The International Monetary Fund (IMF) and the World Bank, created in 1944, reflected the experience of the 1920s and 1930s. The Funds tasks were to adjust current-account imbalances and manage the exchange-rate system. The Banks main tasks were to lend for the reconstruction of Europe and to eliminate the alleged bias against lending to developing countries.
Whatever conditions might have been in the 1940s, the international financial system has found other means of solving the problems that the Fund and the Bank were supposed to solve. Changes in exchange rates are now one common means of adjusting current-account imbalances. Leading countries, including the United States, Japan, Britain, and the European Union, allow their currencies to float. Western Europe now has a common currency and a single central bank in place of fixed but adjustable exchange rates.
Many of the problems or international financial crises of recent years arose because there is too much lending, especially short-term lending, to developing countries, not too little. Recent history gives strong support to the proposition that if a country adopts market-oriented policies of privatization and deregulation, opens its trade to competition in foreign markets and by foreigners in domestic markets, and carefully controls its budget, foreign lenders and investors are willing to finance its development.
Yet the international financial system is crisis-prone. Latin America in the 1980s, Mexico in the mid-1990s, and Asia and Russia most recently present well-known examples of deep, pervasive financial crises that have been costly to the public in the countries with financial problems, to their trading partners, and, often, to much the rest of the world. The past fifteen years have seen ninety serious banking crises, most of them followed by deep recessions. More than twenty of these crises produced direct losses to a developing country exceeding 10 percent of its GDP. In half of these cases, including several Asian countries now, losses exceed 25 percent of GDP (Caprio and Klingabiel 1996, 1997; Calomiris 1998). These losses, relative to GDP, are far larger than the cost of the U.S. savings-and-loan problem to U.S. taxpayers.
The frequency and severity of recent international financial problems, and their occurrence in a period of growth, economic progress, and low inflation should raise a number of questions. Why is there so much financial fragility? Are current international institutions appropriate for current conditions? Have international financial arrangements and institutions adapted appropriately to changes in the world economy? Is it time to agree on new arrangements? What international financial arrangements would serve the world well in coming decades?
I do not pretend to have complete answers to all of these questions, but I believe that economists and policy makers must ask and try to answer them. In this article I attempt to contribute to that discussion. After pointing out some of the failures of current arrangements and the incapacity of those arrangements to resolve current problems, I propose some changes. I concentrate on the IMF, although there is now substantial convergence in the activities of the Bank and the Fund. I omit discussion of policy errors, for example, in Asia in 1997. All policy makers err at times. The important issue is not errors of judgment but structural flaws that exacerbate financial fragility.
Origins and Rationale of the Fund and the Bank
Planning for postwar international monetary cooperation began before the United States entered World War II. The lend-lease agreement, under which Britain and other nations obtained military supplies and equipment on “credit,” provided that the United States could waive postwar repayment if the British agreed to eliminate trade “discrimination.” The term was not further defined, but the objective it expressed included elimination of the prewar system of imperial preference that bound its empire to Britain and favored British exports to its colonies.
As negotiations of postwar arrangements proceeded, trade issues faded into the background, and financial issues moved to center stage. By September 1941, John Maynard Keynes had developed a proposal for an international currency union as part of the British contribution to discussion of postwar arrangements. With minor adjustments, Keyness proposal became the British governments proposal in April 1943, when formal bilateral discussions began (Meltzer 1988, 23637).
Keynes visited the United States in the fall of 1941 and may have discussed his plan informally. In December, a week after the United States entered World War II, Treasury Secretary Henry Morgenthau asked Harry Dexter White to “prepare a memorandum on the establishment of an inter-Allied stabilization fund” as the basis for postwar international monetary arrangement (Blum 1967, 3: 22829). Morgenthaus diary suggests that he had a vague idea about expanding the 1936 Tripartite Pact to avoid competitive devaluations.
Keynes ( 1971) had analyzed the basic problem. Each country acting alone can achieve either stable prices or a fixed exchange rate, but not both. To achieve both requires international cooperation or agreement. The classical gold standard was one such agreement. Keynes, among others, had recognized earlier that the gold standard required deficit countries to bear the cost of adjustment, required procyclical monetary policies, and was inflexible and costly for a country with down-ward wage rigidity. Like White and many others, he considered the inflexibility of the gold standard, the distribution of gold, and the monetary policies of the surplus countries, mainly France and the United States, to have been leading causes of the Great Depression. The aim was to avoid a return to the classical gold standard while retaining enough of its features to solve the coordination problem and while making the arrangement acceptable to prospective surplus and deficit countries.
Both Keyness and Whites plans and the 1944 Bretton Woods agreement imposed costs on surplus countries that would neither expand their imports nor lend to deficit countries. The justification was elimination of an externality. By forcing adjustment on surplus countries, this approach would spare deficit countries the costs of higher interest rates and contraction. This policy would also benefit surplus countries and others, because their exports and incomes would be maintained. If the surplus countries could be made to lend to the deficit countries, fluctuations in economic activity would be damped. Both surplus and deficit countries would gain. Of course, this policy could work only if the imbalance was temporary. Persistent imbalances would require a change in exchange rates, with the consent of the IMF.
Both Keynes and White limited their proposals for the IMF to the financing of current-account deficits. To the extent that the plans discussed financing of capital flows, it was the proposed Bank for Reconstruction and Development (later the World Bank) that would be responsible. White explained privately why the United States favored an international bank. “Many of the loans will be risky and there will be some losses. This is one of the reasons why we insisted that the Bank be an international bank rather than to take the risks ourselves. We felt that the benefits would be worldwide and that other countries should be at risk” (H. D. White to the Board of Governors and Reserve Bank Presidents, Board Minutes, March 2, 1945, 17).
There were two principal arguments for the proposed bank. One was risk sharing, a rationale that continues to be advanced. This is a distributive, not an efficiency argument. The second was the anticipated benefit to the world economy. This claim reflected the belief that economic development was hindered by risk-averse lenders, who restricted the supply of capital and charged premium rates. The supply of capital to developing countries was believed to be suboptimal. The proposed bank would remove the capital-market restriction. Although this second argument was not developed more carefully, it was widely accepted. The experience of the 1930s, when there were sizable defaults by developing countries, was taken as evidence that lenders would be restrictive in the future.
The international system did not evolve as planned. Experience proved that many of the beliefs and conjectures on which the plan had been based were wrong. The United States did not return to its interwar policies. Instead of running large, persistent surpluses and maintaining high tariffs, it ran small surpluses or deficits on current account and worked to reduce tariffs globally. Loans and transfers added to the dollar claims held by foreigners. By 1951, the U.S. gold stock had started a fall that, with a few brief interruptions, continued until the United States suspended gold sales in August 1971.
During the Funds early years, when the United States had its largest current-account surpluses, the Fund had a very modest role in the international payments system. The Marshall Plan redistributed part of the U.S. surplus, and the European Payments Union cleared payments imbalances for the inconvertible European currencies. Historians describe the Funds early years as a period in which its status and even its survival were in question (Presnell 1997, 229; James 1996).
The Fund became more active in the mid-1950s. For a few years, the new system functioned smoothly. Membership increased; more countries drew on the Funds resources, and the amounts drawn increased. The Fund specialized in making relatively small, short-term loans. Because repayment was usually prompt, the Fund recycled its resources (James 1996).
Then came a series of crises or disturbances affecting key currencies, first the pound, then the dollar. By 1968 the U.S. gold stock had fallen so far that a de facto restriction on gold convertibility was in place. The fixed-but-adjustable rate system limped through the next two years. In mid-1971, convertibility ended formally with the closing of the gold window.
If the original plan had been implemented, the IMF would have ceased to exist in 1973, when the fixed-but-adjustable rate system officially ended. The Bretton Woods system left capital-account lending to the World Bank, so the Fund no longer had a reason for being.
In the 1970s and 1980s, the Fund played an important role in capital-account lending to “recycle” the revenues of oil-producing countries after the oil price rise. Its role as advisor to developing countries on macroeconomic adjustment increased. That advice was tied to lending; the Fund made loans to ease the burden of adjustment to the structural and macroeconomic reforms it advocated.
Sebastian Edwards (1989) summarizes the results of an internal IMF study by Mohsin Khan of these policies and actions. “Khan found that Fund programs had a statistically nonsignificant effect on the balance of payments, a nonsignificant effect on inflation, a significant positive effect on the current account, and a significantly negative effect on output” (78). In other words, the Funds programs moved the current account toward balance by reducing output, income, and imports. The conditions for IMF assistance required countries to control inflation, but that condition was not met in many countries. IMF research found no significant effect of its pro-grams on inflation.
Beginning in the summer of 1982, the Fund greatly expanded lending to countries experiencing capital outflow. It continued to lend during the years of financial distress known as the “Latin American debt crisis.” That lending was a major extension of its capital-market program. The size of loans and the time to repayment increased. Many countries became permanently indebted to the Fund.
The Funds main function in the 1980s was lending money to debtor countries so they could pay interest on their outstanding debts to the Fund and commercial banks. As part of those arrangements, commercial banks also increased their loans to indebted countries. Most of this lending was an accounting operation to avoid recognizing defaults. The result was that Latin American debt rose, the creditor banks were protected, and the debtor countries continued to suffer under a rising debt burden and falling income per capita.
IMF programs postponed recognition of losses by U.S. and other international banks at the expense of living standards in Latin America. By the end of the decade, as part of so-called Brady adjustments, banks accepted part of their losses, workout agreements were reached for the remaining debt, and growth resumed in Latin America.
Latin American loans were made under the IMFs conditional lending program, under which governments agreed to follow stabilizing policies. As before, the problem was enforcement against sovereign governments. The IMFs main threats were to cancel the loan program or to withhold payments. In a preview of what was to happen in Russia and Asia, the threats usually proved empty. The main reasons are of interest because they reveal some principal flaws of the system.
First, the IMF becomes committed to the “success” of the program. Cancellation is seen as failure. Governments understand that the IMF is reluctant to withhold funds or to cancel programs. Hence its threats lose force.
Second, IMF officials are judged partly on their contacts with high officials of borrowing governments. Critical reports by an IMF task force dampen the welcome the IMF team can expect on its next visit. The finance minister is “too busy.” Not meeting with principal officials is treated as “failure” harmful to the IMF officials career. Borrowing governments recognize this leverage, so they can keep criticism in check and prevent or delay information from reaching the IMFs top management.
The Latin American program became a model for subsequent capital-market programs. The next large financial crisis occurred in Mexico in 199495. Again the IMF, with assistance from the U.S. Treasury, protected foreign banks and financial institutions by allowing them to avoid portfolio losses. Although many of the foreign commercial banks had made loans in 1994 at interest rates of 20 percent per annum or more, they were not required to bear the risk that they had assumed. Instead the IMF lent money to Mexico at below-market interest rates. Again the international bankers were spared, but the Mexican economy suffered a severe recession.
The IMF and the U.S. Treasury claim success in Mexico. A principal reason, I suspect, is that the loans by the IMF and the U.S. Treasury were repaid. Much of the repayment was effected by the Mexican governments borrowing in capital markets at higher rates of interest, surely not an improvement from the standpoint of Mexican economic welfare.
The welfare losses to Mexicans were much larger than the increased interest payments. In U.S. dollars, Mexicos real GDP per capita in 1997 was only slightly higher than in 1973. In the interim, a period of rising world real incomes, Mexican income was highly variable in the short term but stagnant in the longer term. Meanwhile, debt per capita rose by a factor of three or four, so, after twenty-five years, with income unchanged on average, taxes for debt service were much higher and Mexican disposable income per capita was smaller.
The IMF, the World Bank, and the U.S. Treasury are not responsible for all of Mexicos problems. Oil price changes and mistaken policies of the Mexican government made large contributions. To the extent that IMF support contributed to the durability of mistaken policies and to the burden of the debt, the IMF must bear some responsibility. Mexico is far from the success that the IMF and U.S. Treasury officials proclaim. The IMF did not foresee the Mexican crisis partly because the Mexican government did not provide information, partly because of the incentives within the IMF, discussed earlier.
One promising feature of the Mexican program represented a new departure on which better arrangements can be built. Mexico gave a lien against Mexican oil revenues to guarantee repayment of its U.S. government loan. Use of collateral to support borrowing is a helpful step, but it raises the question of why oil revenues could not have been used to guarantee private lending to mitigate the need for subsidies from the IMF.
A bad feature of the Mexican program was its contribution to the belief that international banks enjoy a safety net not available to investors in equities or to purchasers of real assets in foreign countries. The message implicit in these actions was clear to bankers and investors. Between 1990 and 1996, capital flows to emerging markets rose from $60 billion to $194 billion. After 1995, with the Mexican experience in mind, portfolio investment declined but bank lending increased.
In the past twenty years, the IMF has introduced two new loan facilities. One is the Extended Structural Adjustment Facility (ESAF); the other is the Structural Transformation Facility (STF) to assist former members of the communist bloc, including Russia and states of the former Soviet Union. Both facilities extend the conditional lending procedures in new directions without correcting the flaws.
ESAF offers medium-term loans at an interest rate of 0.5 percent with repayments up to ten years. The borrower agrees to make structural adjustments such as fiscal reform, privatization, and trade liberalization. “A recent IMFWorld Bank study concludes that the results of ESAF loans have been largely negative in terms of reducing budget deficits and inflation and mixed in terms of producing external viability and promoting per capita growth” (Mikesell 1998, 31).
The STF has contributed to transformation in several former communist countries. The bulk of its fundsand other IMFWorld Bank loanshas gone to Russia, where policy mistakes, misjudgments, and corruption have prevented successful transformation. When oil and commodity prices fell in 1998, lenders began to question Russias ability to service its large outstanding debt.
A principal policy mistake in Russia is to neglect the necessary conditions for transformation to a capitalist economy. Those conditions include structural reforms that increase the transparency of business and government reports and statistics and that establish private property, a commercial code, accounting standards, and enforcement through the rule of law. Some commercial and industrial property has been privatized (often by deplorable methods), but 90 percent of agricultural land remained under state or collective controls in 1997, and federal law does not permit the sale of agricultural land. In contrast, China began its more successful reform program by introducing long-term leases that incorporated many features of private ownership of agricultural land.
The structural reforms mentioned, if adopted, would have provided a Hayekian infrastructure, creating incentives for decision-makers to allocate resources efficiently. Capitalism is more than trade at market prices. Successful capitalism requires institutions that provide incentives compatible with economic development, efficient use of resources, and enforcement of contracts. These features of successful capitalist countries are missing in Russia and several countries of the former Soviet Union. The IMF has shown little interest in encouraging appropriate institutional reforms.
Corruption and mismanagement are widespread in Russia. Estimates published in 1997 suggested that capital flight was about $2 billion per month, altogether some $150 billion from Russia since the breakup of the USSR (Jenkins 1998, quoting Global Finance). A report by the director of the Chamber of Accounts of the Russian Federation gives some specific examples: (1) Parliament appropriated $150 million to build planes for sale to India. An audit showed that none of the funding reached the enterprise. (2) Parliament appropriated funds to aid Chechnya after the war ended. The total bill was $3 billion. The audit found documentation for $2 billion; less than $150 million reached Chechnya. No record of the remainder was found. (3) Thirty million dollars of a World Bank loan was allocated to compensate victims of bank frauds and pyramid schemes. Audits showed that after several years not a single victim had received payment. (4) The government issued debt to finance a large fraction of its payments. Forty-five percent of state revenues in 1998 went to pay interest on the debt. “This means no money is left to pay workers or to support education, public health,” or other government services (Sokolov 1998).
These criticisms were made before recent crises drove interest rates above a 100 percent annual rate. The IMF did not publicly condemn these improper and corrupt practices. We now know that the IMF was aware of the problems of inadequate accountability and corruption, yet it failed to enforce the conditions of its own loans or the most elementary standards of accountability and performance.
IMF and U.S. and German government policies encouraged banks and financial institutions to believe that a Russian default on its debt would be avoided. Actions by the IMF and the United States had protected lenders from taking losses in Mexico. If Mexico was critical to the stability of the United States, the reasoning went, surely Russia was more critical. On this reasoning, interest rates of 40 percent, 50 percent, or 100 percent seemed to be a gift to international lendersdefault premiums without comparable default risk.
Moral hazard arises when the private risk to the lender is less than the risk borne by society. This situation was clearly the case, ex ante and ex post, in Mexico and in Asia. Banks and other lenders did not experience the large losses borne by ordinary citizens and by the owners of equity and real assets. Ex ante, banks and financial institutions that made loans to Russia believed that they would be spared losses also. The $22 billion loan promised by the IMF, the World Bank, and others in July 1998 seemed to confirm that view for a time. But the inability of the Russian government to fulfill its commitments soon raised doubts, leading to increased capital outflow.
Moral-hazard lending to Russia, encouraged by the bailout of foreign lenders to Mexico, permitted Russia to finance large unbalanced budgets by borrowing externally. The result is a much larger financial problem for international lenders and for the economies of other countries. The IMF continued lending despite the Russian governments failure to meet the loan conditions. The IMF continued to lend when the Russian government “reduced its deficit” by not paying soldiers, miners, and others. Again, the IMF was committed to “success.” The government understood the nature of that commitment, so “conditionality” failed.
The IMFs mistake was to establish the STF and undertake structural reform of the Russian economy. It had no prior experience and no special competence. The lessons it had learned, mainly in Latin America, concerned macroeconomic adjustment of market economies. The Fund was slow to recognize that structural transformation must involve the development of a Hayekian infrastructure. And it refused to learn a central lesson of its own past experience: sovereign governments cannot be compelled to implement programs that they do not favor. Policy changes will not be implemented unless they are supported by local political institutions and their leaders.
International Monetary Fund Errors and Responses
The IMF had more success in the 1960s, when it limited its efforts to helping countries with current-account deficits. As the Funds scope expanded, its record became ambiguous or worse. Errors or problems pertained to structural-transformation lending and advice to Russia; moral hazard; and the ambiguous effects of IMF (and World Bank) lending. Consideration of these errors and the IMFs responses to them will point us toward worthwhile reforms.
There was no previous experience on which to base transformation policies in Russia. Primarily for political and military reasons, European and U.S. governments hoped to transform Russia into a more democratic society with a market economy. The G-7 governments either were unwilling or believed themselves unable to obtain funding for the transformation from their parliaments. The IMF agreed to accept responsibility. In doing so, it reached far beyond its competence.
Even if its advice had been well founded, the IMF had little scope for enforcing Russian commitments. Like most borrowers, Russian officials understood that the IMF and the G-7 had a stake in reform and transformation. The Fund could threaten to withhold payments. It did, on occasion, delay payments. As in many previous cases, however, the Fund did not want the program to fail. Governments in Russia (and other countries) understand that the Funds commitment to “success” weakens its ability to enforce threats. The IMF was unwilling to call attention to widespread corruption, failure to implement reforms, and cynical maneuvers to reduce reported budget deficits by failing to pay civil servants, coal miners, soldiers, and others.
Russia differed from other countries in many ways, but the IMFs failure in Russia was not unique. The IMF also tolerated corruption in Indonesia. In Korea, Indonesia, and elsewhere, it tolerated continuation of fragile financial systems used to subsidize projects favored by government officials or their political supporters. A main question about these and other failures is whether IMF lending delayed reform both directly by lending and indirectly by encouraging private capital inflows. The additional resources may have contributed to reform, but they also permitted bad policies to continue.
No single answer can be given for all countries. In some countries, IMF loans may have helped reformist politicians to make changes that otherwise would have been delayed or avoided. It seems clear, however, that IMF lending, and the private capital flows that followed, permitted unbalanced budgets, fragile financial systems, government subsidies to specific programs through banking systems, and corruption to continue longer and at higher levels.
Much has been written about moral hazard in Asian lending. I want to distinguish two separate problems. The first is the effect on private lenders of IMF policies toward Mexico and Russia. The second is the decision by governments to delay reform, knowing that IMF loans at low interest rates will be available in a crisis.
IMF officials minimize the role of moral hazard in Asia. A typical statement is that no government pursues policies that lead to the loss of output and employment and to the lower living standards experienced by Mexico, Thailand, Indonesia, Korea, and others. Stanley Fischer (1998) calls “far-fetched” the idea that policy makers will take excessive risks because the IMF stands ready to help.
This defense shows little understanding of how the financial problem develops. Ministers of finance do not set out to generate a crisis. At each critical decision point, however, they can decide to allow an additional increase in short-term foreign borrowing rather than to adopt policies that would avoid the crisis. An example common to Mexico, Korea, and other countries is the decision to offer exchange guarantees to foreign lenders who are reluctant to renew loans. The guarantees may sometimes be followed by stabilizing policies, but in many countries they are not. Rather, their effect is to postpone and enlarge the subsequent crisis.
Such behavior by policy makers is understandable. The crisis is not a “sure thing.” As always, there are risks both ways. A government that adopts restrictive policies early runs the risk that parliament will not approve, that the opposition will claim the policies were unnecessary, and that voters will support the opposition.
This pattern of behavior is not unique to developing countries, and it does not occur only in countries that borrow from the IMF. The U.S. government delayed responding to the savings-and-loan problem for many years. The arguments were similar at the time. Japan delayed a solution to its banking problems. The finance minister who chooses delay may be proved right. Even if he is wrong, a crisis may not occur during his term of office. IMF loans, at subsidized rates, are available. A timely loan may require some retrenchment, but not all countries that go to the IMF have a crisis.
It is sufficient for moral hazard that the existence of subsidized loans from the IMF modify the finance ministers evaluation of the costs he faces. The large increase in the number of countries experiencing large crises in recent years suggests that a change of this kind has occurred. Perhaps the severity of the crises in Indonesia, Thailand, Korea, and Russia will change future behavior. But even if so, institutional re-form is still desirable.
A more problematic defense of IMF procedures compares the IMFs rescue pack-ages for international banks to the rescue of some of the passengers on the Titanic. The comparison is inapt. There is no important difference between individual and social losses when a ship such as the Titanic sinks. There was no learning by other ships or their captains as a result of the rescue. In the Mexican and Asian crises, however, there are large differences between the losses borne by international banks and the losses to the Mexican, Thai, Indonesian, Korean, and other populations. Bankers learned that the IMF and the U.S. Treasury would lend to reduce losses to large banks and financial institutions.
IMF and U.S. Treasury loans to Mexico permitted the Mexican government to honor most of the exchange guarantees on the dollar-guaranteed bonds called tesobonos. By subsequently asking member governments for a large quota increase, the IMF strengthened the belief that similar commitments by governments in Asia, Russia, and elsewhere would be honored. If the IMF was not planning large future rescues of banks and lenders, why was a large increase in funding, about $80 billion, needed?
The IMF is not responsible for all the errors that governments make. Nor is it responsible for all the errors made by lenders and borrowers. At issue is whether IMF lending increases the risks in the international financial system. By pooh-poohing the moral-hazard problem, the IMF avoids recognizing that it is part of the problem. Its behavior promotes too much short-term lending by financial institutions and too few losses on risky loans.
The third problem is that the effects of conditional lending are ambiguous. Lending may encourage reform, for example, by reducing transition costs and strengthening the position of reformers. But lending also may delay reform by permitting governments to continue inappropriate policies. The IMF lacks adequate mechanisms for enforcing desirable change and avoiding retrograde actions.
IMF officials have not proposed effective programs for strengthening the international financial system. Some hint at the possibility of restricting short-term capital movements (Fischer 1998, 7) Other suggestions include improved supervision or more timely information. These suggestions may be helpful, but they are insufficient. Studies from Benston (1973) to Berger, Davies, and Flannery (1998) show that banking supervisors and regulators rarely foresee failures. With modern financial instruments, a banker can change portfolio risk as soon as the examiner leaves the bank. Moreover, better alternatives exist. Based on experience with large-scale failures and moral hazard, some countries have reduced their reliance on supervision and regulation. Recent practice in New Zealand, Chile, the United States, and elsewhere now relies more on bank capital and market-based incentives to increase safety and soundness.
Suggestions for Reform
The IMF was created to assist in the adjustment of current-account imbalances in a world with fixed exchange rates and widespread capital-account restrictions. The World Bank was given responsibility for capital transfers on the presumption that government programs were needed to compensate for a suboptimal volume of development lending.
The conditions under which these institutions were founded no longer exist. It is time to develop the institutions and arrangements that will be useful for current and expected future conditions. The current system has structural flaws, in part because it developed as a response to specific problems and without attention to the longer-term effects of changes in the world economy. To encourage discussion of these issues, I offer some tentative proposals to improve the functioning of the international financial system by increasing efficiency and reducing moral hazard. My aim is to open discussion of arrangements that maintain capital flows and international lending while reducing the frequency and depth of financial panics and the losses to countries and their citizens. The proposed system is designed to reduce moral hazard and avoid both subsidies to lending and regulatory restrictions or taxes that prohibit types of lending. There are seven proposals.
- Increased reliance on fluctuating exchange rates. Many of the exchange rates may be “dirty floats,” but experience has shown that capital mobility, fixed exchange rates, price stability, and full employment are rarely compatible. Kenneth Rogoff (1998) shows that in recent years few countries have been able to maintain a fixed exchange rate for six years or longer. Floating exchange rates shift costs to lenders who withdraw their capital and raise the price borrowers pay for large capital inflows. Hence, they reduce inflows and outflows.
- Improved management of capital flows. Central banks can manage their exchange- rate systems to maintain price or economic stability. If a capital inflow threatens stability, the central bank can (a) permit the exchange rate to appreciate, (b) sterilize the inflow by selling domestic securities, or (c) do some of each. Similarly, the central bank can offset capital outflow by allowing the exchange rate to depreciate or by buying domestic assets.
- Increased reliance on competition in local banking markets. Many producers of goods or services diversify risk in their international operations by locating facilities in many countries. The Coca-Cola Company has expanded internationally since the 1930s. It has prospered despite wars, inflation, and many local or regional crises. The company has learned to manage the risk inherent in an internationally diversified business and has provided a model that others have followed. Many countries do not permit banks to follow this model. They prevent international banks from competing in local markets. International banks are limited to making mainly dollar (or yen or deutsche-mark) loans to local banks. If international banks held portfolios of local assets and issued local currency liabilities, country risks would be part of a diversified international portfolio. Diversification would reduce the cost of bearing risk. Competition would work to improve local banking practices.
- Increased size and diversification. Many countries are too small to achieve optimal diversification of financial assets. Loans to a few industries dominate portfolios of local banks. Suboptimal diversification has been a principal cause of bank failures throughout U.S. history. The admission of foreign banks to local markets should be accompanied by rules for bank capital that prevent international banks from leaving in a crisis. There are different ways to sustain commitment to the country. All involve greater losses from withdrawing than from remaining, hence capital or asset holdings denominated in local currency.
- Establishment of an international quasi-lender of last resort. More than a century ago, Walter Bagehot ( 1962) explained that a financial system requires a lender of last resort to assist financial institutions in a liquidity crisis. Unlike the IMF and many countries central bankers, Bagehot distinguished between liquidity and solvency and provided rules that separated the two. He required the borrower to offer marketable assets as collateral for a loan, and he required the lender to charge a penalty rate on all such loans. The collateral requirement separates insolvent from illiquid banks. The penalty rate eliminates subsidies, reduces moral hazard, and reduces reliance on the lender. Most of the time the lender of last resort would be idle. Markets would function, and borrowers would offer collateral. Hence prospective borrowers would hold such collateral; otherwise they could not get assistance.
- Enforcement of the collateral requirement. This would have stabilizing dynamic properties. Central banks could borrow from the quasi-international lender of last resort only on the presentation of internationally traded assets, so they would be induced to hold such assets. They would lend to domestically chartered banks in the event of bank runs. The international lender of last resort would be barred by statute from making loans without receiving marketable collateral (at a price below last market price). A foreign government that wished to circumvent collateral requirements to assist a developing country would have to obtain an appropriation through its legislature.
- Development lending through capital markets. Experience has shown that capital comes to a country that opens its markets, controls spending and budget deficits, reduces inflation, and deregulates and privatizes. International financial institutions are no longer needed for development lending. A modest role for redistributive lending to reduce poverty would remain.
The combined effect of fluctuating exchange rates, diversified international banks, capital requirements, and a Bagehotian lender would reduce reliance on short-term capital flows and mitigate the crises that occur when several international lenders simultaneously fail to renew their loans. A Bagehotian lender of last resort reduces moral hazard.
At the organizational meeting of the IMF and World Bank, Keynes argued against locating the institutions in Washington. He worried that they would be overly influenced by U.S. domestic politics and the pressures generated by domestic interest groups. It is not known whether such pressures would be reduced by locating the lender of last resort outside the United States. That subject merits more study.
Finally, what should be done about assistance to Russia? The question is primarily a political one, and it should be treated as such. Parliaments should be asked to appropriate transfers, perhaps in exchange for warheads and missiles, as was done in Ukraine. There is a possible collective benefit that has little relation to international development lending. Confounding the two issues was a mistake that is now apparent to all.
Bagehot, Walter.  1962. Lombard Street. Reprint, Homewood, Ill.: Richard D. Irwin.
Benston, George J. 1973. Bank Examination. Bulletin of the Institute of Finance. New York: Graduate School of Business Administration, New York University (May).
Berger, Allen, Sally Davies, and Mark J. Flannery. 1998. Comparing Market and Supervising Assessments of Bank Performance: Who Knows What When? Board of Governors, Federal Reserve System, Finance and Discussion Series, 1998-32.
Blum, John M. 1967. From the Morgenthau Diaries: Years of War. Boston: Houghton Mifflin.
Calomiris, Charles. 1998. The IMFs Imprudent Role as Lender of Last Resort. Cato Journal 17: 27594.
Cavallo, Domingo. 1998. Good Information, Bad Advice. Forbes Global, Internet, March 31.
Edwards, Sebastian. 1989. What Does the International Monetary Fund Really Do? Reply to Goldstein. Carnegie-Rochester Conference Series on Public Policy 31: 7782.
Fischer, Stanley. 1998. The Asian Crisis Countries and International System Improvements. Berlin 1998. New York: Trilateral Commission, pp. 37.
James, Harold. 1996. International Monetary Cooperation since Bretton Woods. Washington, D.C.: International Monetary Fund.
Jenkins, Holman W., Jr. 1998. Russia a Deadbeat? Funny the Banks Never Mentioned It. Wall Street Journal, September 2, p. A19.
Keynes, John Maynard.  1971. A Tract on Monetary Reform. Volume 4 of The Collected Writings of John Maynard Keynes. London: Macmillan and St. Martins Press for the Royal Economic Society.
Meltzer, Allan H. 1988. Keyness Monetary Theory: A Different Interpretation. Cambridge: Cambridge University Press.
. 1995. Sustaining Safety and Soundness: Supervision, Regulation, and Financial Re-form. Washington, D.C.: The World Bank.
. 1998a. Asian Problems and the IMF. Cato Journal 17: 26774.
. 1998b. Financial Structure, Saving, and Growth: Safety Nets, Regulation, and Risk Reduction in Global Financial Markets. Paper presented at First International Conference, Bank of Korea, June.
Mikesell, Raymond F. 1998. The Future of the Bretton Woods Institutions. Jobs and Capital 7: 2733.
Presnell, L. S. 1997. What Went Wrong? The Evolution of the IMF 19411961. Banco Nazionale del Lavoro Quarterly Review 50: 21337.
Rogoff, Kenneth S. 1998. The Risks of Unilateral Exchange Rate Pegs. Paper presented at the First International Conference, Bank of Korea, June.
Sokolov, Venyamin. 1998. The Virus in Russia. New York Times, June 1.
Acknowledgment: This article was originally published in The Asian Financial Crisis: Origins, Implications, and Solutions, edited by William C. Hunter, George G. Kaufman, and Thomas H. Krueger, and published by Kluwer Academic Publishers in May, 1999. Kluwer has granted the right to reprint the chapter in this issue of The Independent Review.
Allan H. Meltzer is the Allan H. Meltzer University Professor of Political Economy at Carnegie Mellon University and a visiting scholar at the American Enterprise Institute.