Impact of Social Protection Programs on Multidimensional Poverty: New Targeting Approaches and Application to Morocco

Two new methodologies are developed to approach the impact of the reform of the social protection system in Morocco on multidimensional poverty as measured by the MPI introduced by OPHI, and its components. These approaches are distinguished by their mechanism for targeting beneciaries. The rst approach bases the identication of beneciaries on random selection. The second approach is a priori more objective and based on a probabilistic model (probit). As an illustration, we use data from Morocco's 2018 Enquête Nationale sur la Population et la Santé Familiale (ENPSF). We consider three reform scenarios targeting health and education indicators. We present the results in ponctual form and by condence intervals constructed using both Monte Carlo and bootstrap approaches. Finally, we perform a distributional analysis to overcome the arbitrariness of setting the poverty line associated with the measures and make robust comparisons. Our results show that the three simulated reforms have a positive eect on the multidimensional poverty measures, regardless of the approach used. At the methodological level, targeting by objective identication does not necessarily dominate random targeting for the simulations we conducted, possibly because of the highly unbalanced sample.