How Individual Investors Trade: Overtrading, the Disposition Effect, and Underperformance
Individual investor behavior describes the systematic patterns—overtrading, the disposition effect, and under-diversification—that distinguish retail trading from rational-agent models. In The Behavior of Individual Investors, Barber and Odean (2011) analyzed 78,000 US brokerage accounts from 1991 to 1996. The most active 20% of investors earned 11.4% annually net of costs versus 18.5% for the least active 20%—a 7 percentage-point gap. The average investor's portfolio earned a Fama–French three-factor alpha of −31.1 bps per month after costs.
What the Study Found
The most active quintile averaged annual turnover of 258% and earned a monthly three-factor alpha of −86.4 bps after costs. Individual investors realize gains at about a 50% higher rate than losses—the disposition effect. In Taiwan, aggregate individual investor trading losses equaled 2.8% of personal income and 2.2% of GDP annually. Men's annual turnover is about 80% versus 50% for women, accounting for virtually all of the gender performance gap. The average investor in the LDB dataset held only 4 stocks, creating high exposure to idiosyncratic risk.
Methodology
The primary dataset is the Large Discount Brokerage (LDB) dataset, individual brokerage accounts from a large US discount broker. The sample covers 78,000 investors over 1991–1996. Performance is measured using calendar-time portfolios and the Fama–French three-factor model. The disposition effect is estimated with Cox proportional hazard rate models, using return-since-purchase categories as time-varying covariates.
Key Statistics
| Metric | Finding | Context |
|---|---|---|
| Annual net return, most active quintile | 11.4% | Top 20% by turnover, LDB 1991–1996 |
| Annual net return, least active quintile | 18.5% | Bottom 20% by turnover, LDB 1991–1996 |
| Three-factor alpha, average investor (after costs) | −31.1 bps/month | LDB 1991–1996 |
| Three-factor alpha, most active quintile (after costs) | −86.4 bps/month | Annual turnover 258%, LDB 1991–1996 |
| Annual turnover, most active quintile | 258% | LDB dataset 1991–1996 |
| Gain realization rate vs. loss realization rate | ~50% higher for gains | Odean (1998), 10,000 accounts 1987–1993 |
| Hazard ratio: selling stock up 18–22% vs. near-zero return | 2.65× | LDB dataset, Cox hazard model |
| Hazard ratio: selling stock up 18–22% vs. down 18–22% | 1.77× | LDB dataset, Cox hazard model |
| Taiwan investor underperformance vs. market | −3.8 pps/year | Barber et al. (2009), Taiwan 1995–1999 |
| Taiwan trading losses as % of personal income | 2.8% | Barber et al. (2009), Taiwan 1995–1999 |
| Men's vs. women's annual turnover | ~80% vs. ~50% | LDB 1991–1996 |
| Average stocks held per investor | 4 stocks | LDB dataset |
| Cox proportional hazard model | h(t, x(t)) = h₀(t) exp(β₁x₁ + ⋯ + βₚxₚ) | Baseline hazard scaled by exponential of covariates |
Why This Matters
The evidence consistently shows that trading activity is the primary driver of retail investor underperformance, not bad stock selection alone. The disposition effect compounds the damage by generating unnecessary tax liabilities and causing investors to exit profitable positions too early. Under-diversification further exposes retail portfolios to idiosyncratic risk that could easily be eliminated. Investors who adopt a passive, buy-and-hold approach in low-cost diversified funds avoid all three sources of value destruction.