Spatial Patterns of Technology Diffusion: An Empirical Analysis Using TFP

We investigate the spatial distribution of TFP growth rates using exploratory spatial data analysis and other spatial econometric techniques. Our sample consists of 73 countries and covers the period 1960-2000. We identify significant positive spatial autocorrelation in TFP growth rates, indicating that high and low values tend to be clustered in space. We also find strong positive spatial autocorrelation in TFP levels, which has increased over the period 1960-2000. This result may be indicative of a tendency towards clustering over time, a conclusion reinforced by our finding of two clusters of high TFP growth rates (in Europe and South East Asia), and two clusters of low TFP growth rates (in the Andean region and Sub-Saharan Africa). We estimate the Nelson and Phelps (1966) model of technology diffusion while allowing for spatial dependence in the error term. Our estimation results suggest that both the growth rate and the level of human capital have an important effect on productivity growth rates. JEL:I2,O4,C21. Keywords: human capital, technology diffusion, spatial econometrics.