The public R&D capital stock is introduced as a quasi-fixed input in a variable cost function. The relative shadow price allows the correct measurement of the equilibrium levels of quasi-fixed inputs thus explicitly assessing the hypothesis of public R&D under (over) investment. By introducing an appropriate R&D price in the long-run equilibrium, the model can also provide empirical evidence on the rationale driving public R&D investment and on the hypothesis that free-riding on public R&D can explain overinvestment. Moreover, the model allows a formal testing of the induced innovation hypothesis and a more accurate calculation of both internal rate of return to R&D and residual exogenous productivity growth. The empirical implications of the model are appraised in the case of Italian agriculture for the period 1960-1995.