EXCHANGE RATES FORECASTING: EVIDENCE FROM EXPORT DEPENDENT ECONOMIES

In this paper, evidence on exchange rate forecasting through linear and non-linear random walk model is examined for export dependent developed countries. The data span for this study ranges from 1995-2019 and data is taken over monthly and quarterly horizon. In-sample forecasting analysis is performed on times series basis and on the basis of in-sample forecasting evaluation it is found that exchange rate of almost one third of export-dependent developed countries is better predicted through their nonlinear random walk model (called Markov Switching Random Walk Model). The findings of the present study are momentous in many aspects for various individual agents including policy makers, portfolio managers, risk managers, financing and investment societies, stock market brokers, foreign exchange brokers, investment banks and international investors.