Tone Management in Conference Calls. An Emprical Analysis

Résumé ou Introduction

This paper adds to the strand of research that examines whether managers can strategically use their wording through specific linguistic devices. I define 'tone management' more broadly than just sentiment, incorporating the forward-looking intensity (FLI), and the linguistic complexity of conference calls, which are recognized as key drivers of qualitative information. I contribute to the existing literature in three ways. First, by taking advantage of recent innovations in machine learning and employing Bidirectional Encoder Representations from Transformers (BERT) for financial context texts, i.e., Financial Analysis with BERT (FinBERT) for sentiment and forward-looking intensity analysis, this study delves deeper into the qualitative content of conference calls and improves the understanding of tone management. Furthermore, by examining the differences between earnings and M&A conference calls as well as between sections of these calls, I shed light on the dynamics of qualitative information in different contexts, which has received limited attention in prior research. Finally, by studying the effect of the aforementioned linguistic devices on bid-ask spreads, this study contributes to the bid-ask spread literature by providing insights into the impact of qualitative information on firms’ information environment and market dynamics.