Large Language Models Excel in Financial Statement Analysis
- AI outperforms financial analysts in predicting the earnings of publicly traded companies according to new research
- Large language models (LLMs) were tasked with computing ratios used to evaluate stocks, including operating efficiency, liquidity and leverage ratios
- LLM correctly predicted whether a company’s earnings would grow or shrink in the subsequent year 60.35% of the time compared to human analysts’ 52.71% accuracy rate
- LLMs predictions were based on financial data alone, while human analysts can incorporate specific narrative contexts like management comments and industry knowledge
- Researchers built theoretical stock portfolios using LLM predictions that returned an average of 12% a year, outperforming the overall market
- LLMs accuracy rate dips during times of economic shock such as the 1974 oil shortage, 2008 financial crisis and Covid-19 pandemic
A new study finds that AI outperforms financial analysts in predicting the earnings of publicly traded companies. Large language models (LLMs) were tasked with computing ratios used to evaluate stocks, including operating efficiency, liquidity and leverage ratios. Based solely on analysis of balance sheets and income statements, the LLM was asked to predict whether individual companies would post higher or lower earnings in the subsequent year. The LLM correctly predicted whether a company’s earnings would grow or shrink in the subsequent year 60.35% of the time compared to human analysts’ accuracy rate of 52.71%. However, human analysts outperform LLMs during times of economic shock such as the 1974 oil shortage, 2008 financial crisis and Covid-19 pandemic due to their ability to incorporate specific narrative contexts like management comments and industry knowledge. Researchers built theoretical stock portfolios using LLM predictions that returned an average of 12% a year, outperforming the overall market.
Factuality Level: 8
Factuality Justification: The article provides accurate and objective information about the research comparing AI’s performance with financial analysts in predicting earnings of publicly traded companies. It presents the findings of the study and discusses the potential implications of AI outperforming human analysts in this area. The article also acknowledges limitations and areas where humans still have an advantage over AI, such as during economic shocks or when evaluating smaller firms.
Noise Level: 3
Noise Justification: The article provides relevant information about the potential of AI in predicting company earnings and compares it with human financial analysts’ performance. It also mentions some limitations of AI during economic shocks. However, it lacks a more in-depth analysis of long-term trends or possibilities and does not explore the consequences of this development on those who bear the risks.
Key People: Valeri V. Nikolaev (Professor of Accounting at the University of Chicago), Nick Fortuna (Writer)
Financial Relevance: Yes
Financial Markets Impacted: Financial analysts and AI’s impact on financial markets
Financial Rating Justification: The article discusses the potential of large language models (LLMs) outperforming human financial analysts in predicting the earnings of publicly traded companies, which could have an impact on the decision-making process in financial markets. It also mentions that LLMs were able to build theoretical stock portfolios with an average return of 12% per year based on their predictions.
Presence Of Extreme Event: No
Nature Of Extreme Event: No
Impact Rating Of The Extreme Event: No
Extreme Rating Justification: The article discusses the performance of AI in financial analysis compared to human analysts, but it does not mention any extreme events such as natural disasters, financial crises, or any other significant incidents.·
