. . "The best argument around for state subsidies for research is that research is an industry that has its impact by *expenditures* rather than *outputs*." . .