ABSTRACT
Stakeholder preference for coastal wetland conservation (preservation) in Muthurajawela Marsh and Negombo Lagoon (MMNL) was revealed in Sri Lanka. The TEV of wetland conservation was estimated using CVM. The field study was used to assess individuals’ perceptions of conservation of mangroves, clean water and fish stocks, possible uses of the wetlands, respondents’ socioeconomic background, and to obtain the individual’s WTP measure for having a well-conserved wetland. The study focused on the MMNL area, as well as adjacent villages and towns. These are the areas immediately threatened by development activities and subsequently have the most to gain from conservation (Wattage and Mardle, 2005). Although those who live relatively close to the MMNL would be affected immediately by any developments, the impact of total benefits may be more widespread. However, given time and resource constraints, the study was limited to the MMNL and adjacent areas in which the most significant impact of conservation-related cultural and provisioning ecosystem service benefits might be expected. Two innovative ideas were used in designing the CVM study in MMNL.2 In this study a relatively new survey-based method known as one-and-one-half bound (OOHB) was implemented to ask questions on WTP values. Typically, in dichotomous survey formats CVM uses single-bound or double-bound approaches. For example, when using the single-bound elicitation format a WTP question is asked as to whether the respondent would like to pay a given amount, say £5, for a given option, where the answer would be ‘yes’ or ‘no’. In doublebound elicitation format, depending on the response to the first question, a second question will be followed with a different value. A criticism of the double-bound approach is that respondents are not told in advance that there will be a second value. As a result, interviews tend to focus on the first price, with the second price coming as something of a surprise when introduced at a later stage. This surprise may cause discrepancies in the responses to the two prices. The OOHB approach3 is suggested to avoid the problem. Cooper et al. (2002) suggested that eliminating the element of surprise has the potential to remove discrepancies in the responses to the two valuation questions, but that it may prevent being able to ask the second valuation question. That is, the second
question will be appropriate half of the time, on average. Due to the advantages of this approach, in this analysis the OOHB has been adopted (Wattage and Mardle, 2008). A series of questions were used in the face-to-face survey to verify a hypothetical conservation programme in the MMNL area and the WTP values. The survey was used to uncover people’s perceptions towards conservation of wetlands in the MMNL area and whether they would be willing to pay the local share of the costs. Briefly, the proposed programme considers the setting up of an institution to manage and conserve a good fish stock, clean (unpolluted) water and well-grown mangrove (i.e. ‘quality A’ type wetland) in MMNL. A question was also asked about individuals’ willingness to have a ‘quality A’ type wetland in the area. For the OOHB dichotomous choice format, the value ranges of bids used were (25, 75), (50, 100), (75, 125), (100, 150) and (125, 175). Starting bid value was selected randomly from the set of prices. Similarly, the first price from the two values in the bracket was also randomly selected. A total of 358 out of 379 questionnaires were accepted for analysis. As such, the response rate to the survey was almost 87 per cent. In order to analyse the responses to the OOHB surveys, a normal cumulative distribution function (CDF ) was applied to the OOHB data. The chi-squared test for significance of regression is 9.66514, which is significant at α = 0.05 level of significance. The income variable was found to be not significant due to the wide variation of income among survey participants. A negative coefficient implies that increases in the parameter have a negative impact on WTP. Data on respondents, such as gender, household size and education level were collected and tested as additional parameters. However, these were found to be not significant, indicating that they are not explanatory variables in the calculation of WTP. Following Cooper et al. (2002), a spike model is used to estimate the mean WTP and is calculated by integrating the cumulative density function. The estimated WTP value for the conservation of wetlands in this study amounts to SLRs.287.02 per month for two years, whereas the estimated median WTP value amounts to SLRs.264.26 per month for two years. Estimating NUV is a controversial issue in valuation studies. The second innovation of this study was the disentanglement of UV and NUV using the method of analytic hierarchy procedure (AHP). The major criticism against the NUV is that it is motivated by a form of altruism termed ‘moral satisfaction’ or ‘warm glow’ (Kahneman and Knestch, 1992). Hence, NUV is not an economic value (Carson et al., 2001). By this argument, the ability to measure NUV is contentious. However, in previous attempts to separate NUV from the total value, several approaches have been tested (Wattage and Mardle, 2008). In this approach participants were first asked to reveal the total value for wetland conservation and then used a preference disaggregation approach (AHP) to recognise the NUVs. The nature of the question format involved in the NUV suggests that personal interviews are the best option to elicit preferences. Decomposition of UVs and NUVs is dependent upon some assumptions. It is assumed that subjects valuing environmental resources have value-motives that are related to
UVs, OVs, EVs and BVs. These motives are important in a decomposition approach of a total value. It is also important that subjects know the values associated with each of these motives exclusively in order to separate UVs and NUVs. It is very difficult to evaluate how subjects might differentiate these estimated values for non-use aspects. Available approaches that are suitable for allocation do not have any solid theoretical foundation (Freeman, 1993), and as such the decomposition is very tedious. This leads to the conclusion that there is no accepted method to decompose total value into UVs and NUVs and to further decompose NUVs into motive-related components (Cummings and Harrison, 1995). AHP could be utilised for this purpose by developing priorities (or weights) for criteria and/or accompanying alternatives. This method was first introduced by Thomas Saaty in the 1970s (Saaty, 1977) and has been used considerably in applications since. The AHP is based on a process of paired comparisons across criteria (or attributes) under analysis. The steps of using the AHP process first include the development of a hierarchy of criteria. A survey is developed for pair-wise comparison of criteria to gain the preferences of individuals towards the criteria selected. Finally, the individual’s results are analysed and the aggregate sets of preferences to evaluate the overriding issue are determined. There are several important points that must be ensured during the process. The most important are that the hierarchy developed must be representative of the system and not be biased to the modeller’s needs, and that the objectives listed must be clear and convey the same meaning to all individuals (Wattage and Mardle, 2008). On completion of the survey, the analysis can be made using standard software such as Expert Choice. AHP has been used in the past to study critical situations scientifically in industry, agriculture and the environment (Mardle et al., 2004). However, it has not been used previously to study a situation such as the decomposition of UVs and NUVs. Total value can be disaggregated to NUVs using the weights allocated to the criteria relating to NUVs. It was assumed that preferences of individuals towards these NUVs indicate the relative importance they are perceived to have compared to the resources. The aim of determining ‘importance’ amongst attributes in the AHP question has clear potential for use as a decomposition method for measuring the impact of the NUVs on the total value. All survey participants who indicated their WTP for wetland conservation were asked to complete the sub-survey of AHP paired comparison to provide preference values for each attribute identified in the analysis (Wattage and Mardle, 2008). Based on the individual responses, a breakdown of the inconsistency in their responses was attained. Standard AHP practice is to accept responses where inconsistency is less than or equal to 10 per cent. From those who responded in the survey, 101 respondents gave responses to the pair-wise comparison survey with inconsistency less than or equal to 10 per cent, and 99 respondents showed inconsistency between 10 per cent and 20 per cent. Responses with high inconsistency were not included in the analysis as the reliability of their responses could not be ascertained. The aggregated preference towards UV is 0.553 and
the NUV is 0.447 in conservation of wetlands in the MMNL area. The implication of this is that a higher stated UV is associated with lower belief of moral responsibility. The total value of wetland conservation in the MMNL area as derived through the OOHB method is SLRs.264.26. So, in this case, TEV can now be split into SLRs.145.34 for UVs and the SLRs.118.92 for NUVs, using the primary results from the AHP (Wattage and Mardle, 2008). The two most important attributes are flood control today (0.254) and future generation use (0.253), where the former is part of the UV and the latter part of the NUV. The range of preference is also given for aggregated UVs and NUVs and shows a particularly insensitive result to changing consistency of responses. The allocation of weights using AHP is a robust method to split total value into UVs and NUVs. Some researchers consider that the OV should be split between UVs and NUVs. If that is the case, then the aggregated preference for the UV is 0.453 and the NUV is 0.547. This implies a stated lower UV, which is associated with higher belief of moral responsibility. So, in this case, TEV would be split into SLRs.119.71 for UVs and SLRs.144.55 for NUVs.
