This paper investigates the impact of event-induced variance on the effectiveness of event-study methods in detecting abnormal returns. The authors simulate an event with stochastic effects on stock returns and compare the performance of several commonly used test statistics. They find that when an event causes even minor increases in variance, the traditional methods frequently reject the null hypothesis of zero average abnormal return when it is true, but they are reasonably powerful when the null is false. To address this issue, the authors propose a 'standardized cross-sectional' test, which combines the standardized-residual and ordinary cross-sectional approaches. This test is shown to have appropriate rejection rates when the null is true and is equally powerful when the null is false. The results are robust to event-date clustering, indicating that the proposed test is a reliable alternative to traditional methods in the presence of event-induced variance.This paper investigates the impact of event-induced variance on the effectiveness of event-study methods in detecting abnormal returns. The authors simulate an event with stochastic effects on stock returns and compare the performance of several commonly used test statistics. They find that when an event causes even minor increases in variance, the traditional methods frequently reject the null hypothesis of zero average abnormal return when it is true, but they are reasonably powerful when the null is false. To address this issue, the authors propose a 'standardized cross-sectional' test, which combines the standardized-residual and ordinary cross-sectional approaches. This test is shown to have appropriate rejection rates when the null is true and is equally powerful when the null is false. The results are robust to event-date clustering, indicating that the proposed test is a reliable alternative to traditional methods in the presence of event-induced variance.