It isn't often that I'm driven to shout at the radio in the mornings, but today was an exception. The Prime Minister is to announce today that the government will start collecting statistics on national well-being. There's one way to do this which is necessary and welcome; but the happiness gurus have latched onto the imbecilic way to approach it, and they were the people who grabbed the airwaves.
It ought to be obvious that measuring 'Gross National Happiness' is a bad idea simply from the fact that its advocates hold up Bhutan as a model. Bhutan? It's one of the poorest countries in the world, with low life expectancy, poor literacy levels and scant political freedom. I don't care how 'happy' its not-very-free people claim to be when they're asked in a survey.
Direct survey measures of happiness have grabbed attention because of the 'Easterlin Paradox': within any country, richer people claim to be happier than do poorer people, on a scale of 1 to 3 or 1 to 5. But the average level of happiness does not rise in proportion with GDP per capita after a level of around $17,000 a year has been reached, either comparing countries at any point in time, or in one country over time. This has been translated into the received wisdom that higher GDP doesn't make us happier. Our prominent advocate of this view is of course Professor Richard Layard, who was one of the first to write about the need for the government to force us all out of the rat race in his book Happiness.
But this supposed fact rests on a statistical error. Measured happiness can never go above 3 – and it's at about 2.6 now in the UK. That's because it comes from a survey – and anyway, happiness isn't a boundless concept: how could we go beyond euphoria? GDP is a statistical construction that can rise without limit (it's a non-stationary time series). To expect happiness to go up at the same pace as GDP is like expecting height to rise in line with economic growth. There is a link – people are taller in richer countries. But not 8 metres tall. A growing number of economists have pointed out this error – for example Helen Johns and Paul Ormerod in their book Happiness, Economics and Public Policy; and recently Betsey Stevenson and Justin Wolfers in a recent paper Economic Growth and Subjective Well-being. When you take the logarithm of GDP to deal correctly with the nature of the data, there's a strong positive link.
Having ranted about the imbecilic version of the happiness agenda, however, and insisted on the importance of continuing economic growth for our well-being (or why would be be bothered about a recession?), it is important to introduce new measures about the state of the economy. There are three kinds of statistics we need.
One is an array of indicators of what people tell the ONS matters to them, on the model of the dashboard of statistics published regularly in Measures of Australia's Progress. This approach is the one recommended recently by Amartya Sen and Joseph Stiglitz in a report for President Sarkozy.
A second is a measure of the nation's comprehensive wealth, including financial and infrastructure assets, and also natural assets and human capital. This is a huge exercise which economists including Ken Arrow and Partha Dasgupta have begun to calculate. Measuring a nation's assets, broadly defined, is an essential step towards taking proper account of the legacy we leave for future generations, and therefore for achieving sustainability meaningfully.
The third is a systematic set of generational accounts so we know what obligations past and present government decisions about pensions and welfare imply for future taxpayers. Without this, we'll never understand the sustainability of government finances and the structure of the state.
It's a huge agenda for statistical offices. Let's hope they're not diverted by nonsense about 'happiness'. For anyone interested in these issues, it's a theme of my forthcoming book, The Economics of Enough: How to Run the Economy as if the Future Matters (already available for pre-order from Amazon and elsewhere!)