Life Expectancy and Wealth in the UK, 2004
I’m doing some research work at the moment on a certain country, and have identified a negative relationship between inequality (measured using the Gini index) and all-cause mortality. I’m not at liberty to identify the country, or the details, of course. This negative relationship means, in essence, that the more unequal a small area becomes, the lower the death rate in that area. The effect is very very weak, so essentially can be dismissed as a policy concern (other aspects of the area’s economy drive much greater increases or reductions in mortality).
Typically, when a result like this occurs, people will dismiss it as either a statistical oddity or will say that the inequality effect is operating as a proxy for other phenomena – for example, the observed values of inequality measure the effect of past rapid economic growth, or are a proxy for some other aspect of the economy or society (social capital, corruption, etc.). The reason for this is simple: it’s very hard to construct a mechanism by which increasing inequality can reduce mortality, but it’s easy to say that inequality increases as wealth does, and that increasing wealth reduces mortality (a fairly solid fact). There is a model for how inequality can increase mortality, through mechanisms such as reduced social welfare support, or inability to seize necessary cultural or physical capital – there are quite a few convoluted mechanisms proposed by which relative inequality can have this effect in wealthy countries like the UK.
I’m a realist in this regard, and I don’t think relative inequality is important to health – I think it’s absolute inequality that determines health, in the vast majority of cases – thus in the short term development is more important than equality (or, in many cases, basic rights – see e.g. China). But this data I’m playing with measures relative inequality – that is, it measures the distribution of wealth, and looks at the relationship between this distribution and mortality after adjusting for absolute wealth (how rich an area is).
So, here’s a challenge to my reader(s): can you propose a social mechanism by which increased relative inequality reduces mortality? This mechanism needs to work across regions with a diverse range of income levels and other forms of social determinants of health. Imagine it in your own country, if you like: whether you’re living in Sao Paulo or on the edge of the Amazon, greater disparity in wealth will reduce mortality. Can you explain how this happens without falling back on arguments that a) the measure of inequality is a proxy for something else or b) higher absolute wealth generates higher relative inequality.
Have at it, if you dare!
Compromise and Conceit 
July 20, 2012 at 8:29 am
OK. Challenge accepted. 5 options have occurred to me so far. In order of likelihood and how well these match your criteria these are:
1. Low income levels forcing poor people to avoid goods with negative health impacts. This would be observed if the absolute level of income of the poor fell to the point where they could not consume Western/modern goods with negative health impacts (i.e. chips, cigarettes, anything badly made with heavy metals in them). If coupled with the rich reaching incomes where they move away from such goods due to awareness then you could end up with everyone either being forced or choosing to avoid things that are bad for you. An example of this would be if an American city had such inequality that the poor couldn’t afford smokes and the rich knew not to buy them (or could afford the cancer treatments). This effect would also be present in less developed countries by preventing the poor from buying into an unhealthy (i.e. lower class Western life style) while enabling the rich to move to a better life style (i.e. Western upper middle class).
1a) This could also be a result of the decreased wealth forcing more sensible decisions on the poor people in the sample. This is the answer that comes closest to expressing an Objectivist viewpoint. Bonus points for anyone who can find a Marxist suggestion for why inequality is good. Wait. I just thought of one. I’ll put it in as idea 4.
2. Social obligation could be driving health patronage. This would suggest that as people become wealthier they are compelled to provide untracked hands outs to the poor people around them. An increase in inequality would sharpen the social obligation. Examples of this would be medieval public works provided by rich merchants to increase their public standing or Bill Gates deciding that he’s rich enough and is willing to share the wealth. In effect this is a reduction in inequality in a way you can’t track.
3. An expansion on the first idea would be if poverty drove a move away from low efficiency health treatments towards medium efficiency treatments. An example of this would be poor people living near natural resources avoiding expensive aspirin and instead using natural herbs. In effect I’m pretty dubious about this one as I’m not a believer in “natural” cures. The effect would probably happen most reliably if the poor were just forced to not go to unhelpful healers (i.e. too poor to go to a witch doctor or faith healer [1]). The scenarios for this are pretty marginal in the real world.
4. Marxism. The revolution could finally have arrived. As is traditional this would lead to massive suffering and reduced income for most while the leaders start to live in disgusting levels of comfort. If the leader took a break from putting people up against the wall long enough to establish a health system (enforced by shooting doctors who don’t work for free) then this may achieve the objective. The downside to this theory is you need to be sampling large areas to have the leaders count in the same bucket as the plebs and I don’t recall hearing of any recent (successful) Marxist uprisings.
5. Religion. If the inequality is due to religiously mandated income levels then the equality effects could be drowned by the health benefits of a peaceful monastic life. Examples of this would be comparing how the Pope lives to how monks live in the Vatican. I suspect that it would also apply in other religious movements where the leader gets a life of luxury and the followers get a life of prayer.
What sort of timeframes did the data contain? Is it possible that there was a die off of poor people at the start of the time or just before it, then inequality increased afterward (possibly as an effect of the die off)? That would make it a just a temporally based anomaly [2].
[1] Why does no one consider the impact of economic downturns on the witch doctors!? Though I’ll bet that funding to homeopathy remained stable during the economic downturn in the UK, so at least they’re ensuring the witch doctors don’t suffer.
[2] I’m pretty sure that the crew of the Enterprise can clean up any temporal anomalies you may want addressed…
July 20, 2012 at 12:10 pm
How about an imitation effect?
Maybe similar to Paul’s first point. Here, the relatively poorer people try to imitate the lifestyle of the richer people. This might include some positive health benefits. For example, it becomes trendy to eat more healthily among the rich because of external influences (media, etc.). The poor, wanting to identify with the rich, start picking up on those habits to the best of their ability.
Another possibility, related to Paul’s second point, is that a relative increase in inequality leads to an decrease in the spread of power, so decisions can be made more quickly. Imagine a hospital needs to be built. In the past, all the stakeholders had more similar incomes so the power of each person’s vote was the same, and because each stakeholder had different preferences no hospital was built. Now that it’s more unequal, some people have more voting power than others (because, say, they invest more in the hospital than others), meaning that the hospital can be built more quickly, even if it’s to more closely meet the preferences of a few people. We would expect positive spillover effects to the other, poorer stakeholders.
July 20, 2012 at 2:48 pm
Another possibility is that there is some finite bound of poverty past which your health doesn’t suffer. In that case both absolute and relative decreases in income would not make the poor worse off while the increase accruing to the wealthy leads to overall improvements.
This would be possible in any setting where subsistence farming and marginal city existence are close by but marginal life styles in the city register as “better off” due to material advantages that don’t impact health (i.e. having a TV doesn’t make you healthier, so losing it and becoming a poor farmer doesn’t impact your health).
July 21, 2012 at 6:51 pm
I’ve updated this post with a simple graph of life expectancy (LE) and disability free life expectancy (DFLE) at small area level in the UK, ordered by the average wealth of the area as measured by Index of Material Deprivation (the British preferred measure of wealth). That shows pretty clearly the relationship between inequality and health, where wealth is measured in absolute terms. Britain has higher inequality than its neighbours in Europe, and lower life expectancy, but when you look at small areas in Britain the poor health is clearly related to absolute wealth.
Thanks for the comments, Alf and Paul. Starting from Paul’s comments:
1. This is likely a non-starter. In almost every country poorer people have higher prevalence of the risk factors for poor health, and there are few places in the world where the rich have more vices than the poor. Maybe at King’s Landing, but that’s about it. In most developed nations, bad health behavior is strongly linked to poverty, and by the time people are so poor that they can’t afford any vices, they are malnourished …
2. I think there are theories that high inequality leads to hidden economies and the like, and patronage such as described here would be a high possibility in a very unequal society. But usually I think this is scene as ameliorating the effects of inequality, rather than reversing them.
3. Maybe in King’s Landing
4. The most successful commie revolution, China, is only just getting around to a universal health system. Amongst the failed models, the USSR had high mortality even when its inequality was very low. I guess the big success in this regard was Cuba, which had much better outcomes than wealthier and less equal countries in its region. But I think everyone would probably agree that it’s a bit of an outlier, and anyway marxist societies, even though authoritarian, usually aren’t very unequal in the numerical sense in which we measure these things…
5. That one would work. I can safely say that I’m not modeling Bhutan, though!
Alf, I think if poor people are able to mimic the lifestyle of rich people there’s not much inequality to speak of, so that probably wouldn’t work. And the decreased spread of power is often suggested as a reason why relative inequality is important even when the poor are wealthy, becasue this decreased spread of power (such as we’re observing in the USA now) enables the rich to seize political and economic resources that secure their rights over others. The most obvious case given for the US is the inability of governments to introduce universal health care systems, which is often said to be because the rich lose out from such schemes, and powerful corporate interests (the insurance lobby) are opposed. So it seems unlikely that this concentration of power would work for poor people unless on were talking about a benign dictatorship. Maybe this happened in Indonesia under Suharto, when they had subsidies for petrol? (I think attempts to reduce those subsidies were a large part of the spark that led to Suharto’s downfall). But in general I don’t think this concentration of power works in favour of poor people…
So, though you guys get marks for effort, I don’t think you’ve presented a convincing explanation yet!!!
July 22, 2012 at 11:50 am
“That one would work. I can safely say that I’m not modeling Bhutan, though!”
Actually, I think that Vatican is about the only place where I’d really expect to see the theory borne out… And even there you have to decide if the pope, as leader of the Catholic faith, is immensely rich while the monks who live next door are immensely poor. After all, they both walk past the same Renaissance masterpieces each day…
“So, though you guys get marks for effort, I don’t think you’ve presented a convincing explanation yet!!!”
What about my second post? That one still looks good to me, though it’d only occur where the health starting point is low to start with and subsistence farming is as reliable for nutrition as marginal city labour.
The real problem I’ve got with suggesting ideas is that most of them are based on real world expectations, which necessarily line up with actual real world places. That constrains the theories to work in a place that really exists, and none of these theories are omni-applicable.
But lets get on to the more fun part. What’s your explanation then mister smarty pants? Are you willing to bet that I can’t find a contrary example that suggests its wrong?
July 22, 2012 at 6:08 pm
I think the problem with that second comment is that it ignores all the distributional effects in between the suggested threshold and the median or the high income earners. Unless the distribution of inequality is really weird (a huge number below the threshold and everyone else well above it) I can’t see it explaining things.
The truth is that I don’t have an explanation for this. I don’t think it’s possible that inequality can reduce mortality. I think inequality can be a proxy for something else, most obviously economic growth or just general vast wealth, but I can’t see how inequality can improve health through a direct mechanism. I think it’s almost as difficult to explain how inequality (as opposed to people being poor) can increase mortality, though. The main possible explanation is that high inequality is often associated with social breakdown, lack of basic public services and the denial of proper social engagement for a large class of people. But such an explanation is often confounded by the simple fact that in such societies large numbers of people are living in absolute as well as relative poverty. I can see that this could be a problem even without large numbers of people being in poverty, but the reality is that no such society exists. Societies where few people live in poverty tend, also, in general, to be more equal. I don’t think the effects of relative inequality can be separated from that simple fact.
July 23, 2012 at 11:44 am
If you don’t think it can and the data shows that it does then either you’re bad at gathering data or you’re bad at thinking…
Or you could subscribe to Fox and insist that you are right in the face of evidence the contrary.
July 26, 2012 at 8:32 pm
That’s what sociologists are for, Paul! I honestly can’t think of a mechanism that isn’t an explanation by proxy. e.g. inequality is a marker of greater economic freedoms which mean more efficient health care blah blah. I can’t think of a direct explanation. And yes, if I found such a result I would be in a pickle – but that’s when you say “more research is needed” and apply for a grant (can you imagine that grant application!?)
July 27, 2012 at 10:19 am
Grant application
Title: The Republicians are right
Details: Recent data suggests that the oppressed workers of the world can improve their health outcomes through further oppression. This research project will test if the old adage of “Treat ’em mean and keep ’em keen” can be applied to the health sphere.
Probable impacts of this research are:
1. Gloating in North Korea as the resident dictator orders himself another yacht
2. A rash of republican ads savaging Obamacare with well researched arguments to back them up