ChatGPT can forecast stock price movement with better accuracy using AI
The finance industry is rapidly adopting artificial intelligence (AI) to transform various sectors, and the latest research from the University of Florida shows that ChatGPT, a large language model, can accurately predict stock market returns using sentiment analysis of news headlines. The study titled "Return Predictability and Large Language Models" found that ChatGPT outperformed traditional sentiment analysis methods provided by leading vendors, thereby displaying its immense potential in the field of finance.
The research was conducted by leveraging ChatGPT's sentiment analysis capabilities to analyze news headlines and forecast whether they were good, bad, or irrelevant news for firms' stock prices. The researchers then computed a numerical score and documented a positive correlation between these "ChatGPT scores" and subsequent daily stock market returns, showcasing better accuracy in predicting such returns.
The study's findings have significant implications for the finance industry as it could help in changing the existing methods used for market prediction and investment decision-making.
Data Set Used in the Study
The researchers used two primary sets of data for their study, namely daily stock returns from the Center for Research in Security Prices (CRSP) and news headlines. They looked at the relationship between the sentiment scores generated by ChatGPT and the corresponding stock market returns to ascertain its predictive power.
They collected news headlines for all companies listed on the NYSE, NASDAQ, and AMEX, which had at least one news story covered by the data vendor. The researchers used a relevance score to determine how closely the news pertains to a specific company and only included full articles and press releases with a relevance score of 100. However, they excluded headlines categorized as 'stock-gain' and 'stock-loss' and removed duplicate headlines for the same company on the same day and extremely similar headlines to avoid any look-ahead bias in their analysis.
Impact on Policies Around AI and Stock Markets
The study's implications go beyond the finance industry, as it could benefit regulators and policymakers in understanding the potential benefits and risks associated with the increasing adoption of large language models (LLMs) in financial markets. The findings can help develop regulatory frameworks that govern the use of AI in finance, information dissemination, market behavior, and price formation, which can be critical areas of concern as these models become more prevalent in the market.
Asset managers and institutional investors can also benefit from the study by providing empirical evidence on the efficacy of LLMs in predicting stock market returns. This insight can help these professionals make more informed decisions about incorporating LLMs into their investment strategies, potentially leading to improved performance and reduced reliance on traditional, more labor-intensive analysis methods.
Future of AI-driven Finance
The paper suggests that as AI-driven finance evolves, more sophisticated models can be designed to improve the accuracy and efficiency of financial decision-making processes. Future research should focus on understanding the mechanisms through which LLMs derive their predictive power, thereby leading to better adoption and development of AI-driven tools in the finance industry.
