A New Direction for Housing Policy

Joe Gyourko

Spring 2015

Whatever one thinks of the Dodd-Frank bill and the Volcker Rule, they are evidence of meaningful reform of the banking sector in the wake of a major failure and crisis. Federal housing policy played a key role in the very same failure and crisis, yet a decade later the architecture of American housing policy remains essentially unchanged.

Just before the financial crisis at the end of 2007, the three primary agents of our housing policy — Fannie Mae, Freddie Mac, and the Federal Housing Administration — either guaranteed the outstanding balances on or owned nearly $5.5 trillion of home mortgages. They did so with about $112 billion in combined equity and excess reserves, meaning they were operating with $49 of aggregate risk exposure per dollar they held in equity or reserves. This also meant that, if a recession or a more serious crisis ever caused borrowers to default on their loans in large numbers, the agencies had an utterly inadequate cushion to handle their losses. By allowing Fannie, Freddie, and the FHA to operate this way, policymakers effectively presumed that the risk of default was negligible.

But the risk of default was very real, of course, and the total risk exposure amounted to nearly 40% of America's $14 trillion national output in 2008. Even modest failure on a base that large with so few reserves and a tiny equity cushion to absorb losses was bound to cause systemic risk, which is why Fannie Mae and Freddie Mac were essentially nationalized and put under conservatorship by the U.S. Treasury Department early in the crisis.

And yet despite all this, remarkably little has been done to prevent another housing crisis. On the contrary, the FHA, for instance, has more than tripled the volume of home mortgages it guarantees since the crisis began, while failing to maintain adequate equity and excess reserves. In 2012, it formally acknowledged that it does not have the resources to pay off all its expected losses and will need a capital infusion from the U.S. Treasury unless losses turn out to be much lower than expected. And this is a rosy scenario compared to what will more likely be the case: Independent researchers have found the degree of FHA's insolvency to be far greater than that reported by the agency's actuary.

The amount of FHA's aggregate mortgage guarantees has skyrocketed from just over $300 billion in fiscal year 2007 to nearly $1.1 trillion in fiscal year 2013. The latter figure amounts to nearly 7% of national output in 2013, so FHA has become important in terms of the overall economy, not just the housing sector. And yet, FHA is still smaller than Fannie and Freddie. Those giants are also badly in need of reform based on the lessons of the crisis, but the FHA is the logical place to start transforming the system. Because it is smaller and more targeted yet also growing swiftly, it offers an appealing and realistic target for fundamental reform.

There is no reason to imagine that depending on the same pre-crisis housing policy will result in anything other than another catastrophe. Real reform must involve more than tinkering with rules and requirements. Instead, policymakers should jumpstart the reform process by fundamentally rethinking how we want to encourage households to become homeowners. Ultimately, the FHA should be not so much changed as replaced with a subsidized savings program for first-time homebuyers and those who are not yet able to amass a meaningful down payment to purchase a home.

This would eliminate a clear source of policy failure that otherwise looks impervious to reform. And it would address the key defect of the current housing-policy infrastructure by encouraging the build-up of equity in the subsidized housing-finance system. Any sound plan for reform must address this equity deficiency in order to dramatically reduce the risk involved in federal housing policy. An FHA reform of this sort would also better target the program's benefits toward the financially fragile households who have been particularly ill-served by current and past policy.

Helping these citizens access homeownership without taking on excessive financial risk is both the right thing to do for them and the right way to protect taxpayers.


The Federal Housing Administration was established back in 1934 to help deal with the fallout of the Great Depression in housing markets, but its current aim and focus were established more recently by the National Affordable Housing Act (NAHA) of 1990. The severe real-estate market downturn of the late 1980s stressed FHA sufficiently that an outside accounting firm deemed it so close to insolvency that restructuring and recapitalization were required in order for the agency to survive. As part of that reform, the NAHA envisioned FHA helping two groups of people — first-time homebuyers and financially constrained households without meaningful down payments — become homeowners by guaranteeing their mortgages in the event the borrowers defaulted. The guarantee permitted lenders to make loans to these riskier borrowers without having to charge prohibitively high interest rates.

The agency's broad policy goals, therefore, have been to create successful and sustainable homeownership experiences for these borrowers and for the guarantee fund itself to be self-supporting — in the sense that it would not require taxpayer bailouts through capital infusions from the Treasury. Unfortunately, FHA has been failing lately on both fronts. It has failed the borrowers it guarantees by operating a program that, according to recent research, has cumulative default rates as high as 35% for borrowers who bought after 2007. Sustainable homeownership is not being realized by far too high a fraction of borrowers whose loans the FHA guarantees. And it has failed the taxpayers because its main Single-Family Mutual Mortgage Insurance Fund has become insolvent, in the sense that it does not have the resources necessary to pay off all its expected losses. My research has suggested this was the case as far back as 2011; the agency's own actuary came to the same conclusion in fiscal year 2012.

The FHA has always experienced relatively high default rates among the borrowers whose mortgages it insures. This is not surprising given that it guarantees loans on homes with very small down payments. Borrowers with positive equity in their homes (meaning that their house value exceeds the balance on the mortgage-debt owed) are much less likely to default.

Statutorily, FHA can backstop loans with equity down payments as small as 3.5% of purchase price, which implies a so-called "leverage ratio" of 29 for the borrower. The leverage ratio is calculated by dividing the value of the asset (the house) by the equity owned in it (the down payment). The higher the leverage ratio, the more risk involved. A leverage ratio of 29 is very high, and such a borrower's true leverage is likely even higher, as FHA permits purchasers to borrow the cost of the upfront insurance premium it charges to help compensate for future default risk.

Data that Fernando Ferreira and I have laid out elsewhere suggest the typical loan-to-value ratio on a new FHA-insured loan is about 98%, for an implied leverage ratio of 50. How risky is a leverage ratio that high? Lehman Brothers and Bear Stearns were found to be operating with leverage ratios in the range of 30 to 40 just before they collapsed. From the perspective of such a highly leveraged borrower, house prices do not have to fall much at all before he has negative equity in his home. Default probabilities increase sharply when that happens, and they spike even higher if having negative equity is coincident with the loss of a job. This is often the case, as recessions tend to increase unemployment and lower house values simultaneously. Thus, default risk on the borrowers whose mortgages FHA guarantees is bound to be high, unless one imagines that house prices will only go up and recessions will not happen in the future.

Even if we concluded that high double-digit default rates were tolerable from a policy perspective (and we should not), the FHA still could shield taxpayers from the fiscal cost by amassing large enough reserves to cover the inevitably high expected losses among owners who start with so little equity in their homes. But, as noted above, this has not happened. The structure of today's FHA is effectively rooted in the hope that default risk would be low so that high premiums would not have to be charged. Hope is rarely a sound basis for policy, and taxpayers are left exposed to the risk of being forced to bail out another federal housing agency

It is not easy to see how the FHA as it now exists can escape this very dangerous dynamic. It is the very structure of our housing policy that is weak. It needs to be replaced with something different, better targeted, and stronger.


Shutting down a program that is fiscally unsound and is leading far too many families into unsustainable homeownership is easy to justify — if only to signal to others that such failure will not be tolerated. But policymakers seeking such a shutdown need to be careful not to harm the program's intended beneficiaries. Some undoubtedly disagree with the very premise and purpose of the FHA. But Congress created the agency, and has continued to authorize it, with the clear purpose of helping first-time buyers and financially constrained households without sizable down payments become successful homeowners. If policymakers still want to pursue that goal, it is time for them to develop different means of doing so.

They should begin by completely phasing out the FHA over some clearly defined period (for example, three to five years) and replacing it with a new subsidy program that would help the two types of households the agency is meant to serve. The new program would help prospective homebuyers amass a 10% down payment that would then allow them to obtain financing at market rates from private lenders. This could be done through a simple system in which qualified households pay into a special savings vehicle and receive some type of match from the government. These funds would accumulate on a tax-free basis until they were large enough to provide a 10% down payment on a home.

It is difficult to estimate precise program costs without a detailed design, but it is possible to perform sensible, back-of-the-envelope calculations that show that this new subsidy would not be more expensive than the FHA's existing single-family mortgage insurance guarantees when they are properly accounted for.

To illustrate the structure of the program, let's assume that the typical eligible household would earn roughly $50,000 per year and would seek to purchase a $150,000 home. A 10% down payment on that home would require $15,000 from the borrower, with the rest financed by a $135,000 mortgage. Further presuming that we would want a household entering the program to be able to buy the home within five years, we can estimate that $3,000 in total savings would be needed each year. How much the household would need to contribute each year would depend upon what one thinks is a reasonable target savings rate. It is safe to assume that no more than a 2% gross savings rate is feasible for such modest-income households. As 2% of $50,000 would be $1,000, we should therefore assume that a $1,000 annual contribution by the household would be achievable, with the government contributing a very generous $2,000 match, so that $3,000 in total savings could be realized each year.

Note that in this example, the equity provided by the targeted household is more or less equal to the amount it would have provided for a standard FHA-insured loan on the same home; the $5,000 they contribute in the new program is equal to a 3.3% down payment on a $150,000 home. Thus, the equity burden on the household is no greater than it would have been under today's system. However, the purchase is much safer for the household (as well as the taxpayer) because there is an additional $10,000 in equity to cushion against any future declines in property values.

FHA has been guaranteeing the mortgages of about 750,000 borrowers per year over the past several years. Contributing $2,000 per year for each of those borrowers would require $1.5 billion in spending by the government, plus the administrative costs of running the individual accounts. If policymakers think those costs are too high, it would be relatively easy to lower them by requiring the subsidized borrowers to repay at least part of the $10,000 in matching funds provided by taxpayers out of future appreciation. This would make the program more complex, as we would not want to take so much away that the household could not fund the down payment for its next home out of housing-capital gains. Naturally, if there were no increase in the value of the home over time, the borrower would not have to pay anything more.

Even without some payback out of future appreciation, this very large, hypothetical subsidy case would not be nearly as expensive as the true costs of operating FHA in its current form. FHA's actuarial review reports its Single-Family Mutual Mortgage Insurance Fund had a net worth of -$7.9 billion at the end of fiscal year 2013 (that is, to be clear, negative $7.9 billion). Its most recent review for fiscal year 2014 claims a positive net worth of $5.9 billion. However, independent researchers believe those numbers are far too optimistic. For example, Edward Pinto's calculations from the end of 2013 suggest that FHA has a negative net worth of $25 billion and is suffering a total capital shortfall of $52 billion. My own research from 2012 suggests that between $50 and $100 billion is needed to recapitalize FHA on a sound basis. The point is that we can help many millions of households amass meaningful down payments over decades for no more than the likely costs of putting FHA on a sound financial footing. And we would get a much safer and less leveraged housing market in the bargain.


No sound reform proposal should want to mimic the way today's policy encourages financially fragile households to make risky, highly leveraged bets to become homeowners. It is essential that we not only reduce risk in housing policy for taxpayers, but also that we design programs that allow much larger fractions of targeted households to achieve sustainable homeownership.

The approach suggested here would do so by conveying the need to sacrifice and defer consumption if one is to sustainably afford an expensive, durable good like a home. Once upon a time, this used to be obvious to homeowners (except perhaps those with the good fortune to be born rich) — until policymakers and their private-sector allies decided during the most recent boom that meaningful equity was no longer necessary to buy homes. The last few years have highlighted the folly of that decision: Recent research and experience have shown how high default rates can become when financially weak households purchase homes with small down payments comprising too little equity and then suffer some negative health or economic shock that impairs their ability to pay their mortgages or maintain their homes. Moreover, the best recent research shows that African-American and Hispanic households were especially vulnerable in this regard. We need to be much more humble about the risks we encourage modest-income households to take. Hippocrates's wisdom for physicians — first do no harm — applies equally well to all housing-affordability advocates.

To reward success and financial discipline, borrowers under this proposed approach would receive the matching funds only if they continued to make their monthly savings contributions toward their target down payment. While no personal savings that households contributed would ever be at risk, monthly statements would clearly indicate that the matching funds from taxpayers could be lost if the participating household withdrew from the program.

This would incorporate an important lesson from behavioral economics regarding what psychologists call "loss aversion": Most people find the pain that results from losing $1,000 to be greater than the joy that results from winning $1,000. Hence, they will work hard to avoid losses. In this case, the rule would make salient the extent of the financial loss resulting from the failure to continue saving, with the obvious goal of nudging households to stay in the program. Households that do withdraw almost certainly should not have become owners in the first place, because if they cannot save $1,000 a year before they buy their home, they would not be able to handle the financial sacrifices necessary to remain homeowners.

Just the monthly debt service on a typical, fully amortizing, 30-year, $135,000 mortgage with a low 4% interest rate is about $645, or nearly $7,740 per year. That is not the only ongoing cost that must be made to avoid default. Housing is an expensive durable good with high operating costs, including property taxes, insurance, and ongoing maintenance and repairs. The sum of just those costs on a $150,000 home will be much more than $1,000 per year. Typical effective property-tax rates are between 0.5% and 1.0% per year, which works out to required payments of $750 to $1,500 per year. Standard fire and hazard insurance on a home would put everyone over $1,000 per year. And that is before normal maintenance spending, which research suggests is at least 1% of a home's value per year. Maintenance can be deferred, but that can cause the home to decline in value. Hence, the screening done by this program could well save these households from life-wrenching defaults in the future.

Households that complete the matched savings program would be in a much stronger position to be successful in paying off their long-term mortgage for two reasons. First, they would have meaningful equity in their homes from the outset. Second, they would have demonstrated the ability to save consistently over a significant period of time leading up to the purchase. Default rates under this program would undoubtedly be much lower.

It is important to note that this proposed new program would focus exclusively on the borrower side of the market. There would be no need for a bureaucracy like FHA to manage the risk associated with a complex and long-term mortgage guarantee. The main program goals — achieving sustainable homeownership among targeted families without taxpayers taking on huge risk — can be achieved simply by encouraging the build-up of equity up front. Equity always is needed to absorb losses when life takes a turn for the worse. As we have recently learned to our great cost, housing markets are no exception to this fundamental truth of financial economics. This proposal would be an important step in de-risking America's housing-policy architecture.

The simplicity of the program structure has other advantages, too. One is that it helps ensure that program benefits accrue to the targeted households without being siphoned off by others. The more complex the program and the bigger the bureaucracy, the more likely it is that some parties in the private sector will figure out how to transfer part of the subsidy benefits to themselves. This is not meant to cast aspersions on those private actors but to acknowledge reality: The senior leaders of those organizations are fiduciaries for their investors, which means they have a legal duty to act in their best interest, and that includes figuring out how to benefit from government subsidy programs. In addition, mutual-fund complexes could serve an important role in this new program, as they are capable of operating these accounts on a large scale at relatively low cost. If we could depend on these organizations, we would not need to pay for a big public bureaucracy (although the Internal Revenue Service probably could play a useful role), and a larger fraction of overall program costs could be devoted to funding the government-match amount.

The costs of this program would also be highly visible. A key reason why Fannie Mae and Freddie Mac were allowed to become so large before the crisis and why FHA was allowed to almost quadruple in size after the crisis was that they were able to "lowball" costs by underestimating program risks within what is a complex mortgage-guarantee business. This allowed politicians, the housing agencies, and their allies in the private sector to claim that substantial benefits were being generated without large costs being incurred. In reality, the risks taken on were huge and so were the expected costs, but the opaque nature of Fannie, Freddie, and FHA allowed that fiction to last far longer than it should have.

The sheer opacity of these programs is further demonstrated by the way the government now exploits the conservatorship status of Fannie and Freddie to undercount the future costs of housing policy in a way that conveniently offsets current reported budget deficits. It does so by treating the net funds of Fannie Mae and Freddie Mac as pure profit that is then used to lower the federal deficit so that spending on other programs can be higher. This is a perverse political fiction that bears no resemblance to financial reality. Essentially, the positive net cash flows of Fannie and Freddie in today's relatively strong economy are treated as riskless profit, in the sense that money will never be needed to cover losses from future defaults. For that to be true, it would also have to be true that we will never again see declines in house prices or recessions and the job losses that come with them. That we seem unable to stop ourselves from engaging in such fantasies is an important reason why the very architecture of housing policy needs to be simplified. The more complex the subsidy program (and valuing the guarantee on a long-term mortgage is very complex), the easier it is to hide silly assumptions in the analysis.

A program like the one proposed here would undoubtedly be far more transparent. It would not be easy to hide the transfer of a matching dollar into a private individual's savings account, so the cost of the program would always be evident. This would help its value be evident, too: Ideally, the last subsidy dollar spent on a program like this should generate social benefits equal to the costs to the taxpayers who funded the subsidy. Without a great deal of transparency, it is impossible to know when that happens and when it does not.


To the extent that reforms of our housing policy have been under discussion lately, they have taken a very different form than the one proposed here. But the danger of those proposals only reinforces the case for a more fundamental transformation of housing policy, beginning with the FHA, because they threaten to make our housing-finance system less safe, not more.

Effective January 26, 2015, the FHA reduced the annual premium it charges on the extremely highly leveraged loans it guarantees. This fee reduction occurred even though the FHA remains in violation of its mandated capital-reserve ratio of 2%. It has been in violation since 2009, and its current capital-reserve ratio is only 0.41%, according to FHA's own annual report for fiscal year 2014. Needless to say, a well-run insurance program would be building a cushion for future defaults in the next downturn, not reducing fees to a level that other research suggests is more appropriate for conventional borrowers with 10% to 20% equity down payments.

Even more disturbing is the recent announcement that Fannie Mae will once again guarantee mortgages with as little as 3% equity down payments. As discussed above, such highly leveraged, 97% loan-to-value loans traditionally have been the province of FHA. It is discouraging to say the least that the government regulator, the Federal Housing Finance Agency, believes that encouraging financially weak households to engage in 33-to-1 leveraged home purchases is a practice that should be expanded to other housing agencies.

These proposals reflect an utter lack of imagination on the part of policymakers, as evidenced by the fact that they are virtual replays of policies implemented before the financial crisis. Worse still, they will not serve the interests of targeted households any better. It is imperative that we move in a new direction.


Some homeownership advocates would likely argue that the government-matching policy proposed here would harm the housing market and the broader economy for two reasons. First, it could lower the homeownership rate by increasing the time many households remain renters while they are saving for their down payments. And second, eliminating the FHA would remove an important tool of countercyclical economic policy. The first claim is true, but should not trouble us. The second is false.

Even with the very generous match rate described above, it still would take most modest-income households years to amass a 10% down payment for a home. Thus, there is no doubt that targeted households will be renters for longer than they would be under the current regime, which allows immediate access to ownership through the subsidization of mortgages with very low down payments. That delayed entry into homeownership would indeed lower the overall rate of homeownership. But that would not harm the broader economy.

Simply put, the targeted households would not cease being part of the economic life of the country while they were saving for down payments. They would just be renters rather than owners. This would cause transfers between businesses in the owner-occupied and rental housing sectors, with the former losing and the latter gaining. But there is no reason to believe that homebuilders are more deserving than rental landlords, so the government should remain impartial. It certainly should not take on substantial risk to benefit homebuilders relative to rental landlords, because having a lower homeownership rate would have no first-order impact on the broader economy. The funds not spent on buying homes would be deployed elsewhere in the economy, including through the lending out of the higher domestic savings generated by this program (yet another benefit of this proposal). Just because a household is not buying a home does not mean its income or wealth disappears or somehow leaks outside the national economy.

Furthermore, eliminating the FHA would not mean surrendering a meaningful countercyclical policy tool. In fact, it is hard to imagine a worse vehicle for countercyclical economic policy than increasing the use of FHA's insurance fund, as has been done in recent years. The fund's structure correlates or concentrates risk, rather than diversifying it. FHA guarantees pools of similar, highly leveraged assets with little or no equity available to absorb losses in the event of a general economic downturn. Because these loans share the same weakness, they are vulnerable to one common, negative shock. This means that if one borrower defaults, many others are likely to do so at the same time. Aggregate risk is higher because of the correlation across individual borrowers' loans. Any countercyclical policy initiatives should be implemented through the much less risky structures and mechanisms available to the Federal Reserve System in monetary policy and Congress in fiscal policy. Depending on the FHA to serve this purpose is nonsensical and counterproductive.


Almost a decade has passed since the housing crisis helped unleash the last recession. There is no excuse for failing to reform American housing policy.

Any reform needs to accomplish two goals. First, it must dramatically reduce the risk to taxpayers inherent in federal housing policy. This requires a subsidy structure that is not easily gamed politically and would not leave taxpayers exposed to large losses in the event of a future economic downturn. Second, it must avoid encouraging financially fragile households to make what are little more than highly leveraged gambles on homes that may or may not increase in value over time.

It would not be easy, and could well be impossible, to achieve these twin goals by just turning the dials of today's FHA. Achieving the agency's goals now requires eliminating the agency entirely and replacing it with a new subsidy scheme that directly encourages modest-income households to save for a 10% down payment by providing matching government funds. This would directly address the lack of equity that characterizes FHA's basic business model so that both taxpayers and modest-income households who want to own homes are better served. And it would be less expensive than the true costs of FHA going forward.

Without some reform of this type, we could not really claim that we have learned the lessons of the last housing crisis — or that we have taken the steps necessary to prevent the next one.

Joe Gyourko is the Martin Bucksbaum Professor of Real Estate, Finance and Business, and Public Policy at the Wharton School of the University of Pennsylvania, where he also serves as director of the Zell / Lurie Real Estate Center.


from the


A weekly newsletter with free essays from past issues of National Affairs and The Public Interest that shed light on the week's pressing issues.


to your National Affairs subscriber account.

Already a subscriber? Activate your account.


Unlimited access to intelligent essays on the nation’s affairs.

Subscribe to National Affairs.