THREE ESSAYS IN THE ECONOMICS OF RECREATION DEMAND (TRAVEL COST METHOD, NATIONAL RESOURCE)

Three essays have been developed to improve the theory and the estimation of the travel cost method which is the most commonly used method to describe the demand for outdoor recreation. Theoretical issues are discussed in the first two essays, and econometric estimation is discussed in the final essay.;In the second essay, the Slutsky equation for the derived input demand function is developed. Using the Slutsky equation, useful prior restrictions and hypotheses of the impacts of input price changes on the demand for inputs are deduced. Further, it is shown that welfare measures, based on the input demand equation, are appropriate measures of the net value of individual recreation sites.;The third essay corrects an estimation problem in previous estimates of the TCM using on-site samples. There are two sources of the problem in on-site samples: truncation and endogenous stratification. Four theoretically correct methods are developed to solve the two problems. A numerical experiment is done to compare the four methods using generated data sets of a known model. Two methods are shown to be superior: (a) the maximum likelihood method solving the two problems together, and (b) normalizing for endogenous stratification and then using a maximum likelihood method to solve the problem of truncation. Using recreation data from sites in Montana, ordinary least squares (OLS) is compared with the maximum likelihood method. OLS significantly overestimates the individual net social benefits by factors of 1.3 (TURN) 2.3. Thus, we should not use OLS and instead should use the maximum likelihood methods developed here to estimate recreational site demand functions using on-site samples.;In the first essay, I demonstrate that historically the travel cost method (TCM) has been used to measure the demand for household commodities (recreation experiences). This approach, unfortunately, is full of empirical hazards mainly due to the endogeneity of household commodity prices. An alternative strategy is to use the TCM as a derived input demand function. In this case, the price endogeneity problem disappears. The appropriate unit for the dependent variable is the consumer's number of trips to the site. Only prices of inputs and full income are appropriate arguments in the demand function. The widespread use of total cost that includes travel cost, on-site time, and money outlays is a misspecification. On-site time and on-site expenditures should be excluded from the model.