The literature on environmental R&D frequently studies innovation as a two-stage process, with a single R&D event leading from a conventional polluting technology to a perfectly clean backstop. We allow for uncertainty in innovation in that the new technology may turn out to generate a new pollution problem. R&D may therefore be optimally undertaken more than once. Using and externding recent results from multi-stage optimal control theory, we provide a full characterization of the optimal pollution and R&D policies. The optimal R&D program is strictly sequential and has an endogenous stopping point. Uncertainty drives total R&D effort and its timing.
stock pollution backstop technology multi-stage optimal control pollution thresholds uncertainty