An Open Letter in response to the World Development Report 2015

Printer-friendly versionPrinter-friendly version

It is with disappointment and bewilderment that we, the undersigned, write this letter in response to the publication of the latest World Development Report Mind, Society and Behavior.

In the lead up to its publication, Robert Chambers of the CLTS Knowledge Hub at the Institute of Development Studies and Frank Greaves of Tearfund UK were invited to advise on a contribution on Community-led Total Sanitation (CLTS). This was being considered for inclusion in the report as a key example of a behavioural change process. Robert and Frank went to London to take part in a video conference with Washington and drafted text on CLTS, the CLTS approach and how and why it works as an input for the report.

The Report draws on this in describing CLTS on pages 17 and 152-3. However, it finds (page 17) that where CLTS was combined with subsidies for toilet construction, its impact on toilet availability within households was much higher and concludes that these findings suggest that: 'CLTS can complement, but perhaps not substitute for, programs that provide resources for building toilets’. We are shocked and puzzled to read this. It is a damaging misrepresentation of CLTS. It is surprising that the authors of the report, after consulting with Robert Chambers and Frank Greaves, then wrote text on CLTS which they did not check with them for representativeness before going to publication. Had they done so, this would immediately have been corrected. They also did not consider it necessary to further their understanding of CLTS by referring to the considerable literature on the subject or corresponding with Dr. Kamal Kar, the pioneer of the approach.

CLTS is a sustained process, much more than just a triggering exercise, dramatic and critical though triggering can be. For CLTS to go to any scale, it requires an environment which is free of hardware subsidies. It is a worldwide experience that provision of hardware subsidies inhibits self-help and does not lead to collective behaviour change, but only to many toilets which are often not used, used for other purposes, or dismantled. These experiences with hardware subsidies and the requirement of a subsidy free policy environment as fundamental for achieving CLTS at scale, has not been understood. More than 20 Governments have adopted CLTS as national policy. Over 30 million people have been estimated to be living in open-defecation free communities as a result of CLTS implemented in enabling environments which are free of hardware subsidy. On the other hand, CLTS spread has been very limited wherever hardware subsidies prevail. The WDR authors, instead of consulting the very extensive CLTS literature (see e.g. the CLTS Knowledge Hub website), have based their conclusions on two Randomised Control Trials, one carried out by the World Bank, and one published in PLoSMed.

The first study is of the Total Sanitation and Sanitation Marketing (TSSM) programme in Indonesia. It was an “impact evaluation” with a baseline in late 2008 and an endline assessment just 2 years later in September 2010. It found that while “treatment communities” that received CLTS triggering showed higher numbers of households constructing toilets than “control communities”, the construction activity was mainly driven by non-poor households. The naïve and simplistic inference that this was somehow a result of simply implementing CLTS processes is very far from the truth. CLTS on its own will always target an open Defecation Free (ODF) environment and not toilet construction by some. These results have to be seen in the context of combining CLTS and sanitation marketing and the fact that this survey was done very early on during programme implementation. TSSM-supported market research by Nielsen (2009)1 had found that poor East Java households aspired to the pour-flush toilets with a ceramic pan, which cost twice as much as what the poorest households were willing to invest. Both the Nielsen study and action research in 20 districts (Mukherjee, 2012)2 found that the rural Indonesians preferred to defecate into rivers until they could afford such a toilet, because they believed that unimproved pit latrines are far more unhygienic than defecation in rivers. The TSSM response to the 2009 market research was to develop the desired product options at lower costs and build local market capacity to deliver them at prices the poor could pay. This took time. Thus poor consumers could find such products in local markets in only a few out of the 29 districts by 2010. The hurried endline assessment in September 2010 failed to capture the impact of improved supply of affordable toilets. That large numbers of the poor did gain access to improved sanitation following both CLTS and sanitation marketing interventions of TSSM – is amply illustrated in the 2200 communities in East Java being verified as ODF by 20113, when the TSSM closed and was replaced by the Government of Indonesia’s national Sanitasi Total Berbasis Masyarakat (STBM) program modelled on the TSSM approach.

The RCT study conveniently asks the poor why they have not constructed toilets and the response needless to say is that the cost is not affordable! There are countless surveys that will report the same findings. And the official knee-jerk reaction is to recommend a policy to subsidize toilets, without any attempt to understand why the poor say toilets are unaffordable, but find multiple cell phones per household entirely affordable !! (another finding from the 2012 action research: Mukherjee et al ,2012). The end result, as studies in India regularly bring out, are toilets on paper only and/or unused toilets on the ground. Clearly surveys of this kind are designed and carried out with little understanding of the importance of behaviour change in sanitation. On the other hand, surveys of ODF communities (where even the poor have toilets) in a subsidy free policy environment (as in Indonesia, since 2008 and in parts of India (Ahmednagar and Nanded districts of Maharashtra in 2004 and 2005 and the whole of Himachal Pradesh between 2006 and 2012), show that the toilets are constructed because the understanding that everyone's excreta must be confined safely for all to reap the benefit, has been internalized and where necessary the community has come forward to help its poor and less able. With the CLTS approach now present in over 60 countries, it is not difficult to access literature which brings out evidence of collective behaviour change being sustained by communities that recognize sanitation as a problem to be addressed by them on their own. Subsidy from above inevitably focuses attention on toilets and not behaviour change. It conveys the impression that the problem of OD is external to the community and breaks up the community into those who need to confine their excreta (those entitled to toilet subsidies) and those who need not!

The second study is of the Government of India’s Total Sanitation Campaign (TSC) in Madhya Pradesh (Patil et al 2013)4. Central to CLTS is collective behaviour change in a subsidy-free environment. The TSC in practice was based on hardware subsidies to individual households. These subsidies were a major inhibitor of collective behaviour change. The study is based on a comparison of 40 treatment villages and 40 controls. The treatment was not CLTS but described in the study as ‘CLTS-like’. What it did was use some ‘CLTS methods’ – only triggering is mentioned - together with the TSC subsidy. This resulted, as would be expected, in only very modest change - 19 per cent more toilets (32 per cent versus 13 per cent). At the end of the period 74 per cent of adults and 84 per cent of children in the treatment villages were still practising open defecation. 41 per cent of intervention households with improved toilets reported that adults were still practising daily open defecation. No villages became open defecation free. The WDR’s conclusion that ‘where CLTS was combined with subsidies for toilet construction, its impact on toilet availability within households was much higher’ is invalid. To justify that statement would have required comparison with a control which received CLTS without subsidies. But the control communities did not receive any treatment (except 10 out of the 40 towards the end of the period) and if they had it would have been with subsidy as national policy. The implication that this research showed CLTS does better with subsidies could not be more wrong. It is contradicted by world-wide experience in many countries which have adopted CLTS with great success. Many of these countries have used CLTS to achieve thousands of ODF communities through an enabling environment of national policies which abolish hardware subsidies. The evidence of the vast scale of the experience of these countries is more than anecdotal.

In privileging Randomised Control Trials over the mass of evidence and recorded experience concerning CLTS that is in the public domain, in failing to distinguish between the TSC and the radically different CLTS, and in neglecting to check out the text and conclusions in the Report before going to press, the WDR fell short of the levels of professional behaviour and rigour we expect from the World Bank. Not only has CLTS been misrepresented, but the credibility of the WDR as a cutting edge, thorough and evidence-based publication has been undermined. Given the focus on mind-sets and behaviour, it is ironic that the report does not critically reflect on the mindsets and behaviours that underpin the report itself. In this case they have led to an erroneous and damaging conclusion.

Signed by
Robert Chambers and Petra Bongartz, CLTS Knowledge Hub at the Institute of Development Studies, UK

Deepak Sanan, CLTS Foundation and Additional Chief Secretary to the Government of Himachal Pradesh, India

Dr. Nilanjana Mukherjee, independent sector specialist, India

Frank Greaves, WASH Adviser, Tearfund, UK

  • 1. Nielsen (2009). Total Sanitation and Sanitation Marketing Report. Prepared for the World Bank Water and Sanitation Program.
  • 2. Mukherjee N. et al (2012) Achieving and Sustaining Open defecation Free Communities : Learning from action research in 80 communities in East Java. Full report on http://www.wsp.org/sites/wsp.org/files/publications/WSP_Indonesia_Action_Research_Report.pdfMukherjee
  • 3. Government of Indonesia’s ODF verification Guidelines require 100 % community households to be owning and using improved sanitation facilities, besides other criteria. That poor households in 2200 ODF communities gained access to improved sanitation , along with their non-poor neighbours, is a verified statistic. Through the TSSM project a total of 1.4 million people gained access to improved sanitation during 2008-11 in East Java province, as verified by the Ministry of Health, Government of Indonesia. CLTS was one of TSSM’s interventions, supplemented with Sanitation Marketing and Enabling Environment building. For details see Results, Impact, and Learning from Improving Sanitation at Scale in East Java, Indonesia (WSP, 2013), on www.wsp.org
  • 4. Patil, S R, Arnold, B F, Salvatore, A et al. (2013) A Randomized, Controlled Study of a Rural Sanitation Behavior Change Program in Madhya Pradesh, India. The Water and Sanitation Program of the World Bank
Date: 18 March 2015
Topics: 

Comments

Submitted by petra on

The World Bank has issued an official response:

Robert Chambers, Petra Bongartz, Deepak Sanan, Nilanjana Mukherjee, and Frank Greaves have
posted a letter expressing disagreement with the discussion of Community-led Total Sanitation
(CLTS) in the World Development Report 2015. The authors are development professionals
whose work on sanitation and other areas of development practice is rightly viewed as original
and important. For this reason, and as they point out, the WDR team consulted with them during
the preparation of the Report. As their own body of work describes, CLTS is an approach to
sanitation that uses emotions and social norms to reduce open defecation and support other
hygiene practices. It is exactly the kind of innovative development intervention that the WDR
aims to highlight.

At the same time, one of the main messages of the Report is that small changes in framing and in
the environment can have large consequences for behavior. Context is enormously important.
What works in one place may not work in another. For that reason, continuous testing of
development interventions and adaptation to design are crucial. There are many ways to evaluate
the impact of an intervention, but the Report takes the view that randomized controlled trials
(RCTs) are particularly important for causal inference and establishing proof of concept.
The Report discusses RCT evaluations of CLTS available at the time of writing. These had
occurred in Indonesia and India. A graph in the Report (Figure 8.4) shows the resulting decrease
in open defecation, and the text summarizes the findings. CLTS without subsidies to toilet
construction reduced open defecation in both Indonesia and India, but it only slightly increased
toilet construction in the one country—Indonesia—in which data were available for a pure CLTS
program: 3 percentage points more households in Indonesia built toilets in treatment
communities than in control communities.

The Report takes care to emphasize that the inference regarding the need for subsidies to toilet
construction was based on RCTs in only two countries. Other countries, and indeed other
communities within those countries, may be different. The Report does not make any general
causal claims that the CLTS approach without subsidies to toilet construction cannot be
effective. Of course, it may be effective elsewhere, and CLTS is indeed a promising and
innovative intervention. That is why the Report describes CLTS in detail. Still, the RCTs that
were available suggested that toilet subsidies can play an important role, at least in some
contexts, in eliminating public defecation.

After the publication of the WDR, a new working paper with results of CLTS interventions in
two additional countries, Mali and Tanzania, was published. The results of that study are in line
with the other RCTs. The study states that “Our results suggest that stronger interventions
that combine intensive health promotional nudges [CLTS] with subsidies for sanitation
construction may be needed to reduce open defecation enough to generate meaningful
improvements in child health studies.” The 2015 study estimates that ending open defecation in
villages where everyone defecates in the open would increase child height by 0.44 standard
deviations.

There is a need for more research on the impact of CLTS in many countries, and on its
interaction with subsidies, information campaigns, and other more traditional interventions. The
WDR team does not mean to prejudge the findings of future research. It is our hope that the
WDR will increase investment in the needed studies, including studies based on productive
collaboration among longstanding members of the CLTS community and researchers drawing on
a rigorous research methods.

References
Cameron, Lisa A. and Shah, Manisha and Olivia, Susan, “Impact evaluation of a large-scale rural
sanitation project in Indonesia” (February 1, 2013). World Bank Policy Research Working Paper
No. 6360. Available at SSRN: http://ssrn.com/abstract=2217339

Gertler, Paul, Manisha Shah, Maria Laura Alzua, Lisa Cameron, Sebastian Martinez. Sumeet
Patil. 2015. “How does health promotion work? evidence from the dirty business of eliminating
open defecation, “ NBER Working Paper 20997.http://www.nber.org/papers/w20997

Patil, Sumeet R., Benjamin F. Arnold, Alicia L. Salvatore, Bertha Briceno, Sandipan Ganguly,

John M. Colford Jr, and Paul J. Gertler. "The effect of India's total sanitation campaign on
defecation behaviors and child health in rural Madhya Pradesh: a cluster randomized controlled
trial." PLoS medicine 11, no. 8 (2014): e1001709. Available at:
http://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1001709

You can download a pdf of the response here

The following extracts are copied directly from https://www.poverty-action.org/node/8771 (to see the full text of the story). The published journal article is gated at: http://www.sciencemag.org/content/early/2015/04/15/science.aaa0491.

Subsidies Key in Improving Sanitation, New Study Finds
Apr 16/15 | Announcement |


April 16, 2015, NEW HAVEN, CT – With poor sanitation estimated to cause 280,000 deaths per year worldwide, improving sanitation is a key policy goal in many developing countries. Yet governments and major development institutions disagree over how to address the problem. A new study released in Science today found that in Bangladesh, a community-motivation model that has been used in over 60 countries to increase use of hygienic latrines had no effect, yet latrine coverage expands substantially when that model is combined with subsidies for hygienic latrines targeted to the poor.

...

Researchers randomly assigned 380 neighborhood communities, or 18,254 households total, to one of four groups. Villages either received a community motivation program, subsidy vouchers with the community motivation program, information and technical support, or none of the above. By comparing outcomes in latrine coverage, investment in hygienic latrines, and open defecation between the groups over time, researchers were able to compare the effect of the different approaches.

The subsidy vouchers, which were only provided to a random subset of households in the second group through a public lottery, could be redeemed for a 75 percent discount on available models of latrines, priced (after subsidy) from $5 to $12. The households were responsible for their own transportation and installation costs, and the richest 25 percent of households were not eligible for vouchers. The community motivation program, called the Latrine Promotion Program (LPP), was modeled after “Community-Led Total Sanitation”, which focuses on behavioral change and community mobilization in eliminating open defection. Such programs have been implemented in over 60 countries worldwide.

Researchers found that the community motivation model alone did not significantly increase adoption of hygienic latrines or reduce open defection relative to the comparison group, nor did providing information and technical support to community members. However, the subsidy had substantial effects when coupled with the community motivation program, increasing hygienic latrine coverage by 22 percentage points among subsidized households and 8.5 percentage points among their unsubsidized neighbors.

...

These results counter the concern among many development practitioners that subsidies undermine intrinsic motivation. Rather, this research shows price is a primary barrier, which is consistent with a growing body of research on adoption of health products.

 

Submitted by petra on

Hi Stephen,

I think it is regrettable that the term 'subsidies' is being used in this context which confuses the issue. It would be better to speak about the vouchers and how they work. In my mind, the voucher system is different from giving households upfront subsidies. India is a case in point that shows that individual household hardware subsidies do not work.  I believe the vouchers could be a good way forward to ensure that the poorest or marginalised in communities are included and can access improved sanitation options. In our forthcoming book on CLTS and Sustainability, we will have a chapter on equity, inclusion and financing in which Andy Robinson writes about the voucher system.

Hi Petra, thanks for the reply and interesting observation. However the vouchers are clearly still a form of subsidy, even if as you say they are a different mechanism of subsidy. I look forward to the book chapter if this goes into more detail. The original letter and also Robert's post on "Minds set in Washington DC" give the impression that no forms of hardware subsidy are compatible with CLTS. If in fact some forms of targeted, partial, voucher-based subsidy are compatible and effective in some contexts then we need to have a debate on what these are and how this could work.  

Submitted by Depinder Kapur (not verified) on

India WASH Forum Newsletter # 40

Maharashtra Rural Sanitation Success: Lessons for Swachh Bharat Mission ODF Benchmarking and for making the CLTS Approach more effective for India

WSSCC organized a workshop in collaboration with the MDWS, for identifying ODF indicators for the Swachh Bharat Mission(Rural) in Pune in April 2015. The workshop had a very good presentation from the ex Rural Development Secretary Mr.Thakre who was involved closely with the Maharashtra state sanitation programme since 2000.  Sanitation coverage achievement in Maharashtra from 2000-2007, is perhaps the most successful CLTS inspired work in India ever done.

The achievement of Maharashtra sanitation programme was the scale of coverage and people engagement in taking sanitation uptake significantly high. Maharashtra experience married the traditional CLTS with promotional and motivational approaches and delivered results using both individual incentives and community incentives(something that CLTS radicals dismiss).

The Maharashtra sanitation success as shared by Mr. Thakre, was achieved in a 2 phase programme mode with full backing at the state level(political and administrative). In the first stage there was basic motivational work done at scale followed by community incentives(prize money at GP, Block, District, Division and State level). In the second stage the targets for community incentives were made more realistic( ODF achievement rewarded in stages and not 100%) and merged with other rural development programmes.

The Maharashtra experience is perhaps ten steps ahead in terms of the richness of experience and results at ODF level, than the traditional CLTS approach in sanitation advocated by some. In terms of both scale and sanitation coverage(up to 7,000 Gram Panchayats out of 28,000 Gram Panchayats in Maharashtra). It is also perhaps the most under reported success story by the CLTS groups worldwide and even by other international WASH agencies.

What and how of the Maharashtra sanitation programme success

Maharashtra had only 6% rural sanitation coverage till 1996 that was achieved with an expenditure of Rs.6 crores. Under a new initiative from 1997-99, Maharshtra spent nearly Rs.750 crores( @Rrs.4000/toilet subsidy) but the results were unsatisfactory. Nearly 55% toilets were not used and 23% toilets were reported as misused.
This led to a rethinking of the sanitation focus and a new state initiative. Focus of the state sanitation vision was on five points: Safety of women while defecation in open from animals, Health impacts, Ambiance in terms of smell and clean village, Dignity of women and Economy in terms of excreta as bio fertilizer.   This five fold strategy had an acronym SHADES.

Operational strategy of Maharashtra sanitation programme involved the following components;
• Motivation of all PHED and Rural Development Staff on sanitation promotion
• Toilet construction demonstration campaigns in the villages to raise awareness and dispel doubts
• A 3-4 month long mass media campaign across the state using all influential people, celebrities from Maharashtra that common people admire
• Introduction of performance incentives(cash awards) in terms of a competition for Gram Panchayats(GPs) to achieve ODF status.
o Cleanest GPs at Block level
o Cleanest GPs at District level
o Cleanest GPs at Division level
o Cleanest GPs at State level
• Joint inspection teams for clean village verification: from departments of Health, Education, NGOs, Media and senior bureaucrats. Ten point verification process, 105 total marks, 25% of the toilets verified in a sample. Verification done four times to identify the best Gram Panchayat at state level.

Maharashtra sanitation programme influence on national sanitation programme

The success of the Maharashtra sanitation vision and strategy of early 2000, resulted in the incorporation of this approach in the national flagship sanitation programme called Nirmal Gram Puraskar in 2002-03. However no one strategy or approach works for a large country like India with a diverse demographic and economic context. There is no magic solution of instant behavior change using any one sanitation promotion approach.

The Maharashtra successful sanitation experience evolved from 2007 onwards into a rural development programme. This was because the incentive/awards driven state programme had reached a plateau with the same set of Gram Panchayats vying for the state incentives and the other GPs feeling demotivated. Because the challenges in different villages are different and some face more challenges than others. Under the Eco Development programme of Maharashtra, where sanitation was one of the sub programme component, the state provided early recognition and incentives to GPs that reached 70-80% ODF status, so as to motivate the more challenging GPs. This practical approach, of marrying certain components of the traditional CLTS approach(Triggering and Health Impacts awareness raising) combined with subsidy/ incentives at less than 100% ODF status, led to a new vigor in rural Maharashtra sanitation programme upscaling.

Important lessons from Maharashtra sanitation success;
• No single sanitation approach works in the long run
• Competition based incentives alone do not work
• Individual subsidy/incentive is important. But should be given after achieving results(toilet construction and usage). Individual incentives can be married with community level ODF incentives. There is no contradiction in giving individual incentives and also securing ODF communities. Individual incentives for toilets are needed to ensure the poorest get some support to build toilets.
• Regulations and motivation are important components of a toilet and promotion behavior change programme
• Sanitation should be part of a longer term village development plan. Merge with more and more development schemes, so as to extend the behavior change window of a sanitation programme over 4 to 5 years and not just a one shot CLTS intervention that may fail.
• Incentives and recognition for significant sanitation improvement(coverage and usage of toilets both), need to be given at less than 100% ODF status for a GP.
• Use a combination of carrot and stick approaches: withholding other government welfare schemes, perks for government staff to promote sanitation, imposing taxes and regulations
• Don’t give total GP level incentive/subsidy for sanitation to an NGO in one go. Give it in intermediate stages of percentage sanitation coverage and usage, and not at the end for 100% ODF.
• De politicise the government programme to make it a social development programme. Maharashtra named it after Sant Gadge Baba and not after a politician.
• Without political commitment at the state level that promotes swachhta as a desirable goal and ably assisted by a bureaucratic leadership – it is impossible to achieve ODF at scale in a programme approach.