Testing Efficiency of the Taiwan-US Forward Exchange Market - A Markov Switching Model

The 30-day Taiwan-US forward exchange market was tested in this study. The hypothesis that the sample was drawn from a finite mixture distributions could not be rejected, which justified the use of a two-state, first-order Markov switching model. The first state rejected the null hypothesis of SEMH, whereas the second state was consistent with the hypothesis. During most of the time, the smoothed probability of state 2 was around 90%, which indicates that the efficient state has the dominant effect. This smoothed probability was observed to change over time.