Analyzing the Performance of Multifactor Investment Strategies under a Multiple Testing Framework

Abstract

This study aims at risk-scaled returns for multi-factor portfolios, many of which appear to systematically have a superior Sharpe ratio. However, given the prevalent use of standalone backtesting research designs in the finance field to evaluate portfolio strategies that are potentially based on numerous combinations of factors, the significance of data mining bias may be among the most pronounced in this area of studies. Therefore, we aim to discriminate against the competing explanation of spuriously significant risk-scaled returns. Our empirical tests provide evidence that the stock picking strategies with certain combined criteria of firm characteristics outperform both value-weighted index and smallcap value portfolio even after we adjust for the multiple testing bias. The occurrences of winning for the superior multi-factor portfolios appear to be more stable than those for the single factor portfolios over different subsamples. Moreover, the outperformance of multifactor rules is robust to alternative definitions of factors.