Let’s start with what I would expect is an uncontroversial sounding principle (call it the progressive principle): if the government is going to run a social assistance program, the average income of the people paying for the program should be higher than the average income of the people benefiting from the program.
The justification for this principle is, as I say, common sensical. Most people support the idea of redistribution, but it’s usually redistribution from the rich to the poor that they have in mind. Redistribution from the poor to the rich (or from the poor to the poor) would, I think, strike most people as perverse.
Yet the odd thing is that quite a large portion of the modern “social assistance state” violates this very principle. Both Medicare and Social Security, for example, involve redistribution not to the poor, but to the old, who are statistically speaking much wealthier than the average American (here is one analysis of the question regarding Social Security; I’m confident that if you ran the numbers for Medicare you’d get the same result). Not only that, but the funding for Social Security and Medicare comes mainly from taxes that are highly regressive.
Similarly, funding for education tends to be only slightly progressive at best, and in many cases (such as funding for higher education) is downright regressive. Much of the funding for public schools comes from property taxes, which means that the people paying for the schooling (either directly or indirectly in the form of increased rents) are the same people getting the benefits from the schools. An increasing proportion of education spending, however, comes from state taxes, which as Ezra Klein notes, tend to be regressive overall. Indeed, a recent trend has been to use revenues from state lotteries (a particularly egregious form of regressive taxation) as a source of education funding. Thus you have a strange situation of states bragging about taking money from the poor to improve the educational experiences of the well to do.
I could go on, but you get the idea.
If we want to reform the social assistance state to bring it in line with the progressive principle, we have two options: we can change the structure of benefits or we can change the structure of how those benefits are funded. For things like Social Security and Medicare, changing the benefits would mean instituting some sort of means testing, to ensure that people aren’t receiving government assistance if they don’t actually need it. The typical argument against means testing is that it will erode public support for the programs. If the wealthy don’t themselves benefit from a program, it is argued, they will agitate for it to be abolished. The problem with this argument is that it has almost no empirical support. We don’t send food stamps to Bill Gates, yet somehow the food stamps program has not been eliminated. Social security remains popular in countries where benefits are means tested (like Australia), and do not seem to be in any danger of being repealed on that account. And if the argument were really true, then it would be hard to see how any attempt to make the funding of these programs more progressive wouldn’t be doomed to failure.
For reasons that I don’t quite fully understand, many people would prefer to keep paying benefits to the rich, while raising their taxes even more to do so. I’m not opposed on principle to the idea of making taxes more progressive, but realistically there are limits to how far one can go with this. States didn’t resort to lotteries because they hate the poor or are just dumb. They did it because there is a wide spread anti-tax sentiment in this country, and that isn’t likely to change soon. Even discounting issues of political viability, it really is the case that higher taxes discourage productivity and growth, and these effects are only going to increase as a global economy makes it easy to move capital from one country to another (one reason, I suspect, that state taxes are so much less progressive than federal taxes is the ease of moving from one state to another). I would therefore submit that if a progressive social assistance state is your goal, it’s going to have to come largely by means testing benefits rather than by raising taxes on the rich.