Estimating Interregional Trade: A Review Emphasizing the “Magic” of Universal Gravity and Data Mining

Abstract

Trade between regions of a single country has re-emerged as a central topic in spatial and international economics. Intranational flows nearly always exceed international flows in volume, yet they have received less attention since global value chains rose to prominence in the mid-1990s. The neglect is understandable. Consistent subnational trade data are scarce, and the usual suspects in trade frictions—tariffs, language, currency—are absent or attenuated within national borders. Studying these flows therefore calls for a particular kind of wizardry: command of spatial economic theory rich enough to handle multilateral resistance and market structure, and command of the mathematics needed to wrestle with sparse data, network dependence, and the more mundane problem of statistical endogeneity. I argue that subnational trade research has matured into a distinctive field, with its own identification strategies, datasets, and policy questions, while remaining tightly linked to advances in international trade, economic geography, and spatial econometrics. To make the case, I review the theories and methods that have crossed over from international work, summarize what three decades of substantive findings imply for interregional trade, engage measurement issues that hold more sway at the subnational scale than internationally—zero flows, zoning artifacts, entrepôt double-counting, and gross-flow versus value-added accounting—and present visually a handful of unusual facts about subnational flows. I close with a research agenda emphasizing subnational value-added decomposition, mode-specific gravity, firm-level data, network methods, and dynamic models capable of addressing resilience and adjustment to large shocks.

The authors' abstract, as published at the source. International Regional Science Review, 2026 · DOI ↗