By Avanidhar Subrahmanyam, University of California, Los Angeles and Sheridan Titman, University of Texas at Austin
- Overconfidence has two aspects: overassessing one’s own information and under-assessing (being skeptical of) others’ information.
- Skeptical traders cause excessive liquidity provision to other informed investors and cause prices to react to stale information; they thus cause momentum.
The relative performance of stocks in the past three to twelve months tends to predict their relative performance in the next three to twelve months. This is generally referred to as the momentum effect. This return pattern was first documented in U.S. equity markets in the early 1990s (Jegadeesh & Titman, 1993). Subsequent research found it to work across several countries (Chui et al., 2010) and in markets for other assets such as government bonds, commodity futures, and foreign currencies (Asness et al., 2013).
The easiest way to understand momentum is as follows. Sort stocks into, say, five or ten portfolios based on their performance in the past three to twelve months. (Any horizon in this interval will work.) This is the portfolio formation period. Then, look at the performance of these portfolios in the next three to twelve months (the holding period). The typical finding is that the performance of the portfolios in the holding period follows the same rank order as that in the formation period. A standard regression analysis also yields the result that past three- to twelve-month returns cross-sectionally predict future returns for up to twelve months.
Momentum does not work well in all market settings. For example, it works better during bull markets and during periods of positive investor sentiment. In the U.S., the momentum effect has been stronger for smaller firms and for growth firms. It also tends to work better in North America and Europe than in Asia. It has failed spectacularly on occasion, typically during recoveries from crises, such as immediately after the Great Depression and the recent global financial crisis. In both cases, momentum generated negative returns because losers, coming out of the crisis, outperformed winner stocks, reversing momentum rankings. Early research also showed that momentum returns tend to reverse when the holding period is extended to two years and beyond. However, this reversal tendency has attenuated in the 1990s and beyond, though the six- to twelve-month momentum effect has persisted.
Why should a strong return pattern such as momentum persist? Empirical attempts at capturing momentum with standard factor models have achieved little success, so a natural alternative rationale is that investors make mistakes, perhaps due to behavioral biases.
Our recent paper (Luo et al., 2020) argues that momentum arises because investors tend to be overconfident. To us, overconfidence takes two forms. First, overconfident investors believe their ability is better than it actually is. Second, overconfident investors are skeptical, or in other words, underestimate the ability of the other investors in the market.
To understand this, let’s assume that there is private value-relevant information about a company and our overconfident investors are actually slow to receive the information. Indeed, other investors receive the information sooner, and trade on the information. However, our slower, overconfident investors tend to believe that if they have not received information that others have also learned little of consequence (i.e., that they have only observed a very noisy version of their information). They therefore provide too much liquidity to those early informed investors. In other words, the prices at which markets clear are too favorable to informed investors: they are too low when the informed have positive information and vice versa, relative to prices in a rational economy.
Thus, under skepticism, the informed trades initially move prices too little. However, when skeptical investors subsequently learn the information, the price moves further in the same direction. There is, therefore, momentum on average, and since overconfident late-informed investors also think their own information is more precise than it actually is, they trade too aggressively, which causes overreaction and long-run reversals.
Skepticism also causes late-informed investors to react to stale information. Specifically, we show that if skeptical investors completely discount the possibility that part of the information has already been traded upon, then they will react to information already in the market price. This implies that the price overshoots fundamentals by repeatedly reacting to the same signal. This also creates momentum, which also is subsequently reversed.
Our theory is consistent with recent research that shows that momentum profits are high when aggregate market liquidity is high (Avramov et al., 2016). Within our setting, when there is greater skepticism, as-yet uninformed but skeptical investors provide more liquidity to the informed investors. This implies that if skepticism varies over time, periods of higher skepticism will coincide with higher liquidity, and we will observe a positive relation between momentum and liquidity.
Our theory is also consistent with the observation that long-run reversals disappeared in recent years while momentum remained. We show that this will happen if information disclosures accelerate, as speedier disclosures tend to mitigate the effects of long-term overreaction. Indeed, one would expect that with the advent of the Internet, information would spread more quickly through the investment community. We also provide insights about why momentum fails (or “crashes”) following financial crises. Such crises are accompanied by large quantities of panic-driven sales (possibly by retail investors) on poorly performing stocks, which are not driven by fundamental information. These sales have to be accommodated by risk-averse investors who demand a premium for accommodating these sales, which means that the market clears at lower prices than those based on fundamentals alone. This premium is temporary until the market makers are able to unwind their positions. As this unwinding occurs, the losers tend to appreciate in value, so that momentum strategies which are based on going long on winners and short on losers fail on the short side. We show that this effect strongly counteracts, and in fact reverses, the tendency of skepticism to create momentum.
Finally, our model is consistent with the observation that stock prices tend to “drift” following corporate events. For example, after positive (negative) earnings or revenue surprises, they tend to drift upward (downward). Similarly, they tend to drift downward following new equity issues. Our setting provides an intuitive explanation. Investors who possess fundamental private information tend to be skeptical of public information sources as well. They tend to think that “the firm’s executives can teach me nothing I don’t already know” and therefore continue to be wedded to their opinions even after the firm conveys information via earnings or security issuances. This implies that the price reaction to the corporate event is incomplete, which leads to drift.
To learn more about these momentum price patterns, we are collaborating with Andy Chui on a study of Chinese stocks. In China, some firms issue both A shares and B shares, which transact in different currencies and attract different clienteles, even though their cash flow and control rights are identical. Our most recent research shows that momentum exists in B shares (where foreign institutions play an important role), but not in A shares (which are mainly traded by local Chinese retail investors). That momentum only obtains in Chinese B shares supports our theory that momentum is driven by informed traders, who are likely to be institutions, rather than retail investors. In contrast, Chinese A shares, which do not exhibit momentum, exhibit much stronger shorter term (monthly) reversals. This last observation is consistent with the idea that reversals arise when there is limited market making capacity available to absorb the trades of uninformed retail investors.
The recognition that overconfidence implies both over-assessment of one’s ability and under-assessment of that of competitors offers much promise in understanding momentum and other patterns in financial markets. It also suggests ways in which skepticism can be counteracted. First, more precise disclosures can help reduce the impact of investors’ overconfidence because they accelerate convergence of prices to fundamentals. Further, sell-side analysts, by directly disclosing information sooner, can help mitigate momentum that arises from the skeptical investors’ tendency to react to stale information.
About the Authors
Dr. Avanidhar Subrahmanyam is a Distinguished Professor of Finance and the Goldyne and Irwin Hearsh Chair in Money and Banking in the Anderson School of Management at the University of California, Los Angeles. He has served as an associate editor of Review of Financial Studies and Journal of Finance. He is also a founding editor of the Journal of Financial Markets. Dr. Subrahmanyam also belongs to the National Bureau of Economic Research’s Working Research Group on Market Microstructure. His research interests include stock market momentum and behavioral finance.
Dr. Sheridan Titman is a professor of finance and the Walter W. McAllister Centennial Chair in Financial Services in the McCombs School of Business at the University of Texas at Austin. At UT, he also directs the Energy Management and Innovation Center. With research interests that include investments and corporate finance, his work has appeared in publications such as the Journal of Finance, Journal of Corporate Finance, and the Journal of Financial and Qualitative Analysis. Dr. Titman also serves as a research associate with the National Bureau of Economic Research.
Asness, C., Moskowitz, T. J., & Pedersen, L. H. (2013). Value and momentum everywhere. Journal of Finance, 68, 929-985.
Avramov, D., Cheng, S., & Hameed, A. (2016). Time-varying liquidity and momentum profits. Journal of Financial and Quantitative Analysis, 51, 1897-1923.
Chui, A. C. W., Titman, S., & Wei, K. C. J. (2010). Individualism and momentum around the world. Journal of Finance, 65, 361-392.
Jegadeesh, N., & Titman, S. (1993). Returns to buying winners and selling losers: Implications for stock market efficiency. Journal of Finance, 48, 65-91.
Luo, S., Subrahmanyam, A., & Titman, S. (2020). Momentum and reversals when overconfident investors underestimate their competition. Review of Financial Studies, forthcoming.