Opinion & Commentary

  • Print
  • Email

When Prophecy Fails: The Spirit Level and the Illusion of Scientific Socialism

Luke Malpass and Peter Saunders | KiwiBlog | 14 September 2011

A couple of Sunday’s ago, on TV One’s Q&A programme, there was a substantial interview followed by a panel discussion about a new book called The Spirit Level by British academics Richard Wilkinson and Kate Pickett. The book claims to ‘prove’ that more equal societies perform better than less equal ones on a range of social indicators including crime, life expectancy, social cohesion, drugs, literacy and obesity.  It concludes that a radical redistribution of incomes would improve the quality of life for everyone, rich as well as poor.

This followed a forum held at Victoria University, Wellington last November to discuss the controversial book, and numerous glowing reports about the book and its authors in the New Zealand Press.

This book has been welcomed by many on the left. In New Zealand, a number of Labour politicians are pedalling it, while in Britain Guardian journalist Polly Toynbee even describes Richard Wilkinson as the 21st century’s equivalent of Charles Darwin!  The left loves the book because it seems to provide scientific backing for their political instincts, ‘proving’ that income inequality is ‘a bad thing.’  But in reality, the book is seriously flawed, and serious left-leaning academics have begun distancing themselves from its claims.

One problem is that the income statistics are faulty.  As reported in the press, the book says Japan is the most equal country in the world.  This is crucial, for Japan performs very well on nearly every social indicator.  But the income data for Japan exclude single-person households and the self-employed, such as farmers.  More complete data supplied by the Organisation for Economic Cooperation and Development shows Japan is actually the 13th most unequal of the 30 OECD nations.  This alone is enough to undermine most of the statistical claims made in the book.

A further difficulty is that the authors ignore some key countries which don’t fit their hypothesis.  They claim to include all the rich nations, but they leave out affluent but unequal countries like Hong Kong and South Korea, and Singapore is missing from many of their graphs as well.  These countries all perform well on the social indicators yet have high income inequality.  If they were included, as they should have been, the authors would struggle to maintain many of their claims.

Their book also ignores social indicators where equal countries tend to perform badly.  They compare homicide rates, but not suicides; teenage births but not divorce rates; government foreign aid budgets but not private charitable donations; drug abuse but not alcohol consumption.  Taking these alternative measures, it could easily be ‘proved’ that equal countries perform much worse than unequal ones.

Perhaps the worst aspect of the book is the analysis of comparative statistics.  The authors present a series of graphs plotting the income distribution in each country against selected social indicators, and in each case they claim to show problems getting worse as we move from less to more unequal countries.  In fact, however, most of their graphs show no such thing.

One plots the homicide rates of 23 countries against their income distribution.  In 22 of the countries, the homicide rate is very similar (indeed, more unequal countries, like Singapore, Britain and New Zealand, actually have lower rates than more equal countries like Sweden and Finland).  One country, however, stands out: the USA has a murder rate three times higher than the others – and the USA is a relatively unequal country.

Wilkinson and Pickett allow this one case (what statisticians call an ‘extreme outlier’) to distort their whole graph.  Even though there is no association between income inequality and homicide rates across 22 countries, they conclude from this one case that there is.  They even suggest that Britain could reduce its murder rate by three-quarters if it had Swedish levels of income inequality, yet Sweden’s homicide rate is higher than Britain’s!

Many of their graphs are skewed in this way, and the problem is compounded by their failure to look for the possible influence of third variables.  For example, they compare infant mortality rates across 50 US states and find those with the greatest income inequality have the higher rates.  But the more unequal states are also those with bigger African-American populations, so is it income distribution or ethnicity that is causing the problem?  When one of us looked into this, we found that ethnicity is 18 times more powerful than income distribution in predicting a state’s infant mortality rate.  When we published this finding, Wilkinson accused us in The Guardian of being a ‘racist’!

It is true that, on the indicators they select, the Scandinavian countries tend to perform better than the Anglo countries.  But this is not because the Scandinavians have a more compressed income distribution.  Rather, it reflects deep historical and cultural differences between the two sets of nations.  We know this because when we look at other countries outside these two blocs, there is no association between social outcomes and income inequality.  Countries like Austria, France, Greece, South Korea and Singapore vary widely on their income distributions but show no pattern on the social indicators.  The claim that inequality causes social pathology is a red herring.

While the recent Q&A panel were sceptical, Professor Jonathan Boston, who chaired the Victoria University forum on the book, was quoted last year as saying: ‘We can have some confidence that more equal societies – other things being equal – have better social outcomes across a range of measures. It may not be absolutely conclusive, but I think it’s reasonably persuasive.’  He’s wrong.

Left academics and politicians have long flirted with the idea that socialism can be made to look ‘scientific,’ and that is why they are drawn to this book.  But all the book really does is dress up political dogma in the garb of science.  Arguments about redistribution are moral, not scientific ones, and this book does nothing to change that.  Don’t let any academic tell you otherwise.

Luke Malpass is Policy Analyst at The Centre for Independent studies. Professor Peter Saunders is Senior Fellow at the CIS and is author of the book When Prophecy Fails: A Critique of the Spirit Level.