Empirical research in the micro productivity literature consistently supports the notion that there is too little R&D. However, the methodology of this literature, based on the neoclassical growth model, is challenged by new growth theory, which emphasizes a richer description of the relationship between R&D and productivity. In particular, it allows for incentives that lead to overinvestment in R&D. We incorporate several distortions to R&D into a general equilibrium growth model that provides a framework for the analytical and empirical analysis of the degree of over- or underinvestment in R&D. We derive the relationship between the social rate of return to R&D and the parameters estimated in the productivity literature. Surprisingly, our results indicate that estimates in the productivity literature represent lower bounds on the social rate of return to R&D and that the bias is limited to the overall growth rate of the economy. Additional supporting evidence for underinvestment is provided by the implied equilibrium R&D share from a calibrated version of the theoretical model.