An examination of sources of sensitivity of consumer surplus estimates in travel cost models.

We examine sensitivity of estimates of recreation demand using the Travel Cost Method (TCM) to four factors. Three of the four have been routinely and widely discussed in the TCM literature: a) Poisson verses negative binomial regression; b) application of Englin correction to account for endogenous stratification; c) truncation of the data set to eliminate outliers. A fourth issue we address has not been widely modeled: the potential effect on recreation demand of the interaction between income and travel cost. We provide a straightforward comparison of all four factors, analyzing the impact of each on regression parameters and consumer surplus estimates. Truncation has a modest effect on estimates obtained from the Poisson models but a radical effect on the estimates obtained by way of the negative binomial. Inclusion of an income-travel cost interaction term generally produces a more conservative but not a statistically significantly different estimate of consumer surplus in both Poisson and negative binomial models. It also generates broader confidence intervals. Application of truncation, the Englin correction and the income-travel cost interaction produced the most conservative estimates of consumer surplus and eliminated the statistical difference between the Poisson and the negative binomial. Use of the income-travel cost interaction term reveals that for visitors who face relatively low travel costs, the relationship between income and travel demand is negative, while it is positive for those who face high travel costs. This provides an explanation of the ambiguities on the findings regarding the role of income widely observed in the TCM literature. Our results suggest that policies that reduce access to publicly owned resources inordinately impact local low income recreationists and are contrary to environmental justice.

An assessment of household willingness to pay for curbside recycling: a comparison of payment card and referendum approaches.

Curbside recycling is a tool that communities use to reduce the need for landfill space. This study provides contingent valuation estimates of household willingness to pay (WTP) to continue a curbside recycling program in the face of budget cuts. Comparisons of two forms of the contingent valuation method (CVM) are provided: a single bounded referendum and a payment card. Neither approach emerges as unambiguously superior. Response rates were virtually identical. Both approaches show that support for curbside recycling is highly sensitive to price. Regression results from the payment card provided a more thorough identification of socio-demographic variables associated with WTP than the referendum, but the explanatory power of the two regressions did not differ significantly. The referendum estimates of mean WTP exceed those from the payment card, although the disparities are less than those typically reported in the CVM literature. Local policy makers cited the CVM results as influencing their decisions regarding funding options for the future of the program, and seemed to appreciate the fact that the two approaches provided a fairly narrow range of estimates of WTP. In an era when more of the burden of financing of environmental programs is being shifted to the local level, use of CVM to estimate the WTP of consumers for highly disaggregated goods and services designed to achieve environmental improvement will likely become more relevant to local decision makers who are interested in understanding their constituents' views.