2.1. Area Description and Drinking Water Sources
The Pavlodar area is one of 14 regions in Kazakhstan. It is situated in the northeast of the country in the Irtysh River Basin. The area has three cities and 412 rural districts. The rural districts have a population of about 270,000 people. The steppe comprises the greatest portion of the region. The climate is highly continental, relatively dry with large temperature amplitude (about −40 °C to +40 °C), a long and cold winter (5.5 months) and a short and hot summer (3 months). This very large intra-annual temperature variations limit the regional water resources that mainly depend on the snowmelt.
The scant precipitation has an uneven distribution within the territory and within the seasons. Up to 80% of annual precipitation fall during the summer period. Most of the rainfall ends up as soil moisture and evapotranspiration. Average annual Class A pan evaporation is about 800 mm. The available water resources of the region are mainly the Irtysh River and groundwater. Smaller rivers usually have a short spring discharge before they dry up.
Main water supply systems can be classified as centralized and decentralized. The main difference is that centralized water supply has a distribution system to provide water from the natural water source with or without treatment to the water user, while decentralized water supply system uses water directly from the water source with or without treatment. Consequently, centralized water supply generally means water provided through pipes to households (tap water) and public standpipes. Protected boreholes, wells, and springs are considered to be decentralized water supply sources. Centralized and decentralized systems can be either private or public. Both groundwater and surface water may be used for the centralized systems. However, groundwater is the most common type for rural centralized systems in the region.
2.2. Survey Design
The survey was performed between October 2011 and January 2012, in eleven villages of Pavlodar region. In total, 168 questionnaires were completed and included in the survey analysis. Since the villages are of different sizes, each household was chosen randomly so that at least half of the respondents would live in four different directions outside of the central part of the village.
Face-to-face interviews with the heads of households were conducted in the respondents’ homes. Standardized questions regarding socio-economic and demographic characteristics, existing drinking water sources and their characteristics, trust to water management types, and direct open-ended and bids questions on WTP were used.
The study adopted both stated preference and revealed preference methods to value the existing and hypothetical water supply service. For stated preference approach, the CV method was used where respondents were directly asked about their WTP for the piped water system. The question asked was: How much is your household willing to pay monthly for the maintenance of a private connection and 24 h a day access to potable water?
For the revealed preference approach, averting behavior method was used. The averting behavior method begins with the recognition that individuals seek to protect themselves when faced with environmental risk such as contaminated drinking water [
17,
18]. The questions asked were: (1) What is your main drinking water source? Is your water source private or public? (2) How do you assess your drinking water quality: turbidity, odor and taste in a scale—bad, satisfactory or good? Are there visible suspended particles in the water? Perceived water quality was assessed as acceptable or bad based on the answers. If the answer is “bad”, or there is a visible suspended particle in the water, then the quality is “bad”, otherwise “acceptable”; (3) Do you treat (boil, filter or other) your water before drinking? Do you use bottled water or get water from a local treatment point?
The CV method was used to identify WTP for individual water connection and public standpipe. One scenario was developed for all categories of water users depending on respondents’ answer to type of water source that they are using. If they do not use standpipe and do not have the individual connection respondents would answer how they would like to pay for one of them when connected and if the family use a standpipe or have a individual connection and how much they would like to pay for potable water available 24 h a day (
Table 1).
Across the questionnaire, two types of questions were asked. The first, open-ended, directly asked about the maximum amount(s) (s)he would be willing to pay for the proposed water supply improvement. The second was a bidding game, when households are asked different prices until settling at a maximum offered price. The reason for having these two question formats is to see whether respondents react similar regardless of type of asked question.
The split-sample experiment was incorporated into the research design; three different bidding games with different starting points were randomly assigned to respondents in the study. All three bidding games were evenly distributed among the respondents in the survey. The purpose of the split-sample experiment was to test whether respondents’ WTP would be influenced by the magnitude of the first price that they received and the sequence of follow-up questions. There are two differing viewpoints on such a “starting point” test. One is that a differing starting point conveys information about the cost of the service provided. From this viewpoint, different starting points will induce different answers from the respondents. Consequently, if the split-sample test elicits different answers, one would conclude that respondents are in fact taking the CV scenario seriously. A second perspective is that a respondent holding the precise WTP amount in his/her mind and receiving different starting points will provide essentially the same WTP answers. If so, one would have greater confidence that they are revealing a “true” WTP [
14].
Table 1.
CV scenario and the choice of elicitation procedure.
Table 1.
CV scenario and the choice of elicitation procedure.
Type of Elicitation Procedure | CV Scenario for Public Standpipe | CV Scenario for Private Connection |
---|
Open-ended maximum WTP | If water is provided for your village with public standpipes on each street and unlimited potable water supply at any time of the day, how much would your household be willing to pay each month? | Besides the use of water from the public standpipe you can have private connection, that is, the water will be in your house. You will not be able to sell water or use it for watering the garden. If you do not pay a monthly fee, your private connection will be disconnected. How much would your household be willing to pay monthly for the maintenance of a private connection and 24 h access to potable water? |
Closed end, bidding game | How much would your household be willing to pay 100, 200, 500, 700 or 1000 KZT * a month for maintaining the standpipe in your district? | Suppose your household pays for the installation of individual connection (taps at home) and there are already public standpipes so that everyone will have at their disposal good drinking water. Would your household be willing to pay 300, 500, 1000, 1500 or 2000 KZT * each month to have a private connection and 24 h access to potable water? |
In order for the CV-based estimates to be reliable, strategic and hypothetical bias sources were considered. Strategic bias will not occur or will be minor if there is no cost associated with telling the truth and little or nothing is gained if the respondent does not tell the truth. The study tried to estimate this type of bias. Two types of respondents’ answers were considered; the first question was explained as a possible future project and the second was explained as a general survey without practical influence. Avoiding hypothetical bias requires the presentation of believable and familiar scenarios for the resource under consideration. One concern in water supply projects is the permanent availability of good quality water. For this reason, questions were designed as offering 24 h access to treated potable water distributed through the piped system.
The enumerators were specially trained students, following the principle that good enumerators make respondents feel comfortable and at ease. Therefore, the enumerator was not supposed to influence or convince the respondent’s WTP by remaining quite neutral about answers. The enumerators were explained what the study was about, so that they would be able to explain what the maximum WTP is as well as to read slowly and clearly the questions.
2.3. Data Analysis
The collected data were used to calculate the mean WTP according to the following:
where
Pr(Zero) is probability that a respondent has zero WTP;
Pr(Positive) = 1 − Pr(Zero) is probability that a respondent has a positive WTP; and
E(WTPWTP>0) is mean WTP for the positive WTPs.
A few extreme outliers had to be excluded in the open-ended format before analyses. Due to more or less equal distribution of WTP among the bids both for standpipe and private, connection WTP was split into two categories. For private connection, low bids were between 300 and 1000 KZT and high bids were between 1500 and 2000 KZT. For standpipe connection, lower bids were 100, 300, and 500 KZT and higher bids were 700 and 1000 KZT. Binary logistic regression commands in the SPSS software were used to find the maximum likelihood estimation of the independent variables (determinants) as regards lower and higher bids (lower and higher WTP) according to:
where
a and
b are constants and
bX is consumer index.