Forecast Dispersion and Forecast Errors across Firms and Time

         
Author Name KAWABATA Hatsu (Keio University) / SENGA Tatsuro (Fellow (Specially Appointed), RIETI)
Creation Date/NO. July 2024 24-E-064
Research Project Firm Dynamics and Aggregate Fluctuations
Download / Links

Abstract

This paper investigates the properties of analysts’ forecast dispersion and forecast errors using a comprehensive dataset of Japanese firms from 1985 to 2023. We construct time-series indices of forecast dispersion and errors and explore their relationships with macroeconomic and financial market indicators. Our analysis reveals that forecast dispersion and errors are positively correlated, indicating that greater disagreement among analysts is associated with larger forecast errors. Forecast dispersion tends to be smaller for larger firms with more analyst coverage, and the number of analysts covering a firm is positively related to its size, age, and stock volatility. We find that the forecast dispersion and error indices are correlated with other popular uncertainty measures like the Economic Policy Uncertainty index, with spikes corresponding to events that heightened uncertainty. The indices are also countercyclical and negatively correlated with stock market performance. Our findings highlight the role of firm-level uncertainty in macroeconomic fluctuations and demonstrate the usefulness of analyst forecast data in studying the relationship between information, uncertainty, and the macroeconomy.