Over the next few months, I’ll be getting stuck into a big Oxfam project on how we understand and work on issues of power and change. As befits its
focus on ‘how change happens’, this is already evolving in unexpected directions, such as a stress on how we support Oxfamistas to work in ‘complex systems’ (aka the real world). Last week we had 30 of our brightest sparks in a room for a half day on this – here are some of the points that came up.
First up, a great way to get people thinking about complex systems and the contrast with the illusory certainties of project planning. Ask participants ‘Think back to your expectations for your life when you were 16. Has it gone to plan?’ Then stand back, enjoy the incredulity and draw out the connections with aid work (influence of random events, importance of relationships etc etc),
Next the thorny question of ‘critical junctures’ – those shocks that contain huge opportunities for changes, to which we have generally been pretty rubbish at responding. Some CJs are foreseeable – eg elections – and there the challenge is just to get a bit more intelligent in our planning. But others – eg Arab Spring – are not in anyone’s calendar. Can we get better at seeing them as they approach?
Based on her understanding of complex systems, complexity physicist Jean Boulton (who also came up with the life expectation tip) suggested a few telltale signs:
- Increased swings and instability
- Trust your instinct/judgement
- Some early indicators might exist – these could include anything from UN Global Pulse type data harvesting from social media to the entry of intermediate/professional organizations into areas formerly dominated by radical activist groups (a sure sign that an issue is going mainstream)
- Otherwise look for ‘weak signals’ – a small sense of ‘we’re getting somewhere’, eg anecdotes of new forms of organization/actor emerging, different constituencies saying the same thing, increasing interest from newspapers, traditional leaders or social media.
But this assumes the existence of neutral observers capable of spotting such signals, whereas we NGO types are all too willing to see what we want to see – alleged ‘weak signals’ are likely to proliferate as every staffer tries to persuade their bosses that the revolution is just around the corner, so please could they have some more budget?
My conclusion was that the more realistic option is to concentrate on the foreseeable junctures such as elections, and try and get better at spotting and responding to the unpredictable events once they occur. That means feedback loops based on savvy staff rooted in local realities, empowered to blow institutional whistles, hit red buttons etc when eg food prices rocket or protest movements start picking up pace.
Persistence is the flip side of jumping opportunistically between critical junctures and it makes my head hurt. If development is in many cases long-term, then aid agencies have to think over 10 or 20 year timespans. It could well be a disaster if a complexity focus led to us hopping madly from issue to issue in a desperate search for the next Arab Spring.
But equally, if you’ve been plugging away at a project for a few years without results, how can you distinguish between a productive long haul that just needs time, and a miserable failure that is going nowhere? Whose judgement can you trust on this? If the answer is only evident in hindsight, how on earth do you make decisions in real time? Suggestions welcome!
Lots of discussion on how to use theories of change to improve our work, but the danger there is that ToCs, complexity etc will become just another toolkit – a checklist that closes down thought and creativity, rather than the opposite. Much the same happened to the logframe back in the day.
The focus needs to be on supporting staff to build the skills, imagination and ability to understand and respond to what is going on around them. That needs lots of discussion, mentoring, and maybe even some suitably non-prescriptive methods, but we need to keep the focus on the people, not the process. With due apologies to Gandhi, ‘Be the toolkit’ seemed to work as a slogan on where to keep the focus.
Another nagging question is how do you fund work based on complex systems? Can you really rock up to DFID and say ‘hey, it’s a complex system, so we have no idea what’s going to happen. Can we have £1m please?’ Answers include:
- Collecting and publishing narratives where donors get it right: the conventional picture of donors hunched jealously over their logframes, hostile to anything smacking of independent thought, is clearly nonsense. But we don’t do enough to capture some great examples of imaginative donorship, such as DFID in Tanzania or the Swiss in Tajikistan. We need to collect these stories, and work with allies in aid agencies to get them out there.
- Trust Fund: Push the trust fund model as a way of turning large chunks of aid into lots of small chunks – the right size for the kind of initiative that needs funding, but would be crushed by the big bucks and paperwork that goes with them.
- Incubation period model: Back to Matt Andrews. How do we get a prolonged experimental incubation phase accepted as a standard (and extended) part of any project application? Interestingly, DFID is starting to do this with some of its research funding – supporting a year-long inception phase during which researchers can clarify and narrow down their agenda.
And finally, a comment from Alan Hudson on my last complexity-related post has stayed with me. ‘One of the things that struck me was that we had a very fruitful exchange without a mention of “complex systems. It just wasn’t necessary.’ Which reminds me of a comment that I think is great, but always seems to produce baffled faces in seminars – ‘talking about ‘non-linear systems’ is like talking about all the ‘non-elephants at the zoo’’: i.e. since non-linearity is the norm, not the exception, why does it need to be picked out? Will the mark of success in getting complex systems taken seriously be when they are merely seen as normal and the off-putting word ‘complexity’ becomes redundant?