Understanding the Role of Managerial Textual Content in the Price Formation Process

August 2013 Elizabeth Demers Darden Graduate School of Business University of Virginia Charlottesville, VA 22903 [email protected] Clara Vega Board of Governors of the Federal Reserve System Washington, DC 20551 [email protected] Abstract Using automated textual analysis algorithms to quantify the content of more than 20,000 managerial press releases from 1998 to 2006, this study presents large sample evidence of the role of difficult-to-verify information in the equity price formation process. We extract two textual content measures, labeled textual net optimism and textual uncertainty, and we show that each measure is associated with numerous alternative proxies for the firm’s valuation fundamentals, that the text of managerial press releases conveys “first-moment” and “secondmoment” information that is excluded from the GAAP earnings summary figure, managers’ earnings forecasts and pro-forma earnings news, and that the market relies upon this information in the price discovery process. We choose and validate an array of proxies for the informativeness of textual content and GAAP earnings, respectively, and use these conditioning variables to study how the price impact of the two signals varies with their expected informativeness about valuation fundamentals. We find that the price impact of textual content increases in settings where the text is more informative about fundamentals and decreases when GAAP earnings are noisy predictors, consistent with the tenets of a Bayesian learning model. However, our analysis also suggests that the relationship between these two sources of information is likely more complex than that of simple substitutes. We find that some firms seem to provide more informative text to complement what they understand to be informationally deficient GAAP earnings. In other words, in some cases, the informativeness of managerial text appears to be endogenously related to the informativeness of GAAP earnings.