Ownership in times of water scarcity: are inhabitants supplied by private utilities consuming more water?

Drinking water availability is an issue becoming more frequent throughout the world, as many countries suffer from scarcity, deterioration, and contamination of the quality of water bodies due to unsustainable management, as well as weather conditions and increasingly uncertain climate (Olmstead & Stavins 2009; Wang et al. 2019; Ferasso et al. 2021; Pinto et al. 2021). Numerous cities globally struggle with the challenge of effectively managing water resources amidst rising populations, heightened consumer demand for water-intensive services, and escalating costs associated with developing new water supplies (Olmstead & Stavins 2009). Brazilian cities are no exception to this trend, since, despite possessing abundant freshwater resources, the country has experienced recent episodes of water shortages (Embasa 2020; Sabesp 2021; Saneago 2021; Sanepar 2023; Sousa & Fouto 2024a). Consequently, several key reservoirs in Brazil have reached critically low capacity, placing many cities on the brink of water rationing (Getirana et al. 2021).

In 2020, the Brazilian Congress enacted a bill that significantly altered the fundamental characteristics of the water and sanitation (W&S) sector. One outcome of this legislation is the increased privatization of water utilities across the country, a movement that is typically considered in three primary scenarios: (i) municipalities with low coverage, where financial resources are needed for the expansion of the infrastructure; (ii) operations with low efficiency, where there should be investments in more efficient equipment; and (iii) places where there is a need to trigger and sustain systematic performance improvements (Beecher & Kalmbach 2013; Homsy & Warner 2020; Tourino et al. 2022).

Over the past decade, extensive research has examined the role of ownership structure in water service provision (Carvalho et al. 2015; Pazzi et al. 2016; Wait & Petrie 2017; Hellwig & Polk 2021). However, there has been limited focus on how different ownership models affect water consumption. Considering water scarcity issues and the current push toward privatization in Brazil, it is crucial to investigate whether the increasing presence of private entities in the sanitation sector influences residential water consumption patterns. This paper aims to explore the following question: do residents served by private water utilities in Brazil use more water compared to those served by public utilities?

The article is organized as follows. In Section 2, we review the economics of water consumption, as well as ownership of water provision. In Section 3, we describe the methodology used in our quantitative analysis. In Section 4, we analyze the findings of the quantitative analysis and discuss the results. The last section draws conclusions, and implications and suggests future lines of research.

A critical area of research lies in understanding the drivers of water consumption, which must be thoroughly considered in the formulation of water-related policies and managerial actions. The challenge of assessing water consumption behavior relies on the fact that different factors contribute to shaping water consumption (da Veiga et al. 2022; Gonzáles et al. 2024; Sousa & Fouto 2024b), for example, psychological factors, available technologies, awareness, socioeconomic features, water availability context, and applied policies. In times of water availability uncertainties, the understanding of these factors and how they affect consumption behaviors becomes an important strategy.

Considering this scenario, demand-side management (DSM) has emerged as a crucial component of water management (Russell & Fielding 2010). DSM represents a managerial approach dedicated to meeting the demand for water by implementing measures and incentives to ensure fair and effective water use (Al-Zahrani et al. 2013). There are different types of DSM tools and techniques currently being implemented by urban water authorities and operators, including non-price-based policies and water pricing-based policies (Biswas et al. 2023).

Non-price-based strategies focus on methods other than economic incentives to manage water demand, such as regulatory measures, educational programs, and other approaches that do not involve direct financial costs to consumers. For example, the deployment of technological advances and the introduction of new home appliances play a significant role in reducing water consumption (Echeverría 2020). Equipment designed to conserve household water consumption includes dual-flush toilets, flow control devices, and water-efficient dishwashers or washing machines (Martínez-Espiñera & García-Valiñas 2013; Ramsey et al. 2017; Roshan & Kumar 2020). Furthermore, the effectiveness of these water-saving technologies is influenced by the level of information provided to consumers (Sousa & Fouto 2024b). Therefore, it is essential to combine technological innovations with educational campaigns to effectively promote water conservation.

Managers may also choose pricing-based strategies as effective water conservation policy tools to reduce consumption, particularly during periods of scarcity (Olmstead et al. 2005; Homsy & Werner 2020; Biswas et al. 2023). This choice is grounded in the understanding that economic factors significantly shape individuals' perceptions of the environmental impacts of water overconsumption (Lucas & Cordery 2019; Stone & Johnson 2022). Notably, the influence of pricing on water consumption is extensively explored in the literature, as evidenced by studies conducted by Dalhuisen et al. (2003), Stavins (2005), Olmstead et al. (2007), Vallès-Casas et al. (2017), and Sousa et al. (2022).

Typically, water prices cover the costs associated with acquiring, treating, and delivering water (Zhang et al. 2022), and, due to the natural monopoly characteristics of water supply systems, water pricing is generally not determined by market forces. As a result, prices often fail to accurately reflect the scarcity of this vital resource (Olmstead 2010). Furthermore, without the existence of a market, per se, water prices end up not automatically responding to changes in supply.

Due to historical developments, water and sanitation services (WSS's) operators in Brazil can be broadly categorized into the following: (i) local public entities (directly administered by municipalities); (ii) local or regional private entities; and (iii) regional public entities (Seroa da Motta & Moreira 2006; Tourino et al. 2022). In 2022, more than 90% of the 5,570 Brazilian municipalities were served by public utilities (state or municipal), covering 81.6% of the Brazilian population. Private companies played a role in supplying water to nearly 450 municipalities, providing services to over 39 million people (18.4% of the total Brazilian population). Despite their relatively modest market share, private companies have gained prominence due to the limited investment capacity of the state (Tourino et al. 2022).

In 2020, a new regulatory framework for the W&S market was introduced in Brazil (Lei n. 14.026; Brasil 2020), ushering in significant changes to the provision of WSS in the country. Some of the most impactful changes include setting mandatory goals for universal access to accessible ‘WSS’, encouraging regionalized provision of these services, and centralizing regulation by assigning the role of general guideline development to the national water regulator.

With regard to utilities ownership, this new regulatory framework actively promotes privatization for several reasons. First, it provides greater legal certainty for the privatization of state-owned companies. Second, it aims to strengthen competition in the market by prohibiting the formalization of contracts between municipalities and state companies without a bidding process. Lastly, it mandates the organization of bidding processes when conceding municipal WSS.

The literature extensively deliberates on the significance of ownership structure in water service provision worldwide. While the analysis of ownership encompasses various dimensions, including effectiveness, equitability, regulation and sustainability, the predominant research focus tends to center around pricing (Barbosa & Brusca 2015; Wait & Petrie 2017; Hellwig & Polk 2021; Zhang et al. 2022) and efficiency (Braadbaart 2002; Renzetti & Dupont 2004; Seroa da Motta & Moreira 2006; Pazzi et al. 2016; Homsy & Warner 2020; Pereira & Marques 2022). Nevertheless, despite the abundance of studies, a considerable number of studies fail to yield definitive conclusions regarding the superiority of water utilities from an ownership perspective (Carvalho et al. 2015).

When it comes to pricing, most studies focus on examining the relationship between privatization and price increases. However, there is no consensus in the literature, as similar research in different countries has produced mixed results – privatization has sometimes led to higher costs and, in other cases, lower costs (Wait & Petrie 2017). Understanding the influence of price changes is crucial, particularly because of the impact on affordability, a pressing issue for low-income households (Zhang et al. 2022) and in developing countries such as Brazil.

Regarding efficiency, the findings in the literature are inconsistent. In some cases, studies suggest that ownership does not significantly impact productivity gains, while in others, evidence indicates that private operators have advanced more quickly than public ones toward greater efficiency (Seroa da Motta & Moreira 2006). The realization that privatization does not always lead to cost savings has fueled interest in the remunicipalization of water services in certain regions (Homsy & Warner 2020). However, some research highlights that public companies may be operating below the optimal scale, as demonstrated by Pereira & Marques (2022) in the Brazilian context.

Still regarding efficiency, much of the research over the past decade has focused on non-revenue water (NRW) (Adams & Lutz Ley 2012; González-Gómez et al. 2012; Carvalho et al. 2015; Maziotis et al. 2020; Tourino et al. 2022). The literature suggests that water utilities with private participation have greater incentives for efficiency than those without private involvement (Carvalho et al. 2015; Tourino et al. 2022) and, thus, have lower NRW indicators.

Ownership has not been extensively examined in recent literature as a determinant influencing behavior related to water consumption, though it may play a significant role. To the best of the authors' knowledge, only Kallis et al. (2010) focused on this aspect, providing insights from a survey conducted in California (US), concluding that privatizations and associated reforms can diminish the conservation potential by depleting users' willingness to cooperate in water-saving efforts.

To examine potential differences in water consumption behavior between residents supplied by private water utilities and those supplied by public companies, we employed a multilevel model framework. This approach was chosen due to the hierarchical characteristics of the data, and it offers a distinct advantage over traditional regression models by considering the natural nesting of data. Multilevel models enable researchers to analyze individual heterogeneities between groups, allowing for the specification of random effects at each level of analysis (Santos et al. 2016; Hair & Fávero 2019).

In our study, we proposed a hierarchical linear model with three levels. At the first level, we considered the year; the second level comprised the municipality itself, accounting for local variations; and the third level involved the federative unit in which the municipality is located, capturing broader regional influences on water consumption behavior.

For this study, we utilized data from ‘SNIS’, Brazil's official database for municipalities' WSS operational and financial information. It's crucial to note that while SNIS is the government's official source, the information is self-declared by municipalities' administrations and is not revised by the federal government or audited by a third party. Recognizing the potential for inconsistencies, we considered it important to meticulously evaluate and refine the collected data before proceeding with the model. The steps employed for the evaluation and refinement of the data are summarized in Table 1, and aimed at enhancing the reliability of the database.

After the treatment process, our database comprised information from 858 Brazilian municipalities, 807 operated by state-owned companies and 51 by private companies, situated in 27 federal units (states), covering the period from 2002 to 2019. Despite the different sizes of samples, the number of municipalities from each group is large enough for the analysis and thus considered adequate for the study (Wait & Petrie 2017). Also, the use of information from municipalities located in all 27 Brazilian states guarantees geographic diversity on a national scale.

As the dependent variable, we considered the annual average per capita water consumption. The utilization of per capita data was selected to normalize consumption across the population, acknowledging the diverse residential characteristics present in each municipality (Vallès-Casas et al. 2017).

The selected explanatory municipal variables, exhibiting variations over the years, encompassed the urban water service index and the NRW index. The urban water service index indicates the percentage of the urban population in each municipality with access to piped drinking water, while the NRW index accounts for the share of water that is extracted and treated but lost in the distribution process, either due to leaks or unbilled consumption resulting from uncorrected measurement or unauthorized use. The inclusion of the urban water service index is justified by Magnusson (2004), who emphasizes that the propensity to respond to demand management diminishes with improved water access.

Also, as a control municipal variable exhibiting variations across the years, we incorporated the municipal per capita gross domestic product (GDP) sourced from the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatística, in Portuguese) database. This variable served as a control, acting as a proxy for local income, given that income has been established as a significant factor influencing water consumption patterns (Sousa & Fouto 2019; Sousa et al. 2022).

As an explanatory variable at level 2, we considered ownership. This variable was represented as a dummy variable, categorizing entities into direct administration, municipal administration, state companies, and mixed capital companies with public administration as ‘state-owned’ (coded as 0), while private companies were designated as ‘private’ (coded as 1).

Table 2 sums up the variables used in this study.

Table 3 shows the descriptive statistics of the variables used in this study, and Table 4 shows the estimation outcomes of the complete model. The findings suggest that, on average, users served by private utilities present a tendency to consume less water compared to those supplied by state-owned utilities.

The result from our analysis, indicating that households supplied by private companies tend to consume less water on average, does not align with the findings of Kallis et al. (2010), who suggested that privatizations and related reforms might diminish water conservation potential. Considering that numerous studies have illustrated that privately owned water utilities often impose higher prices for water services (Martinez-Espiñeira et al. 2009; Beecher & Kalmbach 2013; Wait & Petrie 2017; Zhang et al. 2022), and that increasing prices is related to reducing water consumption (Gilbertson et al. 2011; Sebri 2014; Homsy & Werner 2020), this could be a possible explanation for our findings.

Unfortunately, data on average tariffs was unavailable for inclusion in the model. Moreover, Brazil's specific context – where there still exist by high levels of unbilled water – further complicated the use of this variable. A significant portion of the population in Brazil lacks the financial resources to pay for water usage, often resulting in fraud (Marques & Saraiva 2017; Sampaio & Sampaio 2020; Santi et al. 2021). Given these factors, including a variable related to tariffs, could have introduced bias into the model.

As previously mentioned, literature on water utilities ownership has placed significant emphasis on pricing. For instance, Zhang et al. (2022), in a study of the 500 largest community water systems in the US, concluded that privately owned systems have higher annual bills, an observation also consistent with findings by Beecher & Kalmbach (2013) in their study of the Great Lakes region. Similarly, when examining the Brazilian water market, Barbosa & Brusca (2015) found that privately owned utilities typically impose higher prices, even when subjected to local and regional regulatory agencies' pricing mechanisms. Moreover, there is evidence linking privatization to both price increases and consumption reduction: Vallès-Casas et al. (2017) demonstrated that the substantial rise in water prices in Spain, following the privatization of a regional supplier, resulted in significant decreases in consumption across various metrics, including total consumption, per capita consumption, and consumption per meter.

Contrary to the findings in the previously mentioned studies that link ownership to water prices, it is essential to consider the perspective presented by Marques & Simões (2020), who argue that private operators often impose higher tariffs due to the more stringent requirements they must adhere to, and as a result, higher prices are not necessarily indicative of excessive profits. Furthermore, García-Valiñas et al. (2013) emphasize that, in cases where water services are provided by state-owned companies, especially municipal ones, costs might not be accurately reported, and thus, they are not entirely reflected in the charged prices. While these factors are indeed relevant, it is crucial to acknowledge that what ultimately impacts users' perceptions is the price stated on their water bill. In other words, this final price is a significant determinant shaping users' behavior concerning water consumption and conservation efforts (Sousa & Fouto 2019).

Once the results showed that the population supplied by state-owned water utilities tends to consume, on average, more water, actions should be taken by this kind of utilities to overcome this issue, considering the current need for the global population to consume water rationally. Fostering water conservation can be achieved through public policies that may have diverse focuses. For instance, the literature underscores the significance of information and educational strategies in shaping water consumption (Jethoo & Poonia 2011; Lowe et al. 2014; Quesnel & Ajami 2017; Ramsey et al. 2017). However, while information and education are crucial elements, they are not standalone components sufficient for governmental programs aimed at behavior change. A combination of factors is deemed necessary to effectively promote water-saving behavior (Martínez-Espiñeira & García-Valiñas 2013; March et al. 2015; Sousa & Fouto 2024b; Van Loon et al. 2024).

Additionally, the results from our analysis reveal that the water service index, NRW index, and per capita GDP are all statistically significant in predicting per capita water consumption. Notably, the water service index and NRW index display a negative correlation with the dependent variable, while per capita GDP exhibits a positive correlation.

Regarding the water service index, in municipalities with higher service indexes, individuals tend to consume less water – a finding consistent with existing literature (Ito et al. 2021). Three plausible explanations can be considered for this result. First, a low service index often signifies that the municipality lacks resources to enhance sanitation services (Sousa et al. 2022), having also limited funds for awareness programs promoting water-saving actions. According to Ito et al. (2021), on the other hand, households connected to water systems may possess a heightened awareness of water conservation strategies, allowing them to cope effectively with water scarcity.

Second, ownership exhibits a positive correlation with service indexes, implying that privately owned water utilities typically boast higher service indexes. In our database, for the period from 2002 to 2019, the mean for water services provided by private entities is approximately 96.2%, whereas state-owned utilities have an average of 91.5% (SNIS 2022). This disparity can be attributed to the increased efficiency of water utilities with private involvement compared to those without private intervention (Carvalho et al. 2015; Pereira & Marques 2022; Tourino et al. 2022).

Third, in regions with lower service rates, the level of NRW (water lost during distribution, unauthorized consumption, and/or under-metered water due to metering inaccuracy) tends to be higher. This circumstance implies that information on per capita consumption may be overestimated. As highlighted by Ibrahim et al. (2021), in developing countries, a considerable part of the population does not have access to piped water, meaning that they commonly do not pay for the consumed water and, thus, may not respond to economic incentives, such as price increases. The former argument offers a plausible explanation as well for the observed negative and significant correlation between the NRW index and average per capita water consumption, since illegal residential connection to the network is still a common practice in some areas in Brazil (Marques & Saraiva 2017; Sampaio & Sampaio 2020; Santi et al. 2021).

Finally, with respect to GDP, used as a proxy for local income, the findings align with previous studies, confirming its significant role in influencing water consumption patterns (Sousa & Fouto 2019; Sousa et al. 2022). Notably, in Brazil's North and Northeast regions, which have lower per capita GDP, the indexes for water and sanitation services are also below national averages, while water loss rates are disproportionately high. For example, these regions report NRW rates of 46.2 and 51.2%, respectively (SNIS 2022). In some states within the North region, water losses surpass 70% of total water production (SNIS 2022), meaning more than two-thirds of the water produced is not billed.

A limitation of the present study pertains to the inability to discern whether the mere awareness of the population regarding water utility ownership influences water consumption behavior, independent of the price charged for water. To address this limitation, a qualitative survey would serve as a suitable instrument. Kallis et al. (2010) employed such an approach in their work, conducting a survey of water users in California. Their survey incorporated questions such as: (i) how users might respond if their provider mandated a reduction in water use; (ii) whether users were cognizant of the state/private nature of their water provider; and (iii) if users believed that their responses would be influenced by whether their providers were state or private companies.

Utilizing a multilevel model framework, our study demonstrates that, between 2002 and 2019, individuals served by private water utilities in Brazil, on average, consumed less water compared to those served by public water utilities. In the context of climate changes, water availability uncertainties, and particularly in regions facing water stress, the level of water consumption should be regarded as a crucial factor among other determinants when deciding whether to delegate water services. The findings from this research also underscore the necessity for state-owned water providers to reassess their policies in water demand management, aiming to reduce water consumption. This may involve implementing informational campaigns, promoting voluntary conservation behavior, and adopting economic incentives. Given that state-owned companies continue to be responsible for supplying water to nearly 80% of the Brazilian population, these actions become imperative for sustainable water management in the country.

Future research endeavors should focus on scrutinizing water consumption in Brazil in the forthcoming years, particularly as private participation is anticipated to intensify. A deeper analysis using surveys could enhance the findings of our study by allowing researchers to investigate whether the population's awareness of water utility ownership influences water consumption behavior, irrespective of the billed water price. Such surveys would provide valuable insights into the nuanced factors shaping individual attitudes and behaviors toward water consumption in the evolving landscape of water service management in Brazil.

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

The authors declare there is no conflict.