Can Large language model analyze financial statements well?

Since GPT-3.5’s release, large language models (LLMs) have made significant advancements, including in financial analysis. However, their effectiveness in financial calculations and pre- dictions is still uncertain. This study examines LLMs’ ability to analyze financial reports, fo- cusing on three questions: their accuracy in calculating financial ratios, the use of these metrics in DuPont analysis and the Z-score model for bankruptcy prediction, and their ef- fectiveness in predicting financial indicators with limited knowledge. We used various meth- ods, including zero-shot and few-shot learn- ing, retrieval-augmented generation (RAG), and fine-tuning, in three advanced LLMs and compared their outputs to ground truth and ex- pert predictions to assess their calculation and predictive abilities.The results highlight both the potential and limitations of LLMs in pro- cessing numerical data and performing com- plex financial analyses.