ANALYSIS OF AGRICULTURAL COMMODITY PRICES USING INFORMATION THEORY METHODS
Commodities, Entropy, CECP, Fisher Information.
Agricultural commodities are considered perhaps the most important commodities, as any sharp
rise in food prices has serious consequences for food security and well-being, especially in
developing countries. In this paper we analyze the predictability of Brazilian agricultural
commodity prices during the period after the 2007/2008 food crisis and we also analyze these
prices in sliding windows in five of these commodities, observing the evolution of these prices
over the period before, during and after 2008/2012 food crisis. We use the method based on
the information theory Complexity / Entropy Causality Plane (CECP) which is shown successful
in analyzing the efficiency and predictability of financial markets and Entropy-Fisher's
Information plane (FI). By estimating information quantifiers, a permutation entropy, statistical
complexity and Fisher information we associate each commodity with its CECP and FI position
and compare its efficiency (predictability) using the deviation of a random process. The coffee
market showed greater efficiency (less predictability) while the pork market showed less
efficiency (greater predictability). When we analyzed 5 commodities between the periods (1996-
2018), we observed a temporal evolution of the efficiency index derived from Shannon's entropy
and Fisher's information measure, we saw that the coffee and soybean price series present high
and stable efficiency. For the entire period analyzed, the efficiency of the sugar market shows a
constant increase, while the efficiency of cattle and cotton, first decreases (until the crisis) and
then increases in the post-crisis period. In the case of sugar, live cattle and cotton, it is possible
to observe the period of the food crisis (2007-2011), while for coffee and soy these measures
were not able to capture differences in predictability and disorder in them, which we classify as
commodities which did not suffer distortions in their price series during this period.