This article studies how a government should distribute funds among research institutions and how it should allocate them to basic and applied research. Institutions differ in reputation and efficiency and have an information advantage. The government should award funding for basic research to induce the most productive institutions to carry out more applied research than they would like. Institutions with better reputation a do more research than otherwise identical ones, and applied research is inefficiently concentrated in the most efficient high‐reputation institutions. The article provides theoretical support for a dual‐channel funding mechanism but not for full economic costing.