From Eagan et al, comparison of actual GDP to health-adjusted GDP, using selective VSL estimates that are way too high.A few recent studies have me puzzled. The first is Eagan, Mulligan, and Phillipson (2014) on mortality over the business cycle, which documents that the mortality rate is pro-cyclical, with deaths increasing in expansions and decreasing in recessions. The effect was so large in 2009 that the health value of the lives saved offset a majority of the loss to GDP! Eagan et al aren't the first to note this trend and some plausible theories have been proposed to explain it: work-place injuries and commuter traffic accidents, for example, increase when more people have jobs, resulting in higher fatalities. And yet, here's a research letter in JAMA saying that there was a marked increase in Emergency Department visits by children coinciding with the recession in 2009. So did health improve during the recession or not? While you could spin a story about how recessions increase health and increase ED use by, for example, reducing traffic accidents and workplace injuries while also reducing health insurance coverage, which many believe increases ED use (though, see here). Or perhaps this is just an age thing--the Eagan et al result is heavily weighted towards older adults, who are the only demographic with any substantial mortality risk in their data, whereas JAMA looked at children, and if becoming unemployed causes the reduction in mortality, then this explains why the cyclicality would differ between children and adults.
But here's a third article in American Journal of Public Health finding that becoming unemployed significantly increases an individual's mortality risk during recessions but slightly decreases their risk during expansions! So we have a compositional puzzle on our hands:
Employment decreases during recessions, and
Aggregate mortality decreases during recessions, but
