A Markov-Switching Model of Taka/Rupee Exchange Rate: estimation and forecasting

Abstract

This study considers the validity of a (modified) monetary exchange rate model between monthly Bangladeshi Taka and Indian Rupee exchange rate in a Markov-switching framework. To reflect the beginning of the floating exchange rate regime by Bangladesh Bank, the sample period spans from May 2003 to March 2016. Empirical results lend support for Markov-switching model in capturing the long swings in the observed exchange rate. The results also show that various monetary fundamentals (i.e., interest rate differential, inflation rate differential, money growth differential, and trade balance) are statistically significant determinants of Taka-Rupee exchange rate. It then conducts several out-of-sample forecasting performances of the Markov-switching monetary model against a random walk model. A rolling window Markov-switching model generates better forecasts than a random walk. Policy implications of the results are also discussed.