Gary H. Stern - President, 1985-2009
Published September 1, 2000 | September 2000 issue
Financial modernization has the real potential to expand the safety net, especially for the largest banks that take on new powers, and exacerbate the moral hazard problem. As such, we have extra incentive after the passage of the Gramm-Leach-Bliley Act of 1999 (GLBA) to increase our consideration of plans to address excessive risk taking by too big to fail (TBTF) banks and affiliated financial institutions.
Plans to limit safety net expansion and moral hazard will prove effective only if they are credible or time consistent. That is, these policies must contain features or commitments such that creditors believe they will suffer at least some loss upon the failure of their financial institution.
We categorize methods for achieving credibility into four groups based on the manner in which they address the incentives that policymakers face when deciding to bail out creditors.
Congress prohibits bailouts or passes other simple legal restrictions. This strategy essentially requires regulators to ignore incentives to bail out creditors.
Regulators take steps that reduce the incentives that have led them to bail out creditors in the past. This strategy relies on changing the dynamics that create the time-consistency problem in the first place.
Congress focuses on policymakers themselves and appoints regulators who have a predilection to resist bailouts. This strategy focuses on the incentives internal to the policymaker.
We believe the first two methods are the least likely to be credible because they do not address the underlying factors that give policymakers incentives to bail out creditors at TBTF institutions. As a result, policymakers have significant cause to evade these reforms, making these strategies deficient. The latter two strategies have more potential because they try to alter underlying incentives, although they also raise implementation concerns. (The table summarizes this review.)
Certainly, these initial thoughts do not constitute a very rigorous evaluation of the commitment technologies available to policymakers. Nonetheless, we hope our attempts at categorizing and discussing these plans will encourage more detailed evaluations of them and prompt new suggestions for establishing credibility. Without such efforts we may end up by default with a potentially costly policy of "constructive ambiguity."
The connection between financial modernization, moral hazard and too-big-to-fail may not be intuitive to all. Thankfully, Chairman Greenspan has spent considerable time explaining the problem of moral hazard and its connection to recently passed financial modernization legislation that expands the powers available to banking organizations.
The chairman described the basic problem of moral hazard in banking as follows:
"... As a society we have made the choice to create a safety net for depository institutions, not only to protect the public's deposits, but also to minimize the impact of adverse developments in financial markets on our economy. Although we have clearly been successful in doing so, the safety net has predictably created a moral hazard: The banks determine the level of risk taking and receive the gains therefrom, but do not bear the full cost of that risk; the remainder is borne by the government. Since the sovereign credit of the United States ultimately guarantees the stability of the banking system and the claims of insured depositors, bank creditors do not apply the same self-interest monitoring of banks to protect their own position as they would without discount window access and deposit insurance. ... Put another way, the safety net requires that the government replace with law, regulation and supervision much of the disciplinary role that the market plays for other businesses."1
While some of the safety net the chairman describes is explicit, a major source of moral hazard arises from implicit or discretionary federal coverage. Indeed, it is this form of moral hazard that has received the greatest amount of attention over the last several years. Implied coverage exists for the nominally uninsured bank creditors of too-big-to-fail banking institutions as well the creditors of nonbanking firms. Simply put, despite the fact that these creditors are not entitled to protection from loss when their financial institution fails, they believe and act as if they have such protection. This belief is based both on the historical coverage of such creditors as well as the incentives of policymakers, discussed in more detail below, that could lead them to conduct future bailouts.
Why discuss the problem of moral hazard resulting from explicit and implicit coverage in conjunction with financial modernization? The chairman's concern was that legislation allowing banking organizations to carry out insurance, securities and other financial activities would lead creditors of the insurance firms, securities firms, etc., that were affiliated with banks to believe they were covered by the bank safety net. This belief leads to the mispricing of risk by these creditors, thereby encouraging excessive risk taking by bank affiliates. This outcome was clearly troubling to the chairman, who noted that, "The last thing we should want, therefore, is to widen or spread this unintended, but nevertheless corrosive dimension of the safety net to other financial and business entities and markets. It is clear that to do so would not only spread a subsidy to new forms of risk taking, but ultimately require the expansion of banklike supervision as well."2
The chairman did not see the potential pitfalls of expanded powers for banking organizations as insurmountable. Indeed, he argued that careful organizational design would address concerns about safety net expansion while allowing increased powers for banks. Specifically, the chairman supported the exercise of new powers in subsidiaries of holding companies that own banks but opposed their location in the bank itself or in a subsidiary of the bank. The chairman noted that "... the subsidy that the government provides to banks as a byproduct of the safety net would be directly transferable to their operating subsidiaries to finance powers not currently permissible to the bank or its subsidiaries."3 In contrast, "... the holding company structure is the most appropriate and effective one for limiting transfer of the federal subsidy to new activities and fostering a level playing field both for financial firms affiliated with banks and independent firms. It will also, in our judgment, foster the protection of the safety and soundness of our insured banking system and the taxpayers."4
The version of financial modernization that did pass, the Gramm-Leach-Bliley Act of 1999, contained a compromise between the U.S. Treasury and the Board of Governors rather than a pure version of the chairman's plan. The compromise legislation allowed banks to exercise some new powers in subsidiaries of banks while simultaneously taking steps to limit or discourage the use of this option (for example, the legislation forbids certain activities from being carried out in bank subsidiaries and caps the size of these subsidiaries).
Despite differences, both the chairman's original proposal and the compromise position are premised on the notion that organizational design is an effective tool for addressing a potential increase in moral hazard. But will organizational strategies prove effective? We cannot begin to answer this question without first determining which factors policymakers should consider when judging the likelihood that a proposal will be an appropriate response to moral hazard. This determination should be paramount in our minds for exactly the reasons the chairman has argued. Financial modernization has opened the door for a potentially significant increase in the moral hazard problem.
Let me state from the beginning that I believe the credibility, or time consistency, of efforts to contain moral hazard is the key to determining their effectiveness. This was true before GLBA, is even more true after its passage, and will be the focus of the rest of this column.
Addressing moral hazard appears deceptively simple. Just make creditors at TBTF banks believe they are at risk of loss. Thus, the prices charged by uninsured creditors will reflect the risk taking of the financial institution. These price signals, directly and through appropriate links to regulatory discipline, can alter the behavior of the financial institution, leading to less moral hazard and less risk taking. These results are the benefits to society from a strategy of putting creditors at risk of loss.
However, putting creditors at risk is not as trivial as it sounds because of the so-called time-consistency problem. The problem arises because at the time policymakers must force creditors to bear losses, they will face at least two significant incentives to renege on the commitment to do so. These incentives to bail out creditors include the potential wrath of those who lose money or their political supporters and the prevention of spillover failures or at least serious disruptions at other banks.5 These spillovers can also occur between banks and other financial participants (for example, securities firms) to which they have exposure. Policymakers may fear that this kind of contagion will prevent the efficient operation of the financial sector. Weaknesses in the financial sector could produce a diminution of output and employment in the real economy. By providing full protection against loss to uninsured creditors, policymakers may believe they have potentially saved society from a costly financial shutdown.
Just saying that regulators will not bail out creditors is therefore not time consistent. Creditors will recognize policymakers' incentives to go back on their word and bail them out and behave as if they have nothing to lose. This suggests why a credible commitment to putting creditors at risk of loss is so important. Without this commitment, significant moral hazard is inevitable.
One might question why policymakers don't follow through on their promise, given that their decision to bail out creditors produces moral hazard. In other words, isn't the potential for moral hazard a sufficiently large disincentive to outweigh the incentives to bail out creditors? One answer is that policymakers might think they will act differently in the future, the next time decisions on bailouts must be made. Moral hazard will not occur, policymakers might think, because creditors will recognize the current bailout is a one-time event. However, creditors will actually recognize that each time the decision must be made, the policymaker will face the same incentive that led to the decision to bail out creditors the first time. Alternatively, the policymaker might consider the disincentive of moral hazard to be of less importance than the incentives to bail out because moral hazard costs will be evident in the future while the benefits of bailouts are received today.
A second challenge to our concern about time-consistent behavior comes from research suggesting that the threat of contagion is overblown. Some economists have argued that creditors' losses from bank failures in the past were not so large as to produce significant spillovers. Others have argued that bank creditors exhibit rational behavior even during panics (for example, depositors differentiate between financially sound and unsound banks during panics).6 Policymakers having seen this evidence, the reasoning goes, will have little reason to bail out creditors of large banks in the current period. However, we do not believe that policymakers will change their decision-making based on this research because of the uncertainty they will continue to face. At a minimum, there is disagreement as to the conclusions one should draw from research on banking contagion.7 Moreover, policymakers may not find the conclusions relevant to the particular bank failure that they will face in the future. Policymakers are likely to believe that there is some chance that basing decisions in the current period on past experience will be wrong. The costs of an inappropriate decision (that is, not providing government support in the rare case when it may be needed) will likely appear so high to policymakers as to swamp the rationale for not taking action.
It is clear to us, therefore, that credibility will have to be explicitly and deliberately established through commitment plans that put creditors at risk of loss. We cannot count on addressing moral hazard by jawboning or having creditors spontaneously changing their minds and pricing risk accurately. Consequently, I think it useful to (1) review and categorize strategies conceivably used to establish credibility in addressing moral hazard and (2) provide some general discussion as to which of these strategies are likely to be effective in establishing credibility. We conclude with thoughts about a policy of constructive ambiguity.
To be somewhat systematic about our discussion, we will group strategies by how they affect the incentives to bail out creditors. Strategies include (1) rules or laws that prohibit bailouts of uninsured creditors and thus require policymakers to ignore incentives to provide bailouts, (2) policies that penalize policymakers for bailouts or otherwise increase the explicit disincentives to provide protection, (3) reforms that try to eliminate potential spillovers from the failure of large financial institutions or otherwise mitigate against the very incentives policymakers have for bailing out creditors and (4) appointment of policymakers with a natural disinclination to bail out creditors. (While listed separately, many reform plans use multiple commitment strategies.)
After describing the types of proposals that fall under each category, we will provide some initial thoughts on the likelihood that these proposals will be time consistent. That is, will they be credible and alter incentives enough to make creditors believe they are at risk of loss? These comments are aimed at generating additional discussion and suggestions rather than at presenting a rigorous review. The table summarizes our categories and thoughts.
A direct response to the time-consistency problem is to forbid policymakers from acting on their incentives at the time the decision to bail out creditors must be made. This would be most simply accomplished 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). Despite its benign appearance, this approach is often associated with more radical plans to address moral hazard. For example, plans that eliminate public deposit insurance and replace it with a private insurance system or with nothing at all essentially foreclose government protection.8 Likewise, the collateralized deposit system created by narrow banking plans creates a group of "wide banks" beyond legal coverage. A more recent version of this genre was the system of uninsured wholesale financial institutions that almost made it into the financial modernization legislation. Creditors of these institutions were not covered by deposit insurance.
Chairman Greenspan's organizational approach is also based on legal constructions and rules. He suggests that policymakers will be less likely to bail out creditors, and creditors will be less likely to expect a bailout, because they have a financial relationship with an entity legally distinct from the bank. In other words, these creditors are simply not entitled under any legal regime to ex post protection. As part of that organizational approach, the financial modernization legislation also calls for functional regulation, where bank regulators do not have primary responsibility for monitoring insurance or securities affiliates of the banking organization. Instead, that task falls to securities and insurance regulators. This arrangement also appears to rely on the appearance and legal status of the nonbanking creditors. If bank regulators do not have legal responsibility for the firm, the thinking goes, they will have less legal rationale and ability to effectuate bailouts.
The rule-based approach is quite easy to implement, and it offers a simple and direct way to convey the sentiments of policymakers to creditors. But, is it credible? That is, will it change the decision that policymakers face when a TBTF bank is failing? While these rules might be better than doing nothing at all, I don't think they have a high degree of credibility. To be credible, policymakers must have little desire to evade the rule and such evasion must be difficult. I think both of these tests fail in this case.
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 rationale for evading the rule. Second, there are a fair number of options available to policymakers for circumventing prohibitions, including emergency legislation, lending from the Federal Reserve's discount window and resolution techniques that accomplish the economic substance of a bailout without violating the legal restriction.9 We don't think a complete contract that addresses all the methods that policymakers can use to effectuate bailouts can be written. As such, we think it unreasonable to believe that policymakers will shackle themselves when the incentive benefits of extending the safety net arise, as they surely will some day.
A variety of commitment strategies attempt to raise the costs to policymakers of engaging in bailouts without relying on prohibitions per se. 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 will be reported to the public before bailouts occur. FDICIA raises costs via the extra publicity and formality that would accompany an attempt to bail out creditors. FDICIA combines this approach with a legal rule that requires resolution of a failed institution on a least-cost basis. This explicitly requires the FDIC to quantify the cost of protecting uninsured creditors (as well as creating an explicit ex post levy on banks to pay for bailouts). Resolutions are then subject to review by the General Accounting Office.
Are such disincentives examples of commitment strategies that produce credibility? It does appear that the voting mechanism along with the least-cost test is materially reducing implicit coverage at smaller institutions. However, in previous comments we have noted that the explicit process that FDICIA created for approving bailouts and escaping the least-cost test appears quite similar to the informal process that regulators took prior to FDICIA when they acted in the Continental Illinois case.10 While FDICIA may have raised the disincentives of bailouts, we are not convinced that the disincentives currently exceed the incentives of providing coverage.11
Another commitment strategy that relies on disincentives and has received attention in the literature on credible monetary policy involves explicit contracts. For example, a contract could penalize policymakers monetarily for actions that make implicit coverage more likely or for the coverage itself (as Professor Carl Walsh has suggested for central bankers when they fail to meet inflation targets).12 A second approach would require issuance of debt which pays holders a lump sum if a bailout occurs and nothing otherwise.13 This option raises the explicit costs of bailouts and puts the burden on the taxpayer. Not only could the costs associated with these bonds discourage such bailouts, but also the pricing of these debt instruments would give policymakers the market's assessment of the likelihood of bailouts.
Again, these explicit means of establishing commitment are better than doing nothing. Moreover, these strategies are not as extreme as general prohibitions and thus cannot be dismissed as easily. But, even ignoring their lack of political viability, these contracts may not prove a useful means to alter future decisions of policymakers. The problem here is one of implementation. It appears difficult to write the 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. While we don't think the rationale for evasion remains as strong as it was for prohibition, it is still of concern.
Whereas the previous section reviewed more moderate plans that focused on penalties, this section discusses commitment strategies that try to achieve credibility by reducing the incentives to bail out creditors. Methods to diminish these incentives include reducing the likelihood of spillovers and altering who is at risk of loss.
Spillovers. A major incentive to policymakers to provide coverage is the management of systemic risk and the prevention of spillover failures and contagion. If a plan could eliminate these events, then policymakers would have less incentive to offer such coverage. A few reform efforts of this type focus on the payments system (proponents of this type of plan include Tom Hoenig, president of the Federal Reserve Bank of Kansas City, and Professor Mark Flannery).14 Proponents argue that spillovers from the failure of a large financial institution are transmitted via the payments system. The plans contain methods for limiting or better managing the amount of risk transferred via the payments system, making it less likely, in theory at least, that the failure of one institution will lead to failures or problems at other institutions.
A second type of spillover reduction plan focuses on the amount of loss that uninsured creditors bear. This type of plan does not try to prohibit bailouts, viewing that option as inherently not credible and potentially not desirable. Instead, the goal is to limit the loss that uninsured creditors bear to an amount that motivates them to monitor financial institutions with which they have relationships but which does not lead to their own insolvency. As a result, the failure of one bank would not lead to the failure of other creditor banks. We have proposed a plan which implements this notion through the application of co-insurance for uninsured creditors at TBTF banks.
FDICIA also contains reforms to reduce potential contagion. In particular, the prompt corrective actions (PCA) rules created by FDICIA should limit the number of fixed-income creditors who suffer losses, as equity holders should bear the bulk of those costs.
Finally, a novel suggestion by Professor Frederic Mishkin could address policymakers' concerns about spillovers through a rule.15 Specifically, creditors of the first large institution that fails cannot be bailed out, but creditors of institutions that subsequently fail could receive bailouts. This proposal would allow policymakers to address their fear of spillover failures. At the same time, the plan increases creditors' incentives to price risk correctly so they are compensated for the chance that their bank is the first to fail and they suffer losses.
Risk of loss. A different commitment approach relies on policymakers' disinclination to assist certain types of creditors. Subordinated debt plans exemplify this approach. Holders of this debt are sophisticated and have had ample warning that their low priority in bankruptcy means they should not expect significant proceeds after a bank fails. There is some evidence that holders of banking organizations' subordinated debt already believe they are at some risk of loss.16 Some subordinated debt plans have taken this notion one step further, explicitly requiring particular groups to hold the debt. One plan would have restricted subordinated debt holders to foreign banks. Others fear that these factors may not be enough to prevent the bailout of subordinated debt holders. As such, they also call for something akin to the co-insurance plan described above. For example, the subordinated debt contract could require that the holder of the debt suffer at least some loss if the government provides funds to protect creditors of the failed bank.17
In total, these plans cover a wide range and it is difficult to evaluate their credibility concisely. A few of them seem unlikely to alter incentives and put creditors at risk of loss. For example, PCA as implemented by bank regulators is triggered by book measures of capital, leading to deeply insolvent banks that still report positive capital. And while I am sure distributional matters play some role, I don't think bailouts are largely driven by these issues. Instead, I think policymakers genuinely fear, rightly or wrongly, systemic risk.
The rest of these plans to reduce the underlying incentives to provide bailouts often present quite complex and difficult issues to evaluate. Some have questioned, for example, the ability of policymakers to devise the co-insurance loss rate that would prevent spillovers. Perhaps the focus on payments systems would not adequately limit exposures between financial institutions. I certainly don't intend to review all of these questions, and I'm not sure that all of them are amenable to clear answers, given available evidence. But valid reservations notwithstanding, I would argue that several of these proposals move us in the direction of credibility; that is, they alter 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 those actions. Reforms that focus on prohibitions and disincentives, in contrast, seem likely to fail precisely because they do not reduce the incentives that lead to bailouts in the first place.
So far we have discussed commitment strategies that create new rules or procedures. Another method for addressing the underlying incentives to provide bailouts is to appoint policymakers with significant concerns about moral hazard. These policymakers could be said to have a low discount rate, which means that they will place a high value in the current period on future moral hazard costs. When these policymakers come to decide on bailouts, they will find that the disincentives, namely the future costs of their actions in terms of moral hazard, could very well outweigh the incentives to provide bailouts. As such, their commitment to avoid bailouts will be relatively credible.
This suggestion is a derivation of Professor Kenneth Rogoff's work showing that appointment of a central banker who is more "conservative" than the general population could lead to a credible commitment by the central bank not to inflate.18 This idea has both theoretical and practical appeal, and the reasons that make the Rogoff suggestion welcome for monetary policy would seem to apply here as well.
An interesting question is if the appointment of conservative policymakers is sufficient. Indeed, some may arguebased on insight gained as a policymaker or from close contact with themthat we have overstated the desire of policymakers to provide protection. That is, we could already have fairly conservative policymakers. Even if this were the case, I would argue that society would be better off if steps were taken to establish a credible no-bailout regime.19 Creditors base current decisions on expectations of the future. Ambiguity about future actions by regulators will lead to mispricing that will only be corrected after the regulator allows creditors to absorb losses. Regulators can address this costly learning process by getting creditors to believe in the real potential for future losses and the need to more accurately price risk. Better pricing should lead to a reduction in excessive risk taking. This process could result in fewer potential bank failures, obviating the need for regulators to impose losses in the first place. This reasoning suggests that the appointment of conservative policymakers be accompanied by other commitment strategies.
So far I have discussed plans that take action to establish credibility, even if it were nothing more than one line in legislation that forbade bailouts. This discussion clearly leaves out a policy of constructive ambiguity, where policymakers remain silent as to their policies. As previously suggested, the policy of constructive ambiguity is unnecessarily costly if policymakers oppose bailouts. The lack of an overt strategy to communicate the no-bailout position to creditors creates more doubt than is justified, given the likely actions of the regulator. In contrast, one could argue that ambiguity is better than an explicit extension of the safety net, in that it will produce more market discipline, at least initially.20 But this honeymoon period will come to an end, or at least be greatly attenuated, after a bailout occurs.
Indeed, it is fair to say that we are in such an interim period now. Routinely, those opposed to plans to address moral hazard argue that the FDICIA regime, which includes this type of ambiguity, should be tested before it is changed. Only after we have a period of failures, they argue, should we consider making reforms to the current system.21 I think this policy unnecessarily increases the ultimate cost that society will bear for excessive risk taking by financial institutions. The recent passage of financial modernization legislation only reinforces my conviction that we need to act before losses pile up. Remarks by Chairman Greenspan suggest similar concerns and desire to address perceptions of TBTF.
"There are many that hold the misperception that some American financial institutions are too-big-to- fail. I can certainly envision that in times of crisis the financial implosion of a large intermediary could exacerbate the situation. Accordingly, the monetary and supervisory authorities would doubtlessly endeavor to manage an orderly liquidation of the failed entity, including the unwinding of its positions. But shareholders would not be protected, and I would anticipate appropriate discounts or "haircuts" for other than federally guaranteed liabilities."22
But it is not adequate to only voice these sentiments. Rather than waiting passively to tally costs, I suggest we continue to evaluate reforms in terms of their credibility, as well as other dimensions that will determine their effectiveness. The goal is implementation of steps to increase credibility as soon as feasible.
This column is adapted from a speech Stern presented in April at "The Future of Finance: Globalization, the Internet and Regulatory Reform" conference, sponsored by the University of Chicago's George J. Stigler Center for the Study of the Economy and the State.
2 Ibid, 3.
4 Ibid, 6.
5 Of course, there could be political fallout from providing support as well, but we assume this will play a smaller role in decision making.
6 For a summary of this work see George Benston and George Kaufman, "Is the Banking and Payments System Fragile?" Journal of Financial Services Research 9 (3-4) 1995, pp. 15-46.
7 Stuart Greenbaum, Comment, same issue, pp. 105-108.
8 For examples see: Bank Administration Institute (1996), Building Better Banks: The Case for Performance-Based Regulation, (Chicago, IL: Bank Administration Institute), pp. 63-70; Richard Kovacevich (April 1996), "Deposit Insurance: It's Time to Cage the Monster," Federal Reserve Bank of Minneapolis fedgazette, pp. 14-15; Bankers Roundtable (1997), Deposit Insurance Reform in the Public Interest (Washington, D.C.: Bankers Roundtable).
9 President Alfred Broaddus of the Federal Reserve Bank of Richmond recently expressed concern that discount window lending could offer a way to support failing institutions. See Remarks by J. Alfred Broaddus Jr., "Market Discipline and Fed Lending" for the Federal Reserve Bank of Chicago's Bank Structure Conference, May 5, 2000.
10 Ron Feldman and Arthur Rolnick, "Fixing FDICIA: A Plan to Address the Too-Big-To-Fail Problem," Federal Reserve Bank of Minneapolis Annual Report, March 1998.
11 Others have recently made the point that FDICIA did not eliminate TBTF. For example, see Anthony Santomero and David Eckles, "The Determinants of Success in the New Financial Services Environment," Federal Reserve Bank of New York, Economic Policy Review, October 2000, p. 8. [PDF Format]
12 Carl E. Walsh, "Optimal Contracts for Central Bankers," American Economic Review 85 (March 1995), pp. 150-167.
13 Wall has suggested a similar type of instrument called capital notes that would measure the health of the deposit insurance system. Larry Wall, "Taking Note of the Deposit Insurance Fund: A Plan for the FDIC to Issue Capital Notes," Economic Review, 82, 1997, pp. 14-31. Insurance firms have issued so-called catastrophic bonds that make payouts if a disaster occurs but pay nothing if it does not. In addition, several financial institutions issued so-called goodwill certificates that paid investors some portion of payments the firms would receive if they won a series of legal decisions from the federal governments.
14 Thomas Hoenig (1996), "Rethinking Financial Regulation," Economic Review, 81 (2), pp. 5-13, and Mark Flannery. "Modernizing Financial Regulation: The Relation Between Interbank Transactions and Supervisory Reform," forthcoming in Journal of Financial Services Research.
15 Frederic Mishkin, "Moral Hazard and Reform of the Government Safety Net," paper presented at "Lessons from Recent Global Financial Crises" sponsored by the Federal Reserve Bank of Chicago, October 1, 1999; and Frederic Mishkin and Philip Strahan, "What Will Technology Do to Financial Structure," Brookings-Wharton Papers on Financial Services, 1999.
16 Mark Flannery and Sorin Sorescu, 1996, "Evidence of Bank Market Discipline in Subordinated Debenture Yields: 1983-1991," Journal of Finance LI (4), pp. 1347-1377.
17 William Lang and Douglas Robertson, "Analysis of Proposals for a Minimum Subordinated Debt Requirement," Office of the Comptroller of the Currency Economic and Policy Analysis Working Paper 2000-4, March 2000.
18 Kenneth Rogoff, "The Optimal Degree of Commitment to an Intermediate Monetary Target," Quarterly Journal of Economics 100 (November 1985), pp. 1169-1190.
19 Richard Clarida, Jordi Gali, and Mark Gertler, "The Science of Monetary Policy: A New Keynesian Perspective," Journal of Economic Literature, Vol. XXXVII (December 1999), p. 1680.
20 Douglas Cook and Lewis Spellman,"Taxpayer Resistance, Guarantee Uncertainty, and Housing Finance Subsidies," Journal of Real Estate Finance and Economics, 5, pp. 181-195.
21 This is the implicit theme of recent summary of deposit insurance reforms by the FDIC. See George Hanc, "Deposit Insurance Reform: State of the Debate," FDIC Banking Review (December 1999), pp. 1-26.