Which individual stock pickers beat the market? In their September 2013 paper entitled “The Information Content of Investors’ Expectations for Risk and Return”, Thomas Berry and Keith Jacks Gamble examine the performance of the most recent stock picks of members of the American Association of Individual Investors (AAII) who participated in an online survey. They issued the invitation to participate on June 15, 2011 and collected responses over the next two weeks. Questions solicited demographic data, level of investment experience (excellent, good or limited) and information about the participant’s most recent stock purchase. They exclude responses naming non-U.S. stocks and stocks priced below $5. They measure returns to various cross-sectional portfolios of these stocks over the next six months based on initial equal weighting. They adjust returns for market, size, book-to-market ratio and momentum factors to measure portfolio alphas. Reported returns exclude trading frictions. Using 2,218 survey responses and monthly returns for purchased stocks and risk factors during July 2011 through December 2011, they find that:
- Survey responses indicate that:
- Participants are largely male (93%), older (two-thirds age 60 or older), high-income (half with annual income at least $100,000) and wealthy (60% with net worth at least $1 million dollars). About half made at least ten stock trades in the prior year. About 28%/62%/11% report excellent/good/limited level of investment experience.
- Stocks purchased tend to be large (median market capitalization $19 billion), with median share price $37.
- About 14% (85%) expect an average (above average) return from their pick. About 80% (20%) expect their pick to be no more risky (more risky) than average. 77% expect a positive alpha.
- The actual performance of stock picks indicates that investor expectations are overly optimistic, with average gross annualized four-factor alpha of all picks an insignificant -1.8% over the sample period.
- Investors lose overconfidence only after considerable investment experience. For example, the average annualized four-factor alpha for stocks expected to generate high alpha is a significant 6.7% for investors reporting excellent investment experience, but an insignificant 1.8% (-2.1%) for investors reporting good (limited) experience.
- Three sources of information predict good performance for the high-conviction (high expected alpha) stock picks of individual investors with an excellent level of experience:
- Personal experience with the company’s products indicates a gross annualized alpha of 12%.
- Knowing someone who regularly works for/with the company indicates a gross annualized alpha of 23%.
- Living close to the company’s operations indicates a gross annualized alpha of 11%.
- Conversely, other high-conviction stock picks of investors with an excellent level of experience generate a gross annualized alpha of only 0.4%.
- Predictive value of investor expectations persists for about six months.
- Neither market betas nor dividends of purchased stocks explain findings.
In summary, survey-based evidence indicates that the most experienced individual investors generate significant alpha from high-conviction stock picks based on unique personal knowledge derived from familiarity with company products, personal contacts at the company and/or geographic proximity to company operations.
Cautions regarding findings include:
- As noted, results exclude trading frictions. Reported returns therefore overstate net investor performance.
- Dividing the sample into many subsamples based on experience, conviction and potential success factors introduces data snooping bias, such that the best results likely overstate expected out-of-sample performance.
- The study does not follow up to determine if/when respondents sell specified stocks to measure realized returns.
- While the sample is fairly large, the self-selected set of survey respondents from the AAII membership may not be representative of the broader population of individual investors.
