Benchmarking Real Trade and Multi-Regional Input–Output Databases for Applied Analytical Workflows: An Empirical Comparison of Coverage, Policy Readiness, and Use-Case Fit
Main article
Abstract
The rapid expansion of trade and multi-regional input–output (MRIO) databases has created a new problem for applied researchers: the limiting factor is no longer the absence of data, but the difficulty of selecting a database whose structure matches the intended analytical task. This study benchmarks eight real databases that are widely used in trade, global value chain, and policy analysis—EXIOBASE 3, GTAP 10, GTAP 11, UNCTAD-Eora, the International Trade and Production Database for Estimation (ITPD-E), the WTO Structural Gravity Database, the CEPII Gravity Database, and BACI. Using a structured secondary-data coding design, each database is evaluated on seven dimensions: geographic coverage, granularity, temporal depth, environmental extensions, policy-readiness, openness, and application diversity. The coded dataset is analysed using descriptive statistics, principal component analysis, k-means clustering, and scenario-based suitability scoring. Three application scenarios are examined: sustainability footprint accounting, trade-policy counterfactual analysis, and export diversification intelligence. The results show a clear functional differentiation across database families. EXIOBASE 3 performs best in sustainability-oriented work because of its strong environmental accounts and high analytical richness. GTAP 11 ranks first for trade-policy counterfactual analysis because of its strong integration with computable general equilibrium and simulation workflows. BACI performs best for export diversification intelligence because of its product-level detail and deep time coverage. The article argues that database selection should be treated as a methodological decision with direct consequences for inference quality, reproducibility, and policy relevance. A practical matching framework is offered for researchers and developers who need to build database-aware analytical pipelines.
