Files
Abstract
This dissertation comprises three manuscripts which examine the recreational travel cost literature and consider new approaches to data management and demand modelling that improve the statistical efficiency and accuracy of standard travel cost methods and applications. Together, these papers provide valuable insights and recommendations for future econometric applications of the travel cost model.Revealed preference methods require survey data on past resource use, and numerous studies have found reported recreation frequency to be overestimated and concentrated on prototype values. Our first paper develops two approaches to treat extreme values and rounded responses. We illustrate how, when modeling single-site trip data using a negative binomial (NB) distribution, employing the incomplete beta function simplifies the incorporation of censored intervals. We show the NBs fit is improved by either reassigning rounded responses to censored regimes where reported trip numbers define the intervals upper bounds, or by mixing the NB with a continuous distribution at a cut-point where response behavior begins to exhibit rounding.Much of the travel cost literature uses mixed logit (MXL) models to evaluate recreational site choice data. Multinomial probit (MNP) models are less common, as they have been difficult to work with historically. Our second paper compares these models performances and explores implications for welfare analysis in the case of multi-site trip data. Utilizing a new, more efficient approach (dubbed the Delta Method Approximation) for estimating the distribution of the mean benefit from policy implementation in MNP models, we discuss the merit of increasing MNP models prevalence in non-market valuation studies.North Carolinas beaches are imperiled by coastal erosion, sea level rise, severe storms, and oceanfront development. Proposed solutions to these problems include beach replenishment, coastal retreat, and shoreline armoring. These policies affect the quality and value of coastal resources and recreation, and assessing these welfare impacts is necessary for benefit-cost-analysis of these alternatives. Our third paper analyzes multi-site trip data for North Carolina households using travel costs and site attributes. We employ a MXL model in our recreation demand analysis and discuss the advantages of incorporating a Kuhn-Tucker generalized corner solution model in future extensions of this analysis.