PEMODELAN MARKOV SWITCHING DENGAN TIME-VARYING TRANSITION PROBABILITY

Abstract

Exchange rate or currency is an economic variable which reflects country’s state of economy. It fluctuates over time because of its ability to switch the condition or regime caused by economic and political factors. The changes in the exchange rate are depreciation and appreciation. Therefore, it could be modeled using Markov switching with time-varying transition probability which observe the conditional changes and use information variable. From this model, time-varying transition probability and expected durations are obtained; both are very useful to explain economic growth better and more detailed. This research modeled ln return value of Indonesian Rupiah to U.S Dollars and using return value of Indonesian Rupiah to Euro as information variable. The best model is MS(2) – AR(1). Overall, the mean of transition probability from appreciation to depreciation is 0,025242 and the transition probability from depreciation to appreciation is 0,666369. Expected duration of appreciation is 39,61623 days meanwhile the expected duration of depreciation is 39,18689 days. Keywords: regime switching, Markov switching, time-varying, transition probability, expected duration