ESG 논란과 주식 수익률: 자연어 처리의 활용

This study examines whether corporate value reacts to news and articles disclosing Environmental, Social, and Governance (ESG) controversies by analyzing the daily stock market responses. Based on deep learning techniques, we collect and utilize a large-scale ESG controversy article dataset for the 10-year-long period from 2012 to 2021. By refining the big data and classifying the ESG articles using KoBERT, a natural language processing model, we investigate how the stock market reacts differently depending on the type of ESG controversy issues. We find that stock prices tend to decline in response to the negative news about ESG controversies. We also find that the stock market responds to the news more intensively as the news coverage becomes greater. The financial performance of companies such as operating profit to assets and credit rating affects the extent to which the ESG controversy changes the corporate value in the short term. Under the interdisciplinary framework, our empirical analyses and findings yield insight on how to analyze the news articles containing ESG controversies and broaden our understanding of the influence of ESG controversies on domestic companies.