Investors seeking gains from factor-based strategies aren’t fully appreciating the potential pitfalls, according to a paper co-authored by Research Affiliates founder Rob Arnott.
The returns from these strategies may fall short of expectations due to data mining bias, crowded trades, and an unrealistic view of trading costs, Arnott wrote in a paper co-authored with Research Affiliates partners Campbell Harvey, Vitali Kalesnik, and Juhani Linnainmaa. Investors may also have a false sense of how diversified their factor portfolios are or suffer “downside shocks” far larger than they’d expect due to using “naïve” tools to manage risk, they warned.
Factor-based investing, a strategy that involves choosing stocks based on characteristics such as growth, value, and momentum, has been widely adopted by institutional and individual investors. While weighting portfolios toward certain factors can lead to bigger returns, investors need to be more aware of how they behave in changing markets, according to the paper.
“Factor investing has failed to live up to its many promises,” wrote Arnott, Harvey, Kalesnik, and Linnainmaa. “Advocates for factor investing do a disservice to themselves and those they advise by dismissing the recent disappointing performance of factors.”
[II Deep Dive: Cliff Asness: Stick With Liquid Alts]
The paper urged caution when considering how diversified a factor portfolio may be. Diversification can vanish in stressed markets because factor returns become much more correlated, the authors warned, citing the financial crisis a decade ago as an example.
“Factors did perform well during the beginning of the global financial crisis, as global stock markets crashed in the months of September and October 2008,” they wrote. “The factor crash happened after this,” with the worst factor performance seen after global stock markets had risen sharply in early March 2009.
Investors can also blunder by having a “naive view” of the way factor strategies behave, as returns can stray very far from a normal distribution, according to the paper. It’s a “dangerous misperception” to believe creating a portfolio of factors will remove extreme outliers in performance, the authors said.
In addition, the Research Affiliates group said that some of the thousands of factors that have been tested may only look good due to coding mistakes by researchers or problems with the data.
“Even if a factor has a true structural risk premium, real-world returns can disappoint once the factor becomes crowded,” they wrote. “The backtest results do not reflect the market impact of investors pouring capital into the strategy.”