Integrating NLP tools in corporate governance: a study of disclosure effectiveness

Chaehyun Pyun
DOI: https://doi.org/10.1080/13504851.2024.2363995
2024-06-08
Applied Economics Letters
Abstract:This paper develops a disclosure score focused on several information categories related to effectively communicating the board evaluation process to investors. While both manually hand-collected scores and those generated by ChatGPT show an inverse relationship with information asymmetry, the scores show category- and sector-wise discrepancies. Furthermore, dissecting the negative relationship reveals that different information categories are the main drivers of the results. The findings suggest that while disclosing these information categories deserves future attention, automated tools should be used together with manual methods as complementary strategies rather than replacements for complex tasks.
economics
What problem does this paper attempt to address?