Conference Agenda

In stated preference (SP) surveys, theoretical restrictions can be used in two ways by imposing the restrictions as is often done in attribute non-attendance (ANA) or testing the restrictions as in whether the data meet the specific requirements of ‘adding up’. However, one could just as easily impose the ‘adding up’ restriction rather than test for its existence. We propose a situation where this type of adding-up restriction may be valuable. In our application, survey respondents are asked to gauge their support for a measure to help their struggling municipal wastewater treatment plant meet required discharge levels. First participants are offered a choice between expanding or not expanding (status quo) traditional treatment capacity. As a follow-up, respondents choose whether to support expanding treatment capacity using alternative ecologically engineered treatment technology (EETT) in the form of constructed wetlands. Finally, participants are asked to choose their preferred wastewater treatment expansion choice among the two alternatives. Naïve estimates suggest that households are willing to pay (WTP) roughly the same amount for traditional expansion and wetland treatment expansion, but also exhibit a strong statistically significant WTP premium for wetlands relative to traditional wastewater treatment. Clearly, the results violate the scope test, but after imposing the adding up restriction the WTP estimates decrease for traditional wastewater treatment and increase for wetland alternatives yielding reasonable estimates of the wetland premium. These findings generally hold for a variety of model specifications including mixed logit and ANA. Finally, we introduce a fractional response mixed logit model to accurately handle instances when survey respondents report having no rank-order preference between EETTs and traditional wastewater treatment.