Forecasting food prices: The case of corn, soybeans and wheat

Given the high correlations observed among food prices, we analyse whether the forecasting accuracies of individual food price models can be improved by considering their cross-dependence. We focus on three strongly correlated food prices: corn, soybeans and wheat. We analyse an unstable forecasting period (2008-2014) and apply robust approaches and recursive schemes. Our results indicate forecast improvements from using models that include price interactions.