The Knowledge Capital Model Revisited

The Knowledge Capital model – described in Markusen (2002) - is a general equilibrium model that allows horizontally integrated multinationals, vertically integrated multinationals, and national firms to arise endogenously, based on exogenous country characteristics such as size and relative skill endowments. Using numerical simulations, I show that the model predicts manufacturing affiliate sales 1) to increase in size similarities, 2) to increase in skill-endowment similarities, and 3) to be high when a country is both small and skilled-labor abundant. I estimate these predictions on bilateral panel data for 81 home countries and 76 host countries (period: 2008-2014), using two distinct regression specifications, using both occupational skill levels and human capital as a measure for skilled-labor abundance, and for both manufacturing and services industries. My manufacturing industries results are mixed but slightly supportive of the predictions of the KC model. Results for the services industries are strikingly similar. Since the predicted pattern of affiliate sales is expected to be different for services industries than for manufacturing industries, these similar results suggest that the uncovered effects are due to some unknown mechanisms that are not reflected in the model. The results question the credibility of the model as a tool to predict patterns of multinational activity.