Comment: 30 pages, 1 table, 3 figuresThe econometric literature on treatment-effects typically takes functionals of outcome-distributions as `social welfare' and ignores program-impacts on unobserved utilities. We show how to incorporate aggregate utility within econometric program-evaluation and optimal treatment-targeting for a heterogenous population. In the practically important setting of discrete-choice, under unrestricted preference-heterogeneity and income-effects, the indirect-utility distribution becomes a closed-form functional of average demand. This enables nonparametric cost-benefit analysis of policy-interventions and their optimal targeting based on planners' redistributional preferences. For ordered/continuous choice, utility-distributions can be bounded. Our methods are illustrated with Indian survey-data on private-tuition, where income-paths of usage-maximizing subsidies differ significantly from welfare-maximizing ones.