Machine Learning, Business Analytics, and AI: Decipher Earnings Calls with Artificial Intelligence

Our paper studies whether language patterns elicited by a different level of stimulus during earnings conference calls are informative about firms' earnings. Executives endure various levels of pressure when presenting a scripted managerial discussion and spontaneous answering to analysts' scrutinized questions. Such stimulus induces evasive answers, incoherent responses, and disturbances in emotion and cognition. Our artificial intelligence based measures of language patterns, built upon deep learning and topic modeling, transform analysts' perceptions to FinTech solutions. We find that evasive and incoherent answers forecast firms' earnings and stock return. A trading strategy based on evasiveness yields a positive risk-adjusted return.


Assistant Professor
Information Systems, Business Statistics & Operations Management