Previous event

Cutting Edge Policy - Understanding System Stewardship

Tuesday 20 September 2011, 14:45


  • Paul Ormerod, Director at Synthesis IPS, a partner in Volterra, and author of 3 best-selling books on economics, Death of Economics (1994), Butterfly Economics (1998) and Why Most Things Fail (2005), a Business Week US Business Book of the Year.


  • Michael Hallsworth, Institute for Government, author of the IFG reports System Stewardship and MINDSPACE: Influencing behaviour through public policy.
  • Bridget Rosewell, Chief Economic Adviser to the Mayor of London, responsible for new ways of thinking about the benefits of transport connections, and helping to make more and better information available to Londoners.

Paul Ormerod told a story of how in the 1990 World Cup football hooliganism was a big issue, and England were forced to play all their matches in Sardinia, the massed police started a riot by letting go a round of live ammunition. The normal deterrent effect of firing was overwhelmed by the network effect of rioting. That showed the importance of social networks - especially ‘real world' networks of friends, family, colleagues and groups in influencing our behaviour

Economic and social policy since 1945 ignored those networks. It was based on the assumption of individual agents making decisions rationally based on incentives, usually price. That has had some successes but also serious limitations, especially when it comes to solving social problems.

‘Nudge' and behavioural economics were a smarter and broader way of thinking about these incentives, but they continued to use the same framework of agents making decisions in isolation. They repeated past failures, neglecting to understand the role of networks directly influencing what others do.

Network effects influence by altering individual's preferences, and tastes through can be simply called ‘copying'. In a 2006 experiment, student's preferences for music downloads were dramatically different when there was ‘social influence' - when they had information on what others had chosen - and the influenced distribution was radically different from the uninfluenced one - with a few big winners.

Networks can magnify the effect of an initial change or offset it - they can make the effect of a policy much more unpredictable and less replicable - because the effects depend on what part of the network a policy "hits". The recent London riots showed the importance of network effects: it was not the size of the initial disturbance which determined how far the disorder would spread, but the extent to which those involved were linked to others in different parts of the city.

Networks are a means through which values and cultures spread. Demographically similar communities can have differing rates of worklessness - the idea and norm that being unemployed is acceptable can spread and embed itself within a social network But networks can be part of a solution too: the single most effective way to leave worklessness is hearing about a job through a personal connection. Where worklessness is high people don't have those connections. Networks can be penetrated to change values, or we can try to change the structure of a network, in order to link people to opportunities.

More effective policy solutions to social problems can come through looking at a problem with this understanding, and a key reason for many failures is not recognising that networks and values are a driving factor, not simply incentives. The Big Society as a concept appears to be based on the idea that there is a need for a change in values; a different, broader and potentially more powerful way of thinking about policy. To change values we need an understanding of networks, and progress is being made on getting there.


Michael Hallsworth pointed out that there are two broad intellectual traditions within behavioural economics: one focuses on heuristics and biases in decision-making (and is more focused on individuals), but another which deals more with how external cues mould decisions (and which is more focused on relationships and influence). Network theory is situated closer to this second tradition, rather than being a total contrast to behavioural economics. He added to Paul's analysis the importance of understanding the networks of agents who implement policies. Such networks can create unintended effects, for example in the way in which a change in one part of the criminal justice system caused the whole system to re-align. Additional issues to consider are the need for experimentation and adaptation in policy making, and the need to factor in that various policies can and will combine.

Bridget Rosewell pointed out that policy makers need to ensure that the networks they are accessing are the right ones, and that there can be network failures due to the self-referential nature of groups of policy makers themselves. Additionally, organisations have strong cultures resulting from the networks within them, including the civil service, and that these can be very hard to change. She also highlighted the importance of not looking for quick wins, and of being aware of the physical networks which can bring people together, as demonstrated by agglomeration effects.

After discussion (and some scepticism) we asked each panellist to name an area where they thought policy would have been improved by thinking about networks: Bridget thought transport infrastructure, Michael the spread of clinical best practice within the NHS and Paul thought bank regulators should have better understood connections in the banking system.

More information