RDP 9111: Monthly Movements in the Australian Dollar and Real Short-Term Interest Differentials: An Application of the Kalman Filter 6. Assessments and Conclusions

Direct technical application of Campbell and Clarida's (1987) rational expectations model to Australian post-float data has eliminated real ex ante short-term interest differentials as a source of monthly variation in the real exchange rate. Our empirical results with Australian statistics are consistent with Campbell and Clarida's (1987) original findings for the United States.

The overwhelming conclusion is that, over the post-float period in Australia, the permanent component of the model, namely the expected long-run equilibrium real exchange rate, accounts for the greatest proportion of the month-to-month movements observed in the actual real exchange rate. This is because the transitory component, identified with ex ante short-term real interest differentials, does not. This result does not deny the existence of a weak short-run relationship between Australia's real exchange rate and real interest rate relative to that of the rest of the world. It does, however, indicate that given the influence of all remaining factors, this effect is not substantial enough to be responsible for the observed short-run dynamics of the real exchange rate.

We could interpret this finding as evidence that the monetary approach to real exchange rate determination, which focuses on the role of international capital flows, plays a secondary role in describing the path of the observed real Australian exchange rate. If rational expectations is a valid assumption, then this model suggests that random real shocks to variables like for example, commodity prices, determine the real value of the Australian dollar over the shorter run.

While the solution of Campbell and Clarida's (1987) model is complex, its underlying specification is both transparent and flexible. Its treatment of the unobservable explanatory variables represents an advance on earlier rational expectations models of the real exchange rate. It allows the expected long-run exchange rate to vary over time; provides for the existence of a time-varying risk premium; relaxes any dependence on the expectations hypothesis of the term-structure of interest rates; and allows the vector of explanatory variables to evolve slowly over time, as is appropriate for time-series data.

However, the success of Campbell and Clarida's methodology (although more flexible than earlier rational expectations models) relies on their assumptions about the nature of the foreign exchange market, namely that it is efficient (i.e. that expectations are rational). An efficient market in foreign exchange is one in which the majority of market participants have access to all available information which is reflected in the price; only new information moves the exchange rate (that is, the exchange rate is only affected to the extent that a particular outcome differs from the value expected by the market, even if that outcome is poor^[41]); and all new information is a shock (it is completely random and thus, unpredictable). Allan et al. (1990, p. 97) point out that consistent profits in such a market are impossible to incur since “… no forecasting procedure can give more accurate forecasts than tossing a coin can”.

While it has been both common and tractable to assume market efficiency for empirical work on the short-run determinants of exchange rates, anecdotal evidence from market participants suggests that it is inappropriate.^[42] We feel that the application of Campbell and Clarida's (1987) methodology to Australian data is compromised by the ambiguity surrounding the existence or otherwise of a risk premium. In the absence of a significant risk premium effect, their model would collapse to an uncovered interest parity relation. Such a relation has been discarded empirically. Gruen and Menzies (1991) propose that if the costs of sluggish portfolio adjustment are insignificant, then the failure of uncovered interest parity may result from the survival of near-rational agents in the foreign exchange market. This argument introduces the idea that the essential reason for the failure of uncovered interest parity may be that the assumption of rational expectations is inadequate in the foreign exchange market (Cutler, Poterba and Summers (1990)).

We believe the direction for future research into the short-run dynamics of the real exchange rate should allow for alternative expectations formation.^[43] In so far as Campbell and Clarida's approach found little explanatory power for interest differentials, we feel that this should not necessarily be interpreted as evidence that real shocks determine shorter-run exchange rate movements. Rather, the result may be indicative of the inappropriateness of rational expectations.

In conclusion, this paper has replicated Campbell and Clarida's (1987) rational expectations model of the US foreign exchange market for Australian data. As in their study, little role is found for real interest differentials in real exchange rate determination over the short run. Within the confines of their assumptions, this leaves the long-run equilibrium real exchange rate to do all the explaining of the real exchange rate.

However, given the growing body of literature which now opposes both rational expectations and efficient foreign exchange markets, alternative explanations are possible. As canvassed earlier in the paper, a number of options, some of which entail the inefficient formation of expectations, could be the driving influence behind the monthly movements in the Australian dollar. In so far as monthly movements in the exchange rate motivate resource allocation decisions, less than rational expectations imply that these decisions are suboptimal.

Footnotes

Allan et al. (1990) provide a good example by using market reaction to a monthly current account deficit figure: “… if the market predicts that this figure will be AUDI billion, this information will be reflected in the current exchange rate. [If the actual outcome is] AUD800 million, the dollar is likely to appreciate, even … [though] this is a poor result.” [41]

They argue that consistent profits have been earned from taking positions in the foreign exchange market. [42]

Miller and Weller (1991) provide a concise review of the most recent literature in this area including, for example, the investigation by De Long et al. (1990) into the survival of noise traders in financial markets. [43]