In West African countries, agricultural production per capita has decreased over the past half century. With continued population growth and the diminishing availability of marginal arable land, pressure on land is rapidly increasing and there is now a common view that crop yield must be increased in this region, especially as there is a wide gap between actual and potential yields. Although there are several factors which may explain this yield gap, the fact that agricultural production takes place in resource-constrained farm households exposed to risk is widely recognized as being important. Indeed, risk discourages the adoption of high-risk, high-return agricultural technologies, which in turn impedes the improvement of yields. In order to assess how climatic risk constrains intensification strategy in West Africa, we built and calibrated a bioeconomic farm simulation model predicting the choice to intensify crops or livestock as depending on the availability of key policies in the economic environment of farms, for typical cases in the groundnut basin of Senegal. These cases include two regions contrasted in terms of rainfall (Sine and Saloum) and in each region two typical farms, representing poor and less poor farmers. The model features uncertainty in weather (hence yields) and crop prices, farmer's risk aversion, nine cropping systems representing millet, maize and groundnut with various intensification levels, and the main interactions between crop and livestock: draught animal power, the feeding of animals with suitable crop products (groundnut haulms, cereal straw) and the production of farm manure. Farmers are constraint by land, labour, cash and credit availability. 180 households were surveyed to build the socio-demographic and economic dataset used by the model, and agronomic data were collected from 206 fields. These key policies analysed are (i) weather index insurances against drought impact on crop yields, either subsidised or not, (ii) subsidies