Sanitation Infrastructure

In its broad definition as the separation of human excreta from human contact, sanitation is arguably a public good. An excreta-free living environment benefits entire communities, regardless of who contributes to it, and is therefore susceptible to under-provision when left to individual households. From an efficiency standpoint, the provision of such public goods is typically the responsibility of government. However, in many cash-strapped developing countries, limited fiscal capacity constrains investment in large-scale sanitation infrastructure.

There are also likely political economy reasons why government spending on sanitation remains inefficiently low. Kresch and Schneider (2020) show that in Brazilian municipalities where the mayor is politically aligned with the state governor, spending on sanitation is significantly lower compared with non-aligned municipalities—suggesting that political incentives distort allocation. Glaeser (2004) points out that public involvement in utilities such as sanitation can increase vulnerability to corruption. At the same time, households tend to value sanitation services, which may influence their willingness to comply with tax obligations (Kresch et al. 2023).

Government intervention in the sanitation sector has typically taken four main forms: (1) the provision of large-scale infrastructure, such as treatment plants and networked sanitation systems—primarily in wealthier, formal urban areas; (2) the procurement of sanitation facilities for public institutions, particularly schools; and efforts to stimulate (3) household demand for sanitation uptake and maintenance, and (4) sanitation supply.

In the areas of stimulating household demand and supply, governments often receive support from non-governmental organisations, international organisations, and the private sector. These actors play a central role in financing, designing, and implementing programmes to increase sanitation uptake—particularly among low-income and hard-to-reach populations. Their activities range from microfinance schemes to campaigns to change behaviour and community-based initiatives such as community-led total sanitation (CLTS), all with the goal of encouraging households to invest in sanitation infrastructure. Given limited public budgets, households are often left to shoulder much of the remaining responsibility for sanitation provision.

Networked Sanitation and Sewage Treatment Centres

Government failures in regulation, input provision (training centres), procurement and training can permeate the supply side of the market. The sanitation sector is rife with factors leading to market failure on the supply side including a reliance on government treatment centres that are often poorly maintained or closed, the necessity of investing in large fixed cost investments (such as large trucks) and training, and low levels of formality and training. Because networked water and sanitation are often coupled interventions, in this subsection we discuss network papers that provide us with insights into scaled investments in networked sanitation.

Further expansion of water and sewer networks substantially increases household welfare as they gain access to the network (Devoto et al. 2012), but comes at a large cost as the companies have little motive to offer water and sewage access in communities where payment is uncertain. McRae (2015) shows that subsidising companies to help them expand can lead to unintended effects including degradation in the network, particularly in poor neighborhoods. This is because companies are unwilling to invest in unprofitable areas even if they are willing to expand there when faced with large subsidies to do so. In many cases, this is because payment rates for infrastructure services are very low (Coville et al. 2023), and utilities may not be willing to enforce payment, though this is estimated to be the best way to increase payment. Ashraf et al. (2016) shows that some enforcement of payments may be necessary to generate investment in infrastructure, though there is a risk of incentivising corruption and extortion through large fines.

Informality poses a particular challenge for governments as they determine where to expand infrastructure services, yet the impacts of new access to infrastructure services are substantial. Harari (2024) shows that infrastructure provision is much more likely in informal settlements when these settlements are interspersed with the more wealthy neighborhoods in a city. Gertler et al. (2024) show that a slum-upgrading project in Nairobi, Kenya led to better quality housing both in the upgraded areas and in neighbouring areas, and led to reduced local crime rates. However, even when the utility does expand access to marginalised communities, households may be unwilling to pay for connections to sanitation services which are expensive and come with large positive externalities (Dos Santos and Guidetti 2024).

Mismanagement and ineffective expansion of sewage and water networks can even be harmful on net. Bancalari (2024) shows that unfinished improvements to infrastructure can have negative impacts on health, particularly in children when they play near or fall into inadequately covered ditches or are exposed to raw effluent. Bennett (2012) shows that expansion of piped water networks without corresponding expansion in sewage networks can lead to worsening of sanitation conditions as households are no longer exposed to the impact of poor sanitation since their water is from the utility.

Poor management of utilities is often to blame for service blockages and lack of investment. Privatisation is frequently suggested as a way to improve management and reduce roadblocks to the expansion of the network. In Senegal, Deutschmann et al. (2023) examined the effects of privatising sewage treatment centres in Dakar. Legal dumping rose by 74%, prices for legal disposal fell 5%, and diarrhoea incidence among children dropped by 22 percentage points. These large gains to sanitation infrastructure privatisation in Senegal mirror gains measured in the privatisation of the water sector in Argentina where privatisation led to an average 8% decrease in child mortality—with a 26% decrease in the poorest neighborhoods—as estimated by Galiani et al. (2005). Improvements in the management of sanitation infrastructure can deliver measurable health dividends.

Poor management can also result from unclear ownership and rights over the utility companies. Kresch (2020) shows that designation of the municipal government as the authority over water and sewage provision reduced the threat of takeover of municipal utilities by state-run companies and thereby led to improved management and functioning of the utilities. Total investment in municipality-run utilities doubled following the reform; the municipal-run utilities expanded the network to 6% more households, leading to a 16% increase in the size of the network and substantial reductions in child mortality.

Even in developed countries, when enforcement lags, the high cost of quality sanitation leads utility companies to cut corners and illegally discharge sewage into surface waters, which can leave lasting environmental, health, and welfare impacts. These concerns are doubly important in developing countries where enforcement is extremely difficult. The substantial impacts of lack of sewage collection and treatment degrade surface water quality (Lipscomb and Mobarak 2016, Lepault 2023) with substantial welfare consequences (Garg et al. 2018).

WASH in Schools

Another area in which the government can intervene to help to create a better norm of improved sanitation is in schools. Governments can directly provide better access to improved sanitation in schools and other government buildings and sanitation interventions can be run through schools using the knowledge and leadership of existing teachers.

Schools provide an easily available forum in which the government can provide information and an example to households on the importance of WASH and the impacts of adopting improved WASH. Teachers may be natural leaders to implement sanitation information interventions. For example, Crocker et al. (2017) find that teacher-facilitated CLTS in Ethiopia leads to a 13 percentage point decline in open defecation, though no change in latrine ownership.

One reason for poor attendance in schools, particularly among older girls who are menstruating, may be a lack of adequate sanitation in schools. Adukia (2017) evaluates the extent to which a component of the TSC in India, which devoted substantial financial resources to latrine construction, was able to increase attendance and reduce the probability of students dropping out. She finds that latrine construction is associated with a 5.3 percentage point reduction in dropping out of upper primary and 12.2 percentage point decrease in dropping out of primary school. It leads to an 8% increase in upper primary school enrollment and a 12% increase in primary school enrollment. The benefits in upper primary schools are entirely among latrines separated by sex, while access to latrines was important whether they were separated or not for primary school (though separated latrines did have a 50% larger impact).  Much of this impact on attendance and enrollment may be through increased safety of girls when there is access to gender-separated toilets. Gao et al. (2025) shows that the incidence of child rape reduces by 2.6 percent for a 10% increase in access to in-school toilets. Macours et al. (2024) also show significant learning gains (not through attendance) from providing menstrual health products and toilets at schools.

Figure 4 Sanitation Levels in Schools

Notes: Grey shading denotes no data. Source: UNICEF and WHO (2024); Authors’ compilation.

Though sanitation access in schools is clearly of primary importance to the students and teachers, motivating change can be difficult. Allakulov et al. (2023) examine whether a transparency intervention can instigate change for improvements in WASH service delivery in public schools in rural Bangladesh. The study tests the effectiveness of the Annotated Water Integrity Scan, a participatory transparency tool designed to improve institutional accountability through stakeholder engagement and collaborative workshops in 60 schools in Bangladesh. The findings show that while the transparency intervention had a modest positive impact on WASH-related knowledge and institutional frameworks—such as the adoption of hygiene policies and clarity around responsibilities—it did not improve actual sanitation outcomes, student WASH use, attendance, or academic performance. Notably, WASH access and use among female students declined in treated schools compared with control schools. The authors suggest that the intervention’s limited impact may be due to its short duration, the absence of binding accountability mechanisms, and the lack of participation by formal government authorities.

Household Demand

Given minimal public budgets, governments often rely on households to contribute to—or fully cover—the costs of sanitation investments. As a result, many policies focus on addressing household-level demand constraints, particularly by targeting (i) behavioural barriers, driven by social norms and (lack of) information that discourage adoption, and (ii) financial barriers, including households’ limited ability or willingness to pay.

National Campaigns: The Design and Evidence Around Community-Led Total Sanitation

The most commonly adopted—and evaluated—sanitation policy across the world is CLTS. The intervention, originally developed in Bangladesh, uses a low-resource technique to eliminate open defecation through the use of social network forces at the village level: information, peer pressure, and shaming. To date, it has been rolled out as a national policy by more than 25 governments around the world (Zuin et al. 2019).

CLTS was designed as a community-driven approach for rural areas in which a facilitator analyses community sanitation and health together with the village members through a community map of open defecation, helps the community realise correlations between open defecation and illness, and creates feelings of shame and disgust around open defecation. This community “trigger” event is then commonly followed by community efforts to improve sanitation through their own investment in latrines. To the extent possible, the model eschews subsidies in favour of mobilising the community’s own resources Kar and Pasteur (2005). However, in practice, CLTS is often implemented in combination with other components, most notably subsidies.

Whether implemented in its classic form or with additional components, nearly all rural randomised controlled trials find that CLTS effectively increases sanitation infrastructure ownership—though the magnitude of impact varies widely. When implemented in its purest form, estimated effects range from 2.4 percentage points in Indonesia (Cameron et al. 2019), 3–4 percentage points in Nigeria (Abramovsky et al. 2023), and 5 percentage points in Tanzania (Briceño et al. 2017), to 29 percentage points in Mali (Pickering et al. 2015) and as high as 60 percentage points in Ghana (Harter et al. 2020). The one exception is Bangladesh (Guiteras et al. 2015), where the intervention emphasised hygienic latrines over ending open defecation, and no effect on infrastructure uptake was found.

Three of these RCTs offer insights into the drivers of variation in sanitation uptake. Cameron et al. (2019) and Abramovsky et al. (2023) highlight the role of community characteristics. In Indonesia, Cameron et al. (2019) find that gains were concentrated in villages where the intervention was implemented by the World Bank rather than by local governments, and in communities with high initial levels of social capital. Abramovsky et al. (2023) report that the average increase in sanitation infrastructure ownership (3–4 percentage points) in Nigeria was driven by less-wealthy communities, where effects were stronger and more persistent—reaching up to 9 percentage points.  In contrast, no significant impacts were found in wealthier communities.[1]1. Abramovsky et al. (2023) show that this pattern is not driven by individual poverty status: CLTS led both wealthier and poorer households to invest when they lived in less-wealthy communities, while no such effect was observed in wealthier communities, regardless of household wealth. This contrasts somewhat with Cameron et al. (2019), who find that poorer households in Indonesia were more constrained in their ability to improve sanitation.  This pattern likely reflects the fact that community wealth correlates with factors such as baseline toilet coverage, social cohesion, and local leadership characteristics—all of which are known to influence sanitation investment. To assess generalisability, Abramovsky et al. (2023) pool their data with microdata from four other CLTS evaluations across diverse settings (Guiteras et al. 2015, Pickering et al. 2015,; Briceño et al. 2017, Cameron et al. 2019), and confirm that the inverse relationship between community wealth and CLTS impacts holds more broadly. They argue that such insights can inform better targeting of CLTS interventions.

Finally, Alzúa et al. (2020), using the same data as Pickering et al. (2015) in Mali, explore the mechanisms behind uptake. They find that health information had no measurable effect, whereas information about low-cost technical solutions shifted household perceptions of affordability, likely tipping cost–benefit calculations toward investment. They also suggest that CLTS helped create a new social norm, increasing disapproval of open defecation within the community.

A smaller set of studies address questions of CLTS design by embedding variation within their experimental setups. Crocker et al. (2016) randomised the training of “natural leaders” five months into CLTS implementation in rural Ghana and find that this additional component led to a 20 percentage point reduction in open defecation compared with the standard CLTS intervention. These findings align with Augsburg et al. (2022), who randomise CLTS follow-up activities in rural Pakistan. They show that such follow-ups can help sustain behavioural change—but only in settings with initially poor public infrastructure and limited sanitation facilities. In contrast, Briceño et al. (2017) implemented a 2x2 factorial design in rural Tanzania to test CLTS with and without accompanying handwashing promotion. Their results show that the handwashing intervention alone had no significant effects, and that combining it with CLTS did not produce any additional gains beyond CLTS by itself—suggesting no detectable interaction effects between the two components.

Overall, the evidence suggests that while CLTS can lead to large increases in sanitation ownership, these impacts tend to materialise mainly in poor rural communities and otherwise effect sizes tend to be low. Using a meta-analysis of 14 randomised controlled trials of sanitation, Whittington et al. (2020) suggest that despite these general small impacts, the low cost of running CLTS may justify it on a cost-benefit basis.

However, questions remain about the sustainability of CLTS impacts. Augsburg et al. (2022) show that the behavioural improvements achieved through CLTS tend to erode over time, particularly in areas with poor infrastructure. While their RCT demonstrates that follow-up activities can help sustain sanitation gains, even villages receiving follow-up support exhibit some regression toward pre-intervention levels. Similarly, Orgill Meyer et al. (2019) analyse the long-term effects of a sanitation intervention primarily based on CLTS and find that, a decade after implementation, differences between treatment and control groups had disappeared. This convergence was driven by the acquisition of latrines by control households and deterioration or abandonment of latrines among treated households.

Recognising the limitations of behaviour-focused approaches such as CLTS—particularly in sustaining long-term use and reaching the poorest—subsidy programmes that ease household liquidity constraints and help internalise sanitation externalities are increasingly seen as a complementary strategy. In the following subsection, we explore willingness and ability to pay for sanitation infrastructure and we also discuss the evidence on the design and effectiveness of combining CLTS with subsidy provision. Broadly, the average impacts of sanitation interventions that combine CLTS with subsidies or financial incentives for toilet construction show a similar range of impacts on toilet ownership as CLTS alone, ranging from very modest, close to zero impacts in Punjab, India (where the financial incentives were underutilised due to delays and weak outreach, Deb et al. 2024), 7 percentage points in Bangladesh (Guiteras et al. 2015), to 19 percentage points in Madhya Pradesh (Patil et al., 2014) to 50 percentage points in Odisha (Clasen et al. 2014), but the contribution of each component to this impact and their interaction effects are unclear.

Subsidies and Household Liquidity to Increase Willingness to Pay

Given the positive health externalities associated with many health-improving investments, government subsidies are often used to promote the adoption of health technologies—including sanitation—in order to achieve socially optimal coverage levels. This rationale is particularly compelling in the case of sanitation, where the upfront costs of toilet construction and related infrastructure can be prohibitively high for many households. In India, for example, the average cost of an existing household toilet was reported to exceed 50% of average annual household income (Augsburg et al. 2023b). In line with this, various papers (Peletz et al. 2017, 2019) show that, as with many goods with strong positive externalities, households’ willingness to pay for latrines is below market prices, including in Tanzania (Peletz et al. 2017) and Kenya (Acey et al. 2019, Peletz et al. 2019), with a lack of cash cited as the key underlying reason. Reflecting the importance of public support, a World Bank report estimated that in 2019, governments in LMICs allocated between 1.5% and 2% of their GDP to water and sanitation subsidies (Thibert et al. 2019).

Government subsidies can substantially increase household take up and improve household welfare. Gautam (2023) provides a structural model to show that these public investments are well justified: the cost of subsidising sanitation uptake is outweighed by the large welfare gains from increased coverage. Her findings also suggest that price subsidies are significantly more effective than unconditional cash transfers, which may be diverted to alternative uses that households perceive as more immediately beneficial. Augsburg et al. (2024) further model household decision-making around sanitation investment in a context of shared responsibility between governments and households. One of their model’s key predictions is that monetary transfers and grants should increase household spending on sanitation. This is borne out empirically—studies consistently show that subsidies raise toilet ownership and usage, generally more than programmes without financial support.

Most of the evidence comes from India, where successive government programmes have aimed to increase sanitation coverage through subsidies, beginning with the Central Rural Sanitation Programme (CRSP) in 1986, followed by the TSC in 1999, the Nirmal Bharat Abhiyan (NBA) in 2012, and the current SBM launched in 2014. While all of these programmes had a strong focus on toilet construction and provided subsidies to poor rural households, the TSC added a behavioural change component inspired by CLTS, which was kept under the NBA and SBM.
Experimental evaluations of the TSC report increases in latrine coverage ranging from 8 percentage points in Maharashtra (Hammer and Spears 2016), to around 20 percentage points in Odisha (Pattanayak et al. 2009) and Madhya Pradesh (Patil et al. 2014), and up to just over 50 percentage points in another region of Odisha (Clasen et al. 2014). A more recent evaluation of SBM in Punjab by Deb et al. (2024) shows similarly modest results, comparable to Hammer and Spears (2016).

None of these studies, however, isolates the effect of subsidies from the accompanying efforts to change behaviour inspired by CLTS. Pattanayak et al. (2009) report heterogeneous impacts by subsidy eligibility, finding that even ineligible households increased latrine adoption. They conclude that subsidies “can overcome serious budget constraints but are not necessary to spur action”. However, evidence from Bangladesh highlights the critical role of financial support for poorer households. Guiteras et al. (2015) show that a 75% targeted subsidy, when combined with CLTS programming, led to a 7.3 percentage point increase in toilet ownership, while CLTS alone had no measurable effect. A similar pattern emerges in rural Lao PDR, where Cameron et al. (2021b) find that in the context of CLTS, a partial household-level subsidy (or rebate) significantly increased toilet uptake among poorer households (by 7.1 percentage points). However, the study also shows that village-level incentives actually create a disincentive for the poorest households to build toilets: poor households are 14 percentage points less likely to build a toilet compared with non-poor households.

While subsidies can have an impact—and, in some contexts, be more effective than alternatives such as deposit requirements or earmarked savings accounts, as shown by Lipscomb and Schechter (2018) in the case of waste pit emptying in Senegal—scaling them is far from straightforward. This is especially true in settings with high informality and limited administrative capacity. Augsburg et al. (2023b) highlight this challenge in the context of microcredit for sanitation, showing that unanticipated delays in subsidy disbursement, coupled with high toilet construction costs, constrained conversion of sanitation loans to sanitation investments even among subsidy-eligible households. Deb et al. (2024) similarly argue that the very modest impacts of the Swatch Bharat Mission Gramin in Punjab were driven by underutilised subsides due to delays and weak outreach.

Limited government budgets, combined with the presence of many infra-marginal households who would purchase even without financial support, are further challenges to the design of subsidy programmes that are both affordable and effective at increasing uptake. Johnson and Lipscomb (2021) show that well-designed targeting and cross-subsidisation can generate large improvements in sanitation take-up and health while requiring low-budget outlays from the government, particularly in neighbourhoods with multiple poor households. Deutschmann (2025) shows that short-term subsidies may be more cost-effective than they appear as it may spur take-up long after the period of subsidy distribution.

Sanitation subsidy programmes tend to be focused on the poorest, but even middle-income households may have lower than optimal levels of sanitation. To the extent that households have a high willingness to pay but a liquidity constraint, approaches to increase access to finance, specifically for the purpose of purchasing improved sanitation, should be effective in increasing take-up. Various alternative approaches have been tested to address liquidity constraints, including the use of microfinance—small loans aimed at individuals without access to traditional banking—for sanitation investments. Microfinance can be attractive to governments as it offers a source of finance for subsidy ineligible households which does not require government resources. And indeed, the approach of offering microfinance for sanitation has proven effective in increasing sanitation uptake in rural India when implemented at scale (Augsburg et al. 2023b), particularly in households where the female borrower perceives greater benefits from sanitation than her spouse (Augsburg et al. 2023c). In Cambodia, Ben Yishay et al. (2017) find that offering microcredit alongside sanitation marketing significantly increased households’ willingness to pay for toilets. The programmes differed in design; in Cambodia, households received materials for a standardised toilet model delivered directly to their homes, resulting in a 35–40% conversion rate from loans to completed toilets. In contrast, the Indian programme used a more traditional microfinance approach, providing cash loans that households used to organise construction themselves—leading to a 50% conversion rate to functioning toilets. Both programmes report near-universal repayment rates. Microcredit for sanitation can also be helpful for subsidy eligible households when—as is the case in India under the SBM programme—the subsidy is paid post construction, hence not necessarily fully alleviating liquidity constraints (Augsburg et al. 2023b). However, microfinance institutions (MFIs) tend not to serve the very poor and often offer loan terms that are unattractive to higher-income households (Armendariz and Morduch 2007), making microfinance only a complementary tool in the broader effort to achieve open defecation-free status.

Information and Social Norm Channels for Increasing Willingness to Pay

While financial constraints remain a central barrier to household sanitation investment, a growing body of evidence highlights the importance of behavioural, informational, and social factors in shaping household decisions. Even when subsidies or infrastructure are available, uptake may remain low due to limited awareness, weak social incentives, or intra-household dynamics. Household demand is also inefficiently low because individuals often fail to internalise the positive externalities their investments generate for the broader community. To the extent that governments can leverage social norms to encourage internalisation of these externalities, they may be able to raise sanitation uptake while reducing the fiscal burden. This section reviews recent evidence on behavioural interventions—such as information provision, shame and social norms, public commitments, and gender dynamics—as tools to address these demand-side barriers. It also discusses how social externalities and nonlinear benefits can amplify or constrain individual behaviour. While many of these mechanisms are embedded in CLTS programmes; we review CLTS separately in Section 3.3.1 and focus here on broader behavioural approaches.

Information Provision

Low demand for improved sanitation may stem from limited information about its benefits or from misperceptions about its costs. In such cases, information provision should, in principle, affect take-up of improved sanitation. Information campaigns—including those embedded in CLTS—typically emphasise the benefits of improved sanitation infrastructure. As highlighted in Section 3.3.1, the effectiveness of CLTS varies widely. Additional evidence comes from (Guiteras et al. 2015) who evaluate a health information and community motivation campaign—the Latrine Promotion Programme—in rural Bangladesh. They find no statistically significant impact on take-up of improved sanitation. This suggests that either communities already had the information, or that the latrines were financially out of reach for the households even with full information on their benefits. Another explanation is that households overestimated the costs of sanitation investment—a pattern documented by Augsburg et al. (2023c), who show that a large share of rural Indian households significantly misjudged the actual cost of installing a toilet. Supportive of this mechanism, Alzúa et al. (2020) argue that in a pure CLTS intervention they evaluate, uptake was not driven by health information but by information about affordable technical solutions, which helped shift households’ cost perceptions in favour of investing. Deutschmann et al. (2024) show that within networks, learning is not a major channel for increasing the take-up of mechanised desludging.

Shame and Stigma

We might expect the strongest form of peer pressure to be negative peer pressure: that is, shaming individuals who do not conform to emerging social norms, such as the use of improved sanitation. While the continued use of open defecation is problematic because of the large externalities that it imposes on the community, many may also have ethical concerns related to using shame to reduce open defecation. To the extent that it is the poorest who don’t have access to improved sanitation and they have no option other than open defecation, shame imposes an additional social cost on the most vulnerable. Shame may, however, be particularly effective in contexts where traditional practices, like open defecation, are at odds with evolving community expectations; shame may help to induce  community members still using open defecation only for traditional reasons to switch when they have access to improved sanitation facilities. Guiteras et al. (2016) test the impact of shame and disgust on willingness to pay for chlorine-based water treatment. In their intervention, households were exposed either to a standard health-focused message or to one emphasising disgust at ingesting faecal matter and shame at being judged by neighbours. They find that the treatment led to a modest 3.3 percentage point increase in chlorination of drinking water—an effect that was, however, short-lived. Follow-up qualitative interviews suggest the limited impact may have been due to weak social ties: households placed little importance on the opinions of others in their housing compounds. In contrast, Pattanayak et al. (2009) argue that a 26 percentage point increase in latrine ownership under India’s TSC in Eastern India—an intervention combining CLTS-style messaging with subsidies—was primarily driven by shame rather than the financial incentive. This conclusion is based on the finding that both poor (subsidised) and non-poor (non-subsidised) households responded similarly. However, poor and non-poor households differ in many ways, and other mechanisms—such as information provision, public commitment, or social spillovers—may also have contributed to the observed effects.

An interesting avenue for future research is how the effectiveness of shame may depend on baseline sanitation coverage. It is plausible that shaming becomes more effective only in communities where latrine use is already widespread and non-compliers are visibly in the minority.

Public Commitment and Social Pressure

Public commitments may harness the social pressure element at work in shame by making people unwilling to renege on commitments, while also allowing for a more favourable public approval channel and avoiding some of the negative implications of shame in terms of reducing welfare for potentially the poorest members of society. To test the importance of public commitments in increasing take up of improvements in sanitation, Bakhtiar et al. (2023) implemented a large-scale randomised trial in Bangladesh with 19,000 households. The study tested combinations of financial incentives, public and private commitments, and social recognition. While financial rewards and public commitments both increased hygienic latrine ownership in the short run, only public commitments had sustained effects 15 months later. In contrast, private commitments and social recognition had no significant impact, highlighting the limited role of purely reputational or internal motivations.

Additional evidence on social spillovers comes from Guiteras et al. (2015), who find that subsidies increased latrine adoption not only among recipients (by 22 percentage points) but also among their unsubsidised neighbours (by 8.5 points). These peer effects also translated into a 14-point reduction in open defecation, underscoring the importance of coordinated, community-level interventions. Similarly, Pakhtigian et al. (2022), in an evaluation of India’s SBM programme, report 8.5 times higher latrine use among programme households compared to controls, with a 39.2% increase in the odds of latrine use—though these effects appear to fade over time. They emphasise the role of social multipliers in amplifying the impact of CLTS-style interventions. They suggest that the social multiplier is pivotal in increasing the impact of the CLTS style treatment.

However, the effectiveness of social pressure interventions may depend heavily on context. Deutschmann et al. (2024) find that social network-based strategies—leveraging pressure, reciprocity, and information— are not effective in an urban context.

Differences in Demand Between Genders

While much of the literature on sanitation uptake focuses on community-level influences, such as social pressure and peer effects, important variation also exists within households. In particular, gender differences in preferences, perceptions, and decision-making power can significantly shape sanitation investment outcomes.
Intra-household heterogeneity in preferences can lead to inefficiently low sanitation uptake—particularly when the household member with stronger preferences or better information lacks decision-making power. Women often express higher demand for improved sanitation: for example, in a public goods game, Pakhtigian and Pattanayak (2024) find that all-women groups contribute more to sanitation investments than mixed-gender groups. However, these stated preferences do not always translate into household-level behaviour, likely due to women’s limited bargaining power in many settings.

Augsburg et al. (2023c) demonstrate how the investment perceptions (costs and benefits) of both husband and wife, and in particular the difference in their expectations, play an important role in initial sanitation investments. They consider two key stages in the investment process and show that the take-up of a sanitation loan (targeted to female clients) is higher in households where the wife has a higher perception of the benefits of sanitation, while the successful conversion to a toilet depends on differences in cost perceptions, and in particular on husbands having lower cost perceptions than his wife. These findings highlight the importance of considering gendered expectations in both stages of the investment process.

Gendered social dynamics may also affect demand in more unexpected ways. Andrew et al. (2024) show that women with larger social networks tend to also be less likely to live in households with a toilet inside the home. This pattern supports earlier qualitative evidence from Coffey et al. (2014) and Patil (2019), which suggests that in restrictive societies, open defecation may offer women a rare opportunity for social interaction outside the household.

Externalities and Nonlinearities in Benefits

As discussed in Section 2.1.2, the health and welfare benefits of sanitation appear to be highly non-linear: they often only materialise once sanitation coverage reaches a certain threshold within a community. Evidence from Cameron et al. (2022) and Gautam (2023) suggests that such threshold effects can create multiple equilibria, where few households adopt sanitation unless a critical mass does so first. In this context, subsidies may have larger-than-expected effects if they trigger additional uptake among neighbours who begin to internalise the collective benefits of improved sanitation.

This dynamic is demonstrated in Deutschmann et al. (2024), who show that in Senegal, households living near multiple neighbours receiving high sanitation subsidies (32%) were 1–3 percentage points more likely to purchase mechanised desludging services themselves—even when their own desludging was not subsidised. These findings underscore the importance of considering externalities and social spillovers in the design and targeting of sanitation interventions.

Supply, Access to Markets, and Market Power

While much of the literature has focused on household demand for sanitation, another explanation for low uptake is that the supply of sanitation services remains limited, costly, and of poor quality. Even when households are willing to invest, they may face barriers in accessing affordable and reliable sanitation options—particularly in rural areas or informal urban settlements.

Compared to demand-side factors, the supply-side of the sanitation market has received relatively less attention. Yet constraints on supply—such as limited availability of materials, poorly functioning supply chains, or uncompetitive market structures—can significantly hinder households’ ability to act on their preferences. Some evidence suggests that demand for improved sanitation may be relatively elastic. For example, Deutschmann et al. (2024) find substantial uptake responses to price subsidies for desludging services, implying that reducing supply-side costs could have a meaningful impact on uptake. However, rigorous estimates of price elasticities in this market remain scarce.

A key constraint on the supply-side may be the limited availability of trained masons and service providers, which can raise the cost of latrine construction and reduce quality. Two studies have evaluated interventions aimed at strengthening this link in the sanitation value chain. Guiteras et al. (2015) assess a programme that provided technical assistance and training to selected community members designated as “latrine supply agents” who were tasked with disseminating information on latrine purchase, quality, installation, and maintenance. This intervention had no statistically significant effect, either on its own or when combined with household subsidies.

In contrast, Briceño et al. (2017) combined a CLTS-style intervention with a marketing campaign and mason training in rural Tanzania. The programme aimed to address both demand and supply constraints, under the hypothesis that interested households might still be unable to locate skilled service providers. The intervention led to significant gains in sanitation outcomes: latrine ownership increased by 8.2 percentage points, latrine use by 15.1 percentage points, and the probability of a village being declared open-defecation-free by 8.7 percentage points.

Abramovsky et al. (2025) evaluate a similar intervention, the “SanMark” intervention, the main aim of which is to increase supply from the private sector. The intervention consists of three key activities: (i) subsidising the development of a new low-cost toilet model tailored to local consumer preferences; (ii) engaging and training small local businesses to produce and market the model; and (iii) training locals to become sales agents of these toilets in their communities and linking them to the engaged businesses. They show that treated businesses were more likely to produce and market the new toilet model and to engage in the sales of sanitation products more generally, but there is no statistically significant impact on ownership rates at the household level. The variation in impacts across contexts suggests that any effort to increase supply must begin with a comprehensive understanding of the baseline supply in the market.

Beyond capacity constraints, imperfect competition, corruption, and market power are also common features of the sanitation sector in many LMICs (Plummer and Cross 2006). Houde et al. (2022) show that there is substantial collusion in the market for desludging truck services in Dakar, Senegal, and that this collusion is able to persist to some degree, even after the government scales-up an auction procurement system in which households are able to submit their job to a competitive bidding process for procurement. They show that a group of desludging operators is able to bid round numbers in order to maintain “tacit” collusion, and that the maintained collusion increases prices by approximately 13% relative to the case when the auctions contained only competitive bidders.

Overall, much less is known about the supply side of sanitation markets than about household demand. Research in this area faces unique methodological challenges: the number of suppliers is typically small, they interact within shared markets (Abramovsky et al. 2025), and their behaviour may violate the Stable Unit Treatment Value Assumption (SUTVA), complicating the identification of causal effects. As sanitation policy increasingly incorporates private-sector delivery models, understanding supply-side constraints—and how to address them—will be critical for scaling up effective interventions.

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