Associations between Forecast Errors and Excess Returns near to Earnings Announcements

This paper reassesses the information content of annual earnings announcements using errors in analyst forecasts published within one week of those announcements as the proxy for unexpected earnings. In addition to the use of analyst forecasts near to the announcement date, features which distinguish this study from earlier work include: a more precise dating of earnings announcements; a comparison of analyst forecast errors and changes in fourth-quarter earnings as proxies for unexpected earnings; tests of unusual variability in excess returns at the time of earnings announcements with the influence of forecast errors removed; a separation of early and late disclosers within an industry; and an examination of the properties of forecast range as an ex ante measure of earnings predictability. We conclude that: provided that analyst forecast errors measure unexpected earnings, annual earnings announcements have information content even when compared to market expectations very near to those announcements; analyst forecast errors do not dominate fourth-quarter changes as a proxy for unexpected earnings; other information released concurrently with earnings announcements appears to have significant pricing implications; there is greater information content in earnings announcements of early disclosers than of late disclosers; and forecast ranges may provide a reasonable measure of the error in analyst forecasts, and hence of earnings predictability.