On backstops and boomerangs: Environmental R&D under technological uncertainty

In areas such as climate change, the recent economic literature has been emphasizing and addressing the pervasive presence of uncertainty. This paper considers a new and salient form of uncertainty, namely uncertainty regarding the environmental characteristics of 'green' innovations. Here, R&D may generate both backstop technologies and technologies that turn out to involve a new pollution problem ('boomerangs'). In the optimum, R&D will therefore typically be undertaken more than once. Extending results from multi-stage optimal control theory, we present a tractable model with a full characterization of the optimal pollution and R&D policies and the role of uncertainty. In this setting, (i) the optimal R&D program is defined by a research trigger condition in which the decision-maker's belief about the probability of finding a backstop enters in an intuitive way; (ii) a decreasing probability of finding a backstop leads to the toleration of higher pollution levels, slower R&D, a slower turnover of technologies, and an ambiguous effect on the expected number of innovations; (iii) learning about the probability of a backstop is driven by failures only and leads to decreasing research incentives; and (iv) small to moderate delays in the resolution of technological uncertainty do not affect the optimal policy.