Impact Evaluation and Welfare Evaluation

摘要: The expression “impact evaluation” means different things to people, but most economists now it the use of randomized controlled trials (RCTs) or “quasi-experiments.” I want focus on that side research and argue that, even when statistical circumstances are ideal, constitutes at best an intermediate input into welfare evaluation interventions. That is valuable, should not be confused with final product, a proper costbenefit analysis. summarize arguments here developed greater length, literature citations, in Harrison [2011a][2011b][2013]. There several simple reasons such impact evaluations (IE) fall short. first they limit themselves observables: this price change delivering product leads what revealed demand? second average effects: third their has been partial equilibrium: effect for commodity am able randomize (or study randomization of)?