Abstract

This thesis examines consensus and individual earnings forecasts for the United States, the United Kingdom, Germany, France, Japan, and Canada. The first part of this study is a survey of the characteristics of earnings forecasts in the six markets. The are four main results of this analysis: (1) analyst revisions are concentrated around periods when financial statement information is released; (2) bias and forecast error increase as the forecasted horizon gets further into the future, while the magnitude of both of these measures decreases over the time period studied; (3) analyst forecasts appear to outperform simple extrapolative models in all markets except France; (4) the market response to forecast revisions is generally strong and similar in magnitude across all markets, except for France and Japan. The second part of this study examines individual analysts' forecasts of earnings with the objective of determining the characteristics associated with “superior” forecasters. Variables representing resources available to the analyst, focus of analyst, and experience of analyst are examined. Resource variables represented by characteristics of the analysts' employer-brokerage are the most consistent and significant variables examined. Results are strongest in the United States, with weaker and less consistent results in the other markets examined.