THE TRAVEL COST METHOD: TIME, SPECIFICATION, AND VALIDITY (RECREATION DEMAND, LOGIT MODEL, MARKET VALUATION)

The travel cost method uses the cost of travelling to a recreation site as a proxy for the price variable in a demand study. One problem in this method is the treatment of travel time. It is well established in the literature that travel time needs to be included in the model. To date, it has been included in an ad hoc manner. This dissertation considers travel time in a model of a utility maximizing individual, and attempts to empirically estimate the theoretically correct value.;A third problem is not unique to the travel cost method. It is shared by all methods of imputing a value to a nonmarket good. While theoretically sound, very little work has been done to assess the accuracy of the results of these methods vis-a-vis values generated in a fully functioning market. This dissertation compares values from a travel cost study to values generated in a simulated market.;The traditional travel cost model and the probabilistic travel cost model were estimated using several values of travel time. The best fit of the data was obtained when travel time was valued in the range of 10% to 33% of the wage rate. Consumer surplus estimates from the best-fitting travel cost models were compatible with those observed in the simulated market. In addition, a distribution function estimated from the probabilistic travel cost model was not statistically different from that generated by the simulated market data. The conclusion was that a travel cost model using a low, but nonzero, value of travel time provides a credible indication of the value of a nonmarket good.;A second problem in the travel cost method is a truncated sample problem. Travel cost studies are usually done using samples of users of a recreation site, or participants in an activity, rather than the general population. Ignoring nonparticipants can lead to biased estimates of demand coefficients. Methods are available to correct for the bias in individual travel cost models. Zonal travel cost models, using aggregated data, are more widespread in practice. It is not clear how, or if, these models can be corrected for the bias resulting from ignoring nonparticipants. This dissertation proposes a theoretical framework, capable of incorporating nonparticipants, in which to consider a zonal travel cost model. This probabilistic travel cost model looks at the probability that a person from a given travel zone will visit a recreation site.