A New Estimator for Dynamic Resource Models, with an Application to Reservoir Management

• Environmental and Energy Economics Workshop

Date and Location

Thursday, June 8, 2023, 2:10 PM - 3:30 PM
ARE Library Conference Room, 4101 Social Sciences and Humanities

Abstract

Important innovations in applied econometrics over the past several decades have enabled significant advances in environmental economics. This literature focuses on causal identifi- cation in reduced-form models, and is primarily concerned with causal inference of policy or environmental variables on outcomes of interest. In parallel, a large literature in resource economics uses dynamic models to derive theoretical principles about incentives, institutions, and behavior in dynamic resource settings. Yet only a small, niche literature actually tries to bridge these two approaches to environmental economics. In this paper we derive a new, practical approach to econometric estimation of structural parameters of dynamic resource models. Our estimator leverages a result on stochastic Euler equations that eliminates the expected value function, the main hurdle in the estimation of dynamic structural models. We then apply the GMM approach of Hansen and Singleton (1982) to estimate the stochastic Euler equation under the assumption of rational expectations. In contrast to earlier studies that focus on discrete-choice problems, our estimator can be applied to continuous-choice dynamic problems, consistent with the canonical economic models of natural resources. We demonstrate our approach with an application to reservoir management in California, illustrating how estimation of structural parameters reveals the relative importance placed by managers on water storage and releases.