A new NBER paper (h/t Thoma) by Richard Burkhauser and Kosali Simon says that it does. From the abstract:
We show that ignoring the value of health insurance coverage will substantially understate the level of economic well being of Americans and its upward trend and overstate the level of inequality and its upward trend.
This argument is one of the key ones we often here to rebut the wage/productivity divergence we’ve witnessed in the last thirty years. It’s definitely landed me in uncomfortable “gotcha” moments at times. However, digging into the paper a little bit, the argument is left a bit wanting.
If we accept the methodology, the results are robust. For instance, using the standard measure of income, the 90/10 ratio increases from 8.3 to 8.5 from 1995 to 2008. Using their measure of income including health insurance, the ratio decreases from 9.3 to 9.1 over the time period. Similarly, instead of having the Gini coefficient increase from 0.422 to 0.433, it increases only slightly, from 0.396 to 0.398. With the new measure, income gains are much broader across deciles; except for the lowest 10%, income gains are uniformly higher as one moves down the income distribution.
I have two issues with this study:
1) They use the insurance value of health care in all cases, including Medicare and Medicaid, when we all know that costs have increased over time while quality has remained stagnant. Thus, as a measure of economic well-being, their “total income” metric will almost certainly overstate the reality. Further, health insurance “income” is a much higher share of “total income” at the bottom of the distribution, so the exercise will naturally lead us to less inequality. My suggestion would be using cross-country metrics of health expenditure to capture the well-being. With this change, the added income would likely be reduced by half and the picture would be much different.
2) The authors will inevitably argue that their study is positive economics, not normative. They are right; in fact, much of their paper could be construed as an argument in favor of health care reform. My concern is that many will take their paper as a sign that we should worry less about wage inequality and its systemic drivers, especially if health care is the main explanatory factor. I think their analysis is useful, because of my reservations above, I worry it might be useful for wrong and normative purposes, and the systemic causes of wage inequality are ignored enough as is.
So, to answer my question from the title, I think health insurance does reduce inequality, but less so than the authors state insofar as we are concerned about economic well-being.