An analysis of Japanese earnings forecast revisions with application to seasoned equity offerings

Using the bootstrap method, we explore the characteristics of revisions in Japanese earnings forecast data. We find that forecast revisions exhibit a downward trend over time as the actual earnings announcement date approaches, and are serially correlated with three significant lags. Using these characteristics we develop a model to estimate abnormal forecast revisions, and illustrate the model's use with a sample of Japanese companies announcing seasoned equity offerings (SEOs). In contrast to results obtained by studies using American data, our findings indicate significant positive upward revisions when Japanese firms announce an SEO.