标题：Formulation of Brazilian Sugar Basis Forecasting using Time Series Models: Comparison Between The Northeast and Southeast Spot and Ice Futures Markets
期刊名称：International Research Journal of Finance and Economics
印刷版ISSN：1450-2887
电子版ISSN：1450-2887
出版年度：2019
期号：172
页码：21-35
出版社：European Journals Inc.
摘要：Sugar basis identification indicates a strategic tool for decision-making in Brazil. We formulate an accessible forecast using ARMA (p, q) time series models, comparing Brazilian sugar spot and ICE futures markets. We identify sugar basis in Alagoas and São Paulo with synchronous high and low primary movements, volatility, with a correlation coefficient of 0.70, illustrating a high and positive magnitude, and descriptive statistics for Alagoas that are higher than São Paulo. Alagoas shows seasonality over the year, whereas São Paulo, for nine months. Comparing forecasting errors between ARMA (p, q) and SARMAX (p, q) models, the monthly sugar basis model for Alagoas is SARMAX (1, 0). For São Paulo, the forecast model is ARMA (2, 0). In addition, the Alagoas sugar basis breakpoint month is 2008M09, and São Paulo breakpoint month is 2011M11. These breakpoints coincide with the identifiable sugar basis level and volatility trends in the examined period. Both forecast models are easy to use and implement, and are strategic informational inputs for more efficient allocative decisions by the Brazilian sugar supply chain.
关键词：Brazilian sugar basis; Forecast; Time series; Breakpoints; ARMA (p; q) and SARMAX (p; q).
