Report No: 43344-IN

India

Of Taps and Toilets: Evaluating Community-Demand-Driven Projects in Rural India

Social, Environment & Water Resources Management Unit

Sustainable Development Department

South Asia Region

June 30, 2008

World Bank

i

June 30, 2008

Acknowledgements & Contributions

The study was carried out by the World Bank and RTI International in collaboration with the Government of India’s Rajiv Gandhi National Drinking Water Mission, Maharashtra State Department of Water and Sanitation and the Orissa Water and Sanitation Mission. Project units, health officials, and other stakeholders in those states were also involved. The leadership, support and guidance by government officials and project staff in Orissa and Maharashtra have been critical for progress with this work. The authors are indebted to Aurobindo Behera, Pravin Darade, S.Dey, V.S. Dhumal, P.C. Khatua, Sanjeev Kumar, Ranjan Mallick, Chittaranjan Mohanty, Lalit Mohanty, Sabyasachi Nayak, R.K. Panigrahi, Rashmi Ranjan Pattanaik, S. Rath, Ratnakar Sial, Sudhir Thakre, Surendra Tripathy and Manoj Verma for their help during various stages of the study.

The World Bank formed a multidisciplinary core study team of experts in water supply and sanitation, public health, epidemiology, environment, economic valuation, survey instruments, statistical analysis, and monitoring and evaluation. Research Triangle Institute (RTI International of North Carolina USA) was appointed as principal investigators and led the design, implementation, analysis and dissemination. TNS India collected the data and Knowledge Links (New Delhi, India) conducted the community mobilization program, in collaboration with the Orissa Water and Sanitation Mission. Staff from the World Bank, the joint UNDP-World Bank Water and Sanitation Program and project implementation units in the states supported the research. The study design was reviewed by an advisory group comprising experts from the World Health Organization, U.S. Centers for Disease Control and Prevention (CDC), Agency for International Development (USAID), and health institutions in India, including the Indian Council of Medical Research (ICMR), the National Institute of Epidemiology, and Christian Medical College.

Overall, the study was led by Kseniya Lvovsky (World Bank) and Subhrendu Pattanayak (RTI). Outstanding support by Jeffrey Racki, former Sector Manager; Michael Carter, former Country Director for India; Fayez Omar, Senior Manager, India; and Rachid Benmessaoud, Operations Advisor, India, who have navigated the team through several difficult junctures, have made the completion of this research project possible.

The principal authors of this report are Subhrendu K. Pattanayak, Jui-Chen Yang, Sumeet R. Patil, Christine Poulos, Katherine L. Dickinson, Kseniya Lvovsky and Priti Kumar. Ben Arnold, Stibniati Atmadja, Jonathan Blitstein, Jack Colford, Catherine G. Corey, Nitish Jha, Richard Kwok, Jagdish Krishnappa, Ranjan Mallick, Purujit Praharaj, Vijaya Rao, Erin Sills and Kelly J. Wendland made important contributions to various aspects of this study.

The study team is grateful to, in alphabetical order, Ghanasham Abhyankar, Junaid Ahmad, Oscar Alvarado, Sushenjit Bandopadhyay, Massee Bateman, Peter Berman, Gene Brantley, Jeremy Bray, Kanchan Chopra, Jack Colford, Maureen Cropper, Anil Deolalikar, David Evans, Doug Evans, Jeff Hammer, Michael Hanemann, Kirsten Hommann, Rachel Kaufmann, Ranjiv Khush, Eckhard Kleinau, Peter Kolsky, Soma Ghosh Maulik, Ted Miguel, Smita Misra, Christine Moe, R. R. Mohan, Jack Molyneaux, Prasanta Pattnaik, Christophe Prevost, N. V. V. Raghava, Catherine Revels, Betty Sadoulet, Kurt Schwabe, Chris Timmins, Dale Whittington, and David Zilberman for their contributions to the design and/or implementation of the study and/or interpretation of results.

The study benefited from excellent comments by participants of various seminars at the Institute of Economic Growth (Delhi), Center for Excellence in Health Promotion Economics, Government of Maharashtra, Delhi School of Economics, University of California-Berkeley, University of California-Riverside, University of Minnesota (International Economic Development conference), Cornell University (Infectious Diseases in Poor Countries conference), Emory University, Duke University, the World Bank, and a joint workshop by the Indian Council for Medical Research & the U.S. Centers for Disease Control.

Finally, we would like to gratefully acknowledge the contributions of the donor countries that have supported this study: United Kingdom via the Department for International Development (DfID) Trust Fund, Netherlands via The Bank Netherlands Partnership Program (BNPP), Norway and Finland via The Trust Fund for Environmentally and Socially Sustainable Development (TFESSD) and Spain via Spanish Trust Fund for Impact Evaluation (SIEF).


Abbreviations and Acronyms

APL / Above Poverty line / MIS / Management Information System
ARI / Acute Respiratory Infections / NIE / National Institute of Epidemiology
BPL / Below Poverty line / NGOs / Non-Governmental Organizations
CDD / Community-demand-driven / O&M / Operation & Maintenance
CFU / Colony forming units / RCT / Randomized Community Trials
CLTS / Community-Led Total Sanitation / RTI / Research Triangle Institute
CMC / Christian Medical College / RWSS / Rural Water Supply and Sanitation
DALYs / Disability-Adjusted Life years / SC/ST / Schedule caste/Schedule tribe
DID / Difference-in-difference / TB / Tuberculosis
DWSM / District Water and Sanitation Mission / TNS / Taylor Nelson Sofres
GoO / Government of Orissa / TSC / Total Sanitation Campaign
ICC / Intracluster Correlation Coefficient / UNDP / United Nations Development Program
ICMR / Indian Council of Medical Research / USAID / United States Agency for International Development
IEC / Information, Education, and Communication / VHWSC / Village Health, Water, and Sanitation Committee
IHL / Individual Household Latrines / W&S / Water and sanitation
KAP / Knowledge, Attitudes and Practices / WHO / World Health Organization
LPCD / Liters per capita per day / WSH / Water, sanitation and hygiene
M&E / Monitoring and Evaluation / WSS / Water Supply and Sanitation

Table of Contents

Executive Summary 1

I. Study Methods 2

II. Intensified IEC Campaign in Orissa 2

III. Jalswarajya in Maharashtra 3

IV. Concluding observations – A Way Forward 5

1. Rationale and Objectives 9

I. Why Another Study? 9

II. The Importance of Impact Evaluations 11

III. Outputs, Outcomes, and Impacts 14

IV. Genesis 15

2. Study Design: Evaluation Approach 17

I. Intervention-Based Design 17

II. Treatment and Control Groups 17

A. Randomization in Orissa 18

B. Propensity Score Pre-Matching in Maharashtra 18

III. Sample Size 18

IV. Multiple Data Collection Tools 18

A. Survey Design 19

B. Supplementary Data 19

C. Enumerator Training 20

D. Reflections on Data Collection 20

V. Analysis Plan: DID Estimation of Intention-to-Treat Parameter 21

3. Orissa: Intensified Information, Education and Communication (IEC) Campaign 23

I. The Intervention: Intensive IEC 23

A. Logic behind Intensified IEC Campaign 24

B. What the Intervention Looked Like 26

II. Site Selection 28

III Survey Implementation 30

4. Impacts of Intensive IEC in Bhadrak, Orissa 32

I. Baseline Results of the Intensified IEC Campaign 32

A. Comparison of Treatment and Control Villages 32

B. Using Baseline to Inform Intervention 34

II Results of the Endline Survey 35

A. Latrine Adoption 38

B. Child Diarrhea 39

C. Child Anthropometrics 43

D. Household Welfare 43

III. Summary of Findings 44

IV. Latrine Adoption: Lessons for Moving Forward 44

5. Maharashtra: Jalswarajya 47

I. The Intervention: Jalswarajya 47

II. Evaluation Approach 49

A. Sample Selection and Propensity Score ‘Pre-Matching’ to Identify Controls 49

III. Survey Implementation 56

IV. Adequacy Assessment 57

6. Impacts of Jalswarajya, Maharashtra 59

I. Descriptive Statistics and Bivariate Analysis 59

A. Respondents Characteristics 59

B. Health Outcomes 59

C. Water Sources 61

D. Water Quantity 62

E. Water Quality 63

F. Personal Sanitation 64

G. Environmental Sanitation 65

H. Household WSH Knowledge, Attitudes and Practices (KAP) 65

I. Welfare Outcomes 67

II. Multivariate (Regression) Analysis with Baseline Data 68

A. Open Defecation as Diarrhea Risk Factor 68

B. Behaviors Exposed 68

III. Analytic Statistics: Basic Differences-In-Differences (DID) Estimator 69

A. Inputs: Activities and Resources 69

B. Outputs 70

C. Outcomes 72

D. Impacts 73

IV. How Much Variation is There in the Sample? 74

V. Analytic Statistics Continued: Sub-Intervention Impacts to Uncover Heterogeneity in Interventions 76

A. Water and Sanitation Schemes – Community Leaders & Key Informants 76

B. Pay as You Go 77

C. Tap and/or Toilet Villages 78

VI. Summary of Findings in Maharashtra 80

7. Conclusions: Moving Forward 82

I. In Search of Impacts: Is the Tap half-Full or Half Empty? 82

II. Improving Design of Evaluations and “M&E” of WSH Projects 84

III. Advancing RWSS Policies and Programs: Main Lessons from the Study 87

References 136


Tables

Table 1-1. Generic Model of a Water-Sanitation-Hygiene Program 15

Table 3-1. Summary of Benefits and Costs Associated with Latrine Adoption and Sanitation Improvements 24

Table 3-2. Number of Households Surveyed at Baseline and Endline in Orissa 31

Table 4-1. Differences between Control and Treatment Villages Prior to the Intensified IEC Campaign in Orissa 33

Table 4-2. Differences in Latrine Ownership and Diarrhea Prevalence between 2005 and 2006 in Treatment and Control Villages in Orissa (all numbers rounded) 35

Table 5-1. Propensity Score Estimation of Participation in Jalswarajya 52

Table 5-2. Testing covariate balance across treatment and ‘matched’ control villages using secondary data 54

Table 5-3. Testing balance across treatment and control villages using baseline survey data 55

Table 5-4. Number of Completed Surveys in the Dry Season by Study District 56

Table 5-5. Number of Completed Surveys in the Rainy Season by Study District 56

Table 6-1. DID Analysis of Sub-Groups 75

Table A2-1. Comparison of Means for Outcome Variables 95

Table A2-2. Treatment Effects Analyses for Impact of Intervention on IHL Uptake (marginal effects from probit regressions) 96

Table A3-1. Results of Specification Tests for Impact of Sanitation Campaign on Other Diarrhea-related Behaviors 99

Table A3-2. Estimates of IHL Adoption on Child Health 99

Table A4-1. Information-Gathering Methods and Topics for the Qualitative Study in Orissa 100

Table A4-2. Status of Treatment Villages at the Conclusion of the Intensified IEC Campaign in Orissa, As Reported in the Knowledge Links Report and the August 2006 Community Survey 102

Table A4-3. Attitudes Inimical to the Self-Help Concept in Treatment Villages in Orissa, According to the Knowledge Links Report 104

Table A5-1. Summary Results of the Adequacy Assessment in Maharashtra: Water Supply 108

Table A5-2. Summary Results of the Adequacy Assessment in Maharashtra: Sanitation 109

Table A5-3. Summary Results of the Adequacy Assessment in Maharashtra: Hygiene Education 110

Table A5-4. Results of the Adequacy Assessment in Maharashtra: Community Participation 111

Table A6-1. Descriptive Statistics of Respondent and Household Characteristics 114

Table A6-2. Descriptive Statistics of Health Outcomes 116

Table A6-3. Descriptive Statistics of Water Supply Services 118

Table A6-4. Descriptive Statistics of Sanitation Facilities 120

Table A6-5. Descriptive Statistics of Household WSH Knowledge, Attitudes and Practices (KAP)† 122

Table A6-6. Descriptive Statistics of Welfare Outcomes and Benefits 124

Table A6-7. DID Analysis of Program Inputs and Outputs 125

Table A6-8. DID Analysis of Program Outputs, Outcomes and Impacts 126

Table A6-9A. DID Analysis of Village Categorization Based on Water and Sanitation Schemes by In-District Sub-Sample and Season 128

Table A6-9B. DID Analysis of Village Categorization Based on Water and Sanitation Schemes by Across-District Sub-Sample and Season 129

Table A6-10A. DID Analysis of Village Categorization Based on Receiving Jalswarajya Addendum II Funds by In-District Sub-Sample and Season 130

Table A6-10B. DID Analysis of Village Categorization Based on Receiving Jalswarajya Addendum II Funds by Across-District Sub-Sample and Season 131

Table A6-11A. DID Analysis of Tap or/and Toilet Villages by In-District Sub-Sample and Season 132

Table A6-11B. DID Analysis of Tap or/and Toilet Villages by Across-District Sub-Sample and Season 133

Table A7-1. Summary of Difference-in-Difference (DID) Estimates for Diarrhea in Children < 5 135


Figures

Figure 3-1. Logic Model of the Intensified IEC Campaign in Orissa 25

Figures 3-2A and 3-2B. Intensive IEC activities (community meeting and defecation map) in the intervention villages. 27

Figure 3-3. Location of Bhadrak District within Orissa, India 29

Figure 3-4. Location of Treatment and Control Villages in Tihidi and Chandbali Blocks, Bhadrak, Orissa 30

Figure 4-1. Percent Owning and Using Toilets by Intervention and Year 36

Figure 4-2. Diarrhea Prevalence among Children under 5 by Intervention and Year 36

Figure 4-3. Arm Circumference in centimeters (cm) of Children under 5 by Intervention and Year 37

Figure 4-4. Time Spent Walking to Defecation Site (in minutes) by Intervention and Year 37

Figure 4-5. Percentage of Households with Individual Household Latrines in Treatment Villages in Orissa Between 2004 and 2007 41

Figure 4-6. Percentage of Households with Individual Household Latrines in Control Villages in Orissa Between 2004 and 2007 42

Figure 5-1. Logic Model of the Jalswarajya Program in Maharashtra 48

Figure 6-1. Diarrhea Prevalence among Children under 5 by Intervention, Season, and Year 60

Figure 6-2. Arm Circumference in centimeters (cm) of Children under 5 by Intervention, Season, and Year 60

Figure 6-3. Percentage of Households Reporting Use of Private or Public Tap as Main Water Source by Intervention, Season, and Year 62

Figure 6-4. Household Water Consumption in LPCD by Intervention, Season, and Year 63

Figure 6-5. E. Coli Levels* on log10 Scale in Household Drinking Water by Intervention, Season, and Year 64

Figure 6-6. Percentage of Households Reporting Use of Toilet as Main Sanitation Practice by Intervention, Season, and Year 65

Figure 6-7. Household Coping Costs in Rupees due to Inadequate Water Supply and Sanitation Services by Intervention, Season, and Year 66

Figure 6-8. Household Cost of Illness in Rupees due to Diarrhea by Intervention, Season, and Year 67


Annex

Annex 1. Power Calculations 90

Annex 2. Impact of Intensified IEC Campaign on Adoption of Individual Household Latrines in Orissa 93

Annex 3. Impact of Intensified IEC Campaign on Child Diarrhea in Orissa 97

Annex 4. The Process Evaluation of Intensified IEC Campaign in Orissa 100

Annex 5: Adequacy Assessment 106

Annex 6. Maharashtra Results Tables 114

Annex 7. Summary of methods to adjust for additional covariates in the DID estimates for diarrhea in children under the age of five 134

i

June 30, 2008

Executive Summary

1. The Millennium Development Goals reflect the world’s collective hope and resolve to reverse a particularly pernicious, pervasive, and persistent set of problems in much of the world: high rates of diarrhea (the number one killer of small children), insufficient water and sanitation, and seemingly unsafe and myopic behaviors. Environmental health policies related to water and sanitation (W&S) must address the usual efficiency criteria (e.g., externalities), but also significant equity concerns. Health, time, and energy costs fall disproportionately on the poor, women and children. Furthermore, there is increasing emphasis in the public health community on the deployment of multi-pronged environmental strategies (e.g., water, sanitation, and hygiene) in concert with nutritional interventions to improve child survival (Gakidou et al., 2007). Yet, to date we have few or no rigorous scientific impact evaluations showing that WSS policies are effective in delivering many of the desired outcomes (Poulos et al., 2006). Recently there have also been calls for a closer look at the effectiveness of environmental health interventions (e.g., piped water supply), as contrasted with more conventional clinical or therapeutic interventions (e.g., oral re-hydration salts, vaccination) for combating diarrheal diseases (Kremer et al., 2008).