Abstract
This paper points out some problems of naive travel cost method (TCM) that considers only the travel cost to the study site. TCM needs demand function with regard to the entrance fee of the study site, however, only the demand function with regard to travel cost is observable. We treat two cases (i) there exists substitutes, and (ii) residential land rents are variable among sample data. It is proved that under these cases, the naive TCM, which is based on the demand function with only the travel cost to the study site, brings overestimation or underestimation of the ture benefits. In each of which cases, we propose how to sample data to identify the demand function appropriate for TCM using the market data.