Foreign exchange rate is important as it determines a country's economic condition. It is used to carry out transfers of purchasing power between two or more countries. Volatility in exchange rates may result in difficulty in decision making especially, in financial sectors as high volatility could increase the risk in exchange rates. Thus, Markov switching model is employed in this study as it is believed to be efficient in handling not only volatilility but also nonlinearity characteristics in exchange rates. The aims of this study are to model the foreign exchange rates using two models; Markov Switching (M-S) models and Markov Switching Generalized Autoregressive Conditional Heteroscedasticity (M-S GARCH) and to compare these two models based on log-likelihood, AIC and BIC criteria. This study used the quarterly data of foreign exchange rates for Singapore Dollar (SGD), Korean Won (KRW), China Yuan Renminbi (CNY), Japanese Yen (JPY) and the US Dollar (USD) against Malaysia Ringgit (MYR) which were collected from Quarter 4, 2006 to Quarter 1, 2018. The findings indicate that Markov Switching is the best model since it has the highest log-likelihood value, and the lowest AIC and BIC values. The results show that JPY and SGD have highly persistent trends on regime 1 with probability values 0.96 and 0.84, respectively as compared to CNY, KRW and USD, while the latter have high persistent trends on regime 2 with probability values, 0.99, 0.95, 0.82, respectively.
Can the Markov switching model forecast exchange rates?
