Study of Nonlinearities in the Dynamics of Exchange Rates: Is There Any Evidence of Chaos?

This paper studies the dynamics of six major exchange rates, and runs formal tests to distinguish among different types of nonlinearities. In particular we study exchange rate returns, normalized exchange rates and exchange rate volatilities, classifying these series using BDS test, correlation dimensions and maximum Liapunov exponents. Estimates of dimension indicate high complexity in all series, suggesting that the series are either stochastic processes or high dimensional deterministic processes. Though we obtain a number of positive estimates of Liapunov exponent, they are quite small and it seems more appropriate to interpret them as indicating stochastic origin of the series.