Job Market Paper
The Recreation Information Database (RIDB) is part of the Recreation One Stop (Rec1Stop) project, a web-based single point of access for recreational opportunities including camping on federal lands. We use a random-utility destination choice model to explore the factors that affect demand for visits to particular campgrounds. Trip origins in the available data are identified only at the zip code level, but we use zip code centroids and campground locations to calculate approximate travel distances and travel times. We use zip code median income levels to calculate an approximate measure of the opportunity cost of time as another component in the full round-trip travel costs for alternative destinations. We have detailed information about the various activities that are supported at each destination, and detailed attributes for each campground. The analysis presented is a proof-of-concept for a more comprehensive analysis currently underway. This more comprehensive model examines the role of intertemporal substitution by defining the consideration set over all available sites at all available dates over the coming season. This modeling specification leads to enormous consideration sets (roughly 3500 alternatives per choice occasion) that we sample from, but provides a novel way of accounting for intertemporal substitution.
Works in Progress
Valuing Environmental Amenities Using a Random Utility Maximization Model with Endogenous Trip Timing
Most recreational demand models in the literature assume that visitors choose a particular site from a range of options available on a particular choice occasion. These model substitution between sites that occurs because of differences in site quality, but fails to capture the role of intertemporal substitution in recreational trips. Even models that include the participation option and repeated choice occasions do not fully capture the role of intertemporal substitution. In this paper, I use a novel way of defining the consideration set as over all available sites at all available dates over the coming season. By doing so, I explicitly capture the role of intertemporal substitution in the visitor’s decision making process. The empirical setting uses campground reservation data that details when the reservation was made and the full range of options available at that date. Conditional and mixed logit models are estimated and welfare measures are presented for a variety of policy scenarios. Results are compared for specifications that do and do not allow for intertemporal substitution to reveal the importance of this modeling decision.
Incorporating Endogenous Consideration Set Formation into a Recreational Demand Model of Trip Choice
This research generalizes a previously developed Random Utility Model of recreational destination choice to allow for the endogenous formation of consideration sets. I estimate the probability that each possible destination/date combination "trip" is in an individual's consideration set as a function of the observable attributes of that trip and respondent characteristics, adapting the approach taken by Li et al. (2015). I then compare the marginal utility estimates for different site attributes between (a) models which impose a deterministic consideration set and (b) richer specifications that allow for endogenous consideration sets. Different utility parameters also imply differences in the implied welfare effects of different policy scenarios, such as business-as-usual for climate change versus mitigation efforts, or the closure of a particular destination during a particular part of the season. My illustrative application employs campground reservations from the Recreational Information Database (RIDB) maintained by the U.S. Forest Service.