Notes from the Vault
Lessons from the Housing GSEs for Resolving Too Big to Fail
Larry D. Wall
The future of Fannie Mae and Freddie Mac is again in the news with President Barack Obama's nomination of Congressman Mel Watt to head the Federal Housing Finance Agency (FHFA), which oversees Fannie and Freddie. Additionally, FHFA Acting Director Edward J. DeMarco recently gave a speech that provides an excellent overview of two broad alternatives for the future of housing finance. What one thinks about the future of these institutions to an important degree depends on how one evaluates their past.
However, the history of Fannie and Freddie also has important lessons beyond the future of these two government-sponsored enterprises (GSEs). These GSEs are similar enough to other financial firms so that most of their safety net issues are relevant to the wider set of financial firms. Yet they are also sufficiently different so that some lessons about too-big-to-fail institutions are more easily identified from the two GSEs' experience. This commentary draws three important lessons from their history for the broader regulation of systemically important financial institutions (SIFIs) and the overall financial system.
1. Cutting safety net subsidies will impose costs on some financial services consumers, but that is not a good reason to continue these subsidies.
Fannie Mae and Freddie Mac received a variety of benefits from the federal government that lowered their costs and raised their profits, according to a paper by University of North Carolina–Charlotte Professor W. Scott Frame and this author. In return, the GSEs were supposed to lower the cost of mortgage financing to homebuyers. The GSEs did lower mortgage finance costs for conforming loans but also retained a substantial portion of the subsidy to benefit their stockholders, as would be predicted by theory. If the demand curve for a firm is downward sloping, part of an input subsidy will be used by the firm to reduce its output price and thereby increase quantity demanded. However, theory also predicts that part of the subsidy will also be kept by the firm to increase its profitability. In practice, studies summarized by Scott Frame, New York University Professor Lawrence J. White, and this author estimate that the value of the implicit government subsidy was a reduction in the GSEs' borrowing costs of 35–40 basis points, whereas the benefit the GSEs provided to mortgage borrowers was only 20–25 basis points.
What's worse, the form of the subsidy also resulted in increased risk taking. Ordinarily, as a firm takes on more risk, the cost of its debt goes up to compensate its debt holders for higher expected (predicted) losses. However, with the government implicitly guaranteeing the debt, the GSEs' debt holders did not demand full compensation for increases in the GSEs' risk. As a result, the GSEs were incented to and did take more risk, especially in the form of interest rate risk in their retained portfolios and credit risk through their purchases of higher-risk mortgage-backed securities. Further, to the extent that the GSEs reduced the cost of obtaining a mortgage loan, they encouraged homebuyers to take on more debt, exposing households to more housing price and employment-related risk.
SIFIs, just like the GSEs, typically seek to deflect attempts to reduce their subsidy by saying that such measures will increase borrowers' cost of funds and reduce their access to lending (for example, see Goldman Sachs discussion of higher capital requirements from January 26, 2011). The SIFIs' argument is valid as far as it goes. SIFIs enjoy a lower cost of funding due to the safety net provided by the federal government to banks. SIFIs also face downward sloping demand curves because banks compete with one another for customers. Thus, part of the subsidy is almost surely returned to the banks' customers in the form of lower interest rates on loans.
However, subsidizing SIFI borrowers through safety net subsidies also creates the same problems as did subsidizing mortgage borrowers through the GSEs. Part of the subsidy is retained by the SIFI for some combination of higher employee compensation and better shareholder returns. Moreover, to the extent the safety net subsidy is helping borrowers, it is also encouraging SIFIs and their customers to take more risk.
None of this is to suggest that governments should never provide any subsidies; almost every government in history has provided subsidies to activities that it deems socially desirable. Rather, the point is that providing subsidies through the safety net is inefficient and increases financial system risk. If an activity is worth subsidizing, alternative mechanisms to the safety net should receive careful consideration.
2. A policy of "just say no" to SIFI bailouts is by itself unlikely to end either the perception or the reality that policymakers will view some financial firms too important to be allowed to fail. Instead, the focus should be on credible failure resolution policies.
Fannie Mae and Freddie Mac did not benefit from an explicit "bailout" fund akin to the deposit insurance fund. Indeed, Scott Frame and Larry White note that Fannie and Freddie's securities were required by law to include language indicating that they are not guaranteed by, or otherwise an obligation of, the federal government. Yet market participants acted as if Fannie and Freddie were guaranteed, and these beliefs were ultimately validated by the federal government. Among the reasons why market participants believed in the implicit guarantee, according to Frame and White, were that: (a) these two GSEs operated under special federal charters that gave certain benefits and imposed some responsibilities, (b) the federal government had consistently provided bailouts to distressed GSEs in the past, and (c) the residential housing construction and financing industries were economically important and politically powerful.
While the two GSEs benefited from market expectations of implicit guarantees, the lack of any de jure guarantees likely was also helpful in deflecting efforts to reduce their safety net subsidy. For example, there was no point in talking about the U.S. Treasury charging GSEs a fee to guarantee the GSEs' bonds because the obligations were not explicitly guaranteed. Thus, not only was the "just say no" policy as applied to the GSEs ineffective in stopping safety net subsidies, it was actually helpful in maintaining the subsidies. The key to an effective no-bailout policy is a credible alternative to bailouts and not inconsistent legislation that bans bailouts.
The relevance of this for SIFIs is that some commentators, such as Federal Reserve Bank of Richmond Senior Economist Robert L. Hetzel, argues that, "Credible commitment to limiting the safety net requires taking the bailout decision out of the hands of regulators." Yet as we saw with the GSEs, Congress can and will change the law if its members perceive that failing to do so would have dire economic consequences. A policy of "just say no" will only be credible to the extent that both in reality and in congressional perception, distressed SIFIs can be allowed to fail without major consequences to the real economy. In other words, the focus should be on establishing policies that allow SIFIs to fail without large adverse spillovers to the real economy. While the development of such policies is probably the most important task confronting prudential regulators, there are not simple solutions, as I've previously discussed.1
The danger with banning bailouts is that it amounts to a premature declaration that SIFIs can no longer benefit from implicit guarantees, which may actually result in higher safety net subsidies to the extent the banks use the declaration to reduce necessary regulation. For example, Wells Fargo Chief Executive Officer John Stumpf argues that Dodd-Frank's provisions, including orderly liquidation, solved too big to fail, so no further action is needed to address large banks' size and complexity.
3. The safety net is about more than protecting small depositors. It is also about maintaining the flow of essential financial services.
Although Fannie Mae and Freddie Mac have been put on life support by the Treasury and are in FHFA conservatorship, they have not been wound down to reduce the government's exposure to further losses. Indeed, the two GSEs' share of the market for mortgage securitization has increased since 2007, according to the Federal Housing Finance Agency. Moreover, even those proposals that mandate elimination of the GSEs have typically provided for a multiyear transition to allow private substitutes to be developed rather than risk an extended period of disruption in the supply of mortgage financing (for example, see University of California–Berkeley Professor Dwight M. Jaffee). Absent congressional action to transition to a post-GSE housing finance world, the U.S. Treasury continues to guarantee the two GSEs will have positive net worth (see Federal Housing Finance Agency). And this guarantee of positive net worth allows the GSEs to continue making government guaranteed loans.
This lesson is relevant for those who espouse policy proposals that implicitly assume the way to limit federal guarantees of financial firms is to separate deposits from other risky activities. University of California–Berkeley Professor Robert Reich's call to reinstate the Glass-Steagall prohibitions on banking and commerce is a prime example. What is missed by many of these proposals is that policymakers also use the safety net to preserve what they perceive as essential financial services (for which services this perception is correct is a different discussion). If policymakers determine that the failure of a financial firm will ultimately result in a large reduction in real activity, they are very likely to seek to protect that firm regardless of whether it has insured depositors. Bear Stearns and AIG did not receive extraordinary support to protect their depositors. Indeed, Bear Stearns did not have depositors and AIG's insured thrift was a minor part of its operations. Rather, Bear Stearns and AIG received support because of the services they provided, their interconnections with other large financial firms, and the risk that their failure would spark more widespread runs on other financial firms.
There are, of course, other reasons why limiting the range of SIFI activities may be desirable. For example, MIT Professor Simon Johnson argues some SIFIs are too complex to manage, and Sheila Bair, former chair of the Federal Deposit Insurance Corporation and currently chair of the Systemic Risk Council, argues they are too big to regulate effectively. The important point is that policies to limit SIFI activities should not be implemented on the belief that such limits automatically reduce the scope of the safety net. To restate the conclusion of the second lesson, a premature declaration of victory over implicit guarantees may well result in weaker prudential regulation in the short run and higher safety net costs in the long run.
Larry D. Wall is the executive director of the Center for Financial Innovation and Stability at the Atlanta Fed. The author thanks Paula Tkac for helpful comments on the paper. The views expressed here are the author's and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System. If you wish to comment on this post, please e-mail firstname.lastname@example.org.
1 To be fair to Hetzel, he does recognize the importance of credible commitments. However, he neither addresses the implications of the congressional passage of support for Fannie Mae and Freddie Mac, nor the subsequent congressional passage of the Troubled Asset Relief Program (TARP) for the credibility of ex ante limitations on federal bailouts.