Testing the 'home market effect' in a multi-country world

Abstract

We extend Krugman's (1980) two-country two-sector model to a setup with arbitrary numbers of countries and sectors. The extended model predicts an adequately defined "home market effect" only after controlling for cross-country differential accessibility through a theory-based linear filter. We bring that prediction to data by running a battery of non-parametric sign- and rank-tests that are closely related to those used in factor proportions theory. When applied to production and trade data on a cross-section of OECD and non-OECD countries, we find support for the presence of "home market effects" in a broad number of industries.