Federal Housing Finance Agency.
Notice; input accepted.
The Federal Housing Finance Agency (FHFA) is requesting public input on implementation issues associated with a contemplated reduction in loan purchase limits by the Federal National Mortgage Association (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac) (together, the Enterprises). Each Enterprise must set its loan purchase limits at or below the maximum limits, which are determined by statutory formulas. The maximum limits for 2014 were published by FHFA on November 26, 2013. A decrease in the Enterprises' loan limits below the statutory maximums is one means of reducing the Enterprises' financial market footprint pursuant to FHFA's Strategic Plan for Enterprise Conservatorships. Other means of reducing the Enterprises' footprint relate to their single-family mortgage guarantee business and include increasing guarantee fees and engaging in risk-sharing transactions.
The basic premise of these measures is as follows: with an uncertain future and a desire for private capital to re-enter the market, the Enterprises' market presence should be reduced gradually over time. In addition, at the end of 2012, the amount of taxpayer capital available to support the Enterprises' outstanding debt and mortgage-backed securities obligations became fixed. Limiting their risk exposure is vital to maintaining the adequacy of the remaining capital support through the financial support agreements between the Enterprises and the U.S. Department of the Treasury. Finally, a taxpayer-backed conservatorship provides a significant subsidy to the mortgage market that limits private capital participation and underprices risk in the market.
The contemplated action described below is a plan and not a final decision. The requested public input will be carefully reviewed before FHFA decides whether and how to proceed with the planned reductions in Freddie Mac's and Fannie Mae's loan purchase limits. In short, no final decision on loan purchase limits will be made until all input is reviewed. The changes contemplated in this Request for Public Input will not affect loans originated before October 1, 2014.
The remainder of this Request for Public Input sets forth: FHFA's legal authority for directing the Enterprises to set loan purchase limits below the maximum loan limits; the planned approach to reduce the Enterprises' loan limits; and a request for public input regarding implementation of the plan. An appendix to this Request for Public Input includes analysis describing the potential impact of the plan.
FHFA's Legal Authority for Setting the Enterprises' Loan Purchase Limits
In their chartering acts, the Enterprises are authorized to purchase mortgages up to specified limits, as adjusted annually; 12 U.S.C. 1717(b) and 12 U.S.C. 1454(a). The statutes provide that each Enterprise “. . . shall establish limitations governing the maximum original principal obligation of conventional mortgages that are purchased by it. . . . Such limitations shall not exceed [the loan limits] . . .”Start Printed Page 77451
The Housing and Economic Recovery Act of 2008 (HERA) establishes the maximum loan limits that Fannie Mae and Freddie Mac are permitted to set for mortgage acquisitions. HERA also requires an annual adjustment to these maximums to reflect changes in the national average home price. The maximum general limits are adjusted by a calculation of year-over-year changes to the existing level of home prices. In recent years, FHFA has not selected a specific index, but has noted that all reasonable indexes have declined. On November 26, 2013, FHFA announced maximum loan limits for 2014 and provided a description of the methodology used in determining these limits. The Enterprises, under their charters, then determine whether to set the next year's loan purchase limits at or below the new maximums.
When the Enterprises are in conservatorship, FHFA, as conservator, may take such action as may be: “(i) necessary to put the regulated entity in a sound and solvent condition; and (ii) appropriate to carry on the business of the regulated entity and preserve and conserve the assets and property of the regulated entity.” 12 U.S.C. 4617(b)(2)(D).
In addition, FHFA may “perform all functions of the regulated entity in the name of the regulated entity which are consistent with the appointment as conservator or receiver”; 12 U.S.C. 4617(b)(2)(B)(iii). FHFA's conservator obligation to preserve and conserve the assets includes policies to reduce the Enterprises' presence in the mortgage market and the risks in their business activities. Exercising, as conservator, a business judgment on a core business function of the Enterprises—setting levels of loan amounts below the maximums eligible for purchase by the Enterprises—is consistent with FHFA legal authorities. Therefore, the conservator's legal authority and responsibility to “carry on the business” of the Enterprises supports a decision to direct the setting of new and lower loan purchase limits by the Enterprises.
A Plan for Setting Loan Purchase Limits Lower Than Statutory Maximum Limits
As FHFA announced on November 26, 2013, the maximum loan limits in 2014 for one-unit properties range from $417,000 (the baseline limit) in most locations to $625,500 (the ceiling limit) in certain high-cost areas in the contiguous United States. In accordance with HERA, FHFA will continue to calculate and announce the future annual adjustments to the maximum loan limits in late November of each year.
As described above, the maximum loan limits represent upper bounds to the sizes of loans that the Enterprises can purchase. Through its authority as conservator, FHFA may direct each Enterprise to set new loan purchase limits below the statutory maximum limits and below current limits by the same percentage in every county and county-equivalent area 
in the country. FHFA has developed a plan to gradually reduce loan purchase limits by reducing the baseline loan limit from $417,000 to $400,000, a 4.077 percent decline. The planned ceiling limit in high-cost areas would be lowered by the same percentage from $625,500 to $600,000.
In areas where current purchase limits lie between the baseline and ceiling limits, the planned loan purchase limit would be decreased by the same percentage as the baseline and ceiling purchase limits (i.e., 4.077 percent). The new, lower, purchase limits would only affect loans originated after October 1, 2014. Loans eligible for purchase before the reductions will remain eligible in the future, regardless of whether they exceed the new loan purchase limits.
As FHFA has noted previously, ample notice will be provided to the market before any change in loan purchase limits would be implemented. To meet that goal and provide an opportunity to receive input in response to this Request for Public Input, the approach described above will not, in any event, affect loans originated before October 1, 2014.
Request for Public Input: Implementation Questions
FHFA requests input from the public and interested parties on the following questions associated with implementing the reduction of the Enterprises' loan purchase limits just described:
1. FHFA has promised to provide at least six months advance notice of any reduction of the loan purchase limit. If FHFA makes a determination and announcement by, for example, March 20, would October 1 be a reasonable effective date, or would operational issues suggest that an alternate or later date in 2014 would be preferable?
2. Assuming the Enterprises' loan limit reduction takes effect for purchases of loans originated on or after October 1, 2014, should that reduction be in effect for 12 months or 15 months? In other words, for future announcements on any future change in the loan purchase limits, is a January 1 origination date preferred, or should those announcements be tied to the initial loan purchase limit reduction date?
3. Is it preferable for the Enterprises to announce a multi-year schedule of proposed decreases? If so, should it be a specific percent per year, perhaps five percent, or a specific dollar reduction, perhaps $20,000 each year?
4. Currently, there are several geographic areas with limits between the current baseline loan limit of $417,000 and the ceiling high-cost area limit of $625,500. The maximum limits in these areas are tied to the median house price in those areas. Should FHFA tie future reductions in loan purchase limits in those areas to changes in median house prices in any way, or should reductions in those areas simply be proportional to reductions in the baseline limit?
5. Currently, all loan limits are rounded to the nearest $50. Is this appropriate, or should the loan purchase limits be set at even multiples of either $1,000 or some other dollar amount for greater simplicity?
FHFA will accept public input through its Office of Policy Analysis and Research (OPAR), no later than March 20, 2014. Communications may be addressed to Federal Housing Finance Agency, (OPAR), Constitution Center, 400 Seventh Street SW., Ninth Floor, Washington, DC 20024, or emailed to: email@example.com. Communications to FHFA may be made public and posted without change on the FHFA Web site at http://www.fhfa.gov, and would include any personal information provided, such as name, address (mailing and email), and telephone numbers.
Dated: December 17, 2013.
Edward J. DeMarco,
Acting Director, Federal Housing Finance Agency.
Appendix: Impact Analysis of Reductions in the Enterprises' Loan Purchase Limits
This Appendix provides historical background on loan purchase limits, as well as detail on how they have been calculated. Broadly speaking, this background reveals that the current loan purchase limits (which are set at the maximum loan limits) are historically high and that certain implementation decisions have been made that, in some locations, made those limits higher than they otherwise would have been.
Further, this Appendix provides statistics showing the potential market impact of reducing loan purchase limits by the magnitude described in the Request for Public Input. The focus of the analysis is on evaluating the number and types of borrowers that might have been affected had lower loan purchase limits been in place in Start Printed Page 774522012. The evaluation of 2012 data provides a reasonable indication of likely effects of loan purchase limit reductions in 2014. It is not possible to know with certainty how a different loan purchase limit regime will affect the market environment and specific borrowers, but the analysis suggests a small decline in loan purchase limits will have a modest impact.
Background: Baseline Loan Purchase Limit
Figure 1 plots the time trend in the historical loan purchase limit for one-unit properties in the contiguous United States since 1992.
The graph also shows changes in the ceiling loan limit that has capped limits in certain high-cost areas since 2008. Between 2008 and late 2011, that ceiling was $729,750 for the contiguous U.S. In October 2011, the ceiling was decreased to $625,500.
Figure 1 reveals that the baseline loan limit of $417,000 is at its historical peak. To provide context for the relative size of the loan limit increases shown in Figure 1, Figure 2 plots the growth in baseline loan limits against the growth in several other economic metrics, including median household incomes, consumer prices, and median U.S. home values. The respective values for each of these variables are normalized in the graph so that the value in 1992 for each variable is set equal to 100.
The graph clearly shows the elevated nature of current limits. At $417,000, the 2013 baseline loan limit, for instance, was 206 percent of its level in 1992. The “ceiling” loan limit—the highest loan purchase limit in high-cost areas—was 309 percent of the 1992 limit. By contrast, 2013 data for median home values, inflation, and median household income indicate that those metrics this year have been between 163 percent and 180 percent of their 1992 levels.
Background: Calculation of Loan Purchase Limits in High-Cost Areas
While Figures 1 and 2 provide some indication of the elevated nature of loan limits, they only address the baseline and ceiling loan limits. They do not evaluate the actual calculations that have determined maximum loan limits in high-cost areas. It can be shown that specific implementation decisions have made maximum loan limits higher than they otherwise would be in many high-cost areas. In conservatorship, the Enterprises have set their loan purchase limits equal to the statutory maximum loan limits.
Since 2008, maximum loan limits in high-cost areas have been statutorily set as a function of median local home values. Under HERA, the maximum loan limit in high-cost areas is 115 percent of the local median home value. The resulting limit is bounded between $417,000 and $625,500.
Because maximum loan limits are determined by median home values, the precise method used for estimating median home values can have a significant impact on the actual maximum loan limit. Since 2008, for determining maximum loan limits, FHFA has used median home values produced by the U.S. Department of Housing and Urban Development (HUD).
FHFA has used the HUD-generated median home values because they have full geographic coverage. That is, median home value statistics are included for all counties across the country—something no other single source provides. Also, the introduction of a set of median home values different from those produced by HUD might generate confusion among market participants.
Although HUD's methodology for calculating median values is positive in many respects, for many counties, one of the steps in the process makes Enterprise maximum loan limits, which are based on those median values, quite high relative to what the specific county-level data would suggest.
By law, when determining median home values for counties in Metropolitan and Micropolitan Statistical Areas, HUD's calculation must implement a “high-cost county rule” (HCCR). Under the HCCR, median home values for counties in Metropolitan and Micropolitan areas must reflect the median values in the highest-cost component county. To illustrate—for a Metropolitan Statistical Area (MSA) comprised of 10 counties, HUD begins by separately estimating median home values in each of the 10 counties. Then, after finding the highest of those 10 values, HUD assigns that highest value to all 10 counties in the MSA.
The HCCR tends to lead to an overstatement of local median home values. Washington, DC provides a good example. The two dozen county and county-equivalent areas that comprise the Washington, DC metropolitan area are diverse in terms of their median home values. Over the last several years, median home values in the most expensive counties have been around $600,000, whereas homes values in lower-priced areas were in the $200,000-$300,000 range. If pooled, transactions from the metropolitan area's counties would have generated a DC-wide median home value of roughly $300,000-$400,000. (The precise median home value would have varied over time and would depend on certain technical decisions). Had this median value been used for determination of the maximum loan limit, the area's loan limit likely would have been no higher than $460,000. Because the HCCR was applied, however, the median home value used for the entire metropolitan area was approximately $600,000, which is the median home price in the most expensive county. This means that the maximum Washington, DC loan limit was determined to be $625,500 for the last few years.
Seattle, which is comprised of three counties, including King County (the most expensive) is another example of where actual effects have been present. According to the National Association of Realtors, which does not apply a HCCR in computing median home values, the Seattle-area median was around $300,000 in 2012 and just under that in preceding years. With these median home values, the associated HERA maximum loan limit would have been $417,000. By contrast, because the HCCR only made use of transactions information for King County, where median home values were $400,000 and above, the loan limit for the entire metropolitan area was much higher at $506,000.
However, the overstatement in many places has had no impact on loan limits. In those metropolitan areas, the overstated median home value still was significantly below $362,600, which is the threshold value below which the maximum loan limit is merely set at the baseline level of $417,000.
Impact Analysis: Estimates
Given the elevated nature of existing loan purchase limits, analyzing the possible impact of a loan purchase limit decline is important. This impact analysis evaluates an across-the board decline—i.e., one that reduces loan purchase limits by the same 4.077 percentage in every county and county-equivalent area 
in the country. Per the planned declines, the baseline loan limit is reduced from $417,000 to $400,000, while the ceiling limit is reduced from $625,500 to $600,000.
In areas where loan limits are bounded by the baseline and ceiling, the loan limit has been reduced by the same percentage.
It is impossible to know with certainty the impact these reductions will have in 2014, but one analysis entails counting the number of acquired Enterprise mortgages with loan amounts above the lower loan purchase limits. Using a database of Enterprise loan acquisitions from 2012, Table 1 shows loan counts by state for the number of Enterprise-guaranteed mortgages with original loan amounts above the planned lower limits. Table 2 shows counts for 25 large Metropolitan Statistical Areas.
Table 1 reveals that, in 2012, roughly 170,000 Enterprise mortgages had original loan balances above the lower loan limits described in the Request for Public Input. This represented roughly 2.9 percent of total Enterprise mortgage acquisitions during 2012. About 50,000 purchase-money mortgages had balances above the lower limits.
Across states and MSAs, the share of mortgages with original balances near the applicable current loan purchase limit varied significantly. In Colorado—a state with a relatively large share of potentially impacted loans—roughly 6 percent of Enterprise mortgages (about 9,300 mortgages) had original balances above the reduced loan purchase limit. By contrast, only about one percent of mortgages in West Virginia and Start Printed Page 77453Alaska had balances in the affected range. Because loan amounts tend to be higher in urban areas than they are in states, the data in Table 2 reflect slightly larger shares of affected loans for MSAs. The shares of potentially impacted loans still remain relatively modest.
As indicated earlier, the mortgage counts reflected in the tables likely represent a substantial overstatement of the number of borrowers that might have been unable to obtain an Enterprise-eligible loan, or could be unable to do so in 2014. If loan purchase limits had been lower in 2012, some borrowers who took out loans in excess of the lower limit may have been able to modify their plans and borrow less (i.e., might still have taken out an Enterprise-eligible loan). In other words, whether by either increasing down payment or by taking out a larger second mortgage, some borrowers still would have had the ability to take out a loan that met the lower purchase limit.
A different and more sophisticated analysis would investigate, statistically, the relationship between the loan limit and the distribution of loan amounts. Not surprisingly, a large number of acquired Enterprise loans in 2012 had balances of exactly $417,000. Developing a statistical model that evaluates the size of the spike in the loan count that occurs at exactly the current loan limit would be valuable for estimating the size of the spike that would occur under a lower loan purchase limit. Unlike the prior impact analysis—which assumes that a borrower with a $417,000 mortgage would not have obtained an Enterprise-eligible loan if the limit were $416,999 or lower (i.e., the loan would have been “eliminated”)—a statistical model can implicitly account for borrower adjustments that would take place.
FHFA has been working on a model that might be used for such a purpose. While crude, a preliminary analysis suggests impact estimates that are roughly half of those produced in the simple approach.
Impact Analysis: Loan-Level Inspection
Although a statistical model would represent an improvement over simply counting mortgages in the affected range, an alternative analysis—one that makes use of loan-level information available to FHFA—is also available. Loan-level data can be used to identify options that would have been available to borrowers had loan purchase limits been lower. In doing so, one can remove from the set of eliminated loans mortgages for which borrowers would have had effective ways of responding to lower loan purchase limits. For example, data showing borrower cash reserves can be used to identify borrowers who, in response to a reduced loan purchase limit, would have had the demonstrated capacity to take out a smaller mortgage. Also, information about FICO scores and the loan-to-value ratio at origination can be used to identify borrowers who likely could have qualified for jumbo mortgages. Because interest rates for jumbo mortgages were only modestly higher than rates for Enterprise mortgages,
the “impact” of a borrower receiving a jumbo mortgage as opposed to an Enterprise mortgage would have been minimal. In this analysis, such borrowers are therefore excluded from the counts of impacted borrowers.
Using loan-level data, Table 3 shows the results of this more comprehensive approach for assessing the expected impact. The first row in the table repeats the impact number that was produced in the crude analysis—169,939. The second row estimates the number of mortgages that would have had balances above the new loan purchase limit and had combined loan-to-value (CLTV) ratios and FICO levels that may have made it difficult for the borrower to obtain jumbo financing.
Loans with FICO scores of either less than 720 or CLTV ratios above 80 percent were assumed to present potential difficulties.
The third row uses available information on borrower cash-on-hand to eliminate from the remaining sample borrowers who may have had the ability to take out a smaller mortgage.
Ultimately, after the various filters are applied, row 3 of Table 3 shows roughly 32,000 remaining mortgages. This means that, after accounting for loan characteristics and recognizing that jumbo financing would have been a reasonable alternative for many borrowers, the final impact of a loan purchase limit reduction might have only been about 32,000 loans. This figure is roughly 20 percent of the original crude impact estimate. Assuming that approximately 8.4 million mortgages were originated in 2012, the number reflects less than 0.4 percent of all 2012 loan originations.
It should be noted that the final impact analysis does not account for the availability of mortgages endorsed by the Federal Housing Administration (FHA). Some of the roughly 32,000 impacted loans may have been able to obtain FHA financing. While borrower costs would be higher (vis-à-vis Fannie Mae, Freddie Mac, and jumbo loans), such borrowers would have obtained mortgage rates that still were attractive from a historical perspective.
Impact Analysis: Characteristics of Impacted Loans
Table 4 attempts to answer: “What types of borrowers and what types of loans would be affected by the loan purchase limit reductions?” The table shows summary statistics for loans that the more comprehensive impact analysis suggested might be affected. The first column of the table shows summary data for roughly 32,000 loans identified in the comprehensive impact analysis.
The second column shows statistics for only the purchase-money mortgages contained in that sample. Approximately 40 percent of the affected loans were purchase-money mortgages. The final column shows statistics for only about 13,000 loans.
The table shows that potentially affected borrowers had relatively high incomes. The median 2012 household income for impacted borrowers who took out purchase-money mortgages was about $176,000—more than three times the national median. Twenty-five percent of such borrowers had household incomes of more than $229,000.
In general, the potentially impacted borrowers were attempting to either buy or refinance relatively expensive homes. Across all mortgage types, the median home value was $550,000, while the median sales price for purchased homes was around $520,000. Twenty-five percent of borrowers were attempting to buy homes valued at either $649,000 or more.
Although Table 4 shows many of the affected loans were in California, Illinois, Texas, Florida, and Colorado, these states collectively did not comprise a majority of the impacted loans. Combined, these states accounted for only about 40 percent of affected loans, suggesting that the effects of a loan purchase limit decline might have been geographically dispersed.
Impact Analysis: A Note About Home Prices
In light of the limited number of affected purchase-money mortgages, it would be reasonable to assume the market effects of a small loan purchase limit decline would be modest. Given the millions of single-family property transactions that occur each year in Start Printed Page 77454this country, the influence that around 13,000 purchase-money mortgages might have on home prices would seem to be relatively small.
Though not conclusive, historical evidence supports the expectation that the price effects will be modest. Loan limits decreased in certain high-cost areas in late 2011 with little discernible impact on observable prices. While no comprehensive analysis has been conducted into the effects of that reduction, post-reduction price increases—in many cases large increases—were evident in many of the most affected areas. For instance, Washington, DC, Los Angeles, San Francisco, and San Diego—cities that saw loan limit reductions of more than $100,000—experienced price increases in the following four quarters between 5.2 and 10.0 percent. These appreciation rates compared positively to the national appreciation over that period of 4.0 percent.
The late-2011 loan purchase limit reduction was geographically smaller in scope than the one contemplated for 2014.
In many areas, the 2011 loan limit declines were much larger than the planned 2014 loan purchase limit declines. Moreover, the 2011 reduction occurred in a fragile period for the housing recovery and appeared to have a limited impact during a fragile economic recovery period. This suggests that the impact of the contemplated 2014 loan limit reduction may be quite limited.
Start Printed Page 77455
Start Printed Page 77456
Start Printed Page 77457
Start Printed Page 77458
Start Printed Page 77459
Start Printed Page 77460
Start Printed Page 77461
Start Printed Page 77462
Start Printed Page 77463
Start Printed Page 77464
Start Printed Page 77465
BILLING CODE 8070-01-P
[FR Doc. 2013-30477 Filed 12-20-13; 8:45 am]
BILLING CODE 8070-01-C