A New Bayesian Bootstrap for Quantitative Trade and Spatial Models

2025-05-17

AI Summary

This paper addresses the challenge of quantifying uncertainty in counterfactual predictions within quantitative trade and spatial models—characterized by dyadic (bilateral) data with complex dependence, few interacting units, and predictions that depend jointly on structural parameter estimates and counterfactual equilibrium inputs. We propose the first Bayesian bootstrap tailored to such models, which simultaneously ensures Bayesian validity in finite samples and asymptotic frequentist consistency—overcoming key limitations of classical bootstrap methods in small-scale, strongly dependent network settings. The method is validated across canonical frameworks including Waugh (2010), Caliendo & Parro (2015), and Artuç et al. (2010), demonstrating robustness and improved reliability in policy-relevant counterfactual inference.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Causality
• Game Theory and Economic Paradigms: Mechanism Design

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

Economists use quantitative trade and spatial models to make counterfactual predictions. Because such predictions often inform policy decisions, it is important to communicate the uncertainty surrounding them. Three key challenges arise in this setting: the data are dyadic and exhibit complex dependence; the number of interacting units is typically small; and counterfactual predictions depend on the data in two distinct ways-through the estimation of structural parameters and through their role as inputs into the model's counterfactual equilibrium. I address these challenges by proposing a new Bayesian bootstrap procedure tailored to this context. The method is simple to implement and provides both finite-sample Bayesian and asymptotic frequentist guarantees. Revisiting the results in Waugh (2010), Caliendo and Parro (2015), and Artuc{c} et al. (2010) illustrates the practical advantages of the approach.

Problem

• Address uncertainty in dyadic trade/spatial model predictions
• Handle small sample sizes and complex data dependence
• Improve counterfactual prediction accuracy via Bayesian bootstrap

Innovation

• New Bayesian bootstrap for trade models
• Handles dyadic data with complex dependence
• Provides Bayesian and frequentist guarantees