Two decades of research into educational production functions have produced startlingly consistent results: Variations in school expenditures are not systematically related to variations in student performance. Enormous differences in teacher quality exist, but differences in teacher skill are not strongly related to educational backgrounds, amount of teaching experience, or teaching in small classes. Further, more skilled teachers simply are not regularly paid more than less skilled teachers. These findings suggest that school decision making must move away from traditional "input directed" policies to ones providing performance incentives. The concentration on expenditure differences in, for example, school finance court cases or legislative deliberations, appears misguided given the evidence.