Shorter Papers and Comments The Causes of Real Exchange Rate Variability
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

Historically, real exchange rate variability has been higher under flexible than under fixed exchange rates. Two explanations have been put forward. The neo-Keynesian view focuses on the interaction of volatile nominal exchange rates with sticky prices. The neoclassical approach maintains the flexible price assumption and regards an increased incidence of real shocks as the culprit. Using spectral decomposition, we examine the evidence on sticky prices for the interwar period. Prices and exchange rates are found to be equally flexible, contrary to the sticky price view and consistent with the instantaneous equilibrium approach. [JEL E31, F33, N24]

Abstract

Historically, real exchange rate variability has been higher under flexible than under fixed exchange rates. Two explanations have been put forward. The neo-Keynesian view focuses on the interaction of volatile nominal exchange rates with sticky prices. The neoclassical approach maintains the flexible price assumption and regards an increased incidence of real shocks as the culprit. Using spectral decomposition, we examine the evidence on sticky prices for the interwar period. Prices and exchange rates are found to be equally flexible, contrary to the sticky price view and consistent with the instantaneous equilibrium approach. [JEL E31, F33, N24]

A Heretic View from the Interwar Period

Historically, real exchange rate variability has been higher under flexible than under fixed exchange rates. Two explanations have been put forward. The neo-Keynesian view focuses on the interaction of volatile nominal exchange rates with sticky prices. The neoclassical approach maintains the flexible price assumption and regards an increased incidence of real shocks as the culprit. Using spectral decomposition, we examine the evidence on sticky prices for the interwar period. Prices and exchange rates are found to be equally flexible, contrary to the sticky price view and consistent with the instantaneous equilibrium approach. [JEL E31, F33, N24]

The move from fixed to flexible exchange rates in 1973 was widely expected to lead to a significant decline in real exchange rate variability, as smoothly adjusting nominal rates replaced the occasional large and disruptive changes in par values of the Bretton Woods system. The experience since 1973 refutes the optimistic expectation: real exchange rate variability has increased significantly.

I. Nominal Exchange Rate Variability: Two Views

The gap between classical theory and reality has spawned a new (neo-Keynesian) orthodoxy, which abandons the assumption of instantaneously flexible goods market prices underlying the classical view.1 Flexible exchange rates determined in volatile “news”-driven asset markets combine with sticky-goods market prices to generate highly variable real exchange rates: “The evidence strongly indicates that these differences in the behavior of real exchange rates are intrinsically related to the relative sluggishness of the adjustment of national price levels, in comparison with the rapid adjustment of prices determined in highly organized asset markets” (Mussa (1990, p. 28)). The neo-Keynesian view rejects the classical notion of nominal exchange rate regime neutrality; according to this view, the increased variance of the real exchange rate is caused by the shift from fixed to flexible exchange rates.

Although enjoying broad support, the neo-Keynesian approach has not gone unchallenged. The neoclassical riposte points out that nominal exchange rate neutrality addresses the invariance of the time-series properties of equilibrium real exchange rates with respect to the nominal exchange rate regime. If the move from fixed to flexible exchange rates is accompanied by an increased incidence of real shocks, a higher variability of (equilibrium) real exchange rates is consistent with nominal exchange rate regime neutrality (see Stockman (1983, 1988)).

The question of nominal exchange regime neutrality is of more than academic interest: the neo-Keynesian orthodoxy predicts that nominal shocks are immediately reflected in asset market prices but not in goods market prices, resulting in transitory real exchange rate “overshooting” and welfare costs potentially large enough to motivate the return to some form of fixed exchange rate regime (see Baldwin (1988) and Dixit (1989)). The significant welfare implications warrant an empirical study of nominal exchange rate regime neutrality. In the following sections we provide such an analysis for the interwar period.

II. Methodology

A direct test of the two competing hypotheses requires the calculation of high-frequency equilibrium real exchange rates. Given the lack of consensus within the profession even about low-frequency equilibrium rates, progress along this line of research appears doubtful.

We instead test the crucial assumption of relatively more sticky-goods market prices differentiating the two views. The test is based on a spectral decomposition of the total variability of price and exchange rate series by cycle length: prices are sticky relative to exchange rates if the proportion of the variance of prices explained by long-run movements exceeds the corresponding proportion for exchange rates.

III. Empirical Analysis

We examine the evidence for relative price level sluggishness from 1921 to the abandonment of the gold standard by the United Kingdom in 1931. The period comprises five years of (almost) free floating with six years of a (dirty) gold standard.2 The data set consists of the 12 non hyperinflationary countries for which both exchange rates and wholesale price data were available on a monthly basis from January 1921 onwards.3 The data are taken from Tinbergen (1931, 1936), the League of Nations Monthly Bulletin (various issues), and the Federal Reserve Board Banking and Monetary Statistics (1943). Nominal exchange rates and foreign prices are effective, based on the average export shares in 1921, 1924, 1927, and 1931.

A problem arises in the timing of the two regimes. Unlike the Bretton Woods system, the interwar gold standard lacked both a definite beginning and an unambiguous ending point. The return to gold extended from the early resumption of convertibility by the United States to the adoption of a gold peg by Japan at a time when the standard had already begun to crumble. The abandonment of gold likewise is spread over time, beginning with the imposition of convertibility restrictions in 1929. In this environment the definition of the nominal exchange rate regime for any particular country can either be based on the regime adopted by the country itself or on the regime adopted by most of its trading partners. We use the former definition for the starting point of the gold standard but end our sample uniformly in September 1931, the month in which the sterling block abandoned the gold standard.

Table 1.

Coefficient of Variation

article image

Analysis of Aggregate Variance

Table 1 reports the coefficient of variation of nominal and real exchange rates and relative prices (see also Eichengreen (1988,1989)). The table confirms the stylized fact of higher real exchange rate variability under flexible exchange rates. In contrast to the simple sticky price view, exchange rates and prices exhibit comparable aggregate variability. Furthermore, price behavior is strongly regime dependent, a finding starkly at odds with the results for the post-Bretton Woods period: “ratios of national price levels typically exhibit similar, relatively smooth paths of evolution under both types of a nominal exchange rate regime” (Mussa (1986, p. 118)). The statistics are reinforced by visual impression. Figure 1 plots the log differences of nominal effective exchange rates and relative price levels for a representative group of countries. The two stylized findings of highly similar overall variances of prices and exchange rates and significant regime dependence of relative price variability are apparent.

Figure 1.
Figure 1.

Depreciation and Relative Inflation Rates

(Exchange rate and relative prices)

Citation: IMF Staff Papers 1992, 003; 10.5089/9781451973174.024.A008

Table 2 reports the decomposition of the total real exchange rate variability into fractions attributable to nominal exchange rate variability, relative price variability, and the covariance structure. Under the neo-Keynesian view, the increase in real exchange rate variability is predominantly explained by a higher variability of the nominal rate. In contrast, the table reveals that both prices and nominal exchange rates typically experienced a significant decline in variability as countries moved to the gold standard.

Table 2.

Variability Decomposition

article image
Note: All data scaled by Var (R); P and P* denote the domestic and foreign price level, respectively, and E denotes the nominal exchange rate. “Flexible” and “fixed” period for Sweden and the United States set at February 1920 to December 1926, and January 1927 to August 1931, respectively.

Spectral Decomposition

While the evidence presented to this point tentatively suggests that the distinction between sticky-goods prices and flexible-asset prices finds little support in the interwar data, measures of aggregate variability are limited in their informational content. Spectral estimation, by permitting a disaggregation of the total variance into proportions attributable to cycles of different frequencies, offers one way to sharpen the results. Tables 3 and 4 report the decompositions for nominal and real exchange rates and relative prices.4

Table 3.

Variance Proportion by Cycle Length: Real Effective Exchange Rate

article image

No regime change took place in Spain and the United States.

Table 4.

Variance Proportion by Cycle Length: Nominal Effective Exchange Rate and Prices

article image

On average, roughly 40 percent of the variation in real depreciation rates is explained by cycles lasting less than six months. A comparison across regimes reveals a surprisingly close symmetry: the relative importance of short versus long cycles does not appear to be systematically related to the nominal exchange rate regime, thus casting doubt on models explaining real exchange rate variability as misalignments caused by long swings of nominal exchange rates around stable underlying equilibrium levels.

The decompositions for effective nominal exchange rates and relative prices reveal that between 40 percent and 60 percent of inflation and depreciation variability reflects cycles of less than six months. In sharp contrast to the sticky price view, the decompositions do not differ markedly for the two variables—short- and long-run movements are of similar importance for both. The spectral decompositions thus reject the sticky price view for the interwar period; on statistical grounds no convincing qualitative differences between the two series emerge.

IV. Conclusion

The neo-Keynesian approach attributes the observed increase in real exchange rate variability associated with the move to flexible exchange rates to the interaction between sticky-goods market and volatile asset market prices. Nominal exchange rate regime neutrality fails to hold, once monetary shocks result in potentially very costly overshooting of the real exchange rates. The neoclassical view, in contrast, identifies an increased incidence of real shocks as the culprit. Nominal exchange rate regime neutrality holds, and the increased variability reflects a more volatile equilibrium real exchange rate.

We examined the empirical support for sticky prices. Our results cast serious doubt on the applicability of the neo-Keynesian approach to the interwar period: the spectral properties of prices and exchange rates differ only marginally. Although our findings are consistent with the neoclassical view, price flexibility constitutes only a necessary, not a sufficient, condition for the market-clearing approach. A convincing case for the instantaneous equilibrium school must additionally establish the equivalence between observed and equilibrium real exchange rates.

REFERENCES

  • Baldwin, Richard, “Hysteresis in Import Prices: The Beachhead Effect,” American Economic Review, Vol. 78 (September 1988), pp. 77385.

    • Search Google Scholar
    • Export Citation
  • Bernanke, Ben, and Harold James, “The Gold Standard, Deflation, and Financial Crises in the Great Depression: An International Comparison,” NBER Working Paper No. 3488 (Cambridge, Massachusetts: National Bureau of Economic Research, October 1990).

    • Search Google Scholar
    • Export Citation
  • Dixit, Avinash, “Hysteresis, Import Penetration, and Exchange Rate Pass-Through,” Quarterly Journal of Economics, Vol. 104 (May 1989), pp. 20528.

    • Search Google Scholar
    • Export Citation
  • Dornbusch, Rudiger, and Albert Giovannini, “Monetary Policy in the Open Economy,” Handbook of Monetary Economics, Vol. 2, ed. by Benjamin M. Friedman and Frank H. Hahn (Amsterdam; New York: North-Holland, 1990).

    • Search Google Scholar
    • Export Citation
  • Dornbusch, Rudiger, Federico Sturzenegger, and Holger Wolf, “Extreme Inflation: Dynamics and Stabilization,” Brookings Papers on Economic Activity: 2 (Washington: The Brookings Institution, 1990), pp. 184.

    • Search Google Scholar
    • Export Citation
  • Eichengreen, Barry, ed., The Gold Standard in Theory and History (New York: Methuen, 1985).

  • Eichengreen, Barry, “Real Exchange Rate Behavior Under Alternative International Monetary Regimes: Interwar Evidence,” European Economic Review, Vol. 32 (March 1988), pp. 36371.

    • Search Google Scholar
    • Export Citation
  • Eichengreen, Barry, “The Comparative Performance of Fixed and Flexible Exchange Rate Regimes: Interwar Evidence,” NBER Working Paper No. 3097 (Cambridge, Massachusetts: National Bureau of Economic Research, January 1989).

    • Search Google Scholar
    • Export Citation
  • Federal Reserve Board, Board of Governors, Banking and Monetary Statistics (Washington: Board of Governors of the Federal Reserve Board, 1943).

  • League of Nations, Monthly Bulletin of Statistics (Geneva: League of Nations, various issues).

  • League of Nations, Monthly Bulletin of Statistics, International Currency Experience (Geneva: League of Nations, 1944).

  • Mussa, Michael, “Nominal Exchange Rate Regimes and the Behavior of Real Exchange Rates: Evidence and Implications,” in Real Business Cycles, Real Exchange Rates and Actual Policies, ed. by Karl Brunner and Allan H. Meltzer, Carnegie-Rochester Conference Series on Public Policy, Vol. 25 (1986), pp. 117214.

    • Search Google Scholar
    • Export Citation
  • Mussa, Michael, “Exchange Rates in Theory and Reality” (unpublished; Chicago: University of Chicago, 1990).

  • Stockman, Alan C, “Real Exchange Rates Under Alternative Nominal Exchange Rate Systems,” Journal of International Money and Finance, Vol. 2 (August 1983), pp. 14766.

    • Search Google Scholar
    • Export Citation
  • Stockman, Alan C, “Real Exchange Rate Variability Under Pegged and Floating Nominal Exchange-Rate Systems: An Equilibrium Theory,” in Money Cycles and Exchange Rates: Essays in Honor of Allan H. Meltzer, ed. by Karl Brunner and Bennett T. McCallum, Carnegie-Rochester Conference Series on Public Policy, Vol. 39 (1988), pp. 25994.

    • Search Google Scholar
    • Export Citation
  • Tinbergen, Jan, Abstract of Economic Statistics (London: International Conference of Economic Services, 1931, 1936).

  • Wolf, Holger, “The Exchange Rate During the German Hyperinflation: Center Stage or Side Show?” (unpublished; Cambridge, Massachusetts: Massachusetts Institute of Technology, 1991).

    • Search Google Scholar
    • Export Citation
*

Anne-Marie Guide is an Economist in the European Department. She received her doctorate from the Graduate Institute of International Studies, Geneva, and also holds degrees from the University of Tubingen, the Kiel Institute of World Economics, and Washington University in St. Louis.

Holger C. Wolf is an Assistant Professor at the Stern Business School, New York University. He received his doctorate from the Massachusetts Institute of Technology.

The authors gratefully acknowledge comments from Rudiger Dornbusch, Hans Flickenschild, Hans Genberg, Atish Ghosh, Federico Sturzenegger, and seminar participants at the Massachusetts Institute of Technology and the 1991 European Economic Association Meetings in Cambridge.

1

See Dornbusch and Giovannini (1990, p. 1257): “[T]he dramatic effects of nominal exchange rate movements on relative prices presses the conclusion that stickiness … is an important part of the explanation.”

3

Economies undergoing hyperinflations have been excluded, since pricing behavior alters radically as inflation accelerates, rendering comparisons across time and countries meaningless. See Dornbusch, Sturzenegger, and Wolf (1990), and Wolf (1992).

4

All variables transformed into stationary series. Japan was dropped from the sample due to insufficient data points for the fixed exchange rate period.