The Washington PostDemocracy Dies in Darkness

Lawrence Summers: Yes, the Senate GOP tax plan would cause ‘thousands’ to die

Senate Majority Leader Mitch McConnell (R-Ky.) speaks about tax restructuring Thursday on Capitol Hill. (Saul Loeb/AFP/Getty Images)

I suggested on Friday when it became clear that the tax bill would pass that “thousands would die.” In light of my sharp criticism of other economists' claims regarding the legislation, some have asked whether my statement is well grounded. I think it is, but this should be open to debate.

In reaching my judgment, I relied primarily on work by Kate Baicker, a former colleague now serving as dean of the University of Chicago’s Harris School of Public Policy. Kate served on the Council of Economic Advisers during a Republican administration, so I felt that any political bias would operate against the conclusions I drew.

Kate performed two studies on the impact of being insured on mortality by looking at the effect of moving from uninsured to insured. One peer-reviewed paper, co-written with Benjamin Sommers at the Harvard T.H. Chan School of Public Health and Sharon Long at the Urban Institute based on the experience in Massachusetts, estimated a reduction of one annual death per 830 people insured. Another, co-written with Sommers and Arnold Epstein, also at the Chan School of Public Health and published in the New England Journal of Medicine, looked at the experience of Arizona, Maine and New York and estimated a reduction of one annual death per 176 people.

The Congressional Budget Office estimated that the tax bill could reduce insurance coverage by 13 million people, which to be conservative we can round down to 10 million people. Recognizing all the uncertainties — for example, the fact that the group becoming uninsured as a result of the individual mandate repeal probably is healthier than the group Sommers et al. (2014) studied in Massachusetts — if we treat the 176 to 830 range as implying that it is safe to assume that 1,000 more uninsured means one death, the conclusion would follow that the tax bill would result in 10,000 extra deaths per year.

Of course, there are many issues in the extrapolation. Do changes in private insurance from repealing the mandate have more or less potent effects than changes in Medicaid coverage? How do the effects play out over time? Are there compelling contrary studies? My sense, relying on surveys of the literature, is that if anything, experts think that Medicaid coverage changes have smaller effects on mortality, that effects grow over time, and that Baicker and her colleagues are certainly not very high in their estimates of mortality impacts.

Some may argue that any loss in health care resulting from a repeal of the mandate is voluntary, and that families should have the freedom to make choices about optimal health coverage. I think this ignores two realities. First, for many, the loss of health insurance will not be voluntary: They will lose coverage because premiums will increase, pricing them out of the market. Second, I take seriously the insights of behavioral economics, which suggest that irrational actors may make choices that will lead to worse health outcomes and higher mortality rates.

So, my current judgment is that if anything, my claim that over an unspecified horizon “thousands would die” takes too serene a view of the health consequences of the tax bill.