Universalism and means-testing

Meagan Day points out, correctly, that the pandemic experience has buttressed the case for universalism over means-tested programs:

When eligibility for benefits is conditional, all kinds of bad things happen, ranging from the intentional exclusion of whole (usually maligned and disempowered) demographics to huge numbers of otherwise-eligible people tripping over red tape and falling through the cracks. Another major problem with means testing is political: so long as there’s an income threshold, austerity-minded politicians will always try to lower it, leaving more people out as time goes on. In other words, targeted social programs make easy targets.

Part of the architecture of capitalism is that truly wealthy people are a tiny minority of the population, and the focus on whether this minority gets a minor boost out of a universal program that would bring stability, security, and prosperity to the vast majority (and would be paid for by progressive taxes) has always been a distraction.

Max Sawicky (with whom this blog has tangled on universalism before) responds. In my view Sawicky is far from persuasive:

Means-tested benefits are said to limit eligibility and leave many needy behind, unlike ‘universal benefits.’ Unfortunately, there is no really existing universal benefit that remedies this problem, so we are committing the basic fallacy of criticizing an existing program by comparison to an idealized alternative. There is no reason a benefit founded on a means-testing formula could not be provided to anyone you might like.

The real rationale for means-testing is not that it deprives the unworthy of assistance. That is a straw man often deployed by the less progressive among us. It is that for any sub-group of the population, a means-tested program will be cheaper, or for any given amount of money, it can provide more assistance than a universal one. Advocates for universal benefits must compete with equally righteous left advocates for other types of spending. In practice, public funds are limited. (Also true in an MMT world, by the way.)

It is Sawicky, I think, who is guilty of straw-manning here. That the stated purpose of means-testing is to maximize the benefit-per-unit-dollar to the targeted group is not disputed by anyone. No one suggests that means-testing is imposed explicitly to create bureaucratic hoops in order to deny benefits to the eligible. Universalists suggest that in practice, means-testing of individual programs often has that effect, sometimes unintended, some tacitly intended by public officials who, whether out of ideological pique or in response to fiscal pressures, engineer eligibility criteria to limit disbursals. We had a spectacular example of that when Florida’s Republican governor Ron DeSantis had to acknowledge that his predecessor had deliberately made filing for unemployment benefits difficult, in order to forestall pressure on the insurance fund that might have required raising business taxes. (Unemployment insurance is not a means-tested program, but like means-tested programs it is gated by an application bureaucracy which always has the effect, whether intended or not, of limiting benefits to a subset of the eligible.)

It’s a particularly odd time for Sawicky to write “there is no really existing universal benefit that remedies this problem, so we are committing the basic fallacy of criticizing an existing program by comparison to an idealized alternative.” In the CARES Act, we just had an (almost) universal $1200 cash payment, and despite the fact that the United States lacks the infrastructure for benefits deposits that are a matter of course in more civilized countries, those payments were almost universally received. Someday a comparison will be done of the fraction of the eligible who received their $1200 payments, versus the fraction of eligible people and small businesses that received Federally enhanced unemployment insurance and forgivable PPP loans. Does Sawicky believe the comparison will reflect well upon the latter?

A fundamental misunderstanding undergirds this whole debate, that it’s an argument of “means-testing” versus “universalism”. Universalists don’t dispute the necessity of means-testing, they dispute how and where means should be tested. Partisans of “means-testing” are really arguing for within-program means-testing — that each benefits program should have its own means-based eligibility infrastructure. “Universalists” are also very devoted to means-testing. We just think all programs should share a single, efficiently managed and enforced, means-testing infrastructure rather than jerry-rigging one for each individual benefit. Means-testing is inherently intrusive of privacy. It is costly in time, money, and sometimes dignity, for applicants and for validators. It imposes, in the language of economists, large deadweight costs, from the cost of the exercise itself, and from the social cost of benefits deterred by the exercise and so undelivered. It is simply more efficient to have a single and, um, universal means-test, and use that for all programs.

And we do have a universal means-test. It is called the income tax. Universalists argue that rather than means-testing program by program, we should just provide the benefits unconditionally, but “take them back”, in different degrees, financially, via the income tax. In accounting terms, gross outlays (whether they be cash benefits or “in-kind”) are to be divorced from means-testing and, to the maximum extent possible, any other sort of eligibility bureaucracy that imposes deadweight costs. But net outlays — benefits extended minus taxes paid — are absolutely means tested.

This is just more efficient than multiplying eligibility bureaucracies for every program. But its advantages go beyond eliminating redundancies. Providing benefits up-front and then charging on a sliding scale as income is realized supports the insurance function of social spending. People’s incomes are uncertain and only verifiable after some delay. Requiring demonstration of inadequate means up-front, rather than on the back-end, creates at best a delay between when a shock is experienced and when it can be ameliorated. “Delay” can mean your kid skips meals, you start rationing your insulin, or your family is evicted from its home. It’s a big deal.

Further, from the perspective of recipients, benefit withdrawal is itself a kind of income tax, often imposing very high effective marginal rates at low levels of income. We make programs “cheap” by making people just richer than beneficiaries pay their way, rather than let the burden fall higher on the income distribution. Within-program means testing may be “progressive” relative to no program at all, but it’ll be regressive relative to a universal program funded from the general budget. Bouncy tax rates and cliffs are bad policy that affect human incentives in real, socially costly ways. Restricting means-testing to the tax system allows us to structure incentives rationally and burdens ethically.

Less materially, but perhaps more importantly, universal programs unite and ennoble us while within-program means testing divides and diminishes us. Humans perceive flows of funds and services in gross terms rather than in net terms. We all have the use of the roads “equally”. In net terms, you can argue that poorer people get full use of the roads without much funding them in taxes, so they are lucky duckies. Or you can argue that the rich use the roads much more than even the home-owning, property-tax-paying middle class, so the flow of net benefits is severely regressive. Argue this shit all you want, as a wonk or whatever. Most of us just perceive the roads to be fair and free to all, and the taxes as something else entirely. Roads are common infrastructure, an interest we share, a good reason to support solidaristic government. Given the high salience of gross flows rather than net, it’s a free-lunch in terms of social cohesion to provide services and support to all, rather than make very explicit that some are getting “free stuff” that others are not.

But what about the inflows? If universal outflows to the public are supportive of social cohesion, aren’t the radically varying inflows we require as taxes corrosive of the same? Yes, of course they are. Have you seen how rich people talk about the income tax?

However, these aren’t equal and opposite effects. More affluent Americans still enjoy and support the roads and national parks, despite “overpaying” for them. Many of them would be grateful for the option to send their kids to college less stressfully. Further, developed countries have tremendous flexibility in how we finance universal benefits. There is no one-to-one correspondence between a new dollar spent to support a universal benefit and an increase of divisive dispersion in the tax system. Social democracies support universal benefits with a wide mix of income and payroll taxes that render the benefit system more net-means-tested (when they are not capped) and other taxes (e.g. VATs) that render them less so. Choices about the progressivity of income taxes affect the degree to which the benefits state is net-means-tested, and have potentially complicated effects on social solidarity. (Very progressive taxes leave high-income people disgruntled over the short term, but over the long-term, very progressive of taxation can contribute to a “predistribution” more conducive of solidarity.) Developed countries that control their own currencies have wide latitude to finance benefits with “debt” instruments that are more like preferred equity. If they are careful, they can wisely distribute the inflation risk public expenditure may occasion, rather than let expenditure merely magnify it. A wide variety of regulatory policies (e.g. discouragement of predatory consumer credit arrangements) can reduce the degree to which benefits put pressure on prices and so must be offset with potentially divisive taxes.

Finally, means-testing obligations rather than benefits better aligns social provision with human dignity. When a person applies for means-tested benefits, they are asked to prove their poverty, which in our society shades towards inadequacy, in order to get something that they need. They must make themselves charity cases. When we means-test obligations, no one is humiliated. Detached from “getting a handout”, there’s not much indignity filing a tax return, regardless of ones income. When ones tax bill is high, some people feel pride (“taxes are the price of civilization”) and some people feel ripped off. But no one is diminished in the eyes of their community, or in their own eyes. If you think taxes are theft and the state is predatory, that is a reason to be angry at others. But your high tax bill is not a personal failing. On the contrary. Whether you are content to pay your taxes or outraged that you have to, the cause of your predicament is a source of pride rather than of shame.

From the market or from other sources, the benefits we receive from others without having to beg tell us what we are worth to our society. The taxes we have to pay may be irksome, but they are not deducted from that social expression of value, as long as they are levied on nondiscriminatory terms. There is a free lunch in human welfare embedded in these propositions. We are idiots not to take maximal advantage of it.

Note: While I was writing, Matt Bruenig published a response to Sawicky and Day. I suspect I’ll agree with it, though I haven’t read it carefully yet.

Update History:

  • 28-Sep-2020, 3:55 p.m. EDT: “It is costly in time, money, and sometimes dignity, for applicants and for validators.”
  • 22-Oct-2020, 1:35 p.m. EDT: “…the benefits we receive from others without having to beg tell us what we are worth to others in our society.”

7 Responses to “Universalism and means-testing”

  1. Max Sawicky writes:

    Apologies for responding via Twitter pot-shot. Here is a better effort.

    The objective of means-testing — maximizing the efficiency of a transfer — is in fact ignored in the Day article and forthrightly denied by my friend Matt Bruenig in a previous Twitter dust-up. Your more rational take on means-testing is not echoed among the advocates.


    As a factual matter, the reach of the CARES act is very far from universal. Your claim to the contrary is mystifying. It might compare well in this respect to existing programs, though I doubt it. That aside, CARES is a temporary program launched under extraordinary circumstances, so less of a model for what is really being discussed, the good old UBI. You can be sure under a Biden presidency and a continuing economic collapse, I will be out there promoting a broad and ample NIT whose launch would be well-timed.

    You are really arguing for a more centralized means-testing system, so like Matt you counterpose an idealized alternative to the grubby realities. The simplicity of your ideal is exaggerated in at least two respects.

    1. How would the benefit integrate with the existing, very messy and complicated mix of benefit programs? I encountered a similr problem in my own proposal years ago for a super-EITC. The new credit was relatively simple. Explaining what it would replace was not.

    2. A universal program means the Gov has to chase more money going to and from more people. On the benefit side, identities and addresses and bank accounts would have to be verified, to combat fraud and double-dipping. Higher benefits elevate such incentives. Some appear to think there is some Big List of Everybody in the sky. There isn’t. There could be, but then you have the ham-and-eggs problem. “If we had ham, we could have ham and eggs, if we had eggs.”

    You seem to be more exercised about marginal tax rates than I am, which is funny since I’m a tax economist. After carrying out the task of calculating marginal tax rates under the income tax plus a slew of benefit programs, I am confident ordinary people don’t do that. They compare alternatives founded on fairly big changes in plans (e.g., work full-time or half-time).

    I like the social cohesion argument, made eloquently by Mickey Kaus before he lost his mind, but it applies more to public goods than transfer payments. I wish the U.S. was more congenial to solidarity appeals, but race and now immigration seem to make that objective ever more distant.

    My fear is the UBI chatter distracts from the destruction of TANF-AFDC and the urgent need for some replacement. In that case there is at least some foundation in institutions and program history.

    The burden on UBI advocates, thus far unmet, is to specify how much money, who would get it, how it would integrate with other programs, and what budget adjustments — if any — would be included.


  2. Darius J Wilkins writes:

    1) This country has never had a perfectly universal program, and it has never had one because there are strong constituencies against any such thing. You might talk about promoting unity, but the fact is, plenty don’t want any such positive sense of shared burden.

    2) The tax code of Americans is anything but progressive. We should not lean on the slim progressivitiy of the income tax in exchange for any universal benefit, and frankly, making the overall tax burden of Americans more progressive will have more positive benefits than any UBI or other progressive distribution.

    3) One of the crucial functions of means testing is to prevent better off groups from hogging the benefits of “universal” programs–there are many ways to gatekeep, and means testing is one way to signal what a program is *for*, and the American compact must involve a sense that things are *for* those people in need, and it’s something every American is proud of being able to do.

    In addendum, I view these discussions about “universalism” is really a means to exclude people from active participation in shaping policy with a bit of vague and patronizing “you’ll get yours”. Not to mention all of this is just stupid Sunsteinian social mentality that ultimately doesn’t make good policy in the real world–the real world where the US has plenty of “universal” programs that aren’t so, and for very real reasons.

  3. Detroit Dan writes:

    Public schools and public roads are examples of successful universal benefits.

    Our health care financing system is an example of a nightmarish means tested system, and obviously inferior to the universal systems in other countries.

    The means-tester contingent reminds me of the monetarists. While MMT is a straightforward description of how fiat currency works, the monetarists obscure the simple facts in the service of overly complex market-based solutions. This is the essence of the neoliberal model which is clearly failing around the world, in my view.

  4. hunkerdown writes:

    Or, simply, Max is a shill for the class interest of dead-weight benefit administrators.

  5. BlakeFelix writes:

    Well said. I think that even income is too hard to track, and shifting to pigovian taxes would yield huge benefits. Taxing carbon emissions, for example, would be more progressive than current policy, if both the income and payroll taxes were replaced. And tracking pollution is easier and less invasive than tracking income.

  6. Gram Brinson writes:

    I’m not sure I understand the distinction you draw between instituting means-testing to deny benefits to the eligible and tacitly intending that gatekeeping effect to lower costs. Isn’t the latter just intending to deny benefits for a “fringe benefit” (from miserly policymakers’ perspective) rather than for its own sake?

  7. Bolt writes:

    re: “How would the benefit integrate with the existing, very messy and complicated mix of benefit programs”, isn’t one big benefit of a UBI to get rid of these programs? We wouldn’t need Social Security, unemployment, disability, TANF, SNAP… even child support payment infrastructure could go away if we make the UBI apply from birth.