Sentiment Analysis and Gender Differences in Earnings Conference Calls

Abstract

We apply textual analysis to the transcripts of a sample of nearly 78,000 earnings conference calls between 2004 and 2018, comparing the difference in sentiment between female and male senior managers (CEOs and CFOs). We focus on two main measures of sentiment: tone and vagueness. Our contribution is twofold. Firstly, we show that, ceteris paribus, female executives employ a more positive and less vague tone than their male colleagues during the two sessions that make up the calls. The more positive and less vague tone of female executives does not imply any incremental information content and appears to be a purely linguistic feature that distinguishes female from male executives. Secondly, we find that the financial analysts participating in the calls exhibit gender bias in that they are on average much less positive and more vague when faced with a female executive. In contrast, the stock market reaction, measured by the CAR one day after the call, responds only to the sentiment of the call, not the executive’s gender.