Avaliação do processo de convergência espacial na produtividade da cultura do café brasileiro no período de 2003 a 2023

Abstract

This study analyzed the spatial convergence of coffee productivity in Brazilian microregions from 2003 to 2023, using spatial models to capture the effects of intermicroregional dependence. Through Moran’s I tests, it was possible to diagnose spatial autocorrelation, which requires the use of models that properly address these dependencies. Simple and robust Lagrange Multiplier tests confirmed the simultaneous presence of spatial effects in both the dependent variable and the error term, necessitating simultaneous spatial lag models, indicating the superiority of the SAC model. The results showed that the convergence hypothesis was confirmed for the entire period (2003–2023), for the subperiods 2003–2013 and 2013–2023, and that the structural change test was significant for differences in 𝛽 convergence, with less productive microregions growing faster in the first period, strong spatial competition among neighboring microregions, persistence of spatial autocorrelation in unobserved shocks, and reduction in the second period. Spillover effects were important in explaining the convergence process over the period, corroborating the presence of spillovers between microregions. Although the results indicate a convergence process in coffee productivity over the period, the magnitude of the effect appears to be diminishing over time. The 𝜎-convergence showed a rising trend over time.