A recent paper in Social Science & Medicine makes me wonder whether editorial boards of health journals ever bother including economists as referees.
In Shifting dollars, saving lives (working paper version for those without subscriptions), Kiwi public health profs Blakely and Wilson notice that life expectancy correlates with income but with diminishing marginal returns. They consequently call for massive income redistribution as corrective policy that would reduce overall mortality. They're particularly keen on the expected reduction in the inequality of mortality that would result -- the shifting of mortality risk from the formerly poor to the formerly rich.
The problems in their approach are pretty obvious.
While they correct for age, marital status and education, wouldn't we still expect that unobserved underlying variables might cause both mortality risk and income? We've pretty strong evidence that IQ causes both income and health outcomes: see Gottfredson, for example, on health; Bryan's often talked about IQ and income. Similarly, conscientiousness would correlate positively with both income and health outcomes. If it's the case that underlying variables like intelligence determine both income and health outcomes, then the effect of income on health will be overestimated when IQ's left out of the specification.
They note research findings that transient poverty is less predictive of health outcomes than long-term income. Mightn't this strengthen our expectation that an underlying variable determines both but with random shocks in any particular period.
Doesn't wide-scale redistribution generally result in rather large reductions in economic growth? They nod to the likely deadweight losses but, viewing them as too complicated to deal with, they "used the simplifying assumption that there is no overall change to total income." Unbelievable. They later note that redistribution could be funded via Pigovean excise taxes on things like tobacco and alcohol. Of course, the incidence of these taxes falls primarily on the poor! While one might make the argument that changing the relative prices faced by the poor for different goods via a Pigovean taxation mechanism might have effects on mortality, arguing that it could be the basis for an income equalisation scheme is ridiculous.
The final policy conclusion?
the estimates in this paper tell us that, yes, income redistribution will probably result in a modest reduction in overall mortality rates -- but not a large one, especially if the dead-weight costs of redistributing income are significant. They also tell us that income redistribution should reduce relative inequalities in mortality.
My conclusion: until deadweight costs of taxation, expected increases in mortality due to reductions in total income, and some way of correcting for the effect of IQ on both health and income are added in, it would be ridiculous to advance policy conclusions based on their study.
Does it even make sense to worry about relative inequalities in mortality? It makes sense to worry that some folks have low life expectancy, but the notion of redistributing life expectancy seems repugnant. Why not recommend the forcible transfer of kidneys from the healthy to the sick?