Abstract Standard static hedging models employing futures contracts yield poor results for most commodities, especially when compared with the evidence for financial instruments such as stock indexes and currencies. Moreover, the efforts in the dynamic hedging of commodity prices via GARCH models have found limited success. In this paper, we propose an alternate approach for constructing the ‘optimal’ hedge ratio. The approach differs from previous methods in two respects. First, we incorporate controls for seasonals, time to maturity, inventories, and futures term-structure in the construction of hedge ratio. Second, we adopt a partially linear functional form for the hedge ratio. Employing data from the U.S. markets for corn, cotton, and soybeans, we find that our method substantially outperforms the static, semi-dynamic, and GARCH models.