Men Trade 45% More Than Women and Earn Less for It
Overconfidence causes investors to overestimate the precision of their knowledge, leading to excessive trading. Barber and Odean (2001) test this prediction in Boys Will Be Boys: Gender, Overconfidence, and Common Stock Investment. Using 37,664 U.S. discount brokerage accounts from February 1991 through January 1997, they found men traded 45 percent more than women. Trading reduced men's net returns by 2.65 percentage points per year versus 1.72 percentage points for women.
What the Study Found
Men turned their common stock portfolios over approximately 77 percent annually; women turned theirs over approximately 53 percent. Men's net own-benchmark returns fell 0.221 percent per month; women's fell 0.143 percent per month — a gap of 0.94 percentage points annually. Single men traded 67 percent more than single women and reduced their returns by 1.44 percentage points per year more. After controlling for market risk and firm size, women earned net returns 9 basis points per month higher than men. The stocks men bought underperformed the stocks they sold by 20 basis points per month; the same gap for women was 17 basis points.
Methodology
The data came from 78,000 household accounts at a large U.S. discount brokerage firm. Gender was identified for 37,664 households using demographic data from Infobase Inc. as of June 1997; 79 percent were male-headed accounts. The study period ran from February 1991 through January 1997, yielding 72 months of monthly portfolio and trade data. Key controls included age, marital status, income, presence of children, market beta, and the Fama-French SMB size factor.
Key Statistics
| Metric | Finding | Context |
|---|---|---|
| Men's annual portfolio turnover | ~77% | All male households, Feb 1991–Jan 1997 |
| Women's annual portfolio turnover | ~53% | All female households, Feb 1991–Jan 1997 |
| Men trade more than women | 45% more | All households |
| Net return drag, men | 2.65 pp/year | vs. own beginning-of-year portfolio |
| Net return drag, women | 1.72 pp/year | vs. own beginning-of-year portfolio |
| Gender performance gap (net) | 0.94 pp/year | Men underperform women |
| Single men vs. single women, trading | 67% more | Single households only |
| Single men net return drag | 1.44 pp/year more | vs. single women |
| Risk-adjusted gender gap (2-factor) | 9 bps/month (1.1%/year) | Market + SMB controlled; women outperform |
| Men's purchases vs. sales | −20 bps/month | Stocks bought underperform stocks sold |
| Net return formula | (1 + R_net) = (1 + R_gross)(1 − c_sell)/(1 + c_buy) | Adjusts gross return for bid-ask spread and commissions |
Why This Matters
By partitioning on gender, the study achieves a sharper test of overconfidence models than approaches relying on self-reported investor beliefs. The large sample size allows demographic controls that confirm the gender effect is not a proxy for income, age, or marital status differences. Risk-aversion differences between genders exist but would predict different portfolio compositions, not necessarily different trading frequencies — overconfidence uniquely explains the frequency channel. Practitioners designing trading cost structures or investor education programs have direct empirical support for targeting transaction frequency as a performance lever.