Elizabeth Warren recently announced a college affordability plan that would, among other things, forgive college debt up to a cap of $50k per person. The amount of debt forgiven starts to phase out for people making over $100k a year, at a rate of $1 per $3 in income, so the plan is designed with progressivity in mind. Nonetheless, the benefits of the plan are not well-targeted towards those in need, with the majority going to households with incomes higher than the median.
This is more progressive than it might sound — because income is distributed very unevenly now, the lower quintiles will get more money as a percentage of their income than the higher ones, even though the majority goes to the top half. Depending on your goals for improving equality, the debt relief plan may still help more than hurt those goals. (It would definitely help after considering that Warren proposes paying for it with a tax on millionaires, but funding for any government spending is ultimately fungible so there’s no particular reason not to separate the debt relief and tax as two policies and evaluate each on their merits).
However, progressives ought to be targeting the upper-middle class as sources of government revenue, rather than recipients of government assistance. Soak the rich, yes, but also the upper-middle class. It’s tempting to think that any and all progressive policy goals can be accomplished by only taxing the 1%, or only millionaires and billionaires, but that doesn’t work if you’re trying to build a Scandinavia-style social-democratic welfare state, which require high, broad-based taxes. So the debt relief going to the upper-middle class in this plan represents a lost opportunity to do redistribution away from this class. For this reason, I think that a debt forgiveness plan should start phasing out at a much lower income threshold, perhaps $30k-50k.
As far as I can see, the defenses offered to a policy design of debt relief that helps the affluent middle class fall into two categories— deferring to the general idea that universal (non-means-tested) programs are better than means-tested programs, and arguing that debt relief has some important economic spillovers that would justify helping the well-off due to a trickle-down effect. I’ll focus on the first one here.
Universal programs can be more efficient or effective, but this doesn’t apply to debt relief
One reason to prefer universal programs is if you think the government simply can do something better than the private sector. For example, this is what advocates of single-payer healthcare argue: a single government insurer can use its market power to negotiate healthcare prices down and reduce the administrative complexity of the system, so it’s better to move everyone, poor and rich, onto the same public insurer. But debt forgiveness isn’t a service that the government takes over from the private sector to administer more efficiently on behalf of the upper middle class, so that consideration doesn’t apply. Not means-testing at all would make debt forgiveness somewhat simpler to administer, because there would be no need to verify income (though income tests for a one-time debt forgiveness doesn’t seem like a very onerous administrative requirement). But Warren’s plan already means-tests, so there’s no additional cost to means-testing it more aggressively.
Additionally, class stratification of a given service can be undesirable in some cases. This is true of public education — if we replaced public schools with means-tested vouchers, this would cause even more income segregation in schools (though there certainly is a lot of that already) because the affluent would no longer be subsidized to put their kids in public school. The reason this matters is that education is ongoing and interactive. Because education is ongoing, income segregation can lead to an increasing divergence in school quality over time, similar to how racial segregation leads to more funding and better quality for white schools versus non-white schools. Because education is interactive, putting rich kids and poor kids in the same schools can help combat social stratification. Neither of these considerations apply to debt forgiveness.
“Universal” is subject to framing effects
I’ve used “not means-tested” as a working definition of “universal” for the purposes of this post. But the appeal of a “universal” program is that it identifies something that everyone needs, such as education or healthcare, and simply provides it to everyone, regardless of income. The thought process for designing such a program should be to identify a minimum standard for some good or service that ought to be available to everyone, and then guarantee that standard via a universal program.
But the idea that the current level of student debt that a person owes at one point in time is the key baseline with which to design a policy is really silly and arbitrary. Debt balances vary a lot for reasons that don’t track with desert or need, setting aside the obvious one that not everyone goes to college. Different people choose different mixes of debt vs. out-of-pocket payments to pay for college, colleges don’t all cost the same, people pay off their debts at different rates, and college attendees didn’t all attend college at the same time, so people who attended earlier benefit less from a one-time debt forgiveness. So the idea that forgiving current student debt balances is the universal default and other policy designs are means-tested is illusory.
As an illustrative example, suppose someone proposed a policy that gives everyone an amount of money equal to the last 4 digits of their SSN. The obvious flaw with this policy is that it’s unfair and arbitrary, and at the very least would be somewhat less unfair and arbitrary if it were phased out for people with higher incomes and higher last-4 digits. The advocate of this policy could deride that as means-testing, but this rhetorical trick is only available because they’ve chosen an arbitrary baseline for their “universal” program.