This research compares partial equilibrium and statistical time-series approaches to hedging. The finance literature stresses the former approach, while the applied economics literature has focused on the latter. We compare the out-of-sample hedging effectiveness of the two approaches when hedging commodity price risk using futures contracts. For various methods of parameter estimation and inference, we find that the partial equilibrium models cannot out-perform a vector error-correction model with a GARCH error structure. The partial equilibrium models' unpalatable assumption of deterministically evolving futures volatility seems to impede their hedging effectiveness, even when potentially foresighted option-implied volatility term structures are employed.