Contingent valuation (CV) techniques are commonly used to value the potential effects of a policy change when market-based valuation of those effects is not possible.  Although policy changes often have multiple effects—some of which would generally be considered positive and others negative—many CV studies constrain respondents’ willingness to pay for the effects to be no lower than zero (e.g., Welsh et al. 1995).  That is, individuals are constrained from expressing a positive value to avoid the effects of a policy change.  Such a constraint may be problematic in the case of policy changes where the negative effects of the change have the potential to outweigh the positive ones.   

Using the control of water flowing through the Glen Canyon Dam as our empirical application, in this paper we present an approach that allows survey respondents in CV studies to express a negative valuation of the effects of a policy change.  We also assess whether such an approach has the potential to meaningfully alter willingness to pay estimates.  More specifically, we provide a sample of survey respondents with significant background information on the Glen Canyon Dam and describe how its operation relates to electricity production, as well as natural and cultural resources in the surrounding area.  After providing respondents with this background information, we randomly assign them to one of two groups.  Members of the first group—the control group—are presented with a proposal for changing dam operations and are provided with a list of the estimated effects of the change across multiple dimensions.  Following standard CV procedures, we ask control group members if they would vote for the proposal if it cost them nothing, and then ask if they would vote for the proposal if it cost them a given positive dollar value.  Respondents declining to vote for the proposal at no cost are assumed to have zero willingness to pay for the policy change.

Respondents assigned to the second group—the treatment group—are not presented with a proposal for changing dam operations, but with two options for operating the dam, as well as estimates of the relative effects of each option.  After presenting respondents with this information, we ask them if they would vote for option 1 or option 2 if adoption of either option was guaranteed to cost them nothing.  After casting this vote, we then ask respondents if they would vote for the option if its adoption would cost them a given positive dollar value.  By allowing respondents to first identify their preferred set of effects—and accompanying policy option—and then asking them about their willingness to pay for obtaining the set of effects, we effectively allow respondents to express a negative valuation of the effects of a potential policy change.  We compare the results for the treatment and control groups and discuss their implications for policy decisions and future research utilizing CV techniques.