Healthcare and education exhibit wide variation in spending that is loosely associated with outcomes. We study supply-side explanations for such variation in in healthcare, and extend this discussion to how it might apply to education. In both sectors, variation in risk-adjusted rates could arise from some providers or educators doing too much (overuse) or others are using too little (underuse). Alternatively, the production function varies across providers and educators, so that hospitals and educators with higher returns to treatment deliver more because of comparative advantage. We discuss how a prototypical Roy model can separate these explanations.