The paper by Charles I. Jones and John C. Williams examines the social return to R&D and challenges the empirical findings that suggest a high rate of return to R&D, which could indicate underinvestment. The authors argue that the empirical estimates of the return to R&D represent a lower bound on the true social return, which is significantly higher due to factors like knowledge spillovers, intertemporal effects, and creative destruction. Using a conservative estimate of the social return at 30%, they show that optimal R&D investment is at least four times larger than actual investment. The paper begins by discussing the empirical literature that estimates the return to R&D, which suggests that R&D spillovers are significant and that social returns are higher than private returns. However, the authors argue that the neoclassical approach used in these studies ignores important distortions in the R&D process, such as monopoly pricing, intertemporal knowledge spillovers, and congestion externalities. By linking the new growth theory to the empirical results, they derive the relationship between the social rate of return to R&D and the coefficient estimates from productivity growth regressions. The authors show that the empirical estimates of the return to R&D are lower bounds on the true social return. This is because the empirical approach does not account for the dynamic effects of knowledge spillovers and the capital gain or loss associated with changes in the value of ideas over time. The social rate of return is derived from the production possibilities of the economy, including the production function for ideas and the production function for the consumption/output good. This approach does not rely on specific assumptions about market structure, the patent system, or distortionary taxes. The paper concludes that the findings of the empirical productivity literature are robust, even when considering the distortions highlighted by the new growth theory. The social rate of return function provides a clear mapping between the extent of underinvestment in R&D and the social return. Using a conservative estimate of the social return at 30%, the authors show that optimal R&D investment is more than four times larger than actual investment. This suggests that advanced economies like the U.S. significantly underinvest in R&D.The paper by Charles I. Jones and John C. Williams examines the social return to R&D and challenges the empirical findings that suggest a high rate of return to R&D, which could indicate underinvestment. The authors argue that the empirical estimates of the return to R&D represent a lower bound on the true social return, which is significantly higher due to factors like knowledge spillovers, intertemporal effects, and creative destruction. Using a conservative estimate of the social return at 30%, they show that optimal R&D investment is at least four times larger than actual investment. The paper begins by discussing the empirical literature that estimates the return to R&D, which suggests that R&D spillovers are significant and that social returns are higher than private returns. However, the authors argue that the neoclassical approach used in these studies ignores important distortions in the R&D process, such as monopoly pricing, intertemporal knowledge spillovers, and congestion externalities. By linking the new growth theory to the empirical results, they derive the relationship between the social rate of return to R&D and the coefficient estimates from productivity growth regressions. The authors show that the empirical estimates of the return to R&D are lower bounds on the true social return. This is because the empirical approach does not account for the dynamic effects of knowledge spillovers and the capital gain or loss associated with changes in the value of ideas over time. The social rate of return is derived from the production possibilities of the economy, including the production function for ideas and the production function for the consumption/output good. This approach does not rely on specific assumptions about market structure, the patent system, or distortionary taxes. The paper concludes that the findings of the empirical productivity literature are robust, even when considering the distortions highlighted by the new growth theory. The social rate of return function provides a clear mapping between the extent of underinvestment in R&D and the social return. Using a conservative estimate of the social return at 30%, the authors show that optimal R&D investment is more than four times larger than actual investment. This suggests that advanced economies like the U.S. significantly underinvest in R&D.