Correlations in agricultural portfolio evaluated with Monte Carlo simulation

Abstract

The effect of using correlation coefficients was evaluated, both between grain prices and crop yields, on the estimates of results and associated risks of a portfolio evaluated with a Monte Carlo Simulation model. The diversified portfolio is made up of wheat, corn, sunflower, first and second crop soybeans. The module considered was 1,000 hectares. The yield correlation coefficients were obtained from series elaborated with agronomic simulation models; those of prices from 12-year historical series (2007/08 to 2018/19 campaigns) expressed in constant US dollars, previously homogenized with respect to withholdings and/or export duties. The taxes in force as of May 2019 (28% on soybeans, and 10% on sunflower, corn and wheat) were applied. The inclusion of the correlation coefficients in the chosen portfolio model does not modify the expected result (it is an additive model). Regarding variability, there is a slight increase in it. It is concluded that it is important to evaluate a portfolio of these characteristics, including the correlation between variables, in order not to underestimate the risk faced by the producer.