Bryan Caplan  

Rejoinder on UBI

PRINT
Bribe Bully Beg Borrow Steal... The shocking truth about borde...
Here's my reply to Ed Dolan, point-by-point.  He's in blockquotes, I'm not.


First of all, thank you, Bryan, for the civil, cogent, and detailed response.

Likewise.

1. You say that I acknowledge elsewhere that the incentives are theoretically ambiguous,income effect vs.substitution effect and all that. Fine, but you give the wrong link. The place where I discuss that issue in detail is in the two-part series that starts here. Part 1 of that post deals with theory, and shows that although there is some ambiguity, it requires very special and implausible assumptions for the income effect to outweigh the substition effect. Part 2 looks at the empirical literature, and concludes that the overwhelming weight of evidence suggests that a UBI improves work incentives relative to any means tested program.
My apologies for neglecting your Part 2.  Well-done; I encourage everyone interested to read it.  But I'm puzzled that you describe the evidence you summarize as "overwhelming."  It seems fairly weak overall to me.  And my understanding of the empirical consensus is that, in general, income effects are at least as large as substitution effects.  I'd put more weight on that standard finding than experiments from decades ago.

Even if you're right, you're ignoring my central point: The UBI unambiguously hurts incentives for the vast population that's currently ineligible for most government benefits.

2. You are very right to zero in on the "done properly" proviso as critical. I completely agree that tacking a UBI onto the existing system would not work. I also strenuously object to the line you get from some conservatives that a UBI should replace welfare for the poor, but leave all tax and transfer goodies intact for the rent-seeking middle and upper classes. What is sauce for the goose is sauce for the gander. Does that make a UBI a hard sell politically? Maybe. I'm a lowly economist. As the song says, "If the rocket goes up/who cares where it comes down?/That's not my department/says Werner von Braun."

My point is stronger: Even if we followed your proposal to the letter, the highest income floor you say we can afford is far lower than almost any non-libertarian would accept.  This isn't surprising, because you waste so much money on the able-bodied.

3. Taxpayers have right to attach conditions to public charity. I don't dispute that. Whether pragmatic considerations might lead them to avoid excessive or silly conditions is another matter.

I'm against "silly," too.  But where do you see "excessive" conditions in the U.S. welfare state?  Wherever I look, I see only profligacy.

4. "You shouldn't get aid unless you are poor through absolutely no fault of your own." Yes, that argument has some moral force. However, pragmatically, it is hard to pull off since it requires a huge welfare bureaucracy to decide who qualifies, and the very effort to decide has a Heisenberger-like way of changing the nature of the phenomenon you are trying to evaluate. Exhibit A is our disability system, which tries to follow the principle you suggest, but ends up with massive unintended consequences (UBI vs. disability is subject of a forthcoming post.)

The American disability system's whole problem is that it's gradually moved away from the principle I suggest.  It used to be hard to go on disability; now it's easy.  We should blame the unintended consequences not on standards, but lack of standards.  Reformist libertarians should be pushing to restrict benefits to the truly disabled, not extending them to everyone regardless of need.

Update: Dolan responds in the comments.



Comments and Sharing






COMMENTS (10 to date)
Lawrence D'Anna writes:

"We should blame the unintended consequences not on standards, but lack of standards"

When you have vague standards that require interpretation by capable, trustworthy people, which can be contested by lawyers and interest groups, they tend to decay into lack of standards.

How do you avoid this? You could have congress write down a list of conditions I guess: If you can't see, can't walk, can't use your hands, etc. then you're disabled and you get benefits, but if you have fibromyalgia then you have our sympathy but not our charity. But how well would that work? It would need to be a pretty long list, and congress is not composed of doctors.

George Mason writes:

[Comment removed. Please consult our comment policies and check your email for explanation.--Econlib Ed.]

mariorossi writes:

I find this discussion very interesting.

I tend to agree with you that a UBI is not financially viable, but I find it very frustrating that we don't included benefit reduction as part of marginal tax rate calculation. It seems just wrong to me.

I understand the UK tax system much better, so I will use it as an example. In the UK there is an child benefit of about £1300 for the first child and £900 for each successive child. Considering that free state education costs about £6000/year, it seems a reasonable split between in-kind and cash benefits. The previous conservative government decided to remove it for people earning more than £50,000. The benefit is reduced to zero at £60,000. Even ignoring the fact that the limit is based on the highest earner and not on household income, it means that a family with 2 children faces an effective tax rate of 62% (40%+22%) between 50k and 60k. For a family with 3 children, 71%. You then pay 20% VAT if you want to consume, so basically a 75% comsuption tax rate for a family of 5. Then again at 100k you lose the standard deduction, which is another 15% (pushing it to 55%) for the next 30k. The top rate is 45%. How can this be sensible policy? This micromanagemnet has to be terribly inefficient.

I'd also like to point out the many government benefits are not mean-tested: roads, police, court system, defence, etc... Adding a cash benefit to the in-kind benefits we all receive seems okay to me in principle. The value of those benefits is likely to be thousands of dollare anyway.

Another issue with UBI is that maybe we should really subsidize work instead: you could have a 100% negative income tax for the first say 5k-7.5K and then shift to a flat positive mariginal income tax after that. It might be a better system. You'd probably need to remove minimum wages entirely. I don't think a strongly progressive tax system between middle class and the rich works at all. Unfortunately many people view such changes as a subsidy to employer so that's unlikely to be popular, but I just think they are wrong.

Thaomas writes:

I'm still interested in seeing a discussion of a "UBI" done through a much higher EITC cum lower minimum wage to encourage more employment.

Samuel writes:

It is not "easy" to get on disability. The US process is still one of the strictest in the developed world. The problem is it is also incredibly hard to to get off disability once you're on it (in terms of incentives). That makes it a kind of fly trap.

dolanecon writes:

Caplan: I'm puzzled that you describe the evidence you summarize as "overwhelming." It seems fairly weak overall to me. And my understanding of the empirical consensus is that, in general, income effects are at least as large as substitution effects. I'd put more weight on that standard finding than experiments from decades ago.

I agree that the evidence from the income maintenance experiments of the 1970s and 1980s is old and inconclusive. If it were not so often cited by UBI opponents as “proof” that a UBI could not work, I would not have spent so much time examining it. As for more recent evidence, the CBO working paper that I cited is the most comprehensive literature review I have been able to find. The CBO review reaches the following conclusions regarding labor supply elasticities:

  • Among men and single women, substitution elasticities appear to have increased and now range from 0.1 to 0.3. Income elasticities still appear to be smaller in absolute value than substitution elasticities and remain in the range of -0.1 to zero.
  • Labor supply elasticities of married women—historically much higher than the elasticities of men and unmarried women—have fallen substantially in the last three decades, although they are still higher than elasticities of men and unmarried women. The substitution elasticity of married women appears to range from 0.2 to 0.4, and their income elasticity appears to range from -0.1 to zero.
If you or any readers know of other recent studies that differ from these results, please send me the links.

Really, though, the numerical values of the elasticities are not the whole story. My central point is that when a UBI replaces existing forms of aid, there is no income effect at all. Suppose, for example, that a married couple with two children now get $18,000 per year in food stamps, housing vouchers, childcare subsidies, and other forms of aid, with a benefit reduction rate of 50 percent. We get rid of all of that and instead give the family an $18,000 UBI ($4,500 per person). Their base income does not change, so there is no income effect. However, there is a full substitution effect. Instead of keeping just fifty cents from each added dollar they earn, they get to keep all of it (subject only to payroll and income taxes, if applicable). That very will might be enough to make it worthwhile to put in a few more hours at a minimum wage job—especially for the wife, whose elasticity of supply is higher (according to the CBO).

Caplan: Even if you're right, you're ignoring my central point: The UBI unambiguously hurts incentives for the vast population that's currently ineligible for most government benefits.

Sorry if I’ve ignored this point, because it is a sound one. Isaac Schapiro makes a similar point in a report from the Center for Budget and Policy Priorities, to which I replied at length in an earlier post. To both you and Schapiro, I say, yes, the elasticities argument for a UBI is stronger for households that face high effective marginal tax rates than for those who face lower EMTRs because they receive no benefits. In the extreme case, they are exposed to the full income effect and get no help at all from the substitution effect.

I would make two points here. The first is empirical. I’d like to see a count of how many poor households or individuals get no benefits. Is it a “vast” number, or a relatively small one? And what kind of people are they? If, for example, they are mentally ill and substance-dependent homeless people, I’ll concede that a UBI would be unlikely to get them into the job market. I’ve often said that a UBI is not a magic bullet. It certainly is not an effective weapon against mental illness and substance abuse. I’ll try to do some research on this and post the results. Thanks for pointing out the importance of the question.

Second, I’ll concede that the target group I have the most sympathy for in my writings on the UBI are households that are already at least marginally attached to the labor market and who have incomes in the range from half to double the poverty level. They are the ones who are most screwed over by the current welfare system and who would benefit most from a UBI, IMO.

Caplan: Even if we followed your proposal to the letter, the highest income floor you say we can afford is far lower than almost any non-libertarian would accept. This isn't surprising, because you waste so much money on the able-bodied.

First, I agree, there are some non-libertarians who want a UBI high enough to let everyone live a comfortable middle-class life without working at all. That is particularly the case with some who write about the UBI in the context of an imagined automated utopia in which no one at all has to work. I say, pie in the sky. But let’s not get off topic. This is a debate about why libertarians might take a UBI seriously. I don’t have to worry that my $4,500 UBI is too stingy. I have to make the case that it is not too high.

Second, I think when you write that my version of the UBI “wastes so much money on the able bodied,” you haven’t really thought through the whole program. Again and again, I have emphasized that in order not to “waste money on the able bodied,” a viable UBI I must replace not just welfare for the poor, but “middle-class welfare” as well—much of which comes in the form of tax expenditures.

For example, let’s say a middle-class family of four is getting $18,000 per year in assorted tax deductions—minimum allowance, mortgage interest, retirement savings, employer-provided insurance, an all the rest. If we give that family an $18,000 UBI, and at the same time cancel their $18,000 of tax expenditures, we have not “wasted” any money at all. What we have done is to change the form in which that money is spent, and do it in a completely revenue neutral way. Along the way, we get rid of some of the perverse incentives in current tax expenditure programs, such as “job lock” resulting from deductions for employer-provided healthcare and the bias against renters embodied in the mortgage deduction.

Similarly, to avoid “wasting money on the able bodied,” I propose that all safety net programs for the nonpoor also be integrated into the UBI. Take unemployment benefits, for example. In some cases those are greater than my suggested level for the UBI, in some cases less. I say, give eligible individuals the right to take unemployment benefits or the UBI, whichever is greater, but do not allow “double dipping.” Same for Social Security and some smaller programs.

Caplan: But where do you see "excessive" conditions in the U.S. welfare state?

Example: Subjecting welfare recipients to drug tests. Example: Conditions that restrict interstate mobility, as is often the case with programs administered at the state or municipal level, and which would be intensified with some GOP proposals for “block granting” everything. Example: Provisions that unnecessarily add to the red tape of getting benefits, as with disability programs (see below).

Caplan: The American disability system's whole problem is that it's gradually moved away from the principle I suggest. It used to be hard to go on disability; now it's easy. We should blame the unintended consequences not on standards, but lack of standards. Reformist libertarians should be pushing to restrict benefits to the truly disabled, not extending them to everyone regardless of need.

I know I have been remiss in not dealing with disability at length, and I keep promising to do so. Be patient. Meanwhile, just one point: The real problem, as analyzed by Autor and others, is not that it is too easy to get on disability, but that it is too hard to get off.

It often takes several years and costly legal help to get on disability, especially if the person is (as you suggest) only marginally disabled. Once you succeed, there is a tremendous incentive to stay on, since disability is for life and includes full medical benefits. Even a short spell of work can kick you off for good, and once you are off, it is even harder to get back on. That is why I call disability a UBI with a perverse twist: You get it only if you guarantee that you will never work again.

My basic proposal would be to integrate the UBI and disability through a double-dipping rule. Now, leaving disability is an all-or-nothing, once-in-a-lifetime choice. If you attempt to get back to work is a failure, you are screwed. With a UBI in force, a person getting disability would at least have the UBI to fall back on. The UBI might be less than full disability is now, but the step down in benefits would be less for a person leaving disability. That would improve incentives for getting back to work, and also reduce incentives to spend on the legal representation a marginally disabled person needs to get on disability in the first place. I promise, I’ll expand on this in full in a coming separate post.

Tracy W writes:

@Dolan:

For example, let’s say a middle-class family of four is getting $18,000 per year in assorted tax deductions—minimum allowance, mortgage interest, retirement savings, employer-provided insurance, an all the rest.

From your own figures on your website, tax deductions in the USA, excluding healthcare, average about $1,825 per capita. Therefore a family of four would be averaging about $7,300 per year in tax deductions, less than half of your $18,000 per year.

There presumably is some middle-class family of four that meets your description. But a UBI would affect everyone (the name "universal" is a bit of a giveaway). You have not shown any evidence to disprove Caplan's claim in general.

Ed Dolan writes:

Tracy-- Dang, you have the toughest questions.

I'll have to go over the math for a more complete answer. I need to do an update of my "affordable" post, anyway. And you are right, the example I gave in this comment is just a "what if," I don't mean it as an average.

A couple of questions. One is, what particular post did you get the $1,825 from? Hard to refute since I can't recall the exact source of that number.

Another is, are you including "standard deductions" or only itemized deductions in the $1,825? Standard deductions are way bigger than your number--$12,600 for a married couple, according to that linked source. Also, are you taking Social Security into account? Finally, is your (or my) $1,825 the average for all taxpayers, or for those in the middle class range?

In short, I am not sure how strong your point is without some more numbers to work with. But thanks for bringing it up.

Tracy W writes:

@Dolan:
The $1,825 figure is from http://www.economonitor.com/dolanecon/2014/01/13/could-we-afford-a-universal-basic-income/

To quote from your post linked above:

Even without healthcare deductions, other provisions of the personal income tax came to and estimated$577 billion for 2013.
....
If we eliminate all of these tax expenditures, we can add $1,825 to our per capita UBI grant, bringing it up to $3,408.

You also mention the personal exemption in that post. To quote your post, again:

We could add the personal exemption into the mix as well. The personal exemption for 2013 is $3,900. I have not seen a good estimate of the total budgetary effect of the personal deduction, but we can make an approximation. For someone in the 15 percent tax bracket, the exemption would be worth $585. If 100 million taxpayers get a benefit averaging that amount, the total would be $58 billion, or enough to add another $185 to the UBI grant, bringing it up to $3,591 per capita.

Switching back from your post to your comment here, you say:

In short, I am not sure how strong your point is without some more numbers to work with.

The numbers are yours. And, as you're proposing the policy, the burden of proof is on you too.

Hazel Meade writes:

I have a friend of a friend who is on disability and I have heard, not sure if true, that if he got a job and got off disability he would be required to pay back the disability benefits.
Anyone know if there is any credence to that?
Not sure what the cause of his disability is, I think mental health related.

POST A COMMENT




Return to top