RDP 9811: Effective Real Exchange Rates and Irrelevant Nominal Exchange-Rate Regimes 1. Introduction

When the Bretton Woods system of fixed exchange rates was abandoned in the early 1970s, the subsequent high degree of short-term volatility in nominal exchange rates was largely unexpected. In addition, the short-term volatility of Real Exchange Rates (RERs) increased dramatically post-Bretton Woods. This positive relationship between the degree of flexibility of the nominal exchange-rate regime and the volatility of the RER is important for at least three reasons. First, the high degree of correlation between movements in the nominal exchange rate and the real exchange rate is consistent with the hypothesis that the prices of goods and services adjust sluggishly relative to asset prices, such as the nominal exchange rate.^[1]

Second, lower volatility of the RER is viewed as an advantage of fixed nominal exchange-rate regimes over flexible regimes. In particular, lower volatility of the RER implies greater certainty about this important relative macroeconomic price.

Third, given that the RER is a key relative price in the determination of many real macroeconomic variables – including investment, consumption and trade flows – it seems odd that changes in the behaviour of the RER post-Bretton Woods were not associated with significant changes in the behaviour of these other macroeconomic variables (Baxter and Stockman 1989).

Following the seminal work of Mussa (1986), there has been extensive literature analysing the relationship between nominal exchange-rate regimes and the behaviour of real macroeconomic variables.^[2] Interest in this topic has been heightened by a number of recent events, including currency crises (especially in Asia) which have prompted some countries to move towards more flexible exchange-rate regimes. On the other hand, some countries have moved towards more rigid nominal exchange rate arrangements, including those countries joining the Euro bloc, while others have gone even further by establishing currency boards (Argentina for example).

There are many studies that conclude that RERs exhibit substantially higher short-term volatility under floating nominal exchange-rate regimes than under fixed exchange-rate regimes.^[3] However, most of these studies are based on the analysis of bilateral RERs.^[4] We suggest that from a macroeconomic point of view, an analysis of the behaviour of effective RERs may be of more interest,^[5] and that effective RERs may in fact not alter their behaviour across different nominal exchange-rate regimes.^[6] For example, the top panel of Figure 1 shows the monthly changes in the bilateral RER for Denmark against the United States. The volatility in the bilateral RER increases dramatically moving from the Bretton Woods system of fixed exchange rates to the more flexible post-Bretton Woods era in the early 1970s. However, such behaviour is not evident in the effective RER for Denmark shown in the lower panel of Figure 1.^[7] In this paper we re-examine the effect of nominal exchange-rate regimes on the effective RER across a large sample of countries.

It may be that the effect of averaging bilateral RERs reduces the variance of the effective RER to such a degree that the nominal exchange-rate regime has little impact on the volatility of effective RERs. In other words, the relationship between the flexibility of the nominal exchange-rate regime and RER volatility may be substantially weaker (or perhaps even insignificant) for effective RERs compared with bilateral RERs. Such a result would have a number of implications. First, this may help to explain why the existing studies have failed to find a significant change in the behaviour of a range of macroeconomic variables (other than the bilateral RER) across different nominal exchange-rate regimes. Second, the adoption of a more flexible nominal exchange-rate regime may not be associated with greater uncertainty with regard to the effective RER. Third, it may be that the prices of goods and services adjust sluggishly, but much less so relative to effective nominal exchange rates (compared with bilateral nominal exchange rates).

Implicit in the discussion above is the proposition that the effective RER is more relevant to the behaviour of macroeconomic variables than a country's bilateral RERs. This would appear to be reasonable, except in a world where different firms (and/or consumers) are sensitive to the volatility of specific bilateral exchange rates.

However, even in this case, it may be that the effects of a depreciation of one bilateral RER on one firm are offset by the effect of an appreciation of another bilateral RER on another firm; if the effective RER remains unchanged, the aggregate effect of bilateral RER changes across firms may be insignificant. If instead, firms are affected by more than one bilateral RER then the volatility of bilateral RERs may be less relevant to firms than the volatility of the effective RER.

A feature of many existing studies of the RER behaviour is their reliance on comparisons of the Bretton Woods system and post-Bretton Woods in order to delineate fixed and floating nominal exchange-rate regimes. This is problematic for two reasons. First, there is the problem of identification – it may be that some other event is related to a change in RER behaviour, and that this event is highly correlated with the end of the Bretton Woods system. For example, such an event might be the oil price shocks which led to higher volatility in energy prices and hence, the terms of trade of many countries.^[8]

The other problem with studies that rely on the Bretton Woods era to define fixed exchange rates is that they may not tell us anything about fixed exchange-rate regimes in the post-Bretton Woods era. We suggest that the RER is likely to be less volatile during the Bretton Woods era because all countries were fixing their nominal exchange rates. In contrast, a country with a fixed exchange-rate regime post-Bretton Woods is fixing its exchange rate to currencies which are themselves under floating or managed exchange-rate regimes.

In this paper we present results using a large set of countries over the period 1978 to 1994. In contrast to data sets which are centred on the end of Bretton Woods, this data set exhibits variation in the timing of regime switches across different countries. In addition, results based on more recent experiences of countries with fixed exchange-rate regimes may be more relevant for policy-makers than the experience of fixed exchange rates during Bretton Woods.

The paper proceeds as follows. In Section 2 we use Mussa's sample of 16 industrialised countries to examine the volatility of the RER under fixed and floating regimes which are identified by the Bretton Woods and post-Bretton Woods eras respectively. We show how the use of effective RERs in place of bilateral RERs has an important impact on the results.

We provide a brief description of our data in Section 3 and discuss the problematic task of classifying countries' exchange-rate regimes. In Section 4 we present results for the effective RER across a broad sample of countries over the period from 1978 to 1994. To make a valid determination of the effect of the nominal exchange-rate regime, it is necessary to compare the behaviour of the RER across countries which have similar characteristics. It would be inappropriate to compare experiences across extreme types of economies for which we would naturally expect different RER behaviour irrespective of the nominal exchange-rate regime. We chose a subset of countries based on their inflation experience and on the variability of their growth rates. Both of these factors are likely to influence the underlying behaviour of the RER. Countries with high and variable inflation or variable rates of growth also tend to exhibit high RER volatility across all nominal exchange-rate regimes. Therefore, we concentrated our efforts on countries with both low and stable inflation and stable growth rates. Section 5 concludes with some remarks concerning the interpretation of our results.

Footnotes

The broad class of models that incorporate the assumption of price sluggishness imply that RERs should move relatively slowly under fixed nominal exchange-rates regimes (except for changes in official parities or realignments), while under floating exchange-rate regimes, the RER and the nominal exchange rate should show a high degree of correlation. [1]

See Frankel and Rose (1995) for a summary of this literature. [2]

For example, Mussa (1986), Eichengreen (1988), Baxter and Stockman (1989), and Flood and Rose (1995). See Obstfeld (1998) for a recent exposition of this result. [3]

The IMF (1984) examined the volatility of the effective RER for seven major industrial countries. [4]

This was originally suggested by Black (1986) in his comments on Mussa (1986). [5]

A country's effective RER (also known as a multilateral RER) is the trade-weighted average of bilateral RERs with the country's trading partners (see Section 3). [6]

As we demonstrate later in the paper there is a very slight but statistically insignificant increase in the variance of changes in the effective RER for Denmark post-Bretton Woods. See Section 3 and the Appendix for a description of the data. [7]

Alternatively, existing findings could be due to very different behaviour of the US RER during and after Bretton Woods, with no substantial change in the behaviour of other countries' RERs. O' Connell (1998) makes this point when talking about the time series properties of bilateral RERs versus effective RERs. [8]