The efficiency of the future market of coffee in Brazil: an approach through quantile regression

Abstract

The present paper intends to verify the brazilian future market’s efficiency for the arabica coffee, according to the contracts negotiated in B3 between 2007 and 2017. For this, the method of ordinary least squares and quantile regression was applied. The results indicated that the series are stationary of first order and the test of Engle and Granger indicated that the models are cointegrated. The residue tests made it impossible to interpret the OLS, indicating that its use is not adequate. However, the same does not occur with the quantile regression, which presented in the long term an increase of the optimal rate and the market efficiency, with risk premium where the price oscillations were higher and in the short term the opposite occurs.

Keywords

• eficiência de mercado
• café
• regressão quantílica