Price Density Forecasts in the U.S. Hog Markets: Composite Procedures
研究开发并评估了美国生猪价格的季度样本外个体和组合密度预测，发现对数组合（使用等权和均方误差权重）优于所有个体预测，并展示了组合方法对生产者的经济价值。
Abstract We develop and evaluate quarterly out‐of‐sample individual and composite density forecasts for U.S. hog prices. Individual density forecasts are generated using time series models and the implied distributions of USDA and Iowa State University outlook forecasts. Composite density forecasts are constructed using linear and logarithmic combinations of the individual forecasts and several weighting schemes. Density forecasts are evaluated on predictive accuracy (sharpness), goodness of fit (calibration), and their economic value in a hedging simulation. Logarithmic combinations using equal and mean square error weights outperform all individual density forecasts and are modestly better than linear composites. Comparison of the outlook forecasts to the best composite demonstrates the usefulness of the composite procedure, and identifies the economic value that more accurate expected price probability distributions can provide to producers.