Análise de bolhas nos produtos agricolas: açúcar, café e soja

Abstract

The process of changing market conditions, due to the major shocks of the 1970s and 1980s, as well as the changes in domestic and international market conditions in subsequent years, significantly affected the price volatility of many commodities. The volatility of prices of important commodities such as sugar, coffee and soy, demands a study on the occurrence of bubbles and their characterization. To this end, the present work proposed the identification of speculative bubbles based on the modified unit root test to accommodate the null hypothesis of moderately explosive processes in the prices of the analyzed commodities, using the GSADF methodology for the Monte Carlo simulation structures (MC) and Wild Bootstrapping (WB) with the inclusion of a trend parameter to mitigate the effects of the market trend in the period between January 1964 and December 2016. The results showed the absence of bubbles for soybeans, as it is a commodity with more liquid trading. For sugar, an average of five bubbles were identified in the analyzed period, and the episodes were longer for the trended WB and longer for the trending MC. For coffee, the periods were longer for the simulations without trend. It was concluded that the use of the trend can be used for series of commodities whose production cycle is shorter and, in the case of commodities with long cycles such as coffee, it reduces the occurrence of bubbles when taking into account the long-term trend of such markets.