Our Programmes



Sign up to our newsletter.

Follow ODI

Ten traps to avoid if aid is to take context seriously

Expert comment

Written by Simon Levine

Image credit:Afghanistan couple travelling by donkey. Iain Cochrane/Afghanistan Matters Image license:CC BY 2.0 DEED

When we find ourselves making the same mistakes over and over again, isn’t it time we stopped to ask ourselves why?

That was the main conclusion from our reflections on the (largely failed) efforts over the past two decades to “reconstruct” Afghanistan, after we had been asked to draw lessons that would be useful for current aid efforts there. Our two main findings were: 1) that most of the mistakes were symptoms of the same basic mistake; and 2) that this was particularly visible in Afghanistan because of the scale of the aid operation, but the same problems could be found everywhere.

The basic mistake was that aid efforts were not appropriate for the country – for the economy, for the society, for the politics of the places that they were supposed to help. Because aid interventions are never described as taking places in countries, or places or societies, but only in “contexts”, we can put this into aid jargon: context wasn't taken seriously. That should be surprising. Everyone working in the aid world has heard the mantra a thousand times: you have to pay attention to the context! Pretty much every aid proposal, at whatever level, duly starts with some form of analysis of ‘the context’ – an analysis of the problem being addressed, a conflict analysis, a gender analysis, etc. How is it possible that context is then ignored?

So, it’s not that we don’t know that context is important. We also have a hundred and one ‘tools’ for thinking about different aspects of context, so if the problem isn’t caused either by negligence or ignorance, the diagnosis must lie elsewhere. We found an explanation in 10 patterns of behaviour that were commonly slipped into by all of us without realising it, and which we therefore call traps.

We describe these 10 traps – and how to avoid them – in our new paper. (It’s only a short read!) A quick look at a few them illustrates what we are talking about. The third trap, for example, is a kind of blindness that is caused by letting a semi-identified solution shape what we can see. It’s best illustrated by an example. Agencies such as the World Bank talked about Afghans lacking sources of credit. How was this possible in a country where much had already been written about the many forms of borrowing by which people survived? What was missing were formal institutions with ‘micro-finance’ in their name. These solutions-in-waiting meant that we saw their absence, and thereby missed a thriving, if problematic, sector of what is called ‘informal credit’. We would have done better to look for how people cope when they do not have money when they need it, rather than look for institutions that we recognise as providing credit.

In Afghan reconstruction programming, that trap was compounded by our trap number 2. This is a deviously camouflaged trap, where we think we are using technical theories, which are in fact social theories in disguise. The disguise means that certain theories about how people behave go unquestioned, even when, as in Afghanistan, they were palpably false. Effectively this means not seeing people quite as people, but rather as beings who inevitably follow our predictive rules.

If only Afghans had credit, it was assumed, they’d follow some principle of profit maximisation and be much more entrepreneurial, leading to higher productivity and increased incomes. Except that this flew in the face of everything that was already known about how the Afghan poor survive. They often engage in the lowest-risk enterprises possible, even if they have low returns, because they have to prioritise survival over maximising returns. And that meant that we couldn’t unlock an economic tiger through the market.

If traps 2 and 3 are about getting the problem wrong, trap number 7 – perhaps the most ubiquitous – is about misunderstanding what our interventions can do. The trap is to forget that you can’t impose technical solutions, because context always bites back. We’ve all seen examples of this one. Village governance is dominated by the powerful? – let’s create a democratically elected village committee! Trade is controlled by a small cartel of powerful people? – let’s get farmers to set up some marketing cooperatives! Does anyone really believe that the powerful are going to sit back and watch their power taken away by some aid project? That's highly unlikely. So, to keep going anyway, we now have to deliberately shut our eyes to context, in order to avoid seeing that the world in which our project makes sense is an imaginary one. And we then fall victim to more traps that make that possible, such as traps 8 and 9. These respectively ensure that we put all our attention on what we can control (i.e. the least important things about the lives of people we are trying to work for) and that our monitoring looks narrowly at our own project, and ensures that any glimpse of the real world is excluded from our attention.

Are we perhaps being overly cynical? We very much hope not. The alternative would be much worse. The failure of so much aid investment is well documented, and we do not want to believe that it is because the best qualified people in the aid sector are just not good enough to get things right. If these 10 traps are to blame for many of the limitations of aid investments, then we can be more hopeful. ‘Forewarned is forearmed’, so if we know what to watch out for, we can avoid the traps . That would liberate the passions, intelligence, hard-work, knowledge and analysis of thousands of people, so that their efforts can go in the directions that they judge to be the most helpful.

How can we ensure that no one falls into these traps? Could we not at least have given some guidelines for ensuring that a context analysis is always conducted and used properly? That would indeed have been ironic! If we were to argue that a simple technical tool could ensure that judgement was always based on good context analysis, we would be falling into traps 1, 4, 6 and 7. You can read about those for yourselves.

This blog was originally published on SPARC's website.