A COMPARISON OF INTERVENTION AND RESIDUAL ANALYSIS

By comparing intervention and residual analysis abnormal return metrics, this study illustrates the impact that cross‐sectional return dependencies can have on empirical tests of the semi‐strong form of the efficient market hypothesis. The source of the dependencies encountered in this study appears to be the result of nine value altering information releases (e.g., events) that occurred in a seven‐month period in 1982. For one sample of bank equity returns, the abnormal return metrics obtained with residual analysis are generally smaller than those obtained with intervention analysis. In addition, the intervention analysis results suggest that two of the nine events led to significant shifts in the banks' systematic risk. These findings suggest that tests of the semi‐strong form of the efficient market hypothesis conducted with a less‐than randomly constructed sample need to examine the restrictions that accompany the specification of an expected return metric.