A Note on Conditional Heteroskedasticity in the Market Model

Publisher Website
Google Scholar
Abstract
We examine the usefulness and implications of modeling conditional heteroskedasticity in market model residual returns. Autoregressive conditional heteroskedasticity (ARCH) plays a key role in our approach. To highlight the salient issues, we first provide a case study of one firm, Winn-Dixie Stores. Formal testing procedures reveal strong ARCH effects. ARCH models are then estimated and used to infer the pattern of time-varying volatility; differences in parameter estimates caused by use of the fully efficient estimator are also noted. Next, we provide a systematic examination of the entire New York Stock Exchange (NYSE) market. We find ARCH in roughly 25 percent of NYSE firms; moreover, we cannot reject the null hypothesis that the firms that display ARCH are unrelated to those with unconditional heteroskedasticity. The percentage of firms displaying some form of heteroskedasticity therefore appears quite large.