**2.1 The theoretical model**

The study use Contingent Valuation Method (CVM) to analyze the "willingness to pay (WTP)" of future industrial water users for using reclaimed water. The CVM method surveys user valuation over non-existing transactions of goods or services in the market by way of a direct questionnaire designed basing on hypothetic conditions in the market, therefore is a valuation method over non-market resources. The major feature of the CVM model is the forward-looking (ex ante) decision which evaluates a future event in advance. The price and quality level of the reclaimed water supply mechanism are presumed by the study without actually putting into operation; they are preliminary assumptions of the future supply mechanism of the reclaimed water which may be applicable to the existing factories of the industrial and science parks, for further understanding the WTP level of users regarding water price and the quality level.

The major difference of the CVM approach contrasting with a direct valuation approach is that CVM is specific in combining the survey practice with theories. Popular use of the CVM approach began in 1970's when the Forest Act of UK and the presidential directive #12291 were promulgated; a number of researches were seen conducted with the CVM approach over economic benefits of natural resources. During the Exxon tanker oil spoil incident in 1989, the federal court of USA ordered compensation to be paid by Exxon appraised with the CVM approach, enhancing the authenticity of the same. By 1993, since the US government extensively used CVM to make public policies that concern natural resources, the National Oceanic and Atmospheric Administration (NOAA) therefore promulgated guidelines on the use of the CVM approach, regulating the use of Contingent Valuation Surveys. Research papers show that CVM is applicable to offering a rational valuation over public goods or environmental goods (Smith, 1993). Hutchinson, et al.(1995) also pointed out that as long as the questionnaire is duly designed, CVM is a highly credible means for price evaluation.

The CVM appraisal can be conducted with a Random Utility Model (Hanemann, 1984) or an Expenditure Function (Cameron, 1988). However, Cameron(1988) uses Censored dichotomous choice model to directly estimate the parameter of the Expenditure Function, it directly and easily obtaining the WTP of the public over environmental goods. The microeconomic theory also demonstrates that the indirect utility function has a dyadic relationship with the Expenditure Function; therefore it can also represent the utility preference of the consumer. In order to prevent excessive biases and to make adequate use of all the data acquired from the questionnaire survey, the study employs a close-ended dichotomous choice method design for the questionnaire, and uses the Expenditure Function model (proposed by Cameron (1988) and Cameron & James (1987)) for calculating the WTP function of the reclaimed water.

This study use the questionnaire survey in determining the price level that the factories are willing to pay for the reclaimed water, and to provide incentives for the factories becoming willing to use the reclaimed water, a hypothetic market must be conceptually established for the factories, to create a bidding function based on individual social and economical characters and the level of bidding prices. The main method is to estimate the acceptable price by way of Cameron's expenditure function model based on the WTP of the questionnaire and the percentage of factories that are willing to pay or willing to accept. Follow the defined of resource value by Freeman (1993). We set up the empirical model of reclaimed water as follows:

$$\text{Y}(\text{Q0}, \text{Q1}, \text{U0}, \text{S}) = \text{E}(\text{Q1}, \text{U0}, \text{S}) \cdot \text{E}(\text{Q0}, \text{U0}, \text{S}) \tag{1}$$

Y(Q0,Q1,U0, S) is the bidding function of the factories for the reclaimed water;

E(Q0,U0,S) and E(Q1,U0,S) are the Expenditure Function.

In the formula,

262 Current Issues of Water Management

Taiwan, except adopting the produce/use model - in which factories who promote water saving within the industrial park reclaim their own wastewater for reuse, the environmental assessment requires that wastewater or sewerage within a building to be reclaimed by the building, or a wastewater/sewage treatment plant reclaims a portion of its own effluent for in-plant miscellaneous use - no other industrial development or application reference has been seen. Furthermore, affected by the lacking of experience and the rather low water

The Hydraulic Planning and Experimental Institute made a preliminarily research in 2009 and found that potential candidates for the use of reclaimed water include: 1. Water for secondary livelihood use: using effluent of urban wastewater/sewage treatment plant for watering nearby golf courses, to enhance flexibility of the local supply of water resources. 2. Water for agricultural use: treating effluent from the urban wastewater/sewage treatment plant to meet the standard of "water quality for irrigation" and using the reclaimed water for agricultural irrigation in areas having a water shortage. 3. Water for conservation: using water reclaimed from urban wastewater/sewage treatment plants for groundwater recharge, for artificial recharge of disaster prevention purposes, for agricultural use in substitution for the groundwater which would have originally been extracted, so as to alleviate groundwater extraction. And 4. Water for industrial use. Organizations that may increase water consumption in the future include: Hsinchu Science Park Yilan Base, Taoyuan Aviation City, Taipower Letzer Industrial Park Power Plant, Taipei Harbor Power Plant, Expansion Project of Dragon Steel Corporation, Middle Taiwan Science Park Taichung Base and Houli Base, Taichung Harbor Proprietary Areas (including power plant, petrochemical and industrial areas), Hsinchu Science Park Phase IV Tongluo Base, Yunlin Offshore Fundamental Industrial Park, Taiwan Petroleum Corporation Third Naphtha Cracker Renovation Project, Tainan County Great Hsinyin Industrial Park Development Project, Southern Taiwan Science Park Phase II, Development Project of Southern Taiwan Science Park LCD TV District (Tree Valley Park), Tinnan Industrial Park, and China Steel Corporation. The study carries out questionnaire interview with industrial water users to comprehend their willingness

towards paying for the reclaimed water as well as their methods to use the same.

The study use Contingent Valuation Method (CVM) to analyze the "willingness to pay (WTP)" of future industrial water users for using reclaimed water. The CVM method surveys user valuation over non-existing transactions of goods or services in the market by way of a direct questionnaire designed basing on hypothetic conditions in the market, therefore is a valuation method over non-market resources. The major feature of the CVM model is the forward-looking (ex ante) decision which evaluates a future event in advance. The price and quality level of the reclaimed water supply mechanism are presumed by the study without actually putting into operation; they are preliminary assumptions of the future supply mechanism of the reclaimed water which may be applicable to the existing factories of the industrial and science parks, for further understanding the WTP level of

The major difference of the CVM approach contrasting with a direct valuation approach is that CVM is specific in combining the survey practice with theories. Popular use of the CVM

**2. Method** 

**2.1 The theoretical model** 

users regarding water price and the quality level.

price, the willingness to use reclaimed wastewater has been fairly low.

Q0 is the situation that the factory do not get reclaimed water;

Q1 is the situation that the factory got reclaimed water;

U0 is the utility function of the factory;

S is the price vector of market goods and individual social and economic characteristics vectors.

If the price suggested by the CVM questionnaire is T,

$$\mathbf{Y}(\mathbf{Q}0, \mathbf{Q}1, \mathbf{U}0, \mathbf{S}) \ge \mathbf{T} \tag{2}$$

The Willingness to Pay of Industrial Water Users for Reclaimed Water in Taiwan 265

Where Yi\* is the price of reclaimed water estimated by the supplier under the standard binary probit model; this can be used for the calculation of a reasonable price for the

Assuming *u* to be the logistic distribution, the empirical result can be calculated based on

P(Y)=[1+e-[Yi-Ti]]-1

Where Yi\* is the price of reclaimed water estimated by the supplier under the logistic model;

To the demand end, quality and price of the reclaimed water are the major concern. We detail as follows: Water reclaimed from effluent of large scale wastewater treatment plant by reverse osmosis: capable of reaching quality standard of Taiwan Water Works. For the selection of questionnaire valuation method, the study employs the most easy-to-operate and time saving "Single-bounded dichotomous choice elicitation method" (Boyle & Bishop,

The scenario of this study is as follow: We assumption that "the government guarantees that quality of reclaimed water conforms with city water specifications, no interruption of supply 365 days a year with assured quality and loss indemnification on supply interruption," and that "dedicated pipeline to be installed for reclaimed water delivery, plus with free-of-charge pipe connection," and that "50% deduction on wastewater treatment charge if total consumption of reclaimed water exceeds 40% of total industrial water consumption of the company". Than we ask the manager or boss of the factory " Are you willing to pay for the reclaimed water for the "T" price we suggested on the questionnaire1

The value of reclaimed water depends on its water quality. The quality of "city water" is just the basic requirement of the customer when comparing with more expensive and better quality of "soft water, 1μS/cm". Besides, the assumptions of the following are not yet done but they are the requests of the factories. So we set up the approximate realistic assumption

The selection of the price we suggested on the questionnaire, i.e. "T" in formula(2) in each questionnaire of the scenario is determined based on the current city water price in Taiwan and the costs for reclaiming the wastewater. Furthermore, one or several extreme and median values are set to meet theoretical requirements of the Contingent Valuation Method, the scenario having 12 kinds of "T" prices as shown in Table 1. In another word, the study employs 12 different questionnaires, QA through QL, with different assignment of the "T"

1 we give different "T" price in different type of questionnaires which shows on Table1.

this can be used for the calculation of a reasonable price for the reclaimed water.

Yi\* = Xi′B (8)

reclaimed water.

the logistic model by Cameron(1988).

**2.2 Questionnaire design** 

under the assumption scenario?"

prices for each type of questionnaires scenario.

of the scenario.

Similar to the probit model appraoch, we can obtain

1988) to carry out interviews based on NOAA suggestions.

the probability for the interviewee to check this bid can be expressed by formula (3):

$$\text{Pr=Pr[Y^\*(Q0, Q1, U0, S)-T \ge u]}\tag{3}$$

Where Y\* is observable component, u is observable random component, as shown in Formula (4):

$$\mathbf{Y}(\mathbf{Q}0, \mathbf{Q}1, \mathbf{U}0, \mathbf{S}) = \mathbf{Y}^\*(\mathbf{Q}0, \mathbf{Q}1, \mathbf{U}0, \mathbf{S}) + \mathbf{u} \tag{4}$$

The Bidding Function can be estimated based on the probit model by Cameron & James(1987) as shown below:

Ii=1 if Yi >Ti = 0 otherwise

$$\begin{aligned} \Pr(\text{li} = 1) &= \Pr(\text{Yi} \rhd \text{Ti}) = \Pr(\text{ui} \rhd \text{TI} \text{-Xi} \text{'B}) \\ &= \Pr(\text{ui} \mid \sigma > \text{(TI-Xi'B)} \text{/} \sigma) \\ &= \text{1-} \phi(\text{(Ti-Xi'B/} \sigma) \end{aligned} \tag{5}$$

where Xi′B is exclaiming variable, φ is accumulated probability of intensity function, then the interviewee's bidding valuation can be shown as formula (6) :

$$\mathbf{Y}\mathbf{i} = \mathbf{X}\mathbf{i}'\mathbf{B} + \mathbf{u}\mathbf{i} \tag{6}$$

Yet standard binary probit model shall be

Ii=1 if Yi>0 =0 otherwise

$$\begin{aligned} \Pr(\text{li} \text{=1}) &= \Pr(\text{Yi} \rhd \text{0i}) = \Pr(\text{ui} \rhd \text{-wi} \text{'\'\'\'} \text{\'})\\ &= \Pr(\text{zi} \rhd \text{-wi} \text{'\'\'\'} \text{\'} \text{\'} \text{\'} \text{\'})\\ &= 1 \cdot \phi(\text{-wi} \text{'\'\'} \text{\'} \text{\'} \text{\'} \text{\'}) \end{aligned}$$

at this time,

$$\text{Yi = wri'}\delta + \text{ui }$$

using the following transformation

$$\begin{aligned} \text{-(Ti, Xi')} \begin{bmatrix} -1/\,\sigma \\ B/\,\sigma \end{bmatrix} &= \text{-wi'}\delta \\\\ \delta^\* &= (\alpha, \gamma) = \text{(-1/\,\sigma, B/\,\sigma)} \end{aligned}$$

we obtain

$$\begin{aligned} \mathbf{B} &= -\gamma/\alpha\\ \sigma &= -1/\alpha\\ \mathbf{Y} \mathbf{i}^\* &= \mathbf{X} \mathbf{i}^\* \mathbf{B} \end{aligned} \tag{7}$$

Where Yi\* is the price of reclaimed water estimated by the supplier under the standard binary probit model; this can be used for the calculation of a reasonable price for the reclaimed water.

Assuming *u* to be the logistic distribution, the empirical result can be calculated based on the logistic model by Cameron(1988).

$$\mathbf{P(Y) = [1 + e - [Yi - Ti]] - 1}$$

Similar to the probit model appraoch, we can obtain

$$\mathbf{Y}\mathbf{i}^\* = \mathbf{X}\mathbf{i}'\mathbf{B} \tag{8}$$

Where Yi\* is the price of reclaimed water estimated by the supplier under the logistic model; this can be used for the calculation of a reasonable price for the reclaimed water.
