65

Two new methods have been developed to assess the impact of Morocco’s social protection system reform on multidimensional poverty as measured by the Multidimensional Poverty Index (MPI) and its components. These approaches differ in their way of targeting beneficiaries. The first approach bases the identification of beneficiaries on random selection. The second approach is more objective and is based on a probabilistic model (probit). By way of illustration, we use data from the 2018 Morocco National Survey on Population and Family Health (NPFHS). We present the results for a ponctual estimation and by confidence intervals. Finally, we perform a distributional analysis to overcome the arbitrary nature of the threshold setting associated with the measures and to make robust comparisons. Our results show that objective targeting is not necessarily better than random targeting for the simulations we conducted. Above all, they show that the comparison tool we propose can help to make decisions about targeting systems.