A logical starting point for any discussion with respect to modification of the exchange rate system is an examination of its present status. Such an examination should include at least three elements: (1) an identification of the main descriptive characteristics of the present system; (2) a summary of the operating environment in which the present system has had to function; and (3) an evaluation of the present system in terms of normative criteria applicable to all exchange rate systems. This section of the paper deals with the first two of these elements.

A logical starting point for any discussion with respect to modification of the exchange rate system is an examination of its present status. Such an examination should include at least three elements: (1) an identification of the main descriptive characteristics of the present system; (2) a summary of the operating environment in which the present system has had to function; and (3) an evaluation of the present system in terms of normative criteria applicable to all exchange rate systems. This section of the paper deals with the first two of these elements.

Characteristics of the Present Exchange Rate System

The present exchange rate system can be described by the following four prominent characteristics.

Diversity of Exchange Rate Arrangements

There is a wide diversity of exchange rate arrangements among countries. The peggers far outnumber the floaters, but most of world trade and finance is conducted among countries whose exchange rates float against each other.

Under the Bretton Woods system of adjustable par values, all countries had to adhere to declared par values for their currencies; further, under the Fund’s original Articles of Agreement, these par values could be altered only to correct “a fundamental disequilibrium” and only after “consultation with the Fund.”7

In contrast, the present system allows member countries almost complete freedom of choice about their exchange arrangements; indeed, the only restriction is that if a country chooses to peg its rate, it must not denominate that peg in terms of gold.

This freedom of choice has been vigorously exercised. As of June 30, 1983, 93 countries, almost all of them developing countries, chose to peg their currencies to a single currency or to a currency composite; 17 countries 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 adopted “more flexible” exchange arrangements, including “independent floating” by 4 of the largest industrial countries (Canada, Japan, the United Kingdom, and the United States).8 During the period of floating rates (1973–83), there has been a trend away from pegged exchange arrangements to other arrangements, and within the pegged arrangements there has been a trend away from a single currency to a composite of currencies (with former U.S. dollar peggers accounting for the bulk of the latter shift).9 Thus, if anything, exchange arrangements have become more heterogeneous during the period of floating rates.

The relatively large number of countries maintaining some form of pegging arrangement should not create the impression that the present system is basically one of fixed rates. In fact, in trade-weighted terms, the current system is much better classified as a floating rate system. This is because most of the largest traders maintain either limited flexibility or more flexible exchange arrangements. More specifically, whether “floating trade” is measured as the total trade of those countries having other than pegged exchange arrangements or is measured, in a more sophisticated manner, as the global sum of the portion of each country’s trade that takes place at floating rates, the result is that about two thirds to four fifths of world trade is conducted at floating rates (Goldstein and Young (1979) and Solomon (1983)).

The fact that countries have chosen to adopt such a wide variety of exchange arrangements carries at least two important implications for the following analysis. The first implication is that caution is needed in referring to the present exchange rate system because in reality many exchange rate subsystems exist. This feature may cloud comparisons with the more homogeneous exchange rate systems of the past. The second implication is that the optimal degree of exchange rate flexibility may really be different across countries, owing in large part to their different economic structures and to the different nature of the shocks facing them.10 This means, for example, that any proposal that requires all participants to adhere to the same degree of exchange rate flexibility (whether large or small) may well meet strong resistance from at least some country group or groups. This is apt to make the process of consensus building much more difficult.

Codes of Conduct for Exchange Rate Behavior

Exchange rates continue to be viewed as a matter of international concern. A stable system of exchange rates is now seen, however, to be more dependent on stable macroeconomic policies at the national level than on the form of the exchange rate regime itself. Also, the Fund’s obligations for surveillance over countries’ exchange rate policies are now much greater than before.

Despite frequent complaints that present arrangements are a nonsystem, it is not true that present codes of conduct show less concern about the obligations of countries with respect to their exchange rate policies. Like their immediate predecessor, the Fund’s Articles of Agreement (as amended in 1978) enjoin member countries “to collaborate with the Fund and other members to assure orderly exchange arrangements and to promote a stable system of exchange rates.”11 Further, the Articles are quite specific in their view of why a stable system of exchange rates is beneficial to the world community. Such a system, according to Article IV, Section 1, “facilitates the exchange of goods, services, and capital among countries, and . . . sustains sound economic growth.”

What is different about the present codes of conduct for members can be summarized under three points.

(1) The present codes give clear support to the view that the path to exchange rate stability lies fundamentally in the pursuit of sound domestic policies at the national level, whatever form of exchange arrangements is adopted by a country. Toward this end, each member country should “endeavor to direct its economic and financial policies toward the objective of fostering orderly economic growth with reasonable price stability” and “seek to promote stability by fostering orderly underlying economic and financial conditions and a monetary system that does not tend to produce erratic disruptions.”12 In contrast, the Bretton Woods codes required stability of exchange rates in terms of adherence to declared par values but were not explicit as to how the domestic policies conducive to such stability were to be fostered.

(2) A second key difference is that the present codes recognize explicitly that a system of stable exchange rates can be jeopardized as much by insufficient exchange rate flexibility (in the form of prolonged maintenance of overvalued or undervalued rates) as by excessive flexibility. Thus, whereas the Bretton Woods code directed countries to “avoid competitive exchange alterations,”13 the present Articles warn countries to avoid manipulating exchange rates either “to prevent effective balance of payments adjustment” or “to gain an unfair competitive advantage over other members.”14

(3) The present codes imply a more active role for the Fund itself in monitoring the adherence of countries to their exchange rate policy obligations. Thus, whereas the Bretton Woods code assigned the Fund only the responsibility to concur with or to object to par value changes proposed by a member country, the present Articles require the Fund to “oversee the international monetary system in order to ensure its effective operation,” and in so doing, “to exercise firm surveillance over the exchange rate policies of members” and “adopt specific principles for the guidance of all members with respect to those policies.”15

The actual achievement of a stable system of exchange rates rests, of course, not only on the existence of clearly specified and well understood codes of conduct but also on their implementation. If a stable system of exchange rates has not been forthcoming during the period of floating rates, it is not because of the lack of codes of conduct.

Exchange Rate Variability

Exchange rate variability has been substantial. This statement applies to both nominal and real exchange rates, both bilateral and effective exchange rates, and both short-term and longer-term time horizons. Exchange rate variability has been significantly greater than it was under the adjustable par value system and greater than variability in national price levels but less than the variability of prices of other assets. Also, there are strong indications that most exchange rate changes have been unexpected.

Prior to the advent of managed floating, critics of floating exchange rates predicted that exchange rate changes under such a regime would show considerable volatility. Advocates of floating rates countered with the argument that stabilizing speculation would act to smooth exchange rate movements (especially for real exchange rates) so that the greater freedom of exchange rates to move would not necessarily lead to an abrupt increase in actual variability. After more than ten years of experience under floating rates, it is now clear that exchange rate variability has been more substantial than advocates of the system expected.

Charts 1 and 2 provide convenient pictures of the short-run variability of nominal bilateral exchange rates and of longer-term swings in real effective exchange rates, respectively, for the major currencies during the period of floating rates. It is apparent from these charts that exchange rate fluctuations were often large during that period—even day-to-day changes in nominal bilateral rates sometimes averaging 2–3 percent over a few quarters and real effective rates occasionally moving by 30 percent or even more over two to three years. More broadly, recent studies of exchange rate variability point to the following conclusions about the characteristics of such fluctuations.

Chart 1.
Chart 1.
Chart 1.

Short-Run Variability in Nominal Bilateral Exchange Rates for Five Major Currencies, April 2, 1973–November 30, 1983

Daily percentage changes

Chart 2.
Chart 2.

Nominal and Real Effective Exchange Rates in Major Industrial Countries, First Quarter of 1973 to Second Quarter of 1983 1

1 Index: average for period shown = 100. Indices of effective exchange rates are based on the Fund’s multilateral exchange rate model (MERM). Real effective exchange rates are calculated by adjusting indices of normalized unit labor cost in manufacturing for changes in nominal effective exchange rates.

(1) By almost any measure, exchange rate variability has been much greater during the period of floating rates (1973–82) than it was during the last decade of the adjustable par value system (1963–72). A recent Fund study (IMF (1984 a)), for example, reports that short-term (monthly or quarterly) variability of nominal exchange rates for the seven major currencies was about five times greater under floating rates than under fixed rates.16 Switching to real exchange rates or to longer time horizons (e.g., deviations from four-year to five-year trends) does not alter the qualitative nature of this conclusion.

(2) Within the period of floating rates itself, there has not been a sustained tendency for exchange rate variability to decline over time (e.g., see Shafer and Loopesko (1983), Kenen and Rodrik (1983), and IMF (1984 b)). On most measures, exchange rate variability peaked in 1973, was on a declining trend for the next four or five years, and then rose sharply again during the late 1970s and early 1980s. Thus, the oft-heard prediction that the large variability experienced at the onset of floating exchange rates would fade away as traders and policymakers learned how to operate under the new regime has proved false.

(3) The variability of nominal exchange rates under floating rates has been substantially greater than implied by inflation differentials across countries, thereby yielding sizable changes in real exchange rates as well. The failure of purchasing power parity to hold under floating rates has been particularly marked over the short to medium term-i.e., month to month or quarter to quarter, or even two-year to three-year periods (e.g., Isard (1977) and Katseli (1979)). Thus, only over long periods and only when relative price changes among countries have been quite large, has purchasing power parity served as a useful rule of thumb for explaining actual exchange rate movements.17 Without addressing alternative theories of exchange rate determination, it is enough to note here that such exchange rate behavior is consistent with the view that (a) national price levels are “sticky” and “backward looking” in the short to medium term (often reflecting the existence of previous contracts) while exchange rates are, like other asset prices, “flexible” and “forward looking” (often reflecting expectations about future events); (b) relative goods prices are only one of many factors operating on exchange rates, and they are not the predominant influence over short time horizons; and (c) the vast majority of exchange transactions take place on capital rather than on current account.

(4) Going in the other direction, the variability of nominal exchange rates under floating rates has still been considerably smaller than the variability of some other asset prices. For example, Bergstrand (1983) documents that average absolute monthly changes in nominal exchange rates for the seven major currencies over the 1973–83 period were typically much smaller than changes in national stock market prices, changes in either short-term interest rates or in long-term bond yields, or changes in either commodity prices or prices of commodity baskets.18 Such results are consistent with the view that the period of floating rates was one of sufficient turbulence to make all asset prices (not just exchange rates) fluctuate substantially.

(5) Of direct interest for making the transition from exchange rate variability to exchange rate uncertainty, most exchange rate changes under floating appear to have been unexpected. The evidence for this conclusion is that market indicators of the expected exchange rate, such as the forward rate, have typically turned out to be poor predictors of the actual exchange rate at the time of maturity of the forward contract.19 Further, the forecast errors of such market indicators of expected exchange rate changes have proved to be greatest during the periods when exchange rates have shown the greatest short-run variability (IMF (1982 a,, p.44)). Furthermore, there is evidence that the costs of buying insurance in the forward market against such exchange rate uncertainty increase with the degree of exchange rate uncertainty, with bid-ask spreads in the forward market widening appreciably when exchange rate variability is unusually large (IMF (1982 a, p. 46)). Once more, all of this is consistent with an “asset market view” of exchange rate determination whereby “. . .exchange rates will not adjust slowly and smoothly, but like other asset prices, will display random fluctuations in response to new information that is continually being received by the market.” (Frenkel and Mussa (1980, p. 377))

Official Intervention and the Demand for Reserves

Official intervention in exchange markets has not gone away and 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.

This is another area where actual behavior under floating rates has belied anticipation. Prior to the advent of floating rates, many observers expected that (a) countries would choose to intervene much less in exchange markets so as to secure the extra degree of freedom in policy implementation afforded by not having to defend an exchange rate target, and (b) the increased flexibility of exchange rates itself would permit a significant economy in the demand for reserves because the equilibrating and insulating properties of the exchange rate would reduce the size of payments imbalances.

Although official intervention in exchange markets can be difficult to measure,20 there is now little doubt that official intervention has been very substantial under floating rates—perhaps even greater than under fixed rates (e.g., see Williamson (1976) and Suss (1976)). For example, for the Group of Ten countries21 plus Switzerland, Black (1979) reported that reserve changes averaged about $3.8 billion a month in 1973, $3.3 billion a month in 1974, and $2.6 billion a month in 1975. After that, changes got progressively larger, rising to $4.4 billion a month in 1976, $5.1 billion a month in 1977, and $6.3 billion a month in 1978. Similarly, Lamfalussy’s (1979) figures on gross foreign exchange market intervention by Western central banks show a steady growth in such intervention from $36 billion in 1973/74 to $118 billion in 1978/79. Further, partial data on intervention for 1980, 1981, and 1982—published by the Deutsche Bundesbank (1982) and the Board of Governors of the U.S. Federal Reserve System (1983)—suggest that intervention was substantial for the largest industrial countries (except for the United States) during this period.22

It is not necessary here to enumerate all the reasons why industrial countries have found it useful to intervene so heavily in exchange markets during the past ten years. It is interesting to observe, however, that the recent report of the Working Group on Exchange Market Intervention (1983, pp. 9–10) indicates that such intervention has not been aimed solely at “countering disorder” or at “smoothing day-to-day movements above a certain absolute size,” or even at “leaning against the wind.” Instead, it has also included, inter alia, the objectives of “resisting rate movements ‘which bear no relation to the fundamentals’,” of “resisting depreciation out of concern over its inflation consequences or resisting appreciation in order to maintain competitiveness,” and for European Monetary System participants “. . . to keep rates within parity bands.” All in all, the 1970 prediction of the Fund’s Executive Directors that “. . . national authorities could not be expected in modern conditions to adopt a policy of neutrality with respect to movements in an economic variable of such importance to the domestic economy as the rate of exchange…” (IMF (1970, p. 42)) has proved accurate.

Turning to the overall demand for reserves, several empirical studies (e.g., Heller and Khan (1978) and Frenkel (1983 a)) point to two important conclusions:

(1) A statistically significant shift in the demand for reserves by industrial countries took place about 1972–73, that is, about the time of the move to managed floating. A shift occurs in the sense that some of the factors assumed to determine the demand for reserves (e.g., the variability of payments imbalances, the propensity to import, etc.) have different effects (estimated coefficients) on reserve demand during the period of floating rates than during the period of adjustable par values.

(2) Going the other way, the net effect of these parameter shifts across exchange rate regimes was apparently not very large because the demand-for-reserves equation as fitted to data for the period of fixed rates (1962–72) yields good forecasts of actual reserve demand by industrial countries for the 1973–79 period of floating rates (Frenkel (1983 a)).

Thus, whatever the changes in the reserve supply mechanism between the periods before floating rates and during floating rates,23 the demand for reserves does not seem to have been dramatically affected by the fact that exchange rates are now more flexible than they were previously.

To sum up, the present exchange rate system is one in which there is great diversity in exchange arrangements, in which a stable system of exchange rates (but not necessarily unchanging exchange rates) is expected to be pursued by the application of sound macroeconomic policies at the national level, in which both nominal and real exchange rates in practice have shown a lot of short-term and long-term variability, and in which official intervention in exchange markets has been frequent and substantial. With this outline of the exchange rate system in mind, it is appropriate to enquire next about the economic environment and its influence upon the system over the past ten years.

The Operating Environment

Table 1 compares inflation, unemployment, real income growth, unused industrial capacity, and labor productivity for the seven largest industrial countries during the period of floating rates (1973–82) and the last decade of adjustable par values (1963–72).24 This comparison makes it exceedingly clear that economic performance in the major industrial countries has been far worse during the past ten years of floating rates than during the preceding decade of the adjustable peg system. On average, inflation rates have been more than twice as high, unemployment rates almost twice as high, real income growth less than half as rapid, output gaps in manufacturing more than twice as large, and growth in labor productivity only half as rapid. If, therefore, macroeconomic policy were to be judged only by what happened on the “bottom line,” the conclusion would be inescapable that such policy has been much less successful under floating rates.

Table 1.

Summary Indicators of Macroeconomic Performance: Seven Largest Industrial Countries, 1963–72 and 1973–82

article image
Sources: Data on consumer prices, real gross national product, and unemployment rates are from IMF, World Economic Outlook, various annual issues. Figures on output gap in manufacturing are from Artus and Turner (1978). Data on productivity in manufacturing are from the data file of the Current Studies Division of the Fund’s Research Department.

Unweighted average.

A distinction must be made however, between the period of floating rates and the effects of floating rates themselves. Many features of the global economic environment that are important for macroeconomic performance, but are not proximately related to the exchange rate regime, were also changing during the period of floating rates. Unless these unrelated factors are properly taken into account, outcomes may be attributed to the exchange rate system that are not appropriate.

To appreciate the extent of this identification problem, it is sufficient to note just a few of the other environmental factors that could have contributed to the dismal macroeconomic statistics of the past decade. These environmental factors encompass not only external disturbances but also longer-term structural and institutional changes.

Taking the disturbances first, the most important were clearly the two rounds of large oil price increases in 1973–74 and 1979–80 and the huge expansion (57 percent) in international reserves in 1970–72 associated with the collapse of the Bretton Woods system.25 These disturbances, together with the monetary and fiscal policies adopted in their wake,26 produced important stagflationary effects in most industrial countries despite the recycling of much of the current account surplus of the Organization of Petroleum Exporting Countries (OPEC) and the avoidance of competitive devaluation among the oil importing countries.27 From 1975 to 1982, policymaking authorities in industrial countries had to face high inflation and high unemployment simultaneously.

With respect to structural and institutional changes over the past decade, the list is much longer and more varied. Some of the more important include:

(1) A slowdown in the shift of output and employment away from low-productivity sectors (agriculture, services) toward high-productivity ones (principally, industry) (Denison (1983); Giersch and Wolter (1983));

(2) The reduced opportunities (relative to the 1950s and 1960s) for technological catch-up by Europe and Japan to the United States (Lindbeck (1983 b); Giersch and Wolter (1983));

(3) The spread of indexation of wages and salaries in response to the high and variable rates of inflation (Braun (1976)), with its implications for the magnification of supply shocks (Fischer (1977)) and the reduced scope for real exchange rate changes (Modigliani and Padoa-Schioppa (1978));

(4) The changes in both the growth and demographic composition of industrial country labor forces, with their effects on both unemployment rates and labor productivity (Haveman (1978) and Medoff (1983));

(5) The fall in the profitability of firms and the contemporaneous slowdown of investment growth (Giersch and Wolter (1983));

(6) The increased demand by societies for greater equity and equality at the expense of faster output growth (Okun (1975));

(7) The adoption of preannounced money supply targets by several industrial countries in response to both high inflation and large exchange rate variability, with the consequent effect on, inter alia, the variability of interest rates (Bergtrand (1983) and Fund estimates);

(8) The seeming increase in international competition, spurred by, inter alia, the emergence of the newly industrialized countries as competitors in world markets for manufactured goods, the continuation of trade liberalization, and international technology transfers (Lindbeck (1983 a and 1983 b));

(9) The rapid growth of the Eurodollar market, the switch from asset to liability settlement of external imbalances for countries that are not reserve centers, the liberalization of capital controls (in the United States, the United Kingdom, and Japan), and the technological advances in shifting funds across national borders (Bryant (1980) and Dooley (1980–81));

(10) A deterioration in the functioning of factor markets and in the efficiency of economic incentives, as a result of, inter alia, increases in marginal tax rates, high inflation with a nonindexed tax system, considerable asymmetries and nonneutralities in the tax treatment of different types of assets, and employment legislation restricting labor mobility (Lindbeck (1983 b)).

Changes in these environmental factors are sufficient to invalidate any simple comparison between the period of floating rates and the preceding decade in terms of economic performance. It is even possible to go further and argue that the very successes of the adjustable peg period may have carried the seeds of their own destruction. Lindbeck ((1983 a) and (1983 b)), for example, has emphasized the effect of governmental full-employment guarantees in the 1950s and 1960s on the wage and price aggressiveness of unions and firms during the 1970s. He states (1983 a, p. 17):

  • . . . if unions and firms start to raise wages and prices more aggressively than before because of a belief that the government, in order to guarantee full employment, will accommodate any cost increases by demand expansion, possibly combined with a depreciating currency, the ensuing accentuation of the inflation trend may later on force the government itself to destroy this confidence in a high level of capacity utilization and brisk expansion of demand with restrictive demand management policies and/or price controls.

Perhaps the clearest manifestation of the type of aggressive behavior outlined by Lindbeck was the sharp increase in the real wage rate that occurred in many industrial countries (particularly in Europe) during 1970–75, the lagged effects of which survived throughout the 1970s.28 By Lindbeck’s (1983 b) estimate, the OECD real wage rate in 1979 was still 8 percent higher than warranted by full-employment labor productivity. Sachs’s (1983) corresponding calculation for six major industrial countries in 1978 is 7 percent. This trend of real wages was particularly serious because it was superimposed on historically very low—in fact, often negative—real interest rates; for example, the average ex post short-term real interest rate in the seven major industrial countries was negative every year from 1971 to 1978, except one; the corresponding long-term real interest rate averaged less than 1 percent over this period. By way of comparison, the corresponding average short-term and long-term real rates for 1962–70 were 1½ percent and 2 percent, respectively.29

According to Giersch and Wolter (1983), this distortion between the two key factor prices—real wages being too high and real interest rates being too low—was associated with the following adverse side effects: (1) a decline in the propensity to save; (2) a tendency for governments to run deficits which could be financed by borrowing at low real interest rates; (3) a tendency to invest in real assets rather than financial ones (but not in shares of companies which require a lot of expensive complementary labor); (4) a bias in favor of labor-saving techniques and inventions; and (5) a tendency, therefore, to neglect capital formation for use in productive processes, and hence a slower rate of growth of the productive capital stock.

To summarize, because of the intermingling of disturbances and structural and institutional changes over the past decade, serious limits are imposed on the inferences that can be drawn about the independent effects of floating rates on macroeconomic performance. As such, it is perhaps best to resist the seductive appeal of simple comparisons of economic performance in the period of floating rates with periods of other exchange rate regimes in favor of more specific arguments about how alternative exchange rate systems affect various economic objectives. The answers will be more subjective, but they are less likely to confuse the exchange rate regime with other influences.

Lessons of the Past and Options for the Future
  • View in gallery View in gallery

    Short-Run Variability in Nominal Bilateral Exchange Rates for Five Major Currencies, April 2, 1973–November 30, 1983

    Daily percentage changes

  • View in gallery

    Nominal and Real Effective Exchange Rates in Major Industrial Countries, First Quarter of 1973 to Second Quarter of 1983 1