In this paper, I exploit the cross-sectional and time-series variation in R&D tax credits, and in turn the user cost of R&D, available from U.S. states between 1981-2002 to estimate the elasticity of private R&D with respect to both the within-state (internal) user cost and the out-of-state (external) user cost. To facilitate comparisons to previous studies of the R&D cost elasticity, I first estimate an R&D cost elasticity omitting external R&D costs; the estimated elasticity is negative, above unity (in absolute value), and statistically significant—a finding quite similar to that found by previous studies based on alternative data. Unlike previous studies, however, I then add the external R&D user cost to the regressions. I find the external-cost elasticity is positive and significant, raising concerns about whether having state-level R&D tax credits on top of federal credits is socially desirable. More importantly, I find the aggregate R&D price elasticity—the difference between the internal- and external-cost elasticities—is far smaller than previously estimated. In fact, the preferred specification yields a zero aggregate elasticity, suggesting a zero-sum game among states and raising questions about the efficacy of R&D tax credits more broadly.