Heterogenous Treatment Effect in Development Policy: A Randomized Experiment in Morocco on Those Who Take up the Microcredit

Randomized experiments are widely used in analyzing policy implementation and evaluation as researchers can estimate the average treatment effect between the control and treated groups. Sometimes, treatment effects are variable in different subgroups. The nonrandom variability defined by heterogeneous treatment effect is not like the random variability which is not correlated with explanatory variables and can be fixed by statistical methods, but it estimates individual treatment effects depending on individuals’ characteristics within a subgroup. Using DID estimation, we will uncover treatment effects and heterogeneity in policy evaluation, helping policymakers to evaluate the efficacy of policy implementation. The study is mainly based on the effects of microcredit in 162 villages in Al Amana, a rural area in the Kingdom of Morocco (Crépon et al., 2015).