Travel cost and time measurement in travel cost models TRAVEL COST AND TIME MEASUREMENT IN TRAVEL COST MODELS

In this paper we use the results of a recent on-site recreation survey to know whether it matters to use perceived rather than calculated travel costs. The calculated travel costs (time and money) are derived from detailed GIS computations. The perceived travel times and costs are taken directly from the survey responses. We first find that there is a high non-response rate for the perceived cost question. For those who respond we find that the perceived costs are close to the sum of the calculated time and fuel costs. The relative difference between perceived and calculated costs is negatively related to distance and visit frequency. We explain perceived costs by simple factors such as distance, time, group size, social class, education level, length of stay and general satisfaction with their visit. To examine the effect of the specification of the travel cost and time variables, we estimate several individual trip demand equations for three functional forms of the dependent variable: semi-log, negative binomial and truncated negative binomial. Based on the log-likelihood values of the estimation of the different trip demand equations we show that there are no significant differences in model performance as far as cost specifications are concerned. Comparing cost coefficients and consequently consumer surplus estimates shows that only for the negative binomial functional form there is a statistically significant difference between the specification with perceived costs and the model with total calculated cost measures. 1 The research reported in this paper was financed by the ‘Vlaams Impulsprogramma Natuurontwikkeling’ (contract C96/06). The authors would like to thank P. Nunes and E. Van Ierland for many valuable comments and suggestions during internal workshops. All errors remain the authors’.