Large Language Models and Financial Market Sentiment

60 Pages Posted: 23 Oct 2023

See all articles by Shaun A. Bond

Shaun A. Bond

UQ Business School

Hayden Klok

The University of Queensland

Min Zhu

University of Queensland

Date Written: September 26, 2023


We investigate the predictive capabilities of large language models (LLMs) ChatGPT and BARD in the context of forecasting aggregate stock market returns. We employ these LLMs to extract daily summaries of business news relevant to the S&P 500 Index, from which we construct a market sentiment indicator. Our findings reveal a noteworthy negative correlation between this sentiment indicator and short-term market returns. Notably, LLMs outperform conventional sentiment classifiers, with ChatGPT exhibiting a slight edge over BARD in out-of-sample performance. This analysis underscores the substantial potential of LLMs in text analysis — a relatively underexplored data source — for gaining insights into asset markets.

Keywords: ChatGPT, GPT, BARD, large language model, LLM, Natural Language Processing (NLP), sentiment, behavioural finance, market return predictability, asset pricing

JEL Classification: G10, G14, G17, G41, C53

Suggested Citation

Bond, Shaun Alexander and Klok, Hayden and Zhu, Min, Large Language Models and Financial Market Sentiment (September 26, 2023). Available at SSRN: or

Shaun Alexander Bond

UQ Business School ( email )

The University of Queensland
Brisbane, QLD 4072

Hayden Klok (Contact Author)

The University of Queensland ( email )

Brisbane, Queensland 4072

Min Zhu

University of Queensland ( email )

St Lucia
Brisbane, Queensland 4072

HOME PAGE: http://

Do you have negative results from your research you’d like to share?

Paper statistics

Abstract Views
PlumX Metrics