Lucca Simeoni Pavan

This study aimed to estimate an econometric model to explain the spatial distribution of agricultural productivity between the municipal districts of Paraná in 2010. The hypothesis considered was the existence of autocorrelation and spatial heterogeneity of agricultural productivity and its determinants. We calculated an index of agricultural productivity of the land and the key variables of the model was selected through a literature review. Through Exploratory Spatial Data Analysis was identified the spatial autocorrelation and spatial heterogeneity of the studied variables. Also identified productivity clusters of the type highhigh in the regions Metropolitan, East-Central, West and Southwest. The spatial model used was the lag in the error term, because it showed significant Lagrange Multiplier Robust Test. The type of contiguity weight matrix with the highest Moran index was the array of type queen. Finally, this work concluded that the aspects of location are essential in the study of agricultural productivity and the variables that showed greater relevance in the estimation of the model were soil quality, relative area and spatially lagged error.