Value Line and I/B/E/S Earnings Forecasts

Abstract

This paper compares Value Line and I/B/E/S analyst earnings forecasts in terms of accuracy, rationality, and as proxies for market expectations. Using more recent data and forming consensus forecasts from the I/B/E/S detail files, we reach different conclusions than Philbrick and Ricks [1991], who found that Value Line provided more reliable actual EPS data, but forecasts from the two databases were of similar quality. We find that I/B/E/S actual quarterly EPS data are no longer less reliable than Value Line EPS data, and I/B/E/S quarterly earnings forecasts significantly outperform Value Line in terms of accuracy and as proxies for market expectations. We find that I/B/E/S forecasting superiority can be explained by the combination of I/B/E/S's timing advantage and the mitigation of idiosyncratic error through consensus building. We also evaluate long-term forecasts from both databases and find that I/B/E/S forecasts are less biased and more accurate. Our results have implications for research evaluating the information content of quarterly earnings announcements, research using long-term forecasts in valuation models, and research seeking to generalize evidence regarding Value Line analysts' forecasting behavior.