Predicting Sanitation Behaviors: Comparison of Willingness to Pay Techniques for Sanitation Services in Urban Dhaka,Bangladesh

Abstract

Improving urban sanitation requires understanding user preferences and willingness to pay (WTP) for sewer connections. This study evaluates WTP for sanitation attributes in low-income communities (LICs) of Dhaka, Bangladesh, comparing contingent valuation (CVM) and hedonic pricing (HPM) methods. A cross-sectional study was conducted in 5 LICs, surveying 1000 households and conducting spot checks of toilet facilities. WTP was analyzed using logistic regression models for CVM and generalized linear models (GLM) for HPM, adjusting for socioeconomic factors and cluster-level correlations. The estimated WTP for a sewerage connection varied substantially between the valuation methods. The CVM produced a mean WTP of BDT 87 (95% CI: 79, 95), while the HPM yielded a significantly lower mean WTP of BDT 74 (95% CI: −223, 226). CVM assessments revealed key differences between landlords and tenants. Higher-income individuals were more willing to pay a monthly sewerage bill, but this willingness sharply declines as the bill amount increases, especially among tenants. Over 90% of respondents were willing to pay, considering their neighbors’ agreements and the health benefits for their families. Results from the HPM showed that significant determinants of monthly rental costs included the number of rooms in the household, where each additional room increased WTP by 1549 units (P < .001). Household roof materials also played a crucial role, with significant positive impacts observed for roofs made of kaccha/bamboo/jute/wood (400 units, P < .001), and cement/concrete/tiled (461 units, P < .001). Additionally, specific toilet attributes such as concrete pan materials significantly increased tenant WTP (1002 units, P < .001). Findings highlight that socioeconomic factors influence WTP, emphasizing the need for data-driven urban sanitation policies. Policymakers should integrate these insights to design affordable, inclusive, and sustainable sanitation interventions for LICs.

Plain Language Summary

Comparison of Willingness to Pay Techniques for Sanitation Services in urban Dhaka, Bangladesh

Improving urban sanitation requires understanding user preferences and willingness to pay (WTP) for sewer connections. This study evaluates WTP for sanitation attributes in low-income communities (LICs) of Dhaka, Bangladesh, comparing contingent valuation (CVM) and hedonic pricing (HPM) methods. A cross-sectional study was conducted in five LICs, surveying 1,000 households and conducting spot checks of toilet facilities. WTP was analyzed using logistic regression models for CVM and generalized linear models (GLM) for HPM, adjusting for socioeconomic factors and cluster-level correlations. From the CVM analysis we found that, higher-income individuals showed greater WTP for sewerage but were sensitive to cost increases, especially tenants. Over 90% of respondents expressed WTP, influenced by neighbor participation and perceived health benefits. From HPM analysis we found that, rental values were significantly influenced by household size, roof materials, and toilet attributes. Each additional room increased WTP by 1,452.16 units (p<0.001). Households in Dholpur (-1,013.95 units, p<0.001) and Mohajer Colony (-743.62 units, p<0.001) had lower WTP. Improved toilet features, such as concrete pan materials (+687.47 units, p<0.001) and outside locks (+255.09 units, p<0.001), significantly increased WTP. Findings highlight that socioeconomic factors influence WTP, emphasizing the need for data-driven urban sanitation policies. Policymakers should integrate these insights to design affordable, inclusive, and sustainable sanitation interventions for LICs.

Keywords

urban WASH sewer behavioral economics contingent valuation hedonic pricing model

Background

Access to basic sanitation remains a critical global challenge, affecting over 2 billion people,¹ particularly in impoverished households, where it poses significant threats to public health and impedes economic progress.² Improved sanitation not only reduces health risks but also enhances privacy, dignity, safety, and promotes gender equality.^3,4 However, rapid urbanization worldwide, including in Bangladesh, has strained sanitation infrastructure, leading to inadequate waste management, limited access to safe water, and an increase in waterborne and vector-borne diseases.⁵ Bangladesh has experienced significant urban population growth over recent decades, with more than 36% of its total population now residing in urban areas.⁶ This rapid and unplanned urbanization in Bangladesh has led to a decline in urban Water, Sanitation, and Hygiene (WASH) coverage, particularly affecting smaller and medium-sized urban areas.

Dhaka, the capital of Bangladesh, epitomizes these challenges with a population density exceeding 23 000 people per square kilometer.^7,8 Within Dhaka, low-income communities face profound sanitation deficiencies, characterized by inadequate housing, insufficient sewage and drainage systems, as well as limited access to improved sanitation facilities.⁹ In Bangladesh, while 60% of urban households have access to safely managed sanitation facilities,¹⁰ only 20% of the population is connected to a sewerage network.¹¹ This disparity highlights a significant reliance on the on-site sanitation facilities for the majority of the population, indicating a limited reach and capacity of centralized sewerage systems in urban areas. This limited sewerage network coverage contributes to groundwater contamination, exacerbating health risks and environmental degradation.¹² The United Nations Sustainable Development Goals (SDGs) underscore the importance of comprehensive sanitation management in urban areas to mitigate these challenges.¹³

Addressing the sanitation crisis in low-income urban communities requires a multifaceted approach that encompasses developing essential infrastructure and understanding the behavioral and socioeconomic factors influencing sanitation practices. In recent years, enormous efforts have been made to improve sanitation worldwide, especially in urban areas.¹⁴ Infrastructural improvements have been made in many jurisdictions, many of which encouraged contributions from the user end¹⁵ as user contributions to sanitation services increased the uptake. In Dhaka’s low-income urban settlements, access to basic sanitation remains a critical challenge. Recent studies indicate that a significant proportion of urban households still lack access to hygienic latrines, with coverage dropping below 14% in densely populated informal settlements.^16,17 Unsafe practices such as the widespread use of hanging latrines, open defecation, and unregulated waste disposal both household and medical continue to contaminate the urban environment. Moreover, over 80% of low-income residents face systemic barriers such as lack of secure land tenure, administrative complexity, weakly coordinated government structures, and cost-prohibitive formal service fees which prevent access to legally recognized services.^18,19 These persistent infrastructural and governance deficiencies are compounded by socio-political marginalization, making it difficult to implement sustainable sanitation solutions in Dhaka’s slum areas.¹⁸ A study reported mentioned that sanitation services among the slum residents of Dhaka were obtained through informal channels regulated by mastaans (landlords or middlemen) who are believed to charge extortionate amounts of money.²⁰

Therefore, understanding people’s WTP for sanitation upgrades and sewerage connections is crucial. It provides valuable insights into the financial constraints and priorities of urban residents, which can inform pricing strategies and affordability considerations.²¹ Moreover, socio-cultural factors significantly influence sanitation behaviors and WTP in informal settlements. Landlord-tenant dynamics, social norms, and governance structures within slums can affect decisions related to sanitation investments. For example, a study in Dhaka found that tenants often perceive sanitation improvements as the responsibility of landlords, impacting their willingness to contribute financially.¹⁶ Additionally, issues such as insecure land tenure and lack of formal governance can hinder sanitation interventions.²² Addressing these socio-cultural dimensions is essential for designing interventions that are both effective and socially acceptable. Moreover, recognizing individuals’ WTP is key to achieving sustainable sanitation solutions, and to bridge the gap between low-income households’ ability to pay and the actual cost of sanitation services.^23,24

Assessing the WTP for sanitation attributes specifically for the sewerage connection is an intricate and crucial task that demands careful consideration of numerous factors, including individual preferences, economic circumstances, and community dynamics. To gain a comprehensive understanding of users’ preferences and proficiently quantify their WTP, both stated and revealed preference methods are commonly employed.²⁵ In literature, the CVM has emerged as the most prevalent, albeit with certain limitations such as hypothetical bias.^26,27 On the other hand, revealed preference methods, particularly HPM have been applied to sanitation in a few cases where a large fraction of rent was being estimated, such as for the presence of a toilet.²⁸ The findings from stated preference studies are particularly notable, showing significant WTP for high-quality on-site sanitation in urban areas.²⁹ However, a study in Tanzania revealed preference studies had produced much smaller estimates of WTP, even half the market price, which amounted to less than 5% of their annual income.³⁰ Despite these well-established methodological insights, several critical knowledge gaps remain, especially in high-density, tenure-insecure environments.

Examining different techniques to estimate WTP in Bangladesh provides a unique opportunity to assess the consistency and divergence of WTP estimates. Reliable WTP figures inform decisions on market potential for new products³¹ and optimal government subsidies.³² In 2007, a Bangladesh Rural Advancement Committee (BRAC) study focusing on rural Bangladesh highlighted factors that influenced WTP for improved sanitation among households lacking latrine facilities.³³ However, despite the ongoing efforts to understand WTP for sanitation attributes, there remains a scarcity of evidence in terms of connecting to sewerage network along with other sanitation improvements in the context of low-income informal urban communities in Bangladesh. To explore the WTP for networked sewerage connection is different from WTP for sanitation attributes since it requires collective participation, carries higher capital costs, and is heavily influenced by landlord-tenant dynamics, tenure insecurity, and informal governance structures. Through an examination of estimated WTP with stated and revealed preferences for sanitation attributes, we can deepen our understanding of residents’ valuation of improved sanitation and can identify the most robust and relevant method for estimating WTP among Dhaka’s low-income informal urban communities. This study explored landlord-tenant dynamics and misaligned perceptions of WTP for sewerage connection, a critical yet underexplored barrier to infrastructure investment in informal settlements. The resulting insights will serve as a decisive guide for policymakers, urban planners, and development organizations to design effective and financially sustainable strategies for improving sanitation infrastructure in low-income urban communities around the world.

Aim

This study assesses the willingness to pay for sewerage network connections among low-income communities in Dhaka under the Dhaka Sanitation Improvement Project (DSIP) using Contingent Valuation Method and Hedonic Pricing Method.

Objectives

To understand user preferences around sewer connections (for tenants/landlords) and for different sanitation attributes

To compare stated and revealed preferences methods in terms of WTP for upgrading sanitation services and observe how well they can predict sanitation behavior

Methods

Study Setting and Study Area

A cross-sectional study was conducted in the low-income communities (LICs) of Dhaka, Bangladesh, residing in the eastern trunk main sewerage network catchment area. The Dhaka Sanitation Improvement Project (DSIP), led by the Dhaka Water Supply and Sewage Authority (DWASA), aims to construct a comprehensive sewage network and ensure wastewater treatment citywide by 2035. To support this initiative, we conducted a mixed-method study in 2020 to assess the feasibility of connecting low-income communities in Dhaka city with the sewerage network under the Dhaka Sanitation Improvement Project (DSIP). A part of this study aimed to understand user preferences for sanitation attributes and explore the financial feasibility by assessing the WTP for the sewerage connection through several methods.

The study sites were selected based on the DWASA project’s initial focus on the eastern trunk main sewerage network catchment area. Initially, a list of 100 LICs located in the East sub-catchment area was prepared based on the distance from the trunk main area, where most households were connected to the existing sewerage system. From this list, 100 LICs information related to the number of households, population size, toilet type/connection status, and distance from the trunk main were gathered for the selection of study sites. After the field visit, 5 LICs were selected to cover a range of areas considering diverse population size, availability of at least 200 households where landlords or tenants were abundant, and distance from the proposed sewerage network. Among these 5 LICs, 1 was within 100 feet (30.5 m) of the trunk main (where all households were connected to the existing sewerage network), while the others were within a range of 100 feet (30.5 m) to 2 km (Figure 1). These 5 LICs were finalized upon consultation with WSUP and selected based on population size. Areas with connection to the sewerage network were considered for estimating WTP using HPM, whereas the other areas not connected to the sewerage network were considered as the study sites of CVM.

Figure 1.

Study sites: 5 low-income communities within eastern trunk main area.

Study Population

Study participants included both landlords/homeowners and tenants residing in the selected LICs. Landlords/homeowners were defined as individuals who owned or leased properties and collected rent from tenants. On the other hand, “tenants” were residents paying monthly rent for their accommodations. Adult respondents over 18 years old were chosen from each household, alternating between landlords and tenants. In the selected households where only tenants were present, the tenant was interviewed. Furthermore, in cases where more than 1 adult respondent was available, we chose respondents who were responsible for household decision-making

Sample Size

The sample of the WTP estimation was derived from the broader feasibility study. The sample size of the feasibility study was estimated based on the current population in the Dhaka South City Corporation (DSCC), and calculated at the household level, since a shared toilet is used by the entire household. According to the slum census report, there were 40591 households in the slums of Dhaka South City Corporation.³⁴ The minimum required sample size was estimated using the standard formula for proportions with finite population correction due to small population size.³⁵

n = \frac{z^{2} \times p (1 - p)}{d^{2}} \times d e f f \times \frac{N}{N + n_{0} - 1}

Here, Z is the critical value of the Normal distribution with a 95% confidence level (1.96), p is the expected proportion (consider .5 for maximum variability), d is the margin of error (5%), deff is the design effect, N is the population size, and n₀ is the sample size without finite population correction.

A design effect of 2.5 was considered, adjusting for intra-cluster homogeneity among households in urban informal settings due to shared environmental and socio-economic conditions. Although within clusters, sanitation behaviors can vary, reflecting that these intra-cluster variations increase the likelihood that responses are more correlated than in a simple random sample, thereby inflating the design effect, as aligned with the existing literature.^36,37 Considering 80% power and 5% non-response rate, the estimated sample size was 1000 households for the broader study.

However, the WTP modules were collected from a sub-sample of these 1000 households, which were aligned with the criteria of the WTP modules (CVM and HPM). For the CVM module, data were collected from 800 households that were not yet connected to the sewer network. This sample size was sufficient to provide statistical reliability of WTP estimates as supported by existing literature, which documented that a medium-sized sample (250-400),³⁸ or at least 600 samples are needed for single-bound and at least 400 samples are required for a precise estimate of double-bound questions in the CVM module.³⁹ For the HPM module, 550 tenant households were enrolled to estimate the WTP from the household’s market value or rental value. This sample size is sufficient for estimating HPM predictors, as the common rule of thumb suggests having at least 15 to 20 observations per explanatory variable to ensure stable coefficient estimates and adequate degrees of freedom in cross-sectional hedonic applications with potential multicollinearity and market heterogeneity.⁴⁰

Household Selection

Each selected slums were divided into 4 parts (east, west, north, south) for spatially representative sampling, accounting for variations in sanitation access and behaviors based on location, such as proximity to drainage channels, canals, and sewerage networks (Figure 2- Yellow bordered). This grid-based approach aligned with existing literature on informal settlements and ensured unbiased representation across the entire settlement.^41,42 In these four segments, households were sampled with randomly generated starting points to ensure representation and to avoid the chance of overlapping. These starting points were determined using a GIS-based spatial randomization approach, where GPS coordinates were collected to map the boundaries of each LIC, a grid was overlaid in the GIS environment, and random x- and y-coordinates were generated using R software to serve as enumeration starting points.

Figure 2.

Four parts of selected LIC and Starting points (white circle) of Mohajer Colony (T denotes toilet locations and DW denotes drinking water points; numbers indicate unique facility identifiers within each LIC).

From each selected LIC, a total of 200 households (50 households from each part) were selected, resulting in a study population of 1000 households. Households were selected using a systematic random sampling technique with a skip pattern based on the total number of households in the LIC. We divided the total household number by 200 (required number of households per area) to determine the sampling interval (number of households to skip) to select the households. As a result, slums with smaller numbers of households resulted in smaller intervals compared to other areas. The result provided the household (HH) interval between 2 respondents. Table 1 shows the interval value for the selected LICs:

Table 1.

The Interval Value for the Selected Households in the LICs.

Name	Total HH	Total PoPⁿ	Trunk main distance (m)	Interval between HH (# of households skipped)
City Polli Dholpur Slum	1038	5095	229	5
Madhubagh Wireless Railgate Slum	934	4093	927	5
IG Gate Bank Colony Slum	250	1029	674	1
Mohajer Colony Slum	378	1492	1423	2
Baganbari Bagicha Shyampur Slum	326	1204	23	2

Research Team Training and Pre-testing

A team of 16 enumerators participated in a comprehensive 5-day training session led by the research team. The training covered topics on study objectives, research design, methods, ethical considerations, details of data collection tools, and how to administer WTP modules. To validate the questionnaires and gain an understanding of the initial estimates of bid amounts, 3 pilot tests were conducted in 3 different urban slum areas (similar to the study sites) in Dhaka with approximately 50 households. Feedback from these pilot tests informed subsequent revisions to the questionnaire and bid prices to enhance clarity, cultural appropriateness, and feasibility of the data collection process.

Data Collection

The study employed a rigorous data collection approach in the 5 LICs of Dhaka’s eastern trunk main sewerage network catchment area from January to February 2020. Household surveys were conducted alongside spot-checks of toilet facilities to gather detailed information. Initially, a structured questionnaire was used to collect data on household composition and socio-economic status (eg, household size, number of children, gender of the household head, household income, asset ownership, etc.), housing conditions (ie, rental status, physical condition of housing, electricity access, etc.), current water, sanitation, and sewerage facilities, as well as factors influencing the demand for sewerage connections.

We further assessed willingness of both landlords and tenants to pay for sewer connections to understand the financial feasibility of providing connections to slum residents to improve environmental health conditions. The study used 2 modules (CVM & HPM) to assess participants’ WTP. In the HPM module, WTP for sewerage connection and sanitation attributes were inferred from the actual rent of the households that had access to the sewerage network or not. Households were revealed for WTP through spot-check of different household infrastructure (eg, materials of wall, roof and floor, no of rooms, no of households on plot, presence of electricity etc), toilet facility and toilet attributes (eg, materials of roof, wall, availability of solid door, inside-outside lock, water seal, slab materials, pan materials, flushing facility, handwashing place availability etc). This approach was chosen because hedonic analysis relies on revealed preference through observed market behavior for determining the marginal implicit prices associated with different sanitation attributes, where individuals do not directly acquire toilet components but instead access them through rental decisions. Furthermore, CVM module comprises a set of 5 questions designed to elicit individuals’ preferences concerning different aspects of sanitation facilities. We asked about their preferences regarding the most important aspects of toilets to rank from highest to lowest. We also explored their WTP for a regular fee for connecting their toilet to a sewerage network, considering scenarios such as if their neighbors (those living next-door households to them) agreed to the same, if this connection could improve their health and their family, and if they could avoid paying a larger amount when the existing toilet fills or clogs. Tenants were directly asked about their WTP, while landlords responded about tenant perceived WTP. Each of these questions also assessed their WTP for a predetermined bid price (100 BDT, 200 BDT, 300 BDT, 400 BDT) to be connected with a sewerage connection. The bid amounts were initially determined through a review of relevant literature on WTP for sanitation services in similar low-income urban contexts, as well as in consultation with subject experts. After that these bid amounts were adjusted and validated based on the data from 3 small-scale pilot-tests.

The data collectors randomly selected a single-bounded dichotomous choice question during the data collection process. If respondents replied affirmatively, the data collectors then randomly selected a follow-up predefined bidding price to prompt respondents’ WTP appraisal. They provided the option to either accept or refuse the stated cost value. Data was collected on password-protected devices (tablets/phones). Enumerators uploaded the data to the KoboToolbox server daily, and daily data quality checks were conducted to ensure the accuracy and reliability of the collected data.

Outcome Variable

The key outcome variable of the study was the WTP for sanitation attributes.

Data Analysis

Data cleaning and analysis were conducted using STATA software (Version 14.2, StataCorp LP). Prior to analysis, all the collected data was checked for completeness after each day of data collection. During data collection, all survey questions were marked mandatory in KoboToolbox, preventing any missing responses.

The data analysis focused on understanding the demand for sanitation among Dhaka’s urban poor communities, encompassing both private, non-sewered sanitation and sewerage connections. Unlike previous studies, our emphasis was specifically on Dhaka’s lowest wealth and income groups residing in slum areas, highlighting disparities in sanitation demand and accessibility between landlords and tenants.

For CVM, we illustrated individuals’ willingness to pay for a regular fee for connecting their toilet to a sewerage network using a line graph, with the x-axis representing “Bid Price (in BDT)” and the y-axis representing “Percentage of respondents WTP.”

Furthermore, we assessed individuals’ WTP directly using 4 dichotomous-choice contingent valuation questions related to sewerage connections. A binary variable for WTP of sewerage connection was generated, with the value 1, if the respondent answered “Yes” to any of the 4 sewerage-related valuation scenarios at the offered bid and 0 otherwise. Logistic regression models were used to analyze determinants of WTP, with bid price acceptance as the dependent variable and socio-economic characteristics and bid amounts as covariates, adjusted for cluster-level correlation.

Our analytical approach proceeded in 2 stages to identify the potential determinants of WTP. First, we conducted univariate logistic regressions for each potential socio-economic covariate, which were selected based on existing literature and expert consultation. Variables achieving a significance level of P = .20 in this initial screening were retained for subsequent multivariate analysis.

The regression model utilized to estimate the predictors of WTP using the CVM is given as follows:

\begin{array}{l} l o g i t (P i) = \ln (\frac{P i}{1 - P i}) = β_{0} + β_{1} H H S i z e_{i} + β_{2} R e s p o n d e n t_{i} + β_{3} A g e_{i} + β_{4} S e x_{i} + β_{5} E d u c a t i o n_{i} + l n β_{6} I n c o m e_{i} \\ + β_{7} B i d_{i} + \in_{i} \end{array}

Where P_i is the probability that respondent i is willing to pay for a toilet sewerage connection at the offered bid amount. β₀ is the intercept and β_i denotes the coefficient of the parameter, and € denotes the error term.

Individual WTP for each household was calculated as below:

W T P_{i} = - \frac{β_{0} + \sum_{j = 1}^{n} β_{j} X_{i j}}{β_{b i d}}

Where β_j are coefficients for covariates X_ij and β_bid is the coefficient for the bid amount. Based on the estimated coefficients β_j, for each household, the individual WTP_i for each household was calculated by dividing the negative linear predictor by the estimated bid coefficient. The mean WTP was then computed as the average of individual WTP values across the sample. The 95% confidence interval was determined using bootstrap resampling with 1000 replications of the individual WTP estimates.

In contrast, HPM leveraged market choices made by respondents to assess implicit prices associated with sanitation attributes, particularly their impact on monthly rental payments. Regression analyses within the HPM framework included various housing and toilet attributes such as presence, quality characteristics, and sewerage connection. During the HPMs, we didn’t control for variables related to neighborhood characteristics or proximity to infrastructure such as proximity to city centers, distance to nearest hospitals since our study sites were homogenous in nature such as housing conditions, infrastructure access, toilet conditions and environmental characteristics were similar, which reduced variation in factors like building quality or proximity to city amenities. This approach aligned with the modeling strategy used in prior research in similar urban low-income contexts,²⁸ allowing us to exclude variables such as distance to major facilities or variation in neighborhood characteristics from the model.

For HPM, we conducted 2 regression models: Model 1 incorporating toilet presence as a binary variable alongside housing characteristics, and Model 2 considering detailed measures of toilet quality along with housing characteristics as the predictors. Since the dependent variable “monthly rent” was highly skewed and even if the log transformation of the rent failed the Shapiro-Wilk test of the Normality (W = 0.979, P < .0001). Therefore, the generalized linear model (GLM) was used to estimate each HPM model with monthly rent specified in levels; coefficients represented marginal changes in monthly rent (BDT) associated with sanitation and housing attributes. The GLM framework was selected due to its flexibility in modeling non-normally distributed outcomes and its ease of interpretability in WTP estimation literature. Both models combined the attributes in a standard linear fashion to estimate their influence on the monthly rental payment. This allowed us to quantify the implicit prices associated with different toilet attributes and assess their impact on housing costs. Both of the estimated regression equations are given below:

H_{p} = α_{0} + \sum_{i = 1}^{n} α_{i} H s_{i} + β_{1} t o i l e t_{i} + \in_{i}

H_{p} = α_{0} + \sum_{i = 1}^{n} α_{i} H s_{i} + \sum_{i = 1}^{n} β_{i} T_{i} + \in_{i}

Where H_p denotes the monthly rental cost of the household, α and β denotes the coefficient of parameter estimation. Hsi denotes the vector of household attributes, T_i denotes the vector of toilet attributes and € denotes the error term.

The WTP estimate of sewerage connection was derived from the marginal price coefficient of the sewerage connection variable in the regression model.

Prior to estimation, we assessed multicollinearity using variance inflation factors (VIFs; See Supplemental Table 1). While continuous covariates and binary indicators showed VIFs well below the conventional threshold of 10,⁴³ elevated VIFs were observed among some categorical variables representing mutually exclusive categories of wall and roof materials. This reflects structural multicollinearity, an inherent feature of fully specified categorical variables, rather than model misspecification. As noted in econometric literature, high VIFs among variables from the same conceptual construct inflate standard errors but do not bias coefficient estimates, preserving the validity of inference.^44,45 Consequently, these variables were retained to maintain a theoretically complete model of housing quality a standard practice in applied hedonic studies where categorical built-attribute variables are essential for unbiased hedonic price estimation. In CVM and HPM, the model’s goodness of fit was evaluated using Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) statistics. Model with the lowest AIC & BIC was reported as the final model. Since we have data from five different slums, all models were adjusted for cluster-level correlation using the sandwich estimator.

To compare the mean WTP estimated from CVM and HPM, a 2-sample Welch t-test with the bootstrap estimated 95% confidence interval.

Ethical Approval

This study protocol received approval from the Ethical Review Committee (ERC) of the International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b; #PR-19022). Before the enrollment, enumerators read an information sheet containing a summary of the study, benefits or potential risks of participation, privacy declaration, and contact information to the participants in Bengali. Enumerators answered any questions raised in the field, provided a copy of that information sheet, and obtained written consent from the participants. The participants were free to refuse or withdraw from the study at any time. Confidentiality of information was strictly maintained, and access to the data was restricted among the research team members.

Results

Respondent Characteristics

A total of 800 participants enrolled for CVM and 550 enrolled for HPM. For CVM, participants were evenly distributed (25%) to the 4 areas not connected to the sewerage system, with a similar number of landlord-tenant enrollment (54%, 46% respectively). The majority of respondents (68%) in CVM data reported residing in the slum areas for more than 5 years, while around 40% of participants from HPM reported the same duration of living. Around 70% of the participants were female in both the CVM and HPM modules, and most of them (around 60%) were aged between 20 and 39 years. About 45% of the participants had no formal education, while around 20% didn’t complete their secondary education (Table 2).

Table 2.

Socio-Demographic Characteristics of the Study Participants.

Indicators	CVM	HPM
	N = 800	N = 550
	n (%)	n (%)
Age (y)
18-19	39 (5)	30 (6)
20-29	239 (30)	191 (35)
30-39	241 (30)	177 (32)
40-49	128 (16)	79 (14)
50-59	96 (12)	46 (9)
60-69	46 (6)	23 (3)
70+	11 (1)	4 (1)
Sex
Male	254 (32)	181 (33)
Female	545 (68)	369 (67)
Transgender	1 (0.2)	0 (0)
Education
No formal education	359 (45)	246 (45)
Primary incomplete	117 (15)	84 (15)
Primary complete	120 (15)	86 (16)
Secondary incomplete	152 (19)	103 (19)
Secondary complete or higher	52 (7)	31 (5)
LIC
Mohajer Colony	200 (25)	63 (12)
IG Gate	200 (25)	148 (27)
Madhubag	200 (25)	74 (14)
City Polli Dholpur	200 (25)	86 (16)
Baganbari Bagicha	-	179 (33)
Duration of living in current home
<6 mo	37 (5)	55 (10)
1 y	63 (8)	92 (17)
2 y	44 (6)	55 (10)
3 y	41 (5)	36 (7)
4 y	34 (4)	38 (7)
5 y	37 (5)	49 (7)
>5 y	544 (68)	225 (40)

Household Characteristics

Table 3 exhibited household characteristics across both module participants. In both groups, households were predominantly medium-sized, with 3 to 4 members being the most common category, and the mean household size was similar across methods (4 ± 2). The majority of households in both samples had no children under 5 years of age, with a mean of 0.5 ± 0.6 children. Most households were located in compounds shared by multiple households, typically ranging from 6 to 20 households per compound, reflecting high residential density in the study areas. Housing conditions were generally modest, particularly among HPM households, where nearly 9 in 10 lived in single-room dwellings compared to two-thirds of CVM households. Access to electricity was high in both methods (>95%). Mean monthly household income was BDT 15 477, while monthly rent averaged BDT 2596 among CVM households and BDT 2792 among HPM households. As expected, none of the CVM households were connected to the sewerage network, while one-third (33%) of HPM households had an existing sewerage connection (Table 3).

Table 3.

Household Characteristics of Study Participants.

Indicators	CVM	HPM
	N = 800	N = 550
	n (%)	n (%)
# Of household members
⩽2	101 (13)	94 (17)
3-4	375 (47)	301 (55)
5-6	246 (31)	134 (24)
7-8	51 (6)	20 (3)
9-10	17 (2)	1 (1)
⩾11	10 (1)	0
Mean (±SD) number of household members	4 (± 2)	4 ( ± 2)
# Of under 5 children
0	478 (60)	322 (59)
1-2	314 (39)	228 (41)
3-4	8 (1)	0
Mean (±SD) number of under five children	0.5 (± 0.6)	0.5 (± 0.6)
# Of HH on same compound
1 household	33 (4)	6 (1)
2 households	40 (5)	11 (2)
3-5 households	128 (16)	90 (16)
6-10 households	191 (24)	161 (29)
11-20 households	189 (24)	171 (31)
21-40 households	126 (15)	63 (12)
40-100 households	80 (10)	39 (7)
More than 100 households	13 (2)	9 (2)
# Of rooms in the household
1	529 (66)	482 (88)
2	199 (25)	62 (11)
3	42 (5)	3 (0.5)
4	16 (2)	2 (0.3)
⩾5	14 (2)	1 (0.2)
Availability of electricity	774 (96)	536 (98)
Mean (±SD) monthly household income (in BDT)	15 477 (± 7010)	15 477 (± 6261)
Residency status
Landlord	429 (54)	0
Tenant	371 (46)	550 (100)
Mean (±SD) monthly rent (in BDT)	2596 (± 1147)	2792 (± 1167)
# Of household connected to sewerage network	0	179 (33)

Contingent Valuation Method (CVM)

The results from CVM assessments revealed notable differences between landlords (perceived tenant WTP) and tenants in terms of their WTP for connecting to sewerage at different bid prices. Table 4 showed a clear domination of monthly income over the WTP of the respondents. People with a higher average monthly income are more willing to pay a monthly sewerage bill. Of the 800 respondents for CVM, a randomly selected subgroup of 173 was asked whether they would be willing to pay for sewerage connections, among them 86% agreed to pay and 14% didn’t agree to pay. Of the respondents who were asked to pay a monthly fee of 100 BDT and 200 BDT for the sewerage connections, more than 80% agreed to pay the bill. However, the WTP reduces drastically with the increase of the proposed bid amount, and this reduction is much more stark for the tenants comparing the landlords. More than 90% of the respondents agreed to pay while considering their neighbors' agreements and their family’s overall health benefits.

Table 4.

Stated Willingness to Pay from Contingent Valuation Method.

Agreed to the statement for the toilet use (randomly asked 1 question to each respondents)	Agreed to the statement and agreed to pay on a monthly basis	Landlord (perceived tenant WTP)				Tenant				OverallN = 800
		N = 429				N = 371				OverallN = 800
		Agreed to pay 100 BDT	Agreed to pay 200 BDT	Agreed to pay 300 BDT	Agreed to pay 400 BDT	Agreed to pay 100 BDT	Agreed to pay 200 BDT	Agreed to pay 300 BDT	Agreed to pay 400 BDT	Agreed to pay 100 BDT	Agreed to pay 200 BDT	Agreed to pay 300 BDT	Agreed to pay 400 BDT
		n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)
	N = 173	N = 18	N = 15	N = 23	N = 10	N = 23	N = 19	N = 26	N = 14	N = 41	N = 34	N = 49	N = 24
Would like to pay for a toilet connected to a sewerage network	148 (86)	16 (89)	13 (87)	13 (57)	2 (20)	20 (87)	15 (82)	14 (53)	1 (7.1)	36 (88)	28 (82)	27 (55)	3 (13)
Average monthly family income of those who want to Pay	16 348 (SD ± 7364)	16 188 (SD ± 7204)	17 539 (SD ± 8323)	21 462 (SD ± 12 414)	35 000 (SD ± 21 213)	15 350 (SD ± 5761)	15 667 (SD ± 5407)	15 643 (SD ± 3177)	15 000 (SD ± 0)	15 722 (SD ± 6359)	16 536 (SD ± 6845)	18 444 (SD ± 9217)	28 333 (SD ± 18 930)
Average monthly family income of those who don’t want to pay	13 140 (SD ± 5090)	12 500 (SD ± 7778)	16 000 (SD ± 5657)	15 200 (SD ± 5653)	17 125 (SD ± 5643)	8667 (SD ± 3512)	15 750 (SD ± 7500)	14 667 (SD ± 5245)	15 115 (SD ± 5938)	10 200 (SD ± 5070)	15 833 (SD ± 6338)	14 909 (SD ± 5309)	15 881 (SD ± 5770)
	N = 348	N = 34	N = 43	N = 40	N = 19	N = 39	N = 48	N = 66	N = 28	N = 73	N = 91	N = 106	N = 47
Would like to pay for a toilet connected to a sewerage network if neighbors agreed	317 (91)	31 (91)	39 (91)	25 (63)	8 (42)	29 (74)	37 (77)	35 (53)	6 (21)	60 (82)	76 (84)	60 (57)	14 (30)
Average monthly family income of those who want to Pay	16 186 (SD ± 6779)	16 452 (SD ± 6801)	17 462 (SD ± 8556)	19 920 (SD ± 9014)	18 875 (SD ± 9141)	16 517 (SD ± 4830)	16 297 (SD ± 5792)	15 371 (SD ± 7440)	15 500 (SD ± 3564)	16 483 (SD ± 5882)	16 895 (SD ± 7317)	17 267 (SD ± 8370)	17 429 (SD ± 7272)
Average monthly family income of those who don’t want to pay	14 758 (SD ± 7235)	14 000 (SD ± 5292)	17 500 (SD ± 9713)	14 600 (SD ± 5096)	13 182 (SD ± 3341)	15 500 (SD ± 5583)	14 636 (SD ± 6845)	15 548 (SD ± 4945)	13 818 (SD ± 6500)	15 154 (SD ± 5336)	15 400 (SD ± 7443)	15 239 (SD ± 4958)	13 606 (SD ± 5595)
	N = 312	N = 17	N = 52	N = 34	N = 28	N = 35	N = 51	N = 45	N = 20	N = 52	N = 103	N = 79	N = 48
Would like to pay for a toilet connected to a sewerage network considering the overall health of the family	282 (90)	15 (88)	42 (81)	24 (71)	10 (36)	29 (83)	35 (69)	25 (56)	7 (35)	44 (85)	77 (75)	49 (62)	17 (35)
Average monthly family income of those who want to pay	16 550 (SD ± 7802)	14 400 (SD ± 5475)	17 667 (SD ± 9100)	18 167 (SD ± 6631)	19 100 (SD ± 6191)	16 828 (SD ± 5971)	16 814 (SD ± 10 724)	18 040 (SD ± 8979)	18 286 (SD ± 6873)	16 000 (SD ± 5859)	17 279 (SD ± 9814)	18 102 (SD ± 7835)	18 765 (SD ± 6280)
Average monthly family income of those who don’t want to pay	14 433 (SD ± 6426)	8000 (SD ± 5657)	13 400 (SD ± 6381)	14 200 (SD ± 5371)	18 556 (SD ± 10 118)	14 167 (SD ± 5811)	11 906 (SD ± 4831)	14 200 (SD ± 4641)	18 539 (SD ± 4446)	12 625 (SD ± 6070)	12 481 (SD ± 5405)	14 200 (SD ± 4802)	18 548 (SD ± 8119)
	N = 167	N = 11	N = 33	N = 24	N = 11	N = 18	N = 23	N = 15	N = 5	N = 29	N = 56	N = 39	N = 16
Would like to pay for a toilet connected to a sewerage network to avoid paying a larger amount when the existing toilet fills or clogs	140 (84)	10 (91)	28 (85)	17 (71)	7 (64)	14 (78)	18 (78)	8 (53)	2 (40)	24 (83)	46 (82)	25 (64)	9 (56)
Average monthly family income of those who want to pay	17 404 (SD ± 6510)	18 100 (SD ± 5425)	16 893 (SD ± 7613)	17 882 (SD ± 5011)	24 429 (SD ± 13 252)	16 686 (SD ± 4836)	16 722 (SD ± 4638)	20 250 (SD ± 3845)	25 000 (SD ± 7071)	17 275 (SD ± 5024)	16 826 (SD ± 6550)	18 640 (SD ± 4725)	24 556 (SD ± 11 749)
Average monthly family income of those who don’t want to pay	12 889 (SD ± 4925)	6000 (-)	17 400 (SD ± 5128)	13 714 (SD ± 5283)	17 000 (SD ± 6164)	13 000 (SD ± 2160)	14 400 (SD ± 4336)	19 714 (SD ± 5823)	14 000 (SD ± 7937)	11 600 (SD ± 3647)	15 900 (SD ± 4748)	16 714 (SD ± 6182)	15 714 (SD ± 6525)

Source: https://www.exchange-rates.org/exchange-rate-history/bdt-usd-2020.

1 BDT = 0.01179 USD according to January 31, 2020 exchange rate.

We evaluated the willingness of tenants to pay for a regular fee for connecting the toilet to a sewerage connection and concurrently gathered landlords’ perspectives on tenant WTP (Figure 3). Overall, more than 80% of respondents, irrespective of their residency status (landlord or tenant), expressed some WTP at the lowest bid amount (100 BDT). However, the respondent’s WTP gradually decreased as the bid amount increased and a significant decrease was observed for the highest bid amount (400 BDT).

Figure 3.

Respondents willingness to pay for a regular fee for connecting the toilet to a sewerage connection.

Table 5 provides an analysis of the factors affecting individuals’ WTP for sewerage connections. The univariate model revealed strong association between the WTP and factors like household size, education-level and logarithm of income. In the multivariate model, while the estimated parameters’ magnitudes varied slightly across the analyzed covariates, household size and income consistently showed positive effects on WTP. Moreover, the bid amount for the sewerage connection was significantly negatively correlated, suggesting that as bid amounts rose, WTP decreased slightly by 1%.

Table 5.

Determinants of Willingness to Pay for Sewerage Connection from Contingent Valuation Data.

Outcome = 1 if willing to pay for sewerage connection at offered bid value	Univariate model	Multivariate model
Covariates	Odds ratio (95% CI)	Odds ratio (95% CI)
No of HH members	1.2 (1.1-1.3)***	1.1 (1.1-1.2)***
Respondents type = tenant	0.6 (0.3-1.2)	0.7 (0.3-1.4)
Age	0.9 (0.9-1.0)	0.9 (0.9-1.0)
Sex = male	0.8 (0.5-1.3)	0.7 (0.5-1.1)*
Education in years (ref = No education)
1-5	0.9 (0.7-1.0)	0.7 (0.6-0.8)***
6-10	1.4 (1.1-1.8)**	1.2 (0.6-2.3)
>10	4.4 (2.5-7.7)***	2.7 (0.8-9.6)
Log (income)	2.4 (1.5-3.9)***	2.2 (1.6-3.1)***
Bid amount	0.9 (0.9-1.0)	0.9 (0.9-1.0)***

Statistical significance is denoted at the 1% (***), 5% (**), and 10% (*) levels.

Hedonic Pricing Method

Two HPMs were estimated to identify significant determinants of monthly rental costs (in BDT). Model 1 includes all the household attributes as the predictors, while Model 2 encompasses both household and toilet attributes as predictors. The relative fit of these models was assessed using the Akaike Information Criterion (AIC). Model 2, which incorporates both household and toilet attributes, demonstrates a better fit (AIC = 1636) compared to Model 1 (AIC = 1656), indicating its enhanced ability to capture tenant preferences and rental pricing dynamics associated with household and toilet attributes.

Parameter estimates from Model 2 exhibited that the estimated WTP is significantly associated with the number of rooms in the household, where per room present in the households is the largest contributor to the overall WTP (1549 units; P < .001). Household roof materials also played a crucial role, with significant positive impacts observed for roofs made of kaccha/bamboo/jute/wood (400 units, P < .001) and cement/concrete/tiled (461 units, P < .001) compared to the rudimentary roof. Additionally, in Model 2, specific toilet attributes such as concrete pan materials significantly increased tenant WTP (1002 units, P < .001; Table 6).

Table 6.

Willingness to Pay Estimation Results from Hedonic Pricing Models.

Covariates	Model 1		Model 2
	Coef.	95% CI	Coef.	95% CI
Household attributes
No of households on plot	−21	−239 to 196	−33	−238 to 170
No of rooms	1342***	519-2165	1549***	836-2261
Presence of electricity	305	−273 to 884	357	−194 to 909
Material of household roof
Kaccha/bamboo/thatch/Jute/wood	361	−131 to 853	400**	−41 to 841
Tin	233	−198 to 664	243*	−386 to 873
Cement/concrete/tiled	389*	−49 to 827	461**	26-895
Rudimentary roof	Ref.		Ref.
Other	39	−663 to 742	207	−1297 to 1713
Material of household wall
Kaccha/bamboo/thatch/Jute/wood	−283	−804 to 236	−288	−1071 to 494
Tin	−239	−838 to 359	−290	−1135 to 553
Cement/concrete/tiled	346*	−9 to 702	256	−444 to 956
Rudimentary wall	Ref.		Ref.
Material of household floor
Kaccha/bamboo/thatch/Jute/wood	−966	−21 984 to 266	−698***	−980 to −417
Cement/concrete/tiled	−713*	−1558 to 130	−303	−646 to 38
Rudimentary floor	Ref.		Ref.	-
Toilet attributes
Has household toilets	47	−156 to 250	−73	−417 to 271
No of household sharing toilets	-	-	−0.86	−2 to 0.43
Material of toilet roof
Kaccha/bamboo/thatch/Jute/wood	-	-	Ref.
Tin	-	-	−170	−1552 to 1211
Cement/concrete/tiled	-	-	−18	−1117 to 1081
Rudimentary roof	-	-	−820*	−1636 to −4
Material of toilet wall
Kaccha/bamboo/thatch/Jute/wood	-	-	Ref.
Tin	-	-	−26	−349 to 297
Cement/concrete/tiled	-	-	−306	−609 to −3
Other	-	-	−266*	−588 to 55
Solid door	-	-	−292	−1154 to 568
Inside lock	-	-	-1	−316 to 313
Outside lock	-	-	198	−65 to 463
Water seal	-	-	−62	−239 to 113
Material for slab	-	-
Wood	-	-	Ref.
Cement/concrete	-	-	−175	−344 to −6
Pan materials
Ceramic	-	-	Ref.
Plastic	-	-	188	−248 to 626
Concrete	-	-	1002***	486-1519
Water available for flushing	-	-	−171	−507 to 1519
Place for handwashing	-	-	283	37-530
Toilet connect to sewerage	-	-	74	−223 to 371
AIC: 1656 1636

Source: https://www.exchange-rates.org/exchange-rate-history/bdt-usd-2020.

WTP estimates values were evaluated in BDT, 1 BDT = 0.01179 USD according to January 31, 2020 exchange rate.

Statistical significance is denoted at the 1% (***), 5% (**), and 10% (*) levels.

Comparison of WTP Using CVM and HPM

The estimated WTP for a sewerage connection varied substantially between the valuation methods. The CV method produced a mean WTP of BDT 87 (95% CI: 79, 95). HP method yielded slightly lower and less precise mean WTP of BDT 74 (95% CI: −223, 226). The t-test showed a statistically significant difference in mean WTP (difference = 12 BDT, 95% CI: 1, 25; P = .04), indicating that households stated WTP was slightly higher than revealed WTP (Table 7).

Table 7.

WTP Estimates for Sewerage Connection Across WTP Methods.

Indicator	CVM	HPM	Mean difference	P-value
	WTP estimates in BDT [95% confidence interval]		[95% confidence interval]
Sewerage connection	87 [79, 95]	74 [−223, 226]	12 [1, 25]	.04

Discussion

This study explored the variation in the sanitation preferences of landlords and tenants in the context of urban slum areas in Bangladesh through 2 different techniques, and the prediction of demand for any commodity or service is usually determined by predicting the WTP for that specific good or service.³¹ The findings of this study revealed that individuals’ socioeconomic determinants, such as age, monthly family income, and number of household members, significantly impacted their WTP for the sewerage connection. HPM revealed a significant association between urban slum dwellers’ WTP and various household characteristics, including the type of roof and floor materials, the number of rooms, and specific toilet attributes like the materials of the pan and the presence of outside door locks.

The study revealed that a substantially high percentage of landlords perceived tenants’ WTP, and tenants expressed their WTP for sewerage connection, but their WTP exhibited an inverse relationship with the bill amount. Furthermore, our research showed that a slightly higher percentage of landlords were willing to pay for sewerage connection compared to tenants. This aligns with earlier research from Ghana and Kenya, which found that property owners or landlords are more inclined to pay the sanitation surcharge compared to caretakers or tenants.^4,46 This perception of WTP highlights individuals’ consideration of viewing sanitation as a basic service rather than a value-added amenity. This reconceptualization has significant implications for urban sanitation policy, particularly in low-income informal settlements, where WTP is often shaped more by structural constraints than by preferences. This highlights the need for progressive financing mechanisms, such as targeted subsidies, sliding-scale pricing, or cross-subsidization models wherein higher-income or commercial users offset the cost for the poorest. These approaches have been shown to enhance equity and cost recovery in similar urban sanitation contexts.^47,48

Additionally, the sanitation cost, which can constitute a significant portion of rent in informal settlements, influences tenant preferences and landlord-tenant dynamics, further affecting retention rates and financial decisions.²⁸ The emphasis on cost highlights the significant impact of economic limitations on individuals’ preferences when making choices. Therefore, it is crucial for policymakers to consider these financial thresholds when developing cost-sharing programs targeting consumers who lack the financial means to purchase improved products. This collective action on WTP as evidenced by another literature which emphasize the importance of community-driven sanitation financing models with shared incentives, which is particularly relevant in densely populated settlements where households are interdependent in their use of shared facilities.⁴⁹ Even in tenure-insecure settings, collective upgrading helped build social capital and motivated landlords to participate when trust and benefit-sharing were established⁴⁹ and it may impact long-term investments for sanitation improvement. This is consistent with Loomis’ recognition of the significant role that economic factors play in influencing WTP in different environmental situations.⁵⁰ In addition, various studies conducted in Peru,⁵¹ Kenya,²¹ and the Philippines⁵² have shown that household income plays a crucial role in determining the maximum WTP for sanitation products, such as septage and sewerage services. These studies have found that the WTP varies across different income groups, while the ones with higher incomes generally lead to an increased WTP for sanitation technologies. However, contextual differences across countries, such as urban governance, the informal economy, and service delivery models, along with institutional and infrastructural variations, may hinder comparability in interpretation of WTP. Moreover, The findings of the CVM analysis revealed that certain socio-economic factors, such as age and monthly income, substantially impact individuals’ WTP. Our results showed that respondents’ WTP varied across different income groups for various bid amounts. This finding aligns with existing literature, which suggests that individuals’ income positively influences their WTP as households with higher incomes can afford to pay higher bid amounts.⁵³ Furthermore, previous literature also exhibited on how various socio-economic determinants play a significant role in determining the WTP for sanitation attributes.^21,31,54,55 Former studies have shown that an Individual’s age and sex were observed to be influential determinants in the household decision-making process, including those related to sanitation services.^56,57 Those with younger age might likely prioritize advanced sanitation facilities for health and convenience, while older individuals may have more traditional views on acceptable standards, impacting their willingness to invest. Gender differences also play a role as studies show that women often value sanitation improvements more due to their roles in household health management and hygiene practices and are willing to invest in sanitation facilities effectively.⁵⁸ A survey of rural China also exhibited positive WTP in investing in sanitation upgrades, which were associated with different communal activities since they believed it might benefit the community and the environment.⁵⁹ This implies that gender is a more salient predictor of sanitation investment behavior, along with intra-household decision-making power.^60,61

Results from the CVM underscore the importance of the landlord-tenant relationship in decision-making regarding sanitation in informal settlements. Our findings indicate that landlords often underestimate tenants’ WTP, which can diminish their perceived return on investment in sanitation improvements. This discrepancy is significant, as it illustrates a market failure related to information, where tenants who are willing to contribute may be excluded from enhanced services due to landlords’ inaccurate assumptions. Previous studies in urban sanitation have demonstrated that when landlords are provided with accurate information about tenant demand and potential cost-sharing options, their likelihood of upgrading facilities increases.^47,62

CVM analysis further uncovered those perceptions of neighbors and community play a pivotal role in higher individual WTP, driven by a sense of communal responsibility and shared values. Studies showed that social interactions, peer influences, and community norms are key drivers affecting individuals’ decisions to contribute financially or participate in sanitation infrastructure improvements, emphasizing the importance of leveraging social effects and neighbor’s behaviors in designing effective and sustainable sanitation policies and programs.^63
-65 These findings underscore the significant influence of social norms and perceived collective benefits on WTP in addition to economic capability, which is consistent with social capital theory that plays a powerful role in shaping economic decisions in urban poor settings.⁶⁶

Moreover, concerns about family well-being and economic considerations regarding the prospect of avoiding higher repair costs for clogged toilets through investing in a reliable sewerage connection serve as a compelling economic incentive, aligning with long-term cost-benefit considerations. This association has been identified in previous research where the interplay of social, health, and economic factors collectively shapes individuals’ decisions regarding WTP measurements.^67
-69 The findings indicate that a significant number of landlords and tenants expressed a willingness to enhance the quality of their sanitation facilities, driven by a desire to improve the well-being of their families by avoiding the inconveniences associated with shared toilets⁷⁰ and reducing the transmission of pathogens that cause diarrheal diseases.⁷¹ These determinants are contextualized within the specific socio-cultural environment of an urban slum. Recognizing and understanding these sociocultural nuances is crucial to shed light on the community-level contributions to an individual’s WTP.

The study found that certain features of toilets, such as the material of the pan and the presence of outside door locks, had a significant impact on their perceived value according to the HPM. The focus on concrete and tangible factors in sanitation preferences aligns with theories of consumer behavior, as these factors directly impact consumers’ perceptions and choices regarding sanitation products and services.⁷² The preference for high-quality pan materials and increased security features—such as an external door lock—indicates a prudent assessment of characteristics that enhance the visual appeal and long-term viability of sanitation facilities. By recognizing and addressing these nuanced preferences, policymakers and stakeholders can tailor sanitation infrastructure initiatives to align with consumer expectations, ensuring a more effective and sustainable approach to urban development.

Results from stated and revealed preference methods highlight significant differences that require careful examination within the framework of urban sanitation planning. The higher WTP estimates from the CVM compared to the HPM align with existing evidence on hypothetical bias, where respondents tend to overstate their willingness in hypothetical scenarios.^73,74 While the estimates from HPM might stem biases due to imperfections in market values such as rent controls, information asymmetry, or lack of property rights that suppress the capitalized value of sanitation amenities.⁷⁵ Through the use of CVM, this study examined the impact of socio-economic determinants and economic constraints on WTP. The findings shed light on how socio-economic factors shape individuals’ stated preferences. HPM, in contrast, explored specific factors that influence valuation and emphasized a greater WTP for tangible housing amenities.

Both HPM and CVM show a much smaller WTP compared to their mean income, which is also similar to the findings of a study conducted in Tanzania.³⁰ The presence of negative WTP for specific sanitation attributes indicates potential trade-offs, concerns regarding affordability, and variations in the perceived value among respondents. In addition to this, individuals’ negative WTP with some household amenities and their residency areas may be due to their dissatisfaction with current services or higher expectations that arise from government-funded services. This suggests that WTP is influenced not only by affordability and personal preference but also by institutional credibility and perceived accountability.

The combination of both methods reveals the underlying strengths and limitations of each strategy in studies on WTP. Each of these methods contributes unique perspectives, CVM reflecting individuals’ stated preferences under hypothetical preferences for households not yet connected to the sewerage network.⁷⁶ While HPM is grounded in observed market behavior and reveals WTP embedded in household rental values, making it less susceptible to hypothetical bias, it may not fully reflect the value of sanitation services in informal settings.^66,77 This is because HPM assumes that market prices reflect all relevant attributes, but in low-income rental markets, tenants often do not invest directly and may be unaware of sanitation costs in rental values. Consequently, household rent may not reflect sanitation investments even when households value them, particularly in contexts where rents are shaped by tenure insecurity, limited housing choices, and landlord-tenant dynamics.^28,57 The choice of methodologies may vary depending on the context.⁷⁸ When formulating urban sanitation policies, policymakers must carefully consider the detailed findings from various methods, acknowledging the complex nature of individual valuing mechanisms and context-sensitive sanitation investment in order to implement successful and socially inclusive initiatives.

Limitations of the Study

Despite uniqueness and strength, this study has several limitations as well. We used rigorous methods to estimate WTP through both stated and revealed preference approaches. However, social desirability and hypothetical bias may still affect results. To minimize this bias, we trained enumerators in neutral questioning, and ensured respondent confidentiality. Despite these measures, some residual bias may remain. Secondly, the study was conducted in selected low-income settlements of Dhaka and may not be fully generalizable to all urban contexts in Bangladesh. However, the current study includes most relevant LICs for the current DSIP phase and provides a basis for further assessments that may yield broader insights. Despite these limitations, the study provides valuable insights into the sanitation preferences and WTP among urban slum dwellers and highlights key contextual factors that can inform future policy and infrastructure planning.

Recommendations

This study presents a comprehensive examination of sanitation preferences and WTP in the slum areas of Dhaka, providing in-depth insights that contribute to both theoretical understanding and practical applications. The study findings highlighted disparities in magnitude and signs of the WTP assessment revealed across stated and revealed preferences methods emphasized the complexity of individual valuation mechanisms in the choice preference. However, for planning sustainable peri-urban sanitation services for all, precise assessments of WTP combining appropriate revealed preference methods will be crucial. This study will make a significant contribution to the fields of environmental economics and urban development through its cultural contextualization, multi-method approach, and nuanced exploration of sanitation preferences of urban slum dwellers. By advancing evidence generation, this research, thereby imparting invaluable insights that can inform policy decisions and targeted interventions aimed at fostering inclusive and sustainable urban development.

Supplemental Material

sj-docx-1-ehi-10.1177_11786302261437530 – Supplemental material for Predicting Sanitation Behaviors: Comparison of Willingness to Pay Techniques for Sanitation Services in Urban Dhaka, Bangladesh

Supplemental material, sj-docx-1-ehi-10.1177_11786302261437530 for Predicting Sanitation Behaviors: Comparison of Willingness to Pay Techniques for Sanitation Services in Urban Dhaka, Bangladesh by Mahbub-Ul Alam, Kazy Farhat Tabassum, Ismail Hosen, Tara Devi Laabar, Richard Norman and Ian W. Li in Environmental Health Insights

Footnotes

ORCID iD

Kazy Farhat Tabassum

Author Contributions

MUA conceptualization, methodology, supervision, writing – drafting, review and editing. KFT conceptualization, methodology, data interpretation, writing – drafting, review and editing. IH methodology, data interpretation, writing – drafting, review and editing. TDL conceptualization, methodology, writing – review and editing. RN conceptualization, methodology, writing – review and editing. IW conceptualization, methodology, writing – review and editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We want to acknowledge the contribution of data collectors, study participants, and the Tsuha Foundation for the funding the study.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Dataset used for this study is available from the corresponding author on reasonable request.

Supplemental Material

Supplemental material for this article is available online.

References

WHO and UNICEF. Progress on Household Drinking Water, Sanitation and Hygiene 2000-2024: Special Focus on Inequalities. World Health Organization (WHO) and the United Nations Children’s Fund (UNICEF); 2025.

Fewtrell

Prüss-Ustün

Bos

Gore

. Water, Sanitation and Hygiene: Quantifying the Health Impact at National and Local Levels in Countries With Incomplete Water Supply and Sanitation Coverage. World Health Organization; 2007.

Khanna

Das

Why gender matters in the solution towards safe sanitation? Reflections from rural India. Glob Public Health. 2016;11(10):1185-1201.

Peletz

Delaire

Kones

, et al. Will households invest in safe sanitation? Results from an experimental demand trial in Nakuru, Kenya. Int J Environ Res Public Health. 2021;18(9):4462.

Habib

Impact of urbanization on sanitation management in Pakistan: the Case of Islamabad Capital Territory. Ann Hum Soc Sci. 2022;3(II):495-508.

United Nations. 68% Of the World Population Projected to Live in Urban Areas by 2050. United Nations, Department of Economic and Social Affairs; 2018.

Dhaka Tribune. Dhaka now 4th most populous city globally. Dhaka Tribune; 2022.

World Bank. Urban Population (% of Total Population) - Bangladesh. World Bank; 2022. Accessed December 12, 2024. https://data.worldbank.org/indicator/SP.URB.TOTL.IN.ZS?locations=BD

Bangladesh Bureau of Statistics (BBS). Census of slum areas and floating population 2014. 2015. https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/axvjbnqprylg/b/V2Ministry/o/office-bbs/2024/12/53c9ef0872294083adfda78151eeb460.pdf

10.

Akter

Islam

Akter

, et al. Equity in access to safely managed sanitation and prevalence of diarrheal diseases in Bangladesh: a national and sub-national analysis. BMC Infect Dis. 2022;22(1):871.

11.

Mansour

Islam

Akhtaruzzaman

Situation Analysis of the Urban Sanitation Sector in Bangladesh. WSUP-Water & Sanitation for the Urban Poor; 2017.

12.

World Health Organization. Sanitation. World Health Organization; 2021. Accessed December 12, 2024. https://www.who.int/news-room/fact-sheets/detail/sanitation

13.

Lüthi

Willetts

Hoffmann

City-Wide Sanitation: The Urban Sustainability Challenge. Frontiers Media SA; 2020:585418.

14.

WHO and UNICEF. Progress on sanitation and drinking water: 2015 update and MDG assessment. WHO/UNICEF Joint Water Supply Sanitation Monitoring Programme, World Health Organization: UNICEF; 2015.

15.

Hutton

Global costs and benefits of reaching universal coverage of sanitation and drinking-water supply. J Water Health. 2013;11(1):1-12.

16.

Alam

M-U

Sharior

Ferdous

, et al. Strategies to connect low-income communities with the proposed sewerage network of the Dhaka sanitation improvement project, Bangladesh: a qualitative assessment of the perspectives of stakeholders. Int J Environ Res Public Health. 2020;17(19):7201.

17.

Das

Ghosh

Singha

. Participatory Community hygiene education in Dhaka slums: DSK Experience. South Asia Hygiene Practitioners’ Workshop; 2010.

18.

Farah

Karim

Akther

Begum

Knowledge and practice of personal hygiene and sanitation: a study in selected slums of Dhaka city. Delta Med Coll J. 2015;3(2):68-73.

19.

Sinharoy

Pittluck

Clasen

Review of drivers and barriers of water and sanitation policies for urban informal settlements in low-income and middle-income countries. Utilities Policy. 2019;60:100957.

20.

Arias Granada

Haque

Joseph

Yanez Pagans

. Water and sanitation in Dhaka slums: access, quality, and informality in service provision. World Bank Policy Research working paper 8552. 2018.

21.

Mulatya

Were

Olewe

Mbuvi

Willingness to pay for improvements in rural sanitation: evidence from a cross-sectional survey of three rural counties in Kenya. PLoS One. 2021;16(5):e0248223.

22.

Rahman

Atkins

McFarlane

Factors affecting slum sanitation projects in Dhaka City: learning from the dynamics of social-technological-governance systems. J Water Sanit Hyg Dev. 2014;4(3):346-358.

23.

Delaire

Peletz

Haji

, et al. How much will safe sanitation for all cost? Evidence from five cities. Environ Sci Technol. 2021;55(1):767-777.

24.

Meeks

JV.

Willingness-to-Pay for Maintenance and Improvements to Existing Sanitation Infrastructure: Assessing Community-Led Total Sanitation in Mopti, Mali. University of South Florida; 2012.

25.

Hutton

Varughese

The Costs of Meeting the 2030 Sustainable Development Goal Targets on Drinking Water, Sanitation, and Hygiene. The World Bank; 2016.

26.

Tilley

Logar

Günther

The effect of giving respondents time to think in a choice experiment: a conditional cash transfer programme in South Africa. Environ Dev Econ. 2017;22(2):202-227.

27.

Hausman

Contingent valuation: from dubious to hopeless. J Econ Perspect. 2012;26(4):43-56.

28.

Simiyu

Swilling

Rheingans

Cairncross

Estimating the cost and payment for sanitation in the informal settlements of Kisumu, Kenya: a cross sectional study. Int J Environ Res Public Health. 2017;14:49.

29.

Anjum Altaf

Hughes

. Measuring the demand for improved urban sanitation services: results of a contingent valuation study in Ouagadougou, Burkina Faso. Urban Stud. 1994;31(10):1763-1776.

30.

Peletz

Cock-Esteb

Ysenburg

Haji

Khush

Dupas

Supply and demand for improved sanitation: results from randomized pricing experiments in rural Tanzania. Environ Sci Technol. 2017;51(12):7138-7147.

31.

Van Minh

Nguyen-Viet

Thanh

Yang

JC.

Assessing willingness to pay for improved sanitation in rural Vietnam. Environ Health Prev Med. 2013;18(4):275-284.

32.

Whittington

Lauria

Wright

Choe

Hughes

Swarna

Household demand for improved sanitation services in Kumasi, Ghana: a contingent valuation study. Water Resour Res. 1993;29(6):1539-1560.

33.

Seraj

KFB

. Willingness to Pay for Improved Sanitation Services and Its Implication on Demand Responsive Approach of BRAC Water, Sanitation and Hygiene Programme. BRAC; 2008.

34.

Bangladesh Bureau of Statistics. Preliminary Report on the Census of Slum Areas and Floating Population 2014. 2015.

35.

Cochran

WG.

Sampling techniques. John Wiley & Sons; 1977.

36.

Bennett

Woods

Liyanage

Smith

. A simplified general method for cluster-sample surveys of health in developing countries. World Health Stat Q. 1991;44(3):98-106.

37.

Katz

Zeger

SL.

Estimation of design effects in cluster surveys. Ann Epidemiol. 1994;4(4):295-301.

38.

Calia

Strazzera

Bias and Efficiency of Single vs. Double Bounded Models for Contingent Valuation Studies: A Monte Carlo Analysis. Working Paper. CRENoS; 1998.

39.

Fujita

Fujii

Furukawa

Ogawa

. Estimation of willingness-to-pay (WTP) for water and sanitation services through contingent valuation method (CVM): A case study in Iquitos City, the Republic of Peru. JBICI Rev. 2005;11:59-87.

40.

Linz

Beisteiner

Böttcher

et al. Statistics and Science: Handbook on the Application of Quality Adjustment Methods in the Harmonised Index of Consumer Prices. Federal Statistical Office of Germany; 2009.

41.

Improving Health in Slums Collaborative. A protocol for a multi-site, spatially-referenced household survey in slum settings: methods for access, sampling frame construction, sampling, and field data collection. BMC Med Res Methodol. 2019;19(1):109.

42.

Victor

Ocasio

Cumbe

, et al. Spatial heterogeneity of neighborhood-level water and sanitation access in informal urban settlements: a cross-sectional case study in Beira, Mozambique. PLOS water. 2022;1(6):e0000022.

43.

Kutner

MH.

Applied linear statistical models. 2005.

44.

Belsley

DA.

On the efficient computation of the nonlinear full-information maximum-likelihood estimator. J Econ. 1980;14(2):203-225.

45.

Wooldridge

JM.

Econometric Analysis of Cross Section and Panel Data. MIT Press; 2010.

46.

Appiah-Effah

Duku

Boakye

Oduro

Nyarko

KB.

Navigating Taxpayers Attitude: Exploring support for urban sanitation surcharge in Kumasi. Environ Health Insights. 2024;18:11786302241247797.

47.

Trémolet

Kolsky

Perez

Financing on-site sanitation for the poor: a six-country comparative review and analysis, (Technical paper / WSP). Water and Sanitation Program, WSP; 2010.

48.

Andres

Thibert

Cordoba

Danilenko

Joseph

Borja-Vega

. Doing more with less: Smarter subsidies for water supply and sanitation. 2019.

49.

McGranahan

Realizing the right to sanitation in deprived urban communities: meeting the challenges of collective action, coproduction, affordability, and housing tenure. World Dev. 2015;68:242-253.

50.

Loomis

What’s to know about hypothetical bias in stated preference valuation studies?

J Econ Surv. 2011;25(2):363-370.

51.

Tudela-Mamani

Leos-Rodríguez

JA.

Estimation of economic benefits due to improvements in basic sanitation services through choice experiments. Rev Chapingo Ser Cienc Forestales Ambiente. 2018;24(2):237-250.

52.

Harder

Sajise

AJU

Galing

EM.

Willingness to pay for sanitation services in Dagupan City, Philippines. J Water Sanit Hyg Dev. 2013;3(2):165-180.

53.

Orgill

Shaheed

Brown

Jeuland

Water quality perceptions and willingness to pay for clean water in peri-urban Cambodian communities. J Water Health. 2013;11(3):489-506.

54.

Hazra

Goel

Maitra

Willingness-to-pay and preference heterogeneity for service attributes of solid waste management: an experience in Kolkata, India. Global NEST J. 2014;17(1):82-92.

55.

Dare

AM.

Rural households’ access, willingness to pay (WTP) and factors influencing WTP for safe water and sanitation in Southwest Nigeria. Agricultural and Applied Economics Association (AAEA) Conferences; 2014; Minneapolis, Minnesota.

56.

Sharma

KK.

Do the socioeconomic characteristics influence households’ preferences for improved solid waste management options? (A study in the Tinsukia district of Assam). Int J Sci Technol Res. 2019;8:1572-1576.

57.

Tidwell

Chipungu

Chilengi

Curtis

Aunger

Theory-driven formative research on on-site, shared sanitation quality improvement among landlords and tenants in peri-urban Lusaka, Zambia. Int J Environ Health Res. 2019;29(3):312-325.

58.

Jenkins

Scott

Behavioral indicators of household decision-making and demand for sanitation and potential gains from social marketing in Ghana. Soc Sci Med. 2007;64(12):2427-2442.

59.

Ding

Wang

Residents’ willingness to invest in sanitation: evidence from rural China. Desalination Water Treat. 2020;182:405-413.

60.

Jenkins

Curtis

Achieving the ‘good life’: why some people want latrines in rural Benin. Soc Sci Med. 2005;61(11):2446-2459.

61.

Routray

Schmidt

Boisson

Clasen

Jenkins

MW.

Socio-cultural and behavioural factors constraining latrine adoption in rural coastal Odisha: an exploratory qualitative study. BMC Public Health. 2015;15:880.

62.

Osumanu

KI.

Urban Water and Sanitation in Ghana: How Local Action is Making a Difference. Iied; 2010.

63.

Pakhtigian

Dickinson

Orgill-Meyer

Pattanayak

SK.

Sustaining latrine use: peers, policies, and sanitation behaviors. J Econ Behav Organ. 2022;200:223-242.

64.

Zhang

B-J.

Public willingness to pay for and participate in sanitation infrastructure improvement in western China's rural areas. Front Public Health. 2021;9:788922.

65.

Ray

Ryngnga

Chetia

Willingness to pay and its determinants for improved solid waste management: a case study. Int J Environ Waste Manag. 2023;31(1):119-134.

66.

Duku

Bonso

Appiah-Effah

, et al. Achieving universal sanitation in Ghana: an analysis of key drivers of toilet ownership among property owners in urban areas. PLoS One. 2025;20(1):e0307729.

67.

Safari

Issa

Masanyiwa

ZS.

Community perception and adoption of improved sanitation in Tunduru district, Tanzania. 2020.

68.

Doma

Patrick

Conrad

, et al. An exploration of sanitation-related decision-making, leadership, collective action, and freedom of movement among women in urban Tiruchirappalli, India. Front Water. 2023;5:1048772.

69.

Nobi

MN.

Willingness to pay to prevent water and sanitation-related diseases suffered by slum dwellers and beneficiary households: evidence from Chittagong, Bangladesh. arXiv preprint arXiv:2109.05421, 2021.

70.

Shiras

Cumming

Brown

Muneme

Nala

Dreibelbis

Shared latrines in Maputo, Mozambique: exploring emotional well-being and psychosocial stress. BMC Int Health Hum Rights. 2018;18:30.

71.

Freeman

Garn

Sclar

, et al. The impact of sanitation on infectious disease and nutritional status: a systematic review and meta-analysis. Int J Hyg Environ Health. 2017;220(6):928-949.

72.

Riyanto

Parengkuan

Edbert

Consumer Behavior on buying decision sanitation product in Jakarta, Tembagapura, and bandar Lampung after the end of PPKM. Int J Res Rev. 2023;10:729-747.

73.

Murphy

Stevens

Weatherhead

Is cheap talk effective at eliminating hypothetical bias in a provision point mechanism?

Environ Resour Econ. 2005;30:327-343.

74.

Blumenschein

Johannesson

Yokoyama

Freeman

Par13: hypothetical versus real willingness to pay in the health care sector: results from a field experiment. Value Health. 2001;4(2):79-80.

75.

Freeman

Iii Herriges

Kling

CL.

The Measurement of Environmental and Resource Values: Theory and Methods. Routledge; 2014.

76.

Liziński

Wróblewska

Rauba

Application of CVM method in the evaluation of flood control and water and sewage management projects. J Water Land Dev. 2015;24:41-49.

77.

Carreres-Prieto

García

Cerdán-Cartagena

Suardiaz-Muro

Wastewater quality estimation through spectrophotometry-based statistical models. Sensors. 2020;20(19):5631.

78.

Shin

Study on variability of WTP estimates by the estimation methods using dichotomous choice contingent valuation data. Environ Resour Econ Rev. 2016;25(1):1-25.