In his New York Times newsletter, Paul Krugman discusses the political debate over who should and who shouldn’t receive government aid.

Excerpt:

Every advanced country, the United States included, has an extensive welfare state — a set of government programs intended to give its citizens more economic security than they would have if we just let markets rip. Our safety net has far more holes in it than that of, say, Denmark. Even so, Medicaid and Medicare cover a third of the population, and most Americans over 65 get the majority of their income from Social Security.

We have, in other words, effectively decided as a society that many people should receive substantial aid from the government. But there’s a real debate, often obscured by political rhetoric, over who should, and maybe even more important, who shouldn’t receive that aid. This is a debate that rages both between the parties and, rather differently, within the Democratic Party.

The between-party debate mainly concerns how much we should do for low-income families. Republicans are obsessed with the idea that we’re providing too much aid to the undeserving poor. There are still 14 states, all Republican-controlled, that have refused to expand Medicaid under the Affordable Care Act, even though the federal government would pick up almost all the tab. And Republicans are constantly coming up with new schemes to impose work requirements for programs like Medicaid and food stamps.

But I don’t want to say much in today’s newsletter about all this except to note that the cruelty of such schemes appears to be a feature, not a bug. There’s no evidence that access to public aid is a significant deterrent to work in 21st-century America, and attempts to limit access don’t even save much money, because they involve a lot of administrative costs. All they do is make it harder for people who really do need help to get it.

The debate among Democrats is, as I said, very different: it’s about whether we should be providing aid to affluent people who may not need it. At one end of this debate is Pete Buttigieg, who has attacked proposals for free tuition at public universities on the grounds that some of the beneficiaries would be millionaires, and called for an income limit. At the other end is Andrew Yang, who wants to give everyone a basic income whether they need it or not.

How should we think about these differences? As I see it, there’s a fundamental trade-off between dignity and cost. And if you ask me, Buttigieg and Yang have each moved too far on one or the other side of that trade-off.

Now, there are two advantages to universal programs that aid everyone, no questions asked. The lesser but still important advantage is that they’re cheap to run. Social Security and Medicare have low administrative costs and are surprisingly unbureaucratic, because there aren’t any questions about eligibility: if you’re legally here and the right age, you get the benefit.

The more important advantage is dignity. You don’t have to go cap in hand to the Social Security Administration with proof that you’re poor enough to be entitled to aid. You just get it.

The downside of universal programs is that they do indeed give aid to some people who don’t need it. Next year Bill Gates will become eligible for Medicare; why should middle-class taxpayers be subsidizing his health care?

Which side of this trade-off should dictate policy? My answer is that it’s mainly about the numbers.

Means-testing the basic hospital benefits under Medicare would be a terrible idea, because only a small number of older Americans are actually rich enough to do without it. So you’d be saving at most a small amount of money, while sharply increasing the costs of running the program and taking away the dignity of everyone else by forcing them to prove they need it.

Buttigieg’s demand that we place an income test on free tuition falls into the same category. It would add bureaucracy and intrude on peoples’ lives in return for fairly small savings.

But Yang-type proposals for universal basic income are different: they would cost a lot of money. As of 2012 only 21.3 percent of Americans participated in programs designed to alleviate poverty — like food stamps and Supplemental Security Income — and even so these programs were clearly inadequate for many families. Expanding income support to everyone would either offer such support at a drastically inadequate level or require a huge increase in revenue; Yang is talking about a 10 percent national sales tax, and it’s not clear even that would do it.

This might be worth doing if we really were, as Yang claims, facing mass technological unemployment. But so far, at least, we aren’t.

So, which programs should be universal, which targeted on those who really need them? The unsatisfying answer is, it depends. It’s a messy world, and one-size-fits-all solutions usually fall short.

For more: Sign up for his weekly New York Times newsletter.