Historically, most attention in public programs has been given to the resources devoted to the activity, and resources have been used to index both commitment and quality. Education differs from other areas of public expenditure because direct measures of outcomes are available, making it is possible to consider results and, by implication, to consider the efficiency of provision. Early interpretations of the evidence, emanating from popular interpretations of the Coleman Report that 'schools do not make a difference,' are incorrect, but the basic evidence behind the statement suggests serious performance problems of government supply, because purchased inputs to schools are not closely related to outcomes. This paper reviews that evidence along with providing an evaluation of the various controversial aspects including issues of causality, consumer behavior, and estimation approaches. Two detailed policy areas are discussed in terms of the evidence on performance: public versus private provision and the financing of schools.