An Artificial Neural Network Model to Analyze Maize Price Behavior in Mexico

Abstract

Mexican producers were supported by the Government from 1993-2013 through the PROCAMPO (support program, direct subsidy from the federal government), in order to keep the prices stable. The tool used to analyze the basic products price behavior was Artificial Neural Network (ANN) models. Data from the period 1980 to 2014 was available, and two ANN models were implemented, one including only internal prices of related products and the second takes into account also the international maize price. The purpose of this paper was to analyze the impact of internal related products and international maize price on the behavior of internal maize price. Both models demonstrated the powerful of the ANN for prediction of the maize price 3 and 5 years ahead. Also, a sensitivity analysis was performed using Garson’s algorithm, through which was demonstrated that rice, safflower and sorghum are corn’s complementary products. The rice price was the variable with a major positive impact on the national corn price. On the other hand, wheat, soybean and barley behave as corn’s substitute products, with wheat price being the variable with major impact on the corn price. Including the international corn price improves the model performance, besides there was a negative sensitivity between the international and national corn price; this behavior is explained by the fact that during PROCAMPO existence, the national prices were forced to be stable while the international maize price was increasing. Finally, this study showed the application of the ANN models in a particular product price behavior, with the possibility of being used in the decision making process for the application of subsidies to the agricultural producers in public policies.