Catching growth determinants with the adaptive LASSO

This paper uses the adaptive LASSO estimator to determine the variables important for economic growth. The adaptive LASSO estimator is a computationally very simple procedure that performs at the same time both consistent parameter estimation and model selection. The methodology is applied to three data sets, the data used in Sala-i-Martin et al. (2004), in Fernandez et al. (2001) and a data set for the regions in the European Union. The results for the former two data sets are very similar in many respects to those found in the published papers, yet are obtained at a tiny fraction of computational cost. Furthermore, the results for the regional data highlight the importance of human capital for economic growth. -- adaptive LASSO ; economic convergence ; growth regressions ; model selection