Yesterday at the colloquium, Miranda Perry Fleischer presented the above-titled paper (coauthored by Daniel Hemel). They posit that libertarianism (which they note has multiple strands) and classical liberalism might be consistent with favoring limited redistribution in the form of a safety net protecting the poor, and that this might plausibly take the form of supporting a universal basic income (UBI) or demogrant.
Three great things about the paper and the project are (1) it reminds us that there are more versions of libertarianism (and affiliated sentiments) out there than just Nozick, (2) it addresses real world policy issues based on a normative framework different from the usual one (or mine), thereby expanding the breadth of discourse, and (3) UBI / demogrants is a great topic, even if one doesn’t consider it currently politically practical. Like the long-running income tax vs. consumption tax debate in U.S. tax policy circles, it doesn’t just illuminate a particular instrument choice but is a well-suited vehicle for exploring fundamental underlying issues.
BTW, I’ve written in the past about varieties of libertarianism in relation to the very interesting case of Henry Simons, and about demogrants vs.expressly income-conditioned “welfare,” and (with David Bradford) about cashversus non-cash benefits. So this is territory that I like and have thought about before (albeit not extremely recently).
My thoughts about the paper and the topic are best divided into two headings: using libertarianism to assess the UBI, and UBI design issues. I’ll discuss Part 1 in this blog post, and Part 2 in a follow-up.
1. Using libertarianism to assess the UBI
Given the if-then structure of the paper’s analysis – “If you subscribe to some version of libertarianism or classical liberalism, then you might favor a UBI” – debating the merits of those views is not the most pertinent way to respond. But because one needs to understand the rationales in order to apply different versions of these doctrines, I find it hard to stay away entirely from raising issues that may reflect my skepticism.
1) Hypothetical consent – In general, this family of normative views requires consent in order for people to be subject to limitations on their property rights (if those meet the Lockean, etc. tests for validity). But it can be hypothetical consent that a reasonable person would have to grant.
Sometimes these exercises might involve comparing the actual state of affairs to a version of the state of nature, or to a world in which the state doesn’t exist (but that’s otherwise the same?) or in which property rights don’t exist, or perhaps something else. I admittedly find myself unclear for the rationale for picking a particular counterfactual, which might be a quite fanciful and artificial construct. I’m also inclined to base my own preferred hypothetical consent exercise on Harsanyi’s behind-the-veil analysis, where all you don’t know is which person you are – a device for valuing people’s welfare equally – and which can push one towards utilitarianism based on an expected-utility analysis.
2) How decide the relevance of empirical inputs? – The paper extensively discusses various empirical issues, as it should given their relevance. But given the deontological foundations of at least some versions of libertarianism, it’s hard to reach firm conclusions about how they would affect the analysis. For example, the paper notes evidence in support of the view that, if poor people were given cash in lieu of in-kind benefits (e.g., Food Stamps or housing subsidies), they might tend to use the money wisely rather than foolishly. This might be normatively irrelevant, however, under a view in which people should be given maximum freedom of choice but if they screw it up that is purely on them.
3) Should people be rescued from the consequences of their own mistakes? – Suppose X is poor because X made bad choices. Should government policy help X out? Under utilitarianism, the principle of beneficence – from counting everyone’s welfare positively – says yes. There may be a tradeoff in practice, because rescuing people from the consequences of costly mistakes reduces their incentive to avoid incurring the costs of those mistakes, but the utility gain to X is counted no differently than if there had been no mistake made.
Libertarians often care about (and moralize) choice for its own sake, rather than just instrumentally. A consequence is that it might matter why someone is poor – due to bad choices, or circumstances beyond her control – even without regard to the empirical significance of incentive effects. In this regard, a well-known hypothetical by the philosopher Eric Mack plays an important role in the paper’s analysis. http://murphy.tulane.edu/people/eric-mack
As paraphrased by Perry Fleischer and Hemel, Mack ““ask[s] us to imagine a fully-prepared hiker on a well-planned excursion. Through no fault of her own, she encounters unpredicted fatally cold temperatures. The hiker comes across a locked but unoccupied cabin in the woods whose shelter, fire, and blankets would save her life. Entering the cabin to save her life, however, would violate the owner’s property rights. Yet according to Mack, the ‘most basic intuition’ is that ‘no plausible moral theory’ would require the faultless hiker to freeze to death. ‘Even more clearly,’ Mack writes, ‘no moral theory that builds upon the separate value of each person’s life and well-being can hold that Freezing Hiker is morally bound to grin and bear it.’ For these reasons, Mack believes that libertarianism must tolerate some violation of private property rights—at least in the extreme circumstances of the freezing hiker.” However, “[o]nly instances of extreme need … justify violating the owner’s property rights in Mack’s view. If our hiker were simply tired and sore (and not in fatal peril), no violation would be justified.”
One question I have here is: What does it mean to say that the hiker was fully-prepared and excursion unplanned, yet that, through no fault of her own, she encountered unpredicted fatally cold temperatures? Evidently, there was not, ex ante, a zero percent chance that this would happen. She deliberately took the risk that it would happen by going on the hike, instead of staying home, and not equipping herself with sufficient protective garb for this eventuality. So we presumably are in the realm of evaluating the reasonableness of her precautions, as a function of the likelihood that there would be so severe a cold snap and the costliness of being ready for it.
I myself would want the hiker to be aided even if she planned poorly, and even if she merely faced extreme discomfort, rather than death. Whether I’d want to impose the cost of helping her on the cabin owner, which would affect that individual’s incentive to maintain an intact and well-stocked cabin in the woods despite being less than eager to have invaders take advantage of it, would depend on evaluating the costs and benefits, rather than being decided on the basis of abstract reasoning about property rights. But one broader issue I have with the style of reasoning reflects my sense that it contains discontinuities that I find intuitively unappealing. E.g., certain death between lesser degrees of harm; planning that meets the cost-benefit test (or whatever we use to determine whether the hiker was at fault) versus that which falls just short.
Mack offers a rationale, within libertarianism, for favoring some degree of safety net protection for the poor. But the relevance he ascribes to lack of fault, and the extremity of the harm that the hypothetical posits, might tend to weigh against using his argument to support a UBI. Perry Fleischer and Hemel argue, however, that any such negative implication for the UBI can be overcome, e.g., by reason of mistakes that the government might make in seeking to assess fault.
4) Sufficientarianism and discontinuity – Other discussion in the paper notes that some self-described libertarians or classical liberals support a “sufficientarian” safety net, so that everyone is guaranteed a minimum level of resources even though inequality is not otherwise a relevant social policy concern. The journey from this view to supporting a UBI is certainly a lot more straightforward than that from wanting to help only people who are faultless. Once again I find its discontinuity intuitively uncongenial. Put in my sort of terms, it posits a marginal social value to increasing someone’s welfare that is high until the recipient of the benefit reaches “sufficiency,” at which point the social value of further increases in welfare drops to zero. And “sufficiency” presumably isn’t just a matter of staying alive, but of being able to function at some level of adequacy; it also presumably would be higher in the U.S. in 2017 than 1717, or for that matter anywhere in 2017 BC.
5) What sorts of market failures are relevant to government policy? – While libertarians and classical liberals like market outcomes at least in part for underlying philosophical reasons, pertaining to their particular views about property rights, voluntariness and consent, etc., a utilitarian also will like markets if sympathetic to mainstream economic analysis – subject, however, to believing that the preconditions for markets to work well are sufficiently met. (NoahSmith and others, however, use the term “101ism” or “Econ 101ism” to describe what they regard as undue optimism regarding the frequency with which these preconditions are met in practice.)
Many self-styled libertarians or classical liberals accept, however, not just that market failure can occur, but also that its occurrence may justify government intervention. An example would be using compulsory taxation to fund important public goods that markets would fail to provide due to the free rider problem (arising from the public goods’ nonexcludability). Pigovian taxation to address negative externalities such as pollution (if transaction costs impede just relying on the assignment of property rights) is a second example.
Once one starts down this road, however, one can go pretty far. For example, full-out redistributive taxes and transfers can be rationalized as providing insurance against ability risk (and undiversified human capital risk given the need to specialize) that private markets would provide if not for the adverse selection problem, which governments, unlike private insurance companies, can address by mandating participation. (There is still a moral hazard problem, which governments are stuck with, but they presumably don’t reduce the optimal insurance coverage to zero.)
I’m not clear on how (or how persuasively) one can stave off this line of argument once one has accepted that market failure can make a case for government intervention. Merely being skeptical about the social benefits to be derived in practice from following such a policy can convert the dispute from a purely philosophical one into one with significant empirical content (and potentially within a utilitarian or other welfarist framework).
This can complicate evaluating how a generally pro-market classic liberal who is open to consequentialist arguments based on empirical evidence ought to respond to the case for a more broadly redistributive fiscal system. But a further question about market failure’s significance to the analysis is raised by the paper’s quoting Milton Friedman (who advocated “negative income taxation” that is effectively the same as having a UBI) as follows: “I am distressed by the sight of poverty; I am benefited by its alleviation; but I am benefited equally whether I or someone else pays for its alleviation . . . . To put it differently, we might all of us be willing to contribute to the relief of poverty, provided everyone else did. We might not be willing to contribute to the same amount without such assurance.” Hence, if voluntariness falls short we may have the standard public goods rationale for supplying poverty relief by government mandate.
A utilitarian would certainly be inclined to agree that the altruistic externality cited by Friedman – in that X’s poverty hurts the observer, not just X – is at least presumptively relevant to policy. But it does open a can of worms, for both Friedman and utilitarians. E.g., what if only one minds the poverty of people in one’s own ethnic group? Or observer preferences that are either anti-redistributive or affirmatively malevolent towards others? What if the way in which one wants to help another person reflects either misunderstanding of the inputs to her subjective welfare, or the aim of wanting to impose one’s own moral or aesthetic preferences on her? So one can’t just accept the Friedman point and move on, without getting to a deeper set of issues that are not fully captured merely by deciding to label oneself as a classical liberal, or for that matter as a utilitarian.