A new study makes a wildly exaggerated claim
A new study predicts that Britain’s cost-of-living crisis will cause a sharp rise in mortality. The study has prompted various alarming headlines, such as one in the Guardian that refers to “thousands of premature deaths”. But a closer look suggests that its predictions are unlikely to materialise.
Elizabeth Richardson and colleagues use data from Scotland to model how inflation will affect real household incomes, and how this in turn will affect mortality. Yet one of their key assumptions is highly questionable. In particular, they assume that the effect of income on mortality is given by the cross-sectional relationship between the two variables.
What does this mean? When researchers calculate mortality rates for different income groups at a point in time, they find lower rates in the upper income groups. Richardson and colleagues are assuming this reflects a causal effect of income. Hence they predict that a fall in income of, say, £2,000 will increase mortality by the same amount as the difference in mortality between groups whose average incomes differ by £2,000.
But just because mortality rates are lower in the upper income groups, doesn’t mean higher average income is the cause. Perhaps being in good health leads people to earn a high income. Or perhaps the traits that predispose people to earn a high income also predispose them to live longer.
One way to find out whether income really does have a causal effect on mortality is to follow lottery winners. If they go on to live longer than people who played the lottery but did not win, it’s likely because the extra income caused them to live longer — since winning the lottery is determined by chance. David Cesarini and colleagues used this method in a 2016 study and found “no evidence that wealth impacts mortality”.
Zooming out to the macro-level, there’s no evidence that mortality tends to rise when living standards go down. Paradoxical as it may seem, mortality rates usually fall during economic recessions (though the relationship has become weaker over time).
We can check this for the 2008 financial crisis — when Scotland’s GDP contracted by 4% and the unemployment rate doubled. Looking at figures published by the Scottish government, we can see that the age-standardised mortality rate fell from 2007 to 2008, and then again from 2008 to 2009.
Even if Scotland does see a rise in mortality this year, it’s very unlikely to be as large as Richardson and colleagues predict. They claim that “real-terms income reductions could result in population-wide premature mortality increases of up to 6.4%”. Yet the increase due to Covid in 2020 was ‘only’ 7.4%. So what they’re saying is that the cost-of-living crisis will increase mortality by almost as much as the pandemic did — which I don’t find plausible.
Will the cost-of-living crisis lead to “thousands of premature deaths”? There are strong reasons to doubt Richardson and colleagues’ projections. Of course, this isn’t meant to downplay the crisis: inflation doesn’t have to kill you to be a serious problem.