The Contingent Valuation of Welsh Farm Environment (Tir Gofal Policy)
Contingent valuation is an economic technique for the valuation of public goods currently used around the world both by governments agencies and the World Bank (Navrud 1992, Hanemann 1994). The CV method uses survey questions to investigate people’s preferences for public goods by finding out what they would be willing to pay (WTP) for specified improvements in them. The CV creates hypothetical markets in which the survey respondents could buy the good in question (Venkatachalam 2004, Brookshire et al 1983, Brookshire et al l982, Schulze et al 1981).
The present survey has been designed to value a Welsh government policy (Tir Gofal) that promotes the preservation and creation of the Welsh farm environment. The survey was piloted on a small sample size of 25 persons. The report below summarises the design, administration and findings of the survey, and subsequently performs a critical analysis of the study.
- Pilot Study Report
- Design
The first step was to decide on which characteristics to specify in the design for each Tir Gofal question. This was done by first considering which habitats were affected by the management provisions of Tir Gofal. The list comprised Woodland, Grassland, Wetland, Healthland, Coastal Habitats, as well as archaeological, access and educational access. These characteristics were then tested in the questionnaire.
Survey Mode
The survey mode that has been chosen is in-person interview. This method is consistent with the recommendations of the NOAA panel, as well as those of most CV writers (i.e. Whitehead 1999, Hanemann 1994).
Hypothetical Scenario
As per the recommendations of Hanemann (1994), NOAA (1993), Mitchell and Carson (1989), the WTP was determined by using a hypothetical scenario. The scenario was created to convey a sense of reality to the surveyed. The scenario designed a market where the intangible product in question – the preservation of environment in Wales through the Tir Gofal programme – could be traded.
In order to test the accuracy of the hypothetical scenario, two methods were conceived. In the first method, a direct question was posed in the shape: “Would you be willing to pay an x yearly tax to preserve the Welsh natural landscape and wildlife?” In the second method (cross-tabulation), the measures of preserving the Welsh environment were analysed in detail and the persons surveyed were asked to put a price on each. The two methods would then be compared to determine accuracy and relevancy of the primary WTP question.
Elicitation Technique
As Venkatachalam (2004) has summarized, there are currently four major types of elicitation techniques available in the literature, namely, the bidding game, payment card (PC), dichotomous choice and open-ended (OE). The first three are closed-ended techniques while the latter one is open-ended.
As Hanemann (1994) points out, the open ended questions have been popular until the mid-1980s, when the closed ended question gained preference. Some of the reasons included the ambiguity of open-ended questions, the inability of the respondents to appreciate the scale of the payment, non-response risks, and the tendency to undervalue the good (Whitehead 1999, Hanemann 1994, Desvousges et al.1993). Therefore, this survey has been designed to use close-ended questions except for the first and the last one. This created a ‘referendum’ environment as supported by the NOAA (1993) guidelines.
The closed ended techniques can be bidding game, payment card or dichotomous choice. The bidding game begins with an opening bid and continuously raises and then adjusts the bar until a positive match is found. The payment card contains several WTP amounts out of which the surveyed person must choose one. Finally, the dichotomous choice is a method developed and refined by Hanemann (1984, 1985) and Carson (1985) whereby a maximum willingness to pay is proposed to the surveyed person as a yes/no question, after which, depending on the answer, a follow up WTP is proposed.
Due to the limited scope and sample size of the survey, the author has chosen the bidding game approach. The dichotomous choice has been deemed too complicated for the purpose of the survey. As Venkatachalam (2004) suggested, the dichotomous choice requires a large sample size and sophisticated econometric techniques, parameters that were not available in this exercise. The payment card has also been ruled out as it is affected by range bias and centering bias (Mitchell and Carson, 1989). The bidding game was considered simple enough and effective for this experiment. As per Cummings et al (1986), the bidding technique creates a ‘market-like’ situation to the respondents in which they could research their preferences and because the WTP is maximised. At the same time, the author has acknowledged the fact that this system is prone to starting point bias (Venkatachalam 2004).
Payment vehicle
The payment vehicle that has been chosen is the introduction of a Welsh government tax. It has been recognised that a clear vehicle is necessary in order to avoid abstract questions like “how much would you pay for the preservation of Welsh landscape?” Such questions would be meaningless as they would not reflect any reality in the surveyed person’s mind (Hanemann 1994). The survey author has considered that a tax is the most viable instrument of raising money to preserve the Welsh landscape and wildlife. In order for the tax to be realistic, it was considered that this amount would be raised by the Welsh government from the residents of the area. In line with Whitehead (1999)’s suggestions, the tax could be enforceable on local residents and therefore is realistic.
The Sample
As per the assignment, the questions were piloted on a group of 25 persons, most of which were based in Wales. The Wales residency was considered preferable in order to increase the realism of the WTP choice, as well as the possible knowledge of the Welsh farmland environment.
Choosing Questions
As per Whitehead (1999)’s recommendations, the questions have been kept as short as possible, as clear as possible. In addition, the scenario was kept as short as possible considering in order to minimise time inconvenience for those agreeing to be surveyed.
The author considered that two aspects of Tir Gofal should be investigated: the ‘preservation’ and the ‘creation’ aspect. Tir Gofal plans to compensate both cases. The survey tried to convey the sense that there was a financial difference between the two. It has also presented the two aspects in sequence, as the preservation of the land would be considered more important and stringent than the creation of new land features.
To enhance the realism and accuracy of the questions, the survey attempted to briefly explain each component of Tir Gofal to the respondent. This part was designed carefully as wording could have an impact on the surveyed person and create bias.
No Answer Questions
As per NOAA (1993) guidelines, the surveyed persons have been offered some possibilities of not answering a particular question. This method has been used for the first question.
Bias-Reducing Methods
To reduce bias in the survey, the author has taken the following steps:
- The creation of a sufficiently realistic scenario. As Mitchell and Carson (1989) point out, the biggest challenge of a CV researcher is to develop a realistic, meaningful and understandable scenario. The existence of the Tir Gofal programme has made this task significantly easier to undertake.
- The introduction of a believable payment vehicle, i.e. tax on the Wales residents.
- The piloting of the questionnaire on a majority of Welsh residents. It was considered that this choice would create a more realistic WTP answer, as the respondents would feel that the tax could actually affect them. Moreover, the Welsh residency was deemed to correlate positively with knowledge of the actual farmland environment. Prior knowledge would increase the accuracy and relevancy of the answers.
- The avoidance of open-ended questions. Most research deems this form of questioning as unsatisfactory and biased, unless the surveyed persons are highly familiar with the subject matter (which is not the case in the Tir Gofal example).
- The use of cross-tabulation as per NOAA (1993) guidelines in order to insure accuracy of the original WTP question.
- The reduction of biased or influencing information. For instance, the author has chosen not to use photos as these may distort the surveyed perception.
- Administration
The survey was piloted on 25 persons of different British county origins.Demographic data on the survey participants has been collected (Whitehead 1993). It was considered that the demographic aspect might provide correlation between answers and such variables as gender, income levels and age.
Gender
There were 13 males and 12 females interviewed, giving sufficient balance to possible gender differences.
Income
The income data was also deemed important to collect. Surprisingly the data revealed a mild negative correlation between income levels and the levels of tax: many low-income surveyed affirmed they would pay higher amounts of tax than high-income respondents. This situation deserves further analysis, as it either shows that (a) lower income respondents have a higher environmental citizenship or (b) lower income respondents may overstate their WTP i.e. warm glow or public spiritedness.
Age
The survey attempted to interview a roughly equal number of people on the 1-35 range and 35-80 age range. However, due to the constraints of the survey and the fact that under 20s do not usually have disposable income, most of the 13 under 35 persons interviewed were in the 20-30 age range.
The age offered a vigorous positive correlation to the tax amount chosen. The under 35 respondents chose smaller taxes. Generally, the over 35 chose higher levels of tax.
Current Residence
With a few exceptions, all the respondents are based in Wales (Aberystwyth), therefore they would be directly affected by the implementation of a Welsh government tax. This factor has deemed to foster more realistic answers amongst the persons surveyed.
Knowledge of Farm Environment
It was interesting to note that an overwhelming majority of the respondents declared that they are awareness of the Welsh farm environment is ‘medium’ or ‘high’. Secondly, all respondents with one exception confirmed that they have visited a Welsh farm before. Consequently, the knowledge of the subject has been deemed ‘medium-high’.
- Analysis of Survey Results
The responses of the survey have used the Limdep software to display the sample mean and 95% confidence interval for the true mean income that Welsh citizens are willing to pay for the improvement and maintenance of Welsh farms.
A large majority of the sample were in favour of paying farmers to look after the countryside through income taxation.
With regard to the second question, on average the respondents are willing to pay an £136.04 to preserve natural landscape and wildlife. The confidence interval has a lower bound of £88.98 and an upper bound of £183.10.
For the third question (woodland), on average the respondents are willing to pay 18.54 to allow farmers to protect existing woodlands. The confidence interval was between £11.88 and £25.20.
With regard to the first part of the fourth question (wetland), on average the respondents are willing to pay £25.21 to reward farmers for taking care of existing wetlands. The confidence interval ranged between £16.28 and £34.13. By contrast, on average the respondents are willing to pay an additional £12.08 to reward farmers to create new wetlands. The confidence interval has a lower bound of £7.39 and an upper bound of £16.78.
With regard to the first part of the fifth question (grassland), on average the respondents are willing to pay £18.33. The confidence interval has a lower bound of £10.20 and an upper bound of £26.47. By contrast, on average the respondents are willing to pay £3.75 to increase semi-improved grassland on Welsh farms. The confidence interval has a lower bound of £0.81 and an upper bound of £6.69.
With regard to the first part of the sixth question (heathland), on average the respondents are willing to pay £16.04. The confidence interval has a lower bound of £9.08 and an upper bound of £23.00. By contrast, on average the respondents are willing to pay £6.67 to create lost heathland. The confidence interval has a lower bound of £2.85 and an upper bound of £10.49.
With regard to the first part of the seventh question (coastal habitats), on average the respondents are willing to pay £20.83. The confidence interval has a lower bound of £14.20 and an upper bound of £27.47. By contrast, on average the respondents are willing to pay £11.88 to restore or create new coastal habitats. The confidence interval has a lower bound of £7.35 and an upper bound of £16.40.
With regard to the eighth question, on average the respondents are willing to pay £42.92 to create new or improve current access routes. The confidence interval has a lower bound of £31.80 and an upper bound of £54.04.
With regard to the ninth question, on average the respondents are willing to pay £12.71 to create educational programmes on farms. The confidence interval has a lower bound of £6.36 and an upper bound of £19.06.
With regard to the final question (historical preservation), on average the respondents are willing to pay £4.79 to preserve historical remains on Welsh farms. The confidence interval has a lower bound of £0.46 and an upper bound of £9.13.
- Critical discussion of pilot study
The initial stage in the CVM survey was to conduct a pilot ‘attitudes’ survey amongst 25 respondents from the general public. The sample has revealed strong willingness of the Welsh residents to support the preservation and reconstruction of the farm landscape and wildlife. Consequently, the author deems that the following aspects have been successfully established: (a) a general support toward environmental preservation in Wales and Tir Gofal measures in particular, (b) a willingness to pay for the achievement of these objectives. Significant non-market use values were found, in that people who had not lived in the specific farmland area were still willing to pay to achieve the environmental gains of the programme. Briefly, this survey performed well on a number of validity tests, and yielded mean WTP amount of £136 on the general questions. Another successful aspect has been the closeness of WTP amounts between the direct question (How much are you willing to pay?) and the secondary, cross-tabulated questions. Thus, the combined WTP of the cross-tabulated environmental aspects of Tir Gofal yield £133.33. If the access, education and historical aspects are included, the WTP becomes £193.75.
Next, the study has been compared with the National Oceanographic and Atmospheric Administration (NOAA) guidelines on the design of CVM studies in environmental damage suits (Arrow 1993). In general, the survey design follows NOAA guidelines.
Thus, respondents were repeatedly reminded that they were being asked their WTP for the environmental improvements, and that extra spending would be necessary for all other environmental policies. WTP elicitation rather than compensation format were used, and the questions were designed in the form of a referendum, in accordance with NOAA guidelines.
The NOAA Panel believed it unlikely that reliable estimates of values could be elicited with mail surveys. Face-to-face interviews were recommended, and this is how this survey was conducted. Similarly non-responses were minimised through the use of closed-ended questions (only one non-responsive subject was encountered). Pre-testing has been carried out to some extent, but the pre-testing pool was small and biased, being confined to family and close friends. To the best extent, the questionnaire was designed with a conservative view in mind: it has taken a baseline amount of £75, and then refined the general WTP question by inquiring about each component specifically.
The survey has also attempted to provide a rather accurate image of the programme and policy proposed. However, due to time constraints, the scenario description was restricted. Even so, the author found that the survey was very time consuming, as questions had to be asked several times on occasions before people understood what was being asked. A no-answer question was introduced in the start of the survey to allow for people to abstain. The survey also used pervasive cross-tabulations, by adding numerous questions destined to increase the reliability of the primary WTP question, as well as provide information such as income, knowledge of the site, visitation rates and willingness to perform the task.
Constant reminders of the ‘realism’ of the scenario were used in order to reduce the ‘warm glow’ or public spiritedness of giving rather than actual willingness to pay for the program in question. However, the mild negative correlation between income and tax levels raise questions as to whether the warm glow has been avoided particularly for lower-income respondents.
Other NOAA guidelines such as Reminder of Undamaged Substitute Commodities Present Value Calculations of Interim Losses and Adequate Time Lapse from the Accident were not used, as it was considered that these refer mostly to environmental accidents rather than preservation techniques. Another guideline that has been less used was the debriefing section at the end of the survey. One check on Understanding and Acceptance has been used to inquire on whether the survey questions have been understood and whether perception was changed. It was interesting to note that the survey generally did not change the understanding or perception of the matter for the respondents.
Generally, the author considers that the strengths of the current study rest in its use of close-ended questions, the plentiful cross-tabulations on the initial WTP question, the sampling of different income levels and ages, the use of a WTP payment vehicle that has proven acceptable by the respondents and the efforts made in increasing the realism and reliability of the survey (e.g. by constant reminders, WTP vs compensation techniques and scenario creation). The result between the original WTP question and the cross-tabulation is remarkably similar, which further validates the study.
The main weaknesses of the survey rested in the very small sample size (25 persons) and the limited sampling techniques used. The pre-testing has been very limited in comparison with recommendations. The author found it difficult to provide adequate information to respondents about the policy or programme under the pressure of time. Results have shown that, particularly when the income levels and the tax amounts were compared, ‘warm glow’ and public spiritedness has not been completely eliminated. The use of the bidding game technique suffers from starting point bias, a bias that could only be reduced by careful cross-tabulations, but not eliminated.
To enhance the accuracy of the survey, the author believes that the sample size and diversity should be greatly increased. A careful statistical analysis of the Wales resident population should be undertaken prior to sampling in order to maximise the realism of the survey. Careful and relevant pre-testing should be used on a larger scale, and results recorded. Moreover, on a greater sample size, a double-bounded dichotomous question technique should be used. Several authors recommend this technique as being more complex and accurate. More techniques to eliminate persistent warm glow and public spiritedness should be used, including repeated reminders, comparisons and follow-up questions.
- Conclusions
The CV technique is still a subject of great controversy. Its detractors argue that respondents give answers that are inconsistent with the tenets of rational choice, that these respondents do not understand what it is they are being asked to value (and, thus, that stated values reflect more than that which they are being asked to value), that respondents fail to take CV questions seriously because the results of the surveys are not binding. Nevertheless, as this survey has hoped to show as well, there is a certain level of validity to the WTP technique. The survey has evidenced a general concern for the environment, willingness to pay a tax for its preservation and has successfully produced a WTP amount that was validated by cross-tabulations.
- References
Arrow, K, et al. (1993). Report of the NOAA Panel on Contingent Valuation. Washington, D.C.
Brookshire, D.S., Eubanks, D.S., Randall, A. (1983). Estimating Option Price And Existence Values For Wildlife Resources. Land Economics, 59, Pp. 1–15.
Brookshire, D. S., Thayer, M.A., Schulze, W.D. and d’Arge, R.C. (1982). Valuing Public Goods: A Comparison Of Survey And Hedonic Approaches. American Economic Review, 72, pp. 165-77.
Desvousges WH, Johnson FR, Dunford RW, Boyle KJ, Hudson SP, Wilson N. (1993). Measuring Natural Resource Damages With Contingent Valuation: Tests Of Validity And Reliability. In: Hausman, J.A, Ed. Contingent Valuation: A Critical Assessment. Amsterdam: North Holland, pp. 91– 159.
Hanemann, MW.(1994). Valuing The Environment Through Contingent Valuation. Journal Of Economic Perspectives, 8, pp.19– 43.
Hanemann MW, Loomis JB, Kanninen BJ. (1991). Statistical Efficiency Of Double-Bounded Dichotomous Choice Contingent Valuation. American Journal Of Agricultural Economics, 73, pp. 1255–63.
Mitchell, R.C. & Carson, R.T. (1989). Using Surveys To Value Public Goods: The Contingent Valuation Method. Washington, DC: Resource For The Future.
National Oceanic And Atmospheric Administration (NOAA). (1993). Report Of The NOAA Panel On Contingent Valuation. Federal Register 1993, 58, Pp. 4602–14.
Navrud, S. (1992). Pricing The European Environment. New York: Oxford University Press.
Schulze WD, d’Arge RC, Brookshire DS. (1981). Valuing Environmental Commodities: Some Recent Experiments. Land Economics, 57, Pp. 151– 69.
Venkatachalam, L. (2004). The Contingent Valuation Method: A Review. Environmental Impact Assessment Review, 24, Pp. 89–124.
Whitehead, J. (1999). A Practitioner’s Primer On Contingent Valuation. [Online]. Available at: www.deu.edu.tr/userweb/sedef.akgungor/dosyalar/ecu0008.pdf [Accessed: 3 April 2009].
- Attachments
Attachment 1: Tir Gofal Survey Questionnaire
Attachment 2: Survey Results
Attachment 3: In-detail Survey Results, Limdep Analysis, Demographic Data