In my last post, I said “total net social spending” included net public spending and mandatory private social spending. In fact, it includes voluntary private social expenditures as well. The U.S. has by far the highest voluntary social expenditures in the OECD, so if you subtract those out, the U.S. net public and mandatory private social spending figure is no longer second in the OECD (and thus almost certainly the world, as poorer countries have smaller welfare states), only just about average.
But what does voluntary private social spending include? One big component is employer-provided health insurance. It seems to me that should be included in the size of the U.S. welfare state, even if it is not directly provided by the government, because the government subsidizes it (through the tax code), and because that spending is a substitute for government spending in other countries. If we exclude it for the U.S., we are not comparing like with like, since several other countries provide health insurance mainly or exclusively through the state. Put another way, if the U.S. provides so much social welfare privately, the need for the government to provide it is less. The U.S. welfare state is average-sized in spite of the fact that the private welfare system is enormous.
Now, does that mean the U.S. spends vastly more on the poor than most other OECD countries? Not necessarily. The majority of social spending in the U.S. does not go to the poor – but neither does it anywhere else. The elderly soak up a huge portion of social spending in almost all advanced industrial societies. Indeed, one way to measure how redistributive the U.S. welfare state is is to subtract the “post tax and transfer” Gini ratio from the “pre tax and transfer” Gini ratio. Of course, this is a static measure that does not take into account possibilities for mobility from one income level to another, and the extent to which “poverty traps” can contribute to lost mobility. Still, it’s a suggestive measure.
Using data from
World Development Indicators Standardized World Income Inequality Database, I find that the tax and transfer system in the U.S. shaves only 0.08 points off the Gini ratio, a standard measure of income inequality (“1” means most unequal, “0” perfectly equal). In most other countries, the number is much higher. In Sweden, it is 0.20. In Italy and Germany, is 0.21. Only Switzerland showed (slightly) less progressive redistribution.
So while the U.S. has one of the very largest welfare states in the rich world, it also has one of the least progressive welfare states in the rich world. By the standards of anti-inequality preferences, that’s a terrible record of inefficiency.
Updated with correct source for my data.