property owner's certification of compliance with the applicable income and rent restrictions, rather than having to obtain a certification from the housing program's monitoring entity. Freddie Mac stated that most housing programs that would qualify under the proposal rely on a property owner's certification of compliance.

A trade association commenter opposed the proposal, stating that it would undermine secondary market support for affordable housing by favoring financing of subsidized multifamily properties over affordable non-subsidized multifamily properties.

FHFA Response

Regarding counting rules for rental units in limited equity cooperatives, FHFA has determined that, because of the wide variance among cooperative bylaws that govern the types of rent and occupancy restrictions (if any) that may be imposed on cooperative owners who rent out their units, the counting rule described in this section will not apply to limited equity cooperatives. Instead, the Enterprises will follow the rule's requirements for determining the affordability of a particular cooperative unit's rent. If a limited equity cooperative's bylaws limit the rent and income of tenants who may occupy a cooperative unit at levels that would qualify for housing goals credit, then that can be recognized by the lender or the Enterprise when establishing the comparable rent for the unit, thereby receiving housing goals credit.

Regarding verification of compliance with regulatory agreements, as noted above, FHFA expects the Enterprises to have appropriate procedures in place to ensure the accuracy and reliability of the information they report to FHFA regarding whether units meet the necessary criteria for counting under the housing goals. FHFA agrees that certifications from property owners would be sufficient for purposes of verifying compliance with rent and income restrictions, but it is not necessary to include a specific provision regarding documentation in the regulation itself.

Regarding favoring financing for subsidized over affordable non-subsidized units, FHFA does not believe that allowing the Enterprises to rely on the income and rent compliance determinations of other affordable housing programs would necessarily mean that the Enterprises would, therefore, decide to purchase more loans on properties subsidized by such programs rather than purchasing loans on properties with similarly affordable market rents. Furthermore, the number of subsidized units available to finance is limited by the availability of housing subsidies, whereas the number of affordable market rate units is only limited by market conditions.

5. Counting Unoccupied Units—§ 1282.15(d)(3)

Consistent with the proposed rule, the final rule consolidates the current provisions related to unoccupied units, including model units and rental offices, into a single provision located at § 1282.15(d)(3). As under the current rule, § 1282.15(d)(3) of the final rule provides that a unit in a multifamily property that is unoccupied because it is being used as a model unit or rental office may be counted for purposes of the multifamily housing goals and subgoals only if an Enterprise determines that the number of such units is reasonable and minimal considering the size of the multifamily property. The method for determining affordability for such units is found in the definition of “contract rent” under § 1282.1 of the current regulation.

Consistent with the current regulation, § 1282.15(d)(3) of the final rule also provides that anticipated rent for unoccupied units may be the market rent for similar units in the neighborhood as determined by the lender or appraiser for underwriting purposes.

Comments on Proposed Rule

FHFA did not receive any comments on the proposed changes, and the final rule adopts the changes as proposed.

6. Missing Bedroom Data for Rental Units—§ 1282.15(e)(1)

Consistent with the proposed rule, the final rule revises § 1282.15(e)(1) to provide that a rental unit for which the number of bedrooms is missing shall be considered an efficiency unit for purposes of calculating unit affordability. This provision is moved here from current § 1282.19(f) so that all provisions on missing information are included in the same section of the regulation, and as a result the final rule deletes the current § 1282.19(f). Consistent with the proposed rule, § 1282.1 of the final rule adds a definition for “efficiency” to mean a dwelling unit having no separate bedrooms or 0 bedrooms.

Under § 1282.15(d)(1), the affordability of a rental unit is calculated taking into account adjustment for the unit size under § 1282.19 based on the number of bedrooms in the unit. However, this adjustment is not possible when data on the number of bedrooms is unavailable. Because the Enterprise will have in fact purchased a mortgage secured by the rental unit, consistent with the current regulation, the final rule allows the unit to count towards the housing goals if it qualifies for the lowest-rent unit permitted to receive goals credit under the rule, i.e., as an efficiency.

Comments on Proposed Rule

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

7. Reduction in Cap on Estimating Affordability for Rental Units—§ 1282.15(e)(2)

Consistent with the proposed rule, the final rule revises current § 1282.15(e)(2) to reduce the cap for the number of rental units for which an Enterprise may estimate the rent from 10 percent to 5 percent of the total number of rental units in properties securing multifamily mortgages purchased by the Enterprise in the current year. The final rule does not adopt the proposal to count seniors housing units where additional services are included in the rent toward the 5 percent cap, so such units will continue to be excluded from the cap as under their current treatment. The purpose of lowering the estimation cap to 5 percent is to provide an incentive for the Enterprises to collect rent information for their multifamily mortgage purchases.

Under the current regulation, an Enterprise is permitted to use estimated rent for purposes of determining affordability of a rental unit where both income and rent information are unavailable. The current regulation allows the Enterprises to estimate affordability by multiplying the number of rental units with missing affordability information in each census tract by the percentage of all rental units in the respective tracts that would count toward achievement of each goal and subgoal, as determined by FHFA based on the most recent decennial census. The estimation methodology may currently be used up to a nationwide maximum of 10 percent of the total number of rental units in properties securing multifamily mortgages purchased by the Enterprise in the current year. Rental units in excess of this maximum percentage cap, and any units for which estimation information is not available, may not be counted for purposes of the multifamily housing goal and subgoal. The Enterprises have been permitted to estimate affordability for seniors housing units where additional services are included in the rent because of the difficulty of separating out the housing expenses from the non-housing related services in the rent amount, and those seniors housing units have been excluded from the maximum percentage cap.

As discussed above, under the final rule, the Enterprises will determine the affordability of rental units based on the rents, not on the income of the tenants. Missing rent data rates for multifamily mortgages purchased by the Enterprises are generally very low given the Enterprises' requirements for submission of underwriting and property level information from their lenders as of the date of mortgage acquisition. Historically, the Enterprises' affordability estimations have fallen below 5 percent for units subject to the rent estimation cap. In 2014, Fannie Mae estimated affordability for 5.5 percent of all rental units counted toward the multifamily low-income housing goal (3.8 percent of total acquisitions), but almost all of those units were either seniors housing units or in cooperative buildings and so were excluded from the rent estimation cap. Only 0.01 percent of Fannie Mae's total acquisitions in 2014 were missing data and subject to the rent estimation cap. Freddie Mac estimated affordability for 7.5 percent of rental units counted toward that goal in 2014 (5.6 percent of total acquisitions), but only 0.23 percent of its total acquisitions were subject to the rent estimation cap. In a change from the proposed rule, and consistent with current practice, FHFA has determined that seniors housing units where additional services are included in the rent should continue to be excluded from the affordability estimation cap because the purpose of the cap is to incentivize the Enterprises to obtain rent data but that data cannot be obtained for these seniors housing units because the housing and non-housing expenses are both included in a single rent payment. In addition, as discussed above, the final rule now permits the Enterprises to determine affordability based on the affordability restrictions imposed under other governmental housing programs, which will eliminate the need to estimate affordability in those cases and further lower the number of units counted towards the estimation cap.

In short, given the very few situations where estimation may be necessary, and the exclusion of seniors housing units with additional services included in the rent and subsidized properties with affordability restrictions from the estimation cap, lowering the cap to 5 percent is unlikely to have an impact on Enterprise performance under the multifamily goals as neither Enterprise is likely to exceed the cap. As a result, the final rule reduces the cap for the number of rental units for which an Enterprise may estimate the rent from 10 percent to 5 percent, as in the proposed rule.

Comments on Proposed Rule

Freddie Mac provided the only comment on this proposal. Freddie Mac recommended that the cap on estimating affordability for rental units remain at 10 percent. Freddie Mac stated that two of the other changes discussed in the proposed rule—counting seniors housing units with additional services included in the rent towards the cap and providing goals credit for Enterprise purchases of blanket loans on manufactured housing communities—would increase the number of rental units for which estimation is needed, making it more likely that an Enterprise might reach the cap.

FHFA Response

Separate from and prior to this rulemaking, FHFA has provided guidance to the Enterprises on the appropriate treatment under the housing goals for both seniors housing units and blanket loans on manufactured housing communities. As discussed in more detail in the appropriate section on each issue, the final rule does not make any change to the counting rules treatment for either seniors housing units or blanket loans on manufactured housing communities. As a result, neither seniors housing units nor blanket loans on manufactured housing communities will have any impact on the number of rental units for which estimation is needed.

8. Changes To Reflect U.S. Census Bureau Terminology—§ 1282.15(g)(2)

Consistent with the proposed rule, the final rule revises § 1282.15(g)(2) to eliminate outdated terminology used for purposes of determining split areas in which a dwelling unit is located in determining area median income for affordability determinations. Due to changes implemented by the U.S. Census Bureau, it is no longer necessary to include references to the “block-group enumeration district,” the “nine-digit zip code,” or other geographic divisions partially located in more than one area.

Comments on Proposed Rule

FHFA did not receive any comments on the proposed changes, and the final rule adopts the changes as proposed.

C. Determining Affordability for Blanket Loans on Cooperative Housing—§ 1282.16(c)(5)

The final rule revises § 1282.16(c)(5) to provide that the affordability of units securing a blanket loan on a cooperative property (i.e., a loan that is secured by the entire property) must be determined solely on the basis of comparable market rents that were used by the lender or the Enterprise in underwriting the blanket loan (“underwriting rents”). In response to a comment from Freddie Mac, the final rule permits an Enterprise to use its own underwriting rents, a change from the proposed rule which would have only allowed use of the lender's underwriting rents. If the underwriting rents are not available for the blanket loan on a cooperative property, the units may not be counted towards the multifamily housing goals. Determining affordability for blanket loans on cooperative housing based on the rent estimation methodology will no longer be permitted. Share loans used by residents to finance the purchase of a cooperative unit remain eligible for credit under the single-family housing goals even if the Enterprise also holds a blanket loan on the same cooperative property that may be eligible for multifamily housing goals credit.

As discussed above, the final rule revises § 1282.15(d)(1) to require the Enterprises to use rent levels to determine the affordability of rental units. In the case of blanket loans on housing cooperatives, there is no rent data available because all units are owned by the cooperative in which each unit resident owns shares, which allows the shareholder to occupy one or more units in the property. Shareholders pay a monthly fee to cover expenses for common area upkeep and maintenance and to pay their pro rata share of any blanket loan payments. In 2013, blanket loans on cooperative housing accounted for 2.7 percent and 1.4 percent of multifamily mortgages purchased by Fannie Mae and Freddie Mac, respectively.

Because of the absence of rental data for cooperatives, the Enterprises have used the estimated rent methodology under § 1282.15(e) discussed above to determine whether units in cooperatives count towards the multifamily housing goals. Under § 1282.15(e), this methodology permitted the Enterprises to assume that the same percentage of low- and very low-income affordable rental units (by unit size) as are located in the census tract where the cooperative property is located are also present in the cooperative being financed. For example, if a cooperative property is in a census tract where multifamily properties average a certain percentage of low- and very low-income units, then the cooperative property is assumed to have the same percentage of low- and very low-income affordable units. In some geographic areas, particularly in certain parts of New York City, the rent estimation methodology may significantly overstate the number of low- and very low-income units that are eligible for goals credit in a specific cooperative property. This is because some census tracts in these geographic areas have great variations in unit rents due to the large number of subsidized, rent controlled, and rent stabilized units that are in close proximity to luxury market rate cooperative and rental properties. A luxury building in such a census tract could be determined under the rent estimation methodology to have low- and very low-income units that it does not actually have simply because the census tract has a significant number of such units. Due to these concerns, the final rule provides that the affordability of units in a cooperative property securing a blanket loan shall be determined solely on the basis of comparable rents used by the lender or the Enterprise in underwriting the blanket loan.

Comments on Proposed Rule

Several commenters supported the proposal to require that comparable rents rather than rent estimation be used to determine affordability of units in cooperative properties, although the reasons for their support were not articulated.

Fannie Mae supported the proposal, but also recommended that blanket loans on cooperative housing be permitted to count towards the housing goals if the property is a limited equity cooperative subject to rent restrictions. Fannie Mae stated that the affordability of such cooperative units should be based on the maximum permitted rent levels established under the rent restrictions for those units, as imposed by the cooperative's bylaws.

Freddie Mac opposed the proposal, recommending that the current rent estimation methodology be retained for determining affordability for blanket loans on cooperative housing. Freddie Mac stated that while it is possible that the use of the rent estimation methodology might result in overstating the number of low- and very low-income units in certain census tracts where lower-income cooperatives are in close proximity to luxury market rate housing, it questioned whether there is any data indicating that such overstatement has actually occurred.

Freddie Mac stated that if the proposal is adopted in the final rule, the rule should clarify that it is permissible for Freddie Mac to use its own underwriting rents rather than the rents used by the lenders, for purposes of determining affordability. Freddie Mac stated that it does not rely on a delegated underwriting model and instead re-underwrites each multifamily loan that it purchases.

FHFA Response

Regarding counting rules for rental units in limited equity cooperatives, as discussed in a previous section, FHFA has determined that, because of the wide variance among limited equity cooperative bylaws with respect to the types of rent and occupancy restrictions (if any) that may be imposed on cooperative owners who rent out their units, the Enterprises should follow their standard practice of determining the affordability of a specific unit's rent in limited equity cooperatives.

As to retaining the current rent estimation methodology for cooperatives, FHFA disagrees with Freddie Mac's comments for the reasons stated previously in this section.

As to establishing the underwriting rents for cooperative units, FHFA agrees that relying on an Enterprise's own underwriting rents should be permissible and has adopted this option in the final rule.

D. Mortgages With Unacceptable Terms or Conditions—§ 1282.16(d)

Consistent with the proposed rule, the final rule revises § 1282.16(d), which prohibits the Enterprises from receiving housing goals credit for purchases of “mortgages with unacceptable terms or conditions,” by eliminating the reference to that term, and amends § 1282.1 by removing the definition of “mortgage with unacceptable terms or conditions.” The final rule maintains the current prohibition on receiving housing goals credit for purchases of HOEPA mortgages, defined as mortgages covered by section 103(bb) of the Home Ownership and Equity Protection Act (15 U.S.C. 1602(bb)), as implemented by the Bureau of Consumer Financial Protection (CFPB).

The regulation currently defines “mortgages with unacceptable terms or conditions” to include single-family mortgages with excessive interest rates or costs, mortgages with certain prepayment penalties, and mortgages with prepaid credit life insurance. “Mortgages with unacceptable terms or conditions” also include mortgages with terms contrary to banking regulator guidance on nontraditional and subprime lending and mortgages originated using practices that do not comply with fair lending requirements.

Under the current regulation, “mortgages with unacceptable terms or conditions” and “HOEPA mortgages” must be included in the denominator for purposes of the housing goals. However, such mortgages are excluded from counting in the numerator, regardless of whether the loans would otherwise qualify. This treatment was intended to create a disincentive to purchasing such mortgages, by effectively lowering the goals performance of an Enterprise. In practice, these provisions have not affected the housing goals performance of the Enterprises because the Enterprises have purchased very few such mortgages. For example, in 2014, Fannie Mae reported it purchased one mortgage that met the definition of “mortgages with unacceptable terms or conditions.” Freddie Mac did not purchase any such mortgages in 2014.

Comments on Proposed Rule

Several advocacy groups recommended that high-cost loans should count in both the numerator and denominator for a housing goal because some of these loans can provide access to credit for underserved households if properly underwritten and given CFPB protections. However, the commenters stated that FHFA should monitor these loans closely to ensure consumers are not being overcharged for mortgages.

A housing advocacy group commenter recommended continuing the prohibition on “mortgages with unacceptable terms and conditions.” The commenter stated that keeping the phrase “mortgages with unacceptable terms and conditions” in the regulation would give FHFA the flexibility to address any new abusive loan products entering the market.

FHFA Response

The final rule eliminates the provisions related to “mortgages with unacceptable terms or conditions,” consistent with the proposed rule. As a result of the Enterprises' own mortgage purchase eligibility criteria, the Enterprises purchase virtually no mortgages that would be considered “mortgages with unacceptable terms and conditions” under the current housing goals regulation. Accordingly, the prohibition on receiving housing goals credit for purchases of such mortgages is not necessary in the regulation text.

In addition, the housing goals are not the most effective regulatory tool available for FHFA to discourage purchases of predatory or otherwise unsuitable mortgages. FHFA has regulatory authority to directly prohibit purchases by the Enterprises of any types of mortgages it determines are unsuitable. For example, FHFA prohibits the purchase of HOEPA loans by the Enterprises. FHFA has also required the Enterprises to limit their mortgage purchases to those that meet Qualified Mortgage product characteristics under the regulations implementing the Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank Act). Qualified Mortgage product characteristics are those related to the loan product itself rather than to the borrowers and their debt-to-income ratio. As a result, the Enterprises are generally prohibited from purchasing interest-only or negatively amortizing loans, balloon loans, 40-year loans, or loans with points and fees greater than three percentage points or up to five percentage points for smaller loans. To the extent that FHFA identifies any types of mortgages that meet Qualified Mortgage product criteria yet are not suitable for the Enterprises or for borrowers, FHFA may restrict Enterprise purchases of such mortgages in the future.

Higher Rate Mortgages in FHFA's Measurement of the Market

FHFA's measurement of the single-family mortgage market, which is used to determine the retrospective market share for the single-family housing goals under § 1282.12(b), as well as to set the prospective benchmark levels for the goals, is intended to reflect the portion of the overall single-family market that is eligible for purchase by the Enterprises. FHFA currently excludes mortgages with rate spreads of 150 basis points or more above the applicable average prime offer rate (APOR) as reported in the Home Mortgage Disclosure Act data.

In the proposed rule, FHFA specifically requested comment on whether mortgages with rate spreads that exceed 150 basis points above APOR should continue to be excluded from FHFA's measurement of the market, or whether a higher rate spread threshold should be established.

Comments on Proposed Rule

A housing advocacy group commenter recommended that FHFA continue to exclude loans with rate spreads more than 150 basis points above APOR. A trade association commenter noted that because the Enterprises already purchase mortgages with rate spreads more than 150 basis points above APOR, such loans should be included in the market size calculation. The commenter also stated that loans with rate spreads more than 650 basis points above APOR, which is the HOEPA trigger level for high-cost loans, should not be included.

FHFA Response

The final rule does not make any change to the existing regulation, which excludes loans with rate spreads more than 150 basis points above APOR from the retrospective market measure for the single-family housing goals. FHFA used the same exclusion in determining the size of the market in its analysis supporting the prospective benchmark levels for the single-family housing goals. FHFA recognizes that some mortgages purchased by the Enterprises may have rate spreads that exceed 150 basis points above APOR while still meeting the Enterprises' established underwriting criteria. However, other loans with rate spreads more than 150 basis points above APOR may not meet Enterprise underwriting criteria. While excluding loans with rate spreads more than 150 basis points above APOR is not a perfect substitute for excluding loans that do not meet Enterprise underwriting criteria, FHFA has determined that it is a reasonable approximation given the limited data available under HMDA.

E. Housing Goals Guidance—§ 1282.16(e)

Consistent with the proposed rule, § 1282.16(e) of the final rule adds a new provision requiring FHFA to make available on FHFA's public Web site (www.fhfa.gov) any determinations issued under § 1282.16(e) regarding the appropriate treatment of particular transactions or classes of transactions under the housing goals.

This change is intended to ensure that both Enterprises and any other interested parties are aware of any guidance that FHFA provides to either Enterprise regarding the appropriate housing goals treatment of any transactions in which they may engage, regardless of whether or not those transactions are covered in the housing goals regulation. FHFA and HUD, the Enterprises' predecessor mission regulator, from time to time have issued guidance on particular issues. To promote clear and consistent treatment of all transactions engaged in by either Enterprise, FHFA will make guidance issued to the Enterprises available on FHFA's public Web site.

Comments on Proposed Rule

Fannie Mae commented that Enterprise requests for guidance from FHFA often include confidential Enterprise business information that is subject to limitations on public disclosure. Fannie Mae recommended that the proposal be revised to state explicitly that any confidential business information submitted by an Enterprise in connection with a request will be excluded or redacted from any public release of a determination under this provision.

FHFA Response

FHFA recognizes that any confidential business information submitted by an Enterprise is subject to limitations on its public release. It is not necessary for the housing goals regulation to specifically cross-reference the applicable provisions on confidentiality in order for them to apply. Any public release of a determination under the housing goals would be made subject to the existing limitations on the release of confidential Enterprise information.

X. Seniors Housing Units and Skilled Nursing Units

The proposed rule would have incorporated into the regulation guidance that is currently in effect regarding the treatment of seniors housing units and skilled nursing units under the housing goals. The proposed rule would not have made any substantive changes to the guidance currently in effect.

Currently, seniors housing units are counted towards the housing goals, provided that the units meet the requirements that apply generally for multifamily housing. However, some seniors housing units with additional services included in the rent require that a prospective resident pay an up-front entrance fee as a condition of occupancy in addition to the monthly rent. Units with large up-front entrance fees are excluded from counting towards the housing goals because such fees make it difficult to distinguish between the portion of the up-front entrance fee that constitutes the actual monthly rent for purposes of determining affordability, and because in most instances large up-front entrance fees mean that the units are not affordable to low-income or very low-income families who would not be able to occupy a unit in any case.

Skilled nursing units are generally excluded from counting under the housing goals because their principal purpose is to provide medical services and housing is incidental to those purposes.

After consideration of the comments received on these provisions, FHFA has determined that it is not necessary to include the existing guidance on seniors housing units and skilled nursing units in the regulation itself. FHFA will make the current guidance available to the public on its Web site in accordance with the procedures described above under § 1282.16(e).

Comments on Proposed Rule—Seniors Housing Units

A comment letter signed by several members of Congress supported the proposed housing goals eligibility for seniors housing units with small up-front entrance fees, but stated that FHFA should monitor any adverse impacts on asset-rich seniors with low incomes. An advocacy group, while supporting the proposal, was also concerned with the impact of such fees on asset-rich, but income-poor, seniors.

Fannie Mae commented that it would be difficult to apply the proposal, stating that there is no consistent way of defining what are appropriate up-front entrance fees in the seniors housing industry. Fannie Mae recommended that in lieu of trying to determine which up-front entrance fees would be appropriate, a maximum amount of $12,500 should be established as an appropriate up-front entrance fee, based on current pricing in the seniors housing market.

Freddie Mac stated that the proposed limitation on up-front entrance fees was too broad and would exclude affordable seniors housing units with relatively small up-front “community fees.” Freddie Mac recommended that FHFA revise the proposal to allow units to be counted towards the housing goals unless there are large up-front entrance fees other than application processing fees, first-month advanced rent payments, security deposit fees, community fees, and other similar fees.

FHFA Response

As noted above, no substantive changes to the current guidance are being made at this time. FHFA may issue further guidance at a later date on what constitutes a “large” up-front entrance fee such that a seniors housing unit with services may be excluded from counting towards the housing goals.

Freddie Mac also commented that alternative methods should be permitted for determining affordability in seniors housing units with services rather than relying on the affordability estimation methodology in § 1282.15(e)(2), stating that the current methodology understates their affordability. Freddie Mac recommended that the Enterprises be permitted to determine the level of tenant incomes based on the age of the tenant and the census tract area median income for that age group. Freddie Mac also recommended that the Enterprises be permitted to rely on the receipt of Medicaid benefits as a proxy for income in determining the income level of a resident in a seniors housing unit.

FHFA Response

Under the current regulation, seniors housing units that do not include additional services in the rent are treated as multifamily dwelling units for purposes of the housing goals, with affordability determined based on the unit rent. Seniors housing units that include additional services in the rent are currently treated as multifamily dwelling units with missing data for purposes of determining affordability under the estimation provisions of § 1282.15(e)(2). As discussed above and consistent with current practice, under the final rule, seniors housing units with additional services included in the rent will continue to be excluded from the estimation cap in § 1282.15(e)(3). FHFA will consider whether to conduct further review of the alternatives proposed by Freddie Mac to determine whether they would be appropriate methods for determining affordability. If FHFA changes how affordability is determined for seniors housing units, it will post the revised guidance on FHFA's public Web site in accordance with § 1282.16(e).

Comments on Proposed Rule—Skilled Nursing Units

Fannie Mae recommended that the proposed definition of “skilled nursing unit” be narrowed by distinguishing between units that are principally residential and units with a principal purpose of providing medical services on a temporary basis. Specifically, Fannie Mae suggested revising the definition to mean “a seniors housing unit, the principal purpose of which is to provide 24-hour skilled medical services on a temporary basis rather than to serve as a residence.”

Freddie Mac recommended similar changes to the proposed definition of “skilled nursing unit.” Freddie Mac noted that many facilities provide a range of services and that the market has trended toward continuing care retirement communities. Freddie Mac also noted that the services provided in a particular unit may change over time. Freddie Mac proposed defining “skilled nursing unit” as “a multifamily property unit dedicated to providing tenants aged 55 and over with 24-hour licensed medical services that go beyond assistance with activities of daily living. Activities of daily living may include management of medications, bathing, dressing, toileting, ambulating and eating.”

FHFA Response

The definition of “skilled nursing unit” in the proposed rule was not intended to include other types of continuing care retirement communities where housing is also a principal purpose. FHFA may provide revised guidance at a later date on the definition of “skilled nursing unit.” FHFA will post any revised guidance on its public Web site in accordance with § 1282.16(e).

XI. Blanket Loans on Manufactured Housing Communities

FHFA intends to make available to the public on its Web site, in accordance with the procedures under § 1282.16(e), its existing guidance which provides that blanket loans on manufactured housing communities are excluded from counting under the multifamily housing goals. FHFA specifically requested comment in the proposed rule on whether blanket loans on manufactured housing communities owned by either residents, investors, or cooperatively by residents, should be eligible for multifamily housing goals credit.

The final rule does not revise the current regulation to allow blanket loans on manufactured housing communities to count under the multifamily housing goals. It is difficult to accurately determine a manufactured housing unit's affordability under the housing goals because bedroom count information on individual manufactured housing units in the communities is not collected by the Enterprises, and the pad rent alone does not include the full cost of housing for the residents, which includes paying for their unit financing. Therefore, the practical question of how to determine housing costs and affordability, including how to adjust household size for the number of bedrooms in a unit so as to accurately apply the rent estimation alternative, cannot be answered at this time given available data. FHFA will continue to evaluate the treatment of manufactured housing communities in connection with its rulemaking for the Enterprises' Duty to Serve underserved markets under 12 U.S.C. 4565. FHFA may issue further guidance on the appropriate treatment of blanket loans on manufactured housing communities under the housing goals at a later date.

Comments on Proposed Rule

FHFA received extensive comments in response to its request for comment on the potential inclusion of blanket loans on manufactured housing communities under the multifamily housing goals. All but one of the commenters on this issue recommended counting such loans for goals credit. Fannie Mae noted that purchases of blanket loans on manufactured housing communities are comparable to purchases of blanket loans on cooperative buildings and condominium projects and should be treated similarly for purposes of the housing goals. Both Fannie Mae and Freddie Mac stated that manufactured housing is an important source of low-cost housing, particularly for lower income households. Fannie Mae provided data illustrating the affordability of manufactured housing as compared to other housing types. Freddie Mac stated that manufactured homes account for between 7 and 8 percent of all single-family housing units. Freddie Mac also noted that manufactured housing is particularly important as a source of affordable housing in rural communities, where other housing options often are not available. Fannie Mae and Freddie Mac also provided substantial additional comments on how to define and count blanket loans on manufactured housing communities.

FHFA Response

Due to the practical limitations on determining affordability described above, FHFA has determined not to allow blanket loans on manufactured housing communities to count under the housing goals. FHFA will instead separately consider the treatment of manufactured housing communities in connection with its rulemaking for the Enterprises' Duty to Serve underserved markets.

XII. Paperwork Reduction Act

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

XIII. 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 rule under the Regulatory Flexibility Act. The General Counsel of FHFA certifies that this rule will not 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 amends part 1282 of Title 12 of the Code of Federal Regulations as follows:

PART 1282—ENTERPRISE HOUSING GOALS AND MISSION

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

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

2. Amend § 1282.1(b) as follows:

a. Remove the definition of “Contract rent”;

b. Revise the definition of “Dwelling unit”;

c. Add in alphabetical order a definition of “Efficiency”;

d. Revise the definition of “Families in low-income areas”;

e. Remove the definition of “HMDA”;

f. Revise the definition of “HOEPA mortgage”;

g. Remove the definition of “Mortgage with unacceptable terms or conditions”;

h. Revise the definition of “Rent”;

i. Remove the definition of “Rental housing”;

j. Add in alphabetical order a definition of “Small multifamily property”;

k. Revise the definition of “Utilities”; and

l. Remove the definitions of “Utility allowance,” and “Working day”.

The revisions and additions read as follows:

§ 1282.1

Definitions.

* * * * *

(b)* * *

Dwelling unit means a room or unified combination of rooms with plumbing and kitchen facilities intended for use, in whole or in part, as a dwelling by one or more persons, and includes a dwelling unit in a single-family property, multifamily property, or other residential or mixed-use property.

Efficiency means a dwelling unit having no separate bedrooms or 0 bedrooms.

* * * * *

Families in low-income areas means:

(i) Any family that resides in a census tract in which the median income does not exceed 80 percent of the area median income;

(ii) Any family with an income that does not exceed area median income that resides in a minority census tract; and

(iii) Any family with an income that does not exceed area median income that resides in a designated disaster area.

* * * * *

HOEPA mortgage means a mortgage covered by section 103(bb) of the Home Ownership and Equity Protection Act (HOEPA) (15 U.S.C. 1602(bb)), as implemented by the Bureau of Consumer Financial Protection.

* * * * *

Rent means the actual rent or average rent by unit size for a dwelling unit.

(i) Rent is determined based on the total combined rent for all bedrooms in the dwelling unit, including fees or charges for management and maintenance services and any utility charges that are included.

(A) Rent concessions shall not be considered, i.e., the rent is not decreased by any rent concessions.

(B) Rent is net of rental subsidies, i.e., the rent is decreased by any rental subsidy.

(ii) When the rent does not include all utilities, the rent shall also include:

(A) The actual cost of utilities not included in the rent;

(B) The nationwide average utility allowance, as issued periodically by FHFA;

(C) The utility allowance established under the HUD Section 8 Program (42 U.S.C. 1437f) for the area where the property is located; or

(D) The utility allowance for the area in which the property is located, as established by the state or local housing finance agency for determining the affordability of low-income housing tax credit properties under section 42 of the Internal Revenue Code (26 U.S.C. 42).

* * * * *

Small multifamily property means any multifamily property with at least 5 dwelling units but no more than 50 dwelling units.

Utilities means charges for electricity, piped or bottled gas, water, sewage disposal, fuel (oil, coal, kerosene, wood, solar energy, or other), and garbage and trash collection. Utilities do not include charges for subscription-based television, telephone, or internet service.

* * * * *

3. Amend § 1282.11 by revising paragraph (a)(1) to read as follows:

§ 1282.11

General.

(a) * * *

(1) Three single-family owner-occupied purchase money mortgage housing goals, a single-family owner-occupied purchase money mortgage housing subgoal, a single-family refinancing mortgage housing goal, a multifamily special affordable housing goal, and two multifamily special affordable housing subgoals;

* * * * *

4. 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 2015, 2016, and 2017 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 2015, 2016, and 2017 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 2015, 2016, and 2017 shall be 14 percent of the total number of purchase money mortgages purchased by that Enterprise in each year that finance owner-occupied single-family properties.

(g) 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 2015, 2016, and 2017 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.

5. 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 300,000 dwelling units affordable to low-income families in multifamily residential housing financed by mortgages purchased by the Enterprise in each year for 2015, 2016, and 2017.

(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 2015, 2016, and 2017.

(d) Small multifamily low-income housing subgoal. (1) For the year 2015, the benchmark level for each Enterprise's purchases of mortgages on small multifamily properties affordable to low-income families shall be at least 6,000 dwelling units affordable to low-income families in small multifamily properties financed by mortgages purchased by the Enterprise.

(2) For the year 2016, the benchmark level for each Enterprise's purchases of mortgages on small multifamily properties affordable to low-income families shall be at least 8,000 dwelling units affordable to low-income families in small multifamily properties financed by mortgages purchased by the Enterprise.

(3) For the year 2017, 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.

6. Amend § 1282.15 by revising paragraphs (b), (c), (d), (e) and (g)(2) to read as follows:

§ 1282.15

General counting requirements.

* * * * *

(b) Counting owner-occupied units. (1) Mortgage purchases financing owner-occupied single-family properties shall be evaluated based on the income of the mortgagors and the area median income at the time the mortgage was originated. To determine whether mortgages may be counted under a particular family income level, i.e., low- or very low-income, the income of the mortgagors is compared to the median income for the area at the time the mortgage was originated, using the appropriate percentage factor provided under § 1282.17.

(2) Mortgage purchases financing owner-occupied single-family properties for which the income of the mortgagors is not available shall be included in the denominator for the single-family housing goals and subgoal, but such mortgages shall not be counted in the numerator of any single-family housing goal or subgoal.

(c) Counting dwelling units for multifamily housing goal and subgoals. Performance under the multifamily housing goal and subgoals shall be measured by counting the number of dwelling units that count toward achievement of a particular housing goal or subgoal in all multifamily properties financed by mortgages purchased by an Enterprise in a particular year. Only dwelling units that are financed by mortgage purchases, as defined by FHFA, and that are not specifically excluded as ineligible under § 1282.16(b), may be counted for purposes of the multifamily housing goal and subgoals.

(d) Counting rental units—(1) Use of rent. For purposes of counting rental units toward achievement of the multifamily housing goal and subgoals, mortgage purchases financing such units shall be evaluated based on rent and whether the rent is affordable to the income group targeted by the housing goal and subgoals. A rent is affordable if the rent does not exceed the maximum levels as provided in § 1282.19.

(2) Affordability of rents based on housing program requirements. Where a multifamily property is subject to an affordability restriction under a housing program that establishes the maximum permitted income level for a tenant or a prospective tenant or the maximum permitted rent, the affordability of units in the property may be determined based on the maximum permitted income level or maximum permitted rent established under such housing program for those units. If using income, the maximum income level must be no greater than the maximum income level for each goal, adjusted for family or unit size as provided in § 1282.17 or § 1282.18, as appropriate. If using rent, the maximum rent level must be no greater than the maximum rent level for each goal, adjusted for unit size as provided in § 1282.19.

(3) Unoccupied units. Anticipated rent for unoccupied units may be the market rent for similar units in the neighborhood as determined by the lender or appraiser for underwriting purposes. A unit in a multifamily property that is unoccupied because it is being used as a model unit or rental office may be counted for purposes of the multifamily housing goal and subgoals only if an Enterprise determines that the number of such units is reasonable and minimal considering the size of the multifamily property.

(4) Timeliness of information. In evaluating affordability under the multifamily housing goal and subgoals, each Enterprise shall use tenant and rental information as of the time of mortgage acquisition.

(e) Missing data or information for multifamily housing goal and subgoals. (1) Rental units for which bedroom data are missing shall be considered efficiencies for purposes of calculating unit affordability.

(2) When an Enterprise lacks sufficient information to determine whether a rental unit in a property securing a multifamily mortgage purchased by an Enterprise counts toward achievement of the multifamily housing goal or subgoals because rental data is not available, an Enterprise's performance with respect to such unit may be evaluated using estimated affordability information by multiplying the number of rental units with missing affordability information in properties securing multifamily mortgages purchased by the Enterprise in each census tract by the percentage of all rental dwelling units in the respective tracts that would count toward achievement of each goal and subgoal, as determined by FHFA based on the most recent decennial census.

(3) The estimation methodology in paragraph (e)(2) of this section may be used up to a nationwide maximum of 5 percent of the total number of rental units in properties securing multifamily mortgages purchased by the Enterprise in the current year. Multifamily rental units in excess of this maximum, and any units for which estimation information is not available, shall not be counted for purposes of the multifamily housing goal and subgoals.

* * * * *

(g) * * *

(2) When an Enterprise cannot precisely determine whether a mortgage is on dwelling unit(s) located in one area, the Enterprise shall determine the median income for the split area in the manner prescribed by the Federal Financial Institutions Examination Council for reporting under the Home Mortgage Disclosure Act (12 U.S.C. 2801 et seq.), if the Enterprise can determine that the mortgage is on dwelling unit(s) located in:

(i) A census tract; or

(ii) A census place code.

* * * * *

7. Amend § 1282.16 by revising paragraphs (c)(5), (d), and (e) to read as follows:

§ 1282.16

Special counting requirements.

* * * * *

(c) * * *

(5) Cooperative housing and condominiums. (i) The purchase of a mortgage on a cooperative housing unit (“a share loan”) or a mortgage on a condominium unit shall be treated as a mortgage purchase for purposes of the housing goals. Such a purchase shall be counted in the same manner as a mortgage purchase of single-family owner-occupied units.

(ii) The purchase of a blanket mortgage on a cooperative building or a mortgage on a condominium project shall be treated as a mortgage purchase for purposes of the housing goals. The purchase of a blanket mortgage on a cooperative building shall be counted in the same manner as a mortgage purchase of a multifamily rental property, except that affordability must be determined based solely on the comparable market rents used in underwriting the blanket loan. If the underwriting rents are not available, the loan shall not be treated as a mortgage purchase for purposes of the housing goals. The purchase of a mortgage on a condominium project shall be counted in the same manner as a mortgage purchase of a multifamily rental property.

(iii) Where an Enterprise purchases both a blanket mortgage on a cooperative building and share loans for units in the same building, both the mortgage on the cooperative building and the share loans shall be treated as mortgage purchases for purposes of the housing goals. Where an Enterprise purchases both a mortgage on a condominium project and mortgages on individual dwelling units in the same project, both the mortgage on the condominium project and the mortgages on individual dwelling units shall be treated as mortgage purchases for purposes of the housing goals.

* * * * *

(d) HOEPA mortgages. HOEPA mortgages shall be treated as mortgage purchases for purposes of the housing goals and shall be included in the denominator for each applicable single-family housing goal, but such mortgages shall not be counted in the numerator for any housing goal.

(e) FHFA review of transactions. FHFA may determine whether and how any transaction or class of transactions shall be counted for purposes of the housing goals, including treatment of missing data. FHFA will notify each Enterprise in writing of any determination regarding the treatment of any transaction or class of transactions under the housing goals. FHFA will make any such determinations available to the public on FHFA's Web site, www.fhfa.gov.

§ 1282.17

[Amended]

8. Amend § 1282.17 in the introductory text by removing the phrase “rental housing” and adding in its place the phrase “rental units”.

§ 1282.19

[Amended]

9. Amend § 1282.19 by removing paragraph (f).

10. Amend § 1282.20 by revising paragraph (b) to read as follows:

§ 1282.20

Determination of compliance with housing goals; notice of determination.

* * * * *

(b) Multifamily housing goal and subgoals. The Director shall evaluate each Enterprise's performance under the multifamily low-income housing goal, the multifamily very low-income housing subgoal, and the small multifamily low-income housing subgoal, on an annual basis. If the Director determines that an Enterprise has failed, or there is a substantial probability that an Enterprise will fail, to meet a multifamily housing goal or subgoal established by this subpart, the Director shall notify the Enterprise in writing of such preliminary determination.

* * * * *

Dated: August 13, 2015.

Melvin L. Watt,

Director, Federal Housing Finance Agency.

Footnotes

1.  12 U.S.C. 4561(a).

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

3.  75 FR 55891.

4.  77 FR 67535.

5.  79 FR 54481. The proposed rule was also posted on FHFA's public Web site on August 29, 2014 for public comment.

6.  In addition, FHFA posted in the public comments docket a summary of a meeting on the proposed rule with an individual, a policy advocacy group and a housing advocacy group.

7.  FHFA's determinations regarding Enterprise performance under the housing goals can be accessed from this page: http://www.fhfa.gov/​PolicyProgramsResearch/​Programs/​AffordableHousing/​Pages/​Affordable-Housing-FMandFM.aspx.

8.  12 U.S.C. 4501.

9.  12 U.S.C. 1716(3); 12 U.S.C. 1451(b)(3).

10.  See the Freddie Mac Primary Mortgage Market Survey (PMMS), available at http://www.freddiemac.com/​pmms/​pmms_​archives.html.

11.  See FHFA's house price index (HPI). Historical HPI data are available at http://www.fhfa.gov/​KeyTopics/​Pages/​House-Price-Index.aspx.

12.  The unemployment and labor force participation rates are available in data published by the Bureau of Labor Statistics. State unemployment rates can be found at http://www.bls.gov/​lau/​lauov.htm/​. The U.S.-wide labor force participation rate is available at http://data.bls.gov/​timeseries/​LNS11300000. Household income data are available from the Census Bureau. Recent reports on income growth are available at http://www.census.gov/​content/​dam/​Census/​library/​publications/​2014/​acs/​acsbr13-02.pdf and http://www.census.gov/​content/​dam/​Census/​library/​publications/​2013/​acs/​acsbr12-02.pdf.

13.  See Appendix Tables (Table W-2) in the 2015 “The State of the Nation's Housing,” Joint Center for Housing Studies, available at http://www.jchs.harvard.edu/​research/​state_​nations_​housing.

14.  See http://www.fhfa.gov/​Media/​PublicAffairs/​Pages/​Prepared-Remarks-of-Melvin-L-Watt-2014-NAR-Conference.aspx.

15.  Growth rates calculated by FHFA using data from the New York Federal Reserve Bank's Household Debt and Credit Report Web site, http://www.newyorkfed.org/​microeconomics/​hhdc.html#2014/​q4. The Web site reports that automobile loan debt grew from $0.86 trillion to $0.95 trillion (10.5 percent), whereas student loan debt grew from $1.08 trillion to $1.16 trillion (7.4 percent).

16.  See “State of the Nation's Housing 2015.” In particular, see Table W-9. The data and the full report are available at http://www.jchs.harvard.edu/​state-nations-housing-2015-embargoed.

17.  12 U.S.C. 4562(b).

18.  12 U.S.C. 4562(e)(2)(A).

19.  12 U.S.C. 4562(e)(3).

20.  12 U.S.C. 4564(b)(1), (2).

21.  12 U.S.C. 4566(b).

22.  The housing advocacy groups stated that the statute does not give FHFA authority to administratively adjust the housing goal targets once they have been established by rulemaking without first soliciting public input on any change. The reference in the proposed rule preamble was to FHFA's discretionary authority to reduce a goal in response to a petition from an Enterprise, after notice and comment, as specifically authorized by the statute. See 12 U.S.C. 4564(b); 12 CFR 1282.14(d); 79 FR 54481, 54483 (Sept. 11, 2014).

23.  12 U.S.C. 4562(e)(2).

24.  12 U.S.C. 4562(e)(2)(A).

25.  More detailed explanation of the market estimation models can be found in FHFA's research papers available at http://www.fhfa.gov/​PolicyProgramsResearch/​Research/​.

26.  12 U.S.C. 4562(e)(1).

27.  In 2013, the Enterprises remained the largest issuers of mortgage-backed securities (MBS), guaranteeing 73 percent of single-family MBS, slightly above the average of 72 percent for 2008-2012, but well above the average of 46 percent for 2004-2007, and somewhat above the average of 67 percent for 2000-2003. See Inside Mortgage Finance Publications, “Mortgage Market Statistical Annual,” volume II, “The Secondary Mortgage Market,” p.4 (2013 Edition); see also Inside MBS & ABS, p.4 (April 4, 2014).

28.  The Home Affordable Refinance Program (HARP), which became effective in March 2009 and was expanded in 2011, is an effort to enhance the opportunity for many homeowners to refinance. Homeowners with LTV ratios above 80 percent whose mortgages are owned or guaranteed by Fannie Mae or Freddie Mac and who are current on their mortgages have the opportunity to reduce their monthly mortgage payments to take advantage of historically low mortgage interest rates.

29.  The Enterprise Annual Housing Activity Reports and the summary tables for the AMRs can be accessed from this page: http://www.fhfa.gov/​PolicyProgramsResearch/​Programs/​AffordableHousing/​Pages/​Affordable-Housing-FMandFM.aspx.

30.  12 U.S.C. 4563(a)(4).

31.  U.S. Census Bureau, 2013 American Community Survey, National Table C-12-RO.

http://www.census.gov/​programs-surveys/​ahs/​data/​2013/​national-summary-report-and-tables-ahs-2013.html?​eml=​gd.

32.  MBA/CREF Forecast of Key Multifamily Real Estate Finance Indicators, February 2015.

33.  12 U.S.C. 4563(c); 12 CFR 1282.1.

34.  See “State of the Nation's Housing 2015.” The data and the full report are available at http://www.jchs.harvard.edu/​state-nations-housing-2015-embargoed.

35.  Moody's/Real Capital Analytics, “Composite CPPI Indices” (July 2015), https://www.rcanalytics.com/​Public/​rca_​cppi.aspx.

36.  U.S. Census Bureau, “Current Population Survey/Housing Vacancy Survey, Series H-111, U.S. Census Bureau, Washington, DC 20233.” The vacancy rates reported by the U.S. Census Bureau differ from those reported by some other sources, but trends are similar.

37.  U.S. Census Bureau, “Median Asking Rent for the U.S. and Regions.” The asking rents reported by the U.S. Census Bureau differ from those reported by some other sources, but trends are similar. For example, data from CB Richard Ellis shows average rent rates at $1,191 in 2010, then increasing steadily to $1,339 in 2013 and to $1,457 in 2014.

38.  U.S. Census Bureau, “New Privately Owned Housing Units Authorized by Building Permits in Permit-Issuing Places (In structures with 5 units or more).”

39.  U.S. Census Bureau, “New Privately Owned Housing Started (In structures with 5 units or more).”

40.  U.S. Census Bureau, “New Privately Owned Housing Units Completed (In structures with 5 units or more).”

41.  Enterprise data.

42.  Enterprise data.

43.  “America's Rental Housing Markets: Evolving Markets and Needs,” Harvard Joint Center for Housing Studies (December 2013).

44.  Ibid.

45.  “What Community Banks Are Saying—A Review of Four Community Banks' Small Multifamily Lending Programs,” Community Developments Investments (May 2015), Office of the Comptroller of the Currency: http://www2.occ.gov/​publications/​publications-by-type/​other-publications-reports/​cdi-newsletter/​small-multifamily-rental-spring-2015/​small-multifamily-rental-ezine-article-5-community.html.

46.  The final rule also makes a number of conforming changes throughout part 1282 to reflect the addition of this new small multifamily subgoal.

47.  12 U.S.C. 4563(a)(3).

48.  Id.

49.  12 U.S.C. 4563(a)(4).

50.  U.S. Census Bureau, “2011 American Community Survey.”

51.  “Rental Housing Finance Survey,” Table 3 (March 27, 2013), http://portal.hud.gov/​hudportal/​HUD?​src=​/​press/​press_​releases_​media_​advisories/​2013/​HUDNo.13-035.

52.  “Rental Housing Finance Survey,” Tables 2b, 2c, 2d and 3 (March 27, 2013), http://portal.hud.gov/​hudportal/​HUD?​src=​/​press/​press_​releases_​media_​advisories/​2013/​HUDNo.13-035.

53.  See Fannie Mae, “Fannie Mae's Role in the Small Multifamily Loan Market” (First Quarter 2011), https://www.fanniemae.com/​content/​fact_​sheet/​wpmfloanmkt.pdf.

54.  “Rental Housing Finance Survey” (2012), http://portal.hud.gov/​hudportal/​HUD?​src=​/​press/​press_​releases_​media_​advisories/​2013/​HUDNo.13-035. Although the Rental Housing Finance Survey data do not match FHFA's definition of small multifamily properties precisely (the data use 5 to 49 units instead of 5 to 50 units), the difference is not material.

55.  See 12 U.S.C. 4563(a)(3).

56.  The denominator includes the Enterprise's total purchases of mortgages on owner-occupied single-family properties and is measured separately for home purchase mortgages and for refinancing mortgages. The numerator includes only those purchases of mortgages that actually meet the criteria for a particular housing goal.

57.  12 U.S.C. 4563(c).

58.  Id.