A nonparametric test of the traditional travel cost model

The traditional travel cost model operates on the assumption that choices are made regarding the number of trips to various sites over an entire recreation season. This paper uses actual recreation data to test the consistency of this model with postulates of rational economic choice as embodied in the axioms of revealed preference. The paper assumes consumer rationality and examines the effects of alternative trip prices and model structures on the degree of consistency between them and the axioms of choice. Using tests developed by Varian and Tsur, the authors find that site choices at prices proxied by travel costs for most individuals in the sample violate the axioms. The violations are quite large when time costs, both in traveling and on-site, are omitted from the prices. When they are included, violations are almost as numerous, but not as large. The paper also examines demand heterogeneity by stratifying the sample and repeating the nonparametric tests. These findings suggest caution be used in interpreting welfare measures derived from traditional travel cost models, especially those that do not include measures of time value. The authors' approach provides a method for screening travel cost data and preparing for model selection and estimation.