Choice and temporal welfare impacts: incorporating history into discrete choice models

Static, cross-sectional discrete choice models are often employed to calculate welfare measures in environmental valuation. These measures will be biased if the underlying preference function includes temporally related attributes and lagged dependent variables, nevertheless, static models continue to dominate the literature. In this paper we provide an approach to estimating discrete choice models that includes consideration of prior behavior and past attribute perceptions and allows for testing consistency with random utility maximization in a time-series context. We apply this model to the case of recreational fishing site choice and participation, comparing our time-series approach to static versions. We find that the time-series model provides a richer behavioral characterization of site choice. We also find significant differences between cross-section and time-series welfare measures. The time-series model raises several concerns about the specification of the policy impact and the subsequent welfare measurement that are not apparent in the static case.