ABSTRACT
This study aims to characterise steady-state land allocations between a multiple-stand forest and agriculture, when the forest is subject to a storm risk. The landowner is supposed to have recursive preferences, which permits to distinguish between intertemporal preferences and risk preferences. Using a stochastic dynamic programming model, we show that both land allocation and forest management depend on the risk and on both types of preferences at the steady-state. Risk aversion is shown to favour land allocation to agriculture and to reduce the forest average harvest age while the preference for a regular income is shown to favour forestry and to reduce the average harvest age.