Sanitation Infrastructure

Health Benefits

Improved sanitation infrastructure is widely presumed to deliver substantial welfare gains. The primary expected benefit stems from reducing exposure to pathogens and environmental contamination, thereby improving health outcomes. In this section, we discuss the extent to which these expected benefits are found in the literature.

Large-Scale Sanitation Investments and Mortality

Historically, investments in improved water and sanitation infrastructure have been central to some of the most significant public health gains in modern history. In late 19th century Paris, the expansion of sewer networks is estimated to have added several years to life expectancy by reducing exposure to waterborne pathogens (Kesztenbaum and Rosenthal 2017). Similarly, in the United States, the roll-out of municipal water filtration and sewerage systems in the late 1800s and early 1900s has been described as a “watershed” moment for child health (Alsan and Goldin 2019), with modern infrastructure (the combination of piped water and sewerage) accounting for a large share of the early decline in infant mortality in Boston.

Evidence on more recent sanitation infrastructure investments further underscores these benefits. Watson (2006) evaluates a major federal programme launched in the 1960s to fund sanitation infrastructure on Native American reservations. The study finds that the resulting improvements led to sharp reductions in waterborne gastrointestinal and respiratory diseases among Native American infants. Remarkably, these gains were not only cost-effective but large enough to explain nearly 40% of the convergence in infant mortality rates between Native American and white populations during that period. In Salvador, Brazil, a citywide sewerage expansion in 1997, combined with an education campaign, significantly reduced diarrhoea prevalence by 22% overall and by 43% in high-risk areas (Barreto et al. 2007). The welfare impact of increased urban wastewater treatment in India was found to be even larger by Lepault (2023), who uses granular data on the placement of sewage treatment plants to show that from 2010 to 2020 increased treatment decreased faecal contamination in water by 53% and decreased downstream mortality of children under 6 months by 20%. Lipscomb et al. (2025) show that the addition of water treatment centres in Ecuador reduced hospitalisation from vector-borne diseases, particularly during times of high rain when vector-borne diseases would otherwise have been particularly common. The effects were largest in the most population-dense cantons where the diseases are likely to spread most easily.

In Brazil, large-scale investments at both the utility and municipal levels have proven highly effective in reducing child mortality. Nationally, participatory budgeting reforms that nearly doubled local sanitation investment were associated with a reduction of up to 7% in mortality among children under one year of age (Goncalves 2014), highlighting the importance that households place on sanitation investment by the government. Complementing this evidence, Kresch (2020) uses a natural experiment to demonstrate the importance of institutional clarity: a legal reform that clarified responsibilities within the sanitation sector led to a doubling of system investment, which in turn significantly increased access and further reduced child mortality.

These historical and quasi-experimental studies provide compelling evidence that large-scale sanitation infrastructure has been a powerful driver of public health improvement. Yet despite this well-established potential, the lack of safe sanitation infrastructure remains a pressing global challenge. As of the early 2000s, between 48% and 80% of the world’s wastewater was still being discharged untreated into the environment (Jones et al. 2021, United Nations 2017). By 2022, an estimated 1.6 billion people—primarily in LMICs—still lacked access to safe sanitation or relied on unsafe facilities (WHO UNICEF Joint Monitoring Programme for Water Supply and Hygiene 2022a).

In the decision between networked and isolated sanitation systems, it is important to note that the costs of networked sanitation systems are primarily very large up front investments while isolated sanitation systems require much smaller upfront investments but larger continued maintenance expenses.

Household Sanitation Investments and Health

The challenge of universal access to sanitation gained greater international attention with the launch of the Millennium Development Goals (MDGs) over two decades ago. Under the MDGs, the primary focus was on expanding access to improved sanitation and safe drinking water. This framing often encouraged interventions aimed at individual or household adoption of basic sanitation facilities (such as latrines) and localised water supply solutions. These efforts frequently relied on cost-sharing arrangements, community engagement strategies, or subsidies (Zwane and Kremer 2007, Peletz et al. 2017).[1]1. As noted by Duflo et al. (2015), large-scale infrastructure investment—particularly in rural areas—has often been dismissed as prohibitively expensive.

Economic research on household sanitation has considerably expanded recently: at least 26 ran- domised controlled trials (RCTs) have been conducted in 14 countries since the introduction of the MDGs, exploring a range of constraints to household sanitation investments—on both the demand and supply sides. The evidence from these RCTs (listed in Table 1) on welfare—particularly health—stands in contrast to the demonstrable, large-scale impacts of historical infrastructure investments. Findings are mixed: eleven studies (Panel A) report positive, though often modest, health effects, while another seven (Panel B) find no significant health improvements, even in cases where sanitation coverage and usage increased substantially. Eleven studies report only intermediate outcomes such as latrine adoption and/or willingness to pay.

Variation in impacts is not only a matter of whether effects are detected, but also exists within specific health outcomes—typically morbidity (such as diarrhoea or roundworm infections) and child growth.[2]2. Morbidity refers to infections and symptoms of illness, whereas child physical growth is a longer-term outcome that can be affected by repeated or chronic morbidity. In the context of sanitation, studies often measure diarrhoea as an outcome that is more immediately responsive to improvements in sanitation infrastructure than physical growth. However, diarrhoea is both notoriously difficult to measure—especially in large-scale household surveys (Arnold et al. 2013)—and not a symptom of environmental enteropathy, a subclinical condition increasingly recognised as a key pathway to growth faltering. As a result, relying solely on diarrhoea as an outcome may underestimate the full health impact of poor sanitation.  For instance, four RCTs—conducted in Mali (Pickering et al. 2015), India (Dickinson et al. 2015, Hammer and Spears 2016), and Laos (Cameron et al. 2021b)—report improvements in child growth indicators but no significant reductions in diarrhoea prevalence. In contrast, other studies identify impacts on morbidity rather than growth: Luby et al. (2018) document reductions in diarrhoea incidence in Bangladesh, while Cameron et al. (2019) find decreases in roundworm infections in Indonesia, though neither study observes corresponding improvements in child growth.[3]3. An additional RCT in Kenya (Gimaiyo et al. 2019) found reduced caregiver-reported incidence of child diarrhea. However, the child-focused sanitation and nutrition programme evaluated in this study did not lead to increases in latrine use or improved household sanitation infrastructure. This suggests the observed health benefits may have been driven by behaviour changes such as improved disposal of child feces and more consistent hand-washing practices.

Table 1: Average impacts of Sanitation RCTs

See Page 10 of PDF.

The absence of detectable health impacts in some RCTs is not due to a lack of scale or power. At least four of the studies listed were unprecedented in both the intensity of their interventions and the breadth of their evaluations: the study in Odisha, India (Boisson et al. 2014), the WASH Benefits trials in Kenya (Null et al. 2018) and Bangladesh (Luby et al. 2018), and the SHINE (Sanitation, Hygiene, Infant Nutrition Efficacy) trial in Zimbabwe (Humphrey 2019). Conducted between 2010 and 2013 in India, between 2012 and 2015 in Kenya and Bangladesh, and from 2012 to 2017 in Zimbabwe, these trials delivered comprehensive packages—including latrine construction, water treatment, and hygiene promotion—to thousands of households. Except for the Odisha trial, each featured multiple experimental arms that tested the individual and combined effects of water, sanitation and hygiene (WASH) components. Despite achieving substantial improvements in access to sanitation and hygiene infrastructure, the findings were sobering: sanitation interventions alone did not improve child growth in any setting, and reductions in diarrhoea prevalence were observed only in Bangladesh. In India, Kenya, and Zimbabwe, even the most comprehensive WASH packages failed to produce significant effects on either diarrhoea incidence or height-for-age z-scores.

Several other RCTs have also reported no measurable impact on health outcomes. These include an evaluation of community health clubs promoting WASH practices in Rwanda (Sinharoy et al. 2017), a large-scale, government-led handwashing and hygiene promotion campaign in rural Tanzania (Briceño et al. 2017), and a rigorous assessment of India’s Total Sanitation Campaign (TSC)—a large-scale rural latrine construction and an initiative to change behaviour which was the predecessor to the current Swachh Bharat Mission (SBM)—in Madhya Pradesh (Patil et al. 2014). A fifth RCT, conducted in the slums of two Indian cities, evaluated interventions aimed at improving the quality of community toilet facilities. While it found no statistically significant impact on morbidity, the study documented increased curative health expenditures and a rise in open defecation (Armand et al. 2023).

The emergence of null results in several high-profile studies likely contributed to a shift in focus within the RCT literature on sanitation, with many subsequent trials opting not to collect or report health outcomes. Instead, these studies concentrated on understanding sanitation-related investments and behaviour change. In total, 12 of the identified RCTs either did not collect or did not report health-related endpoints (Panel C in Table 2.1.2).[4]4. For one RCT in India, there are two papers with one-year follow-up results from the main evaluation. One paper focuses on uptake of sanitation infrastructure (Pattanayak et al, 2009), the other reports on health outcomes (Dickinson et al. 2015). The authors subsequently collected a four-year follow-up survey (Orgill-Meyer et al. 2019) and a 14-year follow-up survey that do not report health impacts (Pakhtigian et al. 2022).  In many cases, the absence of health impacts reflected the limited impact of these interventions on sanitation access itself—i.e. they failed to generate substantial increases in the use of improved sanitation facilities, leading to little change in behaviour and no resulting improvement in health outcomes.

Even where access does improve, sustained health benefits depend on continued use and proper maintenance of sanitation infrastructure—particularly in dense urban areas, where latrines can fill rapidly. Motohashi (2024) highlights the potential unintended consequences of inadequate follow-up, showing that each additional latrine built during a major WASH campaign in India was associated with a 3% increase in faecal contamination in local rivers. In contrast, Deutschmann et al. (2024) demonstrate that targeted subsidies for latrine desludging in Senegal not only improved waste management but also led to reductions in diarrhoea rates among nearby households, underscoring the importance of addressing the full sanitation chain.

Why Do Health Impacts Vary? Coverage, Context, and Complementary Factors

The large number of RCTs showing no or lower than expected health effects has sparked intense debate in the sector and led to various commentaries from academics and practitioners; see, for example, Pickering et al. (2019), Coffey and Spears (2018), Cumming and Curtis (2018), WaterAid (2018). The reasons put forward for these results are myriad, but can be broadly categorised into three: (1) inadequate sanitation coverage (or more broadly, continued exposure to faecal contamination from both humans and animals); (2) lack of complementary inputs; and (3) measurement (including the time horizon chosen by the studies).

Continued Exposure to Faecal Contamination due to Inadequate Coverage

Probably the most prominent and tested theory relates to continued exposure to faecal matter due to inadequate coverage. The premise is that because sanitation has strong externalities, partial uptake may not be enough to observe (community-level) health improvements. If only some households in a village build toilets but others continue open defecation, faecal contamination persists in the environment, affecting those with and without sanitation infrastructure.

The literature assessing the empirical relationship between sanitation coverage and health, which is primarily based on non-experimental methods[5]5. Methodological approaches include meta-analysis pooling different studies together (Cameron et al. 2022; Wolf et al. 2019). —looking within primary survey data/RCTs (Hammer and Spears 2016, Harris et al. 2017, Augsburg and Rodríguez-Lesmes 2018, Orgill-Meyer and Pattanayak 2020, Cameron et al. 2021b) and using publicly available household surveys (Buttenheim 2008, Fuller et al. 2016, Vyas et al. 2016, Andrés et al. 2017, Hathi et al. 2017, Geruso and Spears 2018, Cameron et al. 2021a)—draws two main conclusions. First, increases in community sanitation coverage lead to demonstrable health effects, even in studies that found no health effects on average. Second, the relationship appears to be non-linear.

Linear estimates suggest that transitioning from universal open defecation to complete sanitation coverage can substantially improve health outcomes. For example, it has been associated with a 47% reduction in diarrhoea prevalence in India (Andrés et al. 2017), a 0.43 standard deviation (SD) increase in child height-for-age z-scores across four countries (Cameron et al. 2022), and a decline in stunting rates by 10 percentage points in Indonesia (Cameron et al. 2021a) and 67% lower prevalence of stunting in Ecuador (Fuller et al. 2016). Similarly, studies estimate that each 10 percentage point increase in sanitation coverage leads to improvements in height-for-age z- scores of 0.031 SD in Mali (Harris et al. 2017), 0.03–0.05 SD in Cambodia (Vyas et al. 2016), 0.03 SD in India (Hammer and Spears 2016), and 0.17 SD in (semi-)urban India (Augsburg and Rodríguez-Lesmes 2018). In addition, such improvements have been linked to a 3 percentage point reduction in stunting in Laos (Cameron et al. 2021b) and a decline in infant mortality of six deaths per 1,000 live births—about 8% of the national average—in India (Geruso and Spears 2018).

Non-linearity in this relationship has been shown by Wolf et al. (2019) for diarrhoea, demonstrating that diarrhoea reductions were highest at lower faecal contamination levels, and no diarrhoea reduction was found when contamination increased above a certain level. Similarly, Cameron et al. (2022) show this non-linear relationship for child height-for-age and stunting. They estimate that gains occur once village sanitation coverage reaches 50–75%. There do not appear to be additional statistically significant gains beyond this threshold.

Given that most evaluations of sanitation interventions to add latrines find increased coverage of 10 percentage points or less, this modest improvement could explain why the health impacts of latrines differ substantially from those observed in historical infrastructure projects. Such historical projects typically achieved comprehensive community-wide sanitation coverage, reaching the critical threshold required for sharply reducing disease transmission.

Lack of Complementary Inputs

The second commonly cited explanation for the mixed—or often absent—health impacts of sanitation interventions is that improvements in sanitation infrastructure alone may be insufficient to generate health gains. In the absence of complementary investments, such as access to clean water, effective hygiene practices (e.g. handwashing), and safe infant feeding, the full potential of sanitation to reduce disease exposure and improve child health may not be realised.

For example, the medical and epidemiological literature has established that unhygienic conditions are not only a cause of diarrhoea, but also cause a condition known as environmental enteric dysfunction, which reduces the ability to absorb nutrients (Humphrey 2019, Dewey and Mayers 2011, Mbuya and Humphrey 2016) thereby reducing the effectiveness of even the highest quality nutrition in promoting child growth (Prendergast et al. 2014). While results of the WASH Benefits and SHINE studies raise the question of whether there is an additional health benefit from co-programming, Abramovsky et al. (2024) find that hygienic environments and nutrition together produce better child health outcomes in the context of the Philippines using a non-experimental approach.

Similarly, Duflo et al. (2015) evaluate a package consisting of a community water tank, piped water to the home, and household latrines and bathing facilities, and they show that the intervention reduces severe cases of diarrhoea by about 30% to 50%, in a context where sanitation-focused RCTs have failed to detect health effects. They argue that while they cannot say anything about the effectiveness of the individual components of the intervention, the whole package (e.g. providing universal access within a village to hygienic latrines and piped water in homes at the same time) has an impact—essentially in line with the large infrastructure investments discussed at the beginning of this review. Further evidence on the complementarity is provided by Weaver et al. (2024), who demonstrate that cash transfers lead to large increases in calorie consumption and nutrition, but these improvements only translate into health improvements for children in areas where open defecation rates are low.

Understanding which interventions are most likely to have complementarity impacts with sanitation is important to the design of sanitation policy, and more research in this area would be useful.

Measurement

A third commonly cited explanation relates to measurement challenges. Some health benefits of sanitation may take longer to materialise or may be difficult to detect given the sample sizes of typical RCTs. For instance, reductions in morbidity outcomes have been observed in some settings even when no effects on child growth were detected. In Indonesia, a CLTS intervention led to reductions in roundworm infections (Cameron et al. 2019), while in Nepal, Coffey et al. (2018) find that children exposed to improved community sanitation had higher haemoglobin levels. However, neither study found significant impacts on child growth. These outcomes represent meaningful welfare gains, though they may not immediately translate into observable changes in height. In addition, symptoms such as diarrhoea tend to be very difficult to measure, especially in large-scale household surveys (Arnold et al. 2013). There may also be substantial Hawthorne effects, particularly when dealing with hygiene behaviours. For example, Zwane et al. (2011) documents that diarrhoea goes down and water treatment increases among households more frequently surveyed (with no intervention) about their water use and diarrhoea rates.

In short, the general consensus that has emerged is that null findings should not be interpreted as sanitation investments having no value, but rather that context and implementation determine the extent of the health payoff.

Other Welfare Benefits

Although most sanitation investment decisions are motivated by potential health benefits, there is growing consensus that the impacts extend far beyond health alone—and economic research is beginning to provide rigorous evidence on some of these broader effects.

Exposure to improved sanitation in early childhood is hypothesised to support downstream gains in critical developmental domains, such as language acquisition, cognitive ability, and executive function, ultimately enhancing educational outcomes. In support of this, Spears and Lamba (2016) find that six-year-old children who lived in villages exposed to India’s TSC during their first year of life were more likely to recognise letters and simple numbers. Similarly, Orgill Meyer and Pattanayak (2020) show that children in villages with higher latrine coverage scored significantly better on a test of analytical cognitive ability ten years later.[6]6. An earlier review (Sclar et al. 2017) found some support for a positive relationship between sanitation and cognitive development but judged the existing evidence up to that date to be limited in credibility.

Improved health is also theorised to enhance productivity, although direct empirical evidence remains limited. For instance, Czura et al. (2024) find that providing menstrual hygiene products to female garment workers in Bangladesh did not result in measurable productivity gains. Similarly, other WASH interventions have yet to consistently demonstrate labour market effects. However, emerging evidence from China’s Toilet Revolution suggests alternative channels: Wang and Shen (2022) report increases in female labour supply—not as a result of improved health, but due to time reallocation. This highlights the potential for sanitation investments to affect productivity through non-health pathways such as time savings. Access to in-home toilets may also reduce the likelihood of sexual assaults:  Hossain et al. (2022) shows that a 10 percentage point increase in household toilets leads to a 2.5 percentage point reduction in sexual assaults of women (there is also a negative but not statistically significant impact on the incidence of rapes). Similar effects have been documented in India by Dickinson et al. (2015), who also find improvements in household satisfaction—an often-cited, though infrequently quantified, benefit of sanitation, particularly for women and girls.

Other studies have identified welfare gains from sanitation that are likewise independent of health improvements. McCulloch et al. (2025) investigate the effects of sewer access on neighbourhood characteristics, showing that they increase population density, which they argue makes sewer networks as important for the economic geography of cities as transportation networks. Augsburg et al. (2023a), show that India’s TSC, as implemented in the late 1990s, made marriage more attractive for both men and women and altered intra-household bargaining dynamics, leading to a redistribution of gains within marriage. Finally, Kresch et al. (2023) find that access to sewerage services increases households’ favourability ratings of the government and raises their willingness to pay property taxes—suggesting that governments may partially recover the cost of sanitation investments through increased revenue, as improved services strengthen state–citizen relations.

Given the significant potential gains from sanitation investments—many of which appear to materialise only when adoption reaches sufficiently high levels—the remainder of this review turns to the evidence on the constraints that limit increased investment in sanitation. Understanding these barriers is essential for designing more effective interventions that can achieve both higher uptake and broader welfare impacts.

For full reference list see the end of the conclusion chapter.