Abstract
Traditional studies of long-run stock price abnormal performance after corporate events compare the mean returns of an event firm portfolio and a benchmark firm portfolio or index. However, it is well known that long-run abnormal returns are non-normal leading to problems with statistical inference on abnormal performance. Instead in this paper, the entire return distributions of event firms and the benchmark index using non-parametric tests of stochastic dominance are compared. Tests are applied for first and second order stochastic dominance to Ritter's (1991) IPO data. It is found, contrary to results that compare only mean returns, that IPO firms do not underperform a benchmark index. The results are robust to extreme values of buy-and-hold return of IPO firms and underline the fact that long-run abnormal performance measurement is sensitive to the methodology used.