Analyst Information Acquisition via EDGAR

Abstract

We identify analysts’ information acquisition patterns by linking EDGAR (Electronic Data Gathering, Analysis, and Retrieval) server activity to analysts’ brokerage houses. Analysts rely on EDGAR in 24% of their estimate updates with an average of eight filings viewed. We document that analysts’ attention to public information is driven by the demand for information and the analysts’ incentives and career concerns. We find that information acquisition via EDGAR is associated with a significant reduction in analysts’ forecasting error relative to their peers. This relationship is likewise present when we focus on the intensity of analyst research. Attention to public information further enables analysts to provide forecasts for more time periods and more financial metrics. Informed recommendation updates are associated with substantial and persistent abnormal returns, even when the analyst accesses historical filings. Analysts’ use of EDGAR is associated with longer and more informative analysis within recommendation reports. This paper was accepted by Shiva Rajgopal, accounting.