Determinants of the Brazilian coffee export supply function

Abstract

This study aims to estimate the coffee export supply function for Brazil from 1997 to 2016. For this analysis an autoregressive distributed lag model – ARDL – was used with the following variables: quantity of coffee exports, domestic price, export price, exchange rate, Brazilian GDP and dummy variable representing the seasonality of coffee. After identifying cointegration between these variables, the results were expressed for short and long run estimation. There was a direct relationship between the international price and the quantity of exports, whereas for the domestic price this relationship occurred in an inverse way. In the short run, the exchange rate and the Brazilian GDP were directly related to the dependent variable, although in the long run the exchange rate was negatively related to the exported quantity. Seasonality was not considered an influencing factor, given its non- significance for the proposed model. It is also noted the occurrence of fast adjustments towards long run equilibrium within the model estimated for analysis. The results highlight the consideration of implementing new strategies to establish the Brazilian product greater adherence to the dynamics of international prices.