 |
An International Meeting of American Accounting
Association
|
Echo Chambers: How Do They Affect What Analysts Say and What Investors Hear? |
Robert Bloomfield
Cornell University and Harvard Business School
Jeffrey Hales
University of Texas at Austin
Abstract: We show in a laboratory experiment that the quality of available information determines whether consensus forecast accuracy is helped or harmed by allowing forecasters to communicate their views to one another in real time. We find strong evidence that communication decreases forecast dispersion and causes the consensus to be more extreme. When information is less diagnostic (so that extreme forecasts are inappropriate), increased extremity lowers consensus accuracy in a manner consistent with criticisms that equity analysts and the financial press form an “echo chamber” that amplifies unreliable information. When available information is highly reliable, we find that communication improves accuracy, but only among firms that have recently experienced extreme performance. In a second experiment, we examine how a set of investor participants use the forecasts produced in experiment 1.