Current Working Papers

This paper compares the performance of analyst stock recommendations across brokerage houses. We document that the buy recommendations of the largest brokerage houses outperform those of the smallest, as one might expect. Somewhat surprisingly, though, the sell recommendations of the smaller brokers earn more than those of the larger ones. Ranking brokerage houses on the basis of prior-year performance, we find no reliable evidence that the abnormal return on the current recommendations of the top-ranked brokerage houses exceeds that of the bottom-ranked houses. Despite the performance rankings of brokers published in the popular press and the brokerage advertisements that tout prior performance, empirical evidence of performance persistence for brokerage house stock recommendations is weak at best.
Download this paper (pdf version).

We examine changes in the stock trading behavior and investment performance of 1,607 investors who switch from phone based to online trading during the period 1991 to 1996. We compare their trading and performance to that of 1,607 investors with similar size accounts who do not trade online. We find that those who switch to online trading experience unusually strong performance prior to going online, beating the market by more than two percent annually.  After going online, they trade more actively, more speculatively, and less profitably than before -- lagging the market by more than three percent annually. A rational response to lower trading costs, improved execution speed, greater ease of access, or unusual liquidity needs does not explain these findings. The increase in trading and reduction in performance of online investors can be explained by overconfidence augmented by self-attribution bias, the illusion of knowledge, and the illusion of control.
Download this paper (pdf version).

Theoretical models of financial markets built on the assumption that some investors are overconfident yield one central prediction: overconfident investors will trade too much.  We test this prediction by partitioning investors on the basis of a variable that provides a natural proxy for overconfidence – gender.  Psychological research has established that men are more prone to overconfidence than women.  Thus, models of investor overconfidence predict that men will trade more and perform worse than women.  Using account data for over 35,000 households from a large discount brokerage firm, we analyze the common stock investments of men and women from February 1991 through January 1997.  Consistent with the predictions of the overconfidence models, we document that men trade 45 percent more than women and earn annual risk-adjusted net returns that are 1.4 percent less than those earned by women.  These differences are more pronounced between single men and single women; single men trade 67 percent more than single women and earn annual risk-adjusted net returns that are 2.3 percent less than those earned by single women.
Download this paper (pdf version).

Trading is Hazardous to Your Wealth:
The Common Stock Investment Performance of Individual Investors

Individual investors who hold common stocks directly pay a tremendous performance penalty for active trading.  Of 66,465 households with accounts at a large discount broker during 1991 to 1996, those that traded most earned an annual return of 11.4 percent, while the market returned 17.9 percent.  The average household earned an annual return of 16.4 percent, tilted its common stock investment toward high-beta, small, value stocks, and turned over 75 percent of its portfolio annually.  Overconfidence can explain high trading levels and the resulting poor performance of individual investors.  Our central message is that trading is hazardous to your wealth.
Download this paper (pdf version).

We analyze the common stock investment performance of 166 investment clubs using account data from a large discount brokerage firm from February 1991 through January 1997.  The average club tilts its common stock investment toward high-beta, small, growth stocks, and turns over 65 percent of its portfolio annually.  The average club lagged the performance of a broad-based market index by over three percent per year; the average club earned an annualized geometric mean return of 14.1 percent, while a market index returned 17.9 percent.  In addition, 60 percent of the clubs we analyze underperform the index.
Download this paper (pdf version).

We analyze tests for long-run abnormal returns and document that two approaches yield well-specified test statistics in random samples.  The first approach uses a traditional event study framework and buy-and-hold abnormal returns calculated using carefully constructed reference portfolios.  Inference is based on either a skewness-adjusted t statistic or the empirically generated distribution of long-run abnormal returns.  The second approach is based on the calculation of mean monthly abnormal returns using calendar-time portfolios and a time-series t statistic.  Though both approaches perform well in random samples, misspecification in nonrandom samples is pervasive.  Our central message is that the analysis of long-run abnormal returns is treacherous.
Download this paper (pdf version).

 Since 1973, floating exchange rates and significant oil price changes have coincided with dramatic market-share gains (losses) by Japanese (American) automakers in the U.S. market.  This paper analyzes and empirically estimates the extent to which exchange rate and oil price changes have contributed to this market shift.  We first develop a dynamic Cournot model of long-run profit-maximizing firms that operate in a macroeconomy characterized by shocks to income, exchanges rates, oil prices, and firm-specific demands and supplies.  Using the solutions for quantities sold from this model, we then construct a structural vector autoregression (VAR) to estimate and identify a reduced-form VAR.  The empirical results indicate that a strong yen increases quantities sold by American automakers and decreases quantities sold by Japanese automakers; this exchange-rate effect accounts for approximately 4% of the variance of changes in monthly sales quantity for automakers.  Oil-price increases reduce the quantity of automobiles sold by American automakers, but, contrary to the common belief, have little effect on Japanese automakers; this oil-price effect accounts for 6.5% of the variance of changes in monthly sales quantities for American automakers.  Over the two decades we analyze, however, the real value of the dollar has almost steadily declined against the yen, and the real price of oil has ended up unchanged, so these variables cannot explain the decline (rise) of American (Japanese) automakers.  Clearly, automobile sales are exposed to exchange rate, oil price, and income risk; between 10 and 20% of the changes in monthly sales quantities can be explained by the macroeconomic variables that we analyze.  However, we conclude that firm-specific policies likely account for the bulk of gains and losses actually experienced by the automakers.
Download this paper (pdf version).