![[logo] US EPA](https://webarchive.library.unt.edu/eot2008/20081107104058im_/http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Final Report: Stated Preference Valuation Using Real Money for Real Forested Wetlands
EPA Grant Number: R825307Title: Stated Preference Valuation Using Real Money for Real Forested Wetlands
Investigators: Swallow, Stephen K.
Institution: University of Rhode Island
EPA Project Officer: Clark, Matthew
Project Period: October 15, 1996 through October 14, 1998 (Extended to December 31, 2001)
Project Amount: $165,081
RFA: Decision Making and Valuation for Environmental Policy (1996)
Research Category: Economics and Decision Sciences
Description:
Objective:The objectives of this research project were to: (1) identify critical ecosystem attributes of forested wetlands that contribute to ecological quality and to the quality of life; (2) develop a model of public preferences for alternative attributes of wetlands in southern New England using Rhode Island as a case study; and (3) estimate money-measures of value for wetland ecosystem attributes by a survey method that calibrates hypothetical-dollar values to real-dollar values.
Summary/Accomplishments (Outputs/Outcomes):This project developed a list of wetland attributes that may be significant for evaluation of public preferences. This list ranged from focus groups to a review of the conservation biology literature. The attributes listed below do not include all attributes considered important by applied ecologists, because the objective focuses on attributes of concern in public preference analysis. These attributes included: (1) acreage of a wetland; (2) type of land uses surrounding the wetland (wooded, residential, or farm); (3) role of a wetland parcel in a conservation policy (relative to its role in expanding an existing conservation area, connecting two conservation areas, or creating a new but separate conservation area); (4) expectation of the qualitative level of wildlife diversity likely supported by the parcel (low, medium, high); (5) sustainability of the wildlife habitat; (6) location relative to roads; (7) potential for public access; and (8) monetary cost to a resident's household, which is necessary to implement a conservation agreement with the landowner.
The project was designed to implement mail surveys for stated preference assessments of the public's value toward conserving a parcel of wooded wetlands. The survey design presented respondents with choices between two parcels described by the attributes above. Respondents could state their preference to preserve parcel A, parcel B, or their preference to retain the implied costs and forego both parcels. If they forego both parcels, then they are preserving neither of them. These survey questions were implemented in two forms. In the first, all parcel choices involved hypothetical land parcels, and the respondent's cost also was hypothetical. In the second form, the parcel descriptions matched real-land parcels and respondents were told that they would need to write a check for "their cost" to demonstrate that they truly favored land conservation with regard to these costs. This approach established a data set in which responses to hypothetical questions could be analyzed in parallel with real-money questions. This analysis would identify departures in their willingness-to-pay (WTP) or environmental value estimates from the two types of questions.
The study established conservation contracts with two landowners who would permit the investigators (through the University of Rhode Island) to hold development rights for their land for a 10-year period. Wetland parcels under these conservation contracts constituted the parcels described in questions involving real-money. Respondents were told that conservation of a parcel meant the landowner would forego the right to develop his/her land for 10 years. Thus, the WTP that measures from both hypothetical and real-money questions focused on the value of placing a parcel under a 10-year contract.
Survey responses were analyzed using econometrics for discrete-response data. The statistical model assumes respondents identified the choice among parcel A, parcel B, or neither parcel, according to their preferences. The model estimates the probability that a particular parcel (or the neither choice) would be identified by the respondent as most preferred.
Several versions of the survey were designed to allow researchers to compare the effects associated with various aspects of how the survey presents real-money questions. One of the surveys presented respondents with a single-choice question involving real-money, while a second version included a hypothetical-choice question followed by an independent, real-money choice question. A third version of the survey presented two hypothetical-choice questions. These three versions of the survey allow several comparisons.
First, we have performed an analysis that compared stated-preference models estimated using data from different hypothetical questions. A general preference model was estimated, allowing each wetland attribute to carry a different effect on the likelihood that a respondent chose a particular wetland parcel for conservation. These effects could either be positive or negative, and it is denoted as a "marginal utility" of the attribute.
In this analysis, the hypothetical question that preceded the real-money question in one survey version was modeled in comparison to the hypothetical questions presented in the survey version that did not involve real-money. This comparison focuses on the respondents' reaction to the prospect of a real-money choice, but still uses hypothetical questions. The results suggest that the prospect of a real-money choice did not affect the marginal utilities (or relative marginal utilities) estimated for most wetland attributes. In particular, the marginal utility associated with most physical attributes of the wetland were not affected in a statistically significant manner, and the marginal utility of the cost attribute also was not affected. However, while not statistically different at a significant level, the marginal utility of the cost attribute from the hypothetical question in the real-money survey was about 60 percent larger in magnitude than its counterpart estimated from the hypothetical-only survey. This difference is important to note, despite the lack of statistical significance, because the respondents tended to react more negatively to a hypothetical cost question, prior to a real-money question.
However, while the marginal utility estimated for most wetland attributes was not affected by the prospect of a real-money choice question, there are 2 out of 13 exceptions. From the hypothetical question in the real-money survey, the marginal utility of conserving a wetland located in a wooded landscape, rather than in a farm or residential landscape, was positive and highly significant statistically (at the 1 percent level). Its counterpart from the hypothetical-only survey was negative, but not statistically significant at any conventional level. More surprisingly, the respondents' preference with respect to public access changed; indeed, the marginal utility reversed between the two versions of the hypothetical questions. In the hypothetical-only questions, respondents indicated a positive and strongly significant marginal utility for wetland parcels, allowing full public access. In the hypothetical question drawn from the real money survey, this marginal utility was negative, of similar magnitude, and still highly statistically significant.
These results raise the possibility that survey respondents may alter their stated preferences for some attributes of an environmental good, depending on whether the question is related to the prospect of a real-money choice. The results suggest that a preference reversal may occur with respect to some non-monetary attributes; this reversal is associated with changes in presentation prior to a real-money question. These results cast substantial doubt that only the prospect of paying actual money is relevant to departures between estimated WTP and actual WTP, as used for estimation of economic benefits from environmental goods. Even within a hypothetical question, respondents may alter their preferences by a substantial amount for a particular non-monetary, physical attribute of an environmental good.
The second comparison also was analyzed using the survey versions. The project compared the stated preference models that were estimated, based on the real-money questions in the two versions of the survey involving real money. Results indicate that the stated-preference model for the real-money survey questions was not affected by the presence of a hypothetical question (at least not in a statistically significant manner). This result would not be surprising, except that a contrary result was found for the hypothetical survey questions (as discussed above).
The final comparison involved evaluating the differences between real-money and hypothetical WTP in these survey versions. Existing results suggest a statistical difference between the hypothetical and real-money WTP, due to differences in the marginal utility of cost in these questions. It appears that a factor of two to four is observed in the marginal utility of cost between hypothetical and real-money questions. However, this conclusion cannot be drawn too broadly, because the hypothetical questions have proved to be sensitive to the prospect of a real-money choice. Further research clearly is necessary to help identify the factors causing departures in real and hypothetical WTP estimates. These factors may include shifts in preferences that may or may not be associated with standard economic principles, and factors associated with motivations for hypothetical bias as well as free-riding.
A Note on Presentation of Real-Money Questions. Two issues arose in designing the real-money survey questions. First, the project drew on methods in experimental economics where researchers have tested mechanisms that reduce the incentive of individuals to "free ride," or benefit from the financial contributions of others who help pay for an environmental good, such as wetland conservation. In preliminary research for this project, data obtained concerning WTP to monitor water quality in Rhode Island ponds showed that the real-money WTP of respondents and the difference between hypothetical WTP and real-money WTP depended upon how the provider of the environmental good (e.g., the conservation agent) utilized funds collected over and above the minimum needed to provide the good. Uncertainty played a role, but the primary factor affecting these WTP differences was whether respondents expected to receive a rebate of "their share" of any excess funds collected. These mechanisms were used in presenting the real-money questions within the wetland survey above.
The second issue concerned development of the real-money survey as a mail survey. It should be noted that substantial steps were needed to help respondents understand that the money collected would only be used for conservation of a specific wetland parcel (that matched the description given in the survey) and that excess contributions would be refunded to the respondent/contributor. These steps included explicit statements in the cover letter and in the survey questionnaire. However, despite these efforts, a substantial number of real-money survey respondents expressed concern that the survey was just another typical fundraiser. Response rates to the real-money survey were about 30 to 39 percent, while response rates to the hypothetical-only surveys were 39 to 54 percent, depending on the survey version and the Rhode Island town targeted in the survey. These different response rates suggest that the real-money survey may have caused respondents to react in ways not involved in the typical hypothetical, contingent valuation survey. Further research is suggested to understand the implications of this issue.
Journal Articles on this Report: 3 Displayed | Download in RIS Format
| Other project views: | All 14 publications | 3 publications in selected types | All 3 journal articles |
| Type | Citation | ||
|---|---|---|---|
|
|
Spencer MA, Swallow SK, Miller CJ. Valuing water auality monitoring: a contingent valuation experiment involving hypothetical and real payments. Agricultural and Resource Economics Review 1998;27(1):28-42. |
R825307 (1999) R825307 (Final) |
not available |
|
|
Spencer MA, Swallow SK, Shogren JF, List JA. Proportional rebate, random full-rebates, and winner-take-all rules in providing a threshold public good. Journal of Public Economics. |
R825307 (Final) |
not available |
|
|
Whinstanley NL, Swallow SK. Are stated preferences invariant to the prospect of real-money questions? Journal of Environmental Economics and Management. |
R825307 (Final) |
not available |
contingent valuation, nonmarket valuation, choice experiments, contingent choice/valuation, real environmental goods, hypothetical choices, real choices, calibration, RFA, ecology and ecosystems, economics, decisionmaking, social science, biodiversity option values, community involvement, compensation, conservation, cost effectiveness, decision analysis, economic incentives, ecosystem valuation, environmental assets, environmental values, forested wetlands, landscape ecology, policy analysis, preference formation, psychological attitudes, public policy, public values, social psychology, stakeholder, standards-of-value, stated preference, surveys, valuation, valuing environmental quality, Economic, Social, and Behavioral Science Research Program. , Economic, Social, & Behavioral Science Research Program, Scientific Discipline, RFA, Social Science, decision-making, Economics & Decision Making, Ecology and Ecosystems, Environmental Monitoring, Economics, stakeholder, ecosystem valuation, policy analysis, preference formation, psychological attitudes, public policy, forested wetlands, decision analysis, landscape ecology, willingness to pay, conservation, cost effectiveness, stated preference, valuing environmental quality, community involvement, biodiversity option values, public values, social psychology, surveys, compensation, environmental assets, contingent valuation, economic incentives, environmental values
Progress and Final Reports:
1997 Progress Report
1999 Progress Report
Original Abstract