The topic of this paper is the general responsibility of government for the stability of financial markets. Here in the United States we have had in the recent past marked volatility in financial markets—stock market ups and downs, junk bond bubbles, massive thrift institution insolvencies—and all of these have caused questions to be raised about banks and other financial intermediaries. Such an environment could well prompt a central bank to consider what it should be doing to ensure the stability of financial markets and the costs and benefits of its possible actions.

Two potentially conflicting assessments of central bank actions and tools make this a pertinent topic. On the one hand is the sense that central banks and governments have gone too far in their efforts to ensure stability, and by doing so have broken down market discipline and produced a misallocation of resources—even promoted risk in the financial system. This is said most forcefully in the United States about deposit insurance, but often central bank use of the discount window as lender of last resort is cited, and occasionally it is claimed that monetary policy has contributed by trying too hard to cushion fluctuations in financial markets. On the other hand, some hold that central banks (in the United States, the Federal Reserve) may not have the tools and the ability to do what they need to do to safeguard the stability of the financial system. Differences among commercial banks, investment banks, and other types of financial intermediaries have eroded. Our current instruments of policy, particularly central bank instruments, are keyed very much to commercial banks and other depositories. According to this second view, in a world in which finance is global, in which commercial banks, investment banks, even insurance companies are blending together, participating in each other’s markets, forming and reforming in different combinations, it is time to question whether the traditional central bank tools are adequate to the task.

The United States is not alone in facing this problem. In the European context, my understanding is that while some thought has been given to these issues, there is a lot of ambiguity as to how they would be dealt with as monetary union comes closer.¹ Because this is a complex transition period in Europe, the chances of financial failure and the risk of financial problems may be higher than usual. Competitive forces and markets are being reshaped; in the process it is likely that some important players will fail. It may take a while for monetary and fiscal policies to adapt, raising the odds on macroeconomic instability. The chances of contagion—that is, the chance of one failure leading to a more generalized loss of confidence—are higher in this transition period in which people do not know what the steady state market structures are going to look like.

These issues should be front and center for Europe and for other countries as well. Why do we care about financial market stability? What is important about it? What central bank instruments do we have to deal with threats to financial market stability? How have these tools been used? How might these tools be adapted to a changing world?

Our concern with financial market stability is really a concern with the implications of the financial markets for the real economy—for inflation, employment, and output. Fluctuations in stock prices and interest rates and transfers of net wealth among market participants are not problems in themselves. They are the signs of a vigorous economy, of asset markets adjusting to new information, to new developments. The fact is that these fluctuations produce winners and losers; some people do better than others as a reward for information, insight, or good luck. Fluctuating asset prices, and people going into and out of business, are part of a market economy. They should not really concern the central bank, unless they are feeding back on to the real economy and causing instability in what are most important to the residents of any country: their jobs, their real incomes, and the prices they are paying for goods and services.

One way problems can emerge from the ebb and flow of financial markets is when prices tend to overshoot their long-run values. The academics at the 1988 Kansas City conference² on market volatility who were defending a strict interpretation of what economists call the efficient market hypothesis—that is, asset prices are exactly what they ought to be given all available information—were on the defensive, since this was after the crash of October 1987. There was consensus that although an analyst might not be able to predict the next movement in asset prices—that is, those prices might move in a random walk—overall there was excess volatility in the markets. Markets do tend to overshoot in asset prices; stock prices, interest rates, and foreign exchange rates sometimes move in ways that cannot be explained from the fundamentals. This has consequences. A huge rise in the stock market, as the Japanese, for example, have seen, and as the United States experienced several years ago, carries a risk of inflation and of overspending. When people’s assets are worth more, they will increase spending based on that wealth. Excess volatility seems to happen frequently in the real estate markets. If real estate prices go up, the people who own the land and the buildings start spending that gain, and overshooting can contribute to inflation.

Japanese central bankers initially tended to be concerned about the implications for their economy of the increase in stock and land prices in Japan. The United States has been a bit more focused on the downside of prices. When asset prices go down, they destroy wealth, dampen spending, and raise unemployment.

Declines of prices with overshooting on the downside are a particular concern to central banks because they tend to happen faster and more suddenly than increases. Perhaps the reason for this is confidence. As prices start dropping and as people see their wealth being destroyed, they can lose confidence—they may try to liquidate their assets and move to safer investments. In this case, you get a “flight to quality” or “flight to liquidity,” which builds on itself. People try to sell their shares; the stock price goes down; they become more concerned; the stock price goes down even further. This overshooting on the downside can happen very rapidly. As savers withdraw from lending, as they move toward liquidity, and as they move toward safe credits such as Treasury bills, then borrowers lose their access to credit or find it substantially more expensive. Those who are not among the safest borrowers have considerable trouble finding credit; they cut back on their spending and so maintain or accelerate the downward spiral, with effects on spending, employment, and prices.

Banks and other depository intermediaries have a key role in this potential spiral. People deposit savings, and banks lend those savings to those who want to spend. Banks are an important intermediary between those who own the assets and those who borrow and do the spending. Moreover, payments in the society generally go through banks. For these reasons banks are at the heart of the saving-investment process. At the same time, banks are particularly vulnerable to a loss of confidence. The liabilities they borrow tend to be more liquid—of a shorter duration or shorter maturity—than what they lend. Their assets tend to be less liquid, harder to turn into cash. Banks, thus, are particularly vulnerable to people losing confidence and withdrawing funds. If, in the face of withdrawals, particular banks have trouble liquidating their assets, there can be a major loss of confidence in the entire financial intermediary system.

Central banks are expected to forestall and protect against this sort of financial crisis. In the United States, the motivation for founding the Federal Reserve arose importantly out of the financial crises of 1893 and 1907.³

What tools do central banks and governments have to deal with the threat of such a crisis? Traditionally, these are microeconomic tools that apply to the banks or depositors involved. The main means, of course, for the central bank has been the discount window. If financial intermediaries, banks in particular, have illiquid loans on their books and depositors who want to withdraw their funds, one approach is to make those loans liquid—to allow them to be brought to the central bank and for the central bank to lend against those assets to meet depositor withdrawals. Not only does this enable commercial banks to meet their deposit outflows, but by doing so, it prevents a loss of confidence. Depositors may not even ask for their cash if they know the central bank is prepared to make it available.

Traditional theory prescribes that the discount window be available at a penalty rate to discourage people from using it unnecessarily.⁴ In the United States, for many years, the discount window operated with relatively low interest rates—in the past couple of decades at a discount rate that was below market rates. Now, however, after 30 days of borrowing for extended credit, the rate goes above market rates. Moreover, most banks with long-term liquidity problems face some very severe penalties. The Federal Deposit Insurance Corporation (FDIC), working in conjunction with other regulators and the Fed, may take over troubled borrowers if they are not likely to recover.⁵ If this happens, the shareholders will probably lose their value and the managers will lose their jobs.

The discount window is one microtool to deal with loss of confidence. Another, of course, is deposit insurance, originally intended to protect small depositors from losses and therefore to retain their confidence. Deposit insurance, not a central bank function per se, was put into place in the United States when the discount window did not work in the early 1930s and was seen as insufficient for the job.⁶

A third key on the microlevel is supervision and regulation, including capital requirements⁷ and examinations. These mechanisms are designed to reduce the odds that the discount window or deposit insurance will be called on. For banks, being subject to examination is a trade-off for the right to use the stabilizing instruments of the central banks and governments. To have access to the discount window and deposit insurance, a bank or other institution must also endure the examiners checking on its activities and second guessing its management’s decisions.

A second set of tools we have for supporting financial stability is macro-economic: open market operations and changes in the discount rate to affect interest rates and the money supply. The most important way to ensure financial stability is through good macroeconomic policy, for a reasonably sensible macroeconomic policy will alone provide much of the assurance needed to maintain confidence and financial stability.

In many respects, both sets of tools work most effectively together. Some have said that aggregate economic policy—open market operations—can replace the discount window function of providing liquidity. But the discount window and deposit insurance are still essential. The macro- and micro-instruments complement each other. A discount window and deposit insurance add protection against unforeseen shocks to the financial system that cannot quickly be countered by macro policy, or even against a bad macroeconomic monetary policy itself. Without the discount window, deposit insurance, and some other safety features, macroeconomic and monetary policy could be too timid, concerned that tightening credit would bring on a financial crisis. These backstops help give policymakers some insurance against systemic failures even if something unexpected happens.

In applying these instruments, central banks are tugged two ways. One is a desire to ensure financial stability—to avoid contagion effects and the loss of confidence. There must be assurance that the financial markets are going to operate in an efficient way, that confidence is maintained in financial intermediaries, and that people will continue to spend and save.

But leaning too far on the side of ensuring financial stability has its costs. With respect to macroeconomic policies, inflation can increase if the necessary steps to damp demand are not taken when needed. In the late 1980s some worried that the Federal Reserve would allow prices to accelerate because it was too concerned about the leveraging in the economy and the status of banks and thrifts to take steps to tighten interest rates and clamp down on the growth of the money supply. In this view, there is a built-in bias toward an overly expansive monetary policy. Moreover, if the markets think that the Fed is going to ensure financial stability at any price, the results could be the very market behavior it is trying to avoid. Such thinking would promote overleveraging and asset price overshooting on the upside, which inevitably would have to be unwound at some point with greater risk of damage to the financial system.

This problem, associated with running monetary policy in a time of financial fragility and in the context of concern about financial market stability, is a form of what economists call moral hazard—that is, insuring too much against downside risks promotes market inefficiencies. The term “moral hazard”⁸ is more often applied to micro policies, such as the discount window and deposit insurance. When governments protect the financial system too much, people who lend money to financial intermediaries have no incentive to discriminate between good and bad creditors. If they know the government will help them out, why should they penalize those who engage in risky operations? Authorities leaning in one direction to ensure financial market stability promote excessive risk taking by banks and other lenders and distort resource allocation in the economy. This behavior can be avoided, to a certain extent, through supervision and regulation. By replacing market forces with regulations and examiners and by mandating higher capital than the intermediaries would like to have, a degree of control over risk taking is obtained.

Have we gone too far in ensuring financial stability? Have we promoted excessive risk taking and inefficient allocation of resources? Evidence comes in every day as we get new estimates of the huge cost to bail out the thrift institutions. The economic inefficiencies and excessive risk taking have come back to haunt the American taxpayer, who underwrites the Federal deposit insurance system. That system did not have the proper incentives for thrifts or the supervision and regulation that are needed to prevent this kind of behavior. Commercial banks also have engaged in high-risk enterprises, from real estate lending in the 1970s, to lending to less-industrialized countries in the late 1970s and in the early 1980s, to real estate lending again in the 1980s. One would be hard-pressed to argue that commercial banks have been getting the right signals from the market and have been managing their risk in a way that helps society to allocate resources correctly.

In theory, central banks ought to be able to differentiate between the systemic risks—the contagion risks, the confidence risks—and the risk to the individual firm. In theory, the individual firm should be allowed to fail, and the effects on the rest of the financial markets should be contained. But it is extremely difficult to differentiate systemic risks from firm risks, especially when firms grow to become major players in financial markets.

Partly as a result of this, we have commercial banks said to be so large that the government will not let them fail because of their extensive relations with other entities in the markets. Continental Illinois was one example; it had a considerable volume of loans from smaller commercial banks, so if Continental failed, it would bring many other banks down with it. This situation is often referred to as “too big to fail.” Note, however, that the phrase is rather ambiguous. Continental Illinois was not too big to fail in the sense that management lost their jobs. Moreover, the shareholders—the owners—lost all their investment. Still, these penalties may not be sufficient to impose discipline on a financial system.

In addition to direct losses, as in the Continental case, a more subtle contagion can happen when investors and bankers who have learned of the difficulties of one firm start losing confidence in other firms they believe to be in like circumstances. It is very hard for central banks or other regulators to take the risk of allowing one of these latter firms to fail and then try to contain the systemic effects. The systemic risks, after all, are importantly psychological—that is, a loss of confidence—and it is very difficult to predict the psychology of the market at such a time.

The downside penalty for guessing incorrectly and allowing systemic risk to develop is large and swift. In a financial crisis, with a shrinkage of financial intermediation and a decline in lending, employment and prices can be affected rather quickly and severely. On the other hand, the costs of overprotection are longer term; they are the economic inefficiencies of misallocated resources and of excessive risk in the economy. In general, it is harder to see the cost of leaning too far on the side of preventing risk than it is to see the costs of allowing a bank to fail and having a systemic effect follow. The net effect can be a tendency toward excessive use of the discount window, deposit insurance, and “too big to fail.”

The Federal Reserve has tried to deal with this tendency in several ways. One is through supervision and regulation—higher capital requirements and so forth—so that commercial banks less frequently get into a situation in which central banks need to make such judgments. Another is through exploiting ambiguity in the administration of the lender of last resort function. It is never entirely clear how the Fed will handle the discount window. Who has access? How will it work? Laws and regulations are laid out, but in practice how they will be applied is never spelled out completely. This is one way that the central bank has of minimizing the moral hazard problem. If banks and their creditors are not assured that they are going to be made whole, their concerns, working through the market mechanism, will help to control risk. Note that this is not consistent with classic administration of the discount window. In this tradition, ambiguity is discouraged. Instead, the central bank ought to specify very clearly whom it is going to help and whom it is not, so as to maintain public confidence in the presumably solvent institutions it is going to help.

An increasingly pressing question is whether our tools reach far enough to deal with systemic risk. The scope of central bank powers and responsibilities, at least in the United States, has been focused on commercial banks. However, as was noted earlier, distinctions among banks, investment banks, and other intermediaries and across different kinds of instruments have blurred. Maintaining the focus on commercial banks is raising questions about whether the Fed can contain some systemic risk situations. As other institutions become more like commercial banks, we may have to think about the reach of the discount window.

Investment banks, for example, raise money through highly liquid means, and they make loans, such as bridge loans, that very much resemble commercial bank loans. Commercial banks have lost their uniqueness. Consider money market funds that provide payment services, or consider the volume of payments going through clearing and settlement systems in the securities markets and imagine what would happen if confidence were lost there. Clearly, there are conflicts: to the extent we extend the safety net and access to other central bank services, we increase moral hazard and associated problems.

Pressures can be felt in both directions as to how to respond to these conflicts. Some would like to keep focused even more clearly on commercial banks and concentrate on preventing spillovers in terms of perceived risks or protections between commercial banks and related entities. They would, for example, build “fire walls” between commercial banks and their nonbank affiliates. The other choice is to expand access. Mr. Corrigan, formerly President of the Federal Reserve Bank of New York, for one, has raised the question as to whether others should be granted access to the discount window. And the Securities Industry Association, a trade group and not a disinterested or uninterested observer, has recommended that securities firms get access to the discount window. If access to the safety net or the discount window is expanded, we need mechanisms to assure that participants in the market still have to bear risk. For example, coinsurance for creditors and increased levels of capital for any firm that might be at risk of using the safety net would limit the moral hazard problem of extending that net under a wider variety of firms.

COMMENT

ANTHONY HARRIS

We must be careful to distinguish between failures of monetary policy, which might be defined as systemic supervision, and failure of prudential supervision. The securities crash of 1987 was a failure, if it was a failure at all, of monetary policy. A Bank of England official once put it this way: “My job is to go round finding out who is having a good party. Then I go in and spoil it.” That nicely defines the problems of targeting monetary policy; since 1987 the authorities are much more alert to possible financial bubbles as well as to the aggregates.

Prudential supervision is a more complicated matter, and raises the question of what power is required to make it effective. In today’s world of trillion-dollar daily settlements, and the almost daily invention of new instruments, risk assessment is a matter of higher mathematics, and the task of the risk monitor at the systemic level is virtually impossible. Market participants must monitor their own risks, and the authorities must monitor their fitness and their performance. The Bank of England, for example, is now beginning to insist on management audits to ensure that institutions under its supervision are aware of the risks to which they are exposed, and capable of measuring them—something which can by no means be taken for granted.

But effective supervision requires effective sanctions. For example, the dangers facing the U.S. savings and loans, which failed so massively, were well known to their supervisors; but because of improper political pressure combined with a fear of lawsuits for restraint of trade, no timely action was taken. This suggests to me a simple doctrine of “the greater good,” and a single Draconian sanction that flows from it.

The doctrine really underlies all central banking powers: it states that the public interest in the trustworthiness of the whole financial system—the systemic risk that Mr. Kohn addressed—must be paramount and prevail over individual commercial interests. And the simplest sanction that can make it effective is the British “fit and proper person” test. In essence this is arbitrary in the extreme. It gives the supervisor the power to say, without due process or any possible legal challenge: “You may not be a banker, because I don’t like the way your eyes cross.”

This may seem outrageous, but it is no more than an institutional version of the oldest sanction in the financial markets: the power to refuse credit on nothing more than reputation. You may see it being exercised by the protagonist in a classic British film comedy, Kind Hearts and Coronets, as he goes through a sheaf of bills requiring acceptance, “Yes … yes … yes … on no account,” he mutters. In the old intimate markets, doubtful players were simply frozen out. In today’s international markets, few participants have the knowledge to fill this function; the supervisor, with his advantage in information, can, but he is unlikely to do so effectively.

How widely should such powers apply? To banks and quasi-banks, clearly; they have the unique privilege that their liquid debts are regarded by their creditors as ready money. If you lend to a friend, you know the risk, but it is unreasonable to expect retail depositors to form a view on the quality of lending done by a depositary institution. Supervision protects the innocent, and since there are also innocent buyers of insurance, mutual funds, and other financial products, the frontiers of supervision are always creeping outward.

The protection of professionals is another question. Retail depositors can be protected by deposit insurance, though this is not problem-free. But where insurance is provided, the commercial banks that finance it have the right to expect that the industry is effectively policed. Draconian central powers are in the interests of the banks so supervised. But it is here, rather than in the retail markets, that the question of moral hazard may arise. Will faith in central banks’ powers lead market participants into slack risk assessment or questionable lending?

This hazard, I believe, can easily be overstated. We are now in the world of payments systems, a world in which the loss of two or three hours’ of intraday interest could be serious enough to trigger a lawsuit. This is the world of highly geared intermediaries striving to keep their exposures matched, and for these enterprises, risk is measured on a totally different scale. Outside users of the system worry only about failure, but within the interbank market, any departure from smooth functioning is expensive.

The loss of interest when, say, high-coupon CDs are replaced by official bills carrying Treasury rates, as in the S&L bailouts; the loss of liquidity while a weak institution is kept on temporary life support awaiting a determination; these are serious commercial hazards that may arise from the process of effective supervision. Market insiders cannot, then, leave risk assessment to the authorities; for their own protection, they must try to anticipate possible official trouble.

So supposed moral hazard should not inhibit the authorities from responding actively to a market in which risk appears in some new disguise almost every day. Prudential supervision is not a permanent formula to solve an unchanging problem. It must be a flexible, pragmatic response to an evolving panorama of market activity, new products, and new services.