Print
Share
Linking Aid to Results: Why Are Some Development Workers Anxious? (Guest post by Owen Barder)
December 17, 2009
I am pleased to share with our readers at Owen’s request this discussion of Cash on Delivery Aid, which appeared yesterday on his blog, Owen Abroad.
Linking Aid to Results: Why Are Some Development Workers Anxious?
By Owen Barder
The Center for Global Development is working on an idea which they call Cash on Delivery aid, in which donors make a binding commitment to developing country governments to provide aid according to the outputs that the government delivers. I think this is a good idea in principle, and hope that it can be tested to see whether and how it could work in practice. The UK Conservative party have said in their Green Paper that if they are elected they will use Cash on Delivery to link aid to results.
Linking aid more closely to results is attractive from many different perspectives. My own view is that linking aid directly to results will help to change the politics of aid for donors. Many of the most egregiously ineffective behaviours in aid are a direct result of donors’ (very proper) need to show to their taxpayers how money has been used. Because traditional aid is not directly linked to results, donors end up focusing on inputs and micromanaging how aid is spent instead, with all the obvious consequences for transactions costs, poor alignment with developing countries systems and priorities and lack of harmonisation. If we could link aid more directly to results, I think donors will be freed from many of the political pressures they currently face to deliver aid badly; and it would be politically easier to defend large increases in aid budgets.
Other people support Cash on Delivery aid for other reasons. Ministers and officials of developing country governments see it as a way to access more money without the attendant costs of conditionality and foreign interference in domestic policy. Some people see results-based aid as a way to restore the accountability of developing country governments to their own citizens, a social contract in which aid donors too often inadvertently interfere. Especially in the US, some people believe that linking aid to results can create stronger incentives for developing country governments to deliver high quality public services. Others support Cash on Delivery because it will improve the allocation of aid resources, since money flows to the places where services are being delivered and away from the places where money is being wasted. With all these complementary reasons there appears to be the possibility of a broad coalition of people in favour of moving ahead with testing whether Cash on Delivery aid can work in practice.
But there is one group of people for whom these ideas seem to be quite unsettling: development professionals in aid agencies and NGOs.
I recently wrote a response to a brief by CAFOD about some possible concerns about Cash on Delivery aid. As I was doing so I realised that the questions asked by some development professionals reveal some discomfort about the possible impact of results-based aid on the quality and content of their jobs. The “risks” identified in the CAFOD brief are not primarily about the consequences for development but rather risks to the privileged position enjoyed by professional staff in aid agencies and NGOs.
You can judge for yourself whether I am caricaturing the risks set out in the CAFOD paper, but they essentially amount to this: under Cash on Delivery aid money would flow to those governments best able to make use of it; governments would have freedom to decide which services to provide and to whom; governments would be able to decide how to use resources; governments would be accountable for their choices and the results; and progress would be measured according to internationally-agreed targets for impact rather than inputs and intermediate targets negotiated behind closed doors.
All these are necessary steps towards the internationally-agreed agenda for more effective aid set out in Paris and Accra, and necessary for the emergence of capable, accountable and responsive states. Yet when a mechanism is proposed that tries to organise the aid system in a way that means these things could start to come about, these consequences are described as “risks”.
At the heart of these anxieties, it seems to me, is a question about what sectoral advisers in aid agencies are meant to be doing. Take education advisers, for example (I am not picking on this group in particular, but it happens that the current proposals for Cash on Delivery aid are being developed looking specifically at education.) Many people who work for aid agencies managing aid programmes for education are themselves education professionals, often former teachers. Deep down (sometimes also on the surface) many of them want to be educators, not managers of aid programmes. They want to be involved designing the curriculum, reforming the pedagogic approach, training the teachers, buying textbooks, or improving the education management information systems. But it is the job of a community to educate its young, not foreigners. As managers of aid programmes the staff of aid agencies should be ensuring that aid is delivered in ways that increase the accountability of central and local government to the nation’s citizens, keeping transactions costs to a minimum, delivering aid in ways which support the evolution of country systems and priorities, ensuring that the money is used for the purposes intended by the funders, and showing what results have been achieved.
In short, managers of aid programmes should be focusing on the effectiveness of aid, not education policy. If governments need technical advice on education, they can procure that separately, and get advice from people who are more trained to build capacity and who are properly accountable for doing so, not get it as a bundled free offer-that-they-cannot-refuse from the people managing their aid. If it works as intended, Cash on Delivery aid would change the relationship between donors and governments and would turn development professionals back into aid managers instead of would-be educators. And it is this consequence which, I believe, some people find unsettling.
Many of my best friends are development professionals, and I know that everyone who works in development (well, nearly everyone) has the interests of the poor at heart. They often genuinely believe that they need to retain a degree of influence to ensure that developing countries make the kind of progress towards development that they (and I) want to see. There is quite a close parallel with the evolution of the attitudes of politicians, some of whom I also know well and have known since they were young, idealistic students. Nearly all politicians enter politics for the noblest of motives: to contribute to the improvement of the society in which they live. To a very large extent they retain those values through their political career. But over time there can be a gradual erosion of the distinction in their minds between their own interests and the service they give to others: some politicians gradually come to think that increasing their own power is the service of others, because they believe that they will exercise that power better than anyone else.
Politicians are, of course, at their most dangerous when they can no longer distinguish their own interests from the interests of the people they are meant to serve. Similarly we should be concerned when we hear development professionals identifying themselves as speaking for the poor, and arguing that they must retain influence (i.e. power) – purchased by the relative wealth of their country – to promote strategies which the country would not pursue on its own.
To be fair, I also know some development advisers who are focused on improving the effectiveness of aid, who are rightly aghast when they are asked to double up by providing advice on how to manage an education or health system. If I may be permitted a partisan aside, my observation is that DFID sectoral advisers tend to be more respectful of the need to promote effective country systems for policy-making and accountability than professionals from some other donor organisations (both NGOs and official aid agencies), and they are less likely to interfere in the country’s policies and strategies.
This may seem like an elaborate point to build from an innocuous and fairly sensible CAFOD brief about Cash on Delivery aid. But the risks identified by CAFOD, and the questions that have been raised elsewhere, would apply to any system of results-based aid that makes substantive progress towards giving governments more freedom to choose how to deliver their development programmes and making them more accountable to their own citizens for their own success and failure. I think these concerns actually reveal a deep-seated tension between the internationally-agreed agenda for improving aid effectiveness, and the views and interests of development professionals charged with designing and implementing those reforms in practice.
Update from Owen— PS: For the avoidance of doubt, I don’t mean to imply that the author of the CAFOD brief, nor anyone else at CAFOD, is personally concerned about their own career or job description. I am simply reflecting on the nature of the objections that have been raised to Cash on Delivery aid.
Share 
You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.
11 Responses to “Linking Aid to Results: Why Are Some Development Workers Anxious? (Guest post by Owen Barder)”
Post a Comment
We value frank and constructive exchanges and encourage you to use your real name in your comments.
December 18th, 2009 at 5:15 pm
One question is whether Cash on Delivery will bias assistance against TA. It may be easy to disburse cash against vaccinations, but how do you disburse cash against the provision of TA to draft regulations for the power sector. You could do it like the WB’s policy-based lending, where disbursements come after completion of reforms. The risk is the unintended consequences of the maxim that “that which gets measured gets done.”
December 22nd, 2009 at 5:36 pm
PS: I’m no expert on COD aid, but having sat in on a number of presentations about it here and having read various versions of the forthcoming COD Aid handbook, I think I can venture an answer to your question about Technical Assistance (TA).
Answer: TA is an input (like textbooks and school buildings) not an outcome. The point of COD aid is to create incentives for good outcomes and then get out of the way, so that governments in the developing world will more actively pursue the outcomes that we all want.
If we put this in the power sector, the outcome that could be contracted for in a COD Aid agreement could be, for example, reliable electricity delivered to X number of households or businesses at price Y for Z number of days per year w/o interuption. It would then be up to the recipient government to decide whether TA on draft regulations for the power sector is a necessary input–or not.
People (and companies) that provide TA may not like this approach very much. Fortunately (or not) there are plenty of other funding streams that are likely to continue to be available for traditional approaches to development, including TA.
January 5th, 2010 at 5:22 pm
Just by calling the attention to the possibility of insiders sabotaging “Cash on Delivery” aid, you are increasing the chances that it will work, and for this you should be thanked.
January 5th, 2010 at 5:41 pm
Barder’s and MacDonald’s points are well taken. Over several decades of work within USAID in population and health I found that larger and more sustainable gains were made (e.g., rise in prevalence of use of effective methods of family planning, oral rehydration, anthelmentics) in those few instances where we were able to enter AND to enforce formal, multi-year agreements for resource transfers contingent on performance – especially performance of step-wise intermediate steps like host government funding authorization & appropriation, procurement, and delivery of commodities. We did it in Kenya, e.g., in the mid-’80’s and elsewhere in other sectors over past decades.
The process, however, entails pain, particularly within agencies that must bend to strong constituencies for donor-mediated TA as well as local country pressures through host-country counterparts and embassies when pressures build to disburse without actually having achieved key contingent outcomes.
This genre of disbursement – “non-project assistance” or CoD in Barder’s essay – is far tougher and more complicated in practice than in theory and depends critically on negotiating AND enforcing agreements in ways that the USG is not widely capable of constructing. I fully endorse the CoD concept(s) but based on experience am not sanguine about the prospects, alas.
January 6th, 2010 at 7:45 am
I feel very attracted to the CoD concept – it sounds refreshing and pragmatic to me. I have however one comment and one concern. The comment: like in many other areas of development policies (principles, criteria and rules we ask recipients to comply with but we don’t feel to follow for ourselves), I wish we were implementing CoD in our own countries. In the public health sector, that I know better, so many prevention strategies have been implemented for years (decades, in some cases) now without health indicators improving meaningfully that seems we have sadly accepted that they are here to just maintain health problems under certain limits, not to try to reduce them. Take the example of HIV prevention programs in developed countries: when HIV cases rise, people (including and specially NGO’s) tend to anxiously ask for more money to be spent, rather than (re)evaluating programs to see where from the learning should be obtain. In my country, Spain, if CoD would have been used, the National AIDS Program would be have been closed down or re-thought many years ago. But here they are, doing exactly the same things for 20 years now, and blaming the population every time there is an increase in HIV percentages, even if they are within the statistic error margins. A similar comment could be done regarding obesity prevention programs in schools or smoking prevention policies in young people.
The concern: I would like to know better how the CoD idea can be used in combination with a better assessment of impact quality, both in terms of results quality and process quality. This is specially relevant if we want to avoid that Governments feel pressured to obtain results (better health indicators, for instance) by any mean, including in the lest harmful scenario ignoring civil society construction and in the worst case violating human rights. Unfortunately, we have had already numerous examples of that type of action programs with the results we all know.
Joan Tallada, Barcelona, Spain.
January 6th, 2010 at 12:25 pm
All of the risks laid out in the brief by CGD are real – just look at GAVI’s (Global Alliance for Vaccines and Immunizations) experience with paying bonuses for vaccination coverage. (See Lancet 2008, article by Murray analyzing vaccination coverage). This has resulted in inflated numbers in some countries and then GAVI suspended the bonuses. GAVI made a major, clear cut effort to tie results to cash and it incentivised poor data quality. GAVI funds are also being centrally managed and not going to the districts (not cherry picked in this case but identified as not having good vaccination performance). Finally, is CAFOD ready to invest in independent household data collection to validate results?
January 6th, 2010 at 3:51 pm
Excellent observations, Owen. Couldn’t agree more. Of course, getting donor governments – and not simply the front-line aid workers – to change their funding approach is the critical next step. The current USG experience in trying to provide more funding through Pakistani organization, rather than US-based NGOs, is already causing quite a backlash in Washington by the US-based contractors and NGOs who benefit so richly from US project-oriented funding.
January 6th, 2010 at 5:44 pm
CoD sounds interesting in principle, but the devil is in the detail of ensuring the outcomes agreed match the intended effect.
I.e. MDG 2, Universal Primary Education, is an admirable ‘outcome’ that could be targetted by CoD – but with 100 children in a classroom and a focus on enrollment statistics rather than attendance or completion, the real impact of a successful outcome is to my mind somewhat dubious.
January 8th, 2010 at 8:45 am
Cash on Delivery is an attempt to create a new buzz word for a concept that is already well established in the aid industry; it is nothing different from what is usually called General Budget Support (GBS). A number of donor agencies are knowledgeable about the conditions where General Budget Support is successful. The discussion about GBS is more than 5 years old and some agencies have adopted it as the preferred mechanism to channel aid. I would suggest CGD to give space for a discussion that is far advanced than start “marketing” “new ideas” that only adds noise to the debate.
January 12th, 2010 at 12:10 pm
Indeed Cash on delivery sounds more like the concept of General Budget Support but in my opinion; this is well thought as GBS has shown that GBS concept could be reformed better to give the intended results as a preferred mechanism for many countries who have had the opportunity to get such form of Aid. What would I be concerned with is how one would start accessing the results for the intended model? i.e where do u start and how do u start? as in the case of GBS it is a budget pot and redistribution according to government priorities.How will COD function? select sectors that are performing and use as model of encouragement to others or Give Govt the right to choose the priorities of well defined sectors with good outputs and indicators and pump in resources to feel the Gap and let the govt work on the other sectors slowly and gradually. But we must remember that many development agencies have areas of expertise or preference therefore some areas would be suffering with less resources. Alot has to be considered as agencies truly said have experts that would be working less or be of no use . My instincts tell me that they could be the sector monitors of the results for their agencies and evaluators at the same time working in harmony with country and not be police officers in guarding position.
Alot has to be said and done.goodluck
January 15th, 2010 at 12:57 pm
Barder’s blog on COD is good food for thought. Let me add to the mix: I am a ‘development professional’ and I am not at all anxious about COD at all. The devil is in the details:
First, within the COD approach, as presented by Barder, there appears to be a false double assumption that
(1) All or most of development aid is motivated by a genuine interest on the part of the donors in the economic development of the recipients. In reality, there is a great deal of other motives – of which some are purely selfish geo-political or economic interests on the part of the donor.
(2) All developing countries’ governments are genuinely interested in the welfare and development of their citizens, and would respond well to performance incentives.
Because of this, ‘bad’ recipient governments have learned to behave like prostitutes: If donor country X asks for too much accountability, there is always country Y who needs a vote in the UN or a market outlet and who will help out. COD would most likely create more of this ‘aid prostitution’ for bad governments.
Second, it is not always so simple to measure/assess development ‘results’ either in a quantitative or in a qualitative way a year or two after the action. If you are interested in measurable ‘inputs’ (X number of schools were built, X kms of road were built, X patients were correctly treated, etc.) that is easy. But there are more important and more complex development ‘objectives’ that would suffer from a COD approach.
Fourth, there is the question of who will do the assessment and who decides that input X is both good in the long term and therefore should be financed, and how the input costs will be measured, for reimbursement. It appears the most cost-effective approach will be less rewarded. It is not the same as buying pens or hiring a contractor, where you chose the least cost, given a certain quality.
Citizens in countries with the least performing governments would be worse off with a COD approach, and they often have the least capacity to change their governments or force them to perform. Perhaps that is the source of angst among development professionals?
As a carrot, COD is not a bad idea per se, but it is not the only, neither the best solution to the current development aid performance challenges. There is significant scope for more performance in development aid, but I believe the increase in performance requirements has to be gradual, and its application has to be selective, depending on each country’s governance quality, and which segment of the population is targeted. We cannot hold people prisoners’ of their governments’ performance. Especially in developing countries.