Week 35 Day 5: Overconfidence: The Most Dangerous Bias in Finance

Studies show that 74% of fund managers believe they are above average at their job. Mathematically, only 50% can be above average. After fees, fewer than 10% actually beat the market over 15 years. The gap between perceived skill and actual results is overconfidence, and it is the single most expensive bias in investing.

How overconfidence costs you money: (1) Excessive trading. Barber and Odean (2000) found that the most active traders earned 6.5% less per year than the least active traders. Why? Every trade is a bet that you know something the market does not. You usually do not. But overconfidence convinces you otherwise. (2) Under-diversification. Confident investors concentrate bets: 'I know this stock will win.' If they are right, they beat VTI. If they are wrong (more likely than they think), they dramatically underperform. (3) Ignoring base rates. You think: 'I am smart, I work hard, I can pick stocks.' Base rate: over 15 years, 92% of large-cap fund managers (teams of PhDs with Bloomberg terminals and insider access to management) underperform the S&P 500. You, with a brokerage account and some Google searches, will beat them? Overconfidence says yes. Statistics say probably not. (4) Illusion of control. Active management feels like control: you are doing something, making decisions, reading reports. Passive management feels like giving up control. But 'doing something' is not correlated with 'doing better.' Most activity is noise that generates costs. The cure: track your actual performance against VTI. Do it honestly, including all costs and taxes. After 3-5 years of underperformance data, overconfidence fades. This is why most converts to index investing are former active traders who tracked their results.

Overconfidence has three distinct components (Moore and Healy, 2008): (1) overestimation (believing your absolute performance will be better than it actually will be), (2) overplacement (believing your relative performance will rank higher than it actually does), and (3) overprecision (believing your estimates are more accurate than they are -- narrow confidence intervals). All three are well-documented in investment professionals. Gigerenzer, Hoffrage, and Kleinbolting (1991) showed that when people report 90% confidence intervals, the true inclusion rate is approximately 50% -- intervals are dramatically too narrow. In investment forecasting, the same pattern holds: analysts' 90% confidence intervals for earnings estimates contain the actual outcome only approximately 50% of the time (Ben-David, Graham, and Harvey, 2013). Barber and Odean (2001) found that overconfidence is gender-linked: men trade 45% more than women, and the excessive trading reduces men's net returns by 2.65% per year versus 1.72% for women. The mechanism: men are more overconfident about their investment ability, leading to more (costly) trading. Odean (1999) showed that the stocks individual investors buy underperform the stocks they sell by 3.3% over the subsequent 12 months -- investors are not just failing to add value, they are systematically destroying value through overconfident stock selection. The optimal response to overconfidence: pre-commit to a mechanical strategy (VTI, automatic contributions, annual rebalancing) that removes the active decision points where overconfidence operates.