# 13-121/III (2013-08-22)

Bert de Bruijn, Erasmus University Rotterdam; Philip Hans Franses, Erasmus University Rotterdam
Earnings Forecasts, Earnings Announcements, Financial Markets, Financial Analysts.
JEL codes:
G17, G24, M41

We analyze earnings forecasts retrieved from the I/B/E/S database concerning 596 firms for the sample 1995 to 2011, with a specific focus on whether these earnings forecasts can be predicted from available data. Our main result is that earnings forecasts can be predicted quite accurately using publicly available information. Second, we show that earnings forecasts that are less predictable are also less accurate. We also show that earnings forecasters who quote forecasts that are too extreme need to correct these as the earnings announcement approaches. Finally, we show that the unpredictable component of earnings forecasts can contain information which we can use to improve the forecasts.