Commodities agrícolas: análise multifractal do mercado brasileiro

Abstract

The appreciation of the dollar against the real has made consumer items more expensive and pressured inflation in Brazil, since from food to fuel and airline tickets, almost everything is influenced by the exchange rate. In this sense, it is necessary for researchers, analysts and public policy makers to be able to know the behavior of agricultural commodity prices – including the multifractal characteristics – in order to be able to apply more adequate models to make predictions and support decision-making. Therefore, this dissertation aimed to analyze whether the prices of agricultural commodities ox, coffee, corn, wheat, rice and sugar have multifractal characteristics. This investigation was carried out by applying the Multifractal Detrended Fluctuation Analysis (MF-DFA) method to the daily returns of these commodities in the period from 01/04/2010 to 01/12/2023, totaling a sample of 3215 observations, using the statistical software R. It is noteworthy that this method was chosen because it has a simple application, presents reliable results for non-stationary time series and reliably determines the multifractal scale behavior of a time series. The results showed that all studied commodities have multifractal characteristics both in real and in dollar. In more detail, all analyzed commodities showed a higher degree of multifractality in real compared to the dollar, with the exception of coffee; when comparing the degree of multifractality among all commodities only in reais, it was observed that ox stood out with the highest degree followed by wheat and, at the other extreme, coffee showed the lowest degree of multifractality.