NBER WORKING PAPER SERIES
HOW BIG A PROBLEM IS TOO BIG TO FAIL?
Frederic S. Mishkin
Working Paper 11814
晚安温馨的句子www.nber/papers/w11814
形容掌声的词语NATIONAL BUREAU OF ECONOMIC RESEARCH
1050 Massachutts Avenue
Cambridge, MA 02138
December 2005
This review essay has been written for the Journal of Economic Literature. I thank for their helpful comments Beverly Hirtle, Don Morgan and Til Schuerman and Beverly Hirtle for supplying me with data. Any views expresd in this paper are tho of the author only and not tho of Columbia Univer
sity or the National Bureau of Economic Rearch. The views expresd herein are tho of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Rearch.
©2005 by Frederic S. Mishkin. All rights rerved. Short ctions of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given to the source.
How Big a Problem is Too Big to Fail?
Frederic S. Mishkin
崇文门饭店
NBER Working Paper No. 11814
December 2005
JEL No. G21, G28
ABSTRACT
应用型人才
This review essay examines whether too-big-to-fail is as rious a problem as Gary Stern and Ron F
eldman contend. This essay argues that Stern and Feldman overstate the importance of the too-big-to-fail problem and do not give enough credit to the FDICIA legislation of 1991 for improving bank regulation and supervision. However, this criticism of the Stern and Feldman book does not detract from many of its messages. Even if the too-big-to-fail problem is not as rious as they contend, the policies they outline can make it less likely that a banking crisis will occur even if driven by other factors.
Frederic S. Mishkin
Graduate School of Business
Uris Hall 817
Columbia University
New York, NY 10027
and NBER
fsm3@columbia.edu
Banking institutions are especially well suited to minimizing transaction costs and adver lection and moral hazard problems. This is why banks are “special” and play such an important role in the financial system. When banks fail, the information capital they have developed may disappear, and as a result, many borrowers will not have access to funds to pursue productive investment opportunities.2 Indeed, if a large enough number of banks fail at the same time, in other words a banking panic occurs, the economy’s ability to channel funds to tho with productive investment opportunities may be verely hampered, leading to a full-scale financial crisis and a large decline in investment and output. Indeed, the worst economic downturns are almost always associated with bank panics and financial cris, and not only has this been true in the United States when the Great Depression was triggered by banking panics (Friedman and Schwartz, 1963, Bernanke, 1983, and Mishkin, 1991), but it has been true in recent years in emerging market countries (Mishkin, 1996, Bordo and Eichengreen, 2002).
Concerns about bank panics have led most governments throughout the world to provide a safety net for the banking system. Federal government deposit insurance, a guarantee of repayment for depositors, was established in the United States when the FDIC (Federal Deposit Insurance Corporation) started operations in 1934. Deposit insurance can short circuit bank panics by providin
g protection for depositors. When a depositor has fully insured deposits, up to $100,000 of deposits in the United States, the depositor doesn’t need to run to the bank to make a withdrawal when she is worried about the bank’s health becau her deposits will be worth 100 cents on the dollar no matter what. Hence, deposit insurance can short circuit runs on banks and bank panics and can overcome reluctance by depositors to put their funds into the banking system.
For the first thirty years after the FDIC was created, only six countries emulated the United States and adopted deposit insurance. However, this began to change in the late 1960s, with the trend accelerating in the 1990s when the number of countries adopting deposit insurance doubled to over venty and now is clo to 90 (Demirguc-Kunt and Kane, 2002, and Demirguc-Kunt, Kane and Laeven, 2005) . Deposit insurance is now the norm in much of the world.
香菇排骨汤Deposit insurance is not the only way in which governments provide a safety net for depositors. Even without explicit deposit insurance, many countries provide a safety net by providing direct support to domestic banks. This support is sometimes provided by lending from the central bank to troubled institutions as part of the central bank’s lender of last resort role or by direct government infusion of cash into the institutions.
The good news of having a government safety net is that it can prevent bank panics, as it has since the establishment of the FDIC in the 1930s. The bad news is that creates moral hazard incentives for banks to take on greater risk. When a depositor is fully protected, she knows that she will not suffer loss if a bank fails, and thus has little incentive to monitor the bank’s
-1-
activities and withdraw funds if the bank is taking on too much risk. Without this discipline from depositors, banks know that they can engage in risky activities with impunity, and this can increa the probability of bank failures.
The moral hazard created by a government safety net is even more vere for large banking institutions becau when they fail, it can lead to systemic risk in which the whole banking system is threatened. The failure of a large institution not only can cau immediate failures of its counterparties in both the banking and the rest of the financial system, but can also lead to a crisis of confidence that may spill over to other banks and financial institutions, leading to a cascade of failures and a financial crisis. Given the potential costs to the economy from a large bank failure, governments are very reluctant to let large banking institutions fail, or if they do, impo any costs on 小鸡孵化技术
depositors, even if deposit insurance is limited to a fixed amount, say $100,000. A particular manifestation of this phenomenon occurred when Continental Illinois, then one of the ten largest banks in the United States became insolvent in May 1984. Not only did the FDIC guarantee depositors up to the $100,000 insurance limit, but it also guaranteed all accounts exceeding $100,000 and even prevented loss for Continental Illinois bondholders. Shortly thereafter, the Comptroller of the Currency (the regulator of U.S. national banks) testified to Congress that eleven of the largest banks would receive a similar treatment to that of Continental Illinois. Although the Comptroller did not u the term “too-big-to-fail” (which was actually ud by Congressman McKinney in tho hearings), this term now is applied to a policy in which the government provides guarantees of repayment of large uninsured creditors of the largest banks, so that no depositor or creditor suffers a loss, even when they are not automatically entitled to this guarantee. (The “too-big-to-fail” characterization is somewhat of a misnomer becau under the too-big-to-fail policy, banks are often clod or merged into another bank, and then the managers are often fired and the equityholders in the bank lo much of their investment.) The too-big-to fail policy increas the moral hazard problem for big banks. If a deposit insurance agency like the FDIC were willing to clo a bank and pay off depositors only up to the $100.000 insurance limit, large depositors would suffer loss if the bank failed. Thus they would have incentives to monitor the bank’s activities clos
ely and pull their money out if the bank is taking on too much risk. To prevent such a loss of deposits, the bank would be less likely to engage in risky activities. However, once large depositors know that a bank is too-big-to-fail, they have no incentive to monitor the bank becau no matter what the bank does, large depositors will not suffer any loss. The result of the too-big-to-fail policy is that large banks are likely to take on greater risks, thereby making bank failures more likely. Indeed, this is exactly what we saw happen in the United States in the 1980s when large banks took on riskier loans than smaller commercial banks which led to higher loan loss for big banks (Boyd and Gertler, 1993).
-2-
Many analysts have argued that the the landmark legislation in 1991, the Federal Deposit Insurance Corporation Improvement Act (FDICIA), improved banking regulation substantially and has made the too-big-to-fail problem less vere. Gary Stern, the president of the Federal Rerve Bank of Minneapolis, and Ron Feldman, a vice president of the Bank, both insiders in the world of bank regulation and supervision, have written a book that argues that not only was the too-big-to-fail policy a rious problem in the past, but that it was not fixed by FDICIA. Furthermore, it has even gotten wor in recent years becau of the increasing size and complexity of banking organizations.
Given the verity of the problem, Stern and Feldman believe that it is imperative that policymakers adopt policy measures to deal with it.
In this review essay, I examine whether too-big-to-fail is as rious a problem as Stern and Feldman believe it is. My view is that they overstate the importance of the too-big-to-fail problem and do not give enough credit to FDICIA for improving bank regulation and supervision. However, this criticism of the book should not detract from many of its messages. Too-big-to-fail is still a problem of great concern to bank regulators.
I.
How Big a Problem Was Too-Big-To-Fail
After an initial introductory chapter which lays out the basic messages of their book, chapter two provides an insightful discussion of what the problem is all about. Stern and Feldman stress, quite rightly, that the too-big-to-fail problem is due to a lack of credibility of policymakers’ commitment to not bail out large banks. This lack of credibility is just another manifestation of the time-inconsistency problem first discusd by Kydland and Prescott (1977) and Calvo (1978). Policymakers’ pledge not to engage in a bailout of large bank is not time consistent: when a large ba
辰楠旧事nk is about to fail, policymakers will want to renege on their pledge becau they want to avoid the systemic risk that the failure of the bank would entail. Uninsured creditors knowing that policymakers have incentives to renege will assume that the bailout will occur and thus will not monitor large banks sufficiently, leading to the too-big-to fail problem. Stern and Friedman also point out that lower caps on deposit insurance or elimination of deposit insurance altogether are also not credible and are subject to the same time-inconsistency problem.
This innovative way of thinking about too-big-to-fail leads to an important implication for the policy debate. In order to reduce the too-big-to-fail problem, the incentives for policymakers to renege on a no bailout commitment has to be reduced, which requires policy measures that reduce the costs of a failure of a large bank to the financial system by reducing
-3-
the spillovers from such a failure. With less of an incentive for policymakers to renege on no bailout pledges, uninsured creditors will worry that large risk-taking banks will expo the creditors to loss and so creditors will pull funds from the banks, thereby imposing market discipline that will reduce moral hazard risk-taking by the banks.
Chapter three contains an excellent discussion of why a too-big-to-fail policy is costly. Stern and Feldman point out that not only does too-big-to-fail increa risk-taking by banks which increas the likelihood of banking cris, but it also leads to resource misallocation. The possibility of a bank bailout makes it more likely that banks will not operate in a cost-efficient manner and also may innovate less. One important misallocation that they do not mention is that the prence of too-big-to-fail encourages banks to grow in size to take advantage of the too-big-to-fail subsidy, so that banks will be larger than is socially optimal and there will be too many bank mergers.
Chapter four discuss the evidence on whether too-big-to-fail is a pervasive problem. The key to the verity of the too-big-to-fail being a problem is that the market expects that government bailouts occur. The chapter contains a very clear review of the evidence that the market reflects too-big-to-fail. Event studies have shown that the testimony by the Comptroller of the Currency in the aftermath of the Continental Illinois bailout that the eleven largest banks would be subject to the too-big-to fail policy did experience higher returns than other banks after this announcement.3 Furthermore, there is evidence in market reactions that too-big-to-fail coverage spread to other banking institutions not on the Comptroller’s original list. Mergers undertaken by the largest banks result in an increa in market value for shareholders, while this is not the ca for smaller banks, su
ggesting that the market has priced in the subsidy to larger banks from too-big-to-fail. Costs of deposits also appear to be lower for larger banks that benefit from too-big-to-fail. Credit ratings also appear to reflect too-big-to-fail, with larger banks having higher credit ratings when they take account of possible government support. Yields on bonds issued by banks (which are typically quite large) in the early to mid 1980s did not em to reflect much risk.
Stern and Feldman also argue that too-big-to-fail has played an important role in the numerous banking cris throughout the world that have occurred in the last two decades. I find this argument more suspect. They cite statements like that by Honohan and Klingebiel (2000) that “‘Unlimited deposit guarantees, open-ended liquidity support, repeated capitalization, debtor bailouts, and regulatory forbearance’ are associated with a tenfold increa in the fiscal cost of banking cris,” (Stern and Feldman, 2004, p. 40) as supporting their position. Although I agree with Honohan and Klingebiel’s characterization of banking cris, it is more accurate to attribute banking cris not to too-big-to-fail, but rather to “too-politically-important-to-fail”
-4-
>音域是什么意思
