Gary H. Stern
Federal Reserve Bank of Minneapolis
Conference of State Bank Supervisors
San Antonio, Texas
June 2, 2005
Good morning. My colleague Ron Feldman and I wrote the book Too Big to Fail—the Hazards of Bank Bailouts. Obviously, we think too big to fail (TBTF) is an important public policy issue and one which has received insufficient attention to date. So I particularly appreciate the opportunity to address this group.
I want to cover several things this morning, so here is a brief outline of my remarks:
I am, of course, speaking only for myself and not for others in the Federal Reserve.
Let me launch into some substance and specifics by describing what I mean by the term "TBTF," so that we are starting in the same place. TBTF applies when uninsured creditors of large/systemically important and complex banks believe/expect that the government will protect them from loss—i.e., bail them out—if their bank becomes insolvent, or encounters significant financial difficulty.
Creditors holding this view, having this expectation, have less reason, less incentive, than they otherwise would to pay attention to and be concerned with the condition of their bank. After all, they expect to be "protected" irrespective of what happens.
The consequence of this expectation, assuming it is widely held, is that these banks take on more risk than they otherwise would. One way to think about this, in an admittedly extreme but illustrative case, is that suppose uninsured creditors are concerned about the current condition of their bank, or concerned that its financial condition will deteriorate in the future, or don't understand its "business model." These concerns could lead creditors to pull their deposits or other funding from the institution, but to the extent that they anticipate protection they are less likely (than otherwise) to do so, thus enabling the institution to continue to use those funds to finance projects/activities/positions that would otherwise be curtailed.
Making the same point in less extreme form, risk taking by institutions perceived as TBTF is priced too low—because they can attract and retain funding at a price that assumes protection—and therefore too much risk is taken on by such organizations. Or put another way, TBTF is an unauthorized extension of the safety net underpinning banking. And note that even if the bank is not literally TBTF, in that ex post its equity holders experience losses and its managers get fired in the wake of a problem, we still have a significant issue ex ante.
Having put this groundwork in place, let me talk about why we should care about TBTF-why, from my perspective, it is a problem. There are in fact several compelling reasons, which I will cover briefly.
Admittedly, it is difficult to estimate these costs, but our reading of the empirical and anecdotal evidence and of economic analysis, and our general impression suggest that the costs are quite high. However, even if they are relatively modest, we think most of our proposed reforms make sense.
Indeed, some have used these and similar observations to argue that explicit deposit insurance coverage should be increased above the prevailing $100,000 limit, presumably to benefit smaller institutions. In my view, this is not the appropriate response, since other things equal it would encourage even more mispricing of risk. The real solution is to reduce the advantage of institutions perceived to be TBTF.
Until recently, many of those concerned nevertheless dismissed TBTF as one of those issues that simply could not and would not be addressed. But as developments with regard to Fannie Mae and Freddie Mac suggest—two institutions that formerly enjoyed TBTF status—it is time to rethink that conclusion with an eye toward "real reform."
As suggested a moment ago, the heart of the TBTF problem is the fear of spillovers that motivates bailouts. In short, I would describe the situation today along the following lines:
There are in my view three productive avenues for reducing the threat of spillovers and thereby curbing the practice and expectation of TBTF coverage. One way to address spillover concerns is to reduce policymaker uncertainty about the potential for and consequences of such an event. So, to be specific, federal banking supervisory agencies could engage in, and to some extent publicize, planning exercises which simulate a large bank failure and the possible responses to such a failure. Such an exercise would be valuable in and of itself—it should, for example, identify critical exposures to other institutions, holes in documentation and so forth—and would also send a signal to uninsured creditors that the "playing field" was changing. Further along these lines, steps could be taken to clarify and improve the regulatory and legal treatment of creditors, and one could also investigate the possibility of rapid provision of liquidity to creditors, which could work to contain and limit contagion effects.
A second way to reduce the fear of spillovers is to reduce the costs associated with failure. If the magnitude of losses at failure is contained, the chance that such losses will cause other banks to close or markets to function poorly is reduced. Actions that would help to accomplish this objective include
Finally, the payments system has long been viewed as a mechanism by which spillovers are conveyed, and steps have been and can be taken to limit this specific transmission mechanism. For example, policies limiting interbank exposures in the payments system could be tightened or, where appropriate, introduced. And collateral could be required for certain exposures or transactions. I am mindful, however, that such actions must pass a cost-benefit test to assure that the smooth functioning of payments is not disrupted excessively.
Again, the point of these reforms is twofold: to make policymakers more confident that potential spillovers associated with the failure of a systemically important banking organization can be contained, and to make uninsured creditors aware that their expectations about protection need to change. Importantly, to the extent that the latter is achieved and creditors in fact become more intensely interested in the quality of the banks with which they do business, market discipline of such institutions will be enhanced. By way of a virtuous circle, more effective market discipline could reduce the probability of TBTF bailouts.
There are other solutions proposed for the TBTF problem, frequently legal or legislative in nature, which we think would be ineffective or ill-advised. Some have argued, for example, that TBTF bailouts should be prohibited by explicit legislation or by "declaration." Such a prohibition may be poor public policy, but beyond this it is not credible and therefore unlikely to be effective. It is not credible because it does not address the fundamental motivation for bailouts—fear of spillovers—and therefore will be evaded in practice when the case for intervention seems compelling.
Some have maintained that the process in FDICIA requiring sign-off by the FDIC, Federal Reserve Board and Secretary of the Treasury after consultation with the president before a bailout can proceed is sufficient to address TBTF. We would observe that all of these parties were involved in the Continental Illinois case (1984), which remains the poster child for TBTF bailouts. Thus, this process does not represent a particularly high hurdle. Finally, others have suggested that if large, systemically important banks are the problem, why not use the anti-trust laws as a remedy? I am not an attorney, for better or worse, but this seems a misapplication of anti-trust, which has to do with competitive issues in my view.
So, let me quickly summarize these remarks. I first described what I mean by the term "TBTF" and discussed, from several perspectives, the nature of the problem which follows from such a perception. In this discussion, I emphasized the mispricing of risk taking, the resulting resource misallocation in the economy and the concern about spillover and contagion effects following from a problem at a large bank. This latter issue is at the heart of policymakers' concern about the consequences if the viability of a large institution is called into question or if it in fact becomes insolvent, and thus addressing TBTF requires addressing this concern. I proposed a number of policies and practices to deal with this matter, including exercises which simulate the consequences of a failure of a major institution and the policy issues that might arise in such an event; mandatory but limited "haircuts" for uninsured creditors; and possibly payments system reforms to limit exposures.
As I noted at the outset this morning, I regard TBTF as a significant public policy issue. Further, this is an opportune time to begin to address it, because both the economy and the banking system are sound and healthy. I think the current situation puts the responsibility on federal banking regulators to pursue the analysis, processes and infrastructure so that such bailouts can be avoided in the future.