Bringing Order to Chaos: Capturing Relevant Information with Hedge Fund Factor Models

Abstract

We propose using ETF returns as proxies for tradable risk factors in hedge fund performance evaluation, identifying contemporaneously relevant risk factors from the entire universe of ETFs. Our model provides more informative estimates of alpha and beta coefficients for predicting hedge fund out-of-sample performance compared with other widely used hedge fund factor models. Portfolios of top alpha hedge funds selected by our model generate statistically significant out-of-sample performance that is substantially higher compared with portfolios selected by other models. In addition, our beta-weighted clone portfolios exhibit substantially higher out-of-sample correlations with underlying hedge funds than clone portfolios formed using alternative models.

Keywords: active investment, factor selection, ETFs, risk factor exposures, return attribution, hedge funds, alpha, beta, performance prediction

JEL Classification: G11, G23