We use large language models to analyze the content of 4,700 private meetings between a large active asset manager and its portfolio firms. The high-level meetings convey mostly soft information about the firm, and little about industry or market.

We document two new facts about the distributions of answers in famous statistical problems: they are i) multi-modal and ii) unstable with respect to irrelevant changes in the problem.

Winner, EFA Doctoral Tutorial Best Paper Prize 2024 for his paper “The Flattening Demand Curve”. The award was presented at the European Finance Association Meeting 2024, in Bratislava.

Runner-Up, ECSO-CMS 2024 Best Student Paper her paper “American Options With Acceleration Clauses”. The award was presented at the European Conference on Stochastic Optimization and Computational Management Science 2024, in Stockholm.

We provide experimental evidence of a positive memory bias which affects individuals' beliefs, decisions to reinvest, and overconfidence in the stock market.