How Widely Applicable is River Basin Management? An Analysis of Wastewater Management in an Arid Transboundary Case

Abstract

The basin scale has been promoted universally as the optimal management unit that allows for the internalization of all external effects caused by multiple water uses. However, the basin scale has been put forward largely on the basis of experience in temperate zones. Hence whether the basin scale is the best scale for management in other settings remains questionable. To address these questions this paper analyzes the economic viability and the political feasibility of alternative management options in the Kidron/Wadi Nar region. The Kidron/Wadi Nar is a small basin in which wastewater from eastern Jerusalem flows through the desert to the Dead Sea. Various options for managing these wastewater flows were analyzed ex ante on the basis of both a cost benefit and a multi-criteria analysis. The paper finds that due to economies of scale, a pure basin approach is not desirable from a physical and economic perspective. Furthermore, in terms of political feasibility, it seems that the option which prompts the fewest objections from influential stakeholder groups in the two entities under the current asymmetrical political setting is not a basin solution either, but a two plant solution based on an outsourcing arrangement. These findings imply that the river basin management approach can not be considered the best management approach for the arid transboundary case at hand, and hence is not unequivocally universally applicable.

Appendices

Appendix 1

Cost Estimates for Wastewater Treatment Options

Capital costs were calculated on the basis of cost assumptions for an original plant and an adaptation of the costs depending on the size of the WWTP. The transformation equation for the adaptation is given by Eq. 1 (ARIJ 2004):

$$ \hbox{Capital cost} \, j = \hbox{capital\,cost} \, i*\left( {\hbox{capacity}\,j/\hbox{capacity} \, i} \right)^{0.79} $$

(1)

where Capital cost j = the capital cost of a plant with secondary wastewater treatment j treating a different wastewater volume (capacity j) than the original treatment plant i. Capital cost i = the cost of the original wastewater treatment plant i with capacity i.

The initial capital costs were assumed to be 0.10 US$/CM for a capacity of 8.9 MCM/a. This assumption can be considered to be in line with real world evidence for the capital costs of ten secondary wastewater treatment plants in Jordan, which range between 0.03 US$/CM and 0.57 US$/CM (Abu-Madi 2004).

The following considerations were given to the calculation of O&M costs, taking real world experience provided by Abu-Madi (2004) into account. O&M (including conveyance) costs for the Wadi Nar site (1), which is based by and large on an existing conveyor and on existing gravity flows, were assumed to be 0.10 US$/CM. The same assumption was made for the Jerusalem site (3). For the Nebi Musa site (2) O&M costs were assumed to be higher and were estimated to be 0.25 US$/CM, as this treatment option requires redirecting the current wastewater flows and pumping them into the Og-Muqalek basin (apart from the gravity-based option M3 Gravity) (Table 7).

Table 7 Costs of different alternatives

Appendix 2

Calculation of Non-Market Benefits (Total Use Values) for Wastewater Treatment Options

Background

The non-market benefits were estimated on the basis of a Contingent Valuation Method (CVM) survey. The CVM is one of a handful of stated preference techniques used by economists to place a monetary value on non-market goods such as river restoration. An advantage of stated preference techniques such as the CVM is that they can measure both the use values and non-use values of a good.

Although some economists question whether CVM provides a valid estimation of non-market benefits, the method has been accepted by the judicial system for estimation of the damages caused by oil spills and other environmental injuries (Carson and others 2003). In 1993 a panel consisting of several Nobel Laureate economists published guidelines for acceptable practices in the use of the CVM (Arrow and others 1992). CVM studies rely on responses to public opinion surveys to estimate a monetary value for a non-market good. The surveys ask individuals to state their willingness to pay (WTP) for a specified level of change in the provision of a good. These questions are typically preceded by a detailed description of the good and the proposed changes to it, as well as socioeconomic questions and questions about respondents’ use of the good and their perceptions of it (Mitchell and Carson 1989). An individual’s WTP is equivalent to the well-being they receive from knowing that this specific change will be provided to them. Two key assumptions of welfare economics are inherent in measures of WTP. The first assumption is that individuals seek to maximize their utility and that they have clearly defined preferences that can be revealed through the choices they make. The second key assumption, known as consumer sovereignty, is that individuals are the best judges of their own utility.

Survey Design

The survey design consisted in collecting background information, conducting focus groups, developing questionnaires, and pretesting. The first phase of the survey design involved meetings with stakeholders. Two focus groups were then held in different locations to assess the public’s level of knowledge regarding water quality and management of water pollution. The focus groups were also used to test specific survey materials, such as passages of text, photographs, and maps.

The survey design process continued with the drafting of the survey materials. It also included color photos and maps of the river basin. These materials showed the current level of water quality in the basin and a potential future scenario of what water quality in the basin could look like if water quality problems were actively managed. Peers and selected stakeholders reviewed drafts of the survey materials and provided important feedback.

The final phase of survey design involved extensive pretesting of the survey instrument. First, several cognitive interviews were conducted with respondents in a face-to-face setting. In addition to conducting the survey itself, these interviews included extensive debriefing sessions to uncover any potential problems. In all pretesting interviews a behavioral coding technique was used to identify any problematic survey questions (Presser and Blair 1994).

The Contingent Valuation Survey

The first section was devoted to introducing the problem to the respondent. In this section two alternative future scenarios were presented, the Kidron/Wadi Nar dry and the Kidron/Wadi Nar wet options, after which the respondents were asked to rank the dry and the wet river options in terms of their first and second choice respectively.

The second section dealt with the willingness to pay. A payment card was inserted into the survey and the respondent was asked to circle his or her most preferred choice. 32 numbers appeared on the card ranging from 0 to 150 NIS or more. This range was selected on the basis of information provided by the focus group participants who completed an open-ended survey during the first stage. The second section also included questions regarding the respondents’ motives for payment. These motives reflected use values, bequest values, option values and existence values. The survey also included a question designed to elicit respondents’ motives for a zero payment. This is especially important since zero payment may be declared either because the respondent does not place any value on the resource or because they are doing so as a protest bid. The latter motive is not legitimate in CVM and should be dropped, otherwise the results would be biased downward. Owing to their sensitive nature, demographic questions were asked at the end of the interview. The questionnaire can be obtained from the authors upon request.

206 Israeli and 88 Palestinian surveys were completed in different locations. The sample size represents the overall population in both entities. Caution was exercised when validating that the sample mean characteristics were similar to the statistical properties of the overall populations in each case. All sample characteristics and regression coefficients can be obtained from the authors upon request.

When a respondent circles a certain bid on the payment card, their true WTP may be anywhere between that number and the next number up. In order to take account of this, one procedure assumes that the midpoint between the lowest and highest point in the selected range is representative (Cameron and Huppert 1989). This approach has the advantage of enabling a computation of WTP directly from the data, without needing to specify the functional form of the utility function or the appropriate regression model (i.e. probit or logit). We report this method only because running an ordered probit regression did not improve the results.

The results of the WTP study are given in Table 8. We first calculated the mean WTP for the different subsamples and preferences (dry or wet). In order to ascertain the total value we weighted the WTP by the share of respondents who preferred that alternative. However, the rest of the respondents were also assumed to derive a benefit from choosing their less preferred option (that is, a person would not derive zero benefit if a dry river bed were chosen over a wet bed, their 1st choice). We therefore added 50% of the second choice WTP, weighted by the respective share of respondents. The total value was calculated by adding the values for the 1st choice and the 50% values of the 2nd choice. In a next step, we used the answers in the questionnaires to identify the share of respondents who put a use value on the various options. The use value of the various options was derived on this basis in order to gain a more conservative estimate. This number was multiplied by the total number of the relevant households in Israel and the Palestinian West Bank respectively in order to yield the total use value (first column from the right).

Table 8 CVM results summary per household and for entire basin