Author: Sasan Mansouri (Goethe University), Andreas Barth (Goethe University) and Fabian Woebbeking (Goethe University)
It is relatively easy for us humans to detect that a question we asked has not been answered – we teach this skill to a computer. More specifically, we develop a measure that detects the rejection, avoidance or dodging of a question. Using a supervised machine learning framework on a large training set of 48,197 classified responses to questions, we identify 677 trigrams that signal whether or not the respondent tries to avoid answering. We show that this dictionary has economic relevance by applying it to a validation set of contemporaneous stock market reactions after earnings conference calls. Our findings suggest that obstructing the flow of information leads to significantly lower cumulative abnormal stock returns and higher implied volatility. Our metric is designed to be of general applicability for Q&A situations, and hence, can be applied outside the contextual domain of financial earnings conference calls.