RDP 8608: Exchange Rate Regimes and the Volatility, of Financial Prices: The Australian Case 1. Introduction

The collapse of the Bretton Woods Agreement in 1973 and the subsequent failure of the Smithsonian agreement, ushered in a world-wide shift to floating exchange rates. At the time, the weight of the academic literature supported floating exchange rate regimes. However, after more than a decade of experience of (dirty) floating exchange rates, there has been an increasing interest in, and analysis of, arguments for intervention of one form or another – especially in the form of co-ordinated action by a number of countries. The reasons for this are varied. Some are purely technical (e.g., the recent development of techniques that allow for the incorporation of forward-looking expectations into models and the development of game theoretic analysis). Others are experiential, including concern over the volatility of exchange rates and interest rates, the incidence of persistent current account imbalances and the world debt situation.

Tobin (1978), for example, has argued that the advent of floating exchange rates and the subsequent increase in international capital substitutability and mobility have exacerbated the transmission of international disturbances, causing relative price volatility.[1]

On the other hand, theory suggests that when price flexibility is constrained in one part of a simultaneous system, the remaining flexible prices will initially respond more to shocks than if the constrained prices are free to adjust.[2] One implication of this argument is that the volatility of interest rates should be less (and that of exchange rates greater) under a floating exchange rate.[3]

This paper addresses the issue of the empirical volatility of exchange rates and interest rates under different exchange rate regimes. The Australian dollar was floated (and controls on capital flows removed) on 12 December 1983. If the theoretical analysis is correct, interest rates should exhibit less volatility and exchange rates more volatility under the float than under the “fixed” exchange rate regime.[4] Equivalently, interest rates should be relatively easier (and exchange rates relatively harder) to predict (in the statistical sense) undera floating exchange rate regime. Moreover, the volatility in interest rates due to external impulses should be reduced, and that in exchange rates increased, relative to a fixed exchange rate regime.

In order to analyse this question of pre- and post-float volatility of Australian interest rates and exchange rates, we adopt the atheoretical methodology of vector autogressions (VARs). There are two main reasons for taking this approach. Firstly, VARs are well suited to examination of forecasting and forecasting errors.[5] Secondly, the lack of adequate data on price and scale variables, which are typically measured quarterly, prevents a more structural approach.[6]

A VAR model is estimated for both the pre- and post-float periods, on daily data for short-term interest rates and exchange rates for Australia, the US, West Germany and Japan. This allows the forecast error variances of Australian interest rates and exchange rate to be calculated and decomposed into the parts attributable to domestic and foreign sources in each of the two periods.

The results are consistent with one of our two theoretical priors. They support the hypothesis that Australian interest rates have become relatively less volatile (and the exchange rate relatively more volatile) with the move to a floating exchange rate regime. However, there is little evidence to support the hypothesis that this change in volatility has been due to a shift in the incidence of external disturbances. Rather, the results suggest that it was due to a change in the relationship between Australian interest rates and exchange rates. Since the float, the interest rate has been relatively independent of the exchange rate; this has reduced the volatility of the interest rate. This supports the notion that a more independent monetary policy is possible under a floating exchange rate.


Subsequent work has suggested that an increase in capital substitutability and mobility can have this effect only when disturbances are monetary in nature. See, for example, Driskill and McCafferty (1982), Eaton and Turnovsky (1982) and Turnovsky and Bhanderi (1982). [1]

Such applications of Le Chatelier's principle are common in international economics. The overshooting literature stimulated by Dornbusch (1976) is an example of variability in floating exchange rates being induced by slowly adjusting wages or prices. [2]

Kenen (1985) has formulated a model which addresses this issue in the Australian context. [3]

Prior to the floating of the Australian dollar, the exchange rate was pegged to a trade-weighted index. The value of this TWI was set each morning by the Australian authorities. To simplify the exposition, we shall refer to this regime as a “fixed” exchange rate. [4]

There is some controversy over the usefulness of the VAR approach, In particular, Cooley and LeRoy (1985) argue that they are of limited usefulness for counter-factual policy analysis. Nevertheless, the participants of this debate seem to agree that VARs are useful in the realm of forecasting. [5]

This is not a major limitation. Bilson (1984), for example, shows how a vector autoregression for exchange rates and interest rates can be derived from a discrete time, two country version of the Dornbusch (1976) model. [6]