Market models, protest bids, and outliers in contingent valuation

Responses to willingness-to-pay questions in contingent valuation (CV) surveys are being used with increasing frequency to obtain estimates of the value of public goods, such as improved water quality, for use in benefit-cost analyses. Analysts must decide whether the appropriate conceptual model for any particular application is a private or a political market. When estimating willingness-to-pay and total benefits, analysts must decide whether to censor outliers and certain zero responses, called protest bids, from data sets. Issues associated with identification of protest bids are reviewed, and alternate methods for the treatment of outliers are discussed. The significance of these methodological decisions is illustrated with an example involving willingness to pay for new storm water programs in Baltimore County, Md. to help achieve Chesapeake Bay nutrient-reduction objectives. It is argued that decisions concerning analysis of troublesome responses should be consistent with the choice of market model for the application.