Complexity Economics, Evolution and How Change Happens

January 7, 2009

     By Duncan Green     

Eric Beinhocker’s book, ‘The Origin of Wealth: Evolution, Complexity and the Radical Remaking of Economics’ (for review see previous post) challenges our understanding of how change happens and the role of would-be ‘change agents’ like Oxfam.

The evolutionary model of change described in the book is very different from the change models I used in ‘From Poverty to Power’. It neither requires a conscious ’agent’, like a political movement seeking a specific change, nor an arbitrary ‘event’, like a cyclone or a war. Instead, it resembles Richard Dawkins’ ‘blind watchmaker’, shaping order out of chaos through the basic mechanisms of evolution – differentiate, select and amplify. Beinhocker confines himself to economic change, but it is worth thinking about how this applies to social or political change as well.

The challenge in applying these ideas to political and social change, compared to something like the spread of supermarkets, is that the selection mechanism of the market is missing, and other forms of selection (popularity, elections, effective states, active citizens) do not appear as well designed for the job. That absence of an effective selection and amplification mechanism could explain why political and social change has lagged behind the extraordinary spike in economic change since 1750 – so the answer to Jo’s comment on the previous post is that the book’s thinking doesn’t help much with discussions of the non-monetary economy (unpaid caring work etc).

One implication of the book is thus that designing and strengthening institutions to make non-market forms of selection and amplification as effective as possible is a crucial aspect of encouraging change.

What does all this mean for the way NGOs work? As you would expect from a McKinsey man, Beinhocker is at his most insightful when talking about private sector companies, but the lessons are worth applying to non-profits too. He argues that the natural tendency to impose hierarchies, strategic plans etc means that companies are usually better at executing than exploring. In the private sector that doesn’t matter because markets do the job, churning away to ensure that growth proceeds even as companies rise and fall – Schumpeter’s ‘creative destruction’.

In civil society and other areas such as international institutions, that doesn’t apply – when was the last time an international body or NGO went bust – the League of Nations? Beinhocker argues that if companies want to survive, the answer is to ‘bring evolution inside and get the wheels of differentiation, selection and amplification spinning within a company. Rather than thinking of strategy as a single plan built on predictions of the future, we should think of strategy as a portfolio of experiments that competes and evolves over time.’ He holds up Microsoft as a successful example of this approach.

The challenge of doing this in NGOs is not so much in ensuring diversity, (there’s never a shortage of argument and dissent….) but in creating a process of selection and amplification in absence of market signals. How do we get better at spotting and supporting winning ideas, and dropping losers?

Secondly, an intriguing feature of the ‘complex adaptive systems’ that Beinhocker argues are the core of the evolutionary process is that their constituent ‘adaptive agents’, interacting according to a few simple rules, can generate an extraordinarily complex and unpredictable outcome that looks very like the real economy’s trademark pattern of booms, busts and periods of tranquillity. Such complexity requires only:
1. Agents interacting with other agents
2. A few simple goals for agents (e.g. ‘survive’)
3. Rules of thumb (agents’ mental models of ‘if X happens, I do Y’)
4. Feedback and Learning that ensures that rules of thumb evolve

What would be the equivalent with respect to political and social change? In previous discussions on chaos theory and change, I’ve identified a few possible strategies for change agents in chaotic environments:

1. Solidarity with civil society movements (don’t waste time trying to anticipate change, just accompany your partners on the journey)
2. Venture Capitalism (try 20 things/ fund 20 projects, knowing that 19 will fail and one will come good on a large scale. The key is to identify failures and cut losses as quickly as possible to free up resources for further experiments)
3. US Marines (set basic rules, then see what happens)

But Beinhocker’s book suggests a new addition (or a variation on option 3). He says of economics ‘The dream of a clockwork universe ended for science in the twentieth century, and is ending for economics in the twenty-first.…. We may not be able to predict or direct economic evolution, but we can design our institutions and societies to be better or worse evolvers.’

Maybe the same goes for political and social change too. When trying to influence change (e.g. the likelihood of conflict) in complex and chaotic environments, rather than arguing for more specific policies, we may be better advised to try and influence the general rules 2-4 that govern the evolution of the system, namely the goals of agents (in NGO-speak, their ‘attitudes and beliefs’), the rules of thumb (eg on what constitutes good responses to economic crises or natural disasters) and the effectiveness of feedback and learning (how good is their ability to evaluate and adapt?). Beinhocker seems to agree. In the epilogue, he writes: ‘in evolutionary systems, power comes not from the top down, but from the bottom up. Evolution is a blind process, and the evolutionary algorithm will respond to whatever fitness function it is given.’ One task for change agents is thus to alter what constitutes success within the evolutionary process.