This Freakonomics post by Price Fishback on accurately measuring welfare expenditure is very important. The U.S. looks stingy when you look at gross transfers as a portion of GDP. But how much do recipients of assistance really end up with?
The Nordic countries collect income taxes on the cash payments made to social welfare recipients at rates that are four to five times the rates paid by American recipients. Then when the Nordic recipients go out to make purchases, they pay consumption tax rates on their purchases that are 4 to 5 times the rate paid by the poor in America. Furthermore, the U.S. government offers a series of tax breaks to promote social welfare that are not found in the Nordic countries.
So the gap closes when we look at net transfers. How about if we look not at welfare spending as a percentage of GDP, but at transfers per capita?
America in 2003 ranked roughly in the middle of the Nordic countries. Per capita net public social welfare spending in 2003 (in 1990 dollars) in the U.S. was $5,400, while Sweden’s was $6,300, Norway’s $5,900, Denmark’s $5,472, and Finland’s $4,200.
Practically identical to Danish levels! But the system in the Nordic countries is more comprehensive, streamlined, and state-dominated than the U.S.’s more fragmented, variable, and privately-augmented system of assistance.
If we take into account these differences in style, the appropriate measure is net public and private social welfare expenditures per capita. By this metric, the U.S. then leads the way at $7,800, followed by Sweden at $6,700, Norway at $6,300, Denmark at $5,800, and Finland at $4,900.
But health-care in the U.S. is more expensive. If you take that into account, the U.S. drops to Swedish levels and merely ties for first in generosity.
Nevertheless, the Nordic systems are superior by a maximin criterion. Fishback reports that Americans at the 10th percentile do as well as Nordic folk at the same position on their distributions, but Americans below the 10th percentile do much worse.
(Note for Rawls fans: Maximin in Rawls isn’t intended to apply to the unemployed, who are assumed away at the level of ideal theory, but to the productive working class. It’s not at all clear that maximizing the minimum for the working-age unemployed is at all desirable in Rawlsian terms. Insofar as productive participation in the economy is one of the social bases for self-respect, and is required by an ideal of social reciprocity, transfers that discourages labor-market participation may be verboten. If relatively stingy benefit levels for Americans below the 10th percentile lead to lower rates of low-skilled unemployment than in the Nordic countries, that may be a feature and not a bug, even on Rawlsian grounds.)
What to make of all this? First, the U.S. welfare system is very generous. The hackneyed stereotypes are wrong. Second, it’s important to remember that the U.S. has lots of states, lots of variability in state-level policy, and very unevenly distributed pockets of severe poverty. In some states, the patchy federal system is well-filled in and augmented by state and local government and by assistance from non-state civil society orgs. In some states, not so much. In any case, that Americans at the national 10th percentile do as well as anywhere in the world is a testament to the fact that the level of public assistance at this and lower levels in some parts of the U.S. are probably better than anywhere in the world. I’d bet my drinkin’ money that Iowans at the first percentile of the Iowa income distribution are as well compensated for their disadvantage as are similarly struggling Danes.
But what about inequality? What’s striking to me is that lower levels of inequality aren’t buying the Nordic countries higher levels of welfare benefits. If the relative stinginess of the U.S. safety net at the very bottom is a bad thing, but not somehow causally related to income levels at the top half of the distribution, the necessary reforms likely have nothing much to do with reducing the dispersion in incomes.