Decentralisation Blues

The World Bank’s Shanta Devarajan reckons that there needs to be real political demand for capacity building to truly transform dysfunctional developing country institutions, and avoid the trap of isomorphic mimicry. He is surely correct in this assertion, but I fear the rose-tinted spectacles return when he advocates the benefits of decentralisation:

“One reason [for doubting local authorities capacity to manage financial resources] may be that no one has given local authorities the chance to deal with funds.  There may have been no demand for financial management at the local level because the central government has told you what to spend.  If you give them the chance to make the decisions, then they might actually build the capacity or hire that capacity because it’s something they can decide for themselves.

Moreover, if the local governments are accountable to the local population, they will have to build capacity really fast. They can no longer put the blame on central government if things don’t work well.”

To be fair to Devarajan he does qualify his enthusiasm with the requirement that local governments should be accountable to the local population. The trouble is that in the decentralisation that I have witnessed in least developed countries I have never seen much sign of that condition coming true, and certainly not any evidence of it leading to substantially increased capacity. Instead, where local official venality and low capacity are the rule rather than the exception, as is the case around here, such pushes as there are to improve service delivery come from above, although even here there’s a lot more political rhetoric than practical action. Decentralisation thus leads to temporary petty fiefdoms that can go largely unmolested so long as performance is not notably worse than elsewhere in the country (and sometimes even when it is).

This rose-tinted view of decentralisation is not restricted to ill-informed denizens of the embassy district and big donor agencies. I think we field operatives can sometimes be equally guilty in assuming that just because community leaders are that much closer to their constituents they will therefore be that much more responsive to their needs, or that bad leaders will get voted out of office. For even at the community level base party politics and local rivalries so often trump technocratic concerns of executive competence.

I would suggest that demand for good services is predicated on at least some idea of what they should look like and sense that they are properly due. By this I mean not just that a community should want the service, and be prepared to air their grievance to anyone who cares to come by and ask (very common around here), but that their sense of justice should be inflamed at the breach to the perceived social contract, and, as an aggrieved party, they are prepared to act seriously to obtain redress. Consuming many government services, e.g. sending one’s children to the local school, might be a largely passive undertaking, but service quality depends upon a community’s aptitude to pursue their due proactively, and in turn for the rest of society not to regard such direct action as disproportionate.

Some would translate all of that as the need for a large middle class, and trot out that old canard about democracy not being workable without one. I am prepared to be a bit more optimistic than that, but I think we should be cautious about expecting demand for service provision to drive improvements in local capacity. There will always be counter-examples, usually championed by exceptional local leaders, but countrywide I wouldn’t pin your hopes on anything other than slow progress, with plenty of steps back interspersed with the forward ones. Social change is slow and messy.

Hat Tip: Lee Crawfurd

Advertisement

Why I’m a Millennium Villages sceptic

Last week Jeffrey Sachs set out a robust defence of his brainchild, the Millennium Villages Project, although, as Tom Murphy pointed out, it was somewhat low on detail. I’ve never knowingly been near a Millennium Village, but my own experience causes me to doubt the lasting legacy of the MVP, at least in countries with similar problems to where I work.

First the good news, Sachs took on some of this detractors by saying that the MVP was as much about developing systems to improve service delivery (and hence attainment of the Millennium Development Goals) rather than just,  per se, achieving the MDGs in the targeted villages. I’m a big fan of systems approaches, so this gets the thumbs up from me. Systems are definitely more easily replicable and scaled up than individual projects that focus simply on the needs of its target area.

Now for the bad news, systems are not automatically and by definition scalable. A system that works well at one level may not work well a wider scales due to unanticipated problems and bottlenecks. Ben Ramalingam recently blogged on exactly some of the new challenges that occur as one scales up. This doesn’t mean that the original system was designed badly, but simply that good systems management takes an iterative approach, making tweaks and improvements as we go along (what I term the KISI approach).

But that is not the biggest problem that I see. Even the best designed systems need to interact with things outside their control, in particular people; indeed I suspect that the MVP has people playing integral roles at every step in the way (i.e. that mostly what we’re talking about here is systems for organising human work). A system’s output is constrained by the quality of these interactions. In short, as any good businessman knows, you need competent and motivated staff to deliver a high quality of service. And that is where so much service delivery in developing countries falls down, with last mile service delivery particularly badly managed. Unfortunately short-term, local solutions to this are not scalable.

The problems are legion, and not all a result of poor education amongst the workforce. I know some excellent and (when you consider what they are up against) surprisingly motivated local civil servants. But the overall system drags everyone down. Sure you can tinker at the margins with systems to improve paper flow (mostly in local government around here, we’re talking about paper flow), but the elephant in the room is an unmotivated and unsackable workforce.

Of course this problem will apply at the MVP sites, but there you also have a massive aid effort with lots of expat technical advisers and a high level of political interest. I’ve noticed around here, normally sloth-like civil servants who won’t even sit in a meeting without a generous per diem rush around like lauded socialist workers striving manly (or womanly) in the name of their country when a bigwig is due to visit, working into the night and through weekends, all without any per diems.

Thus I fear all the achievements of the MVP will wash up against the great brick wall that is a change resistant bureaucracy. Once the high level of funding, all the expat TAs, and the high level political interest have withdrawn we’ll be back to business as usual, and the MVP will be neither sustainable in the selected pilot villages nor scalable. Maybe this will not apply everywhere, but I would wager a decent sum that it will happen here. The community contributions which Sachs highlights may also be much harder to be elicit when it’s just government staff doing the asking.

The MVP has a laudable goal, and even as an experiment, the idea of resolving various systemic problems in service delivery is a worthy one that definitely deserves some experimentation; marginal changes can lead to marginal improvements, and, as a by product, perhaps a marginal improvement in government staff morale. But if Sachs wants to take a systems approach to achieving the MDGs maybe he should have looked at HR management reform in developing country civil services. It’s a Herculean task to be sure, that, around here at least, the World Bank has been striving at vainly for some time. But until you resolve that problem I fear these sorts of big push attempts to transform service delivery and hence quality of life in developing countries will always be at least one more big push away from succeeding.

The Converse of Accountability

A few years ago a donor asked us (informally) whether we would be prepared to partner up with another local NGO that they were also considering supporting in order to reduce their transaction costs. (An illusory goal – in effect they were asking to pass their transaction costs on to us.) We wanted the donor’s money, so we said yes we’d be prepared to partner in that way, but due to political concerns and doubts about our proposed partner’s capacity, we said that we would not be held accountable for what the supposed partner did or did not do, or how they spent their money. The donor quickly realised the sense of what we were saying and dropped the idea.

Accountability is one of the major pillars of good governance, and a regular buzzword in development-speak. With responsibility comes accountability. But, as the above example shows, with accountability also needs to come responsibility. We rightly object to being held accountable for something we cannot control.

And yet in efforts to improve local governance in developing countries I fear this is exactly what we may be in danger of doing. Around here at least, service delivery over the last mile is often poor to abysmal, so pushing local officials into being accountable for the performance of their departments seems like a good idea. But department heads have remarkably little control over their juniors. They have minimal say over who is hired, and even corrupt officials are very rarely fired, just transferred to another province, to become somebody else’s problem. Patronage is also a problem; disciplining some local bigwig’s distant relative is always going to be a dicey proposition. So officials lapse into the inevitable semi-comatose apathy.

Civil society advocates for accountability should take note. So should donors. My suggested rule of thumb: never give money to someone who lacks professional authority over his/her own staff. Unfortunately that would rule out a lot of government to government aid.

Lighting the Way

I really hope these guys succeed with their prize-winning solar-powered lanterns. Fighting global warming will not always go together with improving the lot of the poor, but this looks like a great win-win. I hope they get the financing they need to scale up (CDM anybody?), and that it doesn’t get stuck in just a few niche markets. The key thing will  be to get the distribution right. Urban areas will be easier than rural, and are the sensible place to start.

Price sensitivity is also a concern. At $10 they are pretty affordable, but at the upper quoted price range ($45) they might struggle to get buyers amongst the poorest, as they then constitute a capital investment. Although the lamps may rapidly save owners more  money than their original purchase cost, the poor typically buy their kerosene at $1-2 a time. $45 is a lot of cash for someone that poor to part with at a time. However, local entrepreneurs (i.e. little shop owners) who know their customers could maybe sell them on a kind of hire-purchase basis if they were able to reach a similar arrangement with their distributor. I’m sure there must have been similar kinds of marketing experiments with other products aimed at the poor, so hopefully D.Light can learn those lessons and apply them well.

Inventing the device is a big step forward, but, in my view, the greater challenge lies ahead: how to get them into the hands of all the world’s poor? As others have commented before me, if Coca Cola can do it, so can they, but we shouldn’t pretend it will be easy. Unlike the iPhone, these kind of devices will not sell themselves.

REDD: to transact or not to transact?

I’ve had various conversations with friends and colleagues over the last few days on the subject of REDD. Most commentators seem to agree that REDD is one of  the few things countries did seem able to agree on at Copenhagen, and that some kind of international agreement on it does seem likely to happen. This is all to the good, our concern has instead been the architecture at the sub-national level. This will be driven to some extent by the framework set out at international level (and I won’t pretend to be up to date on all the latest discussions) but individual countries will also have a certain amount of freedom to choose their own solution.

Essentially there are two different kinds of architecture which can be adopted:

1. Transaction based. In this model individual land owners / forest managers make  some changes to reduce deforestation and/or its cousin forest degradation. These changes and the amount of carbon saved are verified, and then the land owner sells these credits (known as Certified Emissions Reductions, or CERs) to the government in their country, who then sell it on to the international market.
2. Fund based. In this model the government undertakes a variety of actions to reduce deforestation and forest degradation. These lead to an overall carbon saving at the national level, at which point it is verified. The government can then sell these as CERs on the international market, the proceeds of which they use to recover the costs of the original activities, but that the implementers of these activities will be paid according to the cost of the activity, not according to the value of the carbon savings achieved as a result.

The astute amongst you will already have noted that both of these architectures rely on the government as the key intermediary between domestic and international markets, and may reasonably question whether many governments, especially but not necessarily only those in the developing world, are well equipped to play such a role. However, since these agreements are being negotiated by national governments we seem somewhat stuck with this situation. The alternative is the voluntary carbon markets on which valuations tend to be an order of magnitude lower. (Nonetheless I know many protagonists who prefer the predictability of these over the uncertainty of the international regulatory market and the intermediation of their government.)

Clearly the transaction-based model is the more economically pure, and should lead to the most cost-efficient outcomes. However, it is also rather more complex to administer since each individual carbon saving must be independently checked and accounted for, no mean task for governments not great at record keeping, and which may impose significant transaction costs. Furthermore, a key issue for REDD is leakage (will conserving one patch of forest simply lead to another patch being felled more quickly?) which is easier addressed at the national level than at the project level, e.g. through a national energy policy to reduce charcoal usage. Finally, efficient markets rely on reasonably well informed participants, something which is likely to be far from the case when those participants are the rural communities who are the de-facto guardians of their local forests. (Trying to explain carbon dioxide, the greenhouse effect, climate change, carbon markets and REDD to people who have barely completed primary school is a serious challenge!) This calls to mind a time when I was sat in a meeting when someone from FAO was presented their proposed project for using Payment for Watershed Services (PWS) as a means to achieve forest conservation in critical catchment forests; his preferred tool was fungible water rights which might be an economist’s nirvana but struck me as being a development worker’s worst nightmare.

So a fund-based model does appear more attractive. Communities can simply be paid for not cutting down trees (this one is relatively easy to explain), the government can tackle leakage issues through broad national policies and implementation programmes, and the total transaction cost should be significantly lower. Not surprisingly many governments like this option (e.g. Brazil’s Amazon fund) because it gives them much more control and freedom of action, and maybe even the possibility to make a nice profit which can boost their tax revenue. Plus for many of the politicians and senior officials responsible for these negotiations it will not have escaped their notice that a fund-based system would bequeath to them vast empires with large budgets, and plenty of opportunity for patronage. A realisation that leads us to the big BUT in all of this. That is donors have been pumping millions of dollars into forest conservation for quite some time now, with mixed results at best. Why, just because now this money is labelled as REDD, should we expect better results?

There are two counter-arguments to this. Firstly the sums of money that will  be available under REDD are several orders of magnitude higher than institutional funding to date. However, though many times financial resources have acted as a significant constraint to forest conservation in developing countries, plenty of very well funded programmes have struggled to make a significant difference, which would seem to invalidate that argument. Secondly, and more cogently, if significant forest conservation gains are not made then the recipient government will soon stop receiving the money. This is a real paradigm shift that holds out the possibility of solving the sustainability problem. Since REDD is paying for results rather for process (as most aid projects have done in the past), governments will have a real incentive to get things right for once, assuming they can actually out-muscle their timber barons, and deliver a sensible, working energy policy.

The question then comes how do governments go about delivering the carbon savings in the face of the last mile problem? Politicians here frequently bemoan and berate civil servants for their laziness and inefficiency (not necessarily unreasonably, though whether the kettle is blacker than the pot is a moot point). Faced with such inflexible and unreactive bureaucracies developing country politicians are apt to reach for panic button which leads to sweeping and rash policy decisions made in a hurry and with little consultation, such as banning agricultural exports in the face of food shortages (rather than keeping markets open to encourage farmers to produce more). This is what has got many campaigners worried; will governments reverse the hard-won gains of the last twenty years in a flash and renationalise forests, disenfranchising local people en masse of their forest rights? For heavily forested developing countries will REDD be any more of a benefit to their citizens than oil has been for ordinary Nigerians? Hence the focus of  many campaigners on “REDD+” or “REDD++” where the pluses stand for social protection and equity, and  biodiversity conservation.

For these reasons, whilst the optimist in me wishes the fund-based model could work, the pragmatist in me says a strong transaction-based architecture is almost certainly going to be required in most cases, and the pessimist in me thinks that we’ll probably end up with mostly fund-based approaches any way. One colleague of mine said he had heard talk of a hybrid approach, but didn’t really understand how it would work. If anyone does I would love to hear about it; please post details in the comments below. Meanwhile I will focus on my next beer transaction.

Cash on Delivery & the Last Mile

Recently I had a conversation with a donor representative here who was frustrated by the progress of rolling out CBNRM initiatives in this country, and the subject of Cash on Delivery came up. (Ok, I suggested it.) This was well analysed by Owen Barder, but I think he missed one important angle which in many ways sums up  my motivation for writing this blog.

Owen principally talks about COD creating incentives for policy changes (though he is not clear what kind of policy changes he means), and correctly points out that until donor threats carry real credibility these will have little impact on donor country leaders. He also highlights the marginal impact aid conditionality has on the interests of elected politicians, although I believe we should not underestimate their persuasive force when there is no direct negative impact on a politician’s personal interests. (Not all politicians are totally corrupt.) My interest, however, is on solving the delivery problem, and here we are not necessarily talking about the bigwigs. (As the CGD blurb mentions, COD can be used within countries too.)

A lot of aid suffers from the problem of the Last Mile (e.g. see Esther Duflo’s TED talk). This problem is not always well appreciated (or well enough appreciated) by aid officials working in their plush offices in the capital city, or those even further away in donor countries. They tend to engage with aid on two levels. Firstly they get involved in direct bilateral projects; these are intensively funded and can help pilot new approaches, and point the way forward for a certain sector in the host country. Such projects can produce great results (if you can avoid the problem of unsustainability), but do not produce many bangs for the lots of bucks invested owing to the rather concentrated nature of such work. Donors, typically, then seek to convert these narrow projects into broader national programmes; in doing this they team up with the best officials they can find, nurture and support them, and generally get a good feeling that they are working with the right people.

Except that, at least in the country where I work, such talent is at a premium, and most of it heads towards the private sector. All this aid has to be delivered by someone, and here donors are trapped by constraint of low human resources and their understandable, though often misguided, desire to minimise transaction costs. I’m talking here about the teacher who doesn’t turn up to work, or the agricultural extension officer who cannot get to work because his car has been requisitioned by a higher-up. The final mile is often by far the worst managed, and this is the problem we confront daily in our view from the Bottom Up. I think most people working in aid do realise this, but then rather forget it in their desire to push out the latest big idea, or somehow kid themselves that this time they’ve got the right trick to solve that problem.

I think mid-ranking officials, as long as they’re not hopelessly corrupt, might be more susceptible to COD incentives. Most such officials want to be managing bigger departments, with bigger budgets (more opportunity for patronage if you want to be cynical). Under COD, those officials who can better manage delivery will do better, those who cannot will be left by the wayside. And it will also get the donor out from under their noses; as Owen points out, the best thing about COD might be how it incentivises donors to reform. Those who want change should first put their own houses in order, and donors are more culpable than most in this respect.