סמינר במימון חשבונאות

Speaker:     Prof. Reuven Lehavy, University of Michigan

Abstract:

In this study, we employ an advanced topic modeling methodology from computational linguistic research to compare and contrast the thematic content of a large sample of analyst

reports to that of conference calls. This methodology allows us to explicitly identify and empirically quantify the amount of information analysts discover and interpret in their reports,

without referencing to the equity market reaction. Consistent with information discovery, we document that analyst reports issued promptly after conference calls contain a significant amount of discussion on exclusive topics that were not referred to in the conference calls. Moreover, when analysts do discuss the topics covered in the conference call, they frequently use a different vocabulary from that used by managers, consistent with their information interpretation role. Cross-sectionally, we document evidence that analysts respond to investor demand for their services by playing a greater information discovery role when firms’ proprietary cost is high and providing more interpretation when the processing cost of the information in conference calls is high. Finally, we show that investors value both the information interpretation and the information discovery roles played by analysts.