No Silver Bullet for "Too Big to Fail"

November 10, 2009 12:00 PM

Whether to use public money to rescue financial firms because they are too big, too complex, or too interconnected to be allowed to fail is not a new issue. However, in the financial and economic crisis of the last year and a half, the number and nature of such rescues in sophisticated financial markets and systems has dramatically increased the salience of the issue. It is summarized by the initials "TBTF," or "too big to fail." Precedents have been set. There is an understandable desire to limit the application of those precedents. The US Congress is considering remedies to impose such limits in the future. Each remedy comes with costs or with problems. No silver bullet will solve this problem. A combination of steps will be required.

The TBTF issue in the US context is, first, a political issue. Citizens and their representatives see as unfair the "bailouts," or rescues, of large financial institutions using financial resources that ultimately are provided by the taxpayers. They see Wall Street being bailed out by Main Street.

Certainly, going forward, the status quo prior to the crisis is not a viable option not only because of taxpayer revolt but also because of moral hazard considerations that were highlighted and exacerbated by government actions in the crisis. The impacts on the behavior of institutions that are not allowed to fail, or the behavior of their management, are more important than the issue of the cost of bailouts and who bears those costs. Size is just one ingredient in this calculus; the more important ingredient is the lack of the potential to fail. Moral hazard surely will affect behavior in the future even if it did not do so in the past.

If governments are expected to protect private institutions from failure, market discipline will be undermined along with what Schumpeter referred to as the desirable process of creative destruction. Competition is distorted. Capital is misallocated. Poor risk management or, if you like, the taking of excessive risk is rewarded. Losses are socialized. This is where the taxpayers come in, but those losses are only part of the equation. Implicitly or explicitly, governments are injected into the management of financial institutions. This changes the rules of the game. One result is a more concentrated structure of financial entities, more prone to take excessive risk and with little evidence of societal benefits, such as increased efficiency, lower costs, or greater availability of credit.

In the United States, remedies for the TBTF problem fall into four categories of proposals: (1) break up the systemically important institutions so that individually they are not too big and the consequences of any failure are eliminated; (2) separate riskier activities into unrelated institutions whose failure will not have the same direct adverse consequences for the economy and the financial system; (3) employ a combination of regulatory modifications that either discourage excessive risk taking or establish cushions against its consequences; or (4) establish a special resolution mechanism so that the failure of systemically important financial institutions can be managed to minimize the damage to the financial system and economy without the need for a governmental rescue.

Each of these remedies has its merits and supporters, but also its problems and detractors.

The problems with the first category of remedies—the breakup approach—are several. This approach involves more than downsizing institutions by reducing their work forces and selling some of their businesses, as is being done in many countries with and without the encouragement of governments. This approach fully applied involves breaking up institutions into unique entities with their own shareholders and managements and arbitrarily capping their absolute size. We do not know enough to say how small is small, however, or which individual institutions could fail without adverse systemic consequences, in part because those consequences depend on circumstances.

Moreover, if one large institution is broken up into, say, 5 or 25 smaller institutions, one can ask if the system really would be better off. If each of the smaller institutions follows the same or very similar investment strategies, known as herding, and has similar risk management policies, they may all experience similar stresses and losses. The systemic effects of letting them all fail would be the same as the systemic effects of letting the one large institution fail. It is possible that institutions would not herd or that their smaller, reduced size would limit some scope for excessive risk taking, but that is a conjecture, not an established fact.

The empirical evidence on the economies of scale and scope in large financial institutions is at best ambiguous. However, even if there are no economies of scale in the behemoth size range, which I am willing to believe, we also know that today 4 of the 10 largest global financial institutions by market capitalization are chartered in China. These government-owned institutions are by their construction TBTF. Unless their size and status are altered, they would be in a position to distort global competition even if they and their home country do not benefit from economies of scale and scope. One of the major global economic problems with TBTF would remain.

My conclusion is that the break-up approach to TBTF does not have much to recommend it. Many large institutions probably should be encouraged to shrink, but the economic case for breaking them up is not sufficiently strong to overcome the opposing political and practical realities.

The problems with the second category of remedies—separating out riskier activities—are also several. Which activities are too risky? There is no consensus. Should the list of such activities be fixed for all time? Almost certainly, it should not be. In any case, what about institutions outside the perimeter of supervision which are, or become, large and interconnected? By ignoring them, we are merely relocating the TBTF problem and potentially making it worse, at least for the taxpayers, because what in effect may be institutional cross-subsidies in universal banks no longer would be available. One answer is to extend the perimeter of regulation and supervision in order to deal with regulatory arbitrage, the phenomenon of institutions seeking out the least onerous regulator. This answer brings us back to where we started.

My conclusion is that the separation approach to TBTF is an incomplete answer. On the other hand, regulatory modifications—such as capital charges on larger institutions based in part on the nature of their lines of business and, perhaps, their size—should be used to rebalance incentives within large institutions away from an excessive concentration on riskier activities.

The problems with the third category of remedies—enhanced supervision and regulation—are again multiple. One example is the technical problem of designing retooled capital structures. Basel II, the revised set of international banking regulations promulgated in 2004, was more than a decade in the making and has not yet been implemented in all jurisdictions, including the United States. Before Basel II can be fully implemented in all jurisdictions with any confidence, it must be substantially altered or augmented.

A second problem lies in ambiguities about the enforcement of regulations. What makes one so sure that this time it will be different? What is the right balance between rules established in advance and discretion left up to the regulators or supervisors? Could any enhanced regime of supervision and regulation effectively contain the economic, financial, and political costs of moral hazard and eliminate the TBTF problem?

My conclusion is that the regulatory approach to TBTF is more promising than the first two approaches, but the technical problems are substantial. Those problems are compounded by political resistance from the institutions to be regulated, at least in part because of curbs on their future activities that potentially would make them less profitable. Thus, whatever the merits of the first three approaches, large systemically important financial institutions are likely to be here to stay. In any future US regulatory regime, those institutions are likely to face failure, again posing an unattractive choice between an expensive rescue and a disruptive bankruptcy unless other alternatives are present.

Despite the likelihood of future failures of large systemically important financial institutions and the unattractive policy choices under the status quo, skepticism and controversy surround proposals in the United States to establish a special resolution mechanism for such in the event of their potential failure. Some argue that such a mechanism merely would reinforce the notion that certain institutions are too big to fail. In other words, it sends the wrong signal and would exacerbate the basic problem of moral hazard. Others argue that such a mechanism would enhance market discipline while at the same time recognizing the reality that circumstances surrounding potential failures will differ, including the macroeconomic contexts in which the failures occur.1

My conclusion is that a special resolution mechanism for large financial institutions is a necessary complement to any other approach to the TBTF problem, but is not a complete solution by itself.

Thus, all four proposed remedies to the TBTF problem in the United States have merits but also raise concerns. There is no silver bullet. It is preferable to craft a multifaceted approach.

I favor an approach that involves: (1) somewhat more concern for the size of institutions; (2) regulations or limits on certain activities in institutions (but stopping well short of an arbitrary, permanent separation of activities); (3) an expansion of the perimeter of supervision along with other regulatory remedies; and (4) a new resolution mechanism for all systemically important financial institutions. Such a combination is likely to serve the United States reasonably well for a few decades; the regulated and the regulators will be more alert and cautious, aided in their caution by an improved alignment of compensation incentives.

Lest these views are dismissed as Panglossian, I must add a caveat: large US financial institutions, even now, are strongly arguing their case with the US Congress for minimal changes from the status quo before the crisis. Consequently, it is less than fully clear how significant and comprehensive US financial reforms will be. Nevertheless, I am optimistic.

Does what the United States decides about TBTF matter to the rest of the world? History and analysis would suggest that it does. Throughout the world, the decision to allow Lehman Brothers to fail last year is regarded as the biggest mistake that the US authorities made over the 2007–09 period.

The US challenge going forward is to limit the potential need to rescue large, systemically important financial institutions while minimizing the adverse consequences for the financial system, the economy, and the taxpayers if and when such a need occurs. The US challenge is also a challenge for the rest of the world. We lack a global resolution authority even as we lack global consensus on the other three categories of remedies for the TBTF problem.

Problems with a large, global, financial institution can originate in any jurisdiction. The seriousness of those problems and their implications for the stability of national financial systems as well as the global system depend on the circumstances at the time. Recall, for example, the demises of the Bank of Credit and Commerce International (BCCI) in 1990 and Barings Bank 1995. The activities of BCCI were spread around the world, and Barings was brought down by a trader in Singapore. What would have been the consequences for the global financial system and economy if those events had unfolded in late 2008 or early 2009?

Facing these issues, the United States, the G-20 countries, and the international financial community have their work cut out for them as they promote solutions that are both robust and mutually compatible.

This post is adapted from Edwin M. Truman's keynote address, "Lessons from the Global Economic and Financial Crisis," [pdf] at a conference sponsored by the Institute for Global Economics and the International Monetary Fund, in Seoul, South Korea, November 11, 2009.


1. See H. Rodgin Cohen and Morris Goldstein, 2009, The Case for an Orderly Resolution Regime for Systemically Important Financial Institutions [pdf], The Pew Financial Reform Task Force Briefing Paper 13. Cohen and Goldstein argue cogently that such a regime will not increase moral hazard and that it is superior to the approach of amending the existing corporate bankruptcy code in the United States.