AI Investment Ratings: ChatGPT’s Analysis of 1Q2024 Earnings Call Transcripts

open_access
Open Access
Author:
Burke, Thomas
Area of Honors:
Finance
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
Brian Spangler Davis, Thesis Supervisor
Brian Spangler Davis, Thesis Honors Advisor
Keywords:
Artificial Intelligence
AI
ChatGPT
Earnings Call Transcripts
Russell 2000 Index
Sentiment Analysis
Market Effciency
OpenAI
APIs
Abstract:
The rapid pace of innovation in artificial intelligence continues to prove beneficial across various business domains. Recent research studies highlight ChatGPT’s robust ability to interpret text and conduct sentiment analysis. For instance, Lopez-Lira & Tang (2023) demonstrated ChatGPT’s proficiency by observing cumulative returns of 500% over a 14-month period through its analysis of news announcements. Similarly, Pelster & Val (2024) identified a strong positive correlation between ChatGPT’s assessment of web articles for investment attractiveness and the subsequent returns of S&P 500 constituents. The existing body of literature suggests a prevalent phenomenon known as post-earnings announcement drift (PEAD), challenging the notion of market efficiency. According to Fink (2021), there exists an inverse correlation between PEAD and the size of firms, attributed possibly to the heightened trading frictions and worse information environments experienced by smaller companies. This study examines the efficacy of ChatGPT in analyzing earnings call transcripts from companies within the Russell 2000 index, particularly focusing on smaller firms prone to earnings drift. The investment recommendations derived from the analysis serve as a basis for portfolio construction, followed by the calculation of returns. The increasing integration of AI in finance carries significant implications for further enhancing the efficiency of public markets. This trend is likely to accelerate with widespread adoption of APIs provided by prominent financial software companies such as FactSet, Bloomberg, and S&P Capital IQ.