Multicrop farmers must choose variable input levels and land quantity for each crop. Economic researchers to date have analyzed these two decisions separately, either finding the best land use, given crop technologies, or solving for optimal input levels, ignoring the allocation of land. We show that both these approaches lead to suboptimal decision rules under risk aversion. An empirical example demonstrates that a risk-averse farmer who makes these decisions sequentially-first choosing input levels then allocating land-rather than simultaneously, may significantly understates the value of farming.