Moral Hazard and Bank Protection
Asia Pacific Finance Association Conference
July 23, 2001
Good afternoon. The organizers of the conference largely gave me free rein to select a topic, although they suggested that I may want to address "moral hazard and bank protection," in fact one of my favorite themes. And so I will take them up on the suggestion, because I think significant and challenging public policy issues remain in this area.
I essentially will be emphasizing three points this afternoon.
Prospective developments in global financial markets, together with the reforms of the Basel II capital proposal, suggest that over time the international banking system will be characterized both by greater adherence to consistent standards and by sounder banking institutions.
There will over time be greater reliance by supervisors on
disclosure, market data and market discipline of banks. The same factors that have pushed policymakers to support increased disclosure for banking organizations will lead them to take action to address the perverse incentivesthe moral hazardof too-big-to-fail policies.
Effective market discipline which addresses the moral hazard problem requires credibility, in the sense that uninsured creditors of large, complex banking organizations must believe that they are at risk of loss. Establishing and maintaining this credibility are difficult, and I will have some specific suggestions about how this might best be accomplished.
I will spend the next several minutes developing these points. Then, I will be ready to turn to your comments and questions. Let me remind you at the outset, that, as usual, I am speaking only for myself and not for others in the Federal Reserve System.
A productive place to start thinking about the international financial system is with the U.S. financial marketplace. The United States, to be sure, has a number of very large commercial banking organizations. But such institutions face competition from regional and community banks, as well as from investment banks, insurance companies, credit unions, finance companies and others in specialized lines of business. From the perspective of a customer, financing, in theory, can be obtained from a wide range of sources, including of course accessing the debt and equity markets directly in many cases. Moreover, this description fails to do justice to the breadth, depth and competitiveness of the U.S. financial system, because I have yet to mention venture capital firms, "angel" investors and so on. Even the retail customer has an enormous number and variety of options.
Customers worldwide may not currently face as complete a set of alternatives, but in my view they soon will, and probably within the next decade. This is because globalization matters. By the term globalization, I mean the growing integration of the world economy, including the ongoing expansion in international trade, in cross-border location of manufacturers and service providers, and especially in finance. While it may not quite be true that "financial capital knows no boundaries," this is clearly the direction in which things are moving. And I would judge that something akin to the U.S. financial system will prove to be exportable.
From the financial services firms' (the providers') point of view, there is an incentive to expand profitable products geographically and to meet the needs of customers with far-flung operations. Indeed, for firms operating in highly competitive domestic markets, international expansion may represent the most attractive opportunity available. From the customers' viewpoint, there is an incentive to shop for the most attractive terms and conditions irrespective of the national origin of the provider. And with the sharp decline in telecommunications and information processing costs, it is inexpensive to do so.
Both sets of incentives augur well for further financial integration, and greater integration suggests heightened competition in many markets. Obviously, the advent of the European Monetary System, the euro and the European Central Bank is a dramatic case in point. Moreover, we continue to see overseas expansion by major commercial banks, with Citigroup's recently announced purchase of a large institution in Mexico, Bank of Scotland's further growth in the United States, several Spanish banks' significant presence in Brazil and so on.
With increased competition should come a more consistent set of standards, and application of standards, in international banking. Level playing field considerations support such a conclusion, and the interests of customers and counterparties favor standards as well. Both parties have an interest in the caliber of the institutions with which they do business, either because of direct exposure or because they value ready access to liquidity, especially in times of financial stress. That is, nonfinancial firms should want to do business with banks that can provide liquidity when it is sorely needed.
Moreover, financial intermediaries have to fund themselves, and routine access to the capital markets would seem to require adherence to standards as well as voluntary disclosure of information demonstrating such adherence. In other words, there are reasons to believe that normal market practice will go quite some distance toward establishing acceptable standards.
Thus, market developments, left to themselves, should encourage adoption of consistent standards and increased disclosure of relevant information of bank positions and performance. However, there remains a role for government in the management of bank risk taking due to both explicit and implicit support for creditors. As a modest contribution, government can reinforce the market trends discussed previously by better linking bank capital to risk and by promoting greater disclosure.
The new Basel capital accord (Basel II), although complex, should result in sounder banking institutions in the sense that capital levels will be more appropriate for the risks to which the institutions are actually exposed. While there are, to be sure, options in Basel II, it seems that the intent is to move to internal ratings based capital standards over time. Since the internal ratings-based approach should rely on an institution's actual experience, to the extent that risk taking is initially mispriced, and therefore banks take more than they intend, it will be reflected in their results and ultimately in their capital levels.
In principle, customers should demand increased disclosure and sound banks should willingly provide it. Hence there may not be much need for government involvement. But there are several reasons why banks may be hesitant to unilaterally and voluntarily increase disclosure, including:
- Uncertainty about the benefits of disclosure. Assuming there are
costs to providing additional information, bank management would want
to be reasonably confident of the benefits. To the extent they are uncertain,
they are likely to be hesitant.
- Concern about the timing of disclosure. To the extent that firms
cannot or do not coordinate the timing of their releases, there may
be a cost to going first, in that proprietary information may be revealed,
albeit indirectly or inadvertently.
- Externalities. Bank supervisors would benefit from the improvement in market pricing following from increased disclosure, but this benefit is external to the calculations of bank management. Hence, management is likely to provide inefficiently small amounts of information.
As a consequence of these and other considerations, there is a wide variety of official proposals afoot designed to enhance disclosure. I have neither the time, nor the inclination, to review them this afternoon. Instead, let me move to a discussion of a far more important role for government, namely its role in addressing the too-big-to-fail issue. Too-big-to-fail is a serious concern, and, if unchecked, could with time hinder the favorable market trends I previously identified and lead to significant resource misallocation as well. By too-big-to-fail, I mean a set of government practicesin some cases policiesthat protect large banking organizations from the normal discipline of the marketplace because of concerns that such institutions are so important to markets and their positions so intertwined with those of other banks that their failure would be unacceptably disruptive, financially and economically.
Basel II and market pressure for standards may not be sufficient to address effectively too-big-to-fail perceptions surrounding large organizations. Policymakers and bank supervisors recognize this, which explains why there is considerable renewed interest in various proposals to enhance market discipline of such institutions and, in fact, the new Basel accord has disclosure as its third supervisory pillar. While I view disclosure as a lead-in to enhanced market discipline, some appear to see it as market discipline personified.
There is, however, a meaningful distinction between disclosure and market discipline. In fact, the gulf between the two concepts is enormous, because increased disclosure is unlikely to advance market discipline appreciably unless creditors of all institutions, including (and especially) those considered by some to be too-big-to-fail, believe they are at risk of loss. In this regard, let me quote Chairman Greenspan from a recent speech, where he makes this point well:
"Expanded disclosure will be critical to enhanced market discipline, but the additional information will be irrelevant unless counterparties believe that they are, in fact, at risk. That is why ? a prerequisite to effective market discipline is the belief by uninsured creditors that at least they may be at risk of loss. Uninsured counterparties have little reason to engage in risk analysis, let alone act on such analysis, if they believe they will always be made whole under a defacto too-big-to-fail policy. ?"
Admittedly, it is probably neither possible nor wise to eliminate entirely the potential for support of uninsured creditors, but it is possible to reduce both the probability of a bailout and the extent of protection creditors receive when a bailout occurs. That is, it is possible to improve incentives so that creditors of large institutions will in practice take advantage of increased disclosure, thereby mitigating to an extent the moral hazard associated with the public safety net supporting banking.
Putting creditors of large organizations at risk of loss in a credible manner is no easy task. Experience tells us that such credibility will have to be explicitly and deliberately established; we cannot count on addressing moral hazard by jawboning or by having creditors spontaneously decide to pay attention and price risk appropriately.
I will describe and assess four strategies that might contribute to establishing such credibility, that is, that might convince creditors of large banks to pay attention to the caliber of the institutions with which they do business because they may not be fully protected in the event the institution runs into difficulty. The first strategy is rules or laws that prohibit bailouts of uninsured creditors. For success, this approach requires policymakers to ignore the incentives to engage in bailouts, and these can be rather compelling.
There are several ways such a strategy could be pursued. It could be accomplished simply by forbidding in law the coverage of all creditors of a certain type (for example, coverage only allowed for those with bank deposits under $100,000), that is, as we in the United States have done. A variant of the prohibition tactic relies on creditors with legal standing that puts them beyond coverage limits and makes it unlikely that they will receive support when their financial institution fails. Subordinated debt plans exemplify this approach.
The rule-based approach is relatively easy to implement and offers a direct way to convey the sentiments of policymakers to creditors. But is it credible? That is, will it change what uninsured creditors expect, and the decision that policymakers make when a large bank is failing?
While perhaps better than nothing, I doubt that such rules will advance credibility very far. To be credible, policymakers must have little desire to evade the rule, and such evasion must be difficult; these conditions are unlikely to be met. First, these legal regimes do not change any of the incentives to provide bailouts. Policymakers will still fear the economic and political costs associated with large bank failure and will have a strong desire to evade the rule. Second, there are a fair number of options available to policymakers for circumventing prohibitions, ranging from emergency legislation to resolution techniques that accomplish the economic substance of a bailout without violating the legal restriction. Given these options, it seems unreasonable to believe that policymakers will shackle themselves when the benefits of extending the safety net arise, as they surely will some day.
A second commitment strategy attempts to raise the costs to policymakers of engaging in bailouts without relying on prohibitions per se. In the United States, the Federal Deposit Insurance Corp. Improvement Act of 1991 (FDICIA) contains reforms of this nature. In particular, FDICIA requires policymakers to take a series of votes that are reported publicly before bailouts occur. FDICIA thus raises costs via the added publicity and formality that would accompany an attempt to bailout creditors.
Another commitment strategy that relies on disincentives and has received attention in the literature (in this case, the literature on credible monetary policy) involves explicit contracts. For example, a contract could penalize policymakers monetarily or otherwise for actions that make implicit coverage more likely or for the coverage itself (as has been suggested for central bankers when they fail to meet inflation targets). A second such approach would rely on issuance of federal debt, which pays holders a lump sum if a bailout occurs and nothing otherwise.
Again, these explicit means of establishing commitment are better than nothing. Moreover, these strategies are not as extreme as outright prohibitions and thus cannot be dismissed as easily. But even ignoring their obvious political problems, this approach may not prove useful in establishing credibility or changing the decisions of policymakers. One problem is with implementation, because it appears difficult to write contracts such that all contingencies and actions by policymakers are covered. Thus, the same types of evasion that are possible under the rules based reforms could occur here as well.
A third alternative is a commitment strategy that tries to achieve credibility by reducing the incentives to bailout creditors, principally by reducing the likelihood of spillovers. A major incentive to policymakers in providing implicit coverage is the management of systemic risk and the prevention of spillover failures and contagion, the heart of the too-big-to-fail issue. If a plan could significantly reduce or eliminate these possibilities, then policymakers would have less incentive to provide such coverage. A few suggestions of this type focus on payments system reform because proponents argue that spillovers from the failure of a large financial institution are transmitted via exposure in the payments system.
A second type of plan to limit spillovers focuses on the amount of loss borne by uninsured creditors. The goal is to establish the loss that uninsured creditors bear at an amount that, on the one hand, motivates them to monitor financial institutions but which, on the other, does not lead to their own insolvency in the event of a failure at another institution. As a result, the failure of one bank would not lead to the failure of other banks or nonbank creditors. We, at the Minneapolis Fed, have proposed a plan that implements this notion through the application of co-insurance for uninsured creditors of large banks.
These plans present some complex and difficult issues, which make evaluation challenging. Some have questioned, for example, the ability of policymakers to devise an effective coinsurance loss rate. And perhaps the focus on payments systems would not capture adequately the channels by which exposures between financial institutions occur. Despite these concerns, I would argue that some of these plans move us in the direction of credibility: That is, they alter in an appropriate way the underlying incentives that policymakers face. I think it will be extremely difficult to reduce the likelihood of bailouts without reducing the incentives that policymakers have to take such actions. Reforms that focus on prohibitions and disincentives, in contrast, seem likely to fall short precisely because they do not address the incentives that lead to bailouts in the first place.
In any event, a fourth method of addressing the incentive to provide bailouts is through so-called constructive ambiguity. This policy involves general statements that the government will not routinely provide bailouts to creditors of large firms and avoids explicitly detailing the conditions under which a bailout will occur. However, a policy of constructive ambiguity does not directly address policymakers' incentives to provide bailouts. Indeed, this policy risks either creating unnecessary uncertainty for market participants or providing, at best, temporary relief until a large failure looms and the bailout is provided.
Let me summarize and conclude.
At the end of the day, I would expect to see several, perhaps disparate, trends emerge in international banking and finance over the next decade or so. Heightened competition globally, together with reform of Basel capital calculations, should contribute to increased adherence to consistent standards. Such a development is, I believe, in the interest both of bank customers and, certainly, of financially sound banking organizations.
Adherence to consistent standards should help to assure a sound financial infrastructure, and voluntary or mandatory increases in disclosure by banks should contribute as well, but disclosure by itself is not enough. Market discipline must be strengthened so that the incentives confronted by large, possibly too-big-to-fail, banks are appropriate.
It will not be easy to improve effective market discipline of large banks, because doing so requires putting uninsured creditors of such institutions at risk of loss in a credible way. To achieve this, we will need a mechanism to limit spillovers from one bank to another, because these spilloversthe contagion effectare a principal rationale for the practice of too-big-to-fail. Co-insurance, where all bank creditors are potentially exposed to loss but the maximum loss is capped, would help to accomplish this objective. Similarly, payments system reforms to limit interbank exposures would help to contain spillovers. Other constructive proposals may surface, which could alter favorably the underlying incentives policymakers face. In any event, if we are serious about a sound financial system and efficient allocation of resources, we are going to have to come to grips with too-big-to-fail.