Possible Reasons for Variation in the WTP Values

Abstract

Contingent valuation (CV) studies are often criticized for being too hypothetical and therefore eliciting wrong WTP values. Especially the discrete CV method is regarded as being inappropriate. It is true that there are still many unsolved problems in this respect. For example, it is not clear which welfare measure should be favored. This is of great significance since results differ if either mean or median WTP is chosen (see section 1). Furthermore, WTP values depend substantially on the chosen elicitation technique. The PC method, for example, elicits WTP values, which are two to five times smaller than WTPs of the DC method. A possible explanation can be found in the design of the DC method. By giving respondents only a yes/no response alternative, yea-saying and protest answers may be provoked. In section 2 we examine whether the DM method, which is designed to avoid yea-saying and to control for protest answers, elicits smaller WTP values than the DC method. Another drawback for CV studies is that there is enormous potential for manipulating results. Therefore, we test for question ordering and payment vehicle bias in sections 3 and 4. Furthermore, we analyze in section 5 to see whether and how changing the maximum bid influences mean WTP values of discrete CV studies. In sections 6 and 7 we examine whether warm glow or information bias is present.