Journal Issue

Whither the exchange rate system?: Past experience, future options

International Monetary Fund. External Relations Dept.
Published Date:
June 1984
  • ShareShare
Show Summary Details

Lessons of experience and options for the future

Morris Goldstein

The last few years have witnessed a resurgence of calls for a reexamination, or perhaps even a reform, of the international monetary system. In this context the lessons arising from the experience with the present system of managed floating, and the likely advantages and drawbacks of some options for its evolution over the medium term, are of considerable interest. This article summarizes the findings of a recent comprehensive appraisal of the exchange rate system.

Two restrictions limit the scope of the discussion. First, it does not discuss international liquidity. However, many of the important issues in the evolution of the exchange rate system can still be profitably discussed in the context of current reserve and liquidity arrangements. Second, the emphasis is on the larger industrial countries. This reflects the facts that these countries account for a large share of total international trade and capital flows; that much of world trade is denominated in their currencies; that most developing countries and many smaller industrial countries have adopted some form of limited flexibility in their exchange arrangements; that their exchange rate options are limited by “structural” factors; and, finally, that it is the variability of the currencies of major industrial countries that has prompted calls for a reexamination of the system. More stability in the exchange rates of the major currencies would go a long way toward providing greater stability for other currencies.

This article is based on: The Exchange Rate System: Lessons of the Past and Options for the Future, IMF Occasional Paper No. 30, forthcoming.

The present system

The current exchange rate system has four prominent characteristics. First, there is a wide diversity of exchange arrangements among countries. Peggers far outnumber floaters, but most of world trade and finance is conducted among countries whose exchange rates float against each other. In mid-1983, 93 countries, almost all of them developing, were pegging their currencies; 17 countries had opted for what the Fund calls “limited flexibility vis-à-vis a single currency or cooperative arrangements,” including the 8 European countries that operate within the cooperative exchange arrangements of the European Monetary System; and 35 countries had adopted “more flexible” exchange arrangements, including “independent floating” by 4 of the largest industrial countries. During the period of flexible rates since 1973, there has been a trend away from pegged exchange arrangements, and, within these, from single currency to composite pegs (with former U.S. dollar peggers accounting for the bulk of the latter shift).

Despite the large number of countries that peg their currencies, in trade-weighted terms, the current system is better classified as floating because most of the largest traders maintain more flexible forms of exchange arrangements. In fact, in trade-weighted terms, about two thirds to four fifths of world trade is conducted at floating rates.

The second feature of the system is that exchange rates continue to be viewed as a matter of international concern. A stable system of exchange rates is now seen, however, as dependent more on stable macro-economic policies at the national level than on the form of the exchange rate regime itself. Present codes recognize explicitly that a system of stable exchange rates can be jeopardized as much by insufficient as by excessive exchange rate flexibility. Also, the Fund’s obligations for surveillance over countries’ exchange rate policies are now much greater than before.

Exchange rate variability is the third characteristic. This has been substantial—and for both nominal and real exchange rates, bilateral and effective exchange rates, and short- and longer-term time horizons. Exchange rate variability has been significantly greater than under the adjustable par value system and greater than variability in national price levels, but less than the variability of other “asset” prices.

By almost any measure, exchange rate variability has been much greater during the floating rate period (1973–82) than during the last decade of the adjustable par value system (1963–72). Further, the floating rate period has not shown a sustained tendency for exchange rate variability to decline over time. On most measures, it peaked in 1973, was on a declining trend for the next four to five years, and then rose sharply again during the late 1970s and early 1980s. The variability of nominal exchange rates under floating has also been substantially greater than implied by inter-country inflation differentials, yielding sizable changes in real exchange rates as well. The failure of purchasing power parity to hold has been particularly marked over the short to medium term. Going in the other direction, nominal exchange rate variability under floating has been much smaller than variability of some other asset prices (e.g., stock market prices, and changes in interest rates and commodity prices), suggesting that the floating rate period has been sufficiently turbulent to make all asset prices, not just exchange rates, fluctuate substantially.

Finally, although it is difficult to measure, it is clear that official intervention in exchange markets persists and the evidence is that the demand for reserves does not appear to have been appreciably diminished under floating rates. Most countries continue to regard exchange rates, at least in part, as a policy target. Intervention has not been aimed solely at countering disorder, or even at leaning against the wind, but has also included, inter alia, resisting rate movements that bear no relation to the fundamentals and resisting depreciation out of concern for its inflationary consequences or appreciation in order to maintain competitiveness.

Criteria for evaluation

With these broad characteristics of the present exchange rate system in mind, four criteria seem most appropriate in evaluating that system. First, does the system help or hinder macroeconomic policy in pursuit of fundamental domestic economic objectives (price stability, sustainable growth, high employment)? Second, how effective is the system in promoting external payments adjustment? Third, how does the system affect the volume and efficiency of world trade and capital flows (and thereby resource allocation in the international economy at large)? And fourth, how robust or adaptable is the system to significant changes in the global economic environment?

The first criterion reflects the view that the exchange rate system is basically a facilitating mechanism for more fundamental domestic economic objectives, like price stability, high employment, and sustainable economic growth. That is why, in sharp contrast to some earlier analyses of exchange rate systems, the degree of exchange rate variability, for example, is not put forward here as a normative criterion. In other words, exchange rate variability is only important to the extent that it impinges upon, or facilitates, achievement of the ultimate targets of economic policy.

The second criterion introduces considerations of external balance to supplement the internal balance objectives subsumed under the first. Connoting a desirable exchange rate system as one that promotes external payments adjustment implies that the system should set in train an (internationally acceptable) adjustment mechanism, either automatic or discretionary, that eliminates balance of payments disequilibria over a reasonable time period. To make such a criterion operational, it is, of course, necessary to have some definition or concept of balance of payments equilibrium. For the purposes of this article, it is sufficient to think of it as a condition under which the current account position can be financed by normal capital flows without recourse to undue restrictions on trade, special incentives to inflows or outflows of capital, or wholesale unemployment.

The third criterion derives from the proposition, given explicit endorsement in the purposes of the Fund, that global welfare is generally increased by an expansion of world trade and investment. This is another area where one wants the exchange rate system to act as a facilitating mechanism for some more basic economic objective. The criterion refers to the “efficiency” of trade and investment because in the real world, where international traders sometimes react to temporary relative price signals that bear little relation to longer-term changes in comparative advantage, not all increases in the volume of trade will be beneficial.

The rationale for including the fourth criterion is that, as with political constitutions, there are nontrivial costs associated with changing international monetary arrangements, especially under crisis conditions. Other things being equal, it is better to have an exchange rate system that is relatively robust or adaptable to changes in the global economic environment. Such an exchange rate system, for example, may have to work well under conditions of high international mobility of capital, or rapid or abrupt changes in comparative advantage, and accommodate changes in other environmental factors ranging from the degree of real wage flexibility to the preferences for a particular reserve currency or even the assumed behavior of one particular type of economic agent (be it the reserve center country or market speculators).

Lessons of floating

If these criteria are applied to the past decade of experience with managed floating, what is the outcome? The study finds the overall performance of the exchange rate system “remarkably good given the harsh global environment but with plenty of room left for improvement.” At the risk of oversimplifying the complexity of some issues and of ignoring others, the experience with floating rates also suggests the following lessons.

First, the capacity of the exchange rate system per se to do good or harm should not be overestimated. Neither the expectations of the proponents of floating rates nor those of its critics have been confirmed by experience. Floating rates have not provided complete or even good insulation against all types of external disturbances; they have not provided rapid and automatic equilibration of external payments imbalances; they have not eliminated or even significantly reduced the demand for international reserves; and they have not encouraged enough stabilizing speculation to keep real exchange rate movements within narrow bands corresponding to permanent changes in the terms of trade. On the other hand, they have not led to a collapse in international trade and investment; they have not destroyed the discipline to fight inflation; they have not trapped high-inflation and low-inflation countries in vicious and virtuous circles for long time periods, regardless of the authorities’ policy efforts; they have not reduced the size of price elasticities in international trade nor produced perverse long-term effects on current account imbalances; and they have not led to large increases in structural unemployment due to workers shifting back and forth among industries in response to very short-term currency fluctuations. The first lesson has been that the exchange rate system matters, but not as much as was previously thought.

Second, by contrast, the importance of discipline and coordinated macroeconomic policies for the successful operation of floating rates should not be underestimated. Floating rates have allowed more autonomy than fixed rates did in the use or control of policy instruments, but in a world where goods and assets are traded freely and are close substitutes across countries, this increased autonomy will not translate into more effective policies if domestic monetary and fiscal policies are unstable, unbalanced, and uncoordinated with those of other countries. Such policies will eventually destabilize the exchange rate because current rates are heavily dependent on expected exchange rates; the latter are closely tied to expected future macroeconomic policies; and these, in turn, are strongly influenced by past policy behavior. Exchange rate policy can no more be divorced from basic macroeconomic policies under floating than it could be under fixed rates.

The existence of a wide diversity of exchange arrangements does not imply a nonsystem or a lack of logical foundation—this is the third lesson. The optimal degree of exchange rate flexibility differs across countries in large part because of differences in their economic structures. Both theory and empirical evidence indicate that exchange rate changes in the smaller, more open, more highly indexed economies have a proportionally larger impact on domestic prices and give them a less lasting relative price advantage than changes in the larger, less open, and less indexed economies. It is therefore not surprising that the former seek to avoid frequent or substantial movements in exchange rates, while the latter favor increased exchange rate flexibility. As a result, uniform judgments about whether exchange rates fluctuated too much over the past decade are not likely to be valid across countries, and for reasons that go beyond intercountry differences in philosophy about the efficiency of markets. Going one step further, these same cross-country viewpoints about optimal exchange rate flexibility strengthen the case for better coordination of policies. For in the absence of such consultation and coordination, it is unlikely that “common” views about the proper distribution of the adjustment burden between exchange rates and other policy instruments will emerge on their own.

Fourth, in appraising the present exchange rate system, it is crucial to distinguish the effects of floating rates from other developments occurring during the period of floating. Floating rates do not seem to have been responsible for high inflation and high unemployment rates nor for the slow growth in productivity experienced by industrial countries over the past ten years. Nor do they seem to have been the key factor behind the slowdown in the growth of world trade. No exchange rate regime would have emerged unscathed from the combination of shocks, portfolio shifts, and structural and institutional changes of the past ten years. For the same reasons, even if major changes in the exchange rate system could be brought about, such changes would not, by themselves, be likely to reduce unemployment significantly, eliminate pressures for protection, lead to a resurgence in investment or productivity, or make economies immune from future disturbances. The exchange rate system is an important facilitating mechanism for economic interdependence, but it is not a panacea for the world’s current economic troubles.

Fifth, the present system has demonstrated some considerable strengths. Foremost is that exchange rate changes have made a positive contribution to securing effective external payments adjustment over the medium to long run. For example, despite some powerful external disturbances, the average size of payments imbalances of the larger industrial countries has been smaller and their duration shorter during the last decade than they were over the final ten years of the adjustable peg system. Asymmetries in adjustment between surplus and deficit countries, and between the most important reserve center and non-reserve currency countries, have been reduced. Similarly, given slowly adjusting national price levels, nominal exchange rate changes have made it possible for real rates to adjust to, inter alia, permanent changes in the terms of trade, significant natural resource discoveries, and continuing differences in trend rates of growth of labor productivity. Finally, although effective policy coordination among the major industrial countries has been the exception rather than the rule, and there have been serious lapses of discipline in national policymaking, the present system has at least maintained in the foreign exchange market a mechanism of conflict resolution that has involved neither suspension of currency convertibility nor large-scale restrictions on trade and capital flows.

Sixth, a good average performance in the harsh operating environment of the past decade does not mean that the present system has not had serious problems. The most critical one has been that real exchange rate movements (i.e., nominal exchange rate changes adjusted for inflation differences across countries) have sometimes gone far beyond those suggested by best estimates of “fundamentals” and have sometimes stayed out of line for periods of up to two to three years. Such maladjustments have created problems in two major areas.

First, they have handicapped efficient resource allocation. Large real exchange rate changes (for example, of 30 percent or more) over the medium term affect patterns of production, employment, investment, and consumption both within and across countries. Further, when these unsustainable exchange rate and payments positions do unwind, as they eventually must, there are likely to be adjustment costs because resources (especially labor) released from the overexpanded sectors do not quickly find employment elsewhere, particularly in the context of sluggish overall economic activity. Thus, even though such resource misallocation and adjustment costs were probably not the primary determinant of the recent inflation, growth, and unemployment performance, they made a difficult situation even more troublesome.

The second major difficulty has been in the policy reaction to disequilibrium exchange rates. Although the foreign exchange market does provide a decentralized solution to policy inconsistencies across countries, it has become increasingly evident that countries may resort to other more socially destructive administrative mechanisms to adjust to what they regard as a persistently “inequitable” rate. These mechanisms—most of which involve subsidies (overt or hidden), taxes, or quantitative restrictions on exports or imports—not only erode the gains from trade but also make cooperation more difficult in other areas of mutual concern. Again, even though exchange rate distortions have probably not been the prime cause of such restrictive measures, they have certainly not helped. This difficulty also highlights why it is so essential to take sufficient preventive measures in the future to keep an adjustment mechanism as imprecise as the exchange rate from having to shoulder too much of the adjustment burden.

Options for the future

The discussion of the options for change assumes that the rest of the 1980s may be somewhat more hospitable than the 1973–82 period. That is, the exchange rate system will probably still have to contend with, among other things, real and monetary disturbances, high international mobility of capital, and so on, but the average rate of inflation, as well as its dispersion across countries, could well be lower, and there may continue to be a reservoir of goodwill that can be tapped for efforts aimed at greater coordination of policies.

A reasonable intermediate objective of any changes to the present exchange rate system would be to maintain enough flexibility in real exchange rates to aid external adjustment but at the same time create conditions under which they do not stray so far, so often, and for so long as they have from levels consistent with fundamentals. (As long as exchange rate behavior is specified in terms of real exchange rates, this objective can, in principle, be satisfied by fixed rates, flexible rates, or any combination in between.) The first part of this objective might be considered as the present system’s principal strength and the latter, its principal weakness.

The first option is to return to par values. In this context, is it reasonable to envisage the return of conditions under which fixed exchange rates among the major currencies could be restored?

A negative answer rests on the following grounds: (1) the major countries would be unwilling to completely subordinate monetary policy to the dictates of a fixed exchange rate; (2) structural differences among countries are large enough to preclude the emergence of a common rate of inflation; (3) real exchange rate adjustments would be needed to reflect changes in comparative advantage; (4) prices and wages are too inflexible (particularly downward) to obtain the requisite real exchange rate movements without changes in nominal exchange rates; and (5) the absence of a willing or readily acceptable candidate for the central currency in the system.

The case for a positive answer is essentially that: (1) there is already in prospect a significant convergence of inflation rates for the four largest countries in 1984–85; (2) the discipline necessary to coordinate policies among the others will be given much impetus by the establishment of fixed rates; (3) the policy autonomy under alternative systems is largely illusory anyway; and (4) even a partial success (infrequently adjustable rates) would have a strong positive effect on domestic stability and the resumption of world trade growth.

Exchange rate formulas constitute a second option. If nominal exchange rates need to be adjusted to reflect fundamental changes, is there any rule or formula that could help determine the right structure of rates? How useful are “presumptive indicators” for signaling the need for adjustment?

The principal argument against the use of formulas or rules for determining appropriate changes in nominal exchange rates is that the factors that call for changes or that are symptoms of a maladjustment are too varied, too unpredictable, and too unstable over time to be captured ex ante in any formula or rule. While such a formula approach may have represented a reasonable second-best solution to the nominal exchange rigidities of the Bretton Woods era, it is not so in today’s world, where exchange rates are, if anything, too flexible. The main counterargument is that exchange rate formulas represent a reasonable middle ground between the excessive rigidity of rates that are administratively set and those that are market-determined.

Presumptive or “objective” indicators for adjustment are, of course, in principle less restrictive because they usually do not specify which combination of adjustment measures the country should adopt. In brief, the case for them is that the regular examination of a set of multiple indicators could help to detect problems at an early stage and induce a more timely and more symmetrical pattern of adjustment than would occur in their absence (or at least trigger dicussions of policy among countries that make coordinated surveillance workable). As with exchange rate rules, opposition to them is often based on the arguments that there is no simple indicator that will consistently transmit reliable adjustment signals and that even if there were one, practical problems over its precise definition, measurement, and monitoring would severely limit its applicability. (The only exchange rate arrangement that has actually implemented a presumptive indicator is the European Monetary System. In that system, once a country’s actual exchange rate crosses a “threshold of divergence” from the ECU central rate, there is a presumption that the authorities will undertake corrective measures.)

A third option is adjustable par values with narrow margins. But would these be viable for the major currencies given the current high capital mobility?

Although this issue is similar in many respects to the first (on the restoration of fixed exchange rates), it is dealt with separately because capital mobility is widely cited as the key contributory factor to the breakdown of the Bretton Woods system. Indeed, the case against a return to adjustable par values with narrow margins is that none of the factors that made the Bretton Woods system so vulnerable to “hot money” flows would be less problematic today or tomorrow. Large and suddenly changing interest rate differentials would still arise because of the failure to harmonize monetary and fiscal policies across countries, there would be rumors of imminent parity changes due to a whole host of circumstances, and the resources of central banks would be insufficient to cope with the larger resources of private speculators. Some would say par values would be even more vulnerable today because liberalization measures and technological advances have combined to render capital much more mobile than during the 1950s and 1960s. Hence, if such an adjustable peg scheme could work at all, it would need both wider margins and some mechanism to ensure prompt adjustment of par values.

The opposing view is not so much that these problems are less serious today but that their intractability is exaggerated. Specifically, such a system can function even with relatively narrow margins if there are sufficient political commitment, generous support for riding out balance of payments difficulties, active exchange market intervention, a presumptive indicator for adjustment, and the acceptance of the need for occasional, and sometimes large, realignments of central rates.

A fourth option is to reduce excessive exchange rate variability by new taxes or restrictions of international capital flows.

This familiar issue has gained new support through the debate on how to cope with overshooting (or excessive fluctuation) of floating rates. The case against restricting the international capital market includes the following arguments: (1) there is no strong presumption that the costs in terms of resource misallocation from impeding the international flows of capital would be less serious than those arising from restricting the flow of goods; (2) there is no reliable method of separating in advance productive from nonproductive capital flows; (3) any tax on capital flows would make it more difficult for a country to finance a current account imbalance because it would have to raise interest rates enough not only to create a favorable interest rate differential but also to offset the cost of the tax; (4) even aggressive control programs, such as those of the early 1970s, often failed to stem private capital flows, and the subsequent development of offshore banking markets makes their efficacy today less likely; and (5) unless uniform restrictions or taxes could be negotiated and accepted by all parties, there would be a constant danger of escalation and retaliation, with damaging spillovers for other international transactions. The case for such impediments to capital flows does not ignore these costs; instead, it brings out the point that they will be smaller than the macro-economic costs associated with larger exchange rate fluctuations under free mobility of capital.

Greater stability of floating exchange rates should be sought primarily in more stable domestic macroeconomic policies and in better coordination of these and of other policies across countries.

As emphasized earlier, it is now widely accepted that floating exchange rates would be less volatile if medium- and long-term private sector expectations about exchange rates were firmer. The case for stressing the implementation of stable, credible, and balanced policies is simply that, quite apart from their favorable impact on domestic economic objectives, these policies are the single most important ingredient in stabilizing exchange rate expectations. For if market participants cannot gauge the medium-term course of basic policies, and if they cannot be confident that the basic economic objectives can be reconciled across countries without either dramatic shifts in policy mixes and/or in exchange rates, they will have little basis from which to form a view about future exchange rates. In such circumstances, speculative “bubbles” and “bandwagon” effects become more prevalent because there are no natural bounds for the expectations of speculators.

Conversely, where countries have a history of stable policy behavior and where, therefore, forecasts of policy intentions have credibility, neither minor shocks nor short-term deviations of policies from targets are likely to be translated into large exchange rate movements because longer-term expectations about rates will not be much affected. (The relative stability of the floating Canadian dollar over 1950–61 is often cited as an example of such a stabilizing anchor at work.) Because better conduct is widely recognized as improving the functioning of any exchange rate regime, there is, of course, no case against better macro-economic policies. But there are doubts and questions about if and how such improved policy conduct can be brought about within the present system.

A final major option is to establish official “forecasts” or “target zones” for exchange rates, to help both to reduce their variability and to increase the incentives for external adjustment.

The case for official forecasts or target zones rests on two arguments. First, in their absence, it is too difficult for market participants to form a view about future exchange rates. Even where policies are relatively stable, there are just too many factors that affect an exchange rate to make a firm judgment about its value six or eighteen months ahead. Second, because the authorities would be under some pressure either to keep actual rates within the target or forecast zone, or to explain departures from it, it is claimed that the speed of external adjustment would be increased. It is argued that without such official forecasts, authorities have insufficient incentives for adjustment, since they can always equate the “right” rate with the market rate.

The case against official forecasts of exchange rates is: (1) that given stable underlying macroeconomic policies, there is no need for an additional anchor for exchange rate expectations; (2) that negotiation of forecast rates or zones, and changes in them, would be subject to all the centralized management delays of the Bretton Woods era, thereby robbing the scheme of its flexibility; (3) that the best guide for domestic monetary policy is still that of achieving price stability; in contrast, the exchange rate can often give false signals; and (4) that forecast rates or zones would only have credibility if they were backed by broad coordination of macroeconomic policies—and if such coordination could be achieved, no change in the present system would be necessary.

Published by the International Monetary Fund …

World Economic Outlook

1984 Edition

The report on the World Economic Outlook for 1984 is now available. This study, which is prepared by the Fund staff and is published as Occasional Paper No. 27, provides a comprehensive picture of the current international economic situation and its medium-term prospects. After a General Survey of the world economy and an outline of the main features of the staff’s projections for 1984, the report discusses in separate chapters some of the key policy issues at the present time, specifically:

  • Sustaining the recovery in economic activity in the industrial world

  • Problems of growth and adjustment in developing countries

  • External debt and payments adjustment

  • Exchange rate developments among the major currencies.

The report also contains a number of supplementary studies and statistical tables.

The World Economic Outlook for 1984 is available from:

  • Publications Unit, Box A-842

  • International Monetary Fund

  • Washington, D.C. 20431, U.S.A.

  • Telephone: (202) 473-7430

Price: By air mail: US$15.00 a copy (US$10.50 for university libraries, faculty, and students).

By surface mail: US$12.00 a copy (US$7.50 for university libraries, faculty, and students).

Other Resources Citing This Publication