Cheating with Models
Beliefs and decisions are often based on confronting models with data. What is the largest “fake” correlation that a misspecified model can generate, even when it passes an elementary misspecification test? We study an “analyst” who fits a model, represented by a directed acyclic graph, to an objective (multivariate) Gaussian distribution. We characterize the maximal estimated pairwise correlation for generic Gaussian objective distributions, subject to the constraint that the estimated model preserves the marginal distribution of any individual variable. As the number of model variables grows, the estimated correlation can become arbitrarily close to one, regardless of the objective correlation

Link to paper: 297ff237-12bb-4686-9983-60da8eb1eb5d.filesusr.com/ugd/90366b_f660300e9e8a4b1fabd5e35c6c8a78f4.pdf

Please sign up for meetings here: docs.google.com/spreadsheets/d/1G0KdCfEkG4LYBuDSCLxyGRSEULv3_smLEEQMofG4X5U/edit#gid=0
Date: 16 October 2020, 14:15 (Friday, 1st week, Michaelmas 2020)
Venue: Held on Zoom
Speaker: Kfir Eliaz (Tel Aviv)
Organising department: Department of Economics
Part of: Nuffield Economic Theory Seminar
Booking required?: Not required
Audience: Members of the University only
Editor: Melis Clark