Eficiência técnica e de escala dos custos de produção das cidades produtoras de soja do Brasil

Abstract

The cultivation of soybeans plays an important role in the Brazilian agricultural scenario, being one of the main products that compose the Brazilian trade balance. This research aimed to verify the technical and scale efficiencies, of the main Brazilian cities that produced soybean in the crops from 2012/2013 to 2014/2015. The research is characterized as descriptive and applied, using documentary and bibliographic technical procedures and quantitative approach. It was performed the statistical tests of Anderson-Darling, Tukey and Kruskal-Wallis (non-parametric ANOVA), as well as the t-test and Mann-Whitney’s test (non-parametric t-test), to identify which costs have undergone statistically significant variations between observed cities. Data Envelopment Analysis (DEA) was input-oriented to observe the efficiency of soybean-producing cities, ranked them, and indicating which costs to reduce. The DEA output was the quantities of traditional and genetically modified organism (GMO) soybean produced by each city. The inputs were the incurred costs for the soybeans yields. Regarding the ANOVA, the results showed that only the variables of pesticides, agricultural insurance, administrative expenses, periodic maintenance of improvements / facilities, social charges and land presented a behavior of statistically equal costs between the observed cities. The other variables presented statistically different results. In the t-test it was possible to verify that, except for the seed variable, there was no statistical differences regarding the use of the others variables of production costs observed, when was used traditional or GMOs seeds. As for the DEA, the city of Cristalina / GO, in the 2012/2013 crop was the most efficient, presenting efficiency score equal to 1.00. Technical and scale efficiency problems were found in 65% of the observed DMUs, 20% presented problems on scale efficiency, 2% problems on technical efficiency and 14% were classified as efficient. On average, the scale efficiency was 92%, and the pure (technical) efficiency was 95%. In this way, the inefficient cities (DMUs) could, on average, reduce 5% of the inputs that would return the same output. In addition, these cities, to operate at a yield optimal scale, should increase the amount of soybeans produced by 8%.