A Test of Benefits Transfer of Random Utility Models of Recreation Demand

In this paper, I examine the performance of function transfer of random utility models (RUM) and show that the usual notions of reliability and validity need careful examination in the RUM context. Because in RUM models, estimated behavioral parameters are not identifiable independently of the scale parameter, a function transfer carries with it information about preferences and the error structure at the study region. Using a Monte Carlo experiment, I investigate the viability of function transfer when behavioral parameters are assumed to be equal but differences in scale exist for the policy and study regions. The results show that properties of the error structure can impact the reliability and validity of the function transfer method in RUM's under the most ideal of situations- when the behavioral parameters of the two populations are equal.