This study focuses on whether media sentiments can complement the already rich financial data to improve corporate credit analysis. We use natural language processing (NLP) to extract credit-focused media sentiments expressed by three reputable business presses on exchange-listed North American corporates. The NLP techniques deployed include (1) Source-LDA to identify to what extent a news article is credit-related, (2) NER to identify corporate names, and (3) BERT-based supervised classification to assign sentence-specific 5-point sentiment scores. In addition, we devise a sentence-block algorithm to obtain an article's overall sentiment on a corporate and then assign the corporate an aggregate score by factoring in multiple news sources on a single day. The empirical findings reveal that media sentiments on corporates add significant incremental information in predicting defaults and mergers/acquisitions.
Jin-Chuan Duan (https://bizfaculty.nus.edu.sg/faculty-details/?profId=181) is the Jardine Cycle & Carriage Professor of Finance and the Executive Director of the Asian Institute of Digital Finance (https://nuscri.org/en/) at the National University of Singapore (NUS). Duan received a PhD in finance from the University of Wisconsin-Madison, and is an academician of Academia Sinica, a fellow of the Society for Financial Econometrics, and an advisory board member of the International Association of Credit Portfolio Managers. Prior to joining NUS, he held the Manulife Chair in Financial Services at the University of Toronto. In 2009, Duan launched the Credit Research Initiative (CRI), a public good platform offering daily updated and machine-generated default probabilities on over 80,000 exchange-listed firms in 133 economies. He co-founded CriAT in 2017, a FinTech company specializing in deep credit analytical solutions, and has since been its non-executive chairman.
This seminar is organized by the The Centre for a Responsible Wealth Transition (CRWT), which is a new research cluster at the Telfer School of Management that intends to promote cutting-edge and industry-relevant interdisciplinary research to provide new insights and innovative financial solutions fostering a transition towards a more resilient economy and responsible wealth management practices. The cluster includes the following four sub-clusters: “Responsible Investing”, “Climate Finance and Accountability”, “Emerging Technologies”, and “Risk Intelligence and Resilient Solutions”. This seminar is orgainzed by the sub-cluster “Risk Intelligence and Resilient Solutions”. For more information about the cluster please contact Prof. Jonathan Yu-Meng Li (firstname.lastname@example.org).