Basis Forecasting and Performance Analysis on the Selective Long Hedge for Imported Corn

Abstract

The primary objective of this study is to forecast basis for imported corn and evaluate the hedging performance of a selective long hedge strategy. Six models are used to forecast cost and freight (C&F) basis for imported corn: a moving average model (3 and 5 years respectively), a moving average model with current basis information (3 and 5 years respectively), an ordinary least squares (OLS) regression model, and an autoregressive-moving average (ARMA) time series model. Six forecast horizons from 1 to 6 months are considered. The forecasting accuracy of 6 competing basis forecasts is determined in terms of mean absolute error (MAE) and root mean squared error (RMSE). The results show that the ARMA model performs best overall. Based on the results that the ARMA model is the best overall, basis forecasts from the ARMA model are used to simulate a selective long hedge where a long hedge is placed only when the basis will have weakened over the hedging period. The net buying price (NBP) from the selective long hedge is compared in pairs with the objective price of the traditional long hedge. The simulation results show that the NBP from the selective long hedge is significantly lower than the objective price of the traditional long hedge over the hedging period of 5 and 6 months.