Investor Site Visits, Discussion Contents, and Analyst Forecasts: A Machine Learning Approach
Year: 2025
Type: article
Abstract: ABSTRACT We use machine learning to perform a textual analysis of 423,361 Q&As (questions and answers) involved in investor site‐visit reports, in order to explore the impact of information conveyed from them on analysts' forecast errors and revisions. We find that more performance‐ and operations‐related Q&A content discussed during site visits significantly reduces analysts' forecast errors and makes them have a lower degree of revisions. Furthermore, these relations are more pronounced when there are more institutional participants in the site visit. The results remain consistent after addressing endogeneity issues and using alternative calculations for an abnormally larger number of Q&As. Our paper supports the information digesting channel hypothesis of institutional investors and finds that questions raised by them are beneficial to analysts, no matter whether they participate in site visits, due to the timely and accurately conveying information to outside investors.
DOI: 10.1002/ijfe.70094
Authors: Jinyu Liang, Xiaogang Bi
Language: English
Subfield: Accounting
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W4415896836