Friday, 27 January 2012

The Spirit Level debate

It is now three years since Richard Wilkinson and Kate Pickett's book The Spirit Level (TSL) was published, so enough time has now passed to allow the dust to settle on the debate that it generated.

The thesis of TSL is simple.  For most of history, and in many countries today, finding the basic necessities of life has been the principal challenge for most people.  In wealthy modern societies, however, this problem has largely been solved thanks to decades of economic development and growth.  The incidence of social problems, such as ill health and violence, no longer correlates with the general overall wealth of a society.  Instead, it correlates with something else - and that something is relative economic inequality within society.

This is not a new insight.  There is a substantial body of evidence going back many years on the correlation between inequality and ill health and inequality and violence.  However, TSL broke new ground in assembling a fairly comprehensive body of evidence linking inequality with a range of social pathologies, from homicides and mental illness to teenage pregnancies and prison population and from obesity to levels of numeracy and literacy.  The authors also argued that inequality of outcome is related to inequality of opportunity, and that a lack of the former depresses the latter.  In drawing these links, they drew both on international statistics and on figures from the different US states.

Interestingly, the authors also note that there has been a long-term rise in anxiety and depression which appears to be independent of inequality levels and which is attributable to the rise of mass society and the breakup of traditional small communities (which, of course, have their own problems and injustices).

Not surprisingly, TSL has been welcomed with open arms by the political left, but it is worth emphasising that its political implications are wide open to debate.  Great political issues can rarely be settled by graphs and statistics.  It is noteworthy that TSL has been praised by a number of figures on the right, including David Cameron, Michael Gove and writers in the Telegraph and the Economist.  Here are some reasons why the thesis of TSL need not be incompatible with political conservatism:
  • Economic equality is a key issue for the left, but it doesn't follow that inequality need be of no concern for people on the right - particularly if it is the source of social problems which affect potential Conservative voters.  In fact, there is a long history of concern about inequality among British Conservatives.  "One Nation" Toryism takes its name from Disraeli's complaint that the rich and the poor formed two separate nations.
  • Recognising that inequality is problematic does not commit us to a particular political agenda in tackling it.  The old-school socialist model of large-scale redistribution through high taxes and a big state is only one possible response.  Some of the most egalitarian societies in TSL's data sets (Japan, New Hampshire) have unusually low levels of social spending because their markets work in such a way as to distribute wealth equitably without the government getting involved.  Meanwhile, in other, less equal societies, problems which are worsened by inequality, like crime and ill health, end up requiring increased public spending.
  • Inequality only comes to the foreground as a source of social problems in wealthy societies.  It is still more important for developing countries to become rich than to become equal.  To this extent, TSL need not be incompatible with free-market policies on global trade and development.
Having said this, there was something of a knee-jerk response to TSL in some sections of the political right.  The assumption appears to have been that the research leads inexorably to the conclusion that radical redistributive policies ought to be put in place to engineer economic equality.  It is difficult to avoid the conclusion that criticism from this quarter was ideologically motivated.  To be fair, Wilkinson and Pickett themselves are open to this sort of criticism too.  While much of TSL is taken up with rather dry academic discussion, they do not succeed in maintaining a tone of Olympian scholarly objectivity in the book, and there are fairly clear indications that their political sympathies are on the left.  Indeed, there is external evidence that Richard Wilkinson's political commitments are some distance to the left of centre.

There appear to have been three major critiques of the book: Christopher Snowdon's The Spirit Level Delusion, Peter Saunders' Beware False Prophets, and The Spirit Illusion by three Swedish researchers.  Wilkinson and Pickett have produced a combined response to these criticisms (plus some additional material in the second edition of TSL), and the critics have in turn produced counter-responses to the response.

There are basically two categories of criticism of TSL.  The first type of criticism seems to me to be questionable, while the second seems to be valid, at least in part.  The overall result, in my opinion, is that the thesis of TSL is weakened but not discredited.

Critiques in the first category focus on technical questions relating to Wilkinson and Pickett's data.  Sometimes, the critics suggest that the authors ought to have used data from more, or fewer, or different countries.  Sometimes, they argue that they have wrongly identified correlations between inequality and specific social problems, drawn on inappropriate data sources or ignored researchers who have arrived at different conclusions.  Wilkinson and Pickett reject these criticisms and accuse their critics of wanting to move the evidential goalposts because they don't like their conclusions.  They affirm that statistical relationships between inequality and social problems have been repeatedly discovered and confirmed by other researchers using different data sets, and that TSL predicted the results of research that was published after the book came out.

A non-specialist like me isn't in a position to adjudicate this dispute, but my impression is that the criticisms made under this heading don't fundamentally affect the validity of TSL's thesis.  The natural suspicion is that both sides of the debate are, to whatever degree, trying to cherry-pick the evidence and shift the goalposts in their preferred direction.  However, the evidence cited by Wilkinson and Pickett, taken as a whole, is so substantial and varied that the technical criticisms made under this heading don't seem sufficient to refute it.

A more substantial criticism is that TSL mostly ignores the history and culture of different societies.  It has been argued that incidences of the social problems identified by Wilkinson and Pickett correlate not (only) with economic inequality but with international cultural groupings: the Scandinavian countries often appear at one end of the graphs in TSL and the English-speaking countries at the other.  It has similarly been argued that the divergences between different US states are explicable in terms of differences between Northern and Southern culture and the culture of caucasians and African Americans.  It has also been argued that east Asian countries tend to cluster together despite having differing levels of inequality.

Part of the issue here is whether high levels of social cohesion are the result of economic equality or a cause of it.  Peter Saunders writes:
Sweden and Japan... have the income distributions they have because of the kinds of societies they are. They are not cohesive societies because their incomes are equally distributed; their incomes are equally distributed because they evolved as remarkably cohesive societies. To explain why, we have to look to their histories and at factors like social homogeneity and closure, but The Spirit Level resists any such analysis.
Saunders has a partisan political agenda of his own, but these are reasonable criticisms.  A simple causal model which singles out economic inequality as the cause of multifarious social problems is less attractive than a web of causation which posits different factors, including cultural ones, that influence and interact with each other.  Wilkinson and Pickett succeed in establishing the basic point that economic inequality is associated with undesirable things happening in society - that much is difficult to deny in the light of the evidence - but inequality should not be used as a simple monocausal explanatory key for complex social problems, any more than simplistic ideas about sex differences should be used to promote a Mars-and-Venus model of human gender.  To be fair, Wilkinson and Pickett insist that TSL isn't supposed to be a "theory of everything".  But they do focus sharply on inequality without contextualising it, and they are much too quick to downplay the importance of historical and cultural factors.  This approach is perhaps a reflection of their backgrounds, which are in "hard" science (epidemiology) rather than anthropology or the humanities.

So TSL ultimately succeeds in establishing its central thesis, that there is a relationship between inequality and social dysfunction, albeit we are entitled to question whether this relationship is always strong or straightforward.  It is worth reiterating that this does not close the debate between left and right on economic policy, or indeed anything else.  The data presented by Wilkinson and Pickett, however, deserves to shape political thinking on a range of key contemporary issues.  It cannot be ignored.