On the Role of Distance for Outward FDI

This paper focuses on the estimation of three distance-related effects on outward FDI. (i) Distance harms vertical multinationals, since they engage in trade. (ii) It makes non-trading multinationals better off than exporters. (iii) This positive effect on horizontal FDI is expected to rise with bilateral country size due to the home market effect. The use of panel data and related econometric methods is highly recommended to avoid parameter bias from endogenous, unobserved, time-invariant effects. A unified estimation approach to assess all three hypotheses then has to rely on instrumental variable techniques for generalized least-squares methods. In the empirical analysis of 1989-1999 bilateral US outward FDI stocks at the industry level, it is shown that testing and accounting for autocorrelation is extremely important for parameter inference. In sum, the paper lends strong support to the theory of horizontally organized multinationals as outlined in Markusen and Venables (2000).