The Effects of Portfolio Construction on the Performance of Style Factor ETFs or How to Build a Style Factor ETF That Does What It Says on the Tin

Abstract

Most style factor exchange-traded funds (ETFs) use some simple heuristic method of portfolio construction, such as equal weighting or capitalization weighting. This inevitably results in inefficient portfolios that are likely to have significant unintended factor bets, and the returns to these can easily dominate the returns to the chosen style factor. The author uses a simple stock selection rule to create different style factor ETFs using a number of heuristic portfolio construction methods, including equal weighting, capitalization weighting, attribute weighting, inverse volatility, risk parity, and Markowitz optimization. The author also introduces a modified version of Markowitz optimization that identifies the most inefficient holdings in an existing portfolio and only allows limited trading in those stocks; the purpose is to gain a significant improvement in overall efficiency without too much turnover. Each of the strategies is rebalanced quarterly from the beginning of 2006 to 2019. Because the stock selection is always the same, the resulting differences in performance and turnover are due entirely to the different methods of portfolio construction.

TOPICS: Portfolio construction, exchange-traded funds and applications

Key Findings

• • Style factor ETFs purport to offer investors the opportunity to capture the factor risk premium of the target style.

• • However, if these portfolios are not constructed efficiently, the contribution of the target style factor premium to the overall return on the portfolio is very likely to be dominated by other, unintended factor bets.

• • Smart portfolio optimization can maximize the effect of the target style factor premium on the return to the ETF portfolio and minimize all other effects.