- 1Investor overconfidence causes underreaction to public information, creating momentum
- 2Self-attribution bias amplifies overconfidence over time
- 3The model explains both short-term momentum AND long-term reversal
- 4These biases are deeply rooted in human psychology and unlikely to be arbitraged away
- 5Understanding the psychology behind momentum helps investors avoid common mistakes
#The Paper at a Glance
Title: Investor psychology and security market under- and overreactions
Authors: Kent Daniel, David Hirshleifer, and Avanidhar Subrahmanyam
Published: Journal of Finance, 1998
DOI: 10.1111/0022-1082.00077
After Jegadeesh and Titman (1993) documented momentum, the academic world asked: why does it work? The efficient market hypothesis says prices should reflect all available information. If a stock's price already reflects its value, why would past returns predict future returns?
Daniel, Hirshleifer, and Subrahmanyam provided the leading behavioral answer.
#The Two Key Biases
1. Overconfidence
Investors are overconfident about the precision of their private information (their own research, analysis, and opinions) relative to public information (earnings reports, analyst estimates, news).
This means: - When private information suggests a stock is good, investors buy—but not enough to fully incorporate the signal - When public information confirms their view, they overreact - When public information contradicts their view, they underreact
The net effect: prices adjust slowly to new information, creating trends (momentum).
2. Self-Attribution Bias
This bias amplifies overconfidence over time:
- When an investment goes up (confirming the investor's thesis): "I was right! I'm a great analyst."
- When an investment goes down (contradicting the thesis): "Bad luck. The market will come around."
| Event | Attribution | Effect on Confidence |
|---|---|---|
| Stock goes up (thesis confirmed) | "My skill" | Confidence increases |
| Stock goes down (thesis contradicted) | "Bad luck / noise" | Confidence doesn't decrease |
This asymmetry means overconfidence grows over time as investors selectively take credit for wins and dismiss losses.
#How This Creates Momentum and Reversal
Phase 1: Underreaction (Momentum)
- 1New public information arrives (strong earnings, positive news)
- 2Overconfident investors who already own the stock weight their private view more heavily than the new public data
- 3Price adjusts—but not enough
- 4Over the next 3-12 months, the price gradually catches up to reality
- 5Result: Momentum — past winners continue to rise
Phase 2: Overreaction (Long-Term Reversal)
- 1As confirming information arrives, self-attribution bias kicks in
- 2Investors become even more overconfident
- 3They push prices beyond fundamental value
- 4Eventually reality reasserts itself
- 5Result: Reversal — stocks that went up too much eventually come back down (12-36 months)
This elegantly explains both phenomena: - Short-term momentum (3-12 months): Prices gradually incorporate information - Long-term reversal (2-5 years): Overreaction eventually corrects
#Evidence Supporting the Theory
Laboratory Studies
Experimental psychologists have extensively documented both overconfidence and self-attribution: - People are overconfident about predictions in virtually every domain tested - Self-attribution bias is universal across cultures - These biases are resistant to learning and experience
Market Data
The model's predictions match observed patterns:
| Prediction | Observed? |
|---|---|
| Short-term momentum (3-12 months) | Yes — Jegadeesh & Titman (1993) |
| Long-term reversal (2-5 years) | Yes — De Bondt & Thaler (1985) |
| Stronger momentum after confirming events | Yes — subsequent research confirms |
| Momentum stronger in hard-to-value stocks | Yes — Lee & Swaminathan (2000) |
#Why Momentum Won't Be Arbitraged Away
If momentum exists because of behavioral biases, why don't smart investors eliminate it?
1. The Biases Are Hardwired
Overconfidence and self-attribution aren't bugs—they're features of human psychology. Even professional investors exhibit these biases. Decades of awareness haven't eliminated them.
2. Momentum Crashes Create Fear
The strategy experiences severe crashes during market reversals (2009: -55% in three months). This crash risk scares away many would-be momentum traders.
3. Institutional Constraints
Most professional investors are evaluated quarterly. Momentum can underperform for extended periods, making it career-risky to implement consistently.
#Practical Implications
Understanding these biases helps investors in two ways:
1. Exploiting Momentum
Knowing that underreaction creates trends justifies using momentum as a factor. Trends aren't random—they reflect a systematic cognitive bias that affects all humans.
2. Avoiding the Traps
- Don't confuse luck with skill — self-attribution bias makes every investor think they're above average
- Respect public information — when earnings surprise, the market moves, but probably not enough
- Be aware of overconfidence — your private analysis is probably less accurate than you think
#How This Applies to Our Rankings
Our momentum factor (25% weight) exploits exactly the underreaction that Daniel et al. describe. By systematically buying stocks with strong 12-month returns (skipping the most recent month), we capture the gradual price adjustment that overconfident investors create.
We also combine momentum with other factors like value and profitability. This multi-factor approach helps protect against momentum crashes—exactly the reversal phase that the model predicts.
#Academic Source
Daniel, K., Hirshleifer, D., & Subrahmanyam, A. (1998). "Investor psychology and security market under- and overreactions." Journal of Finance, 53(6), 1839-1885.
Last updated: February 1, 2026