Willingness to pay for improved water supply service in coastal urban settings: evidence from Khulna, Bangladesh

Access to and use of improved water has been a pressing global challenge, posing threats to the livelihood option, functions of ecosystems, and human health. The inadequacy and degraded quality of water are perhaps the most critical environmental concerns, especially in developing countries. The crisis of safe drinking and household water usage is predominantly prevailing in the southwestern coastal zone of Bangladesh (Islam et al. 2019).

Urban water supply systems throughout the country are facing acute water supply and quality problems due to various reasons such as high population growth, urbanization, and industrialization (Haque et al. 2020), decreasing groundwater supply, poor water management, increasing water treatment costs, and so on (Asian Development Bank 2009). Besides, anthropogenic wastes, untreated industrial effluents, and toxic and inorganic pollutants are contaminating the aquifers and potable water sources (Hasan et al. 2019), intensifying the scarcity of improved water (Arefin & Mallik 2018) and causing threats to public health (Parvin et al. 2022). It is estimated that 55% of city dwellers must rely upon contaminated traditional sources like ponds, canals, tube-wells, ditches, and dug wells (Sarker & Alam 2013).

Khulna is an expanding coastal urban center in the southwestern part of Bangladesh (Uddin et al. 2006), beset with salinity intrusion in groundwater and aquifers (Sarker et al. 2021). The most menacing challenges to securing fresh and improved water are the salinity intrusion in surface water due to sea-level rise (Mahmud et al. 2020; Asma & Kotani 2021) and the consequent alteration of hydro-chemical properties of water due to salinization (Sarker et al. 2021). A current study reveals that Khulna city has a demand for 240 million liters per day (MLD), yet unmet by the supply of 112 MLD water only (Islam & Ali 2016). In addition, Gunatilake & Tachiiri (2012) estimated that only 22% of households in this urban area have access to piped drinking water, although most of them perceived the water as filthy and of poor quality.

Adequacy, reliability, and quality are crucial for household water supply services (Majuru et al. 2018). A positive correlation has been found between national income and access to improved water supply (Vásquez et al. 2009; Wondimul & Bekele 2011). The World Bank (2018) predicts that access to clean water and sanitation will help reduce the poverty rate in Bangladesh in an accelerated manner. However, the key constraints of connecting households with piped water are the financial backwardness of the consumers and the structural complexities of the supply authority. The provision of clean water involves huge costs to the poorer city households (World Health Organization 2012); nevertheless, the demand for improved water service is constantly rising among the urban inhabitants of developing countries (Boretti & Rosa 2019). For instance, Gunatilake & Tachiiri (2012) determined that the monthly mean willingness to pay (WTP) for the service in Khulna, Bangladesh, is BDT (Bangladeshi Taka) 301.¹Vásquez & Espaillat (2016), applying the contingent valuation method (CVM), estimated that the residents of San Lorenzo, Guatemala, are willing to pay 200% more than their water bill to enjoy the benefits of improved and safe drinking water.

Sustainable Development Goal (SDG)-6 has emphasized the need for ensuring access to improved water and hygienic sanitation for all (United Nations 2015, 2016), which is further linked to the improvement of socioeconomic status and other SDG indicators (Alcamo 2019). More precisely, improved water is not only essential for the improvement in primary health status but also subsidiary for livelihood options and poverty reduction (Odwori 2020). Alcamo (2019) found some positive connections between water quality and the SDG indicators. For instance, improved water quality is a positive indicator of good health and wellbeing (Prüss-Ustün et al. 2014). Furthermore, the currently 844 million people worldwide lacking access to improved water supply (WHO and UNICEF 2015) is predicted to stand to more than five billion due to climate change and incremental demand by 2050 (United Nations 2015). Reasonably, highlighting the demand for and scarcity of improved water has gained special importance for sustainable development (Odwori 2020).

Perceiving the significance of improved water from the perspective of demand, scarcity, and salinity in the coastal zone, the study aims to explore the demand for improved water service for the ore households of Khulna City Corporation (KCC) by eliciting the WTP and the associated determinants of the expressed WTP. The city dwellers are not satisfied with the quantity and improvement of the supplied water for their daily usage (Uddin et al. 2006). Moreover, although the rudimentary source of drinking and sanitation is groundwater, it contains contaminants, iron, arsenic, and salinity (Uddin et al. 2006; Gunatilake & Tachiiri 2012). Therefore, this approach of deriving demand for improved water supply will clarify the understanding regarding the necessity of improved water quality in urban coastal settings. In addition, the WTP determination can be a policy formulating tool in securing improved water service for coastal city dwellers in the backdrop of water scarcity and salinity. Thus, the study can be an initial investigation on how to thrive in SDG 6 by redesigning the water policy in the urban environment. This is because insights into household water demand would deliver a better understanding and way of how to manage water resources to fulfill household demand (Nauges & Whittington 2009). In this regard, the determinants of improved water demand can be helpful in analyzing the sustainable mechanism of how to make the water supply service affordable to all kinds of households (Coster & Otufale 2014).

To facilitate the objective, the study sets the following four specific research questions: (i) Are households in the study areas willing to pay for improved water service? (ii) What are the factors that influence household WTP for improved water supply service?; (iii) How much are residents willing to pay for improved water service and is the amount significant enough for the provider to supply improved water to city dwellers?; and (iv) What is the volume of aggregate welfare and potential revenue based on the expressed WTP amount in the presence of improved water supply service?

In developing countries, many urban areas face unevenness between supply and demand of reliable supply of improved water (Rodríguez-Tapia et al. 2017). Unreliable water supply and shortage of water also negatively affect human life in various ways like health costs, economic costs, and so on (Soto Montes de Oca & Bateman 2006). On the other hand, access to clean water is a precondition for the achievement of several human rights, including those pertaining to people's survival, education, and standard of life (Hundie & Abdisa 2016). Access to clean and improved water for all helps prevent illnesses caused by the use of polluted water such as diarrhea, cholera, guinea worm, typhoid, and fatalities (Eridadi et al. 2021). In addition, other positive externalities of using improved water are verified in agricultural, industrial, and ecosystem development (Sriyana et al. 2020; Eridadi et al. 2021).

Many former studies extracted the WTP for improved water services on different populations and spatial environments (Wondimul & Bekele 2011; Del Saz-Salazar et al. 2015; Tussupova et al. 2015; Vásquez & Espaillat 2016; Rodríguez-Tapia et al. 2017; Wang et al. 2018; Tavárez et al. 2021). Casey et al. (2006) have determined the willingness to pay for improved water services in Manaus, Amazonas, Brazil, worth US$ 6.12 per month. Rahman et al. (2017) operated a similar experiment in Rajshahi city of Bangladesh and concluded that the city households were willing to pay an average of US$ 1.01 per month as the tariff required for the development of a water supply system. Applying the same method, Tenaw & Assfaw (2022) found that the Dire Dawa city residents agreed to pay US$ 0.51 per 20 L of water. Another study in Chennai city also reveals that 28% of the surveyed poor city dwellers are willing to pay for improved water supply services (Venkatachalam 2015). Similarly, Tussupova et al. (2015), studying in the Pavlodar Region, reported that 90% of the consumers agreed to pay for improved water supply services and simultaneously made inferences on the mean WTP amount of US$ 10.6 per month. Shah et al. (2016) identified the monthly amount worth US$ 2.40, totaling US$ 544,000 per annum for the Swat river valley region. Akram & Olmstead (2011) also found a substantial amount of WTP (US$ 7.50–9.00 per month) for improved water supply in Lahore.

Several socioeconomic and water-specific factors are responsible for the WTP for improved water supply services. To illustrate, household income (Vásquez et al. 2009; Coster & Otufale 2014; Tenaw & Assfaw 2022), education (Arouna & Dabbert 2012; Del Saz-Salazar et al. 2015), gender (Coster & Otufale 2014; Odwori 2020), family size (Chatterjee et al. 2017), (Tussupova et al. 2015), and water quality (Tussupova et al. 2015; Tenaw & Assfaw 2022) positively and significantly affect the WTP. However, water quality (Coster & Otufale 2014; Rodríguez-Tapia et al. 2017), bidding price (Coster & Otufale 2014; Odwori 2020), and distance to water sources (Wright 2012), are also deduced as the negative determinants of WTP for improved water supply across different geographical sites.

Although evaluating environmental or non-marketed goods to derive the demand for that good can be conducted by several approaches. The CVM is one of them, which is a stated preference method (Chatterjee et al. 2017; Islam & Farjana 2020). The travel cost method (TCM) is also prominent among other types of observed approaches (Coster & Otufale 2014). The majority of the former studies aiming to unearth the WTP for improved water supply used the CVM approach (Khan et al. 2010; Kanayo et al. 2013; Tenaw & Assfaw 2022). However, a good number of studies followed double-bounded dichotomous choices under the CVM to evaluate the WTP for improved water service (Kim et al. 2021).

In Bangladesh, the challenges inherent to the supply of improved water in urban areas are salinity, climate change, and lack of proper water management (Chowdhury 2010; Chan et al. 2016). Precisely, in Khulna city, education on sanitation and water, household expenditure, and source of water significantly determine the demand for improved water supply services (Gunatilake & Tachiiri 2012). Thus, the problem in Bangladesh is not only from the supply-side issue; instead, the problem is created because of too little emphasis on the demand aspects of water and sanitation.

The objective of the study is to evaluate households' willingness to pay (WTP) for improved water services in the urban areas of Khulna city. To meet the objective, the study applied the contingent valuation (CV) method. CV is a reliable stated preference method to measure WTP for a good not traded in markets. Under this approach, a closed-ended CV approach is used for estimating WTP for improved water service. Closed-ended questions are considered an incentive-compatible and standard approach to simulating WTP. It asks the respondents whether they would pay the minimum amount stated. The question makes respondents answer with either a Yes or No. Thus, this decision format formulates an actual market for a private good or service. The study followed this method to ask the households to bid WTP for improved water supply services.

The survey was carried out in KCC of Khulna District, located in the southwestern region of Bangladesh. KCC consists of 31 wards that hold a total population of 1.50 million, with 70,221 households bearing valid holding numbers.² A total of five wards were selected as the study sites out of 31 wards. The wards were randomly chosen from the ward list. The selected wards were 5, 7, 16, 17, and 30 which are located in Khalishpur, Daulatpur, Boyra, Sonadanga, and Tutpara regions, respectively.

CV survey questions were prepared and formatted as a questionnaire to survey the city households. Based on the presumed selection criteria, two types of households are selected as the sampling units for the survey. The first ones are those connected with piped water, and the second ones are not. Based on these criteria, initial information regarding the households was gathered from the KCC authority. Also, the questionnaire was pretested by conducting a pilot survey on 13 households residing in the city area before initiating the final data collection. Based on the corrected questionnaire, an in-person survey was conducted following a simple random sampling method. First, 20 households were randomly selected from each ward based on the household list provided by the KCC. Second, to maintain a chronological order, data enumerators collected information from the respondents. If any household refused to participate, the following immediate holding number was considered to substitute the previous respondent. The data collection was held for 4 months, from January to April 2019, surveying the five sites once a week, mostly at the weekend. Thus, a total of 100 households participated in the CV survey.

The survey questionnaire was divided into three main sections. The first section asks respondents questions about their socio-demographics, such as age, gender, family size, residence, education, income, and employment status. Section two was designed to record the details of current water sources, water expenditure, and perception of the quality of water. The final part was formulated to understand the WTP for improved water service. In this part, a single-bounded dichotomous question was introduced first to know whether the households were willing to pay. More precisely, the participants were asked: Are you willing to pay for the authority's improved water supply service, which will provide you with a constant supply of potable, fresh, chemical smell-free, and contamination-free water? Based on their response, a single bid amount was opened in front of them, describing the hypothetically improved water quality service to record the amount of WTP. For instance, a respondent whose monthly income falls into the US$ 233–349 class, was asked: are you willing to pay US$ 4.65 for the improved water supply service? Likewise, the study offers five pre-determined bids to the households, depending on the information the participants provided on income (Table 1). If a random amount was offered irrespective of the income level, the lower-income group could be non-agreed to pay if the offered amount is beyond their income level. Therefore, income status is a reliable criterion of the bid as it justifies the affordability and ability of the respondents to respond to the bid agreement (Moffat et al. 2011; Asim & Lohano 2015). More convincingly, the lowest bid was US$ 3.50 for the monthly income group US$ 116–233. The bid was US$ 4.65 for those whose monthly income ranges from US$ 233 to 349. Similarly, the bid was US$ 5.82 for the income earners who fall into the US$ 349–465 range, US$ 6.98 for the US$ 465–698 range, and the highest bid was US$ 8.15 for the highest income class >US$ 698.

Bishop & Heberlein (1979) developed the single-bound dichotomous choice (SBDC) model. Only one question is asked to the respondents under this method, and the amount is treated as a threshold. If the good is valued more highly than the threshold amount, the respondent answers yes; otherwise, no (Lopez-Feldman 2012). The SBDC requires less information. It is easier to implement at the data collection and estimation stages and avoid systematic bias in responses (Hundie & Abdisa 2016).

We can obtain a consistent estimate of mean WTP using and (estimated with the Probit command), even though we do not know the true value of the WTP amount. Equation (6) demonstrates the procedure of estimating a consistent mean WTP based on the regression in which is the vector that comprises the values of interest for the explanatory variables (Lopez-Feldman 2012). Simply, to get the adjusted mean WTP controlling the other explanatory variables in the Probit model, we need to set values to the means of the explanatory variables. After getting the mean values of all the explanatory variables, we created scalars for each explanatory variable used in the Probit model. The estimated coefficients of the explanatory variables (⁠⁠) is divided by the coefficient attached to the bid variable (⁠⁠). Next, we estimated the consistent mean WTP using Equation (6) with the estimate of the Probit model. The purpose of estimating a different mean WTP despite the traditional average WTP obtained from the summary statistics is to obtain a consistent and adjusted mean WTP controlling other characteristics of the respondents. This estimate is more robust than the conventional average.

In summary, the study firstly estimated the parameters of the predictors of WTP decisions for improved water supply services using a Probit model. Secondly, utilizing the Probit estimate, a consistent mean WTP was derived, controlling the explanatory variables. In the final stage, the study extracted the potential revenue and loss from the consistent mean WTP.

Household background characteristics summarized in Table 2 reflect the simple statistics of socio-demographic status, WTP, and water-specific variables. Of the samples, 52% expressed their willingness to pay for improved water supply services. The average bid amount revealed by the respondents was US$ 5.82, which slightly differs from the consistent mean WTP of US$ 5.05 estimated with the help of SBDC and the Probit model in section 4.2. The male-headed households were dominating participants in the survey since they constitute more than half of the samples (52%) while the female counterparts were 48% of the total samples. The average age of the participants is 36 years, with a standard deviation of 7.42 years indicating a slight variation in age among the participants. Similarly, despite the mean schooling years of the respondents being nearly 10 years, the educational accomplishment is uneven (standard deviation = 4.9 years), ranging from no enrollment to 17 years of education. The average household income was US$ 339 which is adjusted by a wide standard deviation of 159. It indicates that the average household income of the participants is inconsistent and lopsided across the lower bound of US$ 116 and the upper bound of US$ 873. The participants required an average of around 14 L of drinking water. They had to spend 20 min on average to collect water from the collection points. Since the average response on dissatisfaction is approximately 3.0, the households are expected to be most dissatisfied with the water quality (1 = not dissatisfied and 4 = highly dissatisfied).

Survey results report that the households' background characteristics and water-specific traits maintain heterogeneity depending on the willingness to pay status for improved water (Table 3). For instance, the average age of the household heads unwilling to pay is around 35 years. In contrast, the household heads willing to pay have a mean age of 38 years, showing a significant difference between the two classes. In the same way, household size, educational attainment of the household head, and household income differ for those willing to pay and those unwilling to pay. As shown in Table 3, the average household income of the agreed-to-pay participants is nearly two times the average household income possessed by the non-agreed-to-pay participants, delineating the higher average income of the former class. The mean and variance test testify this gap is significant at a 1% level. It indicates a positive predictive connection between the level of earnings and WTP.

The significant difference in mean and variance of the water-specific variables subject to the willingness to pay information are also apparent in Table 3. To illustrate, the agreed-to-pay households require 25 min on average, more than 1.5 times the time required for the non-agreed-to-pay households (14 min) to fetch water from the collection points, and the variation is statistically significant at 1% level. Importantly, the finding provides an insight into the demand for reduced collection time which is believed to be met by the demand for improved water service. Therefore, the positive expression of WTP is documented by the households spending much more time compared to the households spending less collection time. Likewise, as expected, the amount of drinking water required by the agreed-to-pay households is higher than that of non-agreed-to-pay households, showing a 1% level of significance.

Gender, an important categorical response, shows that most of the households are male-headed (52%) and the rest is female-headed households (48%). Among the male-headed households, only 22% are willing to pay for improved water service. On the contrary, amongst the female household heads, 30% of participants showed their enthusiasm to pay for improved water supply service. The proportion test and χ² value confirm that the gender percentage over the WTP decision has a significant difference (0.20, significant at 1% level). Among the households with private sources of water (22%), only 2% of the households agreed to pay for the improved water supply service. In other words, among the privately managed water group, the agreed-to-pay households are lower than the non-agreed-to-pay households.

Table 4 displays the multivariate regression of the Probit and single-bound model separately to determine the factors associated with the WTP and bid amount for improved water supply service. According to the Probit regression, as presumed, the bid amount negatively influences the decision for WTP, although meagerly by 0.01 percentage point (significant at 1% level). Conversely, education and household income are positively predicting the decision of WTP. The former variable is significant at 10% and the latter is significant at a 1% level. To elucidate the effect of education and income on WTP decision, Table 4 shows that an extra year of schooling has a 0.07 percentage point positive impact on the likelihood of WTP decision; and a small rise in income is likely to influence the probability of WTP decision by 1.94 percentage points. Meanwhile, gender, a crucial demographic factor, suggests that male household heads are almost 0.3 percentage points less likely to express the WTP for improved water service compared to female-headed households (significant at 1% level). Under the same model analysis, a private source of water, compared to other sources, lessens the expected probability of the WTP decision by 1.66 percentage points. The result is statistically significant at a 1% level. The higher the water collection time, the more likely the WTP decision is positive (significant at 1% level). The dissatisfaction with water quality (degradation) is also predicted to reduce the chance of WTP expression.

Controlling the other variables in the single-bound regression estimate, the monthly mean WTP has been accounted for worth US$ 5.05 using Equation (7) (Table 5). Counting upon the average WTP estimate, the study computes the aggregate monthly welfare worth US$ 0.35 million, and the amount totalizes US$ 4.25 million per annum, as depicted in Table 5. To ascertain the worthiness of such potentials to the respective authority like KWASA requires a cost-benefit analysis portrayed in Table 6. The analysis importantly outlines the consumer surplus (CS), potential revenue, and the respective revenue loss currently incurred by the water supply authority (KWASA), due to the supply shortage of improved water.

Firstly, the monthly and annual CS value in the presence of improved water supply has been extracted based on two types of estimations, i.e., the current level of demand and the potential demand, based on the current price of water supply and the WTP disclosed by the households. The former process states that the authority can easily price the monthly CS (US$ 0.04 million) into revenue by supplying the existing level of water by ensuring improved quality. On the other hand, the potential demand analysis suggests that the authority can hold the capability of supplying 3,000 megaliters of water which induces US$ 0.21 million of revenue monthly, producing a voluminous revenue fund of US$ 4.26 million annually. It indicates that the annual CS will amount to US$ 2.55 million. Most importantly, the authority is forgoing a bulk amount of monthly revenue worth US$ 0.18 million and the annual revenue worth US$ 2.14 million due to the absence of an improved water supply service.

The principal goal of the study was to identify the demand for improved water supply service and its determinants by applying the CVM approach. The key findings provide some striking insights into the importance of improved water supply service to attain SDGs and the role of authority to utilize the pricing opportunity. The study reveals that 52% of the surveyed households were willing to pay for improved water service. Previous studies such as Eridadi et al. (2021) found 66%, Akeju et al. (2018) reported 74.9% and Kim et al. (2021) observed 78.3% of the sampled households were expressive to pay for such improved water service, revealing an apparent discrepancy from the current study outcome owing to its sample size limitation.

Several key positive predictors of WTP for improved water supply service are observed. Years of schooling are one of them that influence the WTP decision, also documented by previous studies (Kanayo et al. 2013; Akeju et al. 2018). Similarly, household income is positively associated with the WTP decision since the higher-earning increases the demand for quality goods, and this is not the exception in the case of improved water supply (Rodríguez-Tapia et al. 2017; Akeju et al. 2018; Dey et al. 2019).

Results also suggest that the households requiring higher time to collect water are expected to express the WTP positively. The possible reason behind such an expression is to save time and associated labor for collecting water from the points. One of the crucial factors affecting the time duration is the distance (Dey et al. 2019). Previous study findings by Tussupova et al. (2015) and Dey et al. (2019) corroborate the current study result. On the contrary, a negative association between the WTP and water point distance was found by Wright (2012).

Among the negative determinants, households with a private source of water (20%) are less willing to pay compared to households having public or other supply connections. It is expected that the privately managed water supply ensures good quality, and the affluent households managing such a source will reduce their demand and bid amount for the current supply of water by the city authority. The finding is strengthened by the similar results revealed by Arouna & Dabbert (2012) and Whittington et al. (2002); while the result is contrasted to that of Tussupova et al. (2015). Degradation or dissatisfaction regarding water quality is found to be negatively related to the WTP and bid amount in the present study, which is supported by Wang et al. (2018) and Akeju et al. (2018).

Regarding the WTP and aggregate value in the presence of improved water supply, the study found an annual WTP of US$ 5.05 which predicts the total revenue for the water supply authority worth US$ 4.26 million. The results also indicate the necessity and demand for improved water in the city household. More impressively, the potential revenue is around 2.5 times the currently earned revenue if the improved water supply is possible to supply. Besides, due to the absence of an improved water supply service, the authority is forgoing approximately US$ 2.14 million in revenue. Many experimental studies determined WTP in different locations in varied contexts (Tussupova et al. 2015). For instance, Rodríguez-Tapia et al. (2017) identified the mean WTP for improved water worth US$ 34.5 million in Mexico City.

Water is regarded as the most valuable natural resource and there is a current focus is improved water quality as a prerequisite for improving wellbeing and sustainable development. A growing scarcity, pollution, and waste of water, therefore, put additional importance on securing improved water and sanitation for all as articulated in the SDG. The present study investigated the demand for improved water supply by applying the CVM approach to portray the economic potential of improved water supply service in one of the major coastal cities in Bangladesh. In addition, by extracting the CS and revenue loss, the study added value to the understanding of the need for improved water supply and pricing options in urban area. The key findings suggested that 52% of the surveyed households agreed to pay a monthly amount of US$ 5.05 on average, producing a yearly CS of US$ 0.41 million, based on current supply information. Based on the potential demand estimation, on the other hand, the city authority can gain a huge amount of US$ 4.26 million annually, leaving the CS worth US$ 2.55 million. It indicates that the water supply authority is incurring a revenue loss equivalent to US$ 2.14 million per annum.

The key predictors of WTP were determined as monthly household income, years of schooling, and extended water collection duration. Contrarily, the negative determinants were the bid amount, dissatisfaction with water quality, household size, gender, and private water supply service source. The study, relying on the empirical results, argues that the coastal city should upgrade the water quality to capture the increased number of household water supply connections and thus capture the CS. As a result, it will increase competition in the private market of water supply. The authority can still gain efficiency and generate profit by supplying improved water since the potential revenue based on WTP is 2.5 times higher than the currently earned revenue. It will maximize the benefits of both parties. The authority should think about the welfare of the urban residents by ensuring pure drinking and sanitation water. However, the primary restrictions of connecting houses with piped water are the financial backwardness of the consumers and structural concerns of the supply authority. The authority should think about cost-efficient technology to reduce the cost burden of poor households to gain greater connections. Besides, the structural problem in the path of supplying improved water should be removed. To attain SDG goal six, there is no alternative way but to fulfill the demand for improved water supply in Bangladesh's coastal and urban zone. The results from this study can be used in cost-benefit analysis to determine the economic viability of investing in the improved water supply system.

Although the study has determined the demand for improved water supply services in a coastal city area, it is not free from limitations. The foremost pitfall is the small sample size which is unlikely to cover maximum variation in explaining the demand for improved water services. However, the current study is an initial assessment that encourages further research with large or longitudinal observations to theorize the current study findings.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.