An empirical study of nonlinear adjustment in the UIP model using a smooth transition regression model

This study considers the nonlinear relationship between the expected exchange rate change and the interest rate differential, using STR models (ESTR and LSTR), with Sharpe ratios, interest rate differentials and exchange rate volatilities as the transition variables. The results generally conclude that UIP holds with the larger Sharpe ratio and higher exchange rate volatility regimes, which is consistent with the transaction costs and limits to speculation hypotheses. However, the interest rate differential (which is generally not used much as a transition variable) when used in this study results in a failure to support UIP in the upper regime, which suggests it is the risk not the pure return that determines the transition.