This study tests the absolute, conditional and sigma convergence hypothesis, for the agricultural GDP per capita (GDPAGpc). In the simulations with the absolute β-convergence, parameters were significant (α =1%) and signs of these theoretically consistent. These results support the convergence of GDPAGpc for each state of the country to a steady state condition. The conditional β-convergence revealed significant (α =1%) of the estimated parameters and their signals theoretically consistent. Therefore, the GDPAGpc of each state of the country converges to its own steady state when the active performance of the human capital is considered. However, the results of σ-convergence did not support the Brazilian GDPpc convergence hypothesis.
convergence; agriculture; neoclassical models