Bryan Caplan  

Caplan-Wilkinson Universal Basic Income Debate

Bernanke on monetary offset an... My Speech at Kansas State Univ...
At this year's 10th Anniversary Students for Liberty conference, I'm debating Will Wilkinson on, "Should libertarians support a Universal Basic Income?"  My answer is an emphatic no.  Not only libertarians but liberals, conservatives, and reality-based socialists should oppose this populist earworm.  In a world of limited resource, means-testing is always the way to go.  Indeed, I struggle to pass the Ideological Turing Test here, but I'll sincerely try.

Question: Any sub-topics you'd like us to address?  Please share in the comments.

Comments and Sharing

COMMENTS (37 to date)
Jim P writes:

Will means testing not introduce implicit marginal taxes that will reduce incentive to work?

David Youngberg writes:

Is your only problems with UBI that it's not means-tested? Would you favor a means-tested basic income rather than a universal one?

Lawrence D'Anna writes:

"Means-testing is always the way to go" isn't just a bad answer, it's not even an answer. The question is "what shape should the gross-income to net-income graph have?". I have no idea what shape "means testing is always the way to go" corresponds to.

Sanjeev Sabhlok writes:

Good to know you're against basic income. Have linked this post to my blog page that has my views (against) and the views of others who oppose (+ some who support):

It amazes me that "efficiency" arguments are applied by some people in support of basic income - at the total exclusion of the consideration of incentives. How can something be efficient if its inevitable consequence is to destroy **all incentives to produce**? Basic income is communism in another guise.

Ben Pacini writes:

Would you support a guaranteed minimum income that tapers off as income goes up? (Mincome?)

Would you support replacing welfare programs with wage subsidies for low-income workers?

Nicholas Weininger writes:

To Jim P's point, what do you think of Casey Mulligan's research on impacts of existing implicit marginal tax rates created by welfare programs? Would your favored means tested approach create lower or higher implicit marginal tax rates than the status quo?

John Strong writes:

How about a UBI that is formulated as dividend payments to resident-shareholders in a private city?

Thomas writes:

I'd like to see the moral case argued, rather than "X is more efficient than Y" -- Y requires theft at gunpoint.

Rajat writes:

Agree with Lawrence.

J Storrs Hall writes:

Re the ITT, have you read Murray's In Our Hands?
It's one of the more surprising-given-its-author proposals I've seen.

Maniel writes:

I think that Milton Friedman called it "negative income tax."

Eli writes:

I think talking about means testing is letting the perfect become the enemy of the good.

The question is whether UBI is better than nothing, not if it's better than means testing.

GregS writes:

1) Is the UBI the same everywhere, or does it vary based on the average income of the region?

2) Do children get an annual payment (the same as everyone else or something different)? If “yes”, doesn’t this potentially encourage having unwanted children, and if “no” doesn’t the UBI fail to address the problem of child poverty? (This trade-off plagues any welfare program, just pointing out here that the UBI doesn’t make it go away.)

3) I second Laurence D’Anna’s comment above.

Matthew Rafat writes:

Means testing won't work--it'll be subject to political manipulation. The point of UBI is to minimize rather than increase subjective political decisions. More here:

Matthew Rafat writes:

You write, "Systematically replacing expensive universal programs with cheap means-tested programs would make the modern welfare state almost unrecognizably small."

You miss the idea of UBI in its ideal form--the replacement of trillions of dollars currently going to old Americans in unsustainable ways and diversion of those dollars to all citizens, increasing freedom by reducing incentives to work in dead-end or menial jobs.

Basically, I agree with you that current welfare programs are too "old-heavy" and need to be reconfigured. In an age where robots can handle almost all routine labor involving servers, food preparers, and retail, what is your justification for keeping those jobs? If teaching work ethic and customer service are your goals, surely there are better ways.

Nodnarb the Nasty writes:

How about True Free Trade versus managed free trade (i.e. NAFTA)?

Joshua Abell writes:

Another vote for high implicit marginal tax rates suffered by poor people due to benefit cutoffs & phaseouts.

Matt writes:

Simple example:

Set UMI (Universal Minimum Income) at the poverty level for a single person (currently, about $11,000). Allowed to earn up to that amount without benefit reduction. Lose $1 in benefits for each $3 of earned income. So a single person is at zero benefit at $44,000.

It's means tested, and while 33% is a high marginal tax rate, there's no "benefits cliff" to worry about.

I'd prefer free markets with minimal regulation and few barriers to entry for both workers and firms, but something like this may be more feasible.

Matthias Görgens writes:

Means testing is fine, but why keep multiple redundant systems for that around? Let's just simply unify (income) taxation and welfare---juat like people want universal basic income to be.

As pointed out above, the current disparate systems lead to crazy high marginal effective tax rates for poor people.

Thaomas writes:

Why not a much higher EITC/wage subsidy, a lower minimum wage and a shift from wage taxation to consumption taxation for SS-Medicare benefits ? We could still have unemployment benefits and SNAP with a work requirement but the EITC would be a powerful incentive to work.

David Condon writes:

I second calls for discussion of replacing existing welfare systems with a guaranteed minimum income. I had to double-check it with google, but basic income is when the income isn't means tested and minimum income is when the income is means tested. Also, with respect to several other comments, Bryan Caplan has already discussed his views on means testing at length:

ColoComment writes:

Because "fair" is such an important concept these days, with either the basic income or the minimum income (as defined per Mr. Condon above), should an individual's benefit be adjusted (perhaps annually) for the variable purchasing power of a dollar in each state/major locality?

E.g., a $10k benefit in Palmdale, California does not have the purchasing power of a $10k benefit in Dalhart, Texas. Therefore, post benefit, the Dalhart resident is "richer" than the California resident. How "fair" is that?

See here:

John Hamilton writes:

Please bring up the experiences of Native Americans with universal basic incomes. Here's the relevant article from The Economist:

Mark S writes:

UBI supporters often use the Nobel-prize-winning-Milton-Friedman-supported-negative-income-tax-which-is-kind-of-basic-income argument and I want to hear your thoughts on that.

simon writes:

Lawrence D'Anna has it right:

"Means-testing is always the way to go" isn't just a bad answer, it's not even an answer. The question is "what shape should the gross-income to net-income graph have?". I have no idea what shape "means testing is always the way to go" corresponds to.

The limited resources thing is a red herring because a means testing can always be replaced by a universal benefit and increased tax to replace the clawback of the means testing. Any measurement you use to means test can equally be used to tax. This results in the same incentives and final resource distribution. On paper, you get a higher gov. spending and tax with the latter method but this is merely on paper and has no basis in reality.

While in principle you can use either method, if you use multiple methods to shape the gross-income-to-net-income curve, you reduce the clarity of the system and are likely to end up with incomplete integration of the tax brackets and clawbacks where the curve ends up having a weird shape for no good reason. Thus in a system where you have progressive taxation like ours, IMO there is a strong case to never means test. Provide universal benefits instead and use the tax to shape the curve.

Miguel Madeira writes:

Sanjeev Sabhlok:

"How can something be efficient if its inevitable consequence is to destroy **all incentives to produce**?"

This doesn't make any sense - if anything, the difference between a universal basic income and the mean tested systems is that the first does not have the desincetive to produce implicit in the second (because you don't lose the benefit if you work more and earn more, you still have an incentive to produce, unlike in a means tested system)

Looking to the exemple in your blog:

"It is a top-up. Say, the social minimum (really frugal) is $5,000 per year per capita in Australia today. Let’s say that a person works very hard and earns $4,800. Then the system would top up with $200, as part of social insurance."

The problem of a top up system (like the portuguese Rendimento Social de Inserção and Brazilian Bolsa-Família) is that, if you work hard and earn $4,800, you receive $200; and if you work not so hard, and earn $2,400, you receive $2,600; earning $5,000 each way, you don't have an incentive to produce (it is, for all efects, a subsidy payavbe to not working, because, the less you work, the more you receive).

In contrast, imagine an UBI of $5,000; if you work hard and earn $4,800, in the end you receive $9,800 ($4,800 of earned income + $5,000 of UBI); if you work not so hard and earn only $2,400, you will receive $7,400 ($2,400 + $5,000)*; then, in an UBI system you have a bigger incentive to be productive

*I am ignoring taxes there - it is true that it will change the numbers

Matthias Görgens writes:
Why not a much higher EITC/wage subsidy, a lower minimum wage and a shift from wage taxation to consumption taxation for SS-Medicare benefits ?

Thaomas, even better would be to tax land and not capital or labour.

There would be no need for traditional means testing for taxes nor benefits in such a system, thus no incentives to work less. Richer people are likely to use more (valuable) land directly or indirectly, but that would be the only source of progression. The UBI would certainly drive up rents, thus partially funding itself.

Miguel Madeira writes:

For me, a "basic income" is any combination of taxes and subsidies with the following properties:

- The liquid tax is negative for an income of zero

- The liquid marginal tax rate is always zero or positive

- The liquid marginal tax rate does not decrease with income

If in some incomes the marginal tax rate is negative, what we have is an EITC system (where an increase of income reduce the net taxation); if lower incomes (specially in the brackets where the tax is negative) have a big marginal tax rate than higher incomes, what we have is a minimum guaranteed income.

JFA writes:

Bryan, could you cite your sources for the data you use in your slides? I would like to take a deeper dive into the numbers.

RPLong writes:

I hope someone captures audio or video of this debate, as I cannot make it but would love to watch. Please let us know if that happens.

JFA writes:

It looks like you got the info for your Northwood talk from Heritage's "Understanding Poverty in the United States: Surprising Facts About America's Poor", which was published in 2011. I'm guessing here since both of you say "one-third have a wide-screen" TVs and "41 percent of poor households own their own homes". You may want to update your numbers since that was based on (in some cases) 2009 data. The American Housing Survey for 2015 (Heritage and you use the 2009 data) has 37 percent of poor households owning their own homes. A difference that may or may not be statistically significant, but if you are going to use point estimates, you should use the most recent numbers.

Philo writes:

I take it that Wilkinson will be arguing that a Universal Basic Income, *replacing (almost) all our present welfare programs*, is better than the welfare system we presently have, and better than any replacement that it is *politically realistic* to hope for in the near future (while UBI itself *is* politically realistic). To argue (as I expect you to do) that there is some conceivable welfare system that would be better than UBI will be irrelevant to the question at issue, unless your preferred system is *not politically very improbable (i.e., unrealistic)*. A better argument by you would be that UBI is not, after all, politically realistic, so libertarians would be wasting their time advocating it.

Tom Jackson writes:

I, too, would like to know if the debate will be recorded and will be made available as a podcast or a video.

JFA writes:

Bryan, it looks like your numbers come from a 2011 Heritage report on poverty in America (I infer this from the many similar numbers used, but perhaps there is a Q source you are both working from). You may want to update the numbers with newer data. You report that 41% of poor households own their home (which comes from the 2009 American Housing Survey). The 2015 survey as it at 37%.

Matthew Rafat writes:

ColoComment, you wrote, "E.g., a $10k benefit in Palmdale, California does not have the purchasing power of a $10k benefit in Dalhart, Texas. Therefore, post benefit, the Dalhart resident is richer' than the California resident. How 'fair' is that?"

The idea behind no geographical adjustments is to encourage people to relocate and bring wealth with them to smaller cities and more rural places. The last thing America needs is more people in its largest cities, which already receive the lion's share of VC, mortgage, and business loans. In addition, if implemented properly, UBI would cause welfare rolls to experience relief, especially in more expensive cities.

John Alcorn writes:

Bryan Caplan wrote: "Any sub-topics you'd like us to address? Please share in the comments."

I would like you to address a sub-topic raised by Scott Alexander (Slate Star Codex blog) in his imaginary graduation speech. Dr. Alexander argues that a universal basic income grant could be funded adequately by eliminating government subsidies to education and reallocating the funds to the grant:

"At $11,000 average per pupil spending per year times thirteen years plus various preschool and college subsidies, the government spends $155,000 on the kindergarten-through-college education of the average American. Inspired by a tweet: what if the government had taken this figure (adjusted for inflation) and invested it in the stock market at the moment of your birth? Today when you graduate college, they remove it from the stock market, put it in a low-risk bond, put a certain percent of the interest from that bond into keeping up with inflation, and hand you the rest each year as a basic income guarantee. How much would you have? And I calculate that the answer would be $15,000 a year, adjusted for interest. We can add the $5,800 basic income guarantee we could already afford onto that for about $20,000 a year, for everyone."
Bryan Caplan's forthcoming book, The Case against Education, argues that, from a social vantage point, expenditures on education are mostly wasteful signaling, and that public subsidies to education should be eliminated.

Many critics of universal basic income argue that the grant would greatly exceed the sum of expenditures on all extant welfare programs. However, Scott Alexander's argument changes the calculus.

Of course, elimination of public expenditures on education in order to fund a universal basic income grant is as imaginary as Scott Alexander's graduation speech!

Nonetheless, this would be a neat sub-topic for the debate, given that Will Wilkinson needs to identify an expenditure-neutral way of funding the grant, and given that Bryan Caplan is the voice of the case against education.

John Alcorn writes:

Next door, at EconTalk, in the comments about Michael Munger's podcast about a universal basic income grant, Kevin Ryan points out that Scott Alexander's calculus depends on unrealistic assumptions about real interest rates.

Perhaps a much more modest, but still worthwhile consideration can be salvaged from Scott Alexander's argument: Re-allocation of government expenditures, from education to the funds for a basic universal income grant, would substantially increase the amount of the feasible grant.


Return to top