This study employs a natural language processing framework to explore how sentiment and tone change in corporate communication in different market conditions. Using sentiment and tone from the text data of SEC filings (10-Q and 10-K filings), fundamental data, and stock returns, we find that expansionary and recessionary market conditions can cause a change in sentiment and tone in corporate communications. The use of negative sentiment, uncertainty, and strong model tone increases during recessionary periods, while constraining tone decreases. Sentiment and tone can play a crucial role in explaining stock returns. For example, we find that positive sentiment has a positive relationship with stock returns regardless of market conditions. The significance of other sentiments and tones varies depending on the market conditions. In addition, different filing types can result in different relationships between sentiments and tones and stock returns. The study has implications for investors, corporate managers, and regulators. They can use these results as guidelines for their decision-making related to corporate communications.
Kwanjai et al. (Wed,) studied this question.