Alternative models for predicting the futures market basis for Tennessee feeder cattle

In a fluctuating market the feeder cattle basis may be subject to changes which reduce the effectiveness of hedging for the Tennessee feeder cattle producer. This variability of the closing basis increases the price risk for the hedger. The purpose of this study was to determine the explanatory variables which affect the basis for feeder cattle at Tennessee locations and quantify those relationships in order to facilitate prediction of the basis for future time periods. An econometric model, simple trend-seasonal model, and an ARIMA model were chosen for use in analyzing and predicting the basis. Explanatory variables in the econometric model included transportation cost, cash feeder cattle prices, cattle inventory, feedlot placements, cattle slaughter, corn (feed) prices, and a set of seasonal dummy variables. The models were estimated using 1972-80 data and forecasts from each model were compared to the actual basis values for 1981. Of the three models, the econometric model provided the best explanation of the basis during 1972-80, and the best forecasts of the basis for 1981 based upon a minimum residual sum of squares criterion. However, even the best model (econometric) did a relatively poor job of explaining the basis for feeder cattle (R²=0.41). Cash price for feeders and transportation cost were found to be significant variables in the econometric model. Results also indicated that a significant seasonal pattern exists in the basis for feeder cattle. The ARIMA model which best explained the 1972-80 basis was a first order autoregressive model with a one-year seasonal component. Even though results provided imperfect basis estimates, the use of these models should improve the hedger's chance of making a good basis forecast, and thus reduce risk in hedging feeder cattle.