🤖 AI Summary
This paper addresses the uncertainty in counterfactual analysis within quantitative trade and spatial models, noting that conventional approaches treat the observed world state as error-free, thereby ignoring measurement error-induced bias in counterfactual predictions. To rectify this, we develop—first in the literature—a systematic empirical Bayes framework for uncertainty quantification, integrating counterfactual sensitivity analysis, structural econometric modeling, and uncertainty propagation techniques to rigorously characterize how observational errors transmit to counterfactual outcomes. Validated under the canonical settings of Adao et al. (2017) and Allen & Arkolakis (2022), our framework reveals substantial and nontrivial uncertainty in counterfactual estimates, demonstrating that standard point estimates are often severely overconfident. The framework combines theoretical rigor with empirical tractability, establishing a new benchmark for robust counterfactual inference in structural models.
📝 Abstract
Counterfactuals in quantitative trade and spatial models are functions of the current state of the world and the model parameters. Common practice treats the current state of the world as perfectly observed, but there is good reason to believe that it is measured with error. This paper provides tools for quantifying uncertainty about counterfactuals when the current state of the world is measured with error. I recommend an empirical Bayes approach to uncertainty quantification, and show that it is both practical and theoretically justified. I apply the proposed method to the settings in Adao, Costinot, and Donaldson (2017) and Allen and Arkolakis (2022) and find non-trivial uncertainty about counterfactuals.
