Testing Stochastic Convergence among Mexican States: A Polynomial Regression Analysis
Pages 36-47
Vicente German-Soto and Natalia Salazar Garza

DOI: http://dx.doi.org/10.6000/1929-7092.2016.05.04

Published: 06 May 2016

Abstract: Another look on the economic convergence among Mexican states is offered examining whether they are approaching along 1940-2010. Methodology is based on polynomial regressions, a method that determines whether predictions can be significantly improved by increasing the complexity of the fitted straight-line model. Estimates from a set of polynomial terms are a theoretical approximation to income differentials, so it constitutes an adequate frame to analyze if different initial conditions tend to diminish in the long-run. We calibrate for each economy the polynomial equation of best adjustment supported in information criteria and a strategy of backward iterative elimination. Empirical results are according with the stochastic convergence, but in a relationship where it changed after trade opening, poorer states are diverging and richer states are converging. A focalized regional policy is necessary with the aim to correct the biases produced in a context where some regions are lagging while others more are advancing.