Growth empirics: evidence from a panel of annual data

Abstract

The paper shows that the use of annual data pooled into a panel of several countries can be a good choice for analysing the properties of long-term economic growth. Of course, an appropriate regression specification must be considered, to account for the short-run components of such high frequency data. The results point out that conditional convergence is important, as well as physical capital accumulation, in growth process; the human capital accumulation, on the contrary, appears to be no longer significant when country fixed-effects are accounted for. More importantly, the estimates are consistent with very plausible values of the input shares.