1. Introduction
A shortage of clean water is one of the world’s most pressing concerns. According to United Nations Water, in 2014 water scarcity affected approximately 700 million people worldwide, with this figure likely to increase to approximately 1.8 billion people by 2025 [
1]. Furthermore, two-thirds of the world’s population live in areas with severe water scarcity [
1], particularly those living in urban areas [
2,
3]. Aside from physical water scarcity, there is also economic water scarcity, which means that water supply is inadequate due to the lack of, or poor water infrastructure, management, and policy [
4]. There is, therefore, a strong imperative to improve domestic water supplies in many urban areas around the world [
5]. Several studies have demonstrated that investment in the domestic water supply can deliver a range of direct and indirect economic co-benefits, including lower healthcare expenditure, time-saving non-health-related benefits (e.g., no queuing at shared water facilities or walking a distance to a water collection site), and water resource protection, among others [
6]. According to the definition of the Joint Monitoring Programme, improved water sources include piped water into dwellings, plots or yards, public taps or standpipes, boreholed, a protected dug well, a protected spring, or rainwater collection, all of which are likely to provide safe drinking-water for communities. Thus, improved domestic water supplies in urban areas should seek to deliver three basic goals, namely: sufficiency in quantity, safe quality of water, and reliable management and service.
Although many households in developing countries lack access to improved domestic water supplies, governments in these countries often cannot afford to offer substantially subsidized improved domestic water to all, or even to the majority of their populations [
7]. As a result, improvements in the domestic water supply in these countries frequently relies mainly on financial contributions (payment of water bills) from households. However, such contributions depend not only on each household’s willingness to pay (WTP), i.e., the maximum amount that households are able to pay for their water supply [
8], but also on each household’s capacity to pay, i.e., the total household income minus the amount to cover basic needs [
9,
10]. WTP information may be used by planners at all levels (national, provincial, city, and rural) to evaluate a project’s economic feasibility, set affordable tariffs, evaluate policy alternatives, assess financial sustainability, and design socially equitable subsidies. Moreover, a cost–benefit analysis would be inadequate without such WTP data; the net economic benefits of an improved domestic water supply are calculated as the difference between the consumers’ maximum WTP for better services and the actual cost of the services [
11]. To estimate the WTP, the Contingent Valuation Method (CVM) is often used. CVM is an economic, non-market valuation method which is particularly useful for determining human preferences for public goods that have no monetary value in the market. CVM is an established method and has found many applications in water-related fields, such as assessing the social value of increasing water quality, reducing risks from drinking water and groundwater contamination, and the provision of drinking water services in developing countries [
12]. In CVM applications to water supply services [
13,
14,
15,
16,
17,
18,
19] the main objectives are to estimate the WTP to improve current water supply services and to explore the factors controlling WTP values via empirical statistical modelling. Mostly, the WTP regression models in these CVM applications are similar in terms of their selection of the WTP as a dependent variable and typically employ demographic factors (usually age, gender, education, income, family size, etc.) as independent variables.
Similar to other developing countries, Vietnam has limited improved domestic water supply, especially in urban areas [
20]. According to the National Environment Report [
21], it is estimated that only about 70% of the population has access to potable water. Hanoi, Vietnam’s capital city, is now facing several water scarcity issues connected to its urban water supply. The fast rate of urbanization and rapid increase in the city’s population (~3.4%/year) are significantly inflating the demand for clean water [
22]. Meanwhile, the quality of the water resources that are being used to supply Hanoi is decreasing. Domestic and industrial effluents have polluted surface water sources in river basins of many major rivers, such as the Red, Nhue, and Day Rivers located in the Red River Delta. It is estimated that between 100,000 and 150,000 m
3/day of untreated industrial wastewater flows directly into the rivers in Hanoi alone [
23]. Furthermore, upstream of the administrative area of Hanoi, the quality and quantity of water resources are significantly affected, affecting abstraction possibilities. Hanoi’s water issues are exacerbated further by high water loss rates, averaging 23% [
24]. As a result, it is critical to improve the water supply system in Hanoi by investing in the necessary supporting infrastructure, adjusting water allocations to fulfill residents’ demands reasonably, and by improving the quality of water supplied so that it meets the Vietnam Ministry of Health Quality Standard QCVN 02-BYT. Thus, the socialization of investment capital is critical to offer financial support for these activities, especially in developing countries [
25].
This study explores these issues within the context of the case study of Hanoi by developing a CVM-based process that focuses on three important points: (1) a naturally exploring WTP technique aligning with the way in which humans think, (2) determining whether the results of the CVM’s social investigation are consistent and reliable, and (3) finding the appropriate variables to include in the WTP regression model based on the groups of key factors that significantly affect WTP and the current circumstances of Hanoi’s urban domestic water supply system’s (HUDWSS) performance. With the application of the developed CVM-based process to Hanoi, it is possible to examine the factors affecting residents’ WTP, thus providing an essential first step in the improvement of the domestic water supply and community expectations in Hanoi.
2. Study Area
Figure 1 displays the study location, consisting of ten urban districts, and the main rivers and lakes of Hanoi. As the economic, political, and social center of the country, Hanoi’s population and its density are very high compared to other developed cities in Vietnam. In 2020, the total population of Hanoi was approximately eight million people; female and male residents accounted for a similar proportion, and almost half of the total population was comprised of urban residents, according to the General Statistics Office of Vietnam. Average monthly income per capita in Hanoi was an estimated 6.3 million dong (about 280 US dollars).
Hanoi’s Urban Domestic Water Supply System (HUDWSS) faces a number of significant challenges in delivering its plans to supply improved water to all residents before 2030. Currently, the per capita average urban water demand is approximately 200 L/day [
26] However, the municipal government is facing several challenges in providing sufficient water sustainably due to: (i) the ever-increasing urban population and density; (ii) serious pollution of both the surface and groundwater resources that serve as the main input water sources for HUDWSS‘s operation, and especially; (iii) the large losses of water (due to leakage) that can then not be used to generate any revenue.
Regarding the first challenge, the rapid urban growth places great pressure on natural resources and the environment, putting particular stress on HUDWSS to meet the growing water demand of customers, causing the fragmentation of Hanoi’s urban water infrastructure [
27]. In recent years, Hanoi’s population has increased rapidly; the density of the nine urban districts has increased up to 11,759 people/km
2, while the urbanization rate reached 40.5% in 2013 [
26]. In recent years, the population grew by approximately 3.4% per year. In just two years, from 2015 to 2017 the population grew from 7.2 to 7.7 million people, and the population density rose from 2000 to 2209 people/km
2 [
28]. The city’s water distribution system, including new and old networks, has a long history of construction and rehabilitation. Hanoi Water Limited Company (HAWACO) is the city’s largest water distribution enterprise and has legal status under the Hanoi Transportation Department. The company was established under Decision No. 546/QDUB dated 4 April 1994 of the People’s Committee of Hanoi, with a history that can be traced back to the nineteenth century. Currently, Hanoi has 12 main water plants and, (including the district of Ha Dong) eight water supply zones managed by HAWACO, Vietnam’s Freshwater Business and Construction Investment Joint Stock Company (VIWACO), and Ha Dong Waterworks [
29]. As reported by HAWACO, the average water supply capacity is 1,462,000 m
3/month, of which 35% is distributed to the old network, which mainly serves the Old Quarter’s communities; and 65% is distributed to the new network, which covers the inner, the west, and the southeast regions of Hanoi, see
Figure 1 [
30]. HAWACO is failing to meet current water needs. With the per capita water demand in the urban districts at approximately 130 L/person/day, the public water utilities failed to supply urban districts approximately every two days per month in 2016 [
30]. Consequently, as mentioned by HAWACO [
31], just 55% of the city’s population has access to HUDWSS even though the public distribution network fully covers all of the urban districts. However, around 30% of urban households use freely accessible well water sources [
32].
Regarding the second challenge, both surface and groundwater resources for HUDWSS are seriously degraded and polluted. This critical situation challenges the water enterprises in terms of how to provide a high-quality water supply for local communities. The majority of input water sources for HUDWSS consist of groundwater harvested from Pleistocene aquifers. Groundwater resources are distributed unevenly, with the largest recharge of 700,000 m
3/day in the south and the smallest of 66,000 m
3/day in the Soc Son district (
Figure 1). The disastrous situation of seriously degraded groundwater quantity and quality as a consequence of inappropriate usage and management has been comprehensively presented in a number of previous studies [
33,
34,
35]. In addition, the surface water sources in rivers and lakes are also seriously contaminated due to solid waste dumping, and domestic and industrial wastewater flowing directly into the water bodies without treatment. The total domestic and industrial wastewater volume in the central area of Hanoi is approximately 600,000–700,000 m
3/day; the combined capacity of all the water treatment plants in Hanoi, however, is only around 245,000 m
3/day. Hence, two-thirds of the generated wastewater is not treated before being discharged into rivers and lakes [
12]. Consequently, water quality parameters are far below the national water quality standards. For instance, as reported by HAWACO in 2016, the observed values of an important water quality parameter of chemical oxygen demand of Quynh lake, which is located in our targeted urban district Hai Ba Trung, is extremely high at 136 mg/L, compared to the recommended value of 10 mg/L, which is the national standard for good surface water quality. Moreover, major rivers such as the Red River, Da River, and Day River are interprovincial water resources, and they have a great impact on the quality and reserves from the watersheds; thus, it is difficult to use these water resources.
Regarding the third challenge, according to HAWACO [
30], the most pressing issue for the Hanoi water supply sector is the large volume of non-revenue water, which is approximately 23%. In fact, considering around 600,000 customers and the total capacity of plants of 534,500 m
3/day, it is estimated that the quantity of per day non-revenue water is approximately 2.8 times the existing supply capacity. The rate of non-revenue water loss was substantially reduced from 38% in 2007 to 23% in 2015 as a result of the efforts of HAWACO’s management. The average water price as of 2020 is 7000 dong/m
3 (~USD 0.31/m
3) in urban districts; thus, this massive wasted budget could be estimated as more than 1 billion dong per day (~USD 43,700/day), while the residents still lack a water supply. The main reasons for this huge loss include the poor maintenance of water pipelines, inaccurate water meters leading to the incorrect recording of water use, or even water theft or illegal water tapping, as observed in other developing countries [
36]. This critical circumstance threatens the sustainability of the financial situation of HUDWSS and thus challenges the ability of HUDWSS to meet the goal of achieving a sufficient water supply for all, as mentioned in the global sustainable development goals.
3. Contingent Valuation Method-Based Process for Exploring the Factors Affecting the WTP of Hanoi’s Urban Domestic Water System
The CVM used here employed a survey of Hanoi water supply service customers (see
Appendix A) that consisted of five main sections: (1) a section asking questions concerning respondents’ public awareness of the water supply service situation, with the aim to understand the community’s response to the service; (2) the presentation of the CVM scenario; (3) several questions for those who do not use HUDWSS; (4) a question asking about the respondent’s WTP for the improvement of the water supply service, and; (5) a series of demographic questions. This survey outline follows the general process of the application of CVM to elicit WTP based on a range of previous studies [
37,
38,
39]. We further developed these general sections to apply them to the specific situation of HUDWSS. The main objectives addressed in the questionnaire were (i) to understand the current situation, public perception, and satisfaction regarding the domestic water supply among local communities; (ii) to explore the WTP of local residents for water supply improvement; and (iii) to enable the subsequent analysis of the key factors affecting the WTP. Conducting a comprehensive investigation, we finally had 402 respondents, which as discussed in
Section 3.3, is sufficient to give adequate statistical power.
In the application of the CVM developed in this study, we developed three methodological innovations. First, we improved the technique of determining the respondents’ WTP values step by step, logically and naturally, in line with natural thinking patterns; the so-called ‘naturally exploring’ WTP technique. Secondly, we checked whether the obtained responses from the CVM social investigation were acceptably consistent and reliable, and this step is essential since the social investigation is usually carried out in diverse situations. Thus, its reliability and consistency are uncertain, depending on awareness, the convenience of the interview process, and even the personalities of both interviewers and respondents. We then clarified the groups of main factors affecting WTP, which have not been mentioned in the previous literature. This step is also crucial because it provides us with the background to propose a list of appropriate variables that should be included in the WTP regression model. The details of these three points are explained in the following subsections.
3.1. Questionnaire Design
In order to gain a better understanding of the current domestic water use situation in urban communities in Hanoi, this study conducted a social investigation based on a questionnaire survey and a face-to-face interview. The questionnaire was set up and completed in two phases. The first phase was a pilot survey, in which 10 samples were collected to test how the respondents understood the primary list of questions and how much information the interviewers could collect from the face-to-face interviews. After the pilot survey, a few questions were changed, making them easier to understand for the respondents and thereby increasing the effectiveness in approaching the problem.
The final questionnaire consisted of five parts, as we needed to deal with respondents with and without a water supply from HUDWSS. The first part sought to understand which water sources were used for domestic purposes; the second part aimed to provide a list of questions for those who used water from HUDWSS. Here, we attempted to measure the satisfaction of the communities in terms of the quantity, quality, and management of the current domestic water supply; the third part addressed communities that did not use water from HUDWSS; the fourth part enquired about which aspects the communities wished to improve regarding the performance of HUDWSS and the amount that the communities were willing to pay for this improvement in the future; and the final part focused on the residential demographic variables such as age, gender, education level, and monthly income. The demographic information provided the basis for understanding the factors affecting the respondents’ WTP for the HUDWSS improvement.
3.2. The Naturally Exploring WTP Technique
To determine respondents’ WTP, one of the most widely used WTP questioning techniques is open-ended questions [
37,
39,
40,
41,
42,
43,
44]. In an open-ended question, the respondents are asked to state the maximum amount they could pay for water-targeted improvement. The advantage of this technique is that the question is easy to understand and gives the respondents freedom in giving their WTP values. However, as highlighted in a previous study [
38], the open-ended question technique can result in several “zero bids”, i.e., respondents indicating a WTP of zero. Therefore, to avoid this zero-WTP situation in our investigation, before asking about the residents’ WTP, we provided them with information about the price they are currently paying for 1 m
3 of water supplied by HUDWSS. By doing this, their WTP values were at least equal to their current price for water use, the so-called WTP
0. Thus, in practice, we used three consecutive questions to determine WTP values as the maximum amount that they were willing to contribute. The first question was “To improve the current HUDWSS to the level of your expectations, are you willing to support the water price?”. Upon answering YES to this question, they were asked the second question, “The current price of water supply is 7000 dong/m
3, how much do you think that this current price could be increased to have a better budget for HUDWSS improvement?”. When answering the second question, respondents were offered several levels by which the current water price WTP
0 could be increased, resulting in pre-WTP values. The third question, “What is the maximum amount which you are willing to support to improve HUDWSS’s performance?”, was used to determine the maximum amount that the respondents were willing to pay, which is referred as their actual WTP values.
3.3. Sampling
The minimum sample size was determined according to Krejcie and Morgan [
45]; for large areas, the number of participants required for the survey was calculated as follows:
in which
S: minimum sample size;
X2: the table value of Chi-square for 1 degree of freedom at the desired confidence level (3.841);
N: the population size;
P: the population proportion (assumed to be 0.5 since this would provide the maximum sample size);
d: the degree of accuracy expressed as a proportion (0.05).
The estimated population in these Hanoi urban communities was 3,962,310 in 2019. Thus, the minimum sample size was calculated as 384 samples. In this study, we first conducted a pre-test survey of ten samples of several staff working at the Hanoi University of Natural Resources and Environment and the residents living near this university in order to determine whether our questionnaire was understandable and appropriate. We then finalized the questionnaire and started conducting the comprehensive social investigation in the three most urbanized district representatives in Hanoi urban communities, including the newly urbanized Ha Dong and two old inner-city districts of Thanh Xuan and Hai Ba Trung (
Figure 1). After three months of surveying from November and December of 2019 to January of 2020, we had randomly collected 454 samples. Among these, we could not use 52 samples, mainly due to missing information for more than 30% of the questions. Hence, the final data set comprised 402 samples, of which 91 samples were from the Ha Dong district (353,200 residents in 2019), 155 samples were from the Thanh Xuan district (286,700 residents in 2019), and 156 samples were from the Hai Ba Trung district (311,800 residents in 2019).
3.4. The Theoretical WTP Regression Model
In the literature, we found that 21 main factors/variables that could be considered to affect the WTP for water supply and water-related service improvement. These factors/variables could be divided into four groups: demographic factors, water supply quantity, quality, and service, as shown in
Table 1. Therefore, the theoretical WTP regression model could be formulated as in Equation (2).
in which
WTP is the dependent variable;
De are demographic variables;
Quan,
Qual, and
Ser are variables regarding water supply quantity, quality, and service, respectively;
CDe,
CQuan,
CQual,
CSer are coefficients; and
e is the random error. The IBM Statistical Package for the Social Sciences (SPSS) was used to generate the WTP regression model and perform data analysis.
Selection of factors/variables in each group: Depending on the targeted water supply issues and the public preferences in a specific study area, the factors/variables of water supply quantity, quality, and service groups could be selected. In terms of demographic factors, this selection depended on the characteristics of the communities that the researchers wished to test for their correlations with the WTP.
3.5. Reliability Check
To validate the reliability of the responses in the questionnaire, we used a statistical measure of confidence consistency, namely Cronbach’s alpha coefficient, for a given sample. The coefficient formula is according to Equation (3), as indicated in [
79].
in which
is the number of components (usual questions);
is the variance of the observed total test scores, and
is the variance of component
i for the sample.
Reliability scale: As mentioned in [
79], the coefficient values in the range of (0.9; 1.0), (0.8; 0.9), (0.7; 0.8), (0.6; 0.7), (0.5; 0.6), and (0; 0.5) indicate respectively excellent, good, acceptable, questionable, poor, and unacceptable internal consistency.
3.6. Application for HUDWSS
Applying our CVM-based process, we obtained 402 useable samples from the three urban districts. Using Equation (3), we find a value for α of 0.879 for the responses to the group of questions regarding the type of water used daily; 0.863 for the responses to the questions regarding the residential satisfaction towards HUDWSS‘s water consumption; 0.943 for the responses to the questions regarding the residential satisfaction towards HUDWSS’s water quality; and 0.865 for the responses to the questions regarding the residential satisfaction towards HUDWSS’s service and management. These coefficients are all higher than a threshold value of 0.8, indicating that the obtained responses are internally consistent and reliable. Therefore, the data set obtained from these 402 reliable responses was used for further analysis.
Table 2 provides further details about the characteristics and status of the survey sample. From the total of 402 people interviewed, 208 (51.7%) were female, and 194 (48.3%) were male. The majority of the respondents’ ages were in the range of 24 to 55 years (66.9%), and only one respondent was under 18 years old (0.2%). Of the five occupation groups mentioned in the questionnaire, business was the most common, accounting for 46.3% of all occupations. The level of education of the surveyed subjects was partly reflected through their occupation. The more educated subjects were aware of the importance of clean water and the impact of inadequate water quality on their health. It was observed during the pilot survey that interviewees did not wish to disclose their actual incomes due to various personal reasons. Consequently, dividing the total family income into ranges of values in this study made the respondents more comfortable in indicating their relative income. The number of respondents with a total family income of between 10 to 30 million VND (approximately between 434 to 1304 USD) accounted for half (51.7%) of the sample.
5. Conclusions
This study successfully proposes a CVM-based process and tests its effectiveness in an investigation of Hanoi urban communities WTP regarding the improvement of HUDWSS. Our results show that Hanoi urban communities were essentially satisfied with the water supply quantity, as our investigation found that more than 90% of surveyed respondents considered their water supply to be mostly sufficient. The water quality was still lower than the quality expectations of the urban residents since most households (80.4%) used advanced purifiers to treat the tap water before drinking and cooking. Almost half of the respondents complained about the lack of maintenance services. From the regression model results, we found that significant factors (at p-value = 0.1, 0.05, and even 0.01) affecting the WTP are gender, age, income, water use, payment method, and maintenance; meanwhile, occupation, family size, quantity and quality satisfaction, treatment, and notification were found to be insignificant factors (at p-value = 0.1). These findings reveal the crucial role of understanding the target problems in selecting and proposing appropriate variables to increase the effectiveness of the WTP regression model. Our results also show that the naturally exploring WTP technique proposed in this study makes it easier for both respondents and interview conductors in determining the WTP values. The average WTP is approximately 1.4% of the average household income, well below the 2.5% threshold established by the US EPA as an indicator of the affordability of monthly water payments among citizens. However, the implementation of policies will take several years, especially in developing countries such as Vietnam, where the average household income is considered low compared to other developed countries, and should take into account the varying income levels among households. This shows the practicality of a future social investment fund contributed to by the communities that is used for upgrading and improving the quality of Hanoi’s urban water supply services. The willingness to pay (WTP) technique offered genuine results that helped to make realistic recommendations to the policy- and decision-makers without any complications. The methodology that we developed for this study can be applied to any similar area, without any geographical limitations, in several STEM and social science subjects and multidisciplinary fields. These characteristics make this model unique and easy to use in all circumstances.
Following are several remarks about how to get highly reliable data in such CVM applications. One thing was asking the cooperative residents who were willing to spend more than 15 min to finish the questionnaire. This thing mainly depended on how the interview conductors started asking the questions. Another factor involved about the investigation approach, which significantly affected the reliability of the collected data. Using email and Google Forms i could reduce the survey time and make it more convenient to complete, but the reliability of the obtained data is usually low in comparison to the face-to-face interviewing approach. In this study, the face-to-face interviewing approach was employed to maximize the possibility of obtaining highly reliable data. The reason is that the respondents usually do not fully understand all the questions, thus the interview process should be like a friendly discussion. Particularly in the CVM investigation, the respondents should imagine the unreal market and give the most proper payment for the goods (domestic water supply, in this case). The friendly face-to-face discussion was thus crucial to help the respondents in finding their appropriate WTPs. Moreover, as experienced from our investigation, mentioning the current water price of WTP0 = 7000 dong when asking about the WTP really affected the respondent’s opinions. All of the obtained Pre-WTPs and WTPs were higher than WTP0. That means that the respondents thought of the WTP0 as their acceptable minimum payment for 1 m3 of water. In addition, the WTP0 could also give us the possible variation range of respondent WTPs. In our case, no respondents gave a WTP of more than three times the WTP0 (i.e., about 21,000 dong). Therefore, in order to obtain a highly reliable data set, it is possible to eliminate the response where the WTPs are higher than 3*WTP0. Regarding future research, we would like to apply the same methodology by combining physical (water quality and quantity) and social science (people income, employment, gender) components in other similar areas in Vietnam. Moreover, we would like to expand our study area to other South Asian countries, such as India, the Philippines, Sri Lanka, and China. Our collaborators have also shown interest in applying this unique method to the countries mentioned above.