The existence of "spillover effects" in financial markets is well documented and multivariate time series techniques have been used to study the transmission of conditional variances among large and small market value firms. Earlier research has suggested that volatility surprises to large capitalization firms are a reliable predictor of the volatility of small capitalization firms. A related line of research has examined how regime shifts in volatility may account for a considerable amount of the persistence in volatility. However, these studies have focused on univariate modeling and many have imposed regime changes on a priori grounds. This paper re-examines the asymmetry in the predictability of the volatilities of large versus small market value firms allowing for sudden changes in variance. Our method of analysis extends the existing literature in two important ways. First, recent advances in time series econometrics allow us to detect the time periods of sudden changes in volatility of large cap and small cap stocks endogenously using the iterated cumulated sums of squares (ICSS) algorithm. Second, we directly incorporate the information obtained on sudden changes in volatility in a Bivariate GARCH model of small and large cap stock returns. Our findings indicate that accounting for volatility shifts considerably reduces the transmission in volatility and, in essence, removes the spillover effects. We conclude that ignoring regime changes may lead one to significantly overestimate the degree of volatility transmission that actually exists between the conditional variances of small and large firms.
- Bivariate GARCH
- ICSS algorithm