Wednesday, November 12, 2008

EPS estimates

David Dreman and Michael Berry in a study published in the FAJ in 1995 and subsequently updated in 1996 examined brokerage analysts quarterly earnings projections as compared to the actual earnings reported between 1972 and 1996. Estimates for the quarter were usually made in the previous 3 months and analysts could revise their estimates up to 2 weeks before quarter end … in all, 94,251 consensus forecasts were used requiring at least 4 separate analyst estimates before including a consensus number in the study. The average EPS estimate error for the study was 44%. In the last eight years of the study the average error was 50% and in two of those years, 57% and 65%, respectively.

No comments: