Heteroskedasticity Models on the Bombay Stock Exchange

In this article, we study conditional heteroskedasticity in a market index on the Bombay Stock Exchange, from April 1979 to March 1995. We nd strong evidence of heteroskedasticity in daily, weekly and monthly returns. The conditional variance of all three data series seem best approximated by a garch(1,1) model. The garch parameter estimates at all data frequencies exhibit strong persistence in variance. In the case of monthly returns, we nd there is seasonality in the volatility, and there is one regime shift in the level of unconditional variance of the data. Remarkably enough, after controlling for these, monthly returns are homoscedastic, and there is no persistence. Both, the regime shift and the seasonality, have clear economic interpretations. The regime shift appears in March 1985, and is associated with a sharp turn towards market-oriented economic policies among other things, this led to an enormous expansion of the domestic IPO market and secondary market trading volumes. The seasonality in the post-March-1985 period is characterised by enhanced volatility associated with each federal budget announcement in end-February. The results with weekly and daily data are not as drastic { while strong evidence of the regime shift and of seasonality is found in daily and weekly data, even after controlling for these, returns are still arch, and still exhibit a fair degree of persistence. Finally, we use our volatility models to test whether the market prices the observed heteroskedasticity using garch-in-mean models. We are unable to reject the null that higher risk is not priced. We o er qualitative arguments suggesting reasons for this behaviour and conjecture that this may change in the near future. This paper is a chapter from my Ph. D. dissertation at USC. I am grateful to Andrew Weiss for guidance and inspiration. Email address [email protected].