Investigating impact of volatility persistence, market asymmetry and information inflow on volatility of stock indices using bivariate GJR-GARCH

Joint dynamics of market index returns, volume traded and volatility of stock market returns can unveil different dimensions of market microstructure. It can be useful for precise volatility estimation and understanding liquidity of the financial market. In this study, the joint dynamics is investigated with the help of bivariate GJR-GARCH methodology given by Bollerslev (1990), as this method helps in jointly estimating volatility equation of return and volume in one step estimation procedure and it also eliminates the regressor problem (Pagan ,1984).Three indices of different market capitalization have been considered where, S&P BSE Sensex represent large capitalization firms, BSE mid-cap represents mid-capitalization firms and BSE small-cap index represents small capitalization firms. The study finds that there exist negative conditional correlation between volume traded and return of large cap index. There is unidirectional relation between index returns and volume traded since change in volume can be explained by lags of index returns. The relation between volume traded and volatility is found to be positive in case of large-cap index but it is negative in the case of mid-cap and small-cap indices. It is observed that there exist bidirectional causality between volatility and volume traded in all the three indices considered. Volatility is affected by pronounced persistence in volatility, mean-reversion of returns and asymmetry in market. The rate of information arrival measured by IDV(Intra-day volatility) is found to be a significant source of the conditional heteroskedasticity in Indian markets since the presence of volume (proxy for information flow) in volatility equation, as an independent variable, marginally reduces the volatility persistence, whereas presence of IDV, as a proxy for information flow, completely vanishes the GARCH effect. Finally, it is observed that volume traded spills over from large cap to mid-cap index, from large-cap to small-cap index and from mid-cap to small-cap index, in response to new information arrival