Rethinking FHA

538 views
421 views

Published on

Published in: Business, Economy & Finance
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
538
On SlideShare
0
From Embeds
0
Number of Embeds
281
Actions
Shares
0
Downloads
2
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Rethinking FHA

  1. 1. RETHINKINGTHE FHAJ U N E 2 0 1 3KEY FINDINGSCurrent Federal Housing Administration (FHA) policy fails in its mission to help many first-time and financially con-strained homebuyers achieve homeownership. To make matters worse, it has failed to remain financially solvent.These twin failures underscore the need for bold reforms:• FHA is a policy failure: Far too many of FHA’s intended beneficiaries fail to achieve sustainable home-ownership. Recent research projects that between 15 and 30 percent of borrowers whose mortgages FHAhas guaranteed since 2007 will default.• FHA is a financial failure: FHA’s main mortgage insurance guarantee fund is under water. It does nothave sufficient funds to cover its expected losses, and its most recent actuarial review puts its net worth at a–$13.5 billion. My research concludes that FHA needs at least a $50 to $100 billion capital infusion to putit on a sound financial footing.• FHA’s business model is fundamentally flawed: Both FHA and the borrowers whose mortgagesit insures are leveraged by more than 30 to 1. To be viable, such a highly leveraged business model virtual-ly requires that housing values never fall. As we have learned from the recent housing crash, this is not a real-istic expectation.How to Reform FHAReplace FHA with a New Subsidized Savings Plan. Phasing out FHA over a period of years and replac-ing it would be far better than trying to reform such a flawed program. The replacement program would offer the fol-lowing benefits:• Simplicity: This straightforward plan would allow qualified households to pay in to a special savings vehi-cle and receive some type of match from the government. This stands in stark contrast to the current system,which requires a large bureaucracy to price a complex mortgage guarantee and to manage a difficult fore-closure process. The funds would accumulate on a tax-free basis until they were large enough to provide a10 percent down payment on a home.• Transparency: Focusing on borrowers helps ensure that program benefits accrue to the targeted house-holds—not to politicians or private actors such as housing financiers, realtors, or builders. Policy should subsi-dize those households directly rather than indirectly through an opaque mortgage insurance guarantee.Because costs would not be hidden (as those for Fannie Mae and Freddie Mac were), policymakers couldmore appropriately balance costs and benefits.• Sustainability: By incentivizing potential homebuyers to demonstrate financial discipline and long-termplanning, this program would help inculcate values that will decrease a household’s likelihood of default onceit purchases a home. And because the household will have developed meaningful equity in a home beforepurchase, it will be less vulnerable to fluctuations in the housing market.• Safety: Helping riskier borrowers build equity over time will result in a much safer and less leveraged hous-ing finance system. This benefits taxpayers and will result in much better housing outcomes for tens of thou-sands of borrowers.
  2. 2. A M E R I C A N E N T E R P R I S E I N S T I T U T ERethinkingthe FHAJoseph GyourkoJune 2013
  3. 3. Executive SummaryiiiThe Federal Housing Administration (FHA) has failedby any reasonable metric. Not only is its main mort-gage insurance guarantee fund insolvent in the sensethat it does not have sufficient capital resources tocover expected losses, but it is also failing far toomany of its intended program beneficiaries in helpingthem achieve sustainable homeownership. These twinfinancial and policy failures call for a fundamentalrethinking of how we might better achieve FHA’s mis-sion to help first-time homebuyers and financiallyconstrained households make minimal down pay-ments. This report argues that phasing out FHA overa period of years and replacing it with a new subsi-dized savings program for these households is far bet-ter than trying to reform FHA.Recent research projects high default rates—between 15 and 30 percent—among borrowerswhose mortgages FHA guaranteed since 2007. Hence,it is quite clear that very large numbers of programbeneficiaries are not successful in becoming stable,long-term owners. FHA’s most recent actuarial reviewfor fiscal year 2012 shows that its Single-FamilyMutual Mortgage Insurance Fund is underwater. Myown research suggests that even this sobering conclu-sion by FHA’s actuarial reviewer is too optimistic bytens of billions of dollars.FHA’s financial failure should not be so surprisingbecause its underlying business model is fundamen-tally flawed. Both FHA and the borrowers whosemortgages it insures are leveraged by more than 30 to1. This was always a financial accident waiting to hap-pen: this leverage ratio is on par with those that wereemployed by Bear Stearns and Lehman Brothers justbefore their collapses. To be viable, such a highlyleveraged business model requires that house valuesnever fall. We have learned the hard way that actualmarket outcomes are not always so obliging.Because it is folly to presume that prices willalways rise in the future, we should take advantage ofthe breathing room provided by the recent housingmarket recovery to fundamentally rethink FHA’s mis-sion to serve its targeted program beneficiaries. Theappropriate policy to support sustainable homeown-ership for these households should focus on buildingequity up front because the prime weakness of theFHA system is the virtual absence of meaningfulequity anywhere along its chain of operations.To do so, I propose replacing FHA with a new sub-sidized savings program that provides matches ofqualified households’ savings. The goal would be tohelp those households achieve a 10 percent downpayment on the home they wish to purchase. As Idescribe in more detail in the report, the benefits ofthis new program include its simplicity, its trans-parency, its encouragement of increased equity inAmerica’s financially fragile housing system, and itssubsidization of financial discipline and perseverance.These benefits stand in stark contrast to the com-plexity and opaqueness of the current FHA system,as well as its subsidization of extremely highlyleveraged bets on homes by financially fragilehouseholds. Not only will this new policy make theoverall housing market less risky, especially for tax-payers, but it will also help many more householdsrealize sustainable homeownership experiences.
  4. 4. IntroductionThe Federal Housing Administration’s (FHA) long his-tory in the housing market dates back to its creation in1934 during the Great Depression. However, its cur-rent policy focus was set by the National AffordableHousing Act (NAHA) of 1990, which encouraged theagency to help two groups—first-time homebuyersand financially constrained households without mean-ingful down payments—become successful owners.The NAHA envisioned FHA doing so by operating amortgage insurance guarantee program that would besustainable in the sense that it was self-supporting andwould not require taxpayer bailouts via capital infu-sions from the US Department of the Treasury.The evidence reviewed in this report shows thatFHA has failed both its intended program beneficiar-ies and US taxpayers. Recent estimates project cumu-lative default rates of between 15 and 30 percentamong borrowers who purchased since 2007, whenFHA began a major program expansion that morethan tripled the size of its mortgage guarantee portfo-lio.1 Thus, sustainable homeownership is not beingrealized by a large fraction of the buyers whose loansFHA guarantees. Moreover, FHA’s Single-FamilyMutual Mortgage Insurance Fund is insolvent.2 Thisinsolvency might still be justified if FHA were suc-cessful at helping its two target populations achievesustainable homeownership experiences. However, thetwin financial and policy failures raise obvious ques-tions of whether FHA should be allowed to continue,and if so, how to reform the agency.I recommend terminating FHA by slowly butsteadily phasing it out over the next few years andreplacing it with a subsidized savings program thatwould provide government matches of individualhousehold savings. A multiyear phaseout of at leastthree to five years is needed to ensure that there is nosudden negative shock to the residential mortgagemarket that would immediately depress the supplyof debt capital to the housing sector. The goal of thenew saving program is to help households build aneconomically meaningful 10 percent down paymentso that the private credit markets would be willingto provide an appropriately priced mortgage for sucha borrower.Eliminating FHA will sound extreme to some,given that the agency has existed for over three-quar-ters of a century. I recommend phasing it out for avariety of reasons beyond the facts that it is broke andhas not been able to discipline its risk-underwritingprocess so that substantially more of the borrowerswhose mortgages it insures are successful in becom-ing long-term owners. In particular, I fear that FHAcannot be successfully reformed. One of the impor-tant lessons of the collapses of the giant housing gov-ernment-sponsored enterprises (GSEs) Fannie Maeand Freddie Mac is that the incentives of politicians ofboth parties, of the senior executives of the GSEs, andof the owner-occupied housing industry to underes-timate program costs and loss estimates are verystrong and extremely difficult to counterbalance overtime. There are strong parallels with FHA.Perhaps even more important is that any govern-ment policy encouraging homeownership should bydesign target the current FHA system’s core weak-ness, which is the virtual absence of equity anywherealong its chain of operations. FHA itself is grosslyundercapitalized, as indicated by its own leverageratio of over 40 to 1.That is, it has over $40 of out-standing insurance guarantees for every $1 of what itcalls total capital resources available to pay off losses.This represents more than a threefold increase fromthe 12 to 1 ratio in fiscal year 2007. The borrowersRethinking the FHA1
  5. 5. whose mortgages it insures are almost as highlyleveraged at about 33 to 1 on average, given that thetypical down payment is no more than 3 percent.That this business model describes a financial acci-dent waiting to happen is highlighted by the fact thatthese leverage ratios are higher than those that wereemployed by Lehman Brothers and Bear Stearns justbefore those firms collapsed. To be viable, such abusiness model requires housing prices to never fall.Housing markets are not always so obliging, leavingtaxpayers to pick up the bill.Focusing homeownership policy on increasingequity via a subsidized savings program is appropriatein a number of other ways, too. First, the equity short-fall can be addressed solely by helping borrowersaccumulate a down payment without requiring a largebureaucracy to price a complex mortgage guaranteeand manage a difficult foreclosure process. Moreover,the experience of big mutual-fund complexes demon-strates that such a savings program can be operated atscale at low cost, allowing more of the required sub-sidy to be used to match borrower savings.Second, the focus on borrowers helps ensure thatprogram benefits are not siphoned off by politicians orprivate actors (for example, realtors and homebuilders)in the housing industry. Third, the program is transpar-ent in the sense that program costs are highly visible.They cannot be hidden as the true costs of Fannie Mae,Freddie Mac, and FHA were (and still are to someextent). This better allows us to right size the programby balancing costs and benefits on the margin.Fourth, it will increase domestic savings, which isa good thing for a country like the United States,which has been borrowing a lot of money from for-eigners. Fifth, it provides appropriate incentives andhelps inculcate useful values in borrowers. House-holds that are able to demonstrate the financial disci-pline and long-term planning demanded by amultiyear savings program are subsidized under thisplan. That is far different from the current program,which subsidizes the risky, highly leveraged gamblesthat Andrew Caplin and colleagues and Edward J.Pinto have shown are not paying off for a disturbinglyhigh share of FHA-insured borrowers.3 Of course,nothing is free, but the benefits of this new subsidyprogram would not be appreciably more expensivethan would FHA under reasonable assumptions. Anda much safer and less leveraged housing marketwould result in the bargain.The first section of this paper reviews recentresearch and analysis of the financial soundness ofFHA’s main mortgage insurance fund, as well as thecurrent and projected default experiences of bor-rowers FHA has insured since the beginning of thefinancial crisis. In the second section, I provide amore detailed discussion of my proposal to replacethe FHA with a new, subsidized savings programdesigned to help first-time and financially con-strained households amass an economically mean-ingful down payment equal to 10 percent of thevalue of the home they desire to purchase. Finally,my conclusion addresses some of the likely com-plaints regarding the termination of the FHA.What Have We Learned about FHAOver the Past Year?The Financial Condition of the Main Single-FamilyMortgage Insurance Fund. The most importantfinancial development over the past year is the formalacknowledgment in the latest actuarial review of FHAfor fiscal year 2012 that the economic value of itsmain single-family mortgage insurance fund is in thered by over $13 billion.4 This was news to many, andconfirmed my 2011 claim that this fund was econom-ically insolvent. It should not have been surprising, asFHA had been in violation of its 2 percent capitalreserve requirement since 2009. This policy guidelineimplies that FHA’s total capital resources available topay off claims should not be below 2.00 percent ofthe aggregate amount of FHA’s outstanding mortgageinsurance guarantee balance; those reserves on itsmain single-family mortgage insurance fund hadfallen to a miniscule 0.18 percent in fiscal year 2011.Obviously, they are now negative.This was virtually preordained by the fact thatFHA had nearly doubled its insurance guaranteeRETHINKING THE FHA2
  6. 6. portfolio since 2009, a time period characterized byfalling house prices and persistently high unemploy-ment. Much research into and experience withdefault tells us that negative equity and incomelosses from unemployment raise the likelihood ofdefault, with the combination of the two for anygiven borrower being particularly influential.5The only reason insolvency was not formallyacknowledged before the 2012 actuarial review wasbecause of systematic underestimation of default risk,which made the economic condition of the maininsurance fund look much less precarious than ittruly was. My 2011 AEI report detailed the four majorreasons why unobserved credit risk is so high in theFHA insured mortgage portfolio: (1) default estimateswere made without any attempt to control for unem-ployment risk; (2) no attempt was made to accountfor the high default risk resulting from unobservedgifts received by borrowers to fund down paymentsassociated with the first-time homebuyer tax creditprogram in 2009–10; (3) program actuarial reviewsunderestimated negative equity in the portfolio ofguaranteed loans because of the inappropriate use ofa house price index based on conventionally financedhome purchases; and (4) program reviews also under-estimated future default risk associated with FHA’sstreamline refinance program.US Department of Housing and Urban Develop-ment (HUD) leadership were for the most part dis-missive of this critique, but FHA’s outside actuarydoes appear to have taken some of my (and others’)input as the constructive criticism it was intended tobe.6 While the latest actuarial review did not addressthe issues of heightened default risk because ofunobserved gifting and underestimated negativeequity, FHA did try to control for unemployment inthe default models it used in the 2012 review byincluding the national unemployment rate (andchanges in that rate) in its estimated specifications.FHA also took some steps to improve its analysis ofstreamline refinance mortgages.7Unfortunately, as my 2013 paper with JosephTracy and my April 2013 AEI report show, theattempt to control for unemployment risk was notsuccessful.8 The two key conclusions are as fol-lows.9 First, FHA is still substantially underestimat-ing the heightened default risk associated with oneof its insured borrowers becoming unemployed. Avery conservative evaluation suggests that the degreeof underestimation at the individual borrower levelis very large—by a factor of at least 25, and possiblyby 100 or more.10 Stated differently, becomingunemployed raises the probability of default for agiven borrower by 25 to 100 times more than FHA’sbaseline models suggest. In economic terms, thismakes unemployment risk more influential in pre-dicting individual default than having even a low,subprime-level credit score (below 580) or a veryhigh loan-to-value ratio at origination.This should not be surprising, as becoming unem-ployed for any meaningful length of time implies ahuge fall in income for the typical household. Once ajobless household depletes its savings, it literally can-not make the monthly mortgage payment, whichcauses a default.11 In addition, FHA’s own surveys ofspecial servicers tell it that the primary reason a mort-gage became delinquent is because of a loss of income(typically from a job loss or substantial cutback inhours).12 Hence, this analysis accords with whatcommon sense, HUD’s own survey data, and basicdefault decision theory show—which is that unem-ployment risk has a huge impact on the probability ofdefault for a given borrower.From a taxpayer’s perspective, it is more impor-tant for FHA to correctly estimate the level of defaultJOSEPH GYOURKO3Much research into and experience withdefault tells us that negative equity andincome losses from unemployment raise thelikelihood of default, with the combinationof the two for any given borrower beingparticularly influential.
  7. 7. risk on its overall portfolio than it is for FHA to accu-rately predict when any one borrower will default.Unfortunately, a second important conclusion of my2013 AEI report and my and Tracy’s 2013 report isthat FHA’s most recent effort to include a market-levelunemployment rate in its default models does notappear to help FHA generate more reliable forecastsof default risk on its aggregate insurance portfolio.The relevance of this conclusion is that it implies thatunemployment risk largely remains unpriced in termsof evaluating the financial soundness of the FHAmortgage guarantee portfolio. Clearly, this is a criticalreason why FHA program risks have been systemati-cally underestimated in the past. This unfortunatelyalso implies that this is likely to be the case going for-ward because, as Tracy and I note in our 2013 paper,there is no immediate, short-term fix to this problem.It is important to note that the very large impact ofunemployment risk on the probability of defaultexists when the credit score of the borrower and theinitial loan-to-value ratio are held constant. Thus,these effects occur even though FHA’s guarantee poolhas a much higher average credit score than it has hadin the past. This may surprise some, but it is readilyunderstood. A credit score is based on a household’spast payment history and the length of that history. Itdoes not change when a household becomes unem-ployed, unless some type of late payment or delin-quency occurs simultaneously.Hence, credit scores should not be expected to cap-ture everything that there is to know about the creditrisk of a borrower; they clearly do not. Nor should oneexpect them to control for an unpriced risk such asthat associated with future unemployment. Hence,one should not anticipate that restricting loan guaran-tees to borrowers with high initial credit scores at thetime of loan origination will protect the insurance fundfrom future losses associated with macroeconomicrisks such as borrowers losing their jobs.In sum, new research helps us better understandwhy FHA systematically underestimates default riskand future losses. However, fixing those problems istechnically challenging. We know that current actuar-ial forecasts still underestimate default risk by a signif-icant amount, but not by exactly how much. Thus,going forward, we should still expect higher lossesthan FHA is currently projecting, which implies thatcontinuing to allow it to expand and essentially bet-ting that its future books of business will be so prof-itable that it can grow out of its present insolvency isunwise. The present insolvency condition will con-tinue and likely worsen absent a large capital infusionby the US Congress or dramatically stronger recover-ies in the labor and housing markets that materiallyreduce negative equity and unemployment risks.The Failure to Support Sustainable Homeowner-ship. The past year has also seen important newresearch that documents an alarmingly high failurerate of supporting sustainable homeownership formany of the buyers whose mortgages FHA guaran-tees. Caplin, and colleagues’ 2012 study of defaultrates on FHA’s 2007–09 books of business, whichconsidered data through September 2011, docu-mented that 15 percent of the borrowers insured byFHA in 2007–09 had already been in serious delin-quency in that they were at least 90 days late on theirloan payments.13 That is an important precursor toforeclosure, as in this data, more than 80 percent ofborrowers who become 90 days delinquent end uplosing their homes.14 Those authors also projectedthat more than 30 percent of all borrowers in the2007–09 pools were likely to default within the nextfive years. That projection is higher than FHA’s, but itis also more credible for reasons discussed earlierand in my 2011 report. That failure rate is also dou-ble the 15 percent of the borrowers that Caplin andRETHINKING THE FHA4Continuing to allow FHA to expand andessentially betting that its future booksof business will be so profitable that itcan grow out of its present insolvencyis unwise.
  8. 8. colleagues project to have paid off their FHA-insuredmortgages at any time within the next five years.15Pinto’s 2012 paper documents that this problemof financing unsustainable homeownership is notconfined to books of business from the beginning ofthe financial crisis. His research focuses on the morerecent 2009 and 2010 books of business guaranteedby FHA, and shows that high failure rates of 15 per-cent or more are projected among borrowers withincomes below their metropolitan-area median level.It is noteworthy that these borrowers tend to be geo-graphically concentrated in working-class neighbor-hoods across the country.16 Pinto points out that thismeans that the adverse ramifications of unsustain-able homeownership spells extend beyond the finan-cial and personal distress of the borrowers who losetheir homes. The negative spillovers on nearby prop-erty values imply that those who do not default arealso harmed.That many of these defaulting borrowers arephysically concentrated in the same low- and mod-erate-income neighborhoods in virtually all of thenation’s major metropolitan areas also increases thelikelihood of poor social, not just economic, out-comes for those places. FHA should not be heldsolely responsible for such outcomes, to be sure. Thefoundation of this problem is the risky subprimemortgage product that was not guaranteed by FHA.Still, there is little doubt that FHA’s policies and prac-tices are making that proplem worse, and that cer-tainly must be considered a serious policy failure.So What Should We Do about FHA?Shutting down a program that is insolvent and that isleading to far too many cases of unsustainable home-ownership is easily justified, and—many mightargue—necessary for signaling to others that suchfailure will not be tolerated. However, we need to becareful not to harm the intended program beneficiar-ies when doing so. Even if one disagrees with thepremise of FHA, the historical fact is that Congresshas charged it with the responsibility of helpingfirst-time buyers and financially constrained house-holds without sizable down payments become suc-cessful homeowners.That is the foundation on which reform propos-als such as that of Pinto’s 2012 paper rest. His rec-ommendations include tightening underwritingstandards so that no loan with a probable defaultrate of greater than 10 percent would be guaranteed.He also recommends that FHA aim for an aggregatedefault rate of 5 percent. Importantly, Pinto’s 2012report calls for risk-based pricing that recognizeslayered risks and prices them appropriately. By thishe means that some combinations of borrowertraits, such as low initial credit scores and high loan-to-value ratios, create higher default risk, and thatthose borrowers should be charged more.Individually and collectively, these recommenda-tions make very good sense. They would make FHAmuch safer, and many lower-income borrowers andthe neighborhoods in which they live would be muchthe better for them. However, I believe we shouldmove in a different direction that involves a more rad-ical reform of FHA and housing policy pertaining toFHA’s intended program beneficiaries. I recommendthat FHA be phased out completely over some well-defined period (for example, three to five years) andreplaced with a new subsidy program that would helpfirst-time homebuyers and other specially targetedgroups amass a 10 percent down payment that wouldthen allow them to obtain financing at market ratesfrom a private market lender.The appropriate housing policy to support trulysustainable homeownership should focus on build-ing equity up front because, as noted previously, theJOSEPH GYOURKO515 percent of the borrowers insured by FHAin 2007–09 had already been in seriousdelinquency in that they were at least90 days late on their loan payments.
  9. 9. dearth of equity throughout the FHA system is itsprime weakness.17 I envisage a simple system in whichqualified households would pay into a special savingsvehicle and receive some type of match from the gov-ernment. These funds would accumulate on a tax-freebasis until they were large enough to provide a 10 per-cent down payment on a home. Thus, if a typicalstarter home in the average market costs about$100,000, the program would be designed to helpqualified households amass $10,000 in savings.To reward success and financial discipline, bor-rowers would receive the matching funds only if theyachieved their targeted down payment by continuingto make their monthly savings contributions. Theirmonthly statements would clearly indicate thematching funds that would be at risk if they were towithdraw from the program. Some private firms inthe home mortgage sector use a similar scheme (forexample, the Loan Value Group).This strategy is alsoin accord with an important lesson learned frombehavioral economics regarding loss aversion: mostpeople find the pain that results from losing $1,000to be greater than the joy that results from winning$1,000. Hence, they will work hard to avoid losses.In this case, the program makes salient the extent ofthe financial loss resulting from the failure to con-tinue saving, with the obvious goal of nudging house-holds to stay in the program.Households that complete the program will be ina strong position to be successful in paying theirmortgage, both because they have meaningful equityin their homes from the outset, and because they havedemonstrated the ability to save consistently over alengthy period of time leading up to the purchase.18Lastly, I strongly prefer means testing of the house-holds to determine program eligibility so that the pro-gram is not open to all. But I leave those details tofuture development.My proposal focuses solely on the borrower sidebecause the entire policy goal can be addressed therewithout having to include a large, poorly run entitysuch as FHA on the other side of the policy intervention.This simplicity and singular focus help ensure that pro-gram benefits accrue to the intended beneficiaries—specific borrowers the government believes wouldbenefit from having access to homeownership—without getting somehow transferred to others in thereal-estate industry or government.It also makes the program costs more visible.This is essential for getting the size of the programright. The economically efficient outcome is a pro-gram size that guarantees that the last subsidy dollarspent generates social benefits equal to the costs tothe taxpayers who funded the subsidy. The cata-strophic failures of Fannie Mae and Freddie Macshow us that they were able to grow far too largebecause they “lowballed” costs by underestimatingprogram risks within their complex fields of opera-tions. For a considerable period of time, this allowedpoliticians, GSEs, and their allies in the owner-occu-pied housing industry to claim that substantial ben-efits were being generated without large costs beingincurred. In reality, the risks taken on were huge andso were the expected costs, but the opaque nature ofFannie and Freddie allowed that fiction to last farlonger than it should have. If this sounds like whathas been happening with FHA, that is because it is.Hence, any new program structure should be astransparent as possible about costs, and my proposalpasses that test because there would be no way tohide whatever matching dollars the federal govern-ment put into the household savings accounts.Moreover, we know there are mutual-fund com-plexes that could operate these accounts in largescale at low cost. Hence, a big public bureaucracy isnot needed (although the Internal Revenue Serviceprobably could play a useful role) so that a largerRETHINKING THE FHA6The appropriate housing policy to supporttruly sustainable homeownership shouldfocus on building equity up front becausethe dearth of equity throughout the FHAsystem is its prime weakness.
  10. 10. fraction of overall program costs could be devoted tothe government match amount.My proposal would also help increase domesticsavings, which would be good for the nation. In theproposal’s current form, the savings would be targetedtoward the owner-occupied housing industry. That isnot an unabashedly good thing, if you believe as I dothat the benefits of homeownership are overestimatedby many households. Given the government’s generalpro-homeownership stance and its explicit encour-agement of highly leveraged purchases by ownerswithin the FHA system in particular, it should not besurprising that some citizens would conclude thatbuying an expensive asset such as a home with littleor no money down must be a pretty good thing ingeneral and not very risky in particular.When combined with the natural incentives ofprivate-sector builders, of mortgage brokers, and ofrealtors to maximize profits by enhancing the visibil-ity of the benefits of owning and of reducing the vis-ibility of the costs, it should also not be surprisingthat many households have excessively optimisticexpectations regarding the financial returns tohomeownership. Thus, I certainly would not beopposed to including an educational component tothe program that helped people understand theactual tradeoffs between owning and renting, as wellas the true long-run costs of maintaining a home.19That well might lead one to allow the accumulatedfunds to be used for another purpose such as educa-tion, but that is a detail best left to future discussion.Last but not least, this program would also helpinstill appropriate incentives for subsidy recipientsby conveying the need to sacrifice and defer con-sumption to sustainably afford an expensive,durable good such as a home. Unless one had thevery good fortune to be born rich, this wise pru-dence has always been true, at least until policymak-ers and bankers decided during the most recentboom that meaningful equity was no longer neces-sary to buy homes. The last few years have high-lighted the folly of that decision. Moreover, thelong-term perspective this type of subsidy programwould encourage in program recipients stands instark contrast to current policy, which facilitateshighly leveraged bets by financially weak house-holds that have not shown the discipline to deferconsumption to save for a meaningful down pay-ment. As Caplin and colleagues documented in their2012 paper and Pinto documented in his 2012study, this has only led to more of those gambles fail-ing in recent years.This new subsidy program and its benefits arenot free, but continuing FHA mortgage guaranteeson highly leveraged borrowers is not free either. I donot provide a detailed cost comparison analysis inthis paper, but a very generous subsidy should notbe appreciably more expensive than FHA is likely tobe under reasonable assumptions.FHA guaranteed the mortgages on nearly750,000 home purchases in the last fiscal year(2012). That is a high number historically, as FHAincreased its market share substantially since theonset of the financial crisis. Hence, a sensible pre-sumption is that FHA will be guaranteeing no morethan 500,000 purchase mortgages a year by the timeit is wound down. With a few added assumptions,an estimate of the subsidy flow required from tax-payers can be provided.My back-of-the-envelope calculation presumes avery generous subsidy rate for modest-income house-holds earning $50,000 per year that purchase homesfor $150,000. Further assuming that no more than a2 percent gross savings rate is feasible for such house-holds and that we want them to be able to own withinfive years of entering the program, taxpayers would beresponsible for $2,000 per year per household in sub-sidy flow. That amounts to $1 billion per year across500,000 program recipients, plus the administrativecosts of managing the individual savings accounts.20Up until now, that would not have been expensivecompared to the true costs of operating FHA overpast housing market cycles. FHA’s latest actuarialreview reported that its single-family mortgage insur-ance fund had a net worth of –$13.5 billion at the endof fiscal year 2012. Others such as AEI’s Edward Pintobelieve that number is far too optimistic. Pinto’s cal-culations suggest a net worth of –$27.4 billion basedJOSEPH GYOURKO7
  11. 11. on generally accepted accounting principles accord-ing to his May 2013 FHA Watch.21 He puts FHA’s totalcapital shortfall in the $47 to $67 billion range. Myown research suggests that between $50 and $100billion is needed to recapitalize FHA. The point is thatwe can help many millions of households amassmeaningful down payments over many years for nomore than the likely costs of putting FHA on a soundfinancial footing after years of mismanagement.The cost comparison is less favorable for the newsubsidy program given recent changes to FHA.Based on FHA’s new upfront insurance premium of175 basis points and its higher annual fee of 125basis points, in their 2012 paper, Caplin and col-leagues calculate that FHA will roughly break even ifcumulative default rates fall much below 14 percent.That is still less than the default rates on recent booksof business according to Caplin and colleagues’paper, but one can imagine better outcomes for FHAas the housing market improves. If this happens, thesubsidy program will not be less expensive, but wewill still get a much safer and less leveraged housingmarket with a lower level of defaults and much bet-ter outcomes for tens of thousands of borrowers—and for relatively little added net cost.22ConclusionsSome homeownership advocates will argue that thispolicy proposal will harm the housing market andthe broader economy for a couple of reasons. One isthat it will lower the homeownership rate by increasingthe time many households remain renters while theyare saving for their down payment. Another is thateliminating FHA removes an important tool of coun-tercyclical economic policy.It will no doubt take most modest-income house-holds many years to amass a 10 percent down pay-ment, even with a generous government match rate.Ten percent of a modest, $100,000 starter home is$10,000. For a household earning $50,000 a year, anot unreasonable 2 percent gross saving rate gener-ates only $1,000 per year. That implies a decade-long period of saving given the lack of current yieldavailable on any safe savings vehicle. A 100 percentgovernment match would halve that period, withshorter savings periods requiring greater subsidies,as illustrated in my most recent example. Thus, tar-geted households will still remain renters for longerthan they would under the current regime, whichallows quicker access to owning via the subsidiza-tion of very low down-payment mortgages. Thatdelayed entry into homeownership will lower theoverall rate of ownership, so the first claim fromhomeownership advocates is accurate.What this would not do, however, is harm thebroader economy. The targeted households do notcease being part of the economic life of the countrywhile they are saving for a down payment. They aremerely renters, rather than owners. This will causetransfers between businesses in the owner-occupiedand rental housing sectors, with the former losingand the latter gaining. However, there is no reasonto believe that homebuilders are more deservingthan rental landlords, and thus that the governmentshould not care about this. More importantly, therewill be no first-order impact on the broader econ-omy. The funds not spent on buying a home will bedeployed elsewhere in the economy, includingthrough the lending out of the higher domestic sav-ings generated by this program. Just because one isnot buying a home does not mean a household’sincome or wealth disappears or somehow leaks out-side the national economy. It does not.In addition, the likelihood that the homeowner-ship experience will be sustainable will be muchRETHINKING THE FHA8The catastrophic failures of Fannie Mae andFreddie Mac show us that they were able togrow far too large because they “lowballed”costs by underestimating program riskswithin their complex fields of operations.
  12. 12. higher because there is a meaningful equity cushionto absorb the effects of negative economic shocks.Even if a lengthy unemployment spell occurs thatprevents payment of monthly debt service, default isnot the only possible outcome because the sale pro-ceeds are much more likely to be sufficient to pay offthe outstanding mortgage balance. Hence, the house-hold would not suffer long-term damage to its creditrating if it could not make monthly debt service pay-ments and had to sell.There is also no meaningful loss of a countercycli-cal policy tool from FHA’s elimination. This is notto deny that FHA and HUD officials have recentlyclaimed that the newly recognized losses on FHA’smain insurance fund are the price that had to be paidfor using FHA as part of a countercyclical economicpolicy. However, it is important to note that no suchclaims were made when FHA was expanding itsguarantee portfolio by nearly 400 percent since fiscalyear 2007, before FHA’s formal recognition of thenegative economic value of its insurance guaranteesthis year. In fact, HUD’s and FHA’s senior leadershipconsistently claimed that the huge increase in FHAactivity involved guaranteeing safe loans that wouldnot generate net losses to the insurance fund. Inaddition to relying on its outside actuary’s conclusioneach year that the main insurance fund was solvent,one of FHA’s standard defenses was to point to theincreasingly high credit scores of the borrowers itinsured, among other factors. Hence, this strikes meas little more than an ex post facto rationalization ofa major risk underwriting failure.23Even if I am wrong about that, it is hard to imag-ine a worse vehicle for countercyclical economic pol-icy than ramping up FHA’s insurance fund. Theproblem is that its structure correlates or concen-trates risk, as opposed to diversifying it. BecauseFHA guarantees pools of highly leveraged assets withlittle or no equity available to absorb losses in theevent of a general economic downturn, many mort-gages are likely to default at the same time. The rea-son is that most of the mortgages FHA guaranteesshare the same weakness (too little equity to absorbany property value declines), and thus are vulnerableto one common, negative economic shock. As I havewritten elsewhere, this means that FHA’s insurancefund shares a key risk trait with so-called CDO2s(collateralized debt obligations squared), which wereamong the first financial securities to fail and helpedinitiate the downward spiral when the financial crisishit.24 Simply put, countercyclical policy initiativesshould be implemented via the much less riskystructures and mechanisms available to the FederalReserve System and Congress. There is absolutely noneed to keep FHA around for this purpose.In addition, the recent claim by one economicconsulting firm (Moody’s Analytics) that FHA “savedthe housing market” through its actions in the after-math of the financial crisis does not justify keepingit around. I have not seen Moody’s Analytics’ report(nor do I know who funded it), so I cannot com-ment on the underlying statistical analysis that leadsit to conclude that house prices would have fallen byanother 25 percent and transactions volumes wouldhave fallen by as much as 40 percent had FHA notdone what it did. However, I do not find it implau-sible that in the midst of the greatest financial crisissince the Great Depression, abruptly shutting downthe one remaining guarantor of what had effectivelybecome a nationalized housing finance systemwould have had major negative consequences forthe housing market. Doing so clearly would havemade the market even more illiquid.25Of course, that still does not justify guaranteeinglarge fractions of poorly underwritten loans that haveeither failed or will do so over the next few years.26 Afar better strategy would have been to allow FannieMae and Freddie Mac to continue serving the marketfor borrowers with more substantial down payments.For political reasons, those two GSEs were virtuallyshut down from doing new business beyond refinanc-ing existing holdings after they were put into conser-vatorship. While politically understandable, it waseconomically unwise, as important liquidity for muchmore soundly underwritten loans could have beenprovided to the market. The data tell us that FannieMae’s and Freddie Mac’s losses were minimal for loansthey guaranteed that had more than a 10 percentJOSEPH GYOURKO9
  13. 13. down payment in front of the mortgage. Hence, thedamage to financially fragile borrowers and many localmarkets from FHA’s poor underwriting policies couldhave been avoided. Finally, even today, critics are notarguing for an abrupt agency shutdown. A multiyearphase out is required even now.Finally, there is no better time to fundamentallyrethink a program than right after it has failed. Andthere is no doubt that FHA has done so. It is runningan insolvent mortgage insurance guarantee platformwith a total risk exposure equal to about 7 percent ofnational output, and has shown itself unable to dis-cipline its underwriting process to prevent a muchtoo high share of its intended beneficiaries fromachieving sustainable homeownership. For FHA tobe broke on a program of this scale and to be failingat such a core part of its policy mission cannot be cat-egorized any other way. Some propose that FHAtoughen and reform underwriting standards so thatsensible risk-based pricing of layered risks canoccur.27 I greatly respect that position and agree thatthose reforms make excellent sense if the agency is tocontinue existing.However, I argue we should go in a differentdirection that involves phasing out FHA completelyover the next few years and replacing it with a newsubsidy scheme that directly encourages borrowersto save for a 10 percent down payment by providingmatching funds from the government. We need toaddress the lack of equity that characterizes FHA’sbasic business model so that taxpayers and house-holds that want to own homes are better served. Ibelieve that is best achieved via a subsidized savingsprogram that will enable deserving households toachieve sustainable homeownership, not via sometemporary access to owning based on extremelyhigh financial leverage that will always be depend-ent on the vagaries of the market and economy.NotesI am grateful to Henry Olsen for encouraging me tobegin studying this topic. I have benefitted from manyconversations and collaborations with Joe Tracy on thisand related issues, but he bears no responsibility for theconclusions (or any errors) in this report.1. For all references to the 2012 paper by Caplin andcolleagues, see Andrew Caplin, Anna Cororaton, and JosephTracy, “Is the FHA Creating Sustainable Homeownership?”(working paper, National Bureau of Economic Research,Cambridge, MA, June 2012); for all references to EdwardJ. Pinto’s 2012 paper, see Edward J. Pinto, How the FHAHurts Working-Class Families and Communities (Washing-ton, DC: AEI, December 2012).2. See Integrated Financial Engineering Inc., ActuarialReview of the Federal Housing Administration Mutual Mort-gage Insurance Fund Forward Loans for Fiscal Year 2012(November 5, 2012). For all references to my 2011 AEIpaper, see Joseph Gyourko, Is FHA the Next HousingBailout? (Washington, DC: AEI, November 2011).3. Caplin, Cororaton, and Tracy, “Is the FHA CreatingSustainable Homeownership?”; and Pinto, How the FHAHurts Working-Class Families.4. Integrated Financial Engineering Inc., Actuarial Reviewof the Federal Housing Administration.5. For a classic paper in this area, see Yongheng Deng,John Quigley, and Rober Van Order, “Mortgage Termina-tions, Heterogeneity and the Exercise of MortgageOptions,” Econometrica 68, no. 2 (2000): 275–307. Formore recent analyses of the so-called “double trigger”hypothesis that negative equity and income loss associatedwith unemployment combine to generate very high defaultprobabilities, see Chris Foote, Kristopher Gerardi, and PaulWillen, “Negative Equity and Foreclosure: Theory and Evi-dence,” Journal of Urban Economics 64, no. 2 (2008):234–45; Chris Foote et al., Reducing Foreclosures: No EasyAnswers (working paper, National Bureau of EconomicResearch, Cambridge, MA, June 2009); and Ronel Elul etal., “What ‘Triggers’ Mortgage Default?” American EconomicReview 100, no. 2 (May 2010): 404–94.6. The assistant secretary for research and developmentat the US Department of Housing and Urban Development(HUD) posted a response on an internal site called theHUDdle blog, which denied each of my claims. It wasavailable at http://blog.hud.gov/2011/11/18/continued-strength-fha, but has since been taken down.RETHINKING THE FHA10
  14. 14. 7. That latter effort is not reviewed here. It is quite com-plex, and we must rely on Caplin and colleagues to workwith FHA to determine how much of the underestimationof default risk relating to streamline refinancing has beenaddressed.8. For this and all other references to my 2013 AEIreport, see Joseph Gyourko, Unfounded Optimism: The Dan-ger of FHA’s Mispriced Unemployment Risk (Washington,DC: AEI, April 2013). For this and all other references tomy and Tracy’s 2013 report, see Joseph Gyourko andJoseph Tracy, “Unemployment and Unobserved CreditRisk in the FHA Single Family Mortgage Insurance Port-folio” (working paper, National Bureau of EconomicResearch, March 2013).9. For more details, see Gyourko and Tracy’s 2013 report.10. The reason for the huge underestimation of the trueimpact of unemployment risk on the probability of defaultis technical in nature and has to do with what statisticianscall “attenuation bias.” The underlying cause is that FHAdoes not observe whether any given borrower whose mort-gage it guarantees is employed at any point in time. Hence,FHA must use the market unemployment rate to proxy fora borrower’s unemployment status. Unfortunately, the cor-relation between the change in the aggregate unemploy-ment rate and any given borrower’s unemployment statusturns out to be quite low. Why? Because even a very highunemployment rate of (say) 10 percent does not mean mostborrowers have been laid off. In this case, 9 out of 10 stillhave jobs. Moreover, increases in that rate to 11 percent donot imply that most employed borrowers lost their jobs.They did not, as only 10 percent more were laid off. Thismakes the aggregate unemployment rate a noisy (measuredwith substantial error) proxy for any given borrower’sunemployment status, and noisy proxies cause attenuationbias, which essentially means that the true impact of becom-ing unemployed is underestimated. In this particular case,the degree of downward bias in the estimate is dramatic. SeeGyourko and Tracy’s 2013 report for the details.11. The combination of suffering a severe negative incomeshock and negative equity is known to be particularlyinfluential in leading households to default. For more onthese two triggers of default in the most recent housingcycle, see Foote, Gerardi, and Willen, “Negative Equity andForeclosure”; and Foote et. al., Reducing Foreclosures.12. For some recent data, see US Department of Housingand Urban Development, Annual Report to Congress, FiscalYear 2011 Financial Status, FHA Mutual Mortgage InsuranceFund (November 15, 2011), table 5.13. FHA’s terminology for a foreclosure event that leadsto an insurance fund loss is a “claim.” Analogously, whatare called “default rates” in this paper, FHA terms “claimsrates.” Technically, a default can be (and sometimes is)cured with no loss to FHA. I use these terms interchange-ably throughout this report.14. Caplin, Cororaton, and Tracy, “Is the FHA CreatingSustainable Homeownership?,” table 2.15. The typical way to pay off the loan is to have it refi-nanced outside the FHA system, which Caplin and col-leagues considered to represent a transition to sustainablehomeownership because it involved a nonsubsidizedfinancing by the private mortgage market.16. In his 2012 report, Pinto uses FHA’s projections offuture cumulative claims rates. I believe those estimatesare too low for the reasons discussed previously in thisreport and in my 2011 and 2013 papers. If this is correct,then his already very somber conclusions are likely toprove optimistic.17. In his 2012 report, Pinto also proposes helpinghouseholds build equity by using shorter-term mortgagesthat have faster amortization speeds. That is another excel-lent idea that could easily be added on to my proposal toamass sufficient equity before one takes out a mortgage.18. The importance of this latter point is demonstratedby the fact that FHA experiences very high default rates onborrowers who did not fund the meager down paymentscurrently required by their program rules. It is a very badcredit-quality sign if a household has to use gifts from anoutside source (including other family members) to fund adown payment.19. The costs of being a homeowner are not restricted tothe upfront equity investment needed for purchase and thesubsequent monthly mortgage payments. Housing is adurable good akin to a complex machine. All suchmachines depreciate in quality over time and ultimatelybreak down. The same is true of a home. Recent research byJames Poterba and Todd Sinai indicates that annual depre-ciation and maintenance costs amount to about 2.5 percentof overall house value. Note that this is not much differentJOSEPH GYOURKO11
  15. 15. RETHINKING THE FHA12from the magnitude of the initial down payment made bythe typical FHA borrower. However, it is an annual cost.Unlike one’s mortgage payment, maintenance can bedeferred, but doing so lowers the house’s value, which onlydiminishes the small amount of home equity most FHA-insured borrowers have in the first place. This is one amongmany reasons why many households underestimate the riskof highly leveraged home purchases—namely, they do notunderstand that the future costs of owning are going to befar higher than the initial down payment or even the regu-larly scheduled mortgage payments. See James Poterba andTodd Sinai, “Tax Expenditures for Owner-Occupied Hous-ing: Deductions for Property Taxes and Mortgage Interestand the Exclusion of Imputed Rental Income,” AmericanEconomic Review 96, no. 2 (May 2008): 84–89.20. A 10 percent down payment on a $150,000 homeamounts to $15,000, which implies $3,000 in savings areneeded per year for five years. A 2 percent gross savingsrate on households earning $50,000 annually implies thatthey would contribute $1,000 per year in savings ($50,000times 0.02 equals $1,000). In this particular case, the other$2,000 per year per household would come from the tax-payers. That reflects an extremely generous match rate, butprovides a ballpark estimate of required subsidy undersuch conditions.21. See Edward J. Pinto, FHA Watch 2, no. 5 (May 2013),www.aei.org/article/economics/financial-services/housing-finance/fha-watch-may-2013-vol-2-no-5/.22. That cost is about $1 billion per year given theassumptions used here. This also presumes that Con-gress and FHA maintain pricing discipline and do notlower their insurance fees as surpluses build before thenext housing market downturn. Otherwise, the cost dis-advantage rapidly turns into a cost advantage for the newsubsidy program. Given the natural incentive to under-price the insurance guarantee for political gain, thisseems a likely outcome and is another reason to replacethe current program with a new, more straightforwardsubsidy scheme.23. Page 1 of the introduction to HUD Secretary ShaunDonovan’s statement in his 2009 Annual Report to CongressRegarding the Financial Status of the FHA Mutual MortgageInsurance Fund is typical in this regard. It claims the 2009book of business has higher underwriting quality com-pared to previous years, with higher borrower credit scoresbeing a prime reason. The introduction to this annualreport and each subsequent one also explicitly notes that“sustainable” homeownership is a key goal and that under-writing standards are such that that goal will be achieved.See US Department of Housing and Urban Development,Annual Report to Congress Regarding the Financial Status ofthe FHA Mutual Mortgage Insurance Fund (November 12,2009).24. This argument and comparison with CDO2s is pre-sented in more detail in Joseph Gyourko, “The Govern-ment’s Overleveraged Housing Bet,” The American, November16, 2011, www.american.com/archive/2011/november/the-governments-overleveraged-housing-bet25. That said, it seems likely that the Moody’s Analyticsreport overestimates FHA’s true impact on prices. One rea-son is that, absent access to FHA’s subsidized guarantees, itis not clear how much further house prices would havefallen before investors would have entered the market insufficient quantity to allow prices to stabilize. Given theexisting supply of homes in the market, house pricesdepend on the overall demand for housing, not just thedemand from owner occupants. Stated differently, under-priced FHA financing guarantees may have crowded out atleast some investor demand for housing that would haveplaced the homes in stronger financial hands and helpedhold down upward pressure on rents. Moreover, whenmany of the FHA-insured mortgages subsequentlydefaulted, the resulting distress sales themselves likelyexerted downward pressure on home prices, thereby pro-longing the period of weakness in house prices.26. Caplin, Cororaton, and Tracy, “Is the FHA CreatingSustainable Homeownership?”27. Pinto, How the FHA Hurts Working-Class Families.
  16. 16. Joseph Gyourko is the Martin Bucksbaum Professor of RealEstate, Finance, and Business Economics & Public Policy at theWharton School of the University of Pennsylvania, where hechairs the real estate department and is director of the Zell/LurieReal Estate Center. He is also a research associate at the NationalBureau of Economic Research (NBER) and served as codirector ofNBER’s special project on housing markets and the financial crisis.13Author Biography
  17. 17. RETHINKINGTHE FHAJ U N E 2 0 1 3KEY FINDINGSCurrent Federal Housing Administration (FHA) policy fails in its mission to help many first-time and financially con-strained homebuyers achieve homeownership. To make matters worse, it has failed to remain financially solvent.These twin failures underscore the need for bold reforms:• FHA is a policy failure: Far too many of FHA’s intended beneficiaries fail to achieve sustainable home-ownership. Recent research projects that between 15 and 30 percent of borrowers whose mortgages FHAhas guaranteed since 2007 will default.• FHA is a financial failure: FHA’s main mortgage insurance guarantee fund is under water. It does nothave sufficient funds to cover its expected losses, and its most recent actuarial review puts its net worth at a–$13.5 billion. My research concludes that FHA needs at least a $50 to $100 billion capital infusion to putit on a sound financial footing.• FHA’s business model is fundamentally flawed: Both FHA and the borrowers whose mortgagesit insures are leveraged by more than 30 to 1. To be viable, such a highly leveraged business model virtual-ly requires that housing values never fall. As we have learned from the recent housing crash, this is not a real-istic expectation.How to Reform FHAReplace FHA with a New Subsidized Savings Plan. Phasing out FHA over a period of years and replac-ing it would be far better than trying to reform such a flawed program. The replacement program would offer the fol-lowing benefits:• Simplicity: This straightforward plan would allow qualified households to pay in to a special savings vehi-cle and receive some type of match from the government. This stands in stark contrast to the current system,which requires a large bureaucracy to price a complex mortgage guarantee and to manage a difficult fore-closure process. The funds would accumulate on a tax-free basis until they were large enough to provide a10 percent down payment on a home.• Transparency: Focusing on borrowers helps ensure that program benefits accrue to the targeted house-holds—not to politicians or private actors such as housing financiers, realtors, or builders. Policy should subsi-dize those households directly rather than indirectly through an opaque mortgage insurance guarantee.Because costs would not be hidden (as those for Fannie Mae and Freddie Mac were), policymakers couldmore appropriately balance costs and benefits.• Sustainability: By incentivizing potential homebuyers to demonstrate financial discipline and long-termplanning, this program would help inculcate values that will decrease a household’s likelihood of default onceit purchases a home. And because the household will have developed meaningful equity in a home beforepurchase, it will be less vulnerable to fluctuations in the housing market.• Safety: Helping riskier borrowers build equity over time will result in a much safer and less leveraged hous-ing finance system. This benefits taxpayers and will result in much better housing outcomes for tens of thou-sands of borrowers.

×