Quantile Regression Estimates of Confidence Intervals for WASDE Price Forecasts

This study uses quantile regressions to estimate historical forecast error distributions forWASDE forecasts of corn, soybean, and wheat prices, and then compute confidencelimits for the forecasts based on the empirical distributions. Quantile regressions with fiterrors expressed as a function of forecast lead time are consistent with theoretical forecastvariance expressions while avoiding assumptions of normality and optimality. Based onout-of-sample accuracy tests over 1995/96–2006/07, quantile regression methods producedintervals consistent with the target confidence level. Overall, this study demonstrates thatempirical approaches may be used to construct accurate confidence intervals for WASDEcorn, soybean, and wheat price forecasts.