MODELAGEM PROBABILÍSTICA DE PREÇOS MÁXIMOS DA COMMODITY BOI GORDO PARA O ESTADO DE SÃO PAULO

• Autor(a) principal: Lucca Filho, Eduardo Lucca Filho
• Data de Publicação: 2023
• Outros Autores: Liska, Gilberto Rodrigues, Santos, Jerônimo Alves, Matiussi, Ana Carolina
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Nativa (Sinop)
• Texto Completo: https://periodicoscientificos.ufmt.br/ojs/index.php/nativa/article/view/13291
• topic: Rural Planning; Applied Economics; Gumbel Distribution; Return Value; Applied Statistics; Planejamento Rural; Economia Aplicada; Distribuição Gumbel; Valor de Retorno; Estatística Aplicada

In the economic scenario, studying the behavior of prices of products, commodities or indicators makes it possible to make forecasts, allowing the elaboration of risk projections with greater precision and, when extreme events of these prices occur, losses or even bankruptcies can occur. In this sense, the Extreme Value Theory (EVT) is more suitable such phenomena. Economic data from CEPEA - ESALQ covering the period from 1997 to 2020 were used, organized in series of monthly maximums and, for each series, the Gumbel, Generalized Extreme Values ​​(GVE) distributions and their non-stationary versions were considered. It could be seen that the Gumbel and GVE distributions fit in every month and the goodness-of-fit attest that the Gumbel distribution is the most suitable in every month. In the months of April to October there is a slight lower probability of prices being exceeded and in the months of November to February are the months with the highest probability of occurrence of high fat ox prices. The Mann-Kendall test was used for testing the trend in all series, which was incorporated into the non-stationary Gumbel distribution, and the likelihood ratio test and AIC were favorable in terms of trend modeling.