Conclusion
We provide new evidence on the performance,persistence and behavior of individual investors. We analyze returns earned on stock, bond and derivatives investments of more than 68,000 investors trading at the largest Dutch online discount broker. Our sample period ranges from January 2000 to March 2006, which covers both the burst of the Internet bubble and the subsequent recovery of financial markets. We show that the average investor earns negative returns after adjusting for risk and style tilts. Individual investors not only underperform in terms of net performance but also earn negative gross returns.
These results are mainly driven by the performance of investors trading derivatives, as gross alphas for non-derivatives traders are close to zero. Due to bad market timing derivativestraders underperforms other investors by on average 1.60% per month in gross terms and 3.30% in terms of net return. Derivatives traders incur higher transaction costs than other investors and speculate on a market decrease when markets start to recover. Furthermore, we document that while the majority of investors do not trade every month, a subset of investors trades very actively. Although gross returns of the most active traders are significantly higher than those of less active traders, the picture is opposite in terms of net performance. The large losses from excessive trading and investments in derivatives stress the need to educate individual investors in financial decision making.
When sorting investors according to gender weconfirm the finding of Barber and Odean (2001) that women earn higher net returns than men due to higher trading costs incurred by men. However, in contrast to Barberand Odean we find that women also outperform in terms of gross performance. In addition, we show that investors with large accounts havehigher gross returns than small accounts and their performance is hurt less by trading. Investor age is not a significant determinant of individual investor returns after controlling for other factors. In general, portfolios of individual investors in our sample are tilted towards high market beta stocks and small firms.
Furthermore, portfolios exhibit a significant positive exposure to the IT sector, particularly during the tech bubble. Factor loadings in the dynamic performance attribution models, estimated using a Kalman filterapproach, show considerable time variation. Our analysis also reveals substantial differences in factor exposures between groups of investors. We find strong evidence of performance persistence among individual investors. Investors ranked in the top decile portfolio based on pastone-year performance continue to outperform investors in the bottom decile by 1.5% (grossalpha) and 2.8% (net alpha) per month over the next year. Persistence in trading costs explains only part of performance persistence. Net alphas are negative for all portfolios and significant for the bottom three deciles. Losers tend to have significantly higher exposures to the market, SMB, UMD and IT factors than winners. Furthermore, the bottom deciles tend to consist of small accounts that have high turnover and are predominantly held by men. Performance persistence is somewhat weaker on shorter horizons but still significant for 6-month periods. In general, our results show that the group of individual investors is extremely heterogeneous and we confirm the theoretical prediction of Black (1986) that most investors would be betteroff by investing in an index fund …
