It's now 8 years since I read Albert-Laslo Barabasi's Linked, and longer since Paul Ormerod published Butterfly Economics, which prefigured some of the applications of network theory and complexity mathematics in economics. A recent high-profile example is Yochai Benkler's The Wealth of Networks. Some of the most creative economists have worked on social norms and their implications for policy – Ed Glaeser at Harvard springs to mind, for example. Over the intervening years there have been quite a few other books on this subject, so I've become a bit blase about new titles and was slow to pick up Connected: The Amazing Power of Social Networks by Nicholas Christakis and James Fowler. However, I think it's well worth reading even if like me you're not a newcomer to network models, and is a terrific introduction for social network novices.

The key message is a vital one for policymakers. It is that policies which focus on individual characteristics – race, sex, IQ, socio-economic status, whatever it might be – as explanatory factors and therefore target variables won't be as informative as looking at people in their structural position in their social network. An effective policy will try to target the people who're in the most influential positions in their network. It's not only who people are that shapes their lives, but who they know. The empirical programme this suggests is mapping networks and trying to identify the small subset who are influential in their networks. This is obvious for public health, for example, when thinking about contagions or how to reduce smoking rates. Many of the examples in the book are drawn from the field of health, the research background of Christakis.

As the authors note: “This approach shifts decades of public health work. It targets neither socioeconomic inequality nor socioeconomic or behavioral vulnerability per se, but rather structural inequality and structural vulnerability.” (p131)

The same applies to other areas of policy. They mention en passant the social nature of contagion in the financial world, although not specifically in the context of the recent crisis. (Andy Haldane's Bank of England paper remains the outstanding analysis of this.) The book doesn't touch as much as as I would have liked on economics more generally, although they note that preferences cannot be taken as fixed and determined independently of others. The authors nod favourably towards behavioural economics – although it seems to me that many behavioural economists also assume that preferences and decisions are independently determined for each individual, it's just that they happen to differ from the ones in conventional economic models. Network models deliver the obviously (sometimes) factually correct prediction that some trends will feature 'avalanches', especially in the modern world of socially networked media.

Finally, the book notes: “The social networks we create are a valuable, shared resource.” They mean the whole can be greater than the sum of the parts in society. Network theory undermines much conventional policy-making but offers, in return, a profoundly optimistic message.

Here's another review of the book, in The Guardian.