Comportamento e previsão dos preços do leite no Triângulo Mineiro e Alto Paranaíba

• Autor(a) principal: Vilela, Eunice Henriques Pereira
• Data de Publicação: 2023
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da UFU
• Texto Completo: https://repositorio.ufu.br/handle/123456789/38508; https://doi.org/10.14393/ufu.te.2023.319
• topic: Produção de Leite; Milk Production; Transmissão de Preços; Price Transmission; Previsão de Preços; Price Forecast; CNPQ::CIENCIAS SOCIAIS APLICADAS; Administração; Leite - Produção - Triângulo Mineiro/Alto Paranaíba (MG : Mesorregião); Leite - Comércio - Triângulo Mineiro/Alto Paranaíba (MG : Mesorregião); Índices de preços ao consumidor - Triângulo Mineiro/Alto Paranaíba (MG : Mesorregião)

Considering that milk production has economic and social importance for the region of Triângulo Mineiro and Alto Paranaíba (TMAP) and milk is considered one of the most volatile agricultural commodities in the international market, this thesis proposed to examine the behavior of milk prices in the region and, more specifically, to develop a model to predict this behavior. For this, the work was divided into seven chapters. The first chapter consisted of an introduction, presenting the contextualization of the theme as well as the objectives and justifications for the elaboration of this thesis. The second chapter is a systematic review of the literature, national and international, published in the last two decades (from 2000 to 2020) that dealt with the issue of milk prices, with the objective of identifying which factors affect the behavior of milk prices. The results indicated that weather conditions, feed and beef prices, milk prices on the international market, prices of dairy derivatives and the imbalance of market relations between producers and processors are the main factors capable of influencing prices. of the milk. Based on these findings, chapters 3, 4 and 5 of the thesis aimed to investigate whether the relationships identified in other markets applied to the TMAP milk market. Chapter 3 verified the existence of spatial transmission processes of prices from the international market, the national market and the regional markets for the prices received by milk producers in the TMAP. Using the Autoregressive Vectors (VAR) methodology, at the international level, it was found that fluctuations in milk prices in Uruguay and the United States are transmitted to prices in Triângulo Mineiro and Alto Paranaíba; at the national level that prices from Goiás and Espírito Santo are transmitted to the TMAP, and, at the regional level, that fluctuations in prices paid to producers in the metropolitan region of Belo Horizonte and in the south of Goiás are transmitted to the TMAP. In chapter 4, the objective was to identify the presence of price transmission processes of the ain dairy products sold in Brazil for milk prices in the region of Triângulo Mineiro and Alto Paranaíba. The results indicated the existence of transmission of the prices of pasteurized milk, UHT milk and mozzarella to the milk prices paid to producers in the TMAP. In chapter 5, the objective was to analyze the existence of processes of transmission of prices of Other commodities to milk prices in the TMAP, namely: corn, soybean, live cattle and calf prices, and energy prices electricity and fuel. The results indicated that, in the short term, changes in calf and electricity prices are transmitted to milk prices paid to producers in Triângulo Mineiro and Alto Paranaíba. In the long term, with the exception of corn and electricity prices, all commodities have their price fluctuations transmitted to milk prices. Based on these findings, chapter 6 aimed to structure a model for predicting the behavior of milk prices paid to producers in Triângulo Mineiro and Alto Paranaíba. For both, forecasting methods were tested, seeking to identify the one that presents the best predictive performance: an Integrated Autoregressive and Moving Averages model (ARIMA) in its standard form and with the inclusion of the seasonal component (SARIMA) and a Networks model Artificial Neurals (ANNs). According to the error metrics adopted to evaluate the models, it was identified that, among the ARIMA and SARIMA models tested, the one that presented the best performance was the SARIMA(0,1,1)(2,1,0)12 model. , with an MSE of 0.285 and an R2 of 0.557. Among the tested ANN models, the one that presented the best predictive performance, with an MSE of 0.009 and an R2 0.1946, was Model 2, which had as input variables the prices of milk in the United States and Uruguay, the prices of pasteurized milk and mozzarella cheese, prices of live cattle at arroba and rural energy, the exchange rates Dollar/Real and Real/Uruguayan Peso, the General Market Price Index (IGPm) and monthly rainfall in the region , considering a lag of three months. Comparing these results, it is possible to state that the model developed from the methodology of Artificial Neural Networks, has a better capacity to predict trends of increase or decrease in milk prices paid to producers in Triângulo Mineiro and Alto Paranaíba, which corroborates studies such as the Shahriary and Mir (2016) and Romão et al. (2020), who identified that the predictions offered by the ANN models are considerably more accurate than those offered by the ARIMA and SARIMA models.