Information about Western Hemisphere Hemisferio Occidental

6 What Future for the International Monetary System?

Jacob Frenkel, and Morris Goldstein
Published Date:
September 1991
  • ShareShare
Information about Western Hemisphere Hemisferio Occidental
Show Summary Details
Richard N. Cooper

This paper discusses some problems with present international arrangements and suggests that the time has come to begin contemplating a common currency among the industrialized democracies. This idea is much too radical to garner much political support in the near future. But it does offer a way to overcome a number of difficulties that these economies now face, ([a-z]+) which are likely to become more, not less, serious in the coming years. So despite its novelty, a common currency provides a focal point for analysis, which may in turn suggest intermediate steps that accomplish some of the same results with less political commitment.

Before we turn to the present and the future, it is worth considering past international monetary arrangements and how they performed. The first section of the paper, therefore, reviews briefly six types of international monetary arrangements that have existed during the past century. It then turns to an evaluation of the strengths and weaknesses of the current arrangement of floating exchange rates subject to ad hoc management. Throughout, the focus will be on the arrangements among the major economies of the day, basically Europe, North America, and in the last 30 years, Japan. During the earlier period, these arrangements encompassed many “peripheral” areas as well, by virtue of their colonial status. How peripheral countries relate to the international monetary core is an interesting and important topic, but it is necessary first to be clear on relationships at the core.

I. Past International Monetary Systems

History is complex, and it would take us too far afield to enter into a detailed exposition of the workings of different international monetary systems. But every system must address two fundamental questions, and it is worthwhile to sketch how different systems addressed these questions, both in theory and in the way the theory was modified in practice. The two fundamental questions concern how the international monetary system envisions that countries will adjust to “disturbances” to economic relations among countries (the adjustment problem) and how it envisions that adequate, internationally acceptable means of payment will be provided (the liquidity problem).

The history of the past century can be divided roughly into six periods, excluding the two world wars and their immediate aftermath, each of which involved a somewhat different international monetary system covering the major countries. Of course, this division inevitably involves some stylization, since history evolves continuously; in some cases the shift from one system to another is gradual and involves antecedents, and not all countries move together.

The six systems are the gold-specie standard (1879–1913); the gold exchange standard (1925–31); freely floating exchange rates (1919–25, 1933–36); the Bretton Woods system, early phase (1947–59); the Bretton Woods system, late phase (1959–73); and managed flexible exchange rates (1973-present).

Gold-Specie Standard

Under the gold-specie standard, national money and acceptable means of payment were the same, namely, gold coins of standard weight and fineness. These might be re-minted into different national coins, although foreign coins sometimes also circulated domestically. Thus both domestic and international liquidity were satisfied in principle from new gold production. If this proved inadequate, as it did for prolonged periods, most notably from the 1870s to the mid-1890s, the increased liquidity was supplemented domestically by a rapid expansion of new forms of payment, notably banknotes and demand deposits in banks (Triffin (1964)) and by a prolonged decline in the price level, which raised the real value (in commodities) of a given volume of gold money. Wholesale prices fell by 40 percent to 55 percent in all the major countries between 1873 and 1896 (Cooper (1982)).

The theory of the adjustment mechanism under the gold standard was simple, and it was clearly and concisely stated by David Hume in 1752 (Cooper (1969)). A transfer (for example, reparations, or a capital investment) from country A to country B raised the money stock in B and lowered it in A. As a result, prices would fall in A and rise in B, and A would in consequence enlarge its trade surplus (or reduce its deficit), make the real transfer in goods, and earn the gold back until monetary and payments equilibrium was restored.

How the adjustment mechanism worked in practice is still controversial. First, much less gold actually moved internationally than the theory would suggest (Bloomfield (1959)). Second, Viner’s (1924) classic study of large capital inflows into Canada before World War I suggested that although price movements could be discerned, especially of nontradable goods against tradable goods, the price movements were far smaller and worked to accomplish changes in trade flows more rapidly, than he and his teacher Taussig expected. Later analysis suggested that a substantial part of the adjustment was accomplished by changes in spending—up in the receiving country, down in the sending country—associated with the transfer itself. Recently, McKinnon (1988a) has argued that there is no presumption one way or the other about which way the sending country’s terms of trade will change, and with a fully integrated capital market, there need not even be any change in domestic (that is, nontradable) prices. The key point is that adjustment is accomplished partly through changes in spending, or absorption, and partly through changes in prices, especially the prices of nontradables.

Gold Exchange Standard

The adjustment mechanism under the gold exchange is similar to that under the gold standard. In this case, the downward pressure on prices in the sending country is mediated by the banking system, which operates on fractional gold reserves. In fact, the period of the gold exchange standard was dominated by economic stagnation in the leading country, the United Kingdom, but that was related to the exchange rate at which the United Kingdom returned to gold convertibility rather than to the nature of the standard itself. Its period was too short to discover if adjustment eventually would have taken place.

The novelty of the gold exchange standard was to conserve gold by removing it from circulation and concentrating it in the hands of the leading banks, even going beyond that by encouraging banks, ([a-z]+) smaller central banks, to hold short-term claims on other countries, mostly in sterling, secondarily in dollars. Thus national currencies began to play a role as international means of payment. Moreover, this period also saw extensive use of loans from one central bank to another as a temporary source of liquidity. Again, the period was too short to reveal the potential problem of relying on growth of the ratio of the gold-convertible currency holdings relative to a more slowly growing stock of monetary gold, a problem underlined later by Triffin (1960).

Floating Exchange Rates

Floating exchange rates during the interwar period were not thought of as a “system” at all, but rather as an unavoidable but temporary expedient during a turbulent time until a more stable system could be re-established. The “adjustment process” worked by a market price—the exchange rate—clearing the market for foreign exchange, much as the price of strawberries clears the market for strawberries. If for any reason home demand for a foreign currency rose, the exchange rate of that currency would appreciate to ration the demand to what was available at the new price. The most comprehensive examination of floating exchange rates during this period (Nurkse (1944)) found them to be a major source of disturbance rather than a source of smooth adjustment, and that view strongly influenced the re-establishment of fixed exchange rates under the Bretton Woods system.

International liquidity in the sense of an internationally accepted means of payment was not necessary under these arrangements, since residents would buy the foreign currencies they needed in the foreign exchange markets, and the banks need not hold international reserves. Because this arrangement was considered temporary, however, central banks had their eye on the longer term and continued to hold gold. In 1919–20 and again in 1929–33 there was a sharp drop in wholesale prices, so the real value of monetary gold rose. In addition, both the United Kingdom and the United States devalued their currencies against gold (that is, they redefined the gold content of a pound and a dollar) in the 1930s, and that also increased the monetary value of gold reserves.

Bretton Woods—Early Phase

The Bretton Woods system was the first international monetary system to be designed from scratch. It stipulated fixed exchange rates, but recognized that with national full employment commitments, “fundamental disequilibria” might arise from time to time and it called for a discrete change in exchange rates, with international approval. It was taken for granted that monetary and fiscal policy would be used to achieve domestic equilibrium, as defined by each country. International disequilibria were to be financed in the short run, drawing if necessary from the new International Monetary Fund.

There was a major adjustment in exchange rates in 1949, but thereafter exchange rate changes of leading currencies were rare. Instead, adjustment was achieved by differential liberalization of trade and payments. Many countries were in a suppressed disequilibrium following World War II and maintained tight restrictions over trade and payments. These restrictions were gradually relaxed, country by country, as conditions permitted. Not until the end of 1958 did Western European countries abandon controls on current account transactions, and many maintained restrictions on outward capital movements long after that.

Curiously, the Bretton Woods system made no provision for a secular rise in international liquidity. Since the United States held a disproportionate share of the world’s monetary gold reserves (more than 70 percent) in the late 1940s, the gold reserves of other countries could be built up in part by drawing on those of the United States, as well as from new production. In addition, however, countries accrued substantial balances of U.S. dollars, as under the gold exchange standard. It should be noted that the world economy grew much more rapidly in the 1950s than anyone dared to expect in the 1940s. The rapid growth in production and trade seemed to call forth a corresponding growth in demand for international reserves, which was satisfied in large part by U.S. Treasury bills, thought to be better than gold since they bore interest and for monetary authorities were convertible into gold at the U.S. Treasury on demand. These circumstances gave rise to the celebrated Triffin (1960) dilemma: how could the world economy grow without additional dollars, yet how could the gold convertibility of the dollar remain credible as U.S. liabilities to foreign monetary authorities grew continually relative to the U.S. gold stock?

Bretton Woods—Late Phase

The conceptual underpinnings of the Bretton Woods system remained the same, but the scope for differential payments liberalization as the mechanism of adjustment diminished. Some modest exchange rate adjustments among major currencies occurred during the 1960s (the deutsche mark in 1961, the pound sterling in 1967, the French franc in 1969), but for the most part, countries were spared the need to adjust by a large and growing U.S. payments deficit, financed by a buildup in dollars held by central banks. When this buildup became unacceptably great, the Bretton Woods system broke down. It broke down basically because discrete changes in exchange rates, the key feature of its adjustment mechanism, are incompatible with the high mobility of private capital, which had resumed by the late 1960s. Any anticipated change in official exchange rates evoked a huge movement of speculative capital, which played havoc with domestic monetary policy. The emergence of any “fundamental disequilibrium,” especially if it was in the world’s largest economy, was bound to be financially destabilizing under these circumstances.

In the late phase of Bretton Woods, the question of international liquidity, omitted earlier, was addressed systematically, and the result was creation of a new kind of international fiat money, the SDR, to be held and used by central banks. The SDR was designated to become the centerpiece of the international monetary system, to be created as needed to serve the world’s need for international liquidity. In fact, total SDR creation has amounted so far to only about $25 billion, less than 10 percent of official foreign exchange reserves.

Managed Floating, 1973-Present

Since 1973 the major currencies have been floating with respect to one another, but the floating has been subject to market intervention by the monetary authorities—sometimes heavy intervention—to influence the movement of exchange rates. And most European currencies since March 1979 have been linked through the European Monetary System (EMS) in a Bretton Woods type system combined with permissible variation in market exchange rates around the central rates, and occasional changes (11 in all from 1979 through 1989) in central rates.

Under floating exchange rate arrangements, it is supposed that the main mechanism of adjustment will be changes in real exchange rates brought about by market-induced changes in nominal rates, supported as appropriate by changes in fiscal policy designed to maintain overall balance in the economy. (Within the EMS, of course, adjustment is similar to what it was supposed to be under the Bretton Woods system.) In fact, it is difficult to interpret the period of floating as behaving in this way, since manifestly exchange rates did not always move in such a way as to reduce current account imbalances. But interpretation of the historical record is complicated by two factors. First, macroeconomic policy, and especially fiscal actions, were not always conducive to international balance. On the contrary, in the early 1980s the United States pursued a markedly expansionist fiscal policy, combined with a tight monetary policy, while Japan, Germany, and the United Kingdom introduced strong fiscal contraction. The combination produced heavy upward pressure on the U.S. dollar, which in turn led to a sharp deterioration of the U.S. current account and corresponding increases in the surpluses of Japan, Germany, and, for a period, the United Kingdom (which was also influenced by increasing production of North Sea oil). In this instance, exchange rate movements “disequilibrated” the current account.

But in doing so, perhaps they were serving the broader role of equilibrating economies. That brings us to the second problem of interpretation: in an integrated world economy, there is no special merit in assuring balance in each country’s current account position. On the contrary, for a variety of reasons at particular times some countries will be net savers and others net investors, and one of the useful functions of an integrated economy is to channel savings to investment. If the integrated economy crosses national boundaries, the savings may also be expected to cross national boundaries, ([a-z]+) sometimes for substantial periods of time. As a consequence, we do not have a clear, operational definition of international equilibrium. At a minimum, “sustainable” capital movements should be set against the current account, ([a-z]+) unfortunately the actual purchases of foreign assets do not carry labels that tell us whether they are sustainable or not. Drawing the line at “long-term” investments will not do, since the definition of short-term (with original maturities under one year) is itself arbitrary, and in an integrated market many financial instruments, whatever their original maturity, may be purchased for short-term or speculative motives, while some short-term credits (for example, trade credits) may be expected to grow predictably in total over time.

Another dividing line often suggested is between official monetary transactions and all others. This suggestion introduces the other characteristic of a monetary system, international liquidity. Under a system of freely floating exchange rates, there should be no need for international liquidity, that is, officially held international means of payment. But in fact we have not had free floating, and the growth of official foreign exchange reserves during the period of floating has been phenomenal (Table 1), despite the widespread view that liquidity was excessive at the end of 1972 before floating began.

The growth in international liquidity has been satisfied overwhelmingly by the acquisition of foreign exchange reserves. There has been only one allocation of SDRs (in three tranches) since 1973, amounting to less than $15 billion. The dollar has continued to be the preferred currency, but the official holdings of deutsche mark and Japanese yen have grown even more rapidly, starting from a much lower base. In addition, the members of the European Monetary Compensation Fund have opened unlimited lines of short-term credit for one another. Gold, the traditional reserve asset, continues to be held by many monetary authorities, but it is virtually never used. Indeed, there is no generally accepted method of valuation, since the official price remains $42 an ounce, while the market price has ranged from $200 to $800, and has remained in the vicinity of $400 an ounce for several years.

Much of the explosive growth in reserve holdings has been voluntary and desired, lending support to Harrod’s long contention that demand for reserves would be higher under floating exchange rates, not lower as economic theorists generally contended. The voluntary acquisition of reserves makes it inappropriate to consider a balance on current account plus net private capital movements a suitable measure of payments equilibrium. Some of the reserve acquisition, particularly the large official acquisition of U.S. dollars in 1987, took place not to satisfy a growing demand for reserves, but to brake the depreciation of the dollar against other leading currencies, particularly the yen and the deutsche mark. So increases in official foreign exchange holdings reflect a mixture of motives, being partly a consequence of defensive exchange rate actions under managed floating.

Table 1.End-of-Year International Reserves(In billions of U.S. dollar equivalents)
U.S. gold holdings20.117.811.111.211.1
Foreign exchange214.318.644.6370.8378.7
U.S. liabilities4.211.123.8157.1172.8
Total reserves47.660.292.5449.1477.7
Addendum: World exports34.2113.4280.31,844.61,783.0
Sources: Board of Governors of the Federal Reserve System,Federal Reserve Bulletin;and International Monetary Fund,International Financial Statistics.

At official prices of $35 an ounce before 1980 and $42 an ounce in 1980 and 1985.

Reported assets differ from U.S. reported liabilities by minor differences in concept, by measurement error, by official holdings of foreign exchange other than dollars, and by official deposits in the Eurocurrency market.

SDRs and reserve positions in the IMF.

Sources: Board of Governors of the Federal Reserve System,Federal Reserve Bulletin;and International Monetary Fund,International Financial Statistics.

At official prices of $35 an ounce before 1980 and $42 an ounce in 1980 and 1985.

Reported assets differ from U.S. reported liabilities by minor differences in concept, by measurement error, by official holdings of foreign exchange other than dollars, and by official deposits in the Eurocurrency market.

SDRs and reserve positions in the IMF.

II. A Brief Evaluation

What do we want of an international monetary system? I suggest (1) that it should contribute to our basic economic objective of growth with low inflation; (2) that to that end and for its own sake it should help reduce the uncertainties economic agents face as close as possible to the minimum intrinsic in nature and the economic system; (3) that it should permit diversity in the national pursuit of economic and social objectives with a view to maintaining harmonious relations among nations; and (4) that it should do all this as unobtrusively as possible.

In terms of such aggregate indicators as real economic growth, variability of inflation, and predictability of growth and inflation, the Bretton Woods system, while it lasted, was a superior performer compared with previous periods and compared with the current managed float (Table 2). In terms of long-term price stability, the gold standard performed best, although the reasons remain poorly understood, since both short- and medium-term price variability was sizable.

III. The Case for Flexible Exchange Rates

Two quite different traditions argue in favor of flexible exchange rates. The monetarist tradition (Friedman (1953)) emphasizes the feature of flexible exchange rates that insulates a national economy from external monetary disturbances, from inflationary impulses that may be coming from abroad. This tradition emphasizes that under fixed exchange rates or market intervention, monetary expansion abroad will lead to reserve outflows for the expanding country, but it will also lead to reserve inflows, hence monetary expansion, for the country that is not expanding. If the economy is operating at full capacity, this expansion will lead to domestic inflation. By eschewing exchange market intervention and allowing the currency to appreciate, this imported inflation can be avoided. In the smooth and frictionless world of economic models, flexible exchange rates provide perfect insulation against monetary impulses, positive or negative, coming from the rest of the world.

Table 2.Variance of Quarterly Forecast Errors (Times 1,000) for the United States



Gold standard
No clear standard
Bretton Woods
Fluctuating rates
Note: Quarterly forecasts are made by using a Kalman filter with respect to expected level and expected rate of change on past data for each series.Source: Meltzer (1986), p. 141.
Note: Quarterly forecasts are made by using a Kalman filter with respect to expected level and expected rate of change on past data for each series.Source: Meltzer (1986), p. 141.

There is also a Keynesian tradition that supports flexible exchange rates, or, more accurately, changes in exchange rates. It operates on the assumption that the economic system will be hit by “exogenous” disturbances from time to time; that for any single country, these disturbances may include the policies of other countries; and that the country will need to “adjust” to these disturbances. Under a competitive market system with continuous market clearing, any needed adjustment is spread throughout the system via price signals, and the loss of output will be minimal and short-lived. But with various rigidities in the formation of prices and wages, such as exist in all modern economies, some of the need for adjustment will be thrown on output and employment, thus producing economic waste. The possibility of changing exchange rates introduces an element of price flexibility into this system replete with price rigidities. Under some circumstances, movements in exchange rates may substitute for product or labor market price flexibility, achieve adjustments in real wages and prices, and thereby avoid some of the loss in output. In this view of the world, the possibility of exchange rate movements introduces an additional element of flexibility into the economy. It is noteworthy that while the monetarist and the Keynesian schools emphasize different aspects of exchange rate flexibility, and the monetarist school in particular underlines the need for full flexibility rather than merely for changes in the exchange rate, the two schools at this level of generalization are quite compatible.

IV. Problems with Flexible Rates

In spite of these arguments for exchange rate flexibility, there are a number of worrisome features about the present arrangements of floating exchange rates subject to occasional official management. First, as Tobin (1988) has pointed out, major adjustments to external disturbances may require changes in the overall price level between one country and another. To the extent that governments treat their price levels as policy targets, adjustment via this mechanism is thwarted. If price level targeting is wholly successful, and if relative prices are sticky—which was the starting point of the Keynesian rationale for floating rates—national action will offset the effect of exchange rate movements. Thus, for example, if the U.S. dollar depreciates in response to some external disturbance, leading to price increases in the United States, the Federal Reserve may tighten monetary policy in order to avoid “inflation.” The Deutsche Bundesbank, in contrast, eases monetary action in order to avoid “deflation.” In this way adjustment is shifted to output after all.

Second, the imperfect competition that leads to sticky prices may have a further implication, as Krugman (1989), elaborating an analysis by Dixit, has recently pointed out. Fixed market entry costs not only will slow the process of adjustment to changes in the exchange rate, but will actually create a band of variation in which, for the relevant industries, no adjustment will take place. The profit incentive must exceed a certain threshhold before entry is worthwhile. Once entry has taken place, and the costs of entry have been sunk, the profit disincentive must move correspondingly far to make exit the proper strategy. This difference between marginal cost on entry and on exit will be reinforced by uncertainty about future exchange rates. It may make sense for a firm to hold onto its market at prices even below its current marginal costs, ([a-z]+) the costs of re-entry after exiting the market are high, and if there is a sufficient probability that the conditions for re-entry—for example, a sufficiently large movement in the exchange rate—will recur.

The implication of this market feature is that exchange rates may have to swing very far in order to achieve real adjustment, since industries subject to significant entry and exit costs will be unresponsive to movements in exchange rates of a relatively minor character. In this respect, flexible exchange rates are not an especially efficient mechanism of adjustment since, when measured against long-run comparative advantage, misallocations of resources may be induced, and then persist for a long period of time, in response to exchange rate fluctuations. One might have in mind the United States in the mid-1980s, for instance, when many foreign firms, attracted by the profit possibilities created by an exceptionally strong dollar, entered the U.S. market profitably and then hung on tenaciously even after the dollar depreciated substantially, in some cases beyond the point at which profitability continued.

Third, a high variability of real exchange rates may reduce total investment in the sectors of the economy open to international competition. When a currency has depreciated strongly, profits in tradables will be high and cash flow will be good, but firms will be reluctant to invest in productive capacity because the cheap currency is not expected to last. They will simply enjoy their windfall profits and perhaps invest more in market opening, as discussed above. When a currency is strongly appreciated, on the other hand, even though the situation is expected by management to be temporary, profits and cash flow will be low, leading to low investment because of credit rationing and because of skeptical, risk-averse boards of directors. Although many other factors have undoubtedly also played a role, it is perhaps not a mere coincidence that investment in plant and equipment in the member countries of the Organization for Economic Cooperation and Development has been depressed since the inauguration of floating exchange rates in 1973, compared with the 1950s and 1960s. In particular, Europe failed to invest much when profits were high and rising as a result of a strong dollar in the period from 1983 to1985.

Fourth, since firms cannot hedge their investments in future production—as distinguished from a particular sale—through financial markets, they will do so by investing abroad, across currency zones, even if that means giving up some of the advantages of cost and scale associated with exporting from their home bases or some other lower-cost location. Because some of this diversification takes place through takeovers and buyouts, one possible further consequence is greater world concentration in certain industries, leading to a reduction of worldwide competition.

Fifth, at the national level businesses will seek to blunt what they consider unequal competition by urging an increase in trade barriers. Business firms generally feel they can cope with the market uncertainty that attends any growing, dynamic economy, as long as their competitors are subject to the same ups and downs. What they find intolerable is being placed at a competitive disadvantage with respect to their leading competitors for reasons unrelated to decision making in the firms, or indeed in the industry. Present exchange rate arrangements violate this strong desire, insofar as a firm can suddenly find itself facing much stiffer competition (or much less, but that is rarely a cause for concern) as a result of an exchange rate movement, which has its origins in the arcane world of finance. Certainly U.S. firms, and organized labor, greatly increased their pressure on the U.S. Government in the period 1983–85 for some form of relief against overcompetitive imports during the period of an exceptionally strong dollar. It is noteworthy that under the gold standard, while national price levels moved substantially, they tended to move in parallel instead of moving relative to one another.

These various factors involve a misallocation of resources arising from the uncertainties associated with exchange rate flexibility. Whether monetary arrangements can be improved is a complex question, which depends in part on whether exchange rate uncertainty is simply the surface manifestation of uncertainties intrinsic to the economic system as a whole, or whether exchange rate dynamics actually add to the uncertainties faced by those who must make decisions on production and investment, that is, on present and future output of the economy. It also depends on whether a superior set of arrangements can be found, one that reduces whatever incremental uncertainties market-determined exchange rates contribute to the economic system while preserving or finding an adequate substitute for the contribution that changes in exchange rates may make to reducing the costs of adjustment to those changes which dynamic economies will inevitably have to make from time to time.

Whether high volatility in floating exchange rates reflects the uncertainties in the economic system and instabilities in economic policies or adds to them is a source of unresolved controversy. It is difficult to sustain the view, however, that foreign exchange markets reflect accurately and faithfully only the disturbances exogenous to the economic system, plus the net impact of governmental actions, which may themselves either dampen volatility or augment it. Chart 1 shows the monthly changes in the U.S. dollar-deutsche mark exchange rate, corrected for differential movements in wholesale prices, during 1960–88. The increase in volatility following the inauguration of floating in 1973 is dramatic.

Chart 1.United States- Germany: Real Exchange Rate Changes

(In percent per month)

Source: Dornbusch (1988).

There is a growing body of evidence to support the view that practitioners have long held, namely, that financial markets have a dynamic of their own and are occasionally subject to bandwagon effects, or speculative bubbles, whether these be rational or irrational. Shiller (1984) has shown that U.S. stock prices have frequently followed paths that are very difficult to explain except in terms of fashion or social psychological dynamics. Indeed, one of the principal bases for stock selection by some specialists, chartism, presupposes that stock prices follow distinctive patterns that are basically unrelated to what is happening in the economy. To the extent that substantial numbers of investors adopt this basis of stock selection, chartism can become self-fulfilling provided the chartist prices stay within the wide bounds set by liquidation value of the firms and the competition for funds provided by yields on long-term bonds.

More recently, Krugman (1985 and 1989) has shown that the appreciation of the dollar in 1983–85 cannot have been based on fundamentals alone, even though some appreciation could have been expected on the basis of the configuration of monetary and fiscal policies in the United States, on the one hand, and in Japan, Germany, and the United Kingdom on the other. Krugman’s analysis starts with the observation that the “market” must have expected a subsequent depreciation of the dollar at a rate no greater than the interest differential between comparable assets denominated in dollars, yen, and deutsche mark. On generous assumptions about the effect of such expected depreciation on trade flows, Krugman shows that U.S. external debt would have grown explosively, that is, unsustainably. This simple calculation suggests that market participants were not paying adequate attention to the underlying fundamentals, but rather were following their own lead in a bubble, until it burst in early 1985, with encouragement from heavy purchases of deutsche mark by the Bundesbank.

Moreover, Frankel and Froot (1987) have shown on the basis of survey data that short-run exchange rate expectations are extrapolative, that is, they project recent rate movements forward into the near future, even when the movement is away from the “long-run”—6 to 12 months—view of what the exchange rate will be. Thus, exchange rate movements are subject to bandwagon effects in the short run. Over a longer period, exchange rate expectations seem to be regressive to recent past levels, according to the survey data. But of course market movements are a series of short runs.

These observations provide circumstantial evidence that the foreign exchange market, with its own dynamic, can introduce disturbances into the real side of the economy. At least on some occasions, it adds to the uncertainties that decision makers on production and investment must face. But is there anything that can be done about it? Are there exchange rate arrangements that are superior to the unstructured managed floating we have had, in the sense of reducing the uncertainty without incurring high costs in some other dimension, and especially in the costs of adjustment to the disturbances that will inevitably occur from time to time?

V. Proposals for Reform

A number of suggestions have been put forward. Tobin (1982) has suggested a modest tax on foreign exchange transactions to discourage short-term transactions of low social utility, with the presumption that that will reduce exchange rate volatility. Williamson and Miller (1987), Kenen (1988), McKinnon (1984 and 1988b), and Cooper (1984) have made proposals for introducing greater stability into exchange rate movements directly, but with different techniques and emphases.

A Transactions Tax

A small tax, say 25 basis points (0.25 percent), could be imposed on all transactions that involve converting one currency into another, including forward transactions. (Some go further and suggest a small tax on all financial transactions, but that is not discussed here.) Such a tax would raise the cost of all cross-currency transactions, but it would be so small that it could be expected to have virtually no effect on trade in goods and services and on long-term capital flows. But the tax would impose a relatively large cost on short-term in-and-out transactions. For instance, 50 basis points for a two-way transaction would require an interest differential of more than 25 percent per annum to cover it for a weekly turnaround and over 100 percent per annum for a daily turnaround. Such a tax would reduce greatly the huge volume of foreign exchange transactions that now occurs ($700 billion a day in New York alone), most of which are interbank transactions. The aim of the tax would be to reduce short-term exchange rate volatility and encourage greater emphasis on longer-term transactions, which presumably are socially more valuable.

In fact, however, the impact of such a tax on short-term volatility is entirely unclear. To the extent that banks, corporate treasurers, and other short-term traders are “market makers,” like securities specialists, their activities should be stabilizing and reduce short-run volatility. Taxing them out of short-term transactions would in that event lead to an increase in volatility. But to the extent that these traders seek quick short-term gains, must guess very short-run market developments, and therefore are subject to bandwagon effects, their activities may increase short-run volatility, and a transactions tax would reduce it.

In either case, a transactions tax would not prevent the emergence of major misalignments, such as those that occurred in the mid-1980s, except insofar as those arise from the cumulative effect of a series of short-run extrapolative expectations, which might not get started in the presence of a tax. Presumably one characteristic of a cumulative exchange rate movement in the “wrong” direction is that each participant feels he can reverse his position at the right moment, before the crowd turns, or at least ahead of it. The crowd could turn at any time, even in the near future. To that extent, a transactions tax could inhibit major speculative currency movements, but it is unclear that it would be enough to prevent them.

To be effective, the transactions tax would have to be introduced in all leading financial centers; otherwise, transactions would move to the tax-free centers, something that is increasingly possible with modern communications. But it would not be necessary to get universal agreement on the tax. It would suffice to stipulate that disputes arising over foreign exchange transactions could not be adjudicated in countries of the leading financial centers unless the tax had been paid. Since it takes years to establish a reputation for fair and impartial dispute settlements, a small tax would be unlikely to drive transactions to tax-free countries without such reputations.

A tax represents one proposal for reducing exchange rate volatility. That is not a certain outcome, however, nor is the avoidance of significant misalignment. Other proposals focus on commitments to affect exchange rate movements directly.

Target Zones

Williamson and Miller (1987) have proposed that the major countries establish “target zones” for their exchange rates. The basis for establishing the target zones would be a calculation for each country of a “fundamental equilibrium exchange rate,” which in turn would be based on mutually agreed upon current account targets for each country. These equilibrium rates would be recalculated at regular intervals so they could move in response to new information, but presumably they would move slowly. They in turn would be translated into a set of mutually consistent nominal exchange rates, which would represent the center of the target zone. At first, the zone around these central rates would be wide, say ±10 percent, to encompass substantial initial deviation from the target current account positions, but the zones could gradually be narrowed over time, once the current account targets were achieved.

Monetary authorities would intervene in exchange markets—and, more important, adjust their monetary policies—as exchange rates reached the edges of the target zone. These edges could be either hard—well-defined rates, which are not to be surpassed—or soft—presumptive points of intervention and “leaning” by authorities, but without a firm commitment that the boundaries will never be crossed.

With n countries and only n—1 exchange rates among them, there would be a degree of freedom in monetary policy. This would be directed toward aggregate demand in the community as a whole. Put another way, if a given exchange rate approached the edge of the zone, whether one country tightened its monetary policy to affect the rate or the other country loosened would depend on the overall state of aggregate demand. Each country, in addition to having a current account target, would also have a target for growth in nominal demand, chosen in a way to be mutually consistent with those of other countries, and national fiscal policies would be directed toward achieving these aggregate demand targets.

The effect of this system in operation would be to limit the movement of exchange rates and thus prevent the emergence of major misalignments of the kind that occurred in the 1980s. It would not, however, eliminate short-run volatility of exchange rates, and the uncertainties they create, unless the target zones were quite narrow, which is not envisaged. While it is possible for businessmen to hedge against unexpected exchange rate movements for particular transactions, it is not possible to hedge for an investment or a commitment to a marketing strategy that will take a number of years to mature. So this source of uncertainty would remain.

The Williamson-Miller proposal is ambitious in its demands on policy coordination among major countries and on the skillful manipulation of monetary and fiscal instruments by national authorities. It is especially ambitious with respect to the coordination of policy targets—real effective equilibrium exchange rates and growth in national nominal aggregate demand. Agreement on the key underpinnings for calculation of real effective equilibrium exchange rates—current account targets, plus the bearing of exchange rates on their achievement—is particularly demanding, not least because the setting of current account targets would be an intrinsically arbitrary exercise in a world of high capital mobility and open markets for goods and services. For concreteness, consider the case of Canada. As a high-income country, it should perhaps run a current account surplus, contributing to the transfer of real resources to the lower-income developing countries. But in every respect except income, Canada itself is a “developing country,” and indeed Canada has run current account deficits throughout most of its existence as an independent state. How should we determine what Canada’s current account position should be in the future? Or, for that matter, the current account position of the United States, which by comparison with Europe and Japan is also a developing country, with its population growing relatively rapidly, increasingly through immigration, as in the nineteenth century.

Kenen (1988) advocates something like an extension of the EMS to include the United States, Japan, and possibly other countries. He supports hard exchange rate margins with a band width of at least 10 percent, much larger than the EMS band, so that changes in central rates, which must occasionally be made, need not affect market rates and hence would not provide a one-way speculative option when accurately anticipated. Kenen would also increase the visible reserves available for intervention, by analogy with the EMS, especially for the United States, by creating a modified substitution account into which the United States would deposit gold and others would deposit dollars.

On the choice of central rates, Kenen takes the view that that is less relevant than the procedures for changing them. If they are not right, that is, are needed for adjustment, they should be altered. He would give weight (but not mechanically) to a host of indicators, especially changes in relative price competitiveness, as in the EMS, but in the end, changes in central rates would be discretionary, to be negotiated among all the relevant parties.

Key Currency Monetary Coordination

McKinnon (1984 and 1988b) concentrates attention on just three countries and their currencies: the United States, Japan, and Germany, although he assumes several other currencies will be linked to the deutsche mark through the EMS. He would at first confine movement of exchange rates among the three key currencies within a 10 percent band, like Kenen, but over time he would gradually reduce the width of the band so that the exchange rates among the three key currencies eventually showed little or no movement. In addition, McKinnon departs from the other two proposals by selecting the central rates to which the exchange rates are to converge on the basis of purchasing power parity. Concretely, McKinnon would construct broad-based indices of tradable goods, and would choose exchange rates that equate the value of a comparable basket of these goods (McKinnon and Ohno (1988)). Nominal exchange rates would then remain fixed except insofar as changes were required to offset relative changes in purchasing power parity as defined by the baskets. Monetary policy in each of the countries would be dedicated to maintaining the fixed exchange rates; monetary policy in the three countries taken together—close coordination would be required—would be dedicated to maintaining stability in the prices of the baskets of tradable goods, as was advocated over half a century ago by Keynes (1930). Such a target implies moderate inflation as measured by the consumer price index, because of its component of services.

A Common Currency

Cooper (1984) would go a step further than McKinnon by institutionalizing the close monetary cooperation in coherent management of a single currency for the industrialized democracies. So long as national currencies are distinct, under distinct management, the possibility of major changes in nominal and real exchange rates exists, and that possibility is itself a source of uncertainty to investors so long as memories of the 1970s and the 1980s persist. The most effective way to eradicate exchange rate uncertainty is to eradicate exchange rates, that is, to introduce a single currency. A single currency would require a single monetary authority, which would represent a bold, even radical step—one that governments and their publics are not yet ready to contemplate seriously, much less undertake. So unlike the proposals discussed above, which are put forward with the near future in mind, this proposal has to be envisioned in a longer time frame, into the next century. Still, it carries the logic of a return to fixed exchange rates to its full conclusion, and for that reason is worth exploring more fully.

The institutional aspects of a common currency are not so difficult to imagine: they could be constructed by analogy with the U.S. Federal Reserve System, which is an amalgam of 12 separate Reserve Banks, each issuing its own currency. One could imagine an open market committee for all or any subset of the industrial democracies that would decide the basic thrust of monetary policy for the group as a whole. On it could sit representatives of all member countries, with votes proportional to gross national product. At one extreme the representatives could be ministers of finance; at the other they could be outstanding citizens chosen by their governments for long terms solely for the purpose of managing the monetary system. An obvious interim (and possibly permanent) step would be to appoint the senior governors of existing central banks.

National central banks could continue as the national components and agents of the new international Board of Bank Governors, and indeed could remain the issuers of currency, just as the 12 District Federal Reserve Banks do in the United States. If national sentiment called for it, currency designations—pounds, deutsche mark, yen, francs, dollars—could even be retained on nationally issued currency. The central point is that monetary policy would be out of the hands of any single government. Governments could not finance their budget deficits through monetary expansion, and the national currencies would exchange at fixed exchange rates. Most commercial transactions do not involve currency at all, so all commercial and bank transactions could take place in a common unit of account. A common currency would of course eliminate exchange rate uncertainty not only for commercial transactions, but also for financial transactions, so a unified capital market would develop throughout the area covered by the currency. By the same token, however, changes in nominal exchange rates could not be used any longer as part of the adjustment process. More will be said about this below.

A common currency could create serious adjustment problems if the exchange rates among the precedent currencies were not right at the time of conversion into the common currency. For this reason also, such a move cannot be consummated until major disequilibria have been eliminated. Purchasing power parity conditions, such as those suggested by McKinnon, must be met at least approximately; in addition, uncovered interest parity over all maturities should obtain at least approximately, implying no expectation of future changes in exchange rates.

For the participating countries, international and national currency would become identical, and liquidity needs would be satisfied by decisions by the Board of Bank Governors to increase the money supply.

VI. Evaluation of the Single-Currency Proposal

One objection that will be raised immediately against a single money for the industrialized democracies is that it involves ceding too much sovereignty to an international entity (in this case, the Board of Bank Governors). This is a misguided objection. Ultimate sovereignty continues to reside with the national governments. It is an exercise of sovereignty, not an abrogation of sovereignty, to agree on a common endeavor with other sovereign nations. The key question that should be asked is whether, on balance, the particular exercise of sovereignty leaves the participants better or worse off. It would limit the freedom of action of individual governments in the monetary arena. But the economies of these nations are becoming increasingly interdependent, and that economic interdependency increasingly limits the efficacy of individual actions in the areas of macroeconomic policy, taxation, and financial regulation. So retaining full freedom of action may turn out to be largely empty short of withdrawal into autarky, which would be extremely costly. A cooperative endeavor, while reducing national freedom of action, will restore effectiveness to joint action.

What of the adjustment process? The principal argument for exchange rate flexibility is that it may reduce the costs of adjustment to various economic disturbances. It should be kept in mind, however, that the requirements for efficient adjustment depend on the nature of the disturbances; and the nature of the disturbances in turn depend in part on the nature of the monetary system.

In particular, disturbances to national economies that are wholly or largely monetary in nature will be greatly diminished or eliminated altogether when the nations share a common currency, with no chance (short of major political disturbance) of changes in exchange rates among national monies. This would apply, for instance, to shifts in asset preferences among national financial claims motivated by expected changes in exchange rates, or to differential inflation among nations in tradable goods. Differences in national wage settlements not based on changes in productivity cannot be ruled out under a common currency, but they are much less likely in the presence of a common currency and extensive trade between nations. So at least some disturbances for which changes in exchange rates might be helpful hardly exist under a common currency.

What about disturbances to the real side of the economy, such as the discovery or exhaustion of natural resources, the technological changes that have differential effects among sectors, or the divergent rates of growth between demand and supply among nations?

With respect to technological change and discovery or exhaustion of resources, three observations can be made concerning economic adjustment in the industrialized democracies. First, most of the adjustment will have to take place within the three regions of the United States, Japan, and the European Community. Sectors, especially resource-based sectors, are often concentrated geographically, and changes in technology or the pattern of demand will require consequential adjustment among them. Such adjustments now take place within these economic areas. But exchange rate changes facilitate adjustment between countries, not within them. Yet these three economies taken as a whole are large and diversified; so little adjustment is likely to be required between them. Second, major disturbances at the global level, such as the oil price increases of the 1970s, are likely to affect all three regions in roughly the same way, although not of course identically. Third, the differential effects that remain once the first two points are taken into account are likely to be manageable within the parameters established by the natural growth and retirement of the labor force and the capital stock. Changes in real income and in domestic relative prices (for example, between tradables and nontradables) will help to bring about the adjustment, without substantial changes in output.1

With respect to differential growth in national (or regional) demand and output, such discrepancies, whether merely cyclical or reflecting more durable changes in saving behavior, can easily be accommodated by capital movements motivated by relatively small differences in yield within an integrated capital market, such as would obtain over a common currency area under modern conditions.

Possibly the greatest source of disturbance between large and diversified economies would arise from significant and opposite changes in fiscal policy. If one country pursued fiscal expansion while another was contracting substantially, that could create significant adjustment problems, and changes in exchange rates might assist the adjustment. But two things should be said about this possibility. First, it is desirable to be able to use fiscal action, within limits, to affect aggregate demand at the national or regional level. The adoption of a common currency, far from preventing this, would enhance the desirability of fiscal flexibility. Thus, it may be undesirable to allow or encourage exchange rates to adjust in response to discrepant fiscal actions, since such adjustment both affects the structure of output (for example, between tradables and nontradables) and weakens the demand effects of the fiscal actions (for example, by leading to a fall in net exports attendent upon a fiscal expansion).

Second, however, fiscal deficits would have to be financed exclusively in the (integrated) capital market. So long as a government’s credit was good, it would have no trouble borrowing. As the ratio of public debt to tax revenues grew, however, the market would require higher yields to be willing to take that government’s securities. There would be market signals indicating when one government was markedly out of line with the others in terms of growing indebtedness. Thus, while fiscal freedom of action would be unimpaired, it would be limited by the ability to service public debt. It would undoubtedly be useful to have informal discussions concerning the framing of fiscal action among all the participants; but formal coordination of fiscal action would not be necessary, and full harmonization would not be desirable.

The above points can be summed up briefly by saying that in the not very distant future nominal exchange rate flexibility among major currencies may create more disturbances for the real productive side of national economies than it corrects.


    Bloomfield,ArthurI.,Monetary Policy Under the International Gold Standard, 1880-1914 (New York: Federal Reserve Bank of New York, 1959).

    • Search Google Scholar
    • Export Citation

    Cooper,RichardN.,“The Gold Standard: Historical Facts and Future Prospects,”Brookings Papers on Economic Activity: 1 (1982), Brookings Institution (Washington), pp. 1–43, reprinted inRichardN. Cooper,The International Monetary System: Essays in World Economics (Cambridge: MIT Press, 1987).

    • Search Google Scholar
    • Export Citation

    Cooper,RichardN.,“A Monetary System for the Future,”Foreign Affairs (New York), Vol. 63 (Fall1984), pp. 166–84, reprinted inRichardN. Cooper,The International Monetary System: Essays in World Economics (Cambridge: MIT Press, 1987).

    • Search Google Scholar
    • Export Citation

    Cooper,RichardN.,ed., International Finance (Baltimore: Penguin Books, 1969).

    Dornbusch,Rudiger.,“The McKinnon Standard: How Persuasive?”Cato journal (Washington), Vol. 8, No. 2 (Fall1988), pp. 375–83.

    Frankel,JeffreyA., andKennethA. Froot,“Using Survey Data to Test Standard Propositions Regarding Exchange Rate Expectations,”American Economic Review (Nashville, Tennessee), Vol. 77 (March1987), pp. 133–53.

    • Search Google Scholar
    • Export Citation

    Friedman,Milton,“The Case for Flexible Exchange Rates,”in Essays in Positive Economics (Chicago: University of Chicago Press, 1953).

    • Search Google Scholar
    • Export Citation

    Jones,R.W.,“Presumption and the Transfer Problem,”Journal of International Economics (Amsterdam), Vol. 5 (August1975), pp. 263–74.

    • Search Google Scholar
    • Export Citation

    Kenen,PeterB.,Managing Exchange Rates (London: Routledge, 1988).

    Keynes,J.M.,A Treatise on Money, Vol. 2 (London: Macmillan, 1930).

    Krugman,Paul,“Is the Strong Dollar Sustainable?”in The U.S. Dollar: Prospects and Policy Options (Kansas City: Federal Reserve Bank of Kansas City, 1985).

    • Search Google Scholar
    • Export Citation

    Krugman,Paul,Exchange-Rate Instability (Cambridge: MIT Press, 1989).

    McKinnon,RonaldI.,An International Standard for Monetary Stabilization (Washington: Institute for International Economics, 1984).

    McKinnon,RonaldI.,(1988a), “An International Gold Standard Without Gold,”Cato Journal (Washington), Vol. 8, No. 2 (Fall1988), pp. 351–73.

    • Search Google Scholar
    • Export Citation

    McKinnon,RonaldI.,(1988b), “Monetary and Exchange Rate Policies for International Financial Stability: A Proposal,”Journal of Economic Perspectives (Nashville, Tennessee), Vol. 2, No. 1 (Winter1988), pp. 83–103.

    • Search Google Scholar
    • Export Citation

    McKinnon,RonaldI., andKenichiOhno,“Purchasing Power Parity as a Monetary Standard” (unpublished; San Francisco: Federal Reserve Bank of San Francisco, October1988).

    • Search Google Scholar
    • Export Citation

    Meltzer,AllanH.,“Some Evidence on Comparative Uncertainty Experienced Under Different Monetary Regimes,”in Alternative Monetary Regimes, ed. by Colin DearbornCampbell andWilliamR. Dougan (Baltimore: Johns Hopkins University Press, 1986).

    • Search Google Scholar
    • Export Citation

    Nurkse,Ragnar,International Currency Experience: Lessons of the lnterwar Period (Geneva: League of Nations, 1944).

    Shiller,RobertJ.,“Stock Prices and Social Dynamics,”Brookings Papers on Economic Activity: 2 (1984), Brookings Institution (Washington), pp. 457–98.

    • Search Google Scholar
    • Export Citation

    Tobin,J.,“A Proposal for International Monetary Reform,”in Essays in Economics (Cambridge: MIT Press, 1982).

    Tobin,J.,“Are There Reliable Adjustment Mechanisms?”in Toward a World of Economic Stability: Optimal Monetary Framework and Policy,ed. by YoshioSuzuki andMitsuakiOkabe (Tokyo: University of Tokyo Press, 1988).

    • Search Google Scholar
    • Export Citation

    Triffin,Robert,Gold and the Dollar Crisis (New Haven: Yale University Press, 1960).

    Triffin,Robert,The Evolution of the International Monetary System: Historical Reappraisal and Future Perspectives, Princeton Studies in International Finance No. 12 (Princeton, New Jersey: Princeton University Press, 1964).

    • Search Google Scholar
    • Export Citation

    Viner,Jacob,Canada’s Balance of International Indebtedness (Cambridge: Harvard University Press, 1924).

    Williamson,John, andMarcusMiller,Targets and Indicators: A Blueprint for the International Coordination of Economic Policy (Washington: Institute for International Economics, 1987).

    • Search Google Scholar
    • Export Citation

In this regard, Krugman (1989) exaggerates the changes in output—or under flexible rates, in the real exchange rate—required to achieve a given adjustment between two large and diversified economies. McKinnon (1988a), based on Jones (1975), argues persuasively that change in the terms of trade need not be large to accommodate a disturbance between two multisectored economies that are reasonably flexible, and it may not be necessary at all.

    Other Resources Citing This Publication