Financial and Environmental Narratives in Earnings Calls: Investor Processing Costs and Market Reactions

Abstract

This dissertation examines the role of soft information in earnings conference calls (ECCs) and its impact on capital market dynamics. By leveraging machine learning, natural language processing, and large language models, it explores three areas: (1) how narrative disclosures contextualize accounting numbers and influence market reactions, (2) the impact of analyst heterogeneity during Q&A sessions on investor disagreement, and (3) the role of environmental transparency in reducing a firm’s cost of capital (CoC). Analysis of ECC transcripts over a decade reveals that narrative attributes enhance the interpretability of financial data, with complex textual elements taking longer to influence market prices. Heterogeneous analyst views further complicate information processing for investors, increasing disagreement. Environmental disclosures, especially those addressing societal externalities, are shown to most effectively lower CoC. This research advances understanding of how narrative elements interact with financial disclosures, highlights the costs investors face when presented with mixed signals, and demonstrates the value of ML tools for analyzing complex market dynamics. It also supports integrating ESG considerations into corporate strategy, offering actionable insights for improving transparency and efficiency in financial markets.