Joint Estimation of Multiple CVM Scenarios under a Double Bounded Questioning Format

Contingent valuation surveys frequently ask the same respondent for willingness to pay (WTP) for either different programs or different levels of provision of a single program. When multiple scenarios are considered by the respondent, the errors in the estimates of WTP are likely to be correlated across scenarios. Failing to account for correlation may lead to erroneous inferences concerning differences in WTP. This paper presents a technique that can be used to jointly estimate WTP for multiple scenarios proposed within a survey when the double-bounded questioning format is used. Monte Carlo simulations are employed to show that estimates derived from the joint model provide lower parameter variances as well as tighter confidence intervals surrounding WTP. The model is used to estimate WTP values for data collected in telephone interviews of California residents concerning WTP for fire reduction programs in Oregon and California. Variance properties of these estimates are shown to be similar to those estimated using simulated data.