A note on the use of quantile regression in beta convergence analysis
Código: WPE – 102
Márcio Poletti Laurini
We discuss how to interpret conflicting results obtained by the use of quantile regression methods in growth regression tests of β-convergence hypothesis and the results obtained by nonparametric methods. We show that the assumption of linearity may cause the non-rejection of the β-convergence hypothesis by quantile regression. We also show that using a nonparametric form of quantile regression, we can reject the hypothesis of β- convergence and confirm the results of divergence and formation of convergence clubs. We illustrate the discussion by using the conflicting results on convergence found in the dataset of per-capita income of Brazilian municipalities between 1970 and 1996.