The Journal of finance l,..., N; N= 16), starting in January 1933(month 85, the " starting month")and up to December 1980, we now compute the cumulative average residual returns of all securities in the portfolio, for the next 36 months (the "test period, months 85 through 120), i.e from t= 1 through t= 36 We find CAR w n, t and CARLn,t. If a security s return is missing in a month ubsequent to portfolio formation, then, from that moment on, the stock is permanently dropped from the portfolio and the Car is an average of the available residual returns. Thus, whenever a stock drops out, the calculations involve an implicit rebalancing 4. Using the CAr's from all 16 test periods, average CAr's are calculated for both portfolios and each month between t=l and t= 36. They are denoted ACARw and ACARL t. The overreaction hypothesis predicts that, for t> 0, ACARw, t<0 and ACARL >0, so that, by implication, [ACARL, ACARw >0. In order to assess whether, at any time t, there is indeed a statistically significant difference in investment performance, we need a pooled estimate of the population variance in CARt ACARL,)2/2(N-1) With two samples of equal size N, the variance of the difference of sample means equals 2S/N and the t-statistic is therefore T=[ACARL:-ACARw1/2S2/N Relevant t-statistics can be found for each of the 36 postformation months but they do not represent independent evidence 5. In order to judge whether, for any month t, the average residual return makes a contribution to either ACARwt or ACARLt, we can test whether it is significantly different from zero. The sample standard deviation of the winner portfolio is equal to √∑N1( ARw-ARw2)2/N-1. Since s //N represents the sample estimate of the standard error of ARw the t-statistic equals T=ARw:/(s:/√N Similar procedures apply for the residuals of the loser portfolio B. Discussion Several aspects of the research design deserve some further comment. The choice of the data base, the CRSP Monthly Return File, is in part justified by 4 Since this study concentrates on companies that experience extraordinary returns, either positive or negative there may be some concern that their attrition rate sufficiently deviates from the rmal"rate so as to cause a survivorship bias. however this concern is unjustified. when a security delisted, suspended or halted, crsP determines whether or not it is possible to trade at the las CRSP tries to find a subsequent quote and uses it to compute a ch quote is available because the stockholders receive nothing for as minus one. If trading continues the last return ends with the last listed price