A Função de produção da agropecuária brasileira: diferenças regionais e evolução no período 1975-1985

Abstract

The main objective of this study was to verify the contribution of the factors of production in the Brazilian agriculture in 1975, 1980 and 1985, using Cobb Douglas and Ulveling-Fletcher type production functions. Specificaly, the aggregated production functions for the Brazilian agriculture and four from its regions were estimated. The data used came from the 1975, 1980 and 1985 agricultural censuses; the 1970 and 1980 demographic censuses; and, special tables of the 1985 agricultural census. The observations are 300 homogeneus microregions integrating the Northeast, Southeast (without the state of São Paulo), the state of São Paulo and Southern regions of the country. The value of agricultural gross product (VP) was the dependent variable of the functions. The explanatory variables were: labor (EH), cropping land (AL), pasture land (AP), land under forests (AM), capital in culture (CC), capital in livestock (CA), value of installations and other constructions, machinery and agricultural equipments (VI), expenditure with fertilizers and correctives (DA), expenditure with livestocks (DM), other expenditures (OD), literacy rate (ESC1), urbanization (URB) and land price (PRT). The results obtained indicate that the production factors which contributed the most for the determination of the gross product value in 1985 were: labor, other expenditures, land price and cropping land. In 1980, the main production factors were: labor, capital in livestock and other expenditures. Finally, in 1975, other expenditures, labor, capital in livestock and cropping land were the most important production factors. Among the production factors included in the models, aggregated labor (EH) as well as its desagregated forms, EHF (family labor) and EHA (hired labor), became the most important ones in the formation of gross product value. Their coefficients of elasticity were allways positive and highly significant for all the models adjusted over the three years. They were followed by capital in livestock, land price and cropping land, in decreasing order of importance. A low marginal productivity per invested cruzeiro was found for fixed capital (CC, CA and VI). Those values were not high enough to compensate for depreciation and services on invested capital. On the contrary, the production factors corresponding to operating capital (DA, DM and OD) presented marginal productivities greater than one in most cases. The production functions estimated for the regions were highly significant at the 5 percent level. However, they presented serious multicolinearity problems. It was also found that the marginal productivity of hired labor in the State of Sao Paulo is approximately 2,2 times larger than the one for the Northeast region, 3,9 times that of Southeast region and close to 1,2 that the of Southern region. This fact is due to the high degree of mechanization and the use of modern inputs in São Paulo's agriculture. It was also determined that the aggregated Brazilian agriculture showed technical change bettween 1975 and 1985, as well as constant returns to scale in 1975, 1980 and 1985. Finally, it was observed that Ulveling-Fletcher fitted models very well, but their goodness of fit was not different from that of the Cobb-Douglas models. However, the Ulveling-Fletcher model permited to verify that the elasticity of cropping land is positively affected by land price and by urbanization. In addition, the elasticity of labor is affected by the average literacy rate in the microregions.