One use of the NLP (natural language processing) features of ChatGPT is to search out patterns in the immense amounts of news, data and other sources of information about specific stocks, and then efficiently convert them into summaries valuable for all types of investors.  Can this be accomplished with useful results? The authors use the Q2_2023 period to test performance around earnings announcements. Earnings announcements and earnings surprises are informationally rich as well as challenging events for investors to analyze.

Can ChatGPT assist in picking stocks?

  • Matthias Pelster, Joel Val
  • Finance Research Letters
  • A version of this paper can be found here
  • Want to read our summaries of academic finance papers? Check out our Academic Research Insight category.

What are the research questions?

  1. Can ChatGPT analyze information obtained from the financial media and provide accurate investment insights?

What are the Academic Insights?

  1. A QUALIFIED YES. ChatGPT (and probably any LLM) provides cost-effective and valuable assistance to stock selection.   The results obtained in the live experiment conducted here included support in various ways.  First, the ratings made by ChatGPT were positively related to future out-of-sample returns on the rated stocks.   Interestingly, even in a weakly efficient market, ChatGPT provided valuable information based on a single rating.   If the ratings were a random process, the next-month average excess returns would have been zero.  Second, the LLM distinguished between positive and negative events in the news.  Further, when a negative event did occur for a specific stock, the LLM changed its “opinion”. Finally, the accuracy of the opinions and recommendations did vary over time.  This suggests that LLMs like ChatGPT and others should be viewed as “assistants” in the stock selection process, not the final judgment.

Why does it matter?

Although the experimental results presented in this research were impressive, there may be misconceptions that an LLM such as ChatGPT should provide the sole opinion for investment decisions. Having said that, it is difficult to disagree with this quote from the authors:

By providing tutoring-like support and helping to interpret data in a conversational manner, ChatGPT and similar technologies are likely to affect the decision-making processes of investors. The rapid integration of ChatGPT and similar technologies in various industries suggests that, in the future, people will more rely on information that has been retrieved, summarized, and analyzed by AI.

 The most important chart from the paper

Abstract

This paper studies whether ChatGPT-4 with access to the internet is able to provide valuable investment advice and evaluate financial information in a timely manner. Using a live experiment, we find a positive correlation between ChatGPT-4 ratings and future earnings announcements and stock returns. We find evidence that ChatGPT-4 adjusts ratings in response to earnings surprises and news events information in a timely manner. An investment strategy based on ‘‘attractiveness ratings’’ by ChatGPT-4 yields positive returns.

About the Author: Tommi Johnsen, PhD

Tommi Johnsen, PhD
Dr. Tommi Johnsen was a past Director of the Reiman School of Finance and a tenured faculty at the Daniels College of Business at the University of Denver. She has worked extensively as a consultant and investment advisor in the areas of quantitative methods and portfolio construction. She taught at the graduate and undergraduate level and published research in several areas: capital markets, portfolio management and performance analysis, financial applications of econometrics, and the analysis of equity securities. Her publications have appeared in numerous peer-reviewed journals.