Do more meritocratic societies have higher health inequality?

The impressive and influential programme of research comparing health inequality in different nations by Prof Mackenbach's team at Erasmus University has produced disappointing results for those of us who think stronger welfare institutions would lead to lower health inequality. I am not going to give a load of URLs to all their papers here, but almost 20 years ago they showed that inequality in mortality was at least as high in Sweden as in, for example, England and Wales. Since then, subsequent studies have looked at changes in inequalities in mortality (and other outcomes, but mortality is the one that seems the most reliable). And while these have not changed much in the Mediterranean, unequal nations and have actually fallen in neo-Liberal Britain, they have risen most in the Nordic nations. So not only did the Nordic nations always have a level of health inequality at least as great as others, but they have experienced more adverse trends.

One explanation that Prof Mackenbach came up with was in a paper subtitled "Now its Personal". In this paper he suggested that maybe the greater social mobility in the nations with stronger welfare states and less unequal income distributions was the answer to this puzzle. Maybe in the more mobile, meritocratic nations, the people who end up in the worst, lowest paid lowest status jobs are those with personal characteristics that have held them back, for example, in education. So no matter where a person begins their life, the less intelligent and conscientious (for example) they are, the more likely they are to end up in a socio-economically less advantaged position. And because of their personal failings, these people are also hypothesised to be less able to understand and implement messages about risky health behaviour. So the causal link is not between socio-economic circumstances and health at all. On the contrary, it is personal characteristics that 'cause' both social position and (via their link with health behaviours) health.

When discussing this with a Swedish epidemiology colleague, he said it can't be true. Sweden has very high social mobility so it should have very high health inequality. My colleague is a clinical epidemiologist and not familiar with the Erasmus work, which has actually shown just that. But I thought OK lets have a look at some other nations.

So I have put together a little graph that takes its measure of health inequality from a recent paper by Prof Mackenbach's group, and its measure of social mobility from an OECD paper on this topic that got a lot of media attention in 2010. By luck, both data sources deal with roughly similar time periods. The measure of social mobility is called "income elasticity" which is the strength of the correlation between fathers' and sons' (usually) incomes. I high elasticity means these are "sticky" (at least that is how I think about it), fathers' incomes have a lot of influence so social mobility is low. And vice versa. And here is what it shows.

The linear correlation is -0.66, which is quite high as my vague memory of these things goes.

So what does everyone think? I don't like the implications of this, but neither do I agree with selective reporting of one's findings. This is a very, very quick and dirty exercise. But it would be really interesting to see what other reckon to it.