VoxEU Column Financial Markets Global crisis

A proposal to reform the US mortgage finance

At the centre of the global financial crisis was a housing boom and bust. This column continues the description of how flaws in the US housing finance sector made the crisis inevitable. Here the authors outline a reform plan to avoid this outcome and contrast it to the US Treasury proposals.

The goal of reforming housing finance should be to ensure economic efficiency, both in the primary mortgage market (origination) as well as in the secondary mortgage market (securitisation). By economic efficiency, we have in mind a housing finance system that:

  • corrects any market failures if they exist; notably the externality from originators and securitisers undertaking too much credit and interest rate risk as this risk is inherently systemic in nature;
  • maintains a level-playing field between the different financial players in the mortgage market to limit a concentrated build-up of systemic risk; and
  • does not engender moral hazard issues in mortgage origination and securitisation.

Motivated by economic theory, we argue that such a mortgage finance system should be primarily private in nature. It should involve origination and securitisation of mortgages that are standardised and conform to reasonable credit quality. The credit risk underlying the mortgages should be borne by market investors, perhaps with some support from private guarantors. There should be few guarantees, if any, from the government.

Genie in the bottle

The question is how do we effectively get to this private system given the current state of mortgage finance? We call this the “genie in the bottle” problem. A quarter of a century ago, the proverbial “genie” was let out of the bottle when mortgage markets were deregulated yet left the government guarantees and special treatment of Fannie Mae and Freddie Mac (the “government-sponsored enterprises”, or GSEs) in place. Capital markets over the past 25 years have developed to be reliant on these guarantees. To wean the system off these guarantees – to put the “genie back in the bottle” – we need to transition away from a government-backed system to a private-based one. The problem is that the transitional process will only succeed if private markets are not crowded out, regulatory capital arbitrage by private guarantors is averted, and systemic risk inherent in mortgage credit and interest rate risks is managed.

We envision that the initial phase of this process would preserve mortgage-default insurance because such guarantees have been essential for the way that the securitisation market for mortgages has developed. However, the government share of these guarantees would be steadily phased out. To achieve this, the transition should include a public-private partnership in which the private sector decides which mortgages to guarantee and sets the price for the mortgage guarantees but insures only a fraction (say 25%), while the government is a silent partner, insuring the remainder and receiving the corresponding market-based premiums. The public sector involvement should be limited to conforming, tightly underwritten mortgages (for example, to mortgages with loan to value ratios that are at most 80%). The private sector mortgage guarantors would have to be regulated to be well-capitalised and subject to an irrefutable resolution authority. This way, the market pricing of mortgage guarantees will reflect neither explicit nor implicit government guarantees. And, the government guarantees being offered in passive partnership to private markets, and importantly, at private market prices, will ensure that the private sector is not crowded out.

We envision that if such a transition plan were followed, then the private sector would be encouraged to shrug off any regulatory uncertainty and allowed to flourish. Financial innovation in these markets could return. New investors that are focused on the credit risk of mortgage pools would emerge. Mortgages would become more standardised and underwriting standards would improve. To help the transition process along its way to an efficient mortgage market in the long run, reliance on the GSEs’ guarantees should be mandated to end in a phased manner. One example of such a mandate would be a gradual reduction of the size limit for conforming mortgages; another would be an increase in the fees that the GSEs charge for their guarantees. These mandates could be implemented sooner if the private capital market develops more quickly.

Although our book, Guaranteed to Fail: Fannie Mae, Freddie Mac, and the Debacle of Mortgage Finance, was written before the Obama Administration’s recently announced plan, there is much common ground between the two: (a) The GSEs should be wound down; and, (b) Efforts to assure housing affordability for low- and moderate-income households should be explicit, on-budget, and primarily the domain of the Federal Housing Administration. However, the Administration does not currently have a specific proposal for the long-run future role of government guarantees in the US housing finance. Instead, the Administration offers three possibilities, without indicating its preference:

  • a wholly private structure;
  • a largely private structure, but with an agency that would provide guarantees to new mortgage-backed securities at times of severe stress in the mortgage markets; or
  • a largely private structure, with a government agency providing “tail-risk” or catastrophic insurance in the event that a private mortgage guarantor defaulted on its obligations.

We believe that the first of these three possibilities is the appropriate long-term goal; but we believe that our transition plan offers a superior means of getting there. In particular, the Federal Reserve is already the agency for dealing with general and severe stress in financial markets, including mortgage-backed securities, so that any additional effort by a new agency would be duplicative. Further, we believe that the “tail-risk” government insurance will inevitably be underpriced and thus will likely end up being a “back-door” means of subsidising general mortgage borrowing. Our proposal calls for the government providing side-by-side guarantees – only in the interim – and would explicitly use the market-based pricing of the private guarantors. Also, our proposal will encourage the private sector to develop tail-risk insurance capabilities, which can then expand and replace the government; the advocates of option iii have no such phase-out scenario.

With these overall remarks, we now provide a more detailed evaluation of the Obama Administration’s plan.

The Obama Administration’s Plan

From a level of 30,000 feet, it is hard to argue against the fundamental premise of the administration’s plan for mortgage finance, and, in particular, for the GSEs. Their various plans all call for effectively winding down and eventually shuttering Fannie and Freddie and, as a replacement, for a privatised system of housing finance with little government involvement:

Under our plan, private markets – subject to strong oversight and standards for consumer and investor protection – will be the primary source of mortgage credit and bear the burden for losses. Banks and other financial institutions will be required to hold more capital to withstand future recessions or significant declines in home prices, and adhere to more conservative underwriting standards that require homeowners to hold more equity in their homes. Securitisation, alongside credit from the banking system, should continue to play a major role in housing finance subject to greater risk retention, disclosure, and other key reforms. Our plan is also designed to eliminate unfair capital, oversight, and accounting advantages and promote a level playing field for all participants in the housing market. The Administration will work with the Federal Housing Finance Agency (“FHFA”) to develop a plan to responsibly reduce the role of the Federal National Mortgage Association (“Fannie Mae”) and the Federal Home Loan Mortgage Corporation (“Freddie Mac”) in the mortgage market and, ultimately, wind down both institutions.”
(“Reforming America’s Housing Finance Market”, Administration Report to Congress.)

That said, there is plenty to quibble about in the report itself and, in particular, with respect to the implementation of the administration’s proposals. We separate our remarks into four areas:

1. The causes of the crisis and the GSE’s role

We agree with the basic notion that we want to look forward and not back. But the role and subsequent failure of the GSEs within the US mortgage finance system provide valuable lessons for how to reform the system.

There is little doubt that the financial crisis was not brought about just by the reckless profit-seeking incentives of then-private-but-always-implicitly-guaranteed Fannie and Freddie. Private-label securitisation of subprime and Alt-A mortgages, which were destined to fail when house prices fell and which were not guaranteed by Fannie and Freddie, left much to be desired as well. There were poor underwriting practices and standards; the capitalisation of originators and securitisers was woefully inadequate and many of these also enjoyed explicit or implicit government guarantees; the private-label mortgage-backed securities were massively mis-rated by the credit rating agencies, whose ratings many investors took on blind faith; and the dispersed owners of sliced-and-diced tranches of risk had little incentive to pursue efficient renegotiation of defaulted or near-default mortgages. In the end, many private players became too big to fail, just like Fannie and Freddie (see Acharya et al. 2010).

But the data do not support the idea that the GSE’s mortgage portfolio and credit guarantees followed safe and sound practices prior to the 2005 period. While the quality of the 2005-2007 vintages were certainly below those of the earlier period, the main determinant of defaults was the collapse in home prices. If such a collapse had occurred prior to the 2005 period, then the GSEs would have failed then too, perhaps not as spectacularly, but failed nonetheless. The data from the Federal Housing Finance Agency (via Fannie Mae and Freddie Mac) show significant increases in the riskiness of their mortgages starting in the 1990s. It is not rocket science to understand that high loan-to-value ratios and lower FICO scores increase risk exposure. The fact that national house prices, according to the Case-Shiller index, increased 132 straight months from 1995 to 2005 is the primary reason mortgage defaults were low in this period. The GSEs, or future variations of the GSEs, need to therefore go back to sound underwriting. We think the 90% loan-to-value ratio cited in the Administration’s report is too high and favour an 80% ratio for first and second liens combined. That is, a household with 20% equity in its house would not be allowed to take on a second mortgage or a home equity line of credit, while a household with 30% equity could take on a second mortgage no larger than 10% of the value of its house.

Also missing from the report, and perhaps most important, is the key problem that the mortgage finance system placed the GSEs as the heads of US mortgage finance. They received implicit government guarantees that enabled them to borrow at near-government rates with little or no capital. If banks involved Fannie or Freddie in the mortgage underwriting and securitisation process, the system allowed for twice the leverage even though the underlying risk was the same. It should not be surprising that the US banking sector, including Fannie and Freddie, held 37% of GSE mortgage-backed securities in 2007. This is the opposite implication of the originate-to-distribute model of securitisation: risks remained on the financial sector’s balance sheet rather than being dispersed to capital-market investors and other intermediaries.

It is important to understand the consequences of this regulatory capital arbitrage. A pool of mortgages, no matter how these mortgages are sliced and diced in securitisation or guaranteed by one counterparty or another, has the same overall risk. If regulators believe a certain amount of capital needs to be held against these mortgages, then sound economics suggests this should be similar at the beginning and end of the securitisation process. Moreover, if systemically important financial institutions hold these mortgages or securitised versions, then the capital requirements should actually go up to reflect the external costs of systemic risk. In equilibrium, one might expect therefore that the less systemic institutions would hold these securities, in contrast to what was observed in the period leading up to 2007.

The importance of this observation cannot be understated. The new and improved mortgage finance system must be a level playing field or systemic risk will once again be built up within a few financial institutions: those who can hold the risk at the lowest cost and whose costs are artificially cheap because of explicit or implicit government guarantees.

2. The unwinding of the GSEs

The report makes four suggestions for unwinding the GSEs. While each of the four recommendations is reasonable, each lacks specifics.

A. Increasing guarantee fees to bring in more private capital.

We support ending the unfair capital advantages that Fannie Mae and Freddie Mac previously enjoyed and recommend FHFA require that they price their guarantees as if they were held to the same capital standards as private banks or financial institutions.”

There is no recognition that the government, no matter how well intended, cannot accurately price default risk. Without accountability, and subject to no market discipline, it is difficult to see how the pricing will improve under the watchful eye of the Federal Housing Finance Agency. Without market pricing, it is not clear how private markets will emerge.

B. Increasing private capital ahead of Fannie Mae and Freddie Mac guarantees.

In addition to increasing guarantee pricing, we will encourage Fannie Mae and Freddie Mac to pursue additional credit-loss protection from private insurers and other capital providers. We also support increasing the level of private capital ahead of Fannie Mae and Freddie Mac’s guarantees by requiring larger down payments by borrowers. Going forward, we support gradually increasing the level of required down payment so that any mortgages insured by Fannie Mae or Freddie Mac eventually have at least a ten% down payment.”

Fannie Mae and Freddie Mac were statutorily required to hold mortgages with at least 20% down payment. The way the GSEs got around this restriction was to have private mortgage insurance. While private mortgage insurance provides protection to Fannie and Freddie, it does not change the fact that mortgages with high loan-to-value ratios are more likely to default. These defaults lead to deadweight costs that push the value of the property down. To lower the mortgage risk, 10% down payment is not sufficient. Encouraging additional credit-loss protection is not a bad idea per se as long as it is structured in a way that does not reduce the overall capital in the system. Mortgage insurers, by their very nature, are systemically risky. For example, leading up to the financial crisis, $960 billion of private mortgage insurance had been written with 80% of the insurance performed by just 6 companies. In 2007 alone, these companies lost 60% of their market value, effectively causing them to suffer a capital shortfall. The key goal should be to prevent a systemic risk build-up anywhere in the financial sector, whether public or private.

C. Reducing conforming loan limits.

In order to further scale back the enterprises’ share of the mortgage market, the Administration recommends that Congress allow the temporary increase in conforming loan limits that was approved in 2008 to expire as scheduled on October 1, 2011 and revert to the limits established under HERA. We will work with Congress to determine appropriate conforming loan limits in the future, taking into account cost-of-living differences across the country.”

While reverting to a loan limit of $625,000 is a necessary action, it is by no means sufficient. The implication of the above paragraph is that these limits would remain in place, otherwise why mention the cost-of-living differences nationwide. A conforming loan limit of $625,000 keeps the government firmly entrenched in the jumbo segment of the housing market. For the private sector to emerge, and as long as the GSEs are “alive”, there has to be a formal way to eventually choke the life out of these GSEs. Gradual loan limit reductions all the way down to zero and according to a clear timeline would seem like a straightforward way to do this.

D. Winding down Fannie Mae and Freddie Mac’s investment portfolio.

“The PSPAs require a reduction in this risk-taking by winding down their investment portfolios at an annual pace of no less than 10 percent.”
(“Reforming America’s Housing Finance Market”, Administration Report to Congress.)

While we prefer a more immediate closing of the GSEs, lest they still crowd out the private sector, and for their portfolio to go into an RTC-like entity, the 10% reduction should be quite manageable given the normal pay-down rate of the mortgage portfolios. We advocate a faster pay-down rate if market conditions permit, and we advocate legislation to that end so that future administrations cannot renege on this plan.

3. The public mission

The report writes that:

we should make sure that all Americans who have the credit history, financial capacity, and desire to own a home have the opportunity to take that step. At the same time, we should ensure that there are a range of affordable options for the 100 million Americans who rent, whether they do so by choice or necessity.”

As one path to the above goal, the report calls for a reformed and strengthened Federal Housing Administration, which includes (i) a commitment to affordable rental housing, (ii) measures to ensure that capital is available to creditworthy borrowers in all communities, including rural areas, economically distressed regions, and low-income communities, and (iii) a flexible and transparent funding source to support targeted access and affordability initiatives.

A reasonable question to ask is why households that have the “credit history and financial capacity” cannot access the mortgage market. Where is the market failure that private markets cannot operate here but can elsewhere?

Is it not mildly worrying that the administration still emphasises a role for government agencies to target creditworthy low and moderate income households, in effect, to continue the “American dream” of homeownership? It is taken as a given that these programmes are socially optimal. The enormous subsidies thrown at housing – the mortgage interest rate tax deductibility, the tax exemption of implicit income from owned housing, the exemption from capital gains upon sale of a house, and the subsidies to Fannie and Freddie – have not served the American people well. Many low-income households ended up losing their little home equity and are left with little but a ruined credit history. As documented in numerous economic studies, most of these programmes involve transfers of wealth to the well-to-do and not to the poor (see Gervais 2001, Krueger et al. 2010, Poterba and Sinai 2010). More generally, there needs to be serious analysis and debate whether this is the best way to redistribute wealth to households in need.

4. The Administration’s three plans

The Administration offers three possibilities, all of which involve efforts to assure housing affordability for low- and moderate-income households -- albeit explicit, on-budget, and primarily the domain of the Federal Housing Administration. Putting aside the above discussion on whether this is the socially optimal goal, returning to the past tighter standards of the Federal Housing Administration and making all of these programmes transparent removes some of the major issues that arose with Fannie and Freddie.
Conditional on a government role through the Federal Housing Administration, the administration presents three plans: (i) a wholly private structure; (ii) a largely private structure, but with an agency that would provide guarantees to new mortgage-backed securities at times of general and severe stress in these securities markets; or (iii) a largely private structure, with a government agency providing “tail risk” or catastrophic insurance in the event that a private guarantor defaulted on its obligations. As is clear, we believe that the first is the appropriate long-term goal; but we believe that our plan is a superior interim means of getting there.

The other two plans offered by the administration allow for the government to slide into the mortgage market through the backdoor and remain permanent fixtures.

In one option, the government comes in like the Lone Ranger (or should we say “Loan Ranger”) and saves the day in a financial crisis. It is not atypical for government entities to act as a lender of last resort, but, to keep the mortgage system in check whatever the government lends out in a crisis should be at exorbitant rates. In other words, it should really be THE last resort. Where to set the threshold for government intervention is difficult and determining the level of guarantee fees that will adequately compensate the government for the default risk of mortgages issued during a crisis is even more difficult. Also, isn’t this the function that we already expect the Federal Reserve to perform?

The final option offered by the administration looks a little like what we currently have in disguise. When mortgages start to default, private mortgage guarantors that sold protection would absorb the first losses. If defaults continue to mount, and private guarantors are wiped out, the government would step in and make the mortgage-backed securities holders whole. In this plan, the government assumes the so-called tail (or economic catastrophe) risk (see Scharfstein and Sunderam 2011). While such a system brings back some market discipline and tries to address systemic risk, it nevertheless is a dangerous idea for four reasons.

First, can we think of any instance in which the government does a good job in pricing credit risk, whether it is deposit insurance for banks or hurricane insurance in Florida? The answer is a resounding no. Even if they had the very best people working for them, setting the correct price would be virtually impossible. Without the interaction and competition amongst market participants and resulting information revelation, the prices of the tail risk will be wrong. And, eventually, moral hazard will rear its ugly head. The problem of pricing the risk accurately is only aggravated by the fact that this is tail risk. By its nature, tail risk only materialises rarely. Lack of data thus plagues the pricing process.

Second, markets (and politicians) become impatient if they have to pay for tail risk insurance which (by its nature) may not materialise for many years. They will call for a reduction in catastrophic risk insurance fees, often at the very moment that more credit (tail) risk is taken on. This is what happened with Federal Deposit Insurance Corporation insurance fees. Banks successfully lobbied to stop contributing to the fund during the quiet time. While its coffers appeared full at the time by standards of expected losses, the fund experienced a massive unexpected shortfall when the catastrophic risk materialised in 2008-2009.

Third, if there is one thing the current financial crisis has taught us, regulators are always one step behind well-paid financiers. The crisis was the poster child for financiers’ coming up with clever ways of pushing risk out into the tails and avoiding capital requirements. It doesn’t take much imagination to see how mortgage financiers will do the same here, given that they in effect control the quantity of tail risk borne by the government.

Fourth, if private mortgage insurers suffer first losses, and an event occurs that triggers a government payout, then by construction all the private mortgage insurers go bust. Next, the government takes over the mortgage market, crowding out private markets from then on. There is a certain “been there, done that” feel to this plan. And we will be back to solving the “genie in the bottle” problem that we face right now!


Acharya, Viral, Matthew Richardson, Stijn Van Nieuwerburgh, and Lawrence White (2011), Guaranteed to Fail: Fannie Mae, Freddie Mac, and the Debacle of Mortgage Finance, Princeton University Press, March.
Acharya, Viral, Thomas Cooley, Matthew Richardson, and Ingo Walter (2010), Regulating Wall Street: The Dodd-Frank Act and the New Architecture of Global Finance, Hoboken, NJ: John Wiley & Sons.
Jeske, Karsten, Dirk Krueger, and Kurt Mitman (2010), “Housing and the Macroeconomy: The Role of Implicit Guarantees for Government Sponsored Enterprises”, University of Pennsylvania Working Paper, February.
Gervais, Martin (2001), “Housing Taxation and Capital Accumulation”, Journal of Monetary Economics, 49(7):1461-1489.
Poterba, James and Todd Sinai (2008), “Tax Expenditures for Owner-Occupied Housing: Deductions for Property Taxes and Mortgage Interest and the Exclusion of Imputed Rental Income”, American Economic Review.
Scharfstein, David and Adi Sunderam (2011), “The Economics of Housing Finance Reform: Privatizing, Regulating and Backstopping Mortgage Markets”, Working Paper, Harvard Business School, February.

420 Reads