How well do supranational regional grouping schemes fit international business research models?

International business (IB) research has long acknowledged the importance of supranational regional factors in building models to explain phenomena such as where multinational corporations (MNCs) choose to locate. Yet criteria for defining regions based on similar factors vary substantially, thus undermining consensus regarding which regional grouping schemes fit IB research models better. In response, we develop and empirically validate a theory of comparative regional scheme assessment for model-building purposes assuming that: (1) schemes can be classified based on their source of similarity; and (2) schemes within the same similarity class can be assessed for their structural coherence, based on group contiguity and compactness. Schemes with better structural coherence will also exhibit better fit with IB research models. We document support for our theory in comparative analyses of regional schemes used to explain where US-based MNCs locate operations around the world. Geography, culture- and trade and investment-based schemes with better structural coherence exhibit better initil fit with MNC location models and less change in fit after modest scheme refinement using a simulated annealing optimization algorithm. Our approach provides criteria for comparing similar regional grouping schemes and identifying "best-in-class" schemes tailored to models of MNC location choice and other IB research models.