Nabil Bouamara

will give a presentation on

Sequential Cauchy Combination Test for Multiple Testing Problems with Financial Applications

Abstract:

We introduce a simple tool to control for false discoveries and identify individual signals when there are many tests, the test statistics are correlated, and the signals are potentially sparse. The tool applies the Cauchy combination test recursively on a sequence of expanding subsets of p-values and is referred to as the sequential Cauchy combination test. While the original Cauchy combination test aims for a global statement over a set of null hypotheses by summing transformed p-values, the sequential version determines which p-values trigger the rejection of the global null. The test achieves strong familywise error rate control and is less conservative than existing controlling procedures when the test statistics are dependent, leading to higher global powers and successful detection rates. As illustrations, we consider two popular financial econometric applications for which the test statistics have either serial dependence or cross-sectional dependence: monitoring drift bursts in asset prices and searching for assets with a nonzero alpha. The sequential Cauchy combination test is a preferable alternative in both cases in simulation settings and leads to higher detection rates than benchmark procedures in empirics.