Hybrid seminar: Mental Health and the Targeting of Social Assistance

Is social assistance well-targeted to people with poor mental health? Mental illness is disproportionately prevalent among the poor, implying a need for access to income-support schemes. However, psychological evidence suggests administrative burdens have higher incidence on people with poor mental health. This generates concern about whether ordeals are screening the mentally-ill efficiently. This question is important because the welfare consequences of such an inefficiency are undocumented so far. Using administrative data from the population of the Netherlands, I find that eligible people with moderately poor mental health receive social assistance more than the healthy, however no evidence that the seriously-ill receive more than moderates. My theoretical framework suggests this is driven by need vs ordeal-incidence. I characterize the normative consequences of selection by deriving sufficient-statistics formulas for marginal welfare effects of policies affecting targeting. I exploit an important benefits-reform and find that increased ordeals disproportionately screen out the mentally-ill. This suggests potential welfare loss from inefficient targeting.