Pitfalls in estimating β-convergence by means of panel data: an empirical test

This article aims to test the conjecture advanced in a recent work by Bianchi and Menegatti (Appl Econ Lett 14:963–967, 2007) that usual β-convergence panel regressions may produce biased evidence, due to their inability to distinguish between actual catching-up across countries and decreasing growth rates over time within countries. The test considers different sub-groups in a dataset of 72 countries for the period 1970–2000 and introduces both human capital and proxies for technological differences into the analysis. The results confirm the conjecture that traditional evidence about β-convergence may be misleading; they also show that catching-up across countries is weaker than usually claimed and that this process occurred only in some sub-groups of countries.