MEASURING SECURITY PRICE PERFORMANCE USING DAILY NASDAQ RETURNS

We evaluate the performance of alternative test statistics in event st udies which include NASDAQ daily security returns. We document varying degrees of test statistic misspecification in NASDAQ samples. In part icular, we find that the commonly used standardized test statistic is misspecified in most settings. Although less pervasive, misspecificati on is also evident in the portfolio test statistic estimated using the time series of portfolio mean abnormal returns. The nonparametric ran k statistic [introduced in Corrado (1989)] performs the best overall i n NASDAQ samples; we recommend its use with market model abnormal retu rns based on an equal-weighted NASDAQ market index.