Three essays in empirical finance.

Summary This doctoral thesis is divided into three distinct chapters. In the first chapter, we study the sheep-like behavior of French individual investors. Our empirical analysis is based on a database of almost 8 million transactions made between 1999 and 2006 by 87,373 French individual investors. We show that sheep-like behavior persists over time and that past performance and level of sophistication influence this behavior. We also attempt to answer a question that has been little addressed in the literature: is adopting sheep-like behavior profitable for the individual investor? Our empirical analysis indicates that contrarian investors obtain more extreme returns (positive or negative) than sheepish investors. In the second chapter, we show that measuring the accuracy of a forecast of the future price of a stock is not sufficient to assess the quality of this forecast because the predictability of prices is likely to evolve over time and depends on the stock considered. We show that the persistence in individual differences in the accuracy of analysts' forecasts, highlighted in the literature, is not evidence of differences in skills between analysts. This persistence is, in fact, caused by a persistence in the volatility of stock returns. We introduce a measure of forecast quality that incorporates both forecast error and price predictability. Options theory provides us with the necessary elements to estimate this predictability. When this is taken into account, there are no longer differences in the skills of analysts. In the third chapter, we show that experienced and inexperienced analysts do not cover the same type of firms. Experienced analysts cover blue chip companies while inexperienced analysts cover small, young, and growing companies. These differences in coverage mean that inexperienced analysts issue price forecasts on companies whose returns are more volatile and therefore less predictable. As a result, forecast accuracy is not a good measure of whether experienced analysts are better or worse than inexperienced analysts. When these differences in hedging are taken into account, we find that experienced analysts nevertheless issue better forecasts. Although statistically significant, the economic impact of analyst experience is small.