Short-term price density forecasts in the lean hog futures market

Abstract

We estimate and evaluate ex-ante density forecasts of lean hog futures prices using two approaches: forward-looking techniques using options market data and time series models. Our findings indicate that risk-neutral and risk-adjusted forward-looking market techniques are better calibrated and have superior predictive accuracy than time series GARCH models based on historical data. Improvements to goodness of fit and accuracy of the forecasts obtained by the calibration from risk-neutral to real-world densities imply that short-term risk premiums may be present in the lean hog futures markets, and they most likely appear in periods of market turmoil.