Professor Finds Investors Who Buy Shares in Local Companies Fail to Enjoy Superior Returns, and Overall, Trades of Individual Investors Perform Poorly
Feb. 1, 2007
Ute S. Frey
UC Berkeley Haas School of Business
UC Berkeley Haas School of Business
The idea of "local knowledge" – or "investing in what you know" – is popular stock-picking advice that doesn't appear to hold true for individual investors, according to new research from the University of California, Berkeley's Haas School of Business.
In a study of almost 1 million transactions from more than 43,000 households, Haas Assistant Professor Mark Seasholes investigated whether investors who buy stocks of local companies have superior information. He found that individuals who buy local stocks fail to outperform the stock market, suggesting that these investors had no superior information about the companies. Seasholes also found that overall, stocks bought by individuals tend to go down, while stocks they sell tend to go up.
As investors revisit their stock portfolios with the new year, Seasholes says his findings suggest they should consider index funds.
"Our findings point to indexing as a straightforward solution to the perils faced by individual investors," Seasholes wrote in a recent working paper co-authored by Ning Zhu of the University of California, Davis.
"It's just more evidence that individuals lose a lot of money," adds Seasholes, a member of the Haas finance group. "Stay away from day trading."
One goal of the working paper by Seasholes and Zhu, titled "Individual Investors and Information Diffusion," was to test the hypothesis that individuals have valuable information about stocks, particularly local stocks. They tested this hypothesis by determining whether individuals buy stocks, including local ones, before they go up and sell them before they go down.
"Our results paint a clear and consistent picture: Stocks individuals buy go down in the future; stocks they sell go up in the future," Seasholes and Zhu wrote. This pattern of poor stock picking ability holds true for investment horizons of two weeks, one month, three months, six months, and one year.
The findings confirm earlier research that has shown individuals lose money when they trade stocks. But Seasholes and Zhu also sought to build on previous research by developing a better understanding of where such well-documented underperformance by individual investors comes from.
They found that investors perform particularly badly when they trade non-S&P 500 stocks whose companies are headquartered more than 250 miles from where the investor lives. Individual underperformance stemmed almost exclusively from trades in these stocks.
"As proscriptive advice, investors should avoid buying remote, non-S&P 500 stocks – i.e. small companies far from home," Seasholes and Zhu wrote.
However, Seasholes and Zhu also found that the typical investor overweights local stocks and has approximately 30 percent of his holdings in stocks located within a 250-mile radius his home. That represents a disproportionately large amount in local holdings given that on average, only 12 percent of companies on the stock market are headquartered within the same radius.
Despite overweighting locals stocks, these stocks failed to outperform the stocks market, Seasholes found. That suggests that investors do not hold superior information about the local companies, he concluded.
Based on their findings, Seasholes and Zhu discouraged investors from buying individual stocks at all and instead suggested index funds.
"An investor who indexes avoids the adverse selection costs associated with remote non-S&P 500 stocks and can also minimize transaction costs – especially for local stocks and S&P 500 stocks for which they have little value-relevant information," Seasholes and Zhu advised.
For a copy of the working paper "Individual Investors and Information Diffusion," visit http://faculty.haas.berkeley.edu/mss/.