Dan Page
As investors have caught on that actively managed funds typically fail to outperform index funds with a similar benchmark, the active market’s share of total assets has shrunk from about 90 percent of the pie at the turn of the century to 50 percent today.
Fees represent a potent headwind for actively managed funds, as it can be hard to generate returns that make up for an annual average charge that is 0.60 percentage points higher, on average, than that for index funds.
But active managers face yet another challenge in generating market-beating performance, suggests research by Chicago Booth PhD student Benjamin Marrow and Booth’s Stefan Nagel. Strategies discovered to exploit market anomalies—such as momentum investing—are not permanently effective. Moreover, the attention generated by academic research papers seems to further undermine these strategies over time.
Marrow and Nagel focused on methods that predicted a stock’s performance for the upcoming month using the previous 10 years’ worth of its monthly returns. With nearly 100 years of data, they studied whether the past and current price data available to an investor at any point in history could have produced a useful signal.
This real-time focus differs from much of the modern-day analysis in academic studies, in which researchers typically analyze a full dataset—including information that wasn’t available to an investor at the time of their decision—to identify a pattern that is presumed to be a permanent feature of the market. Marrow and Nagel, by contrast, continuously updated the analysis with contemporaneous data and assumed that anomalies can change over time.
They created a model that is flexible, they explain. Rather than preselecting strategies that are popular with investment practitioners and researchers, which could be contaminated with the benefit of hindsight, the model entertains even complex patterns that may seem arbitrary. For example, it may predict that a stock will do well next year if it did well last year, badly the year before, and middling the year before that.
With this approach, it identified in the data three well-known trading strategies: long-term reversal, momentum, and short-term reversal. Marrow and Nagel didn’t restrict their analysis to these strategies, but the model found that these anomalies were identifiable in real time, long before being popularized in published studies. It also indicated that the effectiveness of each of these as trading strategies diminished as more investors caught on. The increased demand for stocks that exploited them caused price increases, which in turn reduced potential future returns, the researchers write.
The rise and fall of stock-return predictability
For example, in 1985, DePaul’s Werner De Bondt and Booth’s Richard H. Thaler published an influential paper in the Journal of Finance in which they argued that investors overreact to good and bad news. Their “long-term reversal” effect made a case for being a contrarian: Stocks that had done poorly over the prior 3–5 years outperformed over the ensuing 3–5 years, while those that had done well underperformed.
Marrow and Nagel’s model demonstrates that this anomaly was evident in real-time by the early 1970s, when stocks that had lagged or outperformed for the past 1–6 years were indeed likely to reverse course. But this anomaly became less effective after 1985, though it periodically has become more potent. The net takeaway is that the strategy doesn’t work permanently.
Momentum investing went mainstream in 1993, when Emory’s Narasimhan Jegadeesh and University of Texas’s Sheridan Titman published an article, also in the Journal of Finance, demonstrating how stocks that had done well or poorly over the prior 3–12 months continued on those trajectories.
Marrow and Nagel confirm that this pattern was evident as early as the 1960s. But after Jegadeesh and Titman’s study, shorter-term momentum weakened as the window for outperformance shifted to 7–12 months, in line with research by University of Rochester’s Robert Novy-Marx. The pattern continued to soften, and Marrow and Nagel find it had pretty much petered out by 2008.
The researchers ran through the same exercise for a related momentum effect detailed in a 1990 Jegadeesh study, which documented the predictive power of one-month reversals for stocks. This anomaly could be seen as early as the 1950s, according to their model. But here, too, the strategy grew less effective as investors slowly caught on to the pattern, and weakened further following the paper’s publication in the Journal of Finance, they find.
“Based on these findings, there is little justification for viewing momentum, long-term reversal effects, or other return-based anomalies as permanent features of the cross-section of expected stock returns,” the researchers write. Technology has only sped up the learning curve for investors, they explain, suggesting an even shorter lifespan for other effective anomalies.
- Werner De Bondt and Richard H. Thaler, “Does the Stock Market Overreact?” Journal of Finance, July 1985.
- Narashimhan Jegadeesh, “Evidence of Predictable Behavior of Security Returns,” Journal of Finance, July 1990.
- Narashimhan Jegadeesh and Sheridan Titman, “Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency,” Journal of Finance, March 1993.
- Benjamin Marrow and Stefan Nagel, “Real-Time Discovery and Tracking of Return-Based Anomalies,” Working paper, October 2024.
- Robert Novy-Marx, “Is Momentum Really Momentum?” Journal of Financial Economics, March 2012.
Your Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.