Visiting researcher Maximilian Muhn talks about Large Language Models

On, Wednesday, 3rd July, we had the pleasure to welcome Maximilian Muhn (University of Chicago) in our Accounting Research Workshop (TRR 266 Seminar series). He presented his paper “Financial Statement Analysis with Large Language Models” co-authored with Alex Kim and Valeri Nikolaev. Using financial statements from 1968 to 2021, the paper finds that LLMs such as GPT4 outperform financial analysts in their ability to predict earnings changes, suggesting that LLMs may take a central role in decision-making. The working paper is available here