Some people, especially working on the front-line of delivering aid programmes, are uncomfortable with the idea that aid should be more strongly linked to results.   Some point out that there is no evidence that government officials and aid workers will respond to incentives (see this article by Ngaire Woods and Paolo de Renzio); indeed, the very idea seems to impugn the character of development professionals.   Others are concerned that an increased focus on results will add to the bureaucratic burden of form filling and reporting which plagues the life of front-line staff (see this essay by Andrew Natsios, former USAID Administrator.)  On his blog yesterday, Simon Maxwell lists four further concerns which he says give rise to uncomfortable “seat shifting” about results : that the evidence won’t really be used; that linking aid to results relies on a simplistic, deterministic view of development; that it risks focusing too much on results which can be measured, rather than deeper but less observable changes; and that a results-based approach fails to take account of the complexity of how aid transactions actually feed through into activities.

These are serious and important concerns.  In particular, if measuring results is simply bolted on to the existing systems for the allocation and management of  aid, the danger is that we add to bureaucracy with little real benefit.  But if better measurement of results is used instead by aid agencies to simplify the way they manage aid programmes, rather than just adding new reporting, then the results agenda creates the opportunity to reduce bureaucracy, decentralise decision-making, increase country ownership, increase the focus on outcomes that really matter, step away from linear, deterministic thinking about how results are achieved, focus more on relationships and institutions, and really liberate development workers to work on what really motivates them – delivering change on the ground – and less on managing the bureaucracy at home.

This post sets out how a focus on results might unlock changes in the way aid is managed, which could lead to significant improvements in aid effectiveness.

Let’s deal first with the question of motivation.  In a recent post I reported Dan Pink’s claim that people are generally not motivated by money to do cognitive and creative tasks. This led Nandini to ask:

How do you reconcile these views with those in Cash on Delivery aid, where incentives are clearly being proposed to motivate government officials to get the job done?

I do not think that most people who work in development are in it for the money.  Most of the aid workers could earn much more, and live more comfortable and less stressful lives, if they did something else. Aid workers don’t want to get rich: they want to help make the world a better a place.  And though they are often teased for being self-righteous, it is a genuinely noble and admirable motive.

I also know many fine public servants in developing countries who are similarly motivated by service to their country. They often burn with national pride and want their country succeed. They want to address social injustice and poverty.  They are hard working and often paid very little. They too would be insulted by any suggestion that they are in it for the money.

That isn’t to say that there aren’t some officials in both developed and developing countries who are corrupt and greedy.  In some developing countries where civil service salaries are very low, some people are not fortunate enough to get over Dan Pink’s hurdle of “being paid enough that they don’t have to worry about money”.  For these people, money may be a motivator, but for reasons I’ve set out elsewhere, I think aid plays a relatively minor role in changing their behaviour.

If the people making these decisions are not motivated by money, what is the point of linking aid to results?  The answer is that incentives still matter. People who work in NGOs spend a big part of their time writing grant applications, and sucking up to donors, not because they enjoy it but because they want to increase the funding for their organisation.  Within aid agencies, people spend a lot of time doing things they don’t really value, because that is what they have to do to be allowed to spend money on the programmes they care about.   Clearly people working in development are  responding to incentives, not related to their own incomes but because they want to expand and sustain the programmes they think are important.  So incentives matter, and they do change people’s behaviour.

Incentives to be ineffective

Within the current aid system, donor agencies and their staff face powerful incentives to give less effective aid. For example:

  • Some donors give “tied aid”, even though this is estimated to reduce the value of aid by 15-30%. Tied aid is a consequence of domestic political pressure. Some donors allege that tied aid necessary to maintain public support for their aid programme (e.g. US food aid).
  • Conditions are often imposed on aid, not because anybody believes the conditions really make a difference to what happens – all the evidence suggests to the contrary – but they enable donor governments to claim that they are “doing something”.  Though we know aid works best when we the developing country has real ownership over their policies, in practice the aid system regularly undermines this ownership by setting conditions and imposing its own systems of accountability and control.
  • Donors like to retain the flexibility to turn the aid tap on and off at short notice (a practice often disgracefully known as using a “short leash”) so that they can respond to domestic political pressures, even though the lack of predictability is one of the biggest impediments to aid effectiveness.

This behaviour is driven mainly by the proper and important need for aid agencies to show to their taxpayers that their money has achieved something. Because aid agencies have found it hard to show a clear link to results, sometimes for quite respectable reasons, a complicated bureaucratic culture has grown up to trace how inputs are used, to set milestones and benchmarks, to and to encourage very high levels of engagement with recipients about how they propose to go about spending the money. But this behaviour, while a completely understandable and rational response to the need to be accountable, reduces the effectiveness of aid.

If taxpayers (and their representatives: politicians) cannot observe the results of aid, in a far away country about which they know little, they will be inclined to insist on getting whatever second-best information can be assembled to satisfy themselves, and to demonstrate to taxpayers, that the aid has been well used.  The consequence is that aid is made less effective.   For example:

  • There are political costs (and risks) to making long term aid commitments, and to untying aid, but in the absence of measures of results there is no political cost to the invisible loss of effectiveness that unpredictability and tied aid create.
  • Donors know that country ownership is essential for aid to be used well, but they nonetheless feel that they have to impose their own systems to determine how aid will be used and to track implementation.
  • There are political and strategic benefits to giving aid in every country and every sector; but in the absence of measures of results nobody knows how much inefficiency this proliferation creates, nor is there anything to force donors towards a better division of labour.

It is as though we plant a tree and then dig it up again once a week to see how the roots are doing; and then we wonder why it isn’t growing very fast.

We cannot wish this problem away.  It is right and inescapable that donors must be accountable to taxpayers.  Donors cannot avoid these pressures simply by making international agreements to do better in future (which is why those agreements are implemented only sketchily and slowly, if at all). Nor will they meet the need for accountability merely by improving their communications or changing the narrative.  A stronger link between aid and results looks to be the only sustainable way to solve this problem.

The risk of additional bureaucracy

The audit culture in aid is beginning to create a backlash among development professionals.   Earlier this month, the Institute of Development Studies organised a meeting to “take the first step towards resisting the audit culture of philanthropic foundations and government ministries.”  And in an important essay which is getting a lot of attention, Andrew Natsios (a former Administrator of USAID) wrote about what he sees as the most disruptive obstacles to development work in agencies such as USAID: layers and layers of bureaucracy.  He argues that an emphasis on measurement is at odds with the fact that transformational programs are often the least measurable and involve elements of risk and uncertainty.

And if that is how the staff of donor agencies are feeling, imagine how much worse it must be for hard-pressed officials in developing countries, and NGOs trying to keep their overheads to a minimum.

Some people have a reasonable fear that push for better measurement of results, and a better link between aid and results, is a further and troubling manifestation of this bureaucratic tendency.  There is a significant risk of additional bureaucracy if the need to measure results is simply bolted on to the existing systems. But if donors are willing and able to use the stronger link to results to simplify the way they manage aid, this offers the best – perhaps the only – opportunity to step away from “command and control” management of aid.

Will measuring results make the system better or worse?

The argument for greater use of linking aid to results is not based on the simplistic notion that we should motivate aid workers or officials in developing countries with monetary rewards (which I don’t believe).  Instead it is based on the idea that if we can tackle the problem of lack of information about results, it becomes possible – and indeed desirable – to allow decentralised decision-making by people on the ground who know much more about what is important and what works; so respecting country ownership, and liberating developing countries and aid agencies from much of the bureaucracy of the aid system.

Cash on Delivery aid is a possible example of this approach.  It is a market-like mechanism, in that the donors set a price that they are willing to pay for certain types of result, and they allow governments to decide for themselves whether and how to supply those results.  But it is rightly a long way from a pure market solution.  I would like to see it tested because I hope that if aid is linked explicitly to an independently-verified measure of the results that we care about, this might enable donors to change the way that they relate to recipient governments and the way they manage aid.   There should be no reason to agree milestones and benchmarks, no need to set targets; no need for policy conditions or inspections of public financial management.  There does not need to be a simplistic, linear model of how results are achieved, monitored at every milestone and benchmark.  Financial commitments linked to results could be long-term and predictable.  Developing countries could choose their own priorities, and take responsibility for their own progress.  Cash on Delivery might enable donors to respect country ownership, and the diversity of approaches countries will take, while still being able to report to domestic taxpayers what difference aid has made.

Simon Maxwell is right to point out that too much development thinking simplistically assumes that inputs are translated into outputs and outcomes, as implied by the logframe approach.  But this is surely a critique of the existing approach in which inputs, activities and outputs are obsessively described and tracked and rather too little attention is paid to the actual outcomes.  A results-based approach, by contrast, would enable the development partners to step away from deterministic models, to invest more in relationships and institutions, to be politically savvy and opportunistic, to learn as they go, and to deliver results in ways that are messy, unpredictable and imprecise but which actually work.

Simon Maxwell is also right to highlight the importance of institutions and capacity in developing countries.  Many existing approaches to development pay too little attention to this:  because of our need to track inputs and activities we tie up partner governments in red tape and try to direct proceedings from the outside or – worse – bypass government systems to deliver services through NGOs and the private sector.   This undermines institutions and systems.  By contrast, one of the motives behind Cash on Delivery is that it enables developing countries to figure out their own way to deliver results, and so to strengthen their own national institutions and policy making. Because in this model donors are focused only on results, they do not need to intervene intrusively in how countries make progress, so enabling the institutions of developing countries to draw in outside expertise they actually need, want and use; innovate; test ideas; set their own priorities; adopt their own approaches; and learn.

Sceptics are certainly right to ask if a results-based approach will focus on the wrong results: there is a risk that we will link aid to short term, measurable indicators instead of sustained institutional and economic change.  That is certainly a shortcoming of today’s aid system, which sets targets for all kinds of short term activities which may or may not be important (“set up an anti-corruption commission” or “reduce double shift working by teachers”) and creates a huge industry of people designing ways to monitor these indicators, while accepting with apparent equanimity that we lack even the most rudimentary information about the things we ought really to care about, such as whether corruption is being reduced or children are learning at school.   It is possible that, under the banner of results-based aid, we could simply reinforce this short-term interventionism by linking money to short-term indicators that donors currently believe might matter. But if we resist this bureaucratic temptation and instead link aid to results that really do matter – such as reductions in maternal mortality or improvements in access to clean water – then we reduce, not increase, the risk that the aid system is too focused on the wrong things.

Conclusions

The hypothesis is that by addressing particular information failure – lack of information about results – and if necessary dedicating some resources to it, we might be able to break away from the coping strategies which donors have had to adopt, which create a lot of pressures which make aid far less effective and which waste a lot of time and money on nugatory work.  But this will require some imagination on the part of aid agencies to think about how they can simplify their existing systems for managing aid when they link aid more robustly to results.  Development professionals are right to worry that if results monitoring is simply added to existing systems, the effect will not be greater simplicity but more complexity and bureaucracy.

In my view, it is disrespectful to aid workers and to officials of developing countries to suggest that they are motivated by money.  That is not the logic behind linking aid to results.  It also disrespectful to occupy so much of their time on bureaucratic administration, recording information about inputs and processes, milestones and benchmarks, when they would generally make a greater contribution if they were free to deliver development results.

Aid is made less effective by the incentives which aid agencies face, which they in turn transmit to their staff.  In large part, these unhelpful incentives are a consequence of lack of information about results.  If we can measure results better, and if we can use this to simplify the management of aid (and not simply bolt additional reporting on to existing bureaucratic processes),  this will enable more decentralised decision-making, respect country ownership, make the jobs of aid workers and government officials more rewarding, improve the effectiveness of aid, and so reduce poverty faster.

If you enjoyed this post, please consider leaving a comment below, and perhaps sharing this with other people using the buttons on the left. You can also sign up to have blog posts sent to you by email.

20 Responses to Incentives, results and bureaucracy in aid

  • I think there is actually a problem even if you are motivated by incentives. Consider a salesman selling faxes to small businesses in a high street store. The business recognises and rewards the meeting of monthly targets.

    But what happens is that if you actually meet the target (and sometimes if you don’t), it is increased the following month. Which can be incredibly demotivating to the salesman. This is particularly true when the number of small businessmen coming to the store in a given month is essentially random, so it is hard to measure the actual influence of a given salesman on the monthly sales – they may be more influenced by fluctuations outside of their influence.

    I’m no aid expert, but it strikes me that there could be a large amount of nescience involved in the hitting (or not hitting) any given target within the life of any project.

  • Owen — really enjoyed your thoughtful piece — you’ve addressed some key concerns and brought out the nuances regarding incentives vs money, results vs/with flexibility, very well. And as you know, I am a fan of COD and wish some countries and donors would try it out seriously. Still, I keep thinking why ‘lock’ COD as just a way of doing aid, and not articulate it as a way in which Governments can use their own resources, including general budget support funds, to do COD in a manner that sets up the right incentives within their own countries? Eg getting districts to focus on increasing learning outcomes in education. Of course such a proposition comes better from inside countries, perhaps thats your hesitation, but would it not be a worthwhile effort to have a group of serious people in a country engage with this scenario? Even as ‘pilot’ experiment in say 5% of the country?

    • Rakesh – Thanks. I am indeed reluctant to tell other countries how to manage their internal affairs. But one thing I like about COD aid is that it enables governments to enter into back-to-back arrangements internally, if they so wish. Under COD, if Governments want to pay districts by results, they can. If they want to contract out service delivery to the private sector, they can. This is very hard to do with conventional aid. And if they don’t want to make those arrangements internally, that’s up to them, not donors.

  • I’m still skeptical about COD. What sort of measurements will be used?

  • Owen thanks for a very useful piece which touches on a number of key issues, and summarises different positions very well. Clearly we know we have a measurement issue and associated attribution problem, we also have an aggregation problem, and we have an ownership problem – particularly of front line staff in agencies that Dennis Rondinelli in ‘Development Projects as Policy Experiments’, and many others, have pointed out for some time.

    It seems to me that if we need an ‘intelligent mix’ of ways of addressing these problems we need a) to recognise different ways of viewing the world and diffrent types of change process ranging between: i.e. the world is predictable and objectives can be set and organisations held accountable to them, or the course of events is complex & cannot be predicted or controlled, through to development is about empowering less privileged to hold the powerful to account , b) to acknowledge there are different forms of measurment and accountability that are dominant within these i.e. more focused on principal-client for the first, peer and social accountability for the second and political accountability for the third and c) to be clear there are favoured methods of ‘measurement’ within these schemas i.e. more RCT type methods with the first; more peer reviews and stakeholder assessment type processes with the second; and more social accountability and ‘voice’ type options with the third.

    One important advantage of the last is its potential in providing real time ‘signals’, if harnessed cleverly i.e. by Ushahidi type crowdsourcing, which could provide a much better bottom-up baromoter of change which might in turn inform taxpayers elsewhere though a shortened route of accountability. This might also help address the critique of the evaluation of the first pahse of the Paris declaration which noted the technocratic nature of the process was liable to alienate parliaments and citizens everywhere.

    Of course how we arrive at deciding what is the intelligent mix for a given issue is very much a key part of the debate, along with the perennial question of how organisations provide the ‘enabling environment’ (or incentives) for their staff to listen and repond to information, learning, and feedback particularly from those with the least power in the system.

  • Hi Owen — Thanks for this piece. Clearly written and compelling as usual.

    I appreciate your re-framing of results-based financing (RBF) away from simple motivation and towards the idea that it is about demonstrating results vs inputs — which can then allow for a lot of other desirable changes. But I have a slight different take on one thing you mention, which is that NGO and other workers are not primarily motivated by money and get pissy when it is suggested that they are. I’ve traveled in a few SSA countries over the last year writing about RBF (for health) programs for the World Bank. The front line NGO staff I interview are certainly motivated by money — as you say, they want more of it so they can do more with the programs they care about (and/or they want to get all they’ve budgeted for if the “incentive” is actually a penalty for non-performance, as it is in Liberia and S Sudan). Health facility staff too want more money. It’s the first thing they ask for when you ask them what will motivate them. They’d also like phone scratch cards, housing allowances, school allowances for their kids, and stuff for the clinic like lab coats, computers, an ambulance.

    What they get upset about is the suggestion that they weren’t thinking about performance or results until the World Bank showed up and told them that they should. OF COURSE they care — that’s why they got into the business. They may track it differently than the World Bank does — through best practices based on lessons learned, which may be shared in various forums — but for them, it counts. (Not to mention that contracts are usually based on targets and NGOs risk not having their contracts renewed if they perform badly.)

    I have two concerns. First, COD aid. I’m a fan and I think it should be tried, but I worry that an incentive payment to the top for, say, a reduction in maternal mortality, will not translate to all the many levels and layers of people down that are needed to achieve that result, which, first, is notoriously hard to measure anyway, but also requires people far far away from developing country capitals to make decisions: a woman has to choose to come to a facility rather than deliver at home with the traditional birth attendant who she’s known a long time and is comfortable with. The TBA has to not lure the woman away from the facility. The facility has to have competent staff, a midwife, a physician assistant, and it helps if they’re getting paid on time and reporting into an HMIS. They need an ambulance in case of complications. And a million other little things. I’m torn between the reality of this complexity and the need to keep COD aid “pure” in the sense of little or no donor interference.

    This relates to my second concern, which is about a kind of RBF — the type that gives incentives (fee for service, bonus, penalty) to public and private facilities, a la Rwanda, Burundi. Rwanda is well known for being hugely successful and Burundi is shaping up to see positive results as well. And they are regarded as country led/owned. But they are really, really complex. In Burundi, for example, there are over 150 composite quality indicators alone. All sorts of new structures and comittees have to be established and data collection and monitoring procedures — and there’s a huge need for TA. I’m not the only one who’s worried about the wisdom and sustainability of such complexity, even if results are being achieved.

    Just a few thoughts. Apologies for the length of this comment, and thanks again for this great post. I hope it’s widely read.

    Lindsay

  • The data would be independently verified.

    This is a big concern that many proponents of CoD aid tend to gloss over.

    First of all, independent verification (a.k.a. audit) is not easy, especially in the settings that we are discussing here. To be able to audit e.g. “each additional child who completes primary school and takes (but not necessarily passes) a standardized test”, you need to have:
    1. a verified starting point that is audited for completeness of number of children who have completed school and taken the test;
    2. a system which measures the increase and has sufficient internal controls to make the it sufficiently likely that the resulting numbers are accurate;
    3. a way to verify the design, existence and operation of these internal controls (which would include at least some real-life observation by the putative independent auditor).
    In many of the settings that we discuss here, this is highly unlikely to transpire; e.g. can you imagine an adequate system of nation-wide internal controls in Somalia of Afghanistan — or an independent auditor verifying the existence and operation of those controls?

    Secondly, even if all this would exist, it would become quite expensive. Independent auditors don’t come cheap in the first place (and I say this as a former auditor with one of the big international audit firms), and for this sort of work you would pay quite a premium. Would the advantages of CoD outweigh the extra cost of the audit?

    • Michael

      The issues may not have been addressed to your satisfaction, but I don’t think it is fair to say that COD proponents have “glossed over” the question of independent verification. For example, if you read the CGD book about COD aid, you’ll find a lot of discussion of precisely this question.

      My own view is that it is indeed potentially difficult and expensive to build up a system to get reliable and independent figures; but that it is also essential, whether or not there is Cash on Delivery Aid. How can we make progress towards maternal mortality or educational standards when we have no information about them?

      I am struck by the huge expense and time that is devoted to designing and monitoring a vast range of activities and intermediate indicators associated with aid programmes: almost every week some overseas consultant wants to interview me here in Addis Ababa about some mid-term review or end of cycle review for some donor project that they think I might know about. There is an army of consultants tracking activity indicators from logframes, process indicators, sometimes output indicators. All for what, exactly? If Cash on Delivery Aid enables donors and governments to spend less time recording all this intermediate information, and to devote a small part of that resource to getting robust estimates of outcomes, I think that will be an important step forward.

      So I agree this is both important and non-trivial: but it is important and non-trivial whether or not Cash on Delivery aid is introduced. Until donors and governments focus their attention on outcomes, and make it part of the aid process, it is difficult to see them investing enough in gathering and understanding this crucial data.

      Owen

  • Not to be too pedantic, but I said that ‘many [emphasis added] proponents’ gloss over the issue, not all — with CGD being a clear exception.

    I am struck by the huge expense and time that is devoted to designing and monitoring a vast range of activities and intermediate indicators associated with aid programmes … All for what, exactly?

    I think we fully agree there, but our conclusions are different: I would say that many of those monitoring activities are quite useless as well (and, as somebody who would love to see more accountability in aid, I say this with pain in my heart). The difference is that CoD is impossible without effective monitoring/audit because you will not know whether the delivery actually eventuated — in other words, without verification it becomes just a slightly dressed-up form of cash aid. This is not true for more classical forms of aid: we might not know much about their effectiveness (sadly), but the verification is not essential to their execution.

    So yes, there is an issue around verification/audit that is unique to CoD aid, and that does not get enough attention.

  • But do the COD reuslts necessarily translate to short or long term development? Who is pacified more, the donor who sees great vaule for money for his tax payer, or the citizens who don’t necessarily get a say out of it?

  • Another classic post we can use in the advanced development course.

    I would like to go back to the original issue: do we need to focus on results? It goes back to the old discussion on process approach and goal oriented approach. According to some proponents of the process approach, emancipation was the main development goal, so who cares whether the school is built, or the food produced. By having unclear, vague or multiple objectives, some NGOs, but also bilateral and multilateral agencies, have unmeasurable results. They have a right to choose this approach, and important lessons can be learned in the process.

    However,

    When development is funded by donor money, especially donor government money, some rules have to be followed. One of them is that it should be used for what it is aimed for. Cash on delivery complies with this rule. Working “in the general direction of good” doesn’t. This is why this kind of projects NGOs should rather finance with their own resources.

    This is not because there is a belief that only results count. It is because the tax payer has a legal right to know what his money was used for. Moreover, as the needs are way bigger than the available resources priorities must be set. Normally the first investment will be the one where there is a guaranteed maximum return. Less will be invested in more risky ventures, although I would agree there must always be enough “risk capital”for innovative or risky approaches.

  • Just wanted to add that in a long-term consultancy I did for Sida on supporting some framework NGOs to set up non-linear, outcome-focused RBM systems for their operations in deepening democracy (which, by the way, was a BATTLE! both internally in the NGOs and took more than anticipated ‘sensitisation’ within the donor agency), we recently came to a situation where one of the NGOs who previously resisted the whole RBM-thing said: “Wow, NOW I get it. We could really use this to reclaim power over our own programmes.” So can RBM be an empowering tool rather than a draining, bureucratic, vertical one led by donor interests? Yes, absolutely! But instead of spending time and energy on resisting it, it actually takes some guts to ‘claim’ it, fit the language to one’s own reality and start to communicate results (and to motivate why all are not necessarily quantifiable if that is the case)! I see lots of similarities between RBM and ‘systematic investigative reporting’ (being a former journalist), and that just makes it so much more interesting… :-) Thanks for an interesting article! Charlotte (cor@ncg.dk)

  • What gets me is that so many of the “results” are fuzzy around the edges with everyone in the reporting chain, governments, development partners and donors all wanting to see the positive story. There is little independent verification. So everyone is focusing on a satisfyingly positive story and there is little time to focus on the lessons and what’s not working.

Follow Owen




Email subscription

I want to receive new posts by email

Owen on Twitter