This paper analyses how the Heckscher-Ohlin-Vanek (HOV) model performs in a regional framework, applying the methodology of Leamer (1980). We conduct a complete test for the total, interregional and international trade flows of the Spanish regions in 1995. The empirical evidence shows that the methodology of Leamer works better than that of Bowen et al. (1987) for a partition of the vector of total trade and explains interregional factor trade flows better than international ones. Relaxing the identical technology assumption does not improve either the results for the interregional trade or trade with the OECD countries, though an improvement exists in the results for the international trade vector and especially for the flows with non-OECD countries. These findings are consistent with the HOV model assumptions and explain the contradictory results about the Leontief Paradox present in the Spanish literature examining the factor content of trade.

Key words: Heckscher-Ohlin-Vanek model, Leamer approach, interregional and international trade, production cones, Leontief paradox, Spanish regions.
JEL classification: F11, F14, R12.