Quasi-Bayesian Hierarchical Models

Desmond Fairall, Thomas Glinnan · 2026

Read Quasi-Bayesian Hierarchical Models by Desmond Fairall, Thomas Glinnan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-31930.txt.

Introduction

Fairall and Glinnan develop a way to pool evidence across grouped studies when each group is estimated by the Generalized Method of Moments. Applied economists increasingly estimate the same parameter in several countries, markets or experimental sites, then face the external-validity question of how much information to share. Bayesian hierarchical models answer it naturally, but they need a likelihood, which GMM settings often lack.

The Quasi-Bayesian Hierarchical Model keeps each group's own GMM objective and adds a pooling term for economically comparable parameters, with estimation by a Laplace-type quasi-posterior. For strongly identified parameters, the estimator matches standard GMM asymptotically in distribution. For weakly identified ones, a limit-experiment analysis shows the pooling rule is a Bayes rule under the hierarchy-induced prior, and pooling can reduce mean squared error.

Catalog record

CREATORDesmond Fairall, Thomas Glinnan
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE296 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.31930

Browse the full library · About the Tower of Babel