Do you remember the Spirit Level, written by Kate Pickett and Richard Wilkinson in 2009? It advanced what is still widely seen as an intrinsically plausible theory that more equal societies ‘perform better’ than less equal societies on all sorts of outcome measures that reflect individual well-being. Higher life expectancy, fewer suicides, fewer teenage pregnancies, better child well-being, less mental illness and so on.
Pickett and Wilkinson were concerned with differences between rich societies, and not with the differences between the rich world and the poor world. The claim is that once societies achieve a certain level of affluence (measured, say, by GDP per capita), additional increments of income (measured again at population level) have little or no association with improvements in population health. To understand why rich one society does better than another, we have to take account of the distribution of income within these societies. To couch the matter in terms of life expectancy, the theory starts from the observation of a correlation between mortality risk and income inequality at country level, and proposes the existence of an underlying (causal?) relationship between an individual’s mortality risk and societal income inequality in order to explain the observation. This is known as the income inequality hypothesis, or sometimes the relative income hypothesis, to differentiate it from the view that individual income affects mortality risk, i.e. the absolute income hypothesis. Income inequality is regarded as a health risk in its own right. My chances of living a long and healthy life are affected not only by what I have (and can obtain), but also by what other people have.
As long ago as 1998, it was pointed out by Hugh Gravelle there were serious problems with testing the income inequality hypothesis, namely that the observation could be explained as an artefactual consequence of something rather like the diminishing marginal utility of income. If the effect of individual income on individual mortality risk is smaller at higher incomes than at lower incomes, statistical analysis of population-level data will find a correlation between aggregate mortality and income distribution even if the income inequality hypothesis is wrong.
Despite a BMJ editorial in 2002, which argued that the evidence for the income inequality hypothesis had been unable to withstand closer analysis, the arguments (and the research studies that fuel them) have rumbled on for quite a few years now, and it would be rash to suggest that any single contribution to such a hotly contested topic could be decisive. The publication of The Spirit Level may be seen as a riposte to those who concluded that the hypothesis had been permanently put out of action[i]. For myself, I should confess that I am more inclined to find the hypothesis intrinsically odd rather than intrinsically plausible, and it is in that spirit of less than strict impartiality that I want to draw attention to a paper published last year in the European Journal of Epidemiology. Like the Spirit Level itself and the Gravelle paper[ii] and the Mackenbach editorial[iii], it’s worth the read, along with a very useful editorial discussing the results.
The starting point for the study by Hu et al[iv] is that we should not expect much more clarification from new cross-sectional studies. Most such studies tend to confirm the presence of an association, and sometimes this ceases to be significant if controls are introduced for various kinds of country difference. Sometimes, however, even though the controls weaken the effect, they do not remove it altogether. What Hu and colleagues do is use longitudinal data (1987-2008) for 43 European countries in a fixed effects model. In other words, each country acts as its own control. The approach maps the changes in mortality levels over time in any given country onto the changes in income inequality seen in that same country. They also introduce a number of ‘potential confounders’ into the analysis, and as they explain, the ‘trick’ is to select potential confounding variables that are related to both mortality and income inequality without mediating the relation between them. The conclusions in the paper are stated with some care. It is stressed that the analysis only applies to Europe – and maybe they didn’t get it quite right with their choice of confounding variables. “Within Europe cross-sectional associations between income inequality and mortality probably result from confounding”. In other words, “the reduction of income inequality ... has a limited role for reducing average mortality in Europe” (my italics).
The accompanying editorial[v], which also takes in a couple of other papers on the same question, puts it this way:
“Today, despite renewed public interest for the increasing income inequality trend in our societies, and the collective view that high income inequality is undesirable from a moral or societal point of view, there still exists no strong evidence base to argue that tackling income inequality would be an effective strategy to improve population health”.
[ii] Gravelle H. How much of the relation between population mortality and unequal distribution of income is a statistical artefact? BMJ, 1998, 316: 382-385.
[iii] BMJ 2002;324:1
[iv] Hu Y, van Lenthe FJ, Mackenback JP. Income inequality, life expectancy and cause-specific mortality in 43 European countries, 1987-2008: a fixed effects study. Eur J Epidemiology, 2015, 30:615-625.
[v] Avendano M, Hessel P. The income inequality hypothesis rejected? Eur J Epidemiology, 2015, 30:595-598.
--------
About the Author:
Kenneth Howse is a Senior Research Fellow at the Oxford Institute of Population Ageing. He is also a key member of The Collen Programme on Fertility, Education and the Environment.
--------
Comments Welcome:
We welcome your comments on this or any of the Institute's blog posts. Please feel free to email comments to be posted on your behalf to administrator@ageing.ox.ac.uk or use the Disqus facility linked below.
Opinions of the blogger is their own and not endorsed by the Institute
Comments Welcome: We welcome your comments on this or any of the Institute's blog posts. Please feel free to email comments to be posted on your behalf to administrator@ageing.ox.ac.uk or use the Disqus facility linked below.
