Who’s afraid of complexity in aid?

Complexity is fast becoming a hot topic among development economists and aid bloggers. There have been a number of great presentations and papers on this (Owen Barder and Ben Ramalingam have both written accessibly on this on their blogs and in papers). There has also been some back and forth about what complexity really is, and whether people are really understanding and using it properly, some of which is quite academic, and a little intimidating.

This increased attention is a positive thing since in development many of the real life situations we deal with can indeed be characterized as “complex adaptive systems” and so treating them as if they are engineering problems with a clear linear cause and effect, as has been the tendency in aid planning in the past, will continue to lead to disappointing results.

But I’m also a little concerned that the tone of some of the current discussion can also leave the non-initiated with some unfortunate and mistaken impressions:

1. Complexity is well, complex (or is that complicated?), so to understand it you need to be really, really smart. The rest of us should stay aware for fear of making a faux pas and looking stupid (Maybe the tone and language of some of the academic debate doesn’t help).

2. Given how hard this is to understand, if we are to take it seriously in how we plan and deliver aid, we will need to hire high powered academics and management consultants or create a  specialized cadre of “complexity officers”.

3. Since development is  so complex then we can never be really sure about the results of what we are doing,  why try?. Won’t the system just “evolve” itself into the most suitable outcome over time anyway? and won’t this be much more effective and efficient if we don’ t interfere?

Not quite.

Complexity is indeed a complicated field of study, and like with many other topics in development, one around which there is incomplete knowledge and different schools of thought. So if you want to publish academic papers on it, or debate about it with other intellectuals you need to get yourself up to speed.

BUT even without a deep academic knowledge  it’s quite possible to understand what a complex system is, recognize one when you see one, and to use some simple approaches to deal with it.

Complex systems are not something new, they are as old as humanity and we have all been navigating them mostly successfully before we were even aware that such a topic existed.

One of my favourite analogies for this is that of raising a child. Yes, we can read books and get advice on how to do it. But there is no fail-safe recipe for how to raise a healthy well-adjusted child.

Why is this the case?  There are many actors involved – not just us, but also other relatives, teachers and peers, and most importantly not forgetting the child herself, and they all have different views and interests in the raising of the child. There are many environmental factors which you influence but don’t fully control such as the town you live in, the school you send the child to, what the child eats, exposure to illness, exposure to violence in society, consumption of media etc.  The relationship between the actors and the factors is “complex” and it’s hard to predict what approach will work best at a given time, and its not certain that what worked for one child will work for another. It’s hard (as any parent will tell you), but its certainly possible. People do it all the time.

Turning back to aid and development, there are multiple actors and factors all interconnected in ways which make the outcome of any specific action very hard to predict. But that doesn’t mean that there is nothing we can do, nor that we need deep academic knowledge or expensive consultants. An aid agency is only one of the multiple actors in the system but it can still take action and make an impact, even if the exact nature and scale of the impact can’t be easily predicted in advance.

Here are a few suggestions:

1. Start something that seems a reasonable approach based on what we know at the outset (drawing on information such as what has worked elsewhere, whatever scientific literature exists, what do partners say, what own experience and instinct tells us).

2. Adapt your approach over time in light of your actual experience and how well you are doing. Be prepared to modify and improve your approach continually based on what actually happens rather than what theory or past experience tells us.

3. If things are not working at all, admit it , stop doing it, try something else.

4. Continually collect and widely share data and information on what you are doing and what the results are.

5. Build in feedback mechanisms to see how you are doing including feedback from beneficiaries.

6. Try multiple experiments  – don’t put all your resources into a single approach. This way you can compare different methods and then scale up the most promising one(s).

7. Look out for and be open to unintended outcomes. These could be both negative and positive. It might be that the project has a positive benefit, just not the one you were initially looking for when you started out. Also small changes can sometimes  have large impacts.

And for those who are ready for it there are more sophisticated tools and approaches you can use in practice for programming (e.g. action learning, fail fast) and continual learning and evaluation (e.g. most significant change, outcome mapping).

There are many others who have written much more eloquently than me on this, but I just wanted to get down online that development is a complex problem, but there are still simple things we can do to work on it if we are just prepared to look at it differently from how we ave in the past.

[postscript: by chance, or perhaps by spontaneous order, Bill Easterly also posted on this exact same topic on aidwatch today – I’d suggest you also take a look at what he has to say]