More Spirit Level delusions

Chris Snowdon

2 February 2011

In the preface to their best-selling book The Spirit Level, Richard Wilkinson and Kate Pickett compare what they call their ‘discovery’ that ‘more equal societies almost always do better’ with the scientific breakthroughs of Joseph Lister and Louis Pasteur. The intention may be to equate the work of social scientists with that of biologists, chemists and bacteriologists. By placing themselves in the company of those who made world-changing discoveries in the sphere of biology and germ theory, they perhaps hope to narrow the gap between the social sciences and the physical sciences. Above all, they seem to want to medicalise the issues of income inequality and wealth redistribution.

There are, I think, two reasons for this - firstly, the desire to portray income inequality - and the free market capitalism believed to cause it - as a disease infecting the social state. If viewed in the context of a public health peril, governments can wage war on it as they did with smallpox. Secondly, presenting the case as if it were based on hard science enables it to be portrayed as non-ideological and not politically motivated.

Neither claim stands up to scrutiny. The soft social sciences are always going to provide more equivocal evidence than the physical sciences and while Wilkinson and Pickett are social epidemiologists, The Spirit Level is much closer to sociology than epidemiology. Societies are not lab rats and the simplistic bivariate scatter-plots presented in The Spirit Level could not be further away from double-blind randomised control trials.

The notion that Wilkinson and Pickett are non-ideological is difficult to maintain in the light of their political activity. As well as being involved with the Socialist Health Association and a long-standing campaigner for left-wing causes, Wilkinson has spoken about the need to rebuild the empirical evidence for socialism that was once provided by Marx, and explicitly sees his work on inequality as being one component of this. He and Pickett helped found two political pressure groups (the Equality Trust and the One Society) and by the time The Spirit Level has reached its closing chapters, any pretense of political impartiality has been abandoned.

None of this means that their conclusions are wrong, but empirical claims are there to be tested and in the last year a number of people, including myself, the British sociologist Peter Saunders and the Swedish team of Sanandjai et al., have done so. Despite working independently, all three of the resulting critiques identified similar serious flaws in their analysis. I doubt any of us were expecting Christmas cards from Wilkinson and Pickett for our trouble, but nor could we have anticipated the vitriol of the response. For having the temerity to show that health and social outcomes tend to be worse for black Americans than for whites, Saunders was effectively called a racist. For no reason whatsoever, critics were described as being of the ‘far-right’.

In November 2010, a fresh edition of The Spirit Level was published ‘with a new chapter responding to critics’. More of a counter-attack than a defence, this chapter centred around the twin notions that (a) The Spirit Level reflected a well-established scientific consensus, and (b) criticism of their methods and findings were ‘foul attacks’ from well-funded right-wing groups determined to refute an irrefutable large body of evidence. In my view neither of these claims resemble the truth and as an addendum to The Spirit Level Delusion I have released an additional chapter of my own (for free download) critiquing Wilkinson and Pickett’s interpretation of the academic literature and their defence of their methods.

To read the new chapter, click here.

Tags: Christopher Snowdon, egalitarianism, inequality, The Spirit Level, The Spirt Level Delusion, Welfare

Submitted by Whig on Wed, 02/02/2011 - 14:42.

Keep up the good work in exposing this nonsense. The attempt to make a scientific basis for government intervention in the economy has a long history - from Mercantilism and the Physiocrats through Malthus, Marx, Booth and Rowntree, Keynes and now a large part of the climate change debate and 'spirit levelling'. Socialism or etatism really is a hydra that needs its head cutting off regularly. Unfortuately, its ability to appeal to powerful special interests means it is unlikely to be slain. In the case of the Spirit Level there are two sets problems. The first is evidential - the case for equality would have to be unimpeachable to think about justifying coercion; clearly it is nothing of the sort. The second is implementation - even if (!) we accept that their conclusions are correct to some degree. 'Spirit levellers' seem to make an automatic assumption that inequality is best tackled by state intervention. This is, of course, a hugely problematical assumption which serves only to either show that they are sincere but ignorant, or worse, to reveal their underlying motives - which are not equality, but forcible redistribution of income. No wonder Ed Milliband is so keen! This assumption is nonsense for two reasons, at least: (i) the state is already intervening in the economy in a vast range of areas from income redistribution to regulation, to the extent that much of the inequality present in society is actually a consequence of state action (ii) the overwhelming evidence that shows that state intervention in many cases actually enhances inequality via unintended, but often perfectly predictable or observable consequences. This relates, I think, to the excellent points Mark Pennington has been raising regarding Rawls and concepts of 'social justice' on this blog recently. The best way to achieve equality needs to be discovered - even if we desire it - state intervention prevents such a discovery process and assumes that the best route to equality is already known, even if that known is the most desireable route, which is unknowable anyway! There are many other points to raise. Why should health outcomes be priviledged over other choices such as liberty? How much equality is desireable? Given that some inequality is necessary for social and technological progress to occur, would the spirit levellers desire us to have a little bit, and if so, how much? How is this to be determined? Moreover, they do not take into account the likely pay-off between absolute growth and equality via redistribution. A very equal society that is very poor is likely to have worse absolute health outcomes than a rich society that is unequal. Given that income redistribution and government intervention tends to retard economic growth (viz the Rand curve) we can suggest that societies which are compelled to be equal are likely in the long-run to perform absolutely worse than free ones on all measures anyway.

Submitted by D.R. Myddelton on Wed, 02/02/2011 - 15:51.

Writers, including experts, can make honest mistakes. ' To err is human.' Most of us, however, are not god-like enough to forgive those who, given the opportunity of a new edition, fail to correct, or even acknowledge, their material errors. That is why, for example, the arguments about climate change and its consequences have become so depressing. There is, I venture to assert, a wide range of defensible views both about 'the science' and about the economic implications. But when some people seem to think that their preferred political solutions are so desirable that one is even justified in torturing the data to provide apparent support for them, we are entering the realms of religious hysteria. Hence publicly supported media, such as the BBC, have a duty to be impartial which, unfortunately, it does not always manage to achieve. It is always worth asking those who seem to be strong advocates of particular policies: what new evidence, if any, would cause you to change your mind? If (as sometimes seems to be the case) no conceivable new evidence would make them change their mind, then such people are clearly 'prejudiced'. They first make up their mind, then try to twist (often cherry-picked) evidence to fit their conclusions. That is one reason why freedom of speech and debate matters so much. To argue that something is so important that one cannot allow people who disagree to express their views is to miss the point. The more important something is, the more necessary is becomes to allow people to argue about it.

Submitted by Kris on Fri, 04/02/2011 - 19:12.

See also W&P's letter in this week's Economist: "SIR – You were incorrect in asserting that there is a weak link between inequality and the variety of problems we attribute to it in our book, “The Spirit Level” (“Unbottled Gini”, January 22nd). The relationships between national levels of income inequality and mental illness, children’s well-being, low social mobility, teenage births, prison rates and trust are all extraordinarily close, with correlations of between 0.7 and 0.9. Although the connections we show of inequality to life expectancy, infant mortality and homicide are slightly weaker (though still statistically significant with correlations of between 0.4 to 0.6), there are an additional 200 independent analyses that chart the link between health and inequality and another 50 between violence and inequality. Your point about homicide and gun ownership was also wide of the mark: control statistically for gun ownership and the relationship between homicide and inequality becomes slightly stronger. The issue of outliers is a red herring. Raised as a criticism of some of our studies, it is irrelevant to these other analyses. The politically motivated attacks on our work have been rebutted, not only in a new chapter of ours, but also by others."