GEP Research Paper 08/21
Does Aid Cause Trade? Evidence from an Asymmetric Gravity Model
Douglas Nelson and Simone Juhasz Silva
Summary
This paper develops an asymmetric version of the Anderson-vanWincoop (2003, AER) foundation for the gravity model and extends the Baier-Bergstrand method of local linearization to permit estimation with ols appropriate to the AvW extension. This method is then applied to study the link between foreign aid and trade. We find the effect to be positive, but small.
Abstract
Anderson and vanWincoop (2003) developed what has become the standard framework for framing and interpreting empirical work using the gravity model. Its main advantage is that it recognizes and tackles the issue of endogeneity of prices. Hoverer, two shortcomings of their framework are that 1) it relies heavily on an assumption of symmetry among countries; and 2) it requires nonlinear estimations. For issues related to North-South trade, the assumption of symmetry is problematic. In this paper we develop an asymmetric extension of the Anderson-vanWincoop framework appropriate to the analysis of North-South trade. To avoid nonlinear estimations, we also use an appropriately extended version of Baier and Bergstrand’s (2006) method of estimating a linear approximation to the model—thus permitting estimation using (“good old”) OLS and easily compute comparative statics. As an illustration of its use, we examine the empirical link between foreign aid and trade. The results are striking. The coefficients are positive and significant, matching a long list of empirical results in the aid and trade literature. However, the comparative statics shows that aid affects prices so as to reduce the volume of trade of non-donor Northern exporters. Since most Northern countries are non-donors, the total volume of exports from the North actually decreases.
Issued in July 2008
This paper is available in PDF format