Cross hedge entre etanol e açúcar no Brasil: Uma análise de razão ótima e efetividade

Abstract

The objective of this dissertation was to estimate the effectiveness of the cross hedge between the hydrous ethanol and sugar futures quotations with the hydrous ethanol futures contracts at the BM &FBovespa and the Sugar 11 at ICE Futures. The daily price series of hydrated ethanol and crystal sugar provided by the Center for Advanced Studies in Applied Economics - CEPEA/ESALQ-USP, were used to calculate the future hydrous ethanol contract with financial settlement, BM&FBovespa's Information Retrieval System and the data provided by ICE Futures US for the sugar contract No. 11, in a total of 753 observations per series, divided into two periods: the first, from 03/13/2013 to 09/24/2015 and the second, from 09/25/2015 to 10/10/2016. In the first model of this dissertation, the estimation was based on the minimum variance hedge, using ordinary least squares. In the second, the estimation was performed by a autoregression vector model, and in order to achieve the general objective of this work, the methodological approach used was based on the multivariate model with conditional heteroscedasticity, GARCH-BEKK, justified by the need to consider modeling dynamically through the matrix of variances and conditional covariance of the series. A Granger causality was identified in the series analyzed, except for the hedge between the demanded sugar prices and the sugar futures contracts, which contributed to the hedge inefficiency in this model. Among the models analyzed in the first period, the one that resulted in a greater possibility in mitigating the variance of the hedger's income was the cross between the Brazilian sugar prices and the sugar futures contracts with the OLS model, which estimated an effectiveness of approximately 95%, with the model of minimum variance. Then, the same relationship was the one that presented better effectiveness, only with the GARCH-BEKK model, with 92%. For the second period, the results by the minimum variance model, as well as in Period 1, were the one that presented a greater effectiveness of hedge, with a highlight of a complete coverage in the hedge between the Brazilian sugar quotations and the American futures contract. In the application of GARCH-BEKK, the results pointed out that, when the volatility dynamics between the end of September 2015 and the beginning of October 2016 is considered, the results do not reach 50%, deducing that the hedge in this period was not effective and would not allow hedgers to reduce their risk.