EARLY DETECTION OF CURRENCY CRISIS IN INDONESIA USING A COMBINATION OF VOLATILITY AND MARKOV SWITCHING MODELS BASED ON EXPORT INDICATORS

Abstract

The currency crisis that occurred in Indonesia in 1997/1998 and 2008 had a significant impact on the Indonesian economy. An early detection system for crises is necessary to minimize the impact of such crises. One model that can detect currency crises is a combination of volatility and Markov switching models. There are several indicators that can be used to detect currency crises in a country, and one of them is exports. Research results indicate that the best combination of volatility and Markov switching models for the export indicator is MS-ARCH (2,2) with an assumption of two states. The crises of 1997/1998 and 2008 can be detected using the smoothed probability values with certain limits. Predictions for the period of July 2022-June 2023 based on the export indicator show no signs of a crisis in Indonesia.