Theoretical and Behavioural Economics Seminar - Jakub Steiner (Zurich/CERGE-EI)

Abstract: We explore approximate Bayesian updating and likelihood maximization within a machine learning-inspired framework. We demonstrate that, under certain cognitive constraints, simpler models yield the most effective constrained fit to the actual data-generating process. While more complex models could potentially offer a superior fit, agents facing cognitive constraints may lack the capability to assess this fit accurately.

Location: Physics A1/3

Meet the speaker

Seminar Series