For the 2014-2020 phase of the Common Agricultural Policy, the European Commission has the opportunity to reduce the leakage of public support to landowners and to better target it towards active farmers. Our purpose is to assess whether shifting the basis of direct payments from land towards active farmers will significantly alter agricultural production decisions. In a dynamic and stochastic microeconomic framework, we identify the impact of this shift on the farm household’s production and consumption decisions. In the dynamic setting the production impacts of direct payments are much higher than previously quantified, because the “long run” absolute risk aversion (associated with the value function) is lower than the “short run” one (associated with direct utility). In our dynamic setting, the impact profiles are opposed for initially poor and initially wealthy farmers, due to their different precautionary motives. Leakage to land owners is lower with an active-farmer than with a land subsidy, so that the production impact is higher.