Means Testing

First warnings — as usual I am writing on a topic discussed by many experts and I am not an expert. It is very often debated whether social welfare programs should be means tested (available only to people with low income or to people with low income and low wealth). An alternative is universal programs which are provided also to high income people (Medicare, Social Security old age and survivor pensions, K-12 public school, police protection, fire department putting out fires — most programs)

In particular, the topic of the week is Biden’s student debt forgiveness plan, which forgives $10,000 of student debt only if income is less than $125,000 per year. I do not want to focus on this particular program (because it has been discussed a lot and many many people know more about it than I do).

I want to start with a theoretical discussion of the general issue in public economics (a large literature with many researchers vastly more expert than me). I start here, mostly because the viewpoint is strange and not familiar to most normal people (the discussion will be of little practical use, because the viewpoint is strange, not familiar to most normal people, and not connected to actual policy making in the actual real world).

This means I will pretend I am advising a social planner, philosopher king, benevolent dictator. That is I will discuss policy ignoring politics (at first). In this fantasy world, there is a case against means testing. The argument is simple — if you are worried about income distribution, handle it with an income tax, if you are worried about wealth distribution handle it with a wealth tax. If you have optimal income and wealth taxation, you shouldn’t care about the marginal effect of programs on the distribution of income and wealth.

This is standard economics 101. It very much appeals to the policy with no political limits (if you have optimal — then you are not on this planet). The logic is largely based on models in which market outcomes are Pareto efficient, so the only thing wrong with them is that the distribution of income and wealth is not optimal. There are two real world based arguments for universal benefits. First assessing income for the income tax and then again for program eligibility implies duplicating income assessment and un-necessary administrative costs. Second, and more important, the costs are partly born by people eligible for the program and many eligible people do not, in fact, participate. This means that programs for needy people do not reach those who are unable to handle applications or are too busy surviving to spend the time. So any complication prevents a program from helping some of the people who need it most.

There are two very bad arguments against universal benefits and progressive taxation. They are very bad arguments against progressive taxation. The first is that marginal tax rates other than zero distort incentives. This is a reason that optimal taxation and redistribution definitely does not imply perfect post tax equality — optimally managing the tradeoff between equality and undistorted incentives is the original meaning of “second best.” Straw man might argue that means testing benefits is a way to achieve more equality given the limit on taxation due to its effects on incentives. Straw man is made of straw (I won’t google but I might not be able to find anyone making this silly argument). The reason is that means tests also affect incentives. In theory it doesn’t matter whether we take from those with high income with a high income tax or by cutting off benefits. Also in practice. Straw man then argues that the problem isn’t simple incentives but tax evasion and tax avoidance. Again the argument is that increasing taxes and reducing benefits are the same (have the same effect on household budgets) so have the same effect on cheating or legally avoiding.

I think a benevolent planner would use highly progressive taxes to fund universal benefits.

Now politics. One very strong argument for universal benefits is “programs for poor people are poor programs”. Means tested programs are generally stingy. There is one huge exception — Medicaid. I think a key reason is that many Medicaid recipients are needy because they are sick and had previously been middle or working class. A related issue is that participation in means tested programs is stigmatized. The strange ratio of hatred of welfare to dollars spent on AFDC then and TANF now are key illustrations.

On the other hand, there is anger at benefits going to rich people totally out of proportion with the cost in dollars. This is one example of how stories beat numbers and accounting is alien to most people (I am trying to avoid typing the word “innumerate”).

On the other hand, there is also a strange focus on the Federal Budget and Federal Government employment. This means that there is a political difference between a cash benefit and a tax credit. To a certain extent, I think that Straw man (above) is actually influencing people. It is certainly true that Democrats put huge efforts into describing new welfare programs as tax cuts.

I think that, so far, consideration of politics vastly increases the strength of the case for universal benefits and against means testing. Here “politics” is public opinion and elite rhetorical strategies.

But then there is congressional politics. It is silly to consider optimal income redistribution and then optimal other programs. In practice the one thing we all agree on is that the progressivity of the income tax is all wrong (the problem is most of us think that it is not progressive enough but the minority who think it is too progressive are relatively rich and powerful). In the real world, the effect of policy on income distribution always matters a lot, because the progressivity of taxation is hard to change, because the law is hard to change.

I can predict my position on policy issues basically almost always by considering their effect on the income distribution. In theory that should be one issue separate from all the others, In practice it is one issue connected to all the others.

OK so just about student debt relief. There is no way a philosopher king would even consider Biden’s program. Nor would Biden if it weren’t for the detail that it is one of the few things he can do without Congress.