Previsão e análise dos preços de fertilizantes com formulação NPK para estratégia de compra dos produtores de soja e milho do Paraná

Abstract

The work proposes to analyze and forecast the behavior of price of fertilizers with NPK formulation for the state of Paraná during the period between January 2014 and December 2018. For the analyzed variables, the presence of unit root was tested through the augmented Dickey-Fullerand and the Phillips-Perron tests. During the forecast of the time series, the seasonal integrated ARIMA model (SARIMA) was used, and in one comparison between observed and estimated values, it was observed that the adopted model approaches real values with a maximum error of forecast of 10%. For analyzing the relation between the selected variables,the vector autoregressive model (VAR) was preferred, also analyzing the impulse response and the variance decomposition. The results of this last model revealed a small connection between the commodities and the fertilizer prices, however, the dollar fluctuation might be partially passed onto the fertilizers. The discovered data works as a beacon for agriculturists when making decisions.