A Study on Volatility Modelling of BSE Sensex

The present study has analysed the volatility in the returns of BSE Sensex for the period April 2003-March 2012. It is found that the time series data is stationary but not normally distributed. The return series is serially correlated and thus there is autocorrelation. GARCH (1,1) model is used to study the conditional variances and it is found to be a good fit model as the coefficient value is close to one. This shows that a positive/negative return leads future forecasts of the variance to high/low for a long period of time. TARCH (1,1) model is applied to analyse the leverage effect. The results show the presence of leverage effect and there is also news asymmetry in the market thereby concluding that bad news has more effect on the volatility than the good news.