This paper develops a model analyzing the first best policy for a social planner when pesticide use causes resistance, the level of pest infestation is heterogeneous among farmers, and the development of a backstop technology is an endogenous and uncertain process which depends on the cumulative amount of R&D effort.