By JR on Friday, September 09, 2011
As the British parliament rose for its summer recess this year, Opposition Leader Ed Miliband handed the members of his shadow cabinet some holiday homework. He told them to read a book that has been capturing the attention of the Left, not only in Britain but across the Western world.
Written by a couple of socialist academics, Richard Wilkinson and Kate Pickett, the book is called The Spirit Level. The clue to why so many on the Left have been drawn to it is in the subtitle: Why Equality is Better for Everyone.
This book seeks to reinstate radical income redistribution at the heart of the Left's political agenda. Socialists always have believed in greater equality, of course, but until now their case has rested on an ethical principle that it is morally wrong for some people to have a lot more money than others.
As with all ethical principles, this can be challenged. Why should people who work hard have the fruits of their labour taken away to be given to lazier folk, for example? The Right points out that hard work and risk deserve reward, and equalising shares can be quite immoral.
The Spirit Level aims to break away from these ethical conundrums and to replace them with the authority of science. It says governments should redistribute incomes, not because it is moral but because equality produces happier people and better-functioning societies. It claims everybody stands to benefit from income redistribution, rich and poor alike.
If this claim were true, it would pull the rug from under the feet of the Right. If a radical redistribution of income and wealth really did benefit everybody, how could the Right continue to hold out against it? The case for high taxes, big government and massive income transfers would be unanswerable. But it's not true. This book has many flaws (even though Miliband, and others on the Left appear blind to them).
The book's evidence consists of a series of graphs apparently showing that people in more equal countries live longer, are less likely to get murdered, enjoy higher literacy rates, suffer less mental illness and trust each other more. These findings are repeated for the 50 US states, where the authors find that states with the widest income spread have worse outcomes. But little of this evidence stands up to critical scrutiny.
Their sample of countries is biased. It excludes nations such as South Korea, where strong social outcomes coexist with high income inequality, as well as those such as the Czech Republic, with poor social outcomes despite a compressed income distribution.
Their choice of measures is also biased. Community strength is measured by whether people say they trust their neighbours, but membership of voluntary organisations is ignored. Drug dependency is included as an indicator of social pathology but not alcohol abuse. Murders likewise are in, but suicides are out. Prison numbers are analysed, but not crime figures. Government aid to foreign countries is included as a measure of generosity and compassion, but not private donations to charities. High teenage births are analysed as an indicator of family dysfunction, but not high divorce rates.
What is striking about this list of inclusions and exclusions is that, in every case, the measures that Wilkinson and Pickett selected fit their argument while the alternative measures would have undermined it. In short, they cherry-picked.
Their data analysis, too, is suspect, for they allow extreme cases to create the appearance of an association where there is none. For example, they claim that inequality produces a higher homicide rate, but this depends entirely on the US, where the murder rate is three times higher than anywhere else. Look beyond the US and you often find the most equal countries, such as Sweden and Finland, have a worse murder rate than less equal ones, including Britain and Australia. Yet appealing to their misleading graph, the authors claim Britain's murder rate would be three times lower if it had Scandinavian levels of income inequality.
The Scandinavians, it is true, do fare better than the "Anglo" countries on many of their measures, but this is not because inequality is lower in Scandinavia. It rather reflects the homogeneity of the Nordic countries as against the diversity of the Anglo nations, for the greater the social mix, the weaker the social bonds tend to be. We see this clearly in the variations between US states. Wilkinson and Pickett find the more equal states (usually those in the northeast) do better than the less equal ones (concentrated in the south).
But had they taken account of the ethnic mix of these states, they would have found ethnicity is a much stronger predictor of social outcomes than income distribution. Ethnicity is 18 times more powerful in predicting a state's infant mortality rate, for example.
The issue of equality is important, and it generates strong and impassioned arguments on both sides. But The Spirit Level is little more than polemic. It is to be hoped that we do not allow its spurious claims to scientific status to muddy the waters of our political and moral debate.