'Just Give Money to the Poor: the Development Revolution from the Global South', an excellent overview of cash transfers

May 25, 2010

What is the future of UK development policy?

May 25, 2010

Are aid workers living a lie? And does it matter?

May 25, 2010
empty image
empty image

These are the questions posed by Rosalind Eyben in an intriguing new paper in the ejdr_01European Journal of Development Research (no ungated version, sorry).

Ros, formerly of DFID and now attached to the Institute of Development Studies, knows the aid industry backwards and is struck by “the dissonance between what [aid workers] do and what they report that they do.” The aid industry as institution thinks in terms of logical frameworks, results-based management and an underlying view that Ros calls “substantialist” – a perspective that “sees the world primarily in terms of pre-formed entities in which relations among the entities are only of secondary importance.”

Good aid workers on the other hand, are “relational”: relationships between actors matter, and actors themselves change and evolve through their interaction with eachother. At their best, aid workers surf the unpredictable realities of national politics, spotting opportunities, supporting interesting new initiatives, acting like entrepreneurs or searchers, rather than planners. But when they report back to their bosses, out come the logframes and strategic plans, as messy reality is shoehorned back into the substantialist fantasies of the machine.

Does this “cognitive dissonance” matter? The machine doesn’t seem to think so. “No official aid agency has been prepared to undertake a study that aims to learn about their staff’s everyday practices – what they are doing, as distinct from what they report they are doing – and their effects.” But Ros thinks it is important, not least because it cripples our efforts to understand change processes. Here she draws a distinction between bounded v unbounded problems:

“The first are “difficulties”. With difficulties there is broad agreement on the nature of the problem; there is some mutual understanding of what a solution would look like; and there are limits to what is required in terms of the time and resources required for their resolution. Unbounded problems, on the other hand, are “messes”. There is no agreement about the diagnosis and therefore the actions required; no possibility of an eventual permanent solution because solutions generate new problems; and therefore no way of determining the quantity and type of resources needed. Governments fail to achieve results because they insist on treating messes as difficulties.” Sound familiar?

While some interventions, eg measles vaccinations, lend themselves to the substantialist world of logframes and bounded problems, many others do not. Here change is complex unpredictable and messy. Attributing change to any one intervention by an aid donor or anyone else is impossible. A relational approach is useful, a substantialist one largely futile. But, Ros asks,  “if the case for such an approach to the complex context of international aid would appear to be so convincing, then why is it that top management continue to ignore process and prefer substantialist inputs and/or outputs? Why are economists still preferred over anthropologists?”

What would happen if the aid industry opted for relationalism and binned its logframes?

“Just as glasnost brought about the fall of the Soviet Union, so might an admission of what is really happening in international aid result in its dismantlement with Northern taxpayers refusing to buy into such a contingent and messy process. On the other hand, an energetic clamp-down on relational practices might equally lead to institutional collapse. Practitioners need just sufficient encouragement from top management – as well as from relational advocates like myself – to continue subverting the system for the system’s benefit.”

And she mischievously draws a parallel “with the way that the Soviet Union was able to report that collectivised agriculture was an effective means for sustaining agricultural productivity. In practice, the farm workers put their energies not in the collective farm but in their own small holdings, and pilfered collective-farm resources to invest in them, and it was this that led to sufficient food being produced for the authorities to be able to demonstrate that the overall system was working. Without the farm workers realising it, their subversion was maintaining the very system that they were resisting”.


  1. Very interesting application of complexity theory. My view on this is that you will always need to come to a shared view about what one agency can do to play a useful role in the mess (aka a “programme”). The big idea, the rationale, the theory of change – call it what you will. But it must be there. Be humble about it. Ask people to challenge it. But have one. And make sure that everyone working on that programme understands it.

    The theory should be liable to change, but not on the basis of a whim or fashion or frustration at not having met lofty goals. Be very clear about the kinds of things that will look like progress along the way, but seek surprises and reward people who point out failure.

    As for a relational approach…. I like talking to other people too. But facilitated processes and participatory exercises should be ends to the means of achieving something together, not ends in themselves. Hold that candle up to a lot of agencies’ work and we start to look quite inefficient.

  2. Very interesting. Sent round to colleagues here in DFID Ethiopia 😉 I am sure that some DFID staff might be persuadable to keep a diary of what they do on a daily basis, with each individual’s efforts loosely structured according to a template of questions/prompts. I don’t think that there’s a need to wait for an aid agency to give official blessing to this.

  3. This reminds me of David Mosse’s ethnography of a DFID project in India. He concludes that NGOs and other actors in the aid industry pursue activities and make the relationships that are needed in order to justify and continue their own existence. They then ‘interpret’ these in their reporting to match what the sector has agreed as acceptable policy. It is important to note that the effects of their actions may well be positive for the people they are trying to help – it is not necessarily a negative situation, just that the actual outcomes are unlikely to be those predicted and reported as policy.

    Mosse, David (2004) ‘Is good policy unimplementable? Reflections on the ethnography of aid policy and practice.’ Development and Change, 35(4) . pp. 639-671.

  4. The debate reminds me of a story told in the South African Civil Service a long time ago. A young man was placed in the Kalahari, an arid part of South Africa. There was no clear job description so he had nothing to do. Being afraid of losing his job he decided to start counting the number of flies in his office. Every month he diligently sent his report about the number of flies to his head office. When the next person went there, the first question the head office asked for was the “Fly report” that was not sent to them after his first month.

    This is what I also think that happens with many aid workers. There is very little guidance given them so they develop their own “Fly reports”.

  5. An interesting article, some interesting points made, and i think Ros has highlighted a very important issue of the current dissonance between the formal systems and the real-life practices throughout the aid system. 2 points:

    1) I would disagree that RBM and other systems are generally seen to work (whether due to workers subverting the system or not). An evaluation of RBM at the UN, and others elsewhere, have shown very limited impact of these frameworks on accountability or service delivery. Unfortunately, most of these evaluations interpreted the problem as one of poor compliance, rather than asking whether the correct approach was being used in the first place

    2) Ros’s assertion that it would be dangerous for the public to be made aware of the ‘messy’ realities of aid work certainly chimes with a lot of people’s assumptions about the issue. Better to keep as much as possible under the political radar is the attitude, possibly quite rightly. However, this is (to my knowledge) largely untested. And, to my mind, if we accept that the story about aid can only really be told with “impact” figures, then we are setting ourselves up to to take some big steps backwards in terms of aid effectiveness.

  6. Great post and article.

    I especially like the distinction Ros makes a distinction between trust-based relationships and “un-scrutinised relational approaches [which] can become complicit with the clientelist cultures in which aid practitioners find themselves”. This point perhaps didn’t come through clearly enough in the post.

    It is dangerous to assume that all ‘relationalist’ work is necessarily good, in line with the broader goals of poverty reduction, and working in morally sound ways despite the substantialist pronouncements of the formal system.

    After all, it may well be the case that many of the darker interests that get fulfilled through the aid system – be they commercial, strategic, military – also rely on an invisible, relationalist way of working.

  7. Interesting though concerning article. Of course, aid workers are far from being saints and there will always be a mismatch between intention and action. It is also particularly important perhaps when considering the reported outcomes on the effectiveness of aid reaching those who need it most. If workers report that they are reaching even the most vulnerable intended beneficiaries, is this just their intention or is it grounded in reality?

  8. Since aid in general is not particularly effective, any aid worker who is taking up the slot of another person who would be doing damage for imperialist or nationalist reasons is doing God’s work.

  9. In reply to Stephen Jones: indeed, David Mosse is and excellent teacher and anthropologist, I studied under him. But to me (as a naive student perhaps) he seemed to be leaning too far towards the view that practice and reporting have nothing to do with each others’ realities.

    Sure, they often come from different rationales, and I’m not claiming that self-preservation isn’t part of it (how many aid agencies REALLY want to ‘put themselves out of business’?). But I think this approach – if I’ve interpreted it correctly – does not give enough credit to donors, a proportion of whose staff are ex-aid workers, not lifelong bureaucrats. I think if we were a bit more honest in our reporting, letting on about the mess a little bit more, this would be one of the best ways to influence policy. Not ceasing resistance, as clearly this keeps things dynamic, but shifting the equilibrium so that the system is slightly less reliant on subversion to keep going.

    The debate about contribution vs. attribution is one example – no project can claim to have entirely caused a social or political change, so how do we prove that we contributed, and how much? We have been saying donor outlooks need to change for years – perhaps more realistic, explanatory reporting is a way to push that change from underneath.

Leave a comment

Translate »