100 indicators of well-being or just one? Stiglitz v Layard

October 30, 2009

     By Duncan Green     

The OECD conference I’ve been attending is winding down. Lots of banquets, but not much booze, so I never had to try the hotel’s tempting room service item ‘outer leaves of cabbage broth to chase a hangover.’ What’s the takeaway (ideas rather than food)?

The key debate seems to me to be over complexity. The various presentations described literally hundreds of different indicators already being used to measure progress, and proposed hundreds more. The World Bank reportedly has 27 indicators of governance alone. The Stiglitz Commission proposes ‘dashboards’ of indicators, allowing different people and institutions to combine them in different ways to measure and track the things that matter most to them (mental health, carbon emissions, citizen participation or whatever).

But there’s a cost to that, as Geoff Mulgan pointed out. Decision makers and ordinary people can only keep a limited number of indicators in their heads (Geoff put it at about 5). Above that and they get increasingly baffled. A South African number cruncher lamented that ‘an indicator cloud has descended on Africa, creating a fog of confusion’, while an EC statistician worried that composite indicators rapidly become a political football as each member state argues for the combination that puts its own performance in the best light, and each successive government changes them, meaning you lose comparability both between countries and across time. For similar reasons I started to harbour heretical doubts about the merits of ‘bottom up’ indicators designed by communities – fine, as a way for the community to understand itself, but the cost is that the government won’t be able to compare progress with the village next door (which will have designed its own, different bottom-up indicator).

Geoff’s answer was to combine the merits of simplicity and complexity by picking 3-5 standardized indicators, each of which would be at the centre of a cluster of disaggregated numbers allowing policy makers and researchers to drill down into the relationships between different aspects of people’s lives (eg between income inequality and child well-being). A key to this model’s success is what the conference called ‘visualization’ – efforts to make data much more intelligible and interesting – including the launch of what looks a potentially remarkable ‘wikiprogress’, which among other things is going to use graphics software that’s even more funky than gapminder.

happiness v researchersRichard Layard was more drastic. He argued that ‘we have to go much further than the Stiglitz report and establish a different metric from the metric of money.’ His preference is self-reported life satisfaction (i.e. happiness), which he predicts will replace GDP within the next 25 years. Like Mulgan’s model, satisfaction would be reported across a number of domains (health, family, work, income, community, environment etc).

The conference (and the OECD, judging by its roadmap for future work on well-being, distributed in Busan but not yet online), largely sided with Stiglitz. They had doubts about the robustness of happiness as an indicator – does it really mean the same thing in different countries? If people are happy through ignorance of what they are missing, does that invalidate their view (false consciousness alarm bells)? Don’t things like voice and engagement matter on their own merits? Are we proposing the social lobotomy of a Brave New World of drugged, happy, obedient slaves?

But I’m leaning towards Layard. This was a conference of the modern data magisterium, the ‘counting community’ as one speaker called it. They live and breathe statistics and complexity. But most people are not like that. If you want to measure well-being in a cash-starved low income country, communicate messages to the public, or interest politicians, simplicity is vital.

The Layard v Stiglitz debate set me thinking about paradigm shifts. One typical feature of a paradigm shift is that as the old model starts to get into trouble in describing reality, layers of complexity are added to it to try and keep it afloat. Eventually someone like Copernicus comes along and says ‘let’s put the sun at the centre of the solar system, not the earth’ and the complexity melts away. Layard feels much more like the Copernicus of well-being than does Stiglitz or any other speaker at this conference.

The last word on this conference goes to the wonderfully serene (and enigmatic) abbot of Beomeosa, a breathtaking BeomeosaBuddhist monastery on the outskirts of Busan. I asked him if he could help us by telling us what happiness is and how we could achieve it. He smiled and told me to drink my tea, and then he would answer. ‘Did you like the tea? Yes? That is happiness.’ He also described happiness as ‘thinking about happiness’ and ‘the undivided mind’. Put that into your metrics, guys.

October 30, 2009
Duncan Green