REPRESENTING SUBSTITUTION EFFECTS IN MODELS OF RECREATION DEMAND (CONSUMER SURPLUS, WILLINGNESS-TO-PAY, TRAVEL COST)

The purpose of this study was to determine how sensitive economic estimates of the value of recreation sites are to the manner in which the influence of substitute recreation sites is modeled. The average consumer surplus per person, per trip at eleven different reservoirs operated by the U.S. Army Corps of Engineers was estimated using three different zonal travel cost models (TCMs). Each TCM differed with respect to its treatment of substitute recreation sites.;The reservoirs are located in Kansas and Missouri. The data were drawn from on-site interviews conducted during the summer of 1982, representing approximately 60,000 day-use recreationists. Each version of the TCM used the same 194 origins; most of them being counties in Kansas and Missouri. Logistic regression analysis was used to account for origins from which no visitation was recorded.;An analysis of variance test showed that the three versions of the TCM do not give equivalent estimates of economic value (F = 26.2 with 2,20 degrees of freedom). The overall averages per person, per trip surplus (1982 dollars) for versions one, two, and three were ;In the first version of the TCM, the influence of substitute recreation sites was totally ignored. The second version included proxy variables, representing the price of substitute sites, as predictors in the demand equation. A principal components analysis of the substitute sites' prices was used to form the proxy variables. The third version modeled recreation demand using a gravity model type formulation. The trip distribution portion of this gravity model was estimated with a multinomial logit model.