The key goal of this research is to examine the impact of analyst participation in corporate earnings calls on forecast accuracy using causal forest, a machine learning-based causal inference method. By addressing selection bias inherent in analyst participation, this study aims to provide a more robust estimation of its effects while identifying heterogeneous treatment effects to determine which analysts benefit most. Additionally, it explores the regulatory implications of selective access to management, particularly concerning Regulation Fair Disclosure (Reg FD), to assess whether such interactions provide certain analysts with an unfair informational advantage. By integrating machine learning with causal inference, this research advances empirical methodologies and offers valuable insights into financial information dissemination and market efficiency.

2024

2025