Income, health, and payday deaths

It’s hardly news that poorer people have worse health on average – but teasing apart the link of income and health is harder. The income-health gradient could be because people lose income when they have health problems, or it could be due to common causes (e.g. education) rather than income itself. In fact, the relationship is more complex than you might guess; and having recently stumbled over three wonderful studies that help unpick this, I thought I’d share them with you and see what you thought.

Short-run effects: payday deaths

Probably my favourite of these studies is a lovely paper in press by Evans and Moore, using US data. Their reasoning is that if we want to know the SHORT-RUN impact of income and health, we can look at the first day of the month when most people get paid or receive welfare payments, and see if people are more or less likely to die on payday.  The results are staggering.

In the chart below, you can see how murder rates vary across the month.  The relate daily mortality risk is about 10% higher on payday than it is for the rest of the month, and a little bit higher for 4 days after payday.

What’s particularly striking is that this is only found for causes of death that are related to drink in the short-term.  So for example, if we look at leukemia deaths in the chart below, we see absolutely no change over the month at all.

In other words, people drink more on payday, and this raises the chances of them dying. Otherwise there’s no short-run impact of payday on health.

Long-run effects: lottery winnings

Does this mean that higher incomes are actually bad for health?  Well, not so fast.  The deaths-by-payday charts are good for showing the very short-run impacts of income on health, but they don’t show what happens to a change in permanent income, nor do they show the longer-run impacts. The best way of getting at the causal impact of income here is to look at lottery winnings, which (unlike most sources of income) are completely random within any particular lottery.

In the UK, a paper by Apouey & Clark 2009 looks at the impact of lottery winnings on health. Like Evans and Moore, they find that this source of extra income leads to more smoking and drinking (perhaps unsurprisingly, if you think about how you personally would spend a lottery winning (or me anyway), and also the fact that people who play the lottery are going to have a different attitude to risk than other people).

Yet we also see noticeable improvements in mental health among people who win on the lottery – suggesting that financial problems genuinely cause mental health problems.

The net effect of these two patterns in Apouey & Clark is that there’s no effect of lottery winnings on self-reported general health – but while self-reported health can be useful, it also has lots of problems as a measure. A more objective measure of general health is mortality, and for this we have to turn to some Swedish data analysed by Lindahl 2005 (free version here). Lindahl finds that about £1,000 worth of lottery winnings reduce the probability of dying by 2-3% over the next five years. With this design, it’s pretty unarguable that this shows a genuine causal impact of income.

Back full circle

So this takes us right back to where we started – poor people have worse health, and this is (at least partly) because lower incomes genuinely cause worse health. But through these three very nicely-designed studies, we get a glimpse into the complexity of this pattern across different aspects of health over different time periods – and it’s this that I thought made them particularly worth sharing.

3 responses to “Income, health, and payday deaths”

  1. Nice, Ben. I think it’s worth mentioning Christopher Ruhm’s research that shows that mortality declined during the recessions of the 1980s and 1990s (and conversely increased during the booms). He finds evidence for decreased traffic mortalities and reduced drinking and smoking. So there’s something to this.

    But I’m still not convinced. This may be one of those instances where studying exogenous shocks actually gives us the “wrong” answer — or at least isolates unusual causal mechanisms. We should not try to estimate the effect of extra income independent of education and cultural background, we should try to understand how those other factors allow income to productively get used toward producing health. It’s no surprise that mortality spikes around paydays — but the more important question is what kinds of norms and attitudes support binge drinking? Spikes in income in the context of poverty means something different than income that is produced by developing human capital and skills in different institutional regimes. That’s why I think cross-national research is actually more informative, for all its limitations. Love to get people’s thoughts.

    • Thanks Brendan – yes, I found out about the Evans/Moore study from a review co-authored by Ruhm, but I probably should have mentioned his stuff too.

      More importantly, I think your main comment is crucial. One of the problems is that we think about ‘the effect of income on health’ – but I completely agree with you that effects are likely to depend on context, and vary hugely between one another (as well as impacting different aspects of health differently). Cross-national research is helpful in providing pointers, but it’s challenging to isolate exactly what matters here… I still think that the best way to look at this is to focus on mechanisms in particular instances; and qualitative research may help here (e.g. on how people react to different changes in income, and all the sociological literature on the many meanings of money – see Viviana Zelizer’s work). Or is trying to ‘isolate what matters’ missing the point?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: