دانلود رایگان مقاله پیشنهاد کوتاه در ارزشگذاری مشروط: برنامه برای گونه پرستو در معرض خطر

Abstract

A yes-response function in a contingent valuation study is said to have fat tails if it has a high and slowly declining yes-response rate at high bid levels. Truncated bids refer to the practice of dropping high bid offers before a yes-response rate of near zero is reached. This is a common practice in contingent valuation. We explore the extent and implications of fat tails and truncated bids in a study of an endangered shorebird species. We find, among other things, that mean willingness to pay is quite sensitive to the highest bid offered – so much so that the choice of highest bid nearly dictates outcomes.

5. Discussion

Consider Table 1 again. Based on our findings, we are left wondering what would have happened if higher bids had been considered in many of these studies where the yes-response function is truncated. While we cannot say for sure, we suspect they may have had findings similar to ours: a difficulty pinning down the tail of the yes-response function and a mean willingness to pay estimate that is highly sensitive to choice of maximum bid and perhaps implausibly high at extreme bids. It would be interesting to test their surveys. Consequentially has become an important issue in contingent valuation (Herriges et al., 2010). In order for respondents to provide meaningful data, they need to believe that the survey is consequential and that their responses matter for policy purposes. At least two recent studies listed in Table 1 are designed to address consequentiality (Petrolia et al., 2014; Herriges et al., 2010). Both appear to have fat tails, suggesting that a lack of consequentiality may not be the issue. Obviously, more is need here to draw definitive conclusions. Again, we see fat tails as a manifestation of hypothetical bias (the tendency of people to report a value other than their true value due to the hypothetical nature of a survey) and not an isolated contingent valuation issue. Fat tails is, after all, consistent with most contingent valuation phenomenon believed to cause hypothetical bias: yea saying, anchoring, using valuation questions to express emotive instead of trade-off values, using valuation questions to show support for a program, etc. Viewed in this way, fixing fat tails amounts to fixing the fundamental hypothetical bias presence in contingent valuation. Truncating high-end bids is a tempting response to fat tails. If the tail of the yes-response surface is ignored over its high end, the analyst may offer truncated values using a lower-bound nonparametric estimator as a conservative value. But, this is not a real fix to the underlying problem of hypothetical bias, nor is the resulting willingness to pay truly conservative. Indeed, it “hides” the effects of fat tails. One may falsely believe that she has a reasonable estimate of value when in fact the survey instrument could produce vastly different values with only modest changes in the bid levels offered. Truncating offers nothing new for understanding underlying preferences, explaining why contingent valuation data yield fat tails, or dealing with hypothetical bias.