Adverse Shocks and Divergence of Investors’ Opinion in The Peer-to-Peer Secondary Market.
This paper examines how adverse information shocks affect investors' opinions in the secondary market for peer-to-peer lending (P2P). Using Bondora platform data of 17 million observations between January 2016 and February 2024, we employ DS-Heckman method to solve the selection bias issue. We have found that when there is a negative or peculiar shock (information) regarding a loan, i.e., loan default, investors' opinions diverge. This indicates that investors will begin to value loans differently if an information shock occurs suddenly. However, as the risk increases, investors will re-evaluate their opinions regarding the loan performance, leading to a decrease in dispersion and convergence to herding behaviour. Additionally, the effects of information shocks on investor reactions to holidays are explored, revealing that investors are less sensitive during these periods.
From Earnings Calls to Headlines: Investigating the Alignment Between Corporate Tone and Post-Call Media Sentiment in Kuwait’s Banking Sector.
This paper examines whether corporate disclosure events affect the persistence of media narratives. The analysis is based on a sample of earnings calls and corresponding newspaper coverage from Kuwaiti banks over the period 2018 to 2025. We apply LLM-based sentiment analysis (ChatGPT) on a sample of text from Arabic and English newspapers. Our main results suggest that post-call media tone is most positive when the sentiment in earnings calls reinforces prior media sentiment. This effect is stronger for Islamic banks, larger institutions, and calls that are longer, more complex, or led by senior executives. Results remain robust using binary tone classification and alternative models like FinBERT.