There has been a revival of interest in the effect of risk on economic growth. We quantify both ex ante and ex post effects of risk using a stochastic version of the Ramsey model. We develop a simulation-based econometric methodology which allows us to estimate the model in the structural form suggested by theory. The methodology is applied to micro data from a remarkable long-running panel data set for rural households in Zimbabwe. We find that risk substantially reduces growth: in the ergodic distribution the mean (across households) capital stock is 46% lower than in the absence of risk. This is, we believe, the first micro-based estimate of the effect of shocks on growth. About two-thirds of the impact of risk is due to the ex ante effect (i.e. the behavioral response to risk) which is usually not taken into account in policy design. Our results suggest that the effectiveness of policy interventions which reduce exposure to shocks or help households in risk man!agement may be seriously underestimated.