Mortality and Science: A Comment on Two Articles on the Effects of Health Insurance on Mortality

Recently, two studies on the effect of health insurance on mortality were published in the Quarterly Journal of Economics. Both articles provide evidence consistent with the conclusion that obtaining health insurance entirely eliminates mortality. I examine the evidence in each article and conclude that neither provides much useful information. Both investigations are severely under-powered for detecting a reasonably sized effect of health insurance on mortality; both were likely to grossly overestimate the effect, if not get the direction of the effect wrong. In addition, between the two articles there are many inconsistencies and anomalous results that seem inexplicable from a behavioral or clinical perspective. Further, one study was observational and likely biased by unmeasured confounding, and the other has virtually no external validity.