Using random utility models to incorporate substitution when estimating economic values and impacts of closing a unique recreation site: The case of snowmobiling in Yellowstone (Wyoming, Montana, Idaho)

Random Utility Models (RUMs) are used to incorporate substitution effects into economic impact and benefit estimates associated with closing a unique recreation site. While RUMs have been used extensively in benefit estimates, there has been little systematic research linking substitution effects and regional economic impacts. RUMs are used to quantify how substitution away from Yellowstone (YNP) and Grand Teton National Parks (GTNP) may affect snowmobile recreationist benefits and the economies of communities surrounding the parks, given a proposed snowmobile ban. The first essay analyzes the sample of snowmobile visitors and finds two distinct user types coming to YNP and GTNP. Nearly 60 percent of the sample is comprised of snowmobilers taking multiple snowmobile trips within the region (Wyoming, Idaho and Montana) surrounding the parks, but they typically visit one site per trip (single destination visitors). The other segment can be described as snowmobilers coming to the region for an annual winter vacation during which they visit multiple snowmobile sites (multiple destination segment). Results indicate visitor heterogeneity is likely an important consideration for economic analyses related to site closure. Essay two investigates the prediction of trip estimates of the single destination segment using revealed (RP) and stated preference (SP) data given site closure. Results indicate the SP data predict trip visitation loss to the region surrounding the park at levels five times that of the RP predictions, and the null hypothesis of preference homogeneity is rejected. Essay three draws upon economics, tourism and economic geography literature to develop a tractable and theoretically consistent framework with which to model the multiple destination segment. This model is consistent with theory and statistically significant in explaining behavior. Essay four investigates the policy implications of banning snowmobiles in YNP and GTNP. Ignoring heterogeneity potentially underestimates snowmobiler welfare losses and overestimates regional trip losses associated with site closure. Essay five uses the RUMs developed in earlier essays to drive trip and expenditure predictions used in regional economic impact estimation for the three states surrounding the parks. Results indicate that substitution effects and visitor heterogeneity are important considerations for regional economic impact analyses of recreation site closure.