Spatial Autoregressive Spillovers vs Unobserved Common Factors Models. A Panel Data Analysis of International Technology Diffusion

Mardi | 2013-01-16


This paper provides an econometric examination of geographic R&D spillovers among countries by focusing on the issue of cross-sectional dependence. By applying several unit root tests, we show that when the number of lags of the autoregressive component of augmented DickeyFuller test-type speciffications or the number of common factors is estimated in a model selection framework, the variables (total factor productivity and the R&D capital stocks) appear to be stationary. Then, we estimate the model using two complementary approaches, focusing on generalised spatial autoregression and unobserved common correlated factors. These approaches account for different types of cross-sectional dependence and are related to the notions of weak and strong cross-sectional dependence recently developed in the literature.