DP14100 Cheating with (recursive) models

Author(s): Kfir Eliaz, Ran Spiegler, Yair Weiss
Publication Date: November 2019
Keyword(s):
JEL(s):
Programme Areas: Industrial Organization
Link to this Page: cepr.org/active/publications/discussion_papers/dp.php?dpno=14100

To what extent can misspecified models generate false estimated correlations? We focus on models that take the form of a recursive system of linear regression equations. Each equation is fitted to minimize the sum of squared errors against an arbitrarily large sample. We characterize the maximal pairwise correlation that this procedure can predict given a generic objective covariance matrix, subject to the constraint that the estimated model does not distort the mean and variance of individual variables. We show that as the number of variables in the model grows, the false pairwise correlation can become arbitrarily close to one, regardless of the true correlation.