Luck versus Skill in Mutual Funds: Size and Power Pitfalls in Bootstrap Methods
33 Pages Posted: 8 Nov 2019
Date Written: October 25, 2019
We question the applicability of existing bootstrap methods for mutual fund performance evaluation, where the number of funds is large relative to the sample size of each fund. Theoretically, we derive the asymptotic distribution of the bootstrap test statistics by using Edgeworth expansion to characterize the approximation error in estimating alpha. We demonstrate that the test size is severely distorted as a large fraction of funds has small sample sizes and that the test power depends on the existence of both negative and positive alphas. Empirically, we design a simulation study to support our theoretical findings.
Keywords: Bootstrap, Edgeworth expansion, Estimation error, Mutual fund, Test size and power
JEL Classification: G11, G23, C58
Suggested Citation: Suggested Citation