Convergence in Per-capita Gdp across Eu- Nuts2 Regions Using Panel Data Models Extended to Spatial Autocorrelation Effects

This paper studies the long-run convergence of per-capita GDP across European regions. Most of the empirical works in this area are based on either cross-sectional or a-spatial panel data xed-e ects estimates. Here, we propose the use of panel data econometrics models that incorporate an explicit consideration of spatial dependence e ects (Anselin, 1988; Elhorst, 2001; 2003). This allows us to extend the traditional convergence models to include a rigorous treatment of the regional spillovers and to obtain more reliable estimates of the parameters. Two models are considered in particular based on the introduction of a spatial lag among the esplicatives ( spatial lag model") and imposing a spatial autoregressive structure to the stochastic component ( spatial error model"). We apply such a modelling framework to the long-run convergence of per-capita GDP of 125 EU-NUTS2 regions observed yearly in the period 1977-2002. The paper also provides a comparative study between the results obtained with the two proposed models and those obtained on the same set of data with the standard β-regression, with the standard β-regression augmented with a spatial component, and with the standard xed-e ect panel data model.