degree). These days she is a management consultant and social scientist who has been working to bring ideas of complexity theory into organisations for many years. More recently she has become interested in international development – hence the chat.
Jean argues that facing up to complexity is not an option. Behaving as if the world is stable and predictable when it is not does not make it so. Such mechanical thinking can lead to blindness to change and difficulty in adapting to shocks and fast changes. So a shift in mindset is needed. Don’t assume the world is a smoothly functioning machine: review progress often, pick up on unintended consequences, look for the unexpected, scan for signs of change.
But there is still a dilemma here. Some people use these ideas to suggest that a) The world is complex, so there’s no point trying to understand it – just do what you feel, or b) The world is complex, so we should give up trying to influence change in any particular direction and just pick civil society/other partners and accompany them through thick and thin. In fact, complexity theory has a much richer set of implications for development policy and practice, but they can be hard to nail.
So in addition to Jean’s more general points, here are some more specific candidates:
Firebreaks: forests and forest fires are classic complex systems, in that you can’t predict where the fire will take place, or how it will spread. But you can still introduce ‘circuit breakers’ into the system by clearing firebreaks in the forest that will slow down the spread of fire. In the development world, the closest parallel is perhaps with financial systems – e.g. suspending share trading once a certain level of volatility has been reached. What other examples are there?
No regrets policies: back to the financial system – we can’t be certain what kinds of speculation, if any, increase food prices, but could we take steps that would be effective if speculation is indeed the guilty party, while not harming the useful operations of financial markets if it isn’t?
Decentralization: bringing decision makers closer to the ground makes sense in complex systems where they are required to spot trends and react to them, rather than develop the master plan and implement it.
Enabling environment: rather than ‘picking winners’ – e.g. backing a particular social actor, technology etc, in complex systems where such winners could come from anywhere, it might make more sense to focus on creating a broader enabling environment to support would-be change agents. Things like data transparency, literacy, health and education, communications infrastructure (see internet pic), or even trying to influence the underlying norms and values that guide human behaviour.
Regulation: If the previous point sounds a bit like the Washington Consensus, that’s because complexity theory sometimes risks veering towards blind faith in the ‘invisible hand’ of markets. Jean’s counter-argument is that the self-organising invisible hand does not necessarily lead to ‘the good’. It depends on the values and intentions of the actors. The work of complexity economist Brian Arthur emphasises that free markets tend to lead to the big getting bigger and the powerful more powerful. The voice of the powerless and the voice of the future are soon lost. Governance, social movements, even good old-fashioned regulation, can be crucial in countering this ‘pull to power’.
Run multiple experiments: If you can’t pick winners, why not pick 20 runners and see which ends up being the fastest, then pick that one? This is essentially what we are doing with the Chukua Hatua project in Tanzania, and it seems like a really sensible way to intervene in complex systems.
Real-time data: functioning effectively in complex systems means spotting new (and inherently unpredictable) trends as soon as possible and reacting to them. Better real-time data on everything from nutrition to levels of popular discontent is important, but so is creating the right set of incentives and mindsets to ensure that organizations actually respond to the data and see the patterns within it.
Monitoring and Evaluation: in complex systems, trying to attribute an outcome to a particular activity is often a fool’s errand. But 
that doesn’t mean you give up on measurement altogether. One method focuses on the use of journals, diaries and looking for patterns in what Jean calls ‘narrative fragments’ which can all help detect impact in complex systems, even if they don’t provide the illusory certainty of ‘intervention A is 36% more effective than intervention B’.
Judging the context: Jean is keen on this one. Not all situations are endlessly uncertain and fluid. We need to make some judgements – what parts of our work and context are relatively stable – and the task is to do well what we are doing; what parts are very unstable and the focus is on agility and adaptation; where do we experience rigidity and ‘lock-in’ and the task is to challenge and disrupt?
This week Jean is in Northern Kenya, exploring what her ideas can bring to our work with pastoralists there – should be a fascinating example of theory meets practice.
By the way, regular readers will know that I am a big fan of Eric Beinhocker and was blown away by his book arguing that evolutionary theory was a much better model for the economy than the 19th century physics of equilibrium. Turns out that Thorstein Veblen got there a bit earlier: in 1898 he wrote a paper ‘Why is Economics not an Evolutionary Science’. Sorry Eric.]]>