RDP 8608: Exchange Rate Regimes and the Volatility, of Financial Prices: The Australian Case 5. Conclusions

The empirical results presented for the pre- and post-float VARs containing the eight domestic and foreign financial variables suggest a number of conclusions. The most striking and important one concerns the relative volatility of Australian financial prices under the two exchange rate regimes. It was quite clearly the case that, in terms of the variance of k-step ahead conditional forecasts, the exchange rate was relatively less volatile under the fixed exchange rate regime than under the floating exchange rate regime, while the short-term interest rate was relatively more volatile under the fixed exchange rate regime. These results accord with the first of our two theoretical priors.

The second theoretical prior related to the source of these changes in volatility. The volatility in interest rates due to external impulses was expected to have been reduced, and that in exchange rates increased, in the move to a floating exchange rate regime. Our results do not support this hypothesis. There is little difference in the percentage contribution of foreign shocks to the forecast variance for either the interest rate or the exchange rate in the two periods. At most, there is a slight increase in the percentage contributed to the variance of the exchange rate by the foreign variables under a floating exchange rate regime compared to a fixed regime – especially when considered over very short forecasting horizons. The tests for Granger-causality also indicate that foreign variables (both individually and jointly) had a greater impact on the exchange rate under the floating exchange rate regime. The equivalent tests for the interest rate indicate little difference between the two periods.

The data do not support this hypothesis that external shocks have (directly) affected interest rates less and exchange rates more since the float. They do, however, suggest a change in the nature of the relationship between the exchange rate and interest rate. There was a stronger relationship between these variables under a fixed exchange rate (and capital controls) regime both in terms of the contemporaneous correlation of innovations and the tests for Granger-causality. The result of this was that a larger percentage of the forecast variance of the interest rate was due to innovations in the exchange rate under the fixed exchange rate regime, than under the floating exchange rate. The variance decompositions for the exchange rate also indicate a much weaker relationship between interest rate and exchange rate movements under the floating exchange rate regime. These results support the notion that a floating exchange rate allows a more independent monetary policy, than does a fixed exchange rate regime.

The results are, of course, derived from a reduced form model of relatively few variables. Data limitations prevent a more structural approach and the inclusion of scale (e.g., income) and relative price variables in the VARs. The extent to which our conclusion that exchange rates are relatively more volatile, and interest rates relatively less volatile, under a floating exchange rate regime, is robust to such alternative specifications may remain unknown until sufficient observations become available. It may be possible, however, to proxy some of the activity and relative price effects by adding stock price indexes to the analysis. These are available on a daily basis and probably capture some influences which are absent from the current analysis. Whether our conclusions will be confirmed by this modification remains an area for further research.