This paper studies the cost efficacy of universal over means-tested (targeted) programs, taking advantage of the rich diversity in state rules governing access to state-funded preschool in the U.S.