Dynamic programming is used to examine whether Australia should store wheat for subsequent sale at higher prices. The dynamic programming model is developed assuming that the demand for Australian wheat is perfectly elastic at the world price. An important consequence of this assumption is that the algorithm used in computing the optimal policy takes a very simple form. It is concluded that interest rate and wheat price in the following season are the major determinants of storage policy. When prices are stable, rule-of-thumb policies, such as storing when prices are below average, are sub-optimal.