Résumé (EN)

The evaluation of the reliability of analysts' earnings forecasts is an important aspect of research for different reasons: Many empirical studies employ analysts' consensus forecasts as a proxy for the market's expectations of future earnings in order to identify the unanticipated component of earnings, institutional investors make considerable use of analysts' forecasts when evaluating and selecting individual sharesand the performance of analysts' forecasts sheds light on the process by which agents form expectations about key economic and financial variables. The recent period put forward a well-known phenomenon, namely the existence of a positive bias in experts' anticipations: the latter tend to over-estimate earnings. In this paper, we study the properties of this bias according to various aspects, that is to say according to country, sector, but also according to the size of the companies.