MODELING STOCK RETURNS VOLATILITY USING REGIME SWITCHING MODELS

Abstract

This paper seeks to model the dynamic relationship between stock market returns, volatility and trading volume in both developed and emerging stock markets. Modeling stock returns volatility has a tremendous reflection of the stock market microstructure behavior. We model this relationship using GARCH model, which previously has been used and reproduced most stylized facts of financial time series data, and compare its results with those of Regime Switching and Markov-Switching GARCH. The results indicate evidence of volatility clustering, leverage effects and leptokurtic distribution for the index returns. Moreover, we find that all the three stock markets are characterized by return series process staying in low volatility regime for a long time than in high volatility regime. Markov-Switching GARCH (1, 1) model is reported to be a better model than GARCH (1, 1).