Sanitation Policy Raymond Guiteras, Maryland James Levinsohn, Yale - PowerPoint PPT Presentation

Sanitation Policy Raymond Guiteras, Maryland James Levinsohn, Yale A. Mushfiq Mobarak, Yale Implementation Partners: Innovations for Poverty Action Wateraid, Bangladesh VERC Grantor: The Bill and Melinda Gates Foundation

Sanitation • One billion people, or about 15% of the world’s population, currently practice open defecation. • 2.5 billion do not have access to an improved sanitation facility • MDG goal of 75% coverage by 2015 will be missed by close to one billion people. • Poor sanitation is estimated to cause 280,000 deaths per year • May also be responsible for serious long-term health conditions such as stunting or tropical enteropathy

Policy Interest • In 2012 UNICEF spent $380 million on programs focused on water, sanitation, and hygiene for children. • The World Bank’s Water and Sanitation Program plans to direct US $200 million to improve sanitation for 50 million people during the 2011-2015 period (World Bank WSP, 2013). • In India, over half the population practices open defecation (Census Organization of India, 2011) • Prime Minister Narendra Modi’s pre-election campaign: “pehle shauchalaya, phir devalaya ” (“toilets first, temples later”) • He has pledged to eliminate open defecation in India by 2019

Policy Debates in Sanitation Programming • Is Demand Generation necessary or sufficient? • Is a Supply-Side Push necessary or sufficient? • Are Subsidies necessary? Are they helpful? • Who should you subsidize? How should you subsidize? • Trying to bring some rigorous evidence to bear on these questions for rural Bangladesh

General Research Question: Interdependencies in Household Decisions • Latrine adoption decisions may be interdependent across households – Epidemiological Complementarity – Social spillovers – learning/shame/status • How important are these interdependencies? • Which mechanisms are most important? • If these interdependencies are significant, how can we use them to improve interventions? – e.g. what’s the threshold to push over to the “good equilibrium”?

Context • Rural areas of Tanore district, Bangladesh

Context • Rural areas of Tanore district, Bangladesh • 32% open defecation (cf. 15% nationally) • Increased coverage of basic latrines, but low coverage of improved or hygienic latrines – 19% used improved latrines in 2009 MICS • Study Sample: – 4 of 7 sub-districts (unions), 115 villages, 372 neighborhoods (paras), 18,000 households

Summary of Design: Key Aspects 2 x 2 Demand-Supply Interventions – Control – Supply – Demand – Demand + Supply Supply Side Push None Demand D+S Demand Only Generation (CLTS or “LPP”) None Supply only Control

Summary of Design: Key Aspects 2 x 2 Demand-Supply Interventions – Control – Supply – [LPP+Subsidy] – [LPP+Subsidy] + Supply LPP= Latrine Promotion Program, a version of CLTS

Summary of Design: Key Aspects 2 x 2 Demand-Supply Interventions – Control – Supply – [LPP+Subsidy] – [LPP+Subsidy] + Supply – LPP only (Added this fifth group – LPP without subsidies – to separately identify information vs subsidy effects)

Design: Education / Motivation Latrine Promotion Program (LPP) • First barrier may be that households simply are not aware of the importance of sanitation • With WaterAid and VERC, designed LPP, based on Community-Led Total Sanitation (CLTS) • Transect walk, Open Defecation Mapping, Community Discussion • Not just individual benefits: inter-dependency

Design: Education / Motivation Latrine Promotion Program (LPP)

Design: Subsidy • Identify “eligible” (poorer) households – primarily based on land ownership • Lottery for voucher, roughly 75% of cost • Winners could choose from three models – After-voucher price (approx.) USD 7.5, 10, 15 – All meet standard criteria for “hygienic” • Independent lottery for superstructure (“tin”)

Design: Subsidy

Design: Supply • Even if households are motivated and can afford a latrine, there may be supply-side barriers – Information about quality, installation, link to supply • Latrine Supply Agent (LSA) – Community member respected in technical matters – Provide information on where to purchase, how to assess quality, how to install and maintain – Linked to masons

Design: Summary Control No intervention 65 Clusters Treatment 1 LPP Only 50 Clusters Treatment 2 LPP + Subsidy 110 Clusters Treatment 3 LPP + Subsidy + Supply (LSA) 110 Clusters Treatment 4 Supply Only (LSA) 35 Clusters

Effects on Open Defecation

Design: Refinements to Subsidy • Form of subsidy – Early adopter versus Fixed share • Intensity: share of Eligibles receiving subsidy – Divide subsidy clusters into Low , Medium and High intensity: 25%, 50% and 75% of eligibles receive a voucher • Target “Highly Connected Individuals” (HCIs) – In random subset of clusters, we weighted voucher winners towards those most highly connected

Effects of Proportion of Community Subsidized on Latrine Ownership

Ineligibles • 25% of the population was ineligible for the subsidy – Based on landownership criteria • What are the key constraints for the relatively rich? • Did neighbors getting subsidies affect their behavior?

Measuring Social Network Effects • Thought experiment: • Similar HH A and B, one eligible contact each • A’s contact wins, B’s does not (random) • A’s contact is more likely to adopt • Can use this random variation to estimate: – Whether a household’s adoption decision is influenced by the decisions of their contacts – What type of contacts matter.

Access to Hygienic Latrine VARIABLES Playmates Interact Conflict Resolve Technical Access rate 0.579*** 0.564*** 0.710*** 0.711*** 0.270** 0.365*** 0.189** 0.110 among social (0.0852) (0.0950) (0.0723) (0.0841) (0.125) (0.112) (0.0811) (0.0931) contacts Won Latrine only 0.0625*** 0.0686*** 0.0464*** 0.0485*** 0.0989*** 0.0634*** 0.111*** 0.0745*** (0.0158) (0.0187) (0.0128) (0.0145) (0.0179) (0.0191) (0.0194) (0.0197) Won Tin only -0.00631 -0.00241 -0.0263* -0.0268* 0.00733 -0.0184 0.0209 -0.0126 (0.0185) (0.0231) (0.0142) (0.0162) (0.0190) (0.0201) (0.0206) (0.0222) Won both 0.120*** 0.127*** 0.0923*** 0.0943*** 0.159*** 0.122*** 0.175*** 0.137*** (0.0171) (0.0194) (0.0136) (0.0147) (0.0177) (0.0191) (0.0199) (0.0224) Treatment: Low Intensity & -0.00122 0.00372 -0.00286 -0.0147 Fixed Subsidy (0.00844) (0.00506) (0.0126) (0.0187) -0.00644 -0.00722 0.0141 0.0270 Treatment: Low Intensity & Early Adoption (0.00850) (0.00502) (0.0147) (0.0214) 0.00205 0.000494 0.0148 0.0390** Treatment: Medium Intensity & Fixed Subsidy (0.00724) (0.00532) (0.0158) (0.0165) 0.00117 0.00202 0.0416*** 0.0336** Treatment: Medium Intensity & Early Adoption (0.00714) (0.00554) (0.0126) (0.0155) 0.00259 0.00550 0.0433** 0.0325* Treatment: High Intensity & Fixed Subsidy (0.00796) (0.00567) (0.0172) (0.0168) -0.00867 -0.00501 0.00610 0.0258 Treatment: High Intensity & Early Adoption (0.00730) (0.00599) (0.0140) (0.0165) Observations 7,930 7,930 13,607 13,607 13,445 13,445 11,438 11,438

Results • We see very strong effects: • If the share of your contacts with a hygienic latrine increases by 10 pct. pts., your household’s probability of adoption increases by approximately 5 pct. Pts. • Strongest effects from HH where your children play, HH with which you frequently interact. • Weaker effect from person whose opinion you respect, person you would consult for technical advice.

Concluding Remarks • Subsidies targeted to the poorer segments of the community are a useful complement to CLTS-style programming (community motivation, joint commitment) • Affects both ownership and use of latrines • Strong inter-dependencies in decision-making: – Interventions should target communities, not individuals • The relatively rich benefit from supply-side help • 1300 extra toilets got built in treatment areas (relative to control) as a result of these interventions.