The application of combined Markov regime switching and volatility model in detecting early financial crisis in Indonesia

Abstract

The financial crisis had occurred in Indonesia in the middle of 1997 to 1998 which caused the economy of Indonesia was down suddenly. Because the crisis occurred suddenly and even unexpectedly, the Indonesian government was not ready to deal with it. Therefore, the Indonesian government needs a system that can detect crisis signals; so that, such a financial crisis can be avoided. Export and import indicators have high fluctuation and regime-changing during a crisis so that they also can be used to detect crisis signals. The volatility model is able to explain the volatility which included in the indicators, while the Markov regime switching model can explain the regime-changing. Furthermore, a method that can be used to detect a crisis is the combination of Markov regime switching and volatility model. The value of smoothed probability obtained from the combination of those models is able to detect the financial crisis in Indonesia. The results show that export and import indicators can be modelled successively using MRS-GARCH (2,1,1) and MRS-ARCH (2,1). The results obtained from the prediction of those two models show that there is no financial crisis signal in Indonesia for a year later.