Syndicate content

Three Changes to the Conversation on Service Delivery

Shanta Devarajan's picture

IN054S13 World Bank Back in 2003, when we were writing the 2004 World Development Report, Making Services Work for Poor People, we had no idea that it would spawn so much research, innovation, debate and changes in the delivery of basic services.  Last week, we had a fascinating conference, in collaboration with the Overseas Development Institute, to review this work, and chart the agenda for the coming decade.   Being a blogger, I wanted to speak about what WDR2004 got wrong, but some of my teammates suggested I should start by describing what we got right.  So here are three ways WDR2004 changed the conversation about service delivery (what we got wrong will be the next post).

Services are particularly failing poor people.  Outcomes—child mortality, educational attainment, access to clean water, etc.—are significantly worse for the poorest 20 percent compared with the richest 20 percent.  And service delivery fails where poor people live—in rural Uganda, public-school teachers are absent about 30 percent of the time; in Bihar (India’s poorest state), doctors are absent from public clinics 67 percent of the time.  While this point was even in the title of the report, it has taken on particular salience in light of the discussions about inequality today. 

Poor people may be stuck in a low-level political equilibrium because political support for improving services is weak.  The middle- and upper-classes have opted out of the system: they send their children to private schools, pay for high-quality medical care, and have their own electricity generators and wells.  Even the poor may not lobby for change because they have found ways of coping: they buy water from water vendors (at 5-16 times the meter rate); they organize private schools that charge “a dime a day.”  Champaben, the Gujarati woman I lived and worked with as part of my Village Immersion Program, took her sick child to an unqualified, private doctor because the doctor in the free public clinic “was never there.”  When I asked why he was never there, she replied, “Because the rains didn’t come this year.”  If poor people think that absentee doctors are due to a failure of the weather rather than of public policy, politicians have little incentive to campaign for improved public services, and the system perpetuates itself.

Money is not the solution.  Almost all cross-country studies that control for per-capita income find no relationship between public spending on health and education, and health and education outcomes, such as child mortality and primary completion rates.  Within-country studies explain why.  These subsidies accrue mainly to the non-poor: the lion’s share of public health spending, for instance, goes for hospitals, which are located in urban areas.  Even the small share that goes towards primary clinics or schools often “leaks” (in Chad, the leakage rate for non-wage public health spending is 99 percent).  The wage spending goes to teachers and doctors who are absent about 20-40 percent of the time.  When present, doctors in Tanzania and Senegal spend a total of 29 and 39 minutes a day seeing patients; in Delhi, unqualified private doctors (sometimes referred to as “quacks”) give better quality service than the qualified public doctors.  In light of this evidence, if someone suggests that with, say, an additional $57 billion in aid, we can reach the Millennium Development Goals—don’t believe them!

“The solution” is not the solution.  The traditional view, not just of service delivery but development more generally, can be described as, “When you see a problem, fix it.”  If there are 67 million children out of school, get them into school.  Build schools, give scholarships, make schooling mandatory.  If teachers aren’t teaching, pay them more (even though public school teachers in Punjab, Pakistan are paid more and absenteeism increases with teacher salaries); if they aren’t qualified, give them training; and give them bonuses to be present.  If doctors aren’t showing up in clinics, give them better equipment, medicines (and stop them from migrating).  If poor people aren’t getting water, fix the pipes or increase the supply of water.

The 2004 WDR was part of a different view, namely, that the problem is with the system.  We need to ask: why are the children not in school, the doctors not in the clinic, the pipes never fixed?  The answer typically is that incentives throughout the system are misaligned.  Doctors are paid whether they are present in the clinic or not, and there are huge rents to be earned in private practice (people are desperate when their kid is sick).  Furthermore, doctors’ and teachers’ unions are politically powerful, and have been known to resist efforts that increase their members’ accountability. 

The implication is that solving the proximate problem (“the solution”) may not actually address the systemic issues.  Higher pay for teachers will increase the rents that absentee teachers already earn.  There is plenty of evidence that teacher training doesn’t improve performance—because better performance is not rewarded. 

In fact, “the solution” may make matters worse.  It is now widely recognized that, even in countries with high levels of primary enrolment, learning outcomes are very poor.  Yet, calls for universal enrolment (to get the out-of-school children into school) are widely embraced, whereas calls for universal learning (all children must be able to read by age 7, say) are sometimes resisted.  One reason is that enrolment is much easier to achieve than learning.  Another could be that universal enrolment means more public-sector jobs, whereas a universal learning goal could mean more resources for private schools.  Hence, the focus on enrolment may be detracting from achieving what education is all about, namely, learning.

If poor people are stuck in a low-level equilibrium of poor service delivery, if money is not the solution, and even the proximate solution is not the solution, what can be done?  That was what the conference was about, so stay tuned.  Also, for the three things the 2004 WDR got wrong.


Submitted by Olu Ajakaiye on

Thanks, Shanta. I am staying tuned!!! Meanwhile, are the presentations at the Conference available online?

Submitted by Shanta on

Olu: The materials will be posted on the conference website:

Submitted by Iw Potentate on

It seems glaringly obvious what the problems are as well as any solutions. The crux is implimenting the right solutions. If teachers or doctors or whomever are absent AND still getting paid, it is doubtful that the "public" entities know what the right solutions are. "No work-No pay" is a better incentive. Free markets tend to work when given the chance.

Submitted by Birgit on

Thank you Shanta. Excellent post. It is so easy to forget for anyone, including us, our partners and Government counterparts that throwing money at a problem is not the solution. Yet it is so tempting and quickly demonstrates that leadership cares and is trying to do something: "Look, we increased education / health expenditure!". It proves again that with structural adjustment alone or macroeconomic stability, not much will change. Instead, we need to work on improving institutions and better understanding why the micro fundamentals of an economy do not work. And the best of the findings you posted: they do not only apply to developing countries, but also middle- and high-income countries. So we are all together in the same boat to search for better solutions.

Submitted by Shanta on

Thanks, Birgit for those comments. You're right. These accountability failures occur at all levels of income, as I'm finding out in the Middle East and North Africa. The problem, though, is that in low-income settings, the costs are unimaginable--you could lose your child, or your child may never learn to read and write. We have to fix this! Shanta

Submitted by Ajith Kumar on

There is an urgent need to view all development from the outcomes angle. For too long, it has been about how much money has been spent, i.e. inputs, and processes. The fact that water is not delivered, sanitation is poor, kids don't learn or people are still sick, is a result of this lopsided emphasis. The resistance to focussing on results is not only from the public sector providers (one can understand their incentives), but, perhaps not so surprising, also from non-government development practitioners, donors and civil society. Demanding or writing a cheque is the easiest option, and the least complicated!

Submitted by Shanta on

Ajith: You're right. The problem is that outcomes are harder to measure than inputs and you don't have as much control over them. It's much easier, as a policymaker or development practitioner, to talk about how much money you've spent than how many more kids can read and write. But we have to resist the temptation to take the easy way out. We could start by insisting that any statement about money should be accompanied by a statement about precisely what outcomes that money is supposed to achieve--and then hold the people concerned accountable for those outcomes. For instance, whenever the World Bank's board approves a loan, we issue a press release. It includes some language about outcomes, but rarely is it precise. Nor is there a mechanism by which the readers of the press release can check whether those outcomes were achieved. We can change this! Shanta

Submitted by Peter Quest on

What is the future of Africa, thank you Shanta. I will be happy to learn on any possible strategies that Africa can do to improve on employment especially for the youth. Case in hand, Kenya.

Submitted by Shanta on

Peter: We just released a report on Youth Employment in Africa ( Best, Shanta

Submitted by Jakob Fredensborg on

Thanks Shanta for the sum-up and selected highlight. One thing that struck me is that the WDR and your breakdown above, only looks at the issues of lacking or poor service delivery from the perspective of hard incentives and completely ignores how important soft incentives can be. Pay and promotion are not the only motivators in ensuring performance. How people feel associated to the job and public opinion can also play important roles in motivating teachers and doctors to show up and do their job. If you are the subject of national shaming and constant blame for poor outcomes, higher pay or a different title, won't help you to mentally set yourself up for the job. Nobody wants to be told that they are horrible at what they do and be blamed for poor performance.

I would also argue that problems regarding lacking performance of public servants are not confined to developing countries, they are also present in developed countries albeit in different form and scope. Extending the scope of the argument; I am sad that so little is being "learnt" from how developed countries tackle or tackled these problems. These experiences are usually only taken into consideration on the margin. Instead, the literature always, only, present cases from developing countries, furthering "South-South" exchange. Why is it that only experiences from places where systems barely function are utilized to show other countries with equally defunct systems how to overcome their challenges?

I recently read that behavioural economics is making its way back into the Bank, very good. I should like to see some of the tenants of the discipline be utilised in explaining why public servants do not do their job, looking beyond the wage paradigm. Service delivery remains one of the central functions of government (regardless of how effectuated or organised) and its improvement is crucial to tackling some of the global challenges our species still face.

Submitted by Shanta on

Jakob: Thanks for your thoughtful comments. There are two parts to what you call "soft incentives." One is intrinsic motivation: people show up and do their job because they believe in it, sometimes because of deeply held religious beliefs. We discussed this in the 2004 WDR and some of the evidence is captured in Ritva Reinikka's and Jakob Svensson's paper, "Working for God" ( But there is a second aspect, which has to do with professional norms. You go to work because, if you don't, your peers will not respect you (and vice-versa). As I mention in my next post, I think this is what differentiates much of service delivery in developed and developing countries. The question is: how to you get from one set of professional norms (where people don't abhor absenteeism among their peers)to another? The forthcoming WDR on "Mind and Culture" will, I hope, have some answers. Regards, Shanta

Submitted by Jakob Fredensborg on

Dear Shanta,

Thank you for your reply! I will wait for the forthcoming WDR with anticipation and excitement. I do however feel that you only partly answered my question. Intrinsic motivation and professional norms develop over time, but they do not do so in a vacuum; public opinion (especially in key demographic groups) matter as well, as does the country context.

In the early 90's a combination of events (where decreased learning outcomes was but one) led to a sharp decrease in public opinion of teachers in my own country (Denmark). Colleges had great difficulty in attracting would be teachers, so much that demand by far outweighed supply; gloomy forecasts from the national statistics office questioned whether sustainable supply was even possible. Responding to the crisis the Socialdemocratic government tried to manipulate hard incentives to attract more young people into teacher colleges to no avail. It was only, when an investigation amongst teachers themselves (conducted by the unions and municipal governments association) uncovered that the seriously bad PR status of the teacher profession was the main obstacle to boosting supply, that the issue could be addressed. Via a combination of public awareness campaigns, increased dialogue with parents and other key demographics, as well as reformation of hard incentives, the situation was remedied (of course looking at PISA scores and accompanying surveys there is strong evidence that the crisis put Denmark in the other ditch).

Let me be clear though, I am not at all neither advocating a direct comparison of 90's Denmark and '14s developing world nor am I advocating a cold shift towards looking at soft incentives. I guess I am questioning whether the dialogue being conducted between the World Bank (as well as other development partners in client countries) contains the correct blend of issues or is there an over-emphasis on looking at problems from a single perspective? it seems, the solutions discussed (at least according to the experience I have) centres around wage bill management, teacher qualifications and availability of teaching materials. Usually side issues are: public investment management as well as governance issues. Of course, looking at the institutional set-up in client countries all those issues make a lot of sense. Yet I can't help escape the feeling that this has been the mantra for a long time, and there seems comparatively little to show for it.

Kind Regards,


Submitted by sadashivan on

I visited one village in tamil nadu streets and roads had lights but not houses as they could not afford to pay bill. School were there with one or two teachers. Is this not eye wash by political structure? Poor's first priority is source for survival, as long as this is missing no impact of the services. Many villages do no have hospitals and doctors are scarce. Why can't have mobile low cost hospitals with Bsc nurses with training to serve villages. I have experienced myself visiting many villages and described in my wesite. We need will and accountability to fulfil

Submitted by mehmet on

You are saying no more aid for MDG Then could you tell me how it will work with out money ?

Submitted by Ernesto Bautista on

Very interesting analysis and observations. I agree that addressing the proximate problem does not completely provide the long term solution. the challenge among practitioners is addressing the systemic problems, many if not all of them involve governance and political issues which require long term transformation and mobilization of multi-stakeholders support. and since most donor support is short term, the focus tend to be on those areas which are visible and more manageable within the timeline. it would be useful to present good examples of how some countries have address some of these systemic issues and the lessons we can draw from these experiences. Regards.

Add new comment