Gary H. Stern
Federal Reserve Bank of Minneapolis
University of Minnesota Economics Roundtable
March 11, 1998
An appropriate subtitle for these remarks is Say What You Mean, and Mean What You Say. As you will see, this phrase is essentially an admonition applicable in establishing and managing the safety net underpinning a nation's banking system. For the safety net should promote financial and economic stability and yet, if poorly designed, it may do just the opposite, thereby contributing to excessive risk taking and instability. Indeed, the experience of many countries, industrialized and developing alike, suggests that such perverse effects have in fact occurred from time to time.
We believe that there is ample evidence in support of a safety net to provide banking system stability, but we also believe that its scope and operation must be defined carefully. Ultimately, the issue transcends banking or financial system stability, for there is now considerable evidence that the extent and quality of a country's financial infrastructure is a significant determinant of economic growth.
The discussion today is organized as follows. I first review the late 1980searly 1990s banking problems in the U.S. and emphasize the role of incentivesessentially, the moral hazard problemin our experience. After noting that this problem is not unique to the U.S., I proceed in the next section to discuss the value of the government safety net underpinning banking, arguing that in its absence excessive instability could result. The subsequent section takes up the issue of reconciling the objectives of promotion of stability and limitation of moral hazard costs, and presents a proposal, with coinsurance at its core, to do just that. I conclude, before turning to comments and questions, with a brief summary.
The savings and loan fiasco and increased failures of commercial banks in the U.S. during the 1980s and early 1990s consumed considerable resources. Explicit resolution expenses were quite high, of course, as the S&L cleanup is estimated at 2 1/2 percent or more of GDP. Indeed, this estimate is probably low, since it ignores both deadweight losses resulting from resource misallocation and damage to public confidence in the financial and political system.
Despite its size and significance, the American experience of this period was not unique. The World Bank reports that at least 70 banking crises occurred in the 1980s throughout the developed and developing worlds. The quantitative impact of some of these crises equaled or exceeded the U.S. experience. Interestingly enough, Asia saw its share of banking crises in the 1980s in countries such as Indonesia, Malaysia, Japan, Thailand, and the Philippines.
To be sure, each banking crisis seems as unique to the country affected as the American debacle seemed to us. While detailed understanding of the financial and legal system of each country is required to speak about a given situation with authority, economists have identified a number of factors common to many. In particular, banking crises are frequently associated with:
Macroeconomic shocks that increase the variability of inflation, exchange rates, interest rates, or capital flows;
Deficiencies in supervisory policies in areas such as capital standards or restrictions on insider activities;
An unwillingness to shut down insolvent institutions (i.e. forbearance); and
Poor credit decisions and inept management of credit risk, which are often related to government influence on credit allocation.
The preceding list is far from comprehensive. Staff and consultants at the Bank for International Settlements, the IMF, the Inter-American Development Bank, and the World Bank have all produced reviews in the last several years indicating, not surprisingly, that many factors are responsible for any given banking crisis.
A factor deliberately omitted from the discussion to this point is the breakdown in, or absence of, market discipline, caused by a belief that the government will absorb the losses that bank creditors would otherwise bear. This is the well-known problem of moral hazard. In the U.S., moral hazard resulted in part from the deposit insurance system that provides explicit coverage to $100,000 per account. Protected depositors did not have adequate incentive to price their funding according to the riskiness of the bank, thereby allowing some institutions to attract funds at below market rates. So-called brokered deposits exacerbated this problem significantly, since the $100,000 umbrella could be exploited despite geographic and other natural barriers.
But perhaps more devastating than explicit insurance coverage was the influence of the government's implied support. Often times the coverage of uninsured depositors is referred to as a policy of too-big-to-fail, implying that, for systemic reasons, an institution is too important to expose its creditors to loss. Actually, however, widespread protection of all uninsured depositors is a more accurate description, for the FDIC protected 99.7 percent of all deposits at failed commercial banks from 1979 to 1989. This implicit coverage created widely held expectations of government protection of depositors, especially at large institutions, and further undermined market discipline significantly.
Moral hazard is not unique to the U.S. Indeed, it is credited to some degree with contributing to the banking problems experienced in a long list of countries, including Japan, the Philippines, Taiwan, Indonesia, South Korea, and Thailand.
Equally important, there should be little doubt at this stage that inadequate or incapacitated financial institutions and markets inhibit economic development and industrialization. A recent article in the Journal of Economic Literature persuasively makes this case, and I refer you to that publication.
If a robust financial system is critical to economic development, and if moral hazard serves to damage the financial infrastructure, why not eliminate government protection of depositors? In other words, why not clean up the incentives and achieve enhanced market discipline? Where explicit insurance coverage does not exist currently, it could be made clear that, as a matter of public policy, losses from banking failures will be borne by depositors and other creditors. Admittedly, this reform would be more difficult to accomplish where there are explicit guarantees or well established precedents in place, but extended lead times could be a vehicle for smoothing the necessary adjustments in such circumstances.
Nevertheless, this is not a policy I would advocate, because commercial banks remain, in a very real sense, special. Banks, after all, offer liabilities payable on demand at par but hold a wide range of assets with diverse maturities, degrees of liquidity, and credit quality. Cognizant of these features and recognizing their informational disadvantage relative to bank management, depositors have an incentive to run the bank when other depositors do or if the bank is perceived, rightly or not, to be in financial difficulty.
The answer to the question posed earliernamely, why not eliminate government protection of depositors?is, then, that excessive instability would result. We need to recognize that in the U.S., deposit insurance backed by the government has in fact produced an extended period without a banking panic, albeit at significant cost in terms of moral hazard. In contrast, privatization options, say a private deposit insurance company or a self-insurance scheme lacking the credit quality of the U.S. government, may not be successful in preventing panics because they could be subject to the same disadvantages, and risk of loss, as depositors.
Finally, a policy of simply ruling out protection of depositors and other creditors, in all circumstances, may not reduce moral hazard effectively because such a stance would likely lack credibility. There may in fact be occasions when the size of the failing institution is such that its downfall could lead to severe spillovers throughout the financial system and, ultimately, the real economy. Government intervention in such a case could well pass a cost/benefit test. Further, market participants in countries which have a long history of government intervention in many facets of the economy are not likely to believe statements of nonintervention, and thus are not likely to increase market discipline of financial institutions, no matter what the authorities say.
The challenge to policymakers in this arena is to balance two competing objectives, namely banking system stability and minimizing the cost of moral hazard. There is clearly a trade-off here: stability can be achieved at the expense of high moral hazard costs, or moral hazard can be eliminated at the cost of instability. But how can the two be reconciled?
As it turns out, this is a question to which several of us at the Federal Reserve Bank of Minneapolis (and the University of Minnesota, I might add) have given considerable thought over the years. There are three key features of our plan to limit the probability of banking panics and the size of the moral hazard problem:
The reforms of FDICIAthe Federal Deposit Insurance Corporation Improvement Act (of 1991)designed to reduce excessive risk taking by insured institutions should be retained. These reforms include risk based deposit insurance premia, risk based capital standards, and prompt corrective action with regard to institutions whose capital becomes impaired. While FDICIA is not perfect, it does effectively establish a framework for having higher risk institutions pay higher premia and hold more capital, and for closing banks as they approach, but before they reach, insolvency.
The current level of explicit deposit insurance coverage would be maintained, although we might eventually want to consider limiting it to one account per social security number. The implication of this provision is that small depositors would be protected.
Market discipline would be enhanced by convincing formally uninsured depositors that they will necessarily bear some loss in a failed bank situation. Essentially, we have in mind a coinsurance provision which would be added to the initial FDICIA reforms.
More specifically, FDICIA requires a series of steps before
too-big-to-fail rescues of institutions can occur. These are that: (1) the Secretary of Treasury must find that least cost resolution would have serious adverse effects on economic conditions and financial stability and that the provision of extra legal insurance coverage would avoid or mitigate such adverse effects. (2) The Treasury Secretary must consult with the President in making this determination. (3) Two-thirds of the governors of the Federal Reserve System and two-thirds of the directors of the FDIC must approve the coverage. Under our proposal, the FDICIA legislation would be amended so that uninsured depositors would take a limited but meaningful loss (say, perhaps, 20 percent of the uninsured portion of their deposit) when a too-big-to-fail rescue occurs, pending, of course, the ultimate disposition of the assets of the failed institution. It is obviously crucial that market discipline be introduced carefully, because losses that proved excessively disruptive would doom such a policy.
We see several advantages to this proposal, taken in its entirety. As noted above, small depositors remain protected (and can, therefore, sleep comfortably). Secondly, the approach builds on FDICIA, legislation in place and with an established degree of credibility. What is added is codification of the loss of the uninsured depositors. In this sense, the government says what it means and means what it says. Third, the loss is imposed on uninsured depositors in a way which limits spillovers and, therefore, should not add materially to the probability of banking panics, yet is sufficiently meaningful so that depositors have significant incentive to gather and assess information about banks and, perhaps, to diversify more than otherwise. Under this proposal, we would expect the market for information about the financial condition of banks to broaden and deepen over time. We would also note that the size of the loss imposed on uninsured depositors could be adjusted over time, so disruption could be contained by starting at a low level.
In summary, I would note that banking systems historically are subject to crises, and policies to reduce the probability of crises are frequently quite costly in their own right because of excessive risk taking resulting from moral hazard.
We are not aware of a solution to this dilemma. Nevertheless, there are, we believe, two principles appropriate for managing the trade-off between instability and moral hazard. First, incentives in favor of market discipline are essential to contain the moral hazard problem. As we see it, market discipline is a complement to regulation and supervision of banks and not a substitute for them. But we have ample experience to suggest that supervision alone is incapable of adequate restraint of moral hazard. Second, the government's policy of support for the creditors of a failed bank should be explicit and in legislation. As we have suggested, such a step, if carefully crafted, can enhance market discipline and limit potentially disruptive spillover effects while, at the same time, adding to the government's credibility.