My "featured lunch speaker" presentation at CNREP

Here is the PDF: Download Contingent Valuation - from dubious to hopeless

Here is the text: http://econpapers.repec.org/paper/aplwpaper/13-07.htm

The gist is I don't think the contingent valuation method is all that bad.

via www.env-econ.net

Thanks for taking the lead on this John.

Comments

[Emily Zankman]

In John C. Whiteheads presentation “Contingent Valuation: From Dubious to Hopeless?”, he highlights that Hausman “selectively reviews the CVM literature and fails to find progress over the 18 years since Diamond and Hausman argued that unquantified benefits and costs are preferred to benefits and costs quantified by CVM for policy analysis”. In his presentation, he gives counter examples to show that the CVM debate is ongoing- it is not hopeless as some claim. He also states that the issues of hypothetical bias, WTP vs WTA and scope still exist, but the difficulties they pose are not in any sense insurmountable.

[Cort Hammond]

Admittedly, there are a number of things that I don't quite understand about Hausman's argument. For example, why does he believe that CVM is ok for private firms but not for government work? I can only imagine that he objects to the idea of wasting government money? Anyhow, it seems that as long as CVM undershoots the actual non-market value of something, then it is better than ignoring the value and allowing the tragedy of the commons to continue at the same rate. Of course, Hausman seems to indicate that only revealed preference methods are truly acceptable since they avoid issues with bias. However, he should recognize that these methods require data that may be unrecorded or difficult to collect, especially in developing countries. For example hedonics cannot be applied well to finding the value of a beach when there is a situation where there are few hotels all on a beach. In this case there would be very little to compare with. I would agree that CVM has its place and that as long as the estimated price is not above the actual, it is an imperfect solution, but better than none.

[David Fishman]

The argument over the legitimacy of contingent valuation practice’s ability to quantify willingness to pay curves poses an interesting topic. When using any non-market value valuation technique it is important to remember that there is always some amount of imperfections that will be generated from the technique. These techniques produce estimates and are, as Hausman points out, subject to many biases and other potential issues. However, I do not believe that one can write off contingent valuation methodologies on the bases listed by Hausman. Some of the concerns highlighted by Hausman are factors discussed in class. When conducting a stated preference model, it is not surprising that there would be a discrepancy between willingness to pay and willingness to accept for a good or service. Humans become easily sentimental over their personal belongings and naturally add some amount of value to the price one would accept for the nostalgic asset. Thus, I would guess this could either be controlled for or understood that the wtp values still reflect accurately regardless of a stated wta value. Another principle concern was hypothetical bias. Also discussed in class, by putting detailed comments within the survey that regularly remind the survey takers of the various constraints that they face in reality, rather than hypothetically,the surveyor can minimize the bias risk. Finally, as Whitehead et al points out in his abstract, CVM studies, on average, pass the scope test, providing scant evidence that Hausman’s adding up test can definitively test the validity of CVM models. If this is true, I think Hausman’s work is a poignant example of a researcher data mining for evidence that supports his own hypothesis. According to one of the cited blogs, he “selectively review[s] the contingent valuation literature.” This sort of self-reinforcement can disprove almost anything, and may be part of the reason for Hausman’s results refuting the validity of CVM.

[Wen Xiang Chuah]

I'm concerned about the psychological impacts present within CVM models and how they might impact the stated preferences of interviewed consumers. In particular, while a consumer's stated preference is related to their attitude towards the good/service, does the presence of informational asymmetry affect the optimality of their decision? Even without methodological challenges such as endowment effects, if a consumer is underinformed regarding the properties of a good/service, are their opinions equally valid as a more informed consumer (who may or may not exist).

[Callie Deddens]

I found Knetch’s statement about willingness to pay versus willingness to accept to be thought provoking. He takes conclusions about the discrepancy between the two a step further to point out that they can’t be solely attributed to a failure by the CVM because this leaves out a consideration of the human psyche and the perception of losses as being more serious than gains. In his comment, David also discusses this observation and mention sentimentality and nostalgia and I would agree that these are obviously factors that come into play when using this method to construct people’s willingness to pay and willingness to accept.

In regards to hypothetical bias mitigation, Whitehead concludes that a range of numbers is better than no number at all. This conclusion is fundamental because it serves as a reminder that it is difficult or perhaps impossible to come up with a single value for the resources being examined, but CVM at least makes it possible to have a ballpark.

[Sommer Ireland]

I find it interesting that Hausman claims that CVM is hopeless. I mean sure it has it's flaws, it isn't a perfect form of measurement, but that doesn't make it completely hopeless. I personally think it's a good way to measure WTP and preferences. Sure it's subject to hypothetical bias and there is a discrepancy between WTP and WTA, but overall it isn't that bad. I love how Whitehead's paper is a complete counterargument to Hausman's, and he has a very good point in the abstract that he isn't trying to convince the reader that contingent valuation is the best thing ever. Rather he's just trying to prove that it is not a hopeless method and that there is still research that needs to be done to improve it. Clearly there is a need for more research in CVM within the field of economics, because it really isn't as bad as Hausman makes it out to be.

[Kate LeMasters]

I agree with the other student’s sentiments that Hausman’s negative comments on CVM are extreme. I agree fully with Cort in his comment that disregarding CVM may also disregard any type of non-market valuation in certain situations. When the object of value is hypothetical, there is no way to conduct hedonic pricing or travel cost methods. Also, in developing and underdeveloped areas, it may not be feasible to conduct revealed preference methods. In some situations, it may also not be feasible to conduct choice experiments if there is primarily one valued item you are measuring and there is a lack of comparable items. Additionally, a researcher could make the situation less hypothetical. For instance, in Professor Casey’s Spring Term class, we may conduct an experiment on campus before going to Belize in which students are asked their WTP for conservation of wale sharks and given $10 for participation. They would then be asked to contribute their WTP value from their $10 after the experiment. This would make CVM highly credible. Finding other methods such as this one and realizing that CVM may be the best estimate in certain situations indicates that CVM is not hopeless but necessary. It is not a perfect measure and is subject to biases, but that does not mean we should ignore it and let the tragedy of the commons continue when other methods are not available.

[Tyler Voorhees]

Like everyone else, I agree that the negative comments on CVM may be pretty extreme. I think that when your using any sort of self-reported data to take the results with a grain of salt, but that does not mean that we should dismiss Contingent Valuation Method. I think the problem arises when economists use Contingent Valuation to make exact and specific policy recommendations for certain fees. For example, if contingent valuation comes back with an recommendation for a certain park fee, we should not necessarily take that as an hard and fast truth. If results show consumers are willing to pay more though, we could take that to mean that we should impose more fees.
Especially in the case of government using CVM, I think that line of thinking influences Hausmann’s negative comments on the government using the method versus private firms. Firms could be able to use and perfect CVM without imposing as much costs on society if earlier estimates where wrong. If the government used CVM and was wrong, there could be a deadweight lose imposed on society.

[Julia Seelye]

I agree with earlier commenters: Hausman's declaration that "contingent valuation is hopeless" is extreme and oversimplified. While CVM obviously has its disadvantages, it provides a way to obtain otherwise unmeasurable data. There are certainly issues with generalizing the results of an empirical study that uses CVM data, but that is not to say that every CVM study does not have significant internal validity. It would be problematic to use micro-level data obtained through CVM as support for public policy proposals, but such data would definitely be useful to firms and others with very specific goals. Like Tyler mentioned, hypothetically, firms that use CVM studies do so without as much risk as if the government were to use the same information. While CVM is not perfect, it does not mean that it is "hopeless" and should be further researched to improve its accuracy.

[Matt Rutley]

It seems interesting to me that Hausman deems the Contingent Valuation Method as hopeless. Yes, contingent valuation may not be perfect in determining a non-market value, but it is a necessary step in determining otherwise incalculable data. Contingent valuation is fine for finding such an arbitrary value, and the only caution we must take when using it is to make sure that we do not use the results we get too generically. Like Tyler said, we are safe letting private firms use contingent valuation, as there is less of a risk of imposing a large amount of costs on society, and they are for specific use. We just can’t let the government use contingent valuation as a means of determining public policy due to the greater societal risks doing that would have. Also, while the issues of hypothetical bias and discrepancies between WTP and WTA do exist, contingent valuation still has its place and can hopefully be improved to a point in which it is more externally valid.

[Emily Foggo]

I agree with the comments above in that I believe Hausman’s comments seem aggressively negative, and in my opinion, they may even be unwarranted. When I was looking over this presentation, I was most surprised that Hausman suggests WTA =WTP. To me, this argument is completely flawed. I wholeheartedly agree with the counter arguments that suggest the difference between the two can be explained either by the elasticity of substitution, or by the fact that people value losses more than they value gains. Again, Hausman completely neglected these explanations and draws his conclusions based on economic theory alone.
Despite my concerns with his argument regarding willingness to pay and willingness to accept, I think Hausman correctly identifies other problems with contingent valuation, such as hypothetical bias. We even addressed this issue in class when we discussed different non-market valuation techniques. With that, I agree with Carson in that there may be problems in the CV method, but we should seek to make it better as a technique, not attempt to discredit it completely.

[Katja Kleine]

I agree with the previous posts that Hausman takes an unnecessarily strong stance against Contingent Valuation Method. I particularly agree with what Callie said, it is better to have a range of numbers than no numbers at all. Surely having some sort of sense of how much a good is valued is better than having no sense at all. I do think it is good to question the quality of the practices that we use, but dooming one as "hopeless" when many top profit-maximizing firms use this practice all the time is extreme. Wen does bring up valid concerns over the psychological impacts that people face when using CVM. Ideally, we would like to have perfect information and no hypothetical bias. But, since that is not possible I definitely think that there are precautions that can be made to help combat some of that misinformation and hypothetical bias. Maybe our approach should be more towards correcting possible errors in that arena instead of disregarding a whole evaluation method.

[Eric Notari]

As stated above multiple times, I would agree that the stance against the Contingent Valuation Method is a little over the top and disregards the benefits of CVM. It is better to have some data over none, even though this data is subject to a lot of bias. I agree with Tyler in that we cannot simply use CVM to make policy recommendations, and that the results from a contingent valuation experiment should be used as a suggestion and data point, not a rule. Thinking of using the results as a be all end all for what people are willing to accept or pay and justifying a policy based on that is likely to not turn out well in the end as there are many other factors to include that CVM does not take into account. Thus, CVM is still incredibly useful, as long as it is used as a recommendation and used discretionary, by both businesses and governments.