My prior post touched on a variety of background philosophical issues raised by Miranda Perry Fleischer’s presentation yesterday at the colloquium of her paper (co-authored with Daniel Hemel), Atlas Nods: The Libertarian Case for a Basic Income. Herewith are some brief reflections on what I consider issue 2: how a demogrant or “universal basic income” (UBI) might be designed. One important general point is that, while people often think about UBI in very basic terms – i.e., as just a uniform cash grant – its optimal design raises many of the same issues as designing, say, income-conditioned transfers or income tax rate brackets. For example:
1) Cash vs. non-cash – As noted in the prior post, David Bradford and I wrote about this set of issues some 18 years ago. Despite the consumer sovereignty arguments for giving poor people cash rather than in-kind benefits, the grounds for in-kind are not limited to paternalism. For example, even apart from altruistic externalities (as in the case where the altruist would rather give people food or shelter than items that they reasonably preferred), the need for in-kind benefits may respond to other market failures, as arguably in healthcare and health insurance, or may reflect distributive desert (e.g., as evidence of a poor health endowment.
Issues of cash versus non-cash are not limited to what we may think of as in-kind transfers, but also extend to the provision of goods and services that do not have a purely public goods (i.e., non-rival and non-excludable) character. One could imagine a classical liberal proposing, first, that public schools be replaced by vouchers for private schools, and then, second, that the vouchers be replaced by straight-up cash. But this would obviously be subject to objection.
housing, healthcare, public schools – once allow altruistic externalities, a lot of stuff! Especially
2) Which programs should one trade in for the UBI? – On both the right and the left, views about the UBI often are influenced (either expressly or implicitly) by the view that its adoption might either increase or reduce the overall amount of redistribution through fiscal and other policy. While such effects might certainly be relevant to the evaluation, it is also desirable to think about it, purely as a design matter, from a broadly distribution-neutral standpoint. Hence, as in the paper by Perry Fleischer and Hemel, it is worth asking which government programs might be traded in for the UBI if it were adopted.
This is made more complicated, however, by the fact that various programs combine a vertically redistributive element with addressing issues that are distinct from just poverty. For example, unemployment insurance is not just about being poor because one lost one’s job, but about negative income shocks that may adversely affect people. Social Security has significant distributional effects but is also a mechanism for putting a floor on one’s retirement saving (taken as given one’s lifetime income, net of taxes and transfers).
3) Who should get the UBI? – Here the issues include those around legal and undocumented immigration, residence, etcetera.
4) Age of the recipient – Should babies and young children, via the custodial parents or other caretakers, immediately get the same full amount as adults? Does it matter that their current consumption needs might be lower? Would this affect decisions about having children or about custody, and if so what do we think of these effects? Should seniors get larger annual grants than working-age adults, because they are more likely to be unable to work? (This question would intersect with those of Social Security and Medicare design.)
5) Other tax/transfer design issues – Should household structure matter? (E.g., couples versus singles.) One can’t just assume not, given the complexity of the issues here. Should there be regional cost-of-living adjustments?
6) Conditional vs. unconditional – The UBI is neither income-conditioned nor conditioned on willingness to work. The former is just optics, but the latter is important (albeit, not limited to UBI; one can have the same issue with expressly income-conditioned welfare benefits).
The reason that it’s just optics not to income-condition the UBI is that limiting it based on income is just another way of applying an effective marginal tax rate. For example, suppose that Assyria has a $10,000 demogrant, along with a 50% tax rate on one’s first $20,000 of income (not counting the demogrant), whereas Babylonia has a 0% tax rate on the first $20,000 of income, but also provides $10,000 in “welfare benefits” to people with zero income, ratably reduced to zero in welfare benefits as income increases from 0 to $20,000. No matter what your income level, you’ll end up with the same net benefit in either society. So it is only true optically, not substantively, that Assyria is giving the cash transfer even to its rich people (who presumably don’t need it), whereas Babylonia isn’t.
Work conditioning does matter substantively, however, whether or not the grant for poor people is expressly income-conditioned. And here, getting back to issues from my prior post regarding whether libertarians or classical liberals should like the UBI, things get trickier. If you have the Eric Mack view that rescue should only extend to those who are faultless, a willingness-to-work requirement may make sense. Indeed, even within a utilitarian framework there are arguments for it, although here they would be purely consequentialist. (These might pertain, for example, to positive externalities or internalities associated with requiring people to work if they can.)
Fleischer-Perry and Hemel challenge the case for a work requirement, under a libertarian or classical liberal framework, by arguing that in practice it is likely to be unacceptably intrusive and error-prone. But arguably that puts it a bit strongly, if the filter is not completely ineffectual and there are other reasons for favoring it.
7) A single up-front grant versus periodic (such as annual or monthly) grants – Fleischer Perry and Hemel rightly note that a “luck egalitarian” might dislike the fact that, with periodic UBI payments, the longer-lived end up getting more money than the shorter-lived. Life annuities (as under Social Security) would appear to be anti-insurance if one were thinking purely in terms of overall lifetime welfare. That is, if you’re already luckier in that you get to live longer, you’re made luckier still by getting more money too. What makes a life annuity true insurance, rather than anti-insurance, is that living longer creates the risk that one will need more resources for one’s support. So it increases expected utility despite its rewarding those who are (in an overall rather than a marginal utility sense) already the “winners.”
With incomplete capital markets, the choice between an upfront grant and periodic grants matters for reasons apart from variations in life expectancy and actual lifespan. Ackerman and Alstott note that, if it’s hard to borrow against the value of expected future grants, cash upfront may be more empowering in some circumstances. (Of course, there are other mechanisms for addressing this, e.g., education loans and not currently taxing expected future earnings.) But on the other hand the empowerment might also lead to costly errors in judgment while one is still young. Arguably libertarians, because of the not- just-instrumental value they place on choice (including any notion that it’s just your tough luck if you suffer from choosing poorly), should be more sympathetic than others to allowing people, say, to pledge future UBI in exchange for cash today.
8) Is the UBI “too popular,” from a libertarian or classical liberal standpoint? – Fleischer-Perry and Hemel question the premise by some that libertarians and classical liberals who dislike redistribution, but figure that there is bound to be some of it, should like the UBI as a kind of political second-best, limiting the amount of redistribution and the related intrusion into people’s lives. They base this part on the idea that “universal” programs can become very popular. A case in point is the decision by Social Security’s founders to call it a “universal” program (and to make its transfers between participants relatively opaque). But there is little evidence to date of the risk that it would become, at least from a particular standpoint, too popular. (BTW, I note also that some on the left are skeptical of UBI because they believe it would cause the government’s redistributive mechanisms to end up being smaller, rather than larger.) The two main things that seem to limit UBI’s political appeal are (a) fiscal illusion from its not being expressly income-conditioned (as in, “Why give it to Bill Gates?”, and more substantively (b) notions of more limited distributive desert that are related to conditionality and willingness to work.