Underpinnings for Prospective, Net Revenue Forecasting in Hog Finishing: Characterizing the Joint Distribution of Corn, Soybean Meal and Lean Hogs Time Series

This research focuses on developing a biannual net revenue forecasting model for hog producers based on Monte Carlo simulation of the joint distribution of hog, corn and soybean meal price series. The relative forecasting power of historical volatility, implied volatility and GARCH-based volatility is examined. Consistent with recent research, the performance of these three methods is both commodity and horizon specific, which means there is no single best predictor. However, implied volatility often performs well. Thus, implied volatility is used to forecast variance. Historical covariance is introduced to capture the co-movement of the three price series. Our forecasting model performs well out of sample; most of the realized net revenues fall in 95 percent prediction interval. Based on this forecasting model and the assumption of a utility function, we compare our prospective evaluation with retrospective evaluation of risk management strategies. Though prospective evaluation is not significantly superior to retrospective evaluation for this particular dataset, it is useful because all the market information has been incorporated in this model and because it did protect producers from adverse price movements.