Sentiment Analysis in Bankruptcy as a Predictor for DIP Loan Interest Rates

A number of recent papers have found a significant relationship between the sentiments of 10-K MD&A sections as well as other textual sources and bankruptcy outcomes, such as success. However, there has been no research on whether sentiment is a significant predictor of the interest rate on debtor-in-possession loans, which has been documented as having substantial benefits on the restructuring of firms in Chapter 11. This paper addresses this gap by analyzing whether the sentiment of company-produced texts—such as the MD&A section and press releases—as well as externally-produced texts—such as news articles—significantly predict for the spread on DIP loans. In general, findings were not significant, resulting in an inability to prove a statistically significant relationship between sentiment and DIP loan interest rates. However, moderate, significant improvement results were found for models that added in a sentiment scores for external-source articles and a blended average of sentiment scores from all textual pieces.