Abstract
Meeting the sanitation needs of a growing and urbanizing population in India led the incumbent national government to launch a large-scale sanitation programme, the Swachh Bharat Mission in 2014. Having reduced the deficit in basic rural sanitation in its first phase, the programme is targeting to achieve universal safely managed sanitation in its second phase launched in February 2020. Since the trajectory to achieving safely managed sanitation is inherently path-dependent, the present article utilizes novel survey data from rural India to understand the nature of sanitation systems in transitioning spaces and identify the investments they require to provide safely managed sanitation. It finds that the de facto preference in sanitation systems that households manifest is in direct contrast to the governmental choice of the low-cost twin pit system. The choice of the on-site sanitation system is informed by the household’s economic status and how ‘characteristically’ urban the settlements they reside in are. The article further discusses the implications of these findings for the short-term responsiveness and long-term effectiveness of the massive sanitation investments (USD 19 billion) planned for the next five years in India.
Keywords
Introduction
The urbanization trajectory of India is creating transitory spaces that display a mix of urban and rural characteristics relating to population size and density, the nature of the local economy and the extent of the built-up area, among others. Their socio-spatial nature often belies their formal or ‘administrative’ classification (rural or urban) to lend additional complexity to the understanding of India’s urbanisation and how it is articulated within development policies and programmes. The Constitution of India recommends instituting three types of urban local bodies (and one type of rural local body, namely, gram panchayat) depending on the size of the settlement, that is, municipal corporations for large urban areas, municipal councils for small urban areas and nagar panchayats for transitional areas; it leaves the matter of setting the exact criteria to identify urban settlements to the state government. However, the related state laws do not always have specific or uniform criteria for notifying urban settlements. Resultantly, the notified or statutory urban settlements (called ‘statutory towns (STs)’ by censuses) vary widely in their characteristics.
Studies evaluating the rate of change in the number of STs attribute the laggard pace to a multi-dimensional cost-benefit analysis by the local ‘politics of classification’ that accounts for levels of taxation, electricity rates, bye-laws and access to public schemes and programmes, which often benefit rural local bodies more than urban local bodies (Mukhopadhyay et al., 2016). The Census of 2011 attempts to correct for the discrepancy by identifying ‘census towns (CTs)’ based on a uniform standard to classify ‘urban’ settlements in lieu of a national administratively relevant standard. It considers settlements as CTs if they exhibit all of the following characteristics:
population of more than 5,000, population density of more than 400 persons per sq. km, and more than 75% of its main male workers employed in non-agriculture sectors.
Thus, the Census considers CTs and STs taken together as urban. The former accounted for more than 30% of the total urban increment recorded between the Censuses of 2001 and 2011 (Pradhan, 2017).
Athough CTs are urban settlements, as identified by an objective and relatively strict standard, they are governed as part of rural local bodies and are still deemed villages for all governmental intents and purposes. In addition to CTs, the authors’ previous work registered the unique presence of large and dense villages (LDVs), which are villages that show population densities at the level of CTs (400 persons per sq. km.) but are smaller in size and do not necessarily fulfil the requisite economic criteria. These LDVs number 158,948 (out of the total number of settlements upwards of 650,000) and amount to nearly 40% of the country’s total population (Dasgupta et al., 2017). Together, LDVs and CTs house a significant proportion of India’s population at the urban and rural interfaces. Moreover, since anointment by the state law and not a strict standard has led to many settlements being declared towns, the smallest of these at the tail-end of the spectrum often shares characteristics with CTs and LDVs (if not even smaller and less dense villages) (Balk et al., 2019).
The linkage between the socio-spatial characteristics of settlements and the nature of the sanitation infrastructure becomes apparent on disaggregating data on the type of toilet households have access to (from Census of India 2011) across different settlements. The distribution shows a steadily increasing share of septic tanks on going from villages which are neither LDVs nor CTs, to LDVs, further to CTs and STs until it drastically dips to 19% in favour of sewerage systems in the metropolitan cities. In parallel, the incidence of open defecation declines from 73% to a mere 4% on going from one end of the spectrum to the other. Interestingly, the shift between CTs and the smallest of administratively-urban settlements is imperceptible (Figure 1).

Though it was launched amidst an ongoing discourse on India’s transitory spaces and the differential prevalence of sanitation systems that they produce, the Swachha Bharat Mission (SBM), India’s latest flagship sanitation programme, has continued in the siloed tradition of its predecessors. In the run up to the Central Rural Sanitation Programme (CRSP), 1986, the national government, with support from the United Nations Development Programme (UNDP) and the World Bank, issued the ‘Manual on the design, construction and maintenance of low-cost pour-flush waterseal latrines in India’. The Manual pioneered the binary of ‘sewerage systems for all class I cities (only)’ and ‘twin pits everywhere else’ that continues to inform sanitation programmes in India to date (Dasgupta & Agarwal, 2021). The SBM, with its outlay of USD 11.6 billion between 2014 and 2021, similarly promoted twin pits to meet the massive deficit in urban and rural sanitation.
Following the completion of the first phase of SBM, investments amounting to USD 19.7 billion were allocated to its next phase. In this context, the article aims to understand how socio-spatial transformations have interacted with SBM’s programmatic preferences in sanitation systems to shape their on-ground typology in the oft-ignored transitory spaces of LDVs and CTs. Since the trajectory to achieve safely managed sanitation is path-dependent, the article discusses the implications of the divergence between programmatic preferences and the de facto prevalence of sanitation systems for the upcoming investments. Given that previous research on rural sanitation in India and similar contexts in the Global South has focused on the determinants of access to a toilet and its usage, the article uniquely furthers the evidence base by looking at what types of sanitation systems the increased access to toilets has produced and their implications for achieving universal safely managed sanitation in India.
Data and Methodology
The article utilizes data from a sample survey of 3,112 households covering five Indian states, 15 districts (sub-state administrative units) and a total of 60 LDVs (of which, 29 were CTs) conducted in December 2018. The survey collected information on the socio-economic profile of the household, level of access to and use of a toilet facility, types of sanitation systems in use, design of the sanitation system and access to desludging services for sanitation systems. It was accompanied by qualitative interviews with key stakeholders involved in sanitation service delivery in rural India to triangulate and further contextualize the quantitative findings. These included representatives of the rural local body (gram panchayat), masons and desludging service providers. Multi-stage stratified sampling informed the selection of the survey sites and allocation of the sample. As per Cochran’s formula, the sample size of 3,112 households provides a margin of error of <5% at a confidence level of 95% at both the levels of the state and country.
The authors employ a logistic regression (implemented in Stata 13.0) to investigate the relationship between the socio-spatial characteristics of a settlement, the socio-economic characteristics of the households (profile of surveyed households in Table 1) and the manifest preferences in sanitation systems. The regression considers the ownership of a septic tank as the dependent variable. The variable has two levels or outcomes: 1 = if the household’s toilet is connected to a septic tank and 0 = if it is connected to a leaching pit. Accordingly, if (p) and (1–p) represent the probabilities of each of the two outcomes, respectively, then the linear relationship underlying the modelling is,
Profile of Surveyed Households.
where xt and βt represent the independent variables acting as predictors of probability and the related co-efficients, respectively. Independent variables representing socio-cultural characteristics like religion and caste, which previous studies have reported as determinants of access to and usage of a toilet facility, did not alter our model, and therefore, it does not include these (De, 2018; Gupta et al., 2018; Tiwari & Nayak, 2017). This model had a statistically significant fit to the data, and its results are discussed in the next section.
Results and Discussion
Of all the toilets, 15% were constructed with the support of the government’s sanitation programme, and the ratio of subsidy-led toilets to privately built toilets varied from state to state (Table 2). Tamil Nadu had the highest share of subsidy-led toilets at 44% and Punjab the lowest at 8%. Overall, septic tanks emerged as the most common sanitation system, with 59% of all toilets connected to one, followed by single pits at 22% and twin pits at 8%. The dominance of septic tanks holds out across states and whether the household built the toilet with support from a subsidy or on their own, except for West Bengal and Tamil Nadu. In West Bengal, single pits are seven times more common than septic tanks among privately built toilets, although the latter’s share is higher among subsidy-led toilets. On the other hand, in Tamil Nadu, the share of septic tanks is depressed among both privately constructed and subsidy-led toilets.
Typology of Sanitation Systems by State and Mode of Construction.
On disaggregating the data by whether a given surveyed settlement was a CT or not, it was found that the share of subsidy-led toilets was higher at 18% in non-CT LDVs than at 13% in CTs. Overall, the share of septic tanks among all toilets was higher in non-CT LDVs than CTs. Although initially surprising, the finding is explained by the effect of state-level variations and their interaction with the nature of the individual settlements. At the level of individual states, residing in a CT has a significantly positive correlation with the ownership of septic tanks in Madhya Pradesh and Tamil Nadu. In other states where the share of septic tanks in CTs is less than or in the proximity of that in non-CT LDVs, the relationship between the nature of settlement and ownership of septic tanks is not statistically significant.
To clarify the relationships further, the authors employed a logistic regression, the results of which are presented in Table 3.
Results of Logistic Regression
Pseudo R-squared = 0.4529.
Log-likelihood = 692.13.
Prob > chi-square = 0.0000.
***, ** and * Indicate the 1%, 5%, and 10% significance levels, respectively.
Characteristics of Settlement
Holding all other factors constant, the odds of owning a septic tank for households in Punjab and Madhya Pradesh are 9.88 and 13.98 times higher than in Himachal Pradesh. The odds are depressed by 0.07 times if the household resides in West Bengal. Interviews with masons reveal that this phenomenon may be attributable to network effects. Masons in India typically lack formal training and technical education. They learn their trade by hands-on experience and rely on informally curated local practices and knowledge to perform their job (Dasgupta et al., 2019; Hepzibha et al., 2017). Masons interviewed as part of the study reported that unless the household states specific preferences in the type of sanitation system or its design, they build what they construe as the commonly-suitable sanitation system for the region and customize the size and design only for constraints like capital and space, as they see fit. During the Censuses of 2001 and 2011, the prevalence of pit latrines in West Bengal was among the highest in the country. Therefore, the strong influence of the state may be due to a continuing perpetuation of historical preferences.
On isolating the impact of the type of settlement, the authors find that CTs favour septic tanks more compared to non-CT LDVs. The odds of a household in a non-CT LDV owning a septic tank are 0.62 times than that of a household in the characteristically more urban CTs. Moreover, the farther a settlement is from a primary road network, the lower the odds—for every unit increase in the settlement’s distance from a major highway, the log odds of a household owning a septic tank decrease by 0.07 (Figure 2). This is important considering that 40% of all LDVs and 65% of all CTs in the country lie within 5 km of a major highway, but not surprising, since transportation corridors are linked to economic development and urbanization in India (Dasgupta et al., 2017; Jain, 2018; Maparu & Mazumder, 2017). Overall, the differential prevalence of septic tanks in LDVs driven by settlement-level characteristics appears aligned to the concept of ‘proximate social pressure driven by economic change’ at play in households’ access to toilets (O’Reilly & Louis, 2014).

Availment of Government Programme
Being constructed under a government programme lowers the odds of a toilet being connected to a septic tank by 0.33 times. This is expected since government programmes favour the construction of twin pits in rural areas to obviate downstream off-site management of faecal sludge. Unlike a septic tank or single pit, twin pits allow the prolonged storage and sanitization of faecal sludge over a 2–3 year-long operation cycle. While the sludge dries out in the first pit, households are meant to switch over to the second ‘twin’ pit. Over the system’s complete cycle, twin pits are designed to convert faecal sludge to a humus-like soil conditioner. Overall, the government programme has exerted an influence over the typology of sanitation. The influence of the programme—outside of whether the toilet is subsidy-led or privately constructed—is also observable on looking at the impact of the age of the toilet on the probability of owning a septic tank. A toilet having been constructed during a period concomitant to the programme has lower odds of being connected to a septic tank. The influence is strong enough to attenuate, but not fully negate, the state effect in Madhya Pradesh and Punjab as evinced by the statistical significance of the related interacting variable.
Access to Basic Services
The findings show that household-level characteristics are significant determinants of the sanitation system. If a household resides in a pucca or permanent 1 house, its odds of owning a septic tank are 2.02 times compared to households residing in a kutcha or semi-pucca (temporary or semi-permanent) housing. In a similar vein, the odds of owning a septic tank steadily rise on going from the lowest monthly per capita expenditure (MPCE) quintile to the highest (Figure 3). Barring the rise in odds on going from the lowest to the second lowest, all the jumps are statistically significant. On going from the lowest to the middle, the second highest and the highest quintile, the odds rise by 1.6, 2.5 and 4.09 times, respectively. Additionally, the nature of access to non-potable water a household has a significant determinant of the sanitation system. If the source of non-potable water (or water for cleaning, toilet use and bathing) is located outside the premises the household dwells in, but within the plot, the odds are 0.45 times of when it is located within the premises. Similarly, the odds are 0.52 times when the source is located away from the plot. All these determinants taken together indicate that a household with a stronger economic status and a higher level of access to related services is likelier to own a septic tank than a leaching pit.
Conclusion
The vibrant discourse on India’s hidden, messy and subaltern urbanization has been growing in tandem with the number of transitory spaces, like LDVs and CTs, that evince the phenomenon. The typology of sanitation systems in these urbanizing settlements (and urbanized, in the case of CTs) does not follow the programmatically-imposed binary of urban and rural but instead exhibits a seamless transition along an urban-rural continuum. But, the policymaking and programmes on sanitation, not unlike the others related to infrastructure and service delivery, continue to apply an outmoded binary to categorize, imagine and invest in the different types of settlements and their infrastructure.
In the same vein, the latest national sanitation programme continued the attempt to shift the typology of sanitation systems in rural areas—a mix of low-density and low-population villages, LDVs and CTs—in favour of twin pits. The second phase of the SBM, with distinct arms for ‘urban’ and ‘rural’ settlements launched in 2020 and 2021, respectively, follows its predecessor in advancing a twin leaching pit-centric strategy to achieving safely managed sanitation in rural areas. However, the article shows that as the economic status of ‘rural’ households improves and the settlements in which they live densify and undergo an urban transformation, they are more likely to own a septic tank than a single or a twin leaching pit. The prevalence of septic tanks has continued to prevail at the urban-rural interface—with the unique characteristics of urbanizing settlements combining with the de facto preferences of individual households to attenuate the impact of top-down sanitation programming.
Since the preferences in household-level infrastructure, emerging endogenously and despite countervailing external forces, create unique path dependencies in the trajectory to sustainable and universal safely managed sanitation, they raise several questions for the policymakers: would the mismatch between the expectations and preferences of the policymaker and the household create new interstices in the sanitation ecosystem? Would the policymaker need to recreate a new cycle of investments to meet the unforeseen gaps towards achieving safely managed sanitation? How cost- and time-efficient would the transition to clean, open defecation-free and open discharge-free villages be?
In laying bare the discrepancy, the article underscores the idea that fostering urbanization cannot be only about incremental changes to the services’ landscapes of the largest cities but must also tend to the recognition and development of urbanizing settlements taking a long-term view. For sanitation, it would mean accepting the prevailing state of sanitation that favours septic tanks in LDVs and CTs and addressing the challenges to environment sanitation from there on out, instead of enforcing dichotomous programmatic preferences. In doing so, faecal sludge management (FSM) may emerge as the key area demanding investments for achieving safely managed sanitation in these settlements. Strategies like urban-rural co-location and co-utilization of FSM infrastructure could be efficiently instrumental towards this objective, but to what extent and how national- and state-level policies will account for India’s socio-spatial transformations remains an urgent albeit open question.
Footnotes
Acknowledgements
The present work was supported by the Bill and Melinda Gates Foundation (BMGF), India (under the grant number OPP1136373).
Author Contributions
Shubhagato Dasgupta: Conceptualization, writing—review and editing, funding acquisition, supervision; Neha Agarwal: formal analysis, writing—original draft, writing—review and editing, visualization.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The authors received no financial support for the research, authorship and/or publication of this article.
