Is Factor Investing Worthwhile Pursuing? Practically speaking?
At its peak, factor investing promised to be a scientific solution to market uncertainty. By tilting portfolios towards proven characteristics like value, size, and momentum, investors hoped to capture systematic sources of excess return. But Andrew Chen, a financial economist at the Federal Reserve and co-creator of the Open Source Asset Pricing project, is challenging some of the field’s most cherished assumptions.
This post explores what Chen’s research really says about the viability of factor investing—particularly for retail investors—and compares his conclusions with those of other industry leaders including Eugene Fama, Cliff Asness, Rob Arnott, Dimensional Fund Advisors, and Vanguard.
Sorting the Language: Predictors, Anomalies, and Factors
A key insight from Chen is that the language of finance—particularly in asset pricing research—has become confusing and imprecise. He draws a sharp distinction between three often-interchangeable terms:
Predictors: Observable variables (e.g. book-to-market ratio, past returns, accruals) that are shown to forecast stock returns.
Anomalies: Predictors that contradict standard financial theory, often generating statistically significant alpha in backtests.
Factors: Systematic sources of return and risk that show up in formal models like the CAPM or the Fama–French 3- and 5-factor models.
Chen prefers the term ‘predictors’, as it avoids prematurely assuming economic meaning or theoretical justification. Many so-called ‘factors’ in academic papers are just variables that correlate with returns—not necessarily drivers of risk or reward.
This distinction is more than semantic. It has serious implications for investment strategy. Labelling something a ‘factor’ implies it's a fundamental building block of expected returns, whereas in reality it may be the outcome of data mining or publication bias (Chen, Lopez-Lira, and Zimmermann 2023).
Larry Swedroe (2025) supports this caution, arguing that true investment factors must meet strict criteria: they should be persistent, pervasive, robust, implementable, and grounded in a compelling risk or behavioural rationale. He warns that many anomalies fail to meet these standards but are still marketed as ‘factors’ by commercial products.
Dimensional Fund Advisors (DFA) and Vanguard also draw a hard line. DFA only incorporates characteristics into their models after they pass exhaustive robustness and cost screens. Vanguard explicitly states that ‘factor investing is not for everyone’, emphasising clarity in both terminology and expectations.
In short, Chen’s linguistic clarity is a call for conceptual discipline. Investors should not conflate a statistical fluke with a meaningful source of long-term return.
Are the Predictors Real?
Chen’s meta-research finds that over 90% of published return predictors are statistically valid—they are not just the result of random chance or poor methodology (Chen 2021). This is surprisingly optimistic, especially when compared to prior critiques of ‘data-mined anomalies’.
But there’s a catch. Predictive power falls sharply after publication. Most predictors lose 50–70% of their apparent alpha once they are documented and known in the marketplace (Chen, Lopez-Lira, and Zimmermann 2023). And many fade to insignificance within 5–10 years.
This pattern, first observed by McLean and Pontiff (2016), suggests that markets adapt quickly—arbitraging away anomalies as soon as they become public knowledge.
Chen’s further insight? Theoretical justification doesn’t help. Even predictors based on elegant economic stories—like value and profitability—show the same pattern of decay. Peer-reviewed theory, it turns out, offers little protection from diminishing returns.
Costs Matter—A Lot
Chen doesn’t stop with statistical decay. He adds another layer: real-world trading costs. His findings here are even more dramatic.
When realistic estimates of transaction costs are applied, even the strongest predictors earned zero net alpha post-2005 (Chen 2024). Strategies that look compelling in academic papers—especially those involving small, illiquid stocks or high turnover—turn out to be uninvestable in practice.
This finding is echoed by Mihail Velikov (2016), who shows that trading costs can completely erode the premiums associated with many anomalies, especially those that require monthly or quarterly rebalancing.
For retail investors, this is a game-changer. Without the scale, infrastructure, or execution skill of institutional managers, they are unlikely to capture any of the theoretical edge described in the literature.
Multi-Factor Strategies: A Better Path?
Although Chen’s research focuses on single predictors, he acknowledges that multi-factor strategies may fare better. By combining signals like value, momentum, and quality, investors can reduce portfolio turnover and benefit from diversification across return sources.
Industry leaders strongly support this view:
Cliff Asness of AQR has long argued that multi-factor investing is more stable, more cost-efficient, and more behaviourally tolerable than single-factor bets. He describes it as being ‘greater than the sum of its parts’ (Asness 2024).
Dimensional Fund Advisors integrates multiple characteristics into their core equity funds. Their structure reduces unnecessary trading and lowers cost drag—allowing them to harvest the intended premiums more effectively.
Rob Arnott of Research Affiliates supports a multi-factor approach, particularly one that is valuation-sensitive. After a prolonged stretch of underperformance, he argues that many factors now look historically cheap (Arnott et al. 2019; Arnott 2023).
Should Retail Investors Bother?
This brings us to the central question: Is factor investing worthwhile for the average investor?
Andrew Chen’s answer is cautious. He doesn’t deny that predictors exist, or that they once generated excess returns. But in today’s competitive, efficient markets—where ideas are quickly arbitraged and implementation is costly—he sees little remaining edge for most investors, especially individuals.
That said, well-constructed, low-cost, multi-factor funds from firms like DFA, AQR, and Vanguard may still offer modest premiums, particularly for investors who can stay disciplined through multi-year periods of underperformance.
For everyone else, a global market cap-weighted index fund remains the default benchmark—low-cost, diversified, and remarkably hard to beat.
References
Arnott, Rob. 2023. ‘Is Now the Time to Bet on Factor Investing?’ Research Affiliates.
https://www.researchaffiliates.com/en_us/publications/articles/881-is-now-the-time-to-bet-on-factor-investing.html
Arnott, Rob, Campbell Harvey, Vitali Kalesnik, and Juhani Linnainmaa. 2019. ‘Alice’s Adventures in Factorland: Three Blunders That Plague Factor Investing.’ Journal of Portfolio Management 45 (4): 26–36.
Asness, Clifford. 2024. The Less Efficient Market Hypothesis. AQR Capital Management.
Chen, Andrew Y. 2021. ‘Most Claimed Statistical Findings in Cross-Sectional Return Predictability Are Likely True.’ Federal Reserve Board, Working Paper.
Chen, Andrew Y. 2024. ‘Interview with Andrew Chen.’ Rational Reminder Podcast, Episode 316.
https://rationalreminder.ca/podcast/316
Chen, Andrew Y., Antonio Lopez-Lira, and Tomas Zimmermann. 2023. ‘Does Peer-Reviewed Theory Help Predict Stock Returns?’ Working Paper.
McLean, David R., and Jeffrey Pontiff. 2016. ‘Does Academic Research Destroy Stock Return Predictability?’ Journal of Finance 71 (1): 5–32.
Swedroe, Larry. 2025. ‘Nine Lessons the Market Taught in 2024.’ Alpha Architect Blog.
https://alphaarchitect.com/investor-lessons/
Vanguard. 2023. Not All Factors Are Created Equal. Vanguard Research.
https://investor.vanguard.com
Velikov, Mihail. 2016. ‘Do Transaction Costs Diminish Anomalies? Evidence from Recent Papers.’ Review of Financial Studies 29 (1): 3131–3174.