Exploring the Value of Clean and Safe Drinking Water to Northern Irish Adults

: Unlike many other countries, Northern Ireland residents do not pay separate water charges and there has been significant political and social controversy about their introduction. Therefore, this paper explores how Northern Ireland residents value access to clean, safe drinking water. Using the open-ended contingent valuation method, 205 adults were asked to provide their annual willingness-to-pay to maintain their current level of service provision, their willingness-to-accept to have it taken away, and their willingness-to-give to ensure similar service provision in a developing country. The results show that the value of clean, safe drinking water to Northern Irish adults is £120 per year, far less than both the real cost of provision, or the nominal amount paid through other combined charging instruments, demonstrating that said level of access to water is significantly undervalued by the Northern Irish people. Education levels and gender were found to be statistically significant predictors of willingness-to-pay and willingness-to-accept, respectively. Participants appeared to value their own access slightly more than that of people in developing countries, but not significantly more. The protest zeros observed throughout this study identified participants’ resistance to change. The results suggest introducing water charges in NI will continue to be a fraught process, with significant difficulty accepting rates set. Willingness-to-give


Introduction
This paper seeks to explore the value of access to clean, safe drinking water among residents of Northern Ireland (NI). A review of relevant media shows that, at present, NI residents contribute towards the cost of water through their domestic rates charges. However, this contribution is not enough to cover the total cost (£460 million in 2018/19) and the NI Executive must contribute almost £300 million annually to subsidize this [1]. Rates bills in NI have not been itemized since 1999 and residents are unaware of their actual contributions towards water costs [2] which may result in a lack of appreciation for how much water is worth. Despite repeated attempts to introduce water charges, the main political parties in NI are opposed to the introduction of separate domestic water charges similar to those paid by businesses in the country [3].
McKibbin [4] notes that free water leads to waste and inefficient use, so to achieve sustainable water consumption in NI, as set by the Water Framework Directive [5], NI Water must educate the public about effective water consumption in place of suitable incentives through water pricing charges [6].
Every developed country pays for their water provision through domestic water charges [7], with the exception of NI and the Republic of Ireland [8], and so there is little literature examining WTP for water provision in these countries. Research has instead focused on the economic value of water to those in developing countries where access can often be restricted [9][10][11]. People living in these countries are most likely to be willing to pay more than they currently do to improve their access to a safe and reliable water supply. Those that focused on developed countries investigated WTP for improving the quality of water bodies rather than domestic water [12][13][14].
To the best of the authors' knowledge, there have not been any similar studies to date investigating the economic value of water in NI to domestic household residents. In order to better understand and plan for the introduction of water charges, it is also important to examine the influence of certain socio-demographic characteristics.
People in developing countries have a higher willingnessto-pay (WTP) to improve their access to their water service, as well as improving the quality of the water provided. In many low/middle-income countries, a safe and reliable water supply is not guaranteed even if a basic water delivery service exists [15]. Residents are therefore often more willing to pay for private water provision or utilize other sources such as bottled water to guarantee their access [15].
Age, gender and education may also be relevant in understanding WTP for access to clean, safe drinking water. Studies have shown women are often more risk adverse than men [16], so would be likely to pay more to avoid the risks posed by losing this access. Studies have also indicated younger generations are often willing to pay more for a good or service than their older counterparts which can be attributed to the shorter planning horizons of older populations, i.e. the duration of time into the future that an individual has accounted for [17]. Having a higher level of education often increases WTP for a service/good [18] and may increase WTP for access to clean, safe, drinking water given that it may also enhance appreciation of the costs and effort required to achieve sustainable service provision [19].
Given this background, this paper aims to value how much Northern Irish residents would be willing-to-pay for access to their current standard of clean, safe drinking water, how much they would be willing-to-accept if it were taken away, and how much they would be willing-to-give to those in developing countries to provide a similar level of access. It also hopes to examine the relationship between sociodemographics and these values. This was achieved through an application of the Contingent Valuation Method (CVM) deployed through an online survey using a volunteer sample of Northern Irish adults. Northern Ireland provides a unique opportunity to determine true WTP as there is no pre-existing values to anchor to.
The paper has commenced with a literature review detailing the current situation of water safety, access and charging in NI contrasted with other countries, previous studies on the value of access to clean safe drinking water, and the methods that can be used to derive the economic value to individuals of abstract services such as water access. The use of CVM is presented, followed by the results of the CVM and possible justifications for these findings. The paper concludes with a discussion of the limitations of this study and future directions for such research.

Materials & Methods
The methodology can be described as an online contingent valuation questionnaire using a volunteer sample of Northern Irish adult residents.
The target population was a "Northern Ireland adult resident" defined by the study as an individual of either gender, aged over eighteen, and currently living in Northern Ireland. It was decided not to limit the study to respondents born or having lived in NI for an arbitrary amount of time, as all NI residents will have to pay water charges should they be introduced. Additionally, limiting residence time would introduce bias regarding awareness over the ongoing debate about water charges. The target sample was calculated at 91, with medium sized effect (ƒ 2 = 0.15), a statistical power level of 0.8 and a probability level of 0.05, using a multiple regression model with five a priori predictors.
Respondents were recruited primarily through convenience sampling. Links to the questionnaires were distributed through the researchers' social networks and through email distribution within the home university and partner organizations. Respondents were asked to share the questionnaire with their own social networks, also constituting snowball sampling [20].
Snowball and convenience sampling were used as this creates a large sample size [21] very quickly and at no cost. People do commonly share research participation opportunities with those who have similar characteristics to themselves [22], possibly increasing the risk of a homogenous sample [23]. While this may have limited the representativeness of the sample [20], this was countered through the online deployment, which increases both the participation rate and the diversity of participants [24,25], and initial respondents contacted were deliberately varied in their ages, genders, and education levels.
The continent valuation method (CVM) was chosen as water is a public good in Northern Ireland as the charges included in domestics rates do not reflect the full cost of providing the service or its true market value [9]. The CVM has been successfully used to estimate the value of commodities not exchanged in regular markets or when transactions cannot be observed under the desired conditions due to the hypothetical situations it creates [26].
The CVM requires people to state their WTP for a specified good or service and their willingness-to-accept (WTA) to be compensated give up the good or service. CVM typically employs questionnaires to create a realistic but hypothetical market for respondents to state their WTP for a good or service [9,27].
The CVM risks protest zero responses from respondents whereby they value the service/good yet state its economic value to them is zero [28]. Here, these protest zeros were anticipated given the contentious nature of the topic, but were not discounted, to avoid selection bias [29]. Respondents were allowed to remain anonymous to increase both participation rate and truthfulness in their responses [24].
Following ethical approval, data collection took place over a period of four weeks from January to February 2020. Potential respondents were emailed or had their attention drawn to a link to the online questionnaire. No incentives were offered for participation in the study. When the link was clicked, respondents were presented with an online information sheet and consent form. Once consent was signified, they were taken to the questionnaire proper.
For each question, context of the topic was provided to ensure the respondent would understand what they were being asked about, such as the meaning of clean, safe, drinking water. The term 'hypothetically speaking' was purposely included so to avoid reification fallacy, whereby people may believe an abstract concept is reality [30] and thus assume that a water charge was being imminently introduced. Participants were first asked to provide their WTP: Access to safely managed clean drinking water is defined as a source of drinking water located on the premises, available when needed and free of contamination [31]. In Northern Ireland the cost of this access is built into the domestic rates of each household, as set by the Northern Ireland Assembly and local councils.
Imagine a separate annual charge was introduced. Hypothetically speaking, how much would you be willing to pay per year to keep your current level of access to clean, safe drinking water? Please give your answer in pounds.
Respondents were then asked about their WTA: Water is abundant in Northern Ireland, so it is hard to imagine restricted access to this resource. Northern Irish residents have access to safely managed water resources, meaning they have access to clean, safe water wherever and whenever they want. But what would happen if this access was removed?
Hypothetically speaking, how much would you be willing to accept as annually paid compensation to have your current level of access to clean, safe drinking water removed? Please give your answer in pounds.
Lastly, respondents were then asked about their willingness-to-give (WTG), a variation of WTP related to donations: Access to clean drinking water is not even as it is often those living in poor rural regions who are the worst affected. In 2017, 785 million people lacked even a basic water service where a water source is located within 30 minutes of their house [31]. Now imagine, as part of the aid provided by the United Kingdom to developing countries, a compulsory annual fee or tax is introduced in which you pay for other people's access to clean, safe drinking water in these countries.
Hypothetically speaking, how much would you be willing to pay per year to provide similar access to clean, safe drinking water as your own in developing countries? Please give your answer in pounds.
Respondents then completed a short demographic questionnaire to obtain their gender (Male/Female/Other), their highest level of education (GCSES/A-Levels/Bachelor's Degree/Master's Degree/PhD or equivalent), their age in years, and if their first language was English or not. These questions were compulsory, but there was also an optional question to provide their postcode. This was intended to be mapped to the Northern Ireland Multiple Deprivation Measures to give an indication of socio-economic status. The questionnaire concluded with a debriefing sheet and a final opportunity to withdraw participation.

Results
There were 288 respondents in total. 83 respondents were removed, having withdrawn their consent or failed to complete the entirety of the required questions, leaving a realized sample size of 205 and a completion rate of 71.18%, which exceeded the minimum power requirement.
The make-up of the sample is described in Table 1. Gender was roughly evenly balanced between male (47.32%) and female (52.68%), with no respondents identifying as a gender other than these two categories. The majority spoke English as a first language (87.8%) and were educated to undergraduate level (35.12%), with an even spread of educational achievement. The mean age was 44. Very few respondents provided their postcode, so this variable was discarded from further analysis. A Tobit model was estimated to examine the relationship between the predictor variables and WTP (Model I, Table 3). Women, people who do not speak English as a first language and older people are willing to pay more to maintain their current level of access, though none of these predictors were statistically significant. Education was the sole statistically significant variable in predicting WTP (p < 0.01) and so a more parsimonious model was estimated with education alone (Model II, Table 3). In this model, as education increased, so did WTP. An individual with a PhD was willing to pay £183.50 more than someone educated to GCSE level. Despite the significance of the respondents' level of education, the low Pseudo R 2 values imply there are variables which were not tested which influence a respondent's WTP.  Table 4 displays how much respondents are willing to accept as compensation for the removal of their current access to clean safe drinking water. Protest zeros were most clearly evident within this question as the respondents reject the removal of the service and answer with £0 despite valuing the amenity/service [28,32]. The mean WTA was calculated to be £2,342.20. The majority of respondents indicated their WTA between £1-2000 per annum, with some respondents expecting much larger sums in compensation. A Tobit model was estimated for each of the variables possibly predicting WTA (Model I, Table 5). Older people, better educated people, and people who do not speak English as a first language are willing to pay more to maintain their current level of access, though none of these predictors were statistically significant. Gender was the only significant predicator of an individual's WTA (p<0.01), so the model was re-estimated with it alone (Model IV, Table 3). Here, women were willing to accept £1,037.84 less in compensation than men. (p < 0.001). Despite the significance of the respondents' gender, the low Pseudo R 2 values imply there are variables which were not tested which influence respondents' WTA.
WTA and WTP converged at £120, indicating that the value of access to clean, safe drinking water to Northern Irish adults is £120 per annum. Respondents' WTG per annum to those in developing countries to provide access similar to their own is displayed in Table 6. The mean WTG was £100.38. WTG was clustered around the lower end; as the amount being paid increased, less people were willing to pay it. A Tobit model was estimated for each of the variables that could possibly affect WTG (Model V, Table 3). Language was determined to be the only significant predictor (p<0.001) and so the model was re-estimated as a parsimonious model (Model VI, Table 3). Those who spoke English as their second language were willing to give £133.18 more than someone whose first language is English. Despite the significance of the respondents' first language, the low Pseudo R 2 values imply there are variables which were not tested which influence respondents' WTG.

Discussion
This paper investigates attitudes of adults in Northern Ireland to clean safe drinking water by exploring their valuation of their current provision using a contingent valuation model. This is the first study of its kind in a Northern Irish context and represents a significant contribution to the lack of literature published about WTP for clean safe water in developed countries. Northern Ireland is unique in its lack of domestic water charges, making it the ideal study area for this line of research.
The mean WTP for access to clean, safe drinking water was calculated to be £165.58 per annum. This WTP was significantly predicted by the individual's level of education (p<0.01). and the better educated a respondent was, the higher their WTP. This confirms the findings that of Kayaga et al., [19] that increased education also increases appreciation of the cost of sustainable service provision. Those who have higher educational backgrounds may be more aware of the steps involved in getting the water from source to tap and as a result, they are willing to pay more for their water. Alternatively, education may have acted as a proxy for income as higher levels of education are often linked with better economic outcomes [33] and WTP generally increases as income does [18]. The mean WTA as compensation for the removal of their access to clean, safe drinking water was estimated to be £2,342.20 per year. This question had significant number of protest zeros, when respondents state their economic value to be zero despite valuing the good/service in actuality [28], as evidenced by the fact that many justified their responses, despite it not being requested in the question. Many implied that they did not agree with removal of this service provision in any form and others stated that they "would not voluntarily give up their access". Given the ease of access to clean, safe drinking water in Northern Ireland, it may be ingrained in participants as a fundamental human right [34], or purely the responsibility of public authorities, regardless of the cost [35]. Some of these zero answers may not be protest zeros, but an inability to actually imagine a NI without this level of access. It was still important to retain these zero responses to avoid any form of selection bias [28,29].
WTA was significantly predicted by the gender of the respondent, with female participants willing to accept £1037.84 less per annum in compensation than men. These findings contrast with a study by Wieland et al. [36] which concluded that gender differences do not exist in risk aversion when discussing WTA; women will not accept less in compensation than men. The findings of this study may be attributed to their different reactions to their property rights potentially being infringed, or that the risks associated with water supply interference are valued differently. It may also be that male respondents were more assertive and reactive to the perceived infringement of their rights of access [37].
The mean WTG to provide a similar level of access was calculated to be £100.80 per annum. The only significant predictor of WTG was first language, with those whose first language was not English willing to give £133 more per annum than those whose native language is English. Language here may operate as a proxy for nationality or ethnicity, where participants may have lived in countries where service provision was poor and so had a better understanding of the needs of those in developing countries.
Altruism may have had a role here. Altruism is defined as the desire to benefit someone else for their sake rather than one's own [38] and often impacts WTP/WTG [39]. Given that it had no tangible benefit to them, the less altruistic participants may have chosen to give less or nothing in donation when compared to more altruistic participants [39]. Some respondents explicitly stated that they "would prefer to pay to maintain/upgrade our own water system here" whereas others provided information regarding their donation history e.g. "I already donate each month to WaterAid". The discrepancies indicate they value their own access far greater than that of others.
As this was a hypothetical exercise, many participants may have offered amounts that they would not have given in real life, constituting hypothetical bias [40]. Additionally, participants may have offered large amounts to benefit from "warm-glow" [41,42] or as a form of "virtue signaling" [43].
It would logically follow that if respondents were giving their true value, there would be no difference between WTP and WTA. This was the case for some participants (n = 47) but, largely, participants expected far more in compensation that they would be willing to pay for the same good. It is clear people value their access to clean safe water by the high compensation amount required for them to give up their access yet, conversely, they are unwilling to pay even a quarter of that in charges to ensure that access. This disparity between WTP and WTA is one of the most commonly observed phenomena in behavioral economics [44]. In the context of this study the WTP/WTA disparity could be due to the endowment effect [45] which suggests that individuals value goods more when they have ownership of them [46]. This is also known as loss aversion [47,48] whereby when required to give up a good/service, individuals require more in compensation than they would be willing to pay for it in the first place [49]. As water is a nonmarket good, the significant divergence between WTP and WTA is to be expected as there are no close substitutes for a safe and reliable water service [50].
In order to provide clean, safe, drinking water, NI Water conducts over 500 water quality tests a day, spending around £480 million annually, of which businesses contribute only £79.2 million. The householder contribution via rates bills is only £101.9 million, leaving the NI Executive to pay a subsidy of £299.9 million to cover the excess balance [1]. Based on the value of £120 found here, across the 487,850 households in NI, that would decrease the householder contribution to £58.5 million, further increasing the subsidy required. Introducing the suggested fees of around £400 per household per year, which are in line with the charges currently implemented in the UK, would greatly reduce the subsidy required and better reflect the true cost of this water provision.
The authors acknowledge the following limitations of the paper. Time and financial constraints associated with this study restricted the size and representativeness of the sample population. While the questionnaire yielded results of sufficient statistical power, it is possible that certain demographics were not represented. The mode of distribution may have hindered the collection of a representative sample. There were many eligible people who do not have access to or use the internet, typically older generations, and could not complete the survey [51] as was reflected in the demographics of the sample population (12.68% were aged 61 and above). Overall, online surveys typically have lower response rates than other modes of distribution; on average 11% lower [52], and CVM is usually expected to have between 300 -1000 participants. Additionally, it is evident that other explanatory variables were not collected.
While this study has provided effective exploratory work, the authors recommend future research with a larger, more representative sample, possibly sampled using probability methods, which aims to collect omitted variables such as deprivation measures and household income. CVM is particularly prone to the overstating or understating of WTP [53], so more incentive-compatible, demanding revealing methods such as experimental auctions could be used.

Conclusion
In conclusion, it is clear that people in NI are unaware of the true cost of providing their water and would be particularly resistant to the introduction of even minimal charges as evidenced by the protest zeros and high WTA values provided. Studies employing the CVM have traditionally focused on developing countries where service provision is restricted. NI is unique in its lack of domestic water charges, which has provided the study with a unique opportunity to study the previously unexplored WTP of the residents without pre-existing values to anchor to.
The value of £120 per year derived here would not be feasible in reducing the strain placed on the NI Executive in paying the annual subsidy to NI Water. Concurrently, previous attempts at introducing separate domestic water charges have been blocked by the NI Executive and residents [3]. Policy makers are thus faced with a lengthy battle to educate the people of NI and arrive at a rate of charge that would amenable to many different sectors of the population. In Northern Ireland, they greatly value their water, but not enough to pay for it.