Evaluating the Efficacy of Multiple Testing Adjustments in Empirical Asset Pricing

Multiple testing adjustments have become increasingly prevalent in empirical asset pricing. The purpose of these adjustments is to correct the tendency to reject the null hypothesis erroneously. The use of such adjustments, however, involves balancing the benefits of controlling false positives against the drawbacks of reduced power. This paper critically examines the statistical underpinnings of these adjustments and the consequences of multiple testing controlling on individual tests. It highlights the deficiencies of adjusting thresholds for statistical significance when the theoretical basis for the adjustment is questionable in finance research. Overall, the study suggests that multiple testing correction may not be an effective means of defending against data mining in finance