For the audience who is unfamiliar with contingent valuation (CV), it is a method of estimating the value for each individual on a good. The experimental design of this kind is simple: the interviewer would directly report each willingness to pay to obtain a specific good. This method has been extremely pronounced in situations where there is no hypothetical marketplace to facilitate physical transactions to happen. In recent years, public institutions have been focused on the evaluation of raising the quality of life and public programs in public countries.
There is also an urge to impose a monetary value on environmental goods, and CV is probably the most appropriate method to measure the value of such goods. Most environmental damages cannot be estimated by the use of market prices since there is no seller for permission to damage the environment. For example, when there is an existence of externalities, such as in cases where a good is classified as a public good, it is inconvenient to impose a fair valuation on that product (Hanemann, 1994). CV is thus widely considered as the standard approach to evaluate environmental damages.
However, since its inception, contingent valuation (CV) has not been totally immune to criticism, especially with regards to implementation problems and biases. One of the central critiques on this method is its insensitivity to scope: the results from these studies should vary with the extent, quantity, and the scope of the environmental good. This article would discuss how the insensitivity to scope is such a hindrance to the development of CV, and potential solutions to make the CV method more robust.
Earlier studies on CV studies, for example, by Kahneman and Knetsch (1992) refutes CV method as being appropriate for sensitivity to scope, adding that as the responses might reflect only the aspect of “moral satisfaction”, and not the true “economic value” of the good, and disregard its importance in economics. To address this issue, the National Oceanic and Atmospheric Administration proposes that every CV study should be checked by the use of the scope test (Arrow, et al., 1993). The test should be conducted by introducing a reduced proportion of the environmental good, then recording both WTP for both the original and the reduced level. These levels can be subsequently tested by statistical means to assess whether these two values are significantly different. However, there are many studies that do not satisfy the criterion to be evaluated for responsiveness to scope. The reason to why studies are insensitive to scope are not yet clarified, however, even in recent studies.
Adequacy and Plausibility of Scope Effect
One of the requirements that is needed to ensure the reliability of the measurement in the damage assessment is listed as adequacy. Desvouges, Mathews, & Train (2012) conduct an investigation into CV studies to evaluate whether a CV study successfully passes or fails the scope test. However, they assert that there is little to infer about the verdict of each CV study since simply testing for statistical significance does not reveal whether the CV study adequately reflects the true estimation of the goods involved. For example, in Berrens et al. (2000), the average value of saving one fish is $57 while the value of saving 11 fish is only $74. In this case, it does not be necessarily the case that there should be a perfect linear relationship of the compensated value with respect to the number of fishes saved, but the results are still quite surprising.
The authors provide that the adding-up criterion can be used to evaluate the adequacy of a CV study. The adding-up criterion, however, can only be applied when the goods are classified as being incremental to each other, so it is only presented in a limited of investigated studies. Using such valuation method means that the study will be deemed inadequate when this hypothesis is rejected.
Another criterion that is used to evaluate a CV study is plausibility, as proposed by Whitehead (2016). Plausibility is understood as all the possible ranges of values of the measure Whitehead criticized the adding-up test as being the inappropriate technique to find economic meaning, but to only detect statistical significance. In this regard, the elasticity of willingness-to-pay can be used alternatively to measure economic significance, and its plausibility ranges from 0 to 1. In application, the “arc elasticity” of a point is used, since it is quite intricate to transform it into a functional form. While most studies cannot be applied to evaluate adequacy as it violates the criterion of incrementals, more studies can be checked for plausibility.
Because of such discrepancy in the methods applied and there is no general principle of what to be used, CV studies in environmental economics are often considered unreliable. Are there ways to mitigate this problem?
a) Presentation of size changes
In order to mitigate the problem of insensitivity to scope, it is the case that the design and implementation issues are the two of the most important features. One recommendation that is offered by Ojea and Loureiro (2010) in their meta-analyses is that they present the scopes (or sizes) in different ways. They notice that across different empirical studies, the experimental designs do not adhere to any standard of presentation with respect to the size changes. There are two possible different illustrations of size changes: the relative change and the absolute change.
Results show that the coefficient on size is only significant when the sizes are illustrated as an absolute change and not relative changes. So, the CV studies imply that size changes, when illustrated as the absolute changes, are more sensitive to scope. Thus it would be more sensible to phrase the fish problem as “How much more would you be willing to pay to conserve ten additional fish?” rather than “How much would you be willing to pay to save 11 fish (instead of 1 fish)?
b) Statistical assumption
When the data are collected from respondents, it is necessary Borzykowski, Baranzini, & Maradan (2018) contribute to this research by applying several statistical distribution assumptions for WTPs of respondents, which include both parametric and non-parametric distributions. They elicit responses from participants by a split-sample survey on the WTP of two forest conservation programs in Switzerland, on two different scales: the national (Swiss) forests and the regional (Geneva) forests. However, because of the selection of geographical areas, these two programs cannot be seen as incremental to each other and thus could not be checked for adequacy by the adding-up test, as illustrated by Desvouges, Mathews, & Trains (2012). The authors assert that similarly to the problem with the presentation of the differences between different scopes, there is also no scientific consensus to use a uniform the statistical distribution of the WTPs for these studies.
In terms of testing for scope effect, they find that the WTPs of respondents are susceptible to the assumption of statistical distribution. They find that non-parametric methods would be more informative of revealing the scope effect than parametric models. Therefore, the general recommendation for testing responsiveness to scope, is that to fit WTPs to a number of different statistical distributions, as long as there is not yet a consensus to be reached at this issue.
The scope effect of contingent valuation on environmental damages, albeit not a new issue, remains to be unsolved. Even with decades of empirical studies, there is still a lot of vagueness in terms of the construction of CV survey across different studies. In principle, testing for adequacy and plausibility of responsiveness to scope is considered extremely useful to evaluate the reliability In order to make the measurements more adequate and plausible to the scope effect, further studies, especially empirical studies, could explore variants of other potential developments to the design of similar experiments. If CV works well with environmental goods, then it should not be so surprising that CV to be extremely useful when it is applied in other similar valuation methods.