Beware the seductive simplicity of the Spirit Level

By Matt Nolan 19/05/2014


I see that the Spirit Level authors are in town, and as a result there was a recent Herald article took aim at income inequality in New Zealand, relying strongly on the book ‘The Spirit Level’.  A conversation about the inequalities society believes are fair, or at least justifiable, is a good thing.  However, the Spirit Level’s claims that simply targeting measures like the Gini coefficient will make everyone better off is a misleading, and dangerous, place to start this conversation.

In their initial book Wilkinson and Pikett make the claim that the relative distance between incomes (which they in turn call inequality) in a country/region causes a variety of social ills (worse health outcomes, higher crime rates, etc).  They stated that this implies everyone, even those with higher tax burdens, would be better off if we increased taxes and transfers and lowered income inequality.

When I initially reviewed the book I found that their claims were significantly oversold, the book was filled with inconsistencies, and their policy conclusions were unjustified.  This disappointed me, not because I think we should ignore inequality, but because I believe that asking why income inequality has changed and who has been hurt is an incredibly important question – one that has not been given enough attention.

It turns out that there are a number of left-leaning economists found the claims oversold.  For example, in the Oxford Economic Handbook of Economic Inequality, three authors (Leigh, Jencks, and Smeeding) point out that a relationship between health outcomes and inequality does not seem to exist.

Many economists view of inequality tends to be based on a more operational view about policy.  In the mid-1990s, the economist Serge-Christophe Kolm said “Few concepts are as meaninglessly used as that of inequality”.  However, he didn’t say this because inequality didn’t matter – to give it context in 1976 he said:

“Many people consider the reduction of economic inequalites as a basic aim of society. Such ideas are, however, largely nonoperational, sterile, and even meaningless, as long as what is called inequality is not stated with precision. This is so because, …  different measures of inequality give widely different, and even opposite, results. Such policy which diminishes some apparently reasonable measure increases other ones.”

Wilkinson and Pikett focus on one measure of income inequality – the Gini coefficient.  This measure ranks incomes based on their distance from each other and tells us nothing about why we may value this distance, or why the distance may have changed through time.  To ensure we determine policy that helps to deal with real injustices in society, we need to answer these other questions – not compare an arbitrary measure of the distance between incomes with another arbitrary measure of social outcomes.

Since then, Wilkinson states in the Herald article that they have become more focused on the idea of “status competition” as the driver of this result.  This is more commonly called “keeping up with the Joneses”.

However, if this is the driver of our concern about inequality, we have to be careful when thinking about policy.  Status goods matter with reference to the group you view yourself in, and other groups you perceive you may be competing with.  If upper middle class people are competing for status with larger cars, how does this impact upon lower socio-economic groups at all?  The fact is that it often doesn’t – these groups have their own sets of status goods, and it is inequality within these groups that can lead to outcomes such as excessive borrowing and excessive stress.

However, it is far worse than that.  If status competition is the issue, then transferring income may simply lead to wasteful bidding up in prices for status goods for those on low incomes, and increased wasteful status competition in terms of non-monetary factors (types of ‘superstar’ jobs, types of investments made) by higher socioeconomic groups.  Status competition is inherent in the nature of individuals and groups – and it is only by understanding those specific groups that we can design policy.

I realise that my comments will be unpopular for those who do want an easy and marketable solution for social injustice, and for colleagues of mine who believe that such simplicity is a good way of getting social justice on the table as a talking point.  However, the price we pay for such simplicity in a subject that requires complicated ethical judgements is far too high.  When discussing targeting income inequality, Amartya Sen stated:

As a result of that assumption, we are made to overlook the substantive inequalities in, say, well-being and freedom that may directly result from an equal distribution of incomes (given our variable needs and disparate personal and social circumstances).

Many real injustices are the result of “variable needs and disparate personal and social circumstances”.  If we want to discuss fairness, this is a better place to start than a Gini coefficient.

Ultimately, policies that help to improve social justice might lower the Gini coefficient – but this in no way implies that policies that aim to lower the Gini coefficient are just.