Poverty, Risk and Accumulation: Pro-Poor Policies when Dynamics Matter

(CERDI), Cornell and the World Bank for very helpful discussions. Poverty and vulnerability measures provide a basis for the targeting of policy interventions. In applied work typically static versions of these measures are used so that the effect of asset changes on welfare is ignored. Unless the policymaker’s concern with household welfare is myopic, policy design requires welfare measures that incorporate the dynamics of asset accumulation. We use a simple microeconomic growth model to show that the use of a static or dynamic welfare measure affects the choice between three types of interventions (transfers, productivity enhancements and reductions in risk). We also show (using data for rural households in Zimbabwe) that the composition of the group of low-welfare households may change dramatically over time. For both reasons static poverty (or vulnerability) measures may be a poor basis for policy. Using the Zimbabwe example we show that when it is not feasible to estimate a structural dynamic model then the accuracy of existing regression-based vulnerability measures can be greatly improved by including asset measures in the regression.