Bryan Caplan  

The Many Faces of Means-Testing

PRINT
My Speech at Kansas State Univ... Ending the Requirement to Show...
Isn't a Universal Basic Income just another name for a negative income tax, such as Tax = -$10,000 + .3*Income?  If so, isn't a Universal Basic Income means-tested by definition?

The answer to the first question is Yes.  UBI is just Milton Friedman's negative income tax in new packaging.

The answer to the second question, however, is more equivocal.  The UBI is means-tested in the weak sense that your net payment falls with income.  But the UBI dispenses with many other traditional forms of means-testing.  Most notably:

1. Means-testing by age.  Most welfare states prioritize children and the elderly.  The implicit theory is that, unlike prime-age adults, the very young and the very old are unable to provide for themselves.

2. Means-testing by dependents and marital status.  Most welfare states prioritize single moms with minor children.  The implicit theory is that single moms have reduced opportunities to work due to their family responsibilities.

3. Means-testing by health.  Most welfare states prioritize the disabled.  The implicit theory is that they're not healthy enough to work.

4. Means-testing by job history.  Most welfare states prioritize people who recently lost their jobs over people who have never worked, or lost their jobs a long time ago.  The implicit theory is that the short-term unemployed are unlucky, while the long-term unemployed are lazy.

If your UBI proposal includes factors like these in its formula, it's very hard to see what makes it a UBI. 

If your UBI proposal dispenses with most or all these factors, then it is a distinctive reform indeed.  But "distinctive" is a far cry from "good."

Advocates correctly note that dropping multi-faceted means-testing reduces moral hazard: If your monthly payment doesn't depend on your health, you have no reason to fake bad health.

But there is also an gargantuan disadvantage: Dropping multi-faceted means-testing greatly increases the number of eligible recipients.  If perfectly able-bodied, childless adults are eligible for free money, plenty will take it - and many won't work at all.  Taxes on remaining workers have to rise to pay for them.  This probably won't create a "UBI death spiral," but a milder sloth spiral definitely kicks in, especially over the longer run as stigma against idleness erodes.  And the burden of supporting able-bodied non-workers is also very likely to cut into funding for the more deserving poor.

Frankly, given the bleak long-run fiscal forecast for the U.S., I'm baffled that anyone with libertarian sympathies takes the UBI seriously.  The welfare state is already unsustainable, largely because our means-testing by age and health isn't stringent enough.  The elderly may have trouble working now, but since they had a lifetime to save for their own retirements, few of the indigent elderly are victims of circumstance.  And given the huge long-run rise in the share of U.S. adults on disability despite rising health and less strenuous jobs, its clearly far too easy to plead disability.

What's especially strange is that the bleak long-run fiscal forecast makes old-school libertarian austerity more relevant than ever.  Why are so many libertarians running away from our core ideas when conditions are nearly ripe for mainstream America to finally listen to us?




COMMENTS (38 to date)
Matthias Görgens writes:

Means testing is not free. The bureaucracy to administer the means testing is non-trivial.

Since we are talking about hypothetical fiscals anyway, replacing most taxes on labour and capital with a land value tax would solve basically all fiscal problems, without causing any dead weight loss on the economy. (Ie a libertarians dream.)

Yes, keeping the total amount of redistribution low, is basically the best argument against UBI.

The argument in favour of folding means testing of taxes to pay and benefits to receive into a single bureaucracy remain valid, at any level of total redistribution aimed for.

yarbel writes:

To add to Matthias' point, there is also a cost for false negatives. This cost may be very high if means testing is based on fuzzy concepts ('ability to work', 'health'), and subjecting a large segment of the population to means testing implies a high overall cost of false negatives.

Bryan's point is a good one, but I worry that he makes too strong of an argument by making it categorically superior to have means testing, without reference to actual administration costs.

I agree with the two comments above and I add: leaving moralistic prejudices aside, what's the difference between being incapable to work due to a lazy character and due to a low IQ? Or due to a weak body and due to a weak will?
It's also highly probable that keeping chronically dysfunctional people away from workplaces can be good for productivity, just think about all the time wasted in the office with people that frequently create trouble and help very little.
Milton Friedman is definitively right on this one.

Hauu writes:
I'm baffled that anyone with libertarian sympathies takes the UBI seriously.

The consequent libertarian case for a UBI is one of justice, and it's a compromise towards the center. While the more radically libertarian solution is nothing short of anarchism. Not all libertarians support anarchism, and not all people want to do a 180 degree turn with their philosophies, championing paternalism or aristocracy.

UBI comes in there in the middle between those extremes, those extremes being:

1. Consequent abolishing of property, seeking justice in an impossible situation of nonviolent, merit-based negotiations for use of anything, and the other,

2. declaring individual property an eternal justice, that no one would ever admit to having obtained from a pool of previously publicly available space of things and ideas.

The welfare state is already unsustainable, largely because our means-testing by age and health isn't stringent enough

As I see it, that the welfare state as we see it today is unsustainable is owed to government enforced ownership and natural and circumstancial monopolies, paired with an unwillingness to much more burden ownership directly or indirectly.

Brad D writes:

You say, "If perfectly able-bodied, childless adults are eligible for free money, plenty will take it - and many won't work at all."

Perhaps, but if this occurred en masse, wouldn't the market self-correct? A shortage of workers would drive up wages, enticing many to re-enter the work force.

AntiSchiff writes:

Brad D,

The problem is that it could hurt demand too. Those content to live with less work and less income would also have less demand.

Pithlord writes:

You are using "means test" in a non-standard way. Normally it refers only to a very high clawback based on additional income or wealth. You could have a NIT disability benefit or age-based benefit.

The real point is that means tests imply really high effective marginal tax rates, sometimes > 100%. That is bad for familiar disincentive reasons. Not extending the benefit to able-bodied adults makes sense. Also it wouldn't be politically possible to give Unconditional money to people who just don't want to work

David Condon writes:

"replacing most taxes on labour and capital with a land value tax would solve basically all fiscal problems, without causing any dead weight loss on the economy."

You're arguing for the dramatic increase in price of a single type of good. Urbanization would increase as property ownership costs would increase, necessitating much smaller dwellings. Any type of business which relies on owning large swaths of land would struggle, particularly when competing internationally. Domestic extraction of resources would plummet. Farming would decrease in favor of fishing, which could plausibly lead to significant overfishing issues.

"To add to Matthias' point, there is also a cost for false negatives."

If an individual qualifies for a government program, but they don't receive the benefits because of an administration error, won't that reduce their benefits and therefore reduce costs? Are you referring to the costs of correcting the false negative? It is not necessarily the case that the administration costs of correcting the error will be higher than the reduced costs from not having to pay benefits for a period of time. Did you mean a false positive? The cost of providing benefits to the occasional false positive will be lower than the cost of providing benefits to everyone.

You both assume that the cost of determining eligibility of an individual for a government program would be higher than the cost of paying that person their benefits from the government program without checking eligibility. I would assume just the opposite. It would seem very implausible to have an eligibility checking system that would cost thousands of dollars. If the program is efficiently designed, then administration costs should be fairly low relative to the cost of providing benefits. Administration costs represent less than 1% of social security's costs for instance. https://www.ssa.gov/oact/STATS/admin.html

simon writes:

As noted in the previous thread, means testing via a clawback is equivalent effect to a universal benefit plus a tax equal to the clawback.

So if you say:

Frankly, given the bleak long-run fiscal forecast for the U.S., I'm baffled that anyone with libertarian sympathies takes the UBI seriously.

where your proposed alternative is clawbacks based on age, dependents, health, job history etc. then that is in effect equivalent to saying:

Frankly, given the bleak long-run fiscal forecast for the U.S., particularly with a UBI, I'm baffled that anyone with libertarian sympathies takes seriously the idea of not levying taxes based on age, dependents, health, job history etc.

Yes we would need higher taxes to pay for a UBI (probably, depending on what stuff is canceled) - but that doesn't mean the taxes should be specialized taxes on those sorts of things (and means-testing clawbacks as you propose are, in effect, specialized taxes on those things). Much more efficient to have broad-base, moderate marginal rate taxes on income, consumption etc.

Tracy W writes:

Matthias

Since we are talking about hypothetical fiscals anyway, replacing most taxes on labour and capital with a land value tax would solve basically all fiscal problems, without causing any dead weight loss on the economy.

How? I mean, I get that a land value tax has no deadweight loss, but I don't get how it can raise unlimited amounts of funding. If I owned land and the land tax got too high for me to fund comfortably out of my income (and I couldn't make more productive use of the land), I'd sell or failing that, abandon the land.

Brad

Perhaps, but if this occurred en masse, wouldn't the market self-correct? A shortage of workers would drive up wages, enticing many to re-enter the work force.

If wages rise without increased productivity, to attract these people back to the workforce, then prices will rise, making it harder for the disabled who can't work to live on the basic income.

Tracy W writes:

@Pedro:

leaving moralistic prejudices aside, what's the difference between being incapable to work due to a lazy character and due to a low IQ? Or due to a weak body and due to a weak will?

Why do you think that a low IQ or a weak body leaves someone unable to work? Comparative advantage! A low-IQ person or a weak-bodied person can add a lot of value freeing up average-IQ/able-bodied people to do work that requires those capacities.

Yes there are people with problems so extensive that they can't work. But when problems are that bad we can't really distinguish between "extraordinarily lazy/weak-willed" and "serious mental problems". The concern is the moderately lazy, moderately weak-bodied, moderately weak-willed.

It's also highly probable that keeping chronically dysfunctional people away from workplaces can be good for productivity

It would be even better for productivity if the chronically dysfunctional could turn themselves into functional people. Hitting bottom can be useful for that. Giving addicts money to buy more drugs generally doesn't.

Matthias Görgens writes:

Tracy W,

You are totally right.

What we actually would want to be taxing is land rent, ie the yearly economic benefit derived just from the location (not from any capital investment like a house on top or labour).

The land rent is seldom directly observed on the market. Eg an owner occupier doesn't pay herself rent.

As a proxy the Georgists suggest taxing land by value.

A standard arbitrage argument between bonds and land prices gives (for a fixed plot of land per year):

(Price of land) * (interest rate + land value tax) = (land rent)

That arbitrage equation tells us a few things:

Given a fixed land rent and a land value tax rate, even if that rate is above 100%, there's always a resulting land price that solves the arbitrage condition. Rearranging the equation gives:

(Effective tax on land rent) = (land value tax)/(land value tax + interest rate)

This never goes above 100%, land won't be abandoned. There's probably some other considerations for how to set the exact land value tax rate, but as long as it's a decent multiple of prevailing interest rate (eg ten times?) it will capture most of the land rent. On the other hand, making it too high isn't too much of an economic problem either. (It does have an influence on equilibrium land prices, and there might be political considerations.)

Next comment will be about how much actual revenue this tax can raise. (It's not unlimited, but large.)

Tracy, you appear to not realize that your argument supports a minimum universal revenue system. You say:
"Why do you think that a low IQ or a weak body leaves someone unable to work?"
The fact is that those that are able to work will go back into the labor force under a minimum revenue system because the marginal rates at the bottom will be much lower. On the contrary, it's the current system that you and Caplan are defending that keeps them from working and favors restrictions in freedom and regulatory overreach and waste.

mariorossi writes:

I personally find the arguement that you cannot eliminate situation based benefis very convincing. You will never be able to raise UBI high enough to support someone with significant disabilities (either health or age related). The numbers would not add up: it would be too generous for someone without such needs. And requirements come in a wide spectrum, so the problem is hardly simple.

The burocracy needed to determine if someone is truly in need of help (either due to bad health or accidents) is required. This would not be different in a privatised system. Society needs someones to perform such checks to pay insurance claims at the very least, and I think designing a market to perform such functions is difficult. So I think the cost advantage of UBI is oversold.

I am not a strong believer in personal responsability to be honest, that's just my prior, so I think it is very unlikely I would see eye to eye with Prof Caplan. I doubt we could reconcile that prior, so there is little point in arguing about it.

I disagree with the idea expressed in older posts that marginal tax rates for low income people are unimportant. Massive jumps in marignal tax rates at magic numbers are in my mind ugly. I am not totally against higher marginal rates for people that receive circumstance-based benefits, but I think such marginal rates should never exceed the top marginal tax rate or at least not exceeded it by much (definetly less than 5%).

This might seem inconsistent with my previous statement but I believe it not to be. I see personal productivity as largely determined by fate/luck, but I still somewhat believe in incentives.

As a commenter in the previous post said, politicians should be required to submit a clear function between market income and available income. Since we all receive many fixed in-kind benefits anyway (e.g. police protection, public roads and paks and so on...) I find cash fixed benefits hardly objectionable. We should get a clear statements that highlights fixed in-kind and cash benefits, plus marginal taxes and conditional taxes and benefits. Reducing the number and types of conditional benefits would be great as well (e.g. mortgage deductions, untaxed employer benefits)...

Amy Willis writes:

For another (libertarian) perspective on UBI, you might also be interested in this EconTalk episode with Duke's Mike Munger.

Chris Wegener writes:
If perfectly able-bodied, childless adults are eligible for free money, plenty will take it - and many won't work at all.

This statement runs counter to everything I have ever read about human nature. Everyone wants to work to some degree. The "lazy poor people" meme mostly comes from communities where no work is available.

A UBI would allow people to work in even marginal ways. If it becomes since it would reduce or eliminate minimum wage people and companies could hire people for nice to have jobs that are not directly economically viable.

Thaomas writes:

The sustainability or not of SS-Medicare depends on the deadweight loss of the tax increases needed to make them grow no faster than GDP. Considering how much loser those costs would be with a shift from the current mix of wage taxes, corporate and personal income taxes compared to a progressive personal consumption tax, long run deficit panicking seems misplaced.

Dave's ideas for reducing the inefficiencies in health care production to which I'd include allowing Medicare-Medicaid to pay for care performed abroad and freer immigration and certification of health care professionals would also make long term financing of age-related safety net programs easier.

Ed Dolan writes:

I am a libertarian and I take UBI seriously. I think you are confused on some points.

1. A true UBI is not means tested. It is universal in the sense that it goes to everyone regardless of income, age, health, etc. Don't mix up a UBI and an NIT. (Sometime people use the term "basic income guarantee" or BIG to cover both NIT (means tested) and UBI (not means tested).

2. UBI can be affordable if it is a replacement for existing government benefits. UBI replaces welfare for the poor, benefits like mortgage deduction for middle class and higher. For benefits like Social Security, unemployment, and disability, people have choice of taking existing benefit OR UBI but not double dipping.

3. Appeal to libertarians:

a. Pragmatic: Maybe we don't like to give government benefits, but if we are going to, a UBI is more efficient and has fewer disincentives to work.
b. Philosophical: Fits with a classical liberal view, a la Hayek, that we need some kind of social safety net
c. Lifestyle: UBI, unlike current welfare, leaves recipient free to live life according to own values rather than those imposed by a welfare bureaucracy.

More here.

Tracy W writes:

@Matthias: your second equation comes out of nowhere. It is not a rearrangement of the first equation, the terms "price of land" and "land rent" were in the first equation and now have just disappeared from the second equation.

As for your second equation in its own terms, I don't get its relevance. My argument was that if the land tax gets too high I would abandon the land. Your second equation merely says that if you take the land tax (rate I presume) and add on something else, the ratio of the land tax to the land tax + something else will never get above 100%. But we could equally say:

Effective tax on income) = (income tax rate)/(income tax rate + interest rate).

That equation will also never get above 100℅ as long as the interest rate is positive. So, by your logic, making income tax too high wouldn't be much of an economic problem either.

@Pedro: that may be so. I don't actually know what a "minimum universal revenue" system is and googling it didn't turn up anything. If this was a typo for "minimum universal income", my objections to UMI is that it either requires cutting the income of the poorest drastically, or raising taxes drastically, or, depending on how you define "drastic", both.

simon writes:

Tracy: Matthias's argument is that, if the tax is based on the market price of (unimproved) land, a high tax rate on it will depress the market price to lower the tax collected on it. This will keep the tax collected always below the rent.

That argument is correct* as far as it goes, and the equation is also valid (he is defining "Effective tax on land rent" as (land value tax)*(price of land)/(land rent), you can see with that substitution the second equation is indeed a rearrangement from the first).

Nonetheless, that argument doesn't address your original point that the land tax can't raise unlimited funding. I await Matthias' following post where he will argue that the finite funding it can raise is large - I am skeptical.

*assuming land rent, land tax rate and interest rate are all constant. If they vary, land tax collected can exceed land rents temporarily. Land values will still remain positive.

Matthias Görgens writes:

Tracy W,

Sorry for being a bit confusing.

Let's be a bit more explicit.

A piece of land yield an economic benefit per year, call that (land rent). Call the taxes paid on that (land value tax rate * land price).

The net benefit for a land owner is: land rent - (land value tax rate * land price).

In equilibrium market participants will value that piece of land at the same value as a perpetual bond with the same coupon: bond price * interest rate = bond coupon

So we get:

bond price * interest rate = bond coupon = land rent - (land value tax rate * land price)

Renaming:

land price * interest rate = land rent - (land value tax rate * land price)

As you can see, a higher interest rate will decrease land prices (and all assets prices in general); and a higher land value tax rate will also decrease asset prices.

Solving for land price gives:

land price * interest rate + (land value tax rate * land price) = land rent

land price * (interest rate + land value tax rate) = land rent

land price = land rent / (interest rate + land value tax rate)

Which is basically the first equation I gave you. I hope you can see its validity now, and why it makes sense to talk about interest rates here?

(Coincidentally, bringing income taxes and interest rates into the same equation would only make sense when you are trying to figure out how much a human earning a certain amount of money is worth as an asset. But I don't think the economics of slavery are useful in this discussion.)

So where does my second equation come from?

I was trying to work out how a specific land value tax rate translate to as an effective tax rate on the land rent.

By definition:

(tax rate on land rent) := (land rent) / (price of land * land value tax rate)

Substitute the formula for the price of land derived above and you get this monstrosity:

(tax rate on land rent) := (land rent) / (price of land * land value tax rate) = (land rent) / (((land rent) / (interest rate + land value tax rate)) * land value tax rate)

Simplify to get:

(tax rate on land rent) := (land value tax rate) / (interest rate + land value tax tax)

All that equation tells you is that no matter what the land value tax rate, there's a natural limit to how much revenue it can generate: the whole of the land rent---it never goes above 100%. So land will never be abandoned. (Of course, in practice there might be transaction costs and bureaucratic overheads etc.)

Are you with me so far?

Just to be sure, I'll address your reductio ad absurdum directly:

Your example would make economic sense in a slavery economic (or think of a tax on robots, if that makes you feel more comfortable).

An income tax can go above 100%, and thus would lead to people abandoning their robots. But if there was an annual tax on the value of robots instead of their income, the price of robots would adjust so that the tax paid would be less than 100% of the robots income.

Anyway, this is all a minor side to the argument. Even if you don't follow---if you can just assume for the rest of the discussion that there are some ways to tax most land rent without leading people to abandon their land, that's good enough for the argument.

More important: how much revenue could a land tax raise?

In 1977 Josepth Stiglitz published an interesting result, aptly nick-named the Henry George theorem: basically sensible public investments in public infrastructure raise aggregate land rents by more than the cost of the investment.

In practice we see that land prices rise next to new public infrastructure like a train line etc.

Of course, economic theory doesn't much care about whether the public infrastructure invested is an actual train, or whether it's eg a decrease in income tax: anything that makes an area more desirable to live and do business in will increase land rents there.

In a simple model with perfect labour and capital mobility, all relative difference in quality of infrastructure and institutions between different places will be captured by land rents. (Because we assume that labour and capital move until they get the same wages / return everywhere.)

So thanks to the 'Henry George theorem' in a land value tax system all sensible public infrastructure spending is 'self funding' by increasing land rents that will be taxed away to finance the investment. And all untaxing of labour and capital will be self-funding as well. By removing dead-weight losses this untaxing of labour and capital will even increase the total tax take while being less of a burden on the economy.

A UBI in a certain area will certainly raise land rents in the area: it will be a more desirable place to live so people will bid up rents. If people were perfectly mobile and uniform, the increase in land rents would probably exactly offset the cost of the UBI? In reality probably only citizens will get UBI, and someone who uses less than the average amount of land by value (directly, by living in a small apartment in a high rise, or in the middle of nowhere; indirectly but consuming less products and services that take a lot of high value land to produce) would come out ahead in net terms. And someone who uses more than the average amount of land by value would be a net payer of the UBI.

That's pretty similar to a income tax financed UBI: the average tax payer (by definitional virtue of being average) will pay as much in additional taxes as she gets UBI. People on less income will get a net payment; people with more income will make a net payment.

The difference is that for the income tax financed UBI the implicit means testing is by income (duh..). And for the land value tax financed UBI the implicit means test is by amount of land used by value.

I hope this all makes sense? Any holes left?

Please pardon the length..

Matthias Görgens writes:

Tracy, I wrote a very long comment, but I don't know if it posted. But yes, basically what Simon said.

As for how much revenue a land value tax can raise: given reasonable levels of labour and capital mobility, all differences in infrastructure, institutions and taxation between places go to to the land owners.

The aptly nicknamed Henry George Theorem proved by Joseph Stiglitz talks about conditions under which public infrastructure investments raise aggregate land rents by more than they cost.

Basically any sensible public infrastructure investments will do this. And a land value tax will capture that increase, making the investment self funding.

As an example we can see in practice, a new train line usually raises house prices next to the new stations. A land value tax can capture that increase to finance the train line.

The border area between Germany and Switzerland shows another effect: labour and capital is quite mobile around there, and Switzerland has lower taxes. As a consequence the Swiss land just across the border sells for a considerable higher price.

A land value tax not only captures the benefit of physical public infrastructure improvements, but also of eg income tax decreases.

Hence, we could untax labour and capital completely, and collect all that revenue via land taxes instead. (And perhaps even more, because we would avoid dead weight losses.)

A UBI in an area would also raise land rents, juar like anything else that makes living or doing business in that area more attractive.

If we had perfect mobility for people, the UBI would perhaps be perfectly offset by higher land rents? In practice probably only citizens will get UBI, so land rents will probably increase by a somewhat smaller but still substantial amount.

What that means is that for the average person a UBI will probably almost be eaten up by higher land rent. (Just like a large chunk of the greater warning opportunities for programmers in silicon valley are eaten up by local landlords there.)

A land value tax would allow to capture that UBI driven increase in land rents to partially finance the UBI---perhaps even most of it.

Tracy W writes:

@Simon, thanks for explaining that.

I agree that a high land tax rate would depress land prices, but as the land tax rises above the land rent, I would expect distortions: firstly rising land rents, then, once rents can't increase any more, people abandoning land.

simon writes:

@Tracy: The attraction of land tax is that the supply of land is (mostly) fixed and land tax also doesn't alter the uses to which land can be put. So, it shouldn't* affect land rents.

Also, it seems you didn't understand (or don't believe?) the argument. If land tax is based on land value, then for any given land tax rate, no matter how high, the land value will adjust downward so that the amount of land tax collected remains below land rents.

In practice, I would expect issues to arise with a high land tax due to inaccuracies in land value measurement. In particular, a too-high value assessment could indeed lead to land being abandoned. But in the ideal case, perfect accuracy in valuation, that can't happen.

*considering only the land tax itself and not the spending/tax cuts its introduction enables.

Tracy W writes:

@Simon. The supply of land may be fixed. That doesn't tell us anything about the demand for land.

As for the given land tax rate, that only makes sense if the land tax is set as a percentage of land value. I agree that if the land tax rate is 3000% and that drives land prices down so the land value is only 3 cents per hectare then the land tax in $ terms is only $90 per hectare and that's presumably easily affordable. But that hardly supports the argument that unlimited amounts of money can be raised. 100,000,000% of nothing is nothing.

I suspect Matthias got confused between a land tax as a percentage of the value of land and a land tax in terms of dollars per hectare.

But hey, it's possible I'm missing something. Wouldn't be the first time.

Matthias Görgens writes:

Tracy, Simon,

The good people of EconLog rescued my comments from the automatic filter. Please see atop.

Tracy, I am not confused about tax rate proportional to value vs absolute by area.

Simon, you hint why I think the revenue raised can be very high: lots of things to spend on raise aggregate land rents. Be that infrastructure improvements, or cuts for tax on labour and capital, or even a UBI.

(And without a land value tax that's actually a reasonable decent argument against UBI: any UBI will be eaten by rent increases to a large extent.)

simon writes:
The supply of land may be fixed. That doesn't tell us anything about the demand for land.

The demand for the use of land (as opposed to demand for the ownership of land) should be unaffected by a land tax. The incidence of the tax falls on the owner, but from the perspective of the owner, it's a sunk cost (land tax doesn't depend on whether the land is rented or the owner uses it themself).

It reduces demand for the ownership of land of course, but demand for ownership of land has no effect on demand for the use of land - so long as the supply of land is fixed this should have no effect on land rents.

But that hardly supports the argument that unlimited amounts of money can be raised

Yes, as I said in my first post in this thread derailment.

I suspect Matthias got confused between a land tax as a percentage of the value of land and a land tax in terms of dollars per hectare.

No, he has definitely always been talking about land tax as a percentage of the value of land.


@Matthias

The argument from labour and capital mobility works great for already very dense places like Hong Kong or Singapore. But for low density countries, an enormous amount of immigration could in principle be accepted before land rents rise enough to balance the loss of revenue from the cut to other taxes. In the case of, say, the US trying to replace its income tax, I expect it would take more than the entire world's population.

Strict constraints on development could change that, but if you put in those constraints, that has a cost and you have to factor that cost in. The more the immigration pressure, the higher the cost of the constraints.

Matthias Görgens writes:

Simon, thanks for taking the time to decipher my ramblings. Yes, your interpretation is exactly what I intended to say.

And indeed, I mostly thought about these things from the point of view of cities. Think Hong Kong, Singapore, or even less autonomous areas like London or New York City.

But land rent is already an enormous problem in less urban Silicon Valley. (https://techcrunch.com/2014/04/14/sf-housing/)

And you are also right that public investments in infrastructure make less sense in rural areas. (And making this explicit is probably a feature of the LVT, not a bug.)

We could try to figure out how much increase in taxable land rents we would get in rural states from untaxing of labour and capital (and from introducing a UBI). But I suspect the proper answer lies with federalism anyway: if UBI and taxation issues are decided locally, then NYC doesn't need to care whether the mechanism would work in Idaho. (And it might even make sense to pay more UBI in NYC than in Idaho, because a bigger part of the UBI will need to gowards land rent (apartment rent + indirect cost of living expenses) in NYC than in Idaho. But that's a moral problem for voters to decide.)

(I do suspect that even in rural areas loss of revenue from abolishing income tax will be more than offset by increases in taxable land rent. But I don't want to argue that point right now.)

Simon, another agreement that I kind-of derailed the discussion---but only kind-of. Now that we have the preliminaries out of the way, I can succinctly formulate my actual point that strikes at the heart of Bryan Caplan's article:

Direct means testing is neither desirable nor necessary. And the massive revenue necessary to pay a UBI in rich countries can be raised without distorting the economy.

I'm all for old-school libertarian "keep the government small" austerity that Bryan extols. And for that reason I am in favor of giving out a large proportion of the collected land value tax directly to the people as a UBI---instead of leaving the decision over what to do with the bulk of the money in the hands of politicians.

It is remarkable, that Georgist ideas (plus NGDP targetting / free banking) present an easy synthesis of classic libertarian goals (we still call that 'liberal' where I'm from) and Fabian-Society-style social-democracy. No wonder Marx hated the Georgist's guts.

I should probably just write my own blog post instead of hogging all the space on this thread. Many thanks for your thoughtful remarks, Simon and Tracy!

(I am currently working out how to set up a system that mimics LVT-as-single-tax, if all you managed to do is convince a county to implement your ideas or even only an eccentric billionaire who's agreed to do some real estate investment on your behalf. The main idea revolves around giving people discounts in land rent tax for their income tax paid. This way nobody paying their income tax, but the disincentives are neutralized. (And the land rents in the area should rise to reflect the generous neutralization of income tax offered.))

Matthias Görgens writes:

@Simon

The argument from labour and capital mobility works great for already very dense places like Hong Kong or Singapore. But for low density countries, an enormous amount of immigration could in principle be accepted before land rents rise enough to balance the loss of revenue from the cut to other taxes. In the case of, say, the US trying to replace its income tax, I expect it would take more than the entire world's population.

Have a look at this data from Australia, where a land value tax has been around for ages (alas a toothless one with many exceptions and too low a rate):

http://www.henrygeorge.org/rem42.htm

Private land rent in Australia was around 26% in 2010, and that's before untax labour or capital. As a whole Australia is even more sparsely populated than the USA.

Lauren writes:

Matthias Görgens says:

No wonder Marx hated the Georgist's guts.

I can't speak to Marx's feelings toward Henry George, but Henry George revered Karl Marx. See http://www.georgistjournal.org/2012/09/25/henry-georges-letter-at-the-funeral-of-karl-marx/.

George's Progress and Poverty argues for taxing land not because the tax would be economically efficient but because--so he argues--private land ownership is the single largest deterrent to economic growth. But, he argues in the final chapters, because private land ownership likely cannot be abolished on political or cultural grounds, taxing it to death would be the next best thing to do.

Modern economic Georgists sometimes come from a different perspective, one based on economic efficiency of tax collection, relative inelasticity of land expansion, or perhaps that basing taxes and resultant tax expenditures on local rather than on national, far-reaching uber-directives makes for more control over both taxes and how those tax monies are spent. It is useful to remember that these are modern reinterpretations, and were not at all what Henry George himself argued.

Matthias Görgens writes:

Lauren, thanks for the history lesson! Sorry, I was being sloppy. My comment about the Marxist views are based only on secondary sources like the Criticism section of https://en.m.wikipedia.org/wiki/Georgism#Criticism (since mostly I don't spend much time thinking about Marxism, and this was just a throwaway remark.)

From Wikipedia:

Karl Marx considered the Single Tax platform as a regression from the transition to communism and referred to Georgism as "Capitalism’s last ditch."[93] Marx argued that, "The whole thing is... simply an attempt, decked out with socialism, to save capitalist domination and indeed to establish it afresh on an even wider basis than its present one."[94] Marx also criticized the way land value tax theory emphasizes the value of land, arguing that, "His fundamental dogma is that everything would be all right if ground rent were paid to the state."[94] Fred Harrison replies to these Marxist objections in "Gronlund and other Marxists – Part III: nineteenth-century Americas critics", American Journal of Economics and Sociology.[95]

In any case, I like the modern reinterpretations better---and will refer to them unless otherwise noted. Same with Adam Smith: good ideas, but we've learned more since then.

(And Henry George's ideas on eg the business cycle weren't really worth preserving, far as I know. The market monetarists seem to have a better grasp. But I am only a consumer of blog posts and the occasional paper, not a researcher.)

For another history lesson: it's a bit of a shame Silvio Gesell didn't have more of a lasting influence. He proposed UBI financed by land rent around a hundred years ago. (And a price-level-targeting fiat currency with negative interest on cash.)

Tracy W writes:

@Matthias:
Your earlier claim was, to quote you:

Since we are talking about hypothetical fiscals anyway, replacing most taxes on labour and capital with a land value tax would solve basically all fiscal problems, without causing any dead weight loss on the economy.

I queried whether a land value tax could solve "basically all fiscal problems" (the wording I used was "unlimited amounts of funding", looking back on that, it was wrong of me to reword your comment, I apologise).
You responded with an equation. I presumed that equation was intended to have something to do with your claim that "a land value tax would solve basically all fiscal problems." Some discussion with Simon resulted in me seeing that that the land tax rate in your equation was a tax rate as a percentage of land price.

As you yourself said, in your moderater-rescued comment: " there's a natural limit to how much revenue it can generate: the whole of the land rent---it never goes above 100%."

Your further discussion in your moderater-rescued comment is that the price of land tends to rise with the increase in incomes that can be generated by the land. To put it into my words, a land value tax is an income tax, in that it rises with incomes, but without the deadweight loss. The whole bit about land taxes as a percentage of land prices being able to be raised infinitely was just a distraction. (Out of curiousity, if you were not confused between land taxes as dollars, and land taxes as percentages, why did you even bother putting that in?)

The agreed fact that a land tax is always going to be limited to no more than 100% of the land rent is what makes me doubt your claim that a land tax would solve basically all fiscal problems.

To turn to your other comments about UBI:

A UBI in a certain area will certainly raise land rents in the area: it will be a more desirable place to live so people will bid up rents. If people were perfectly mobile and uniform, the increase in land rents would probably exactly offset the cost of the UBI?

In which case, why bother with a UBI? Why give people $10,000 a year when it's just going to be taken away via them having to pay higher rents?

That's pretty similar to a income tax financed UBI: the average tax payer (by definitional virtue of being average) will pay as much in additional taxes as she gets UBI.

This depends on your definition of average, and indeed of tax payer.

But, leaving that aside, if there are people actually coming out ahead on the deal, ie they receive more money from the UBI than they pay in additional taxes, then, mathematically, there must be some people who pay more money than they receive in the UBI. The above-average taxpayers, the UBI losers.

If some of these losers decide to respond to their higher taxes by dropping their work effort (or emmigrating), then we can have a downward spiral in affordability: to keep paying the UBI we have to raise taxes again (or lower the UBI income), shifting more people into the losers' category, and/or increasing the cost on the people remaining in the losers' category, encouraging some of them to drop out, so rates have to rise/UBI fall, so ....

Matthias Görgens writes:

@Tracy, thanks for engaging.

Let's go one by one. I start with the easiest:

In which case, why bother with a UBI? Why give people $10,000 a year when it's just going to be taken away via them having to pay higher rents?

That's a problem with any UBI.

A land value tax doesn't create land rents, it just taxes them. The observation that a good chunk of UBI will be eaten by higher land rents applies whether land lords get to keep the rent, or have it taxed away.

Introducing LVT and UBI together has the benefit of making the chunk of UBI that inevitably gets gobbled up by higher rent essentially free for the fisc.

Let's forget about my equation---it's redundant if you accept that a LVT will be able to capture essentially all land rent. (And not a single Dollar more than that! as you point out.)

To make my claim of "solving basically all fiscal problems" more explicit and empiric. I am relying on the concept of tax incidence:

  1. With perfect labour and capital mobility, all tax increases/decreases ultimately result in increasing/decreasing land rents
  2. thus any kind of tax revenue available to be generated by any system can also be achieved with an LVT

In practice, we don't have perfect labour and capital mobility---so only an approximation applies. On the flip side, we agree that moving from income tax on labour to a land value tax will remove deadweight losses; and that growth in the economy will also grow the tax base. (And even if the LVT only captures 90% of what the income tax could generate ignoring deadweight losses---the 'lost' 10% would go to workers.)

Thus all fiscal problems become easy to solve:

On the revenue side: no more agony about whether the tax system is efficient versus equitable. Or whether it's progressive enough (or whether flat would be better for the economy). And whether we are past the hump on the laffer curve.

On the spending side decisions become easier to make to. Any public investment project that raises aggregate land rents by more than it costs is automatically worth doing. So no agony about whether we should rather pay for a railroad or lower income taxes instead.

Other investments fall broadly into the following criteria:

  • investments that don't produce enough benefit to bother
  • investments that produce enough benefit, but those go mainly to labour or capital, not increased land rent: let the private sector do them and capture the benefits
  • investments without economic benefit, but voters want them for moral reasons. (Think eg protecting endangered species.) That's just consumption spending. Basically, present the voters with the bill, and let them decide. (Yes, that's a bit messy.)

Yes, rich countries will have more land rent to use than poor countries. And thus they will be able to afford more goodies. An LVT system won't give Greece the ability to spend like Germany.

The land value tax just makes fiscal policy much easier in the same way that nominal GDP level targetting makes monetary policy much easier.

To talk about your affordability spiral example: with the LVT the government won't raise or lower taxes. The idea is to always essentially tax all land rents anyway. And those are set by the markets. No discretion for the government.

If less people would be working thanks to UBI, land rents might fall. (Depends a bit on what people economize on when they decide to forgo labour income. If they decide to make do with a smaller place to live, land rents would fall. If they decided to just get rid of their car, land rents might stay the same---or even increase.) But in any case, the real economy can only consume what the real economy produces.

An LVT doesn't have any deadweight loss. That means, there's no disincentive to work or consume. There's no "If some of these losers decide to respond to their higher taxes by dropping their work effort (or emmigrating),".

The only way LVT plus UBI leads to a change in work effort is if via income effects---ie someone who already makes enough UBI might decide they don't need as much extra money.

(If anything, the part about untaxing labour and capital and capturing the extra gains via LVT will lead people to work harder: for an individually the rent payment is independently fixed of how hard they work, and 0% income tax is a hell-of-a-lot better incentive to put in some more effort than the current ~50%.

Of course, if everybody is working harder, land rents increase---because they are outbidding each other. But that's a 'tragedy of the commons' for you---in this case utilized to improve the economy for a change.)

It might be interesting to set the UBI not at any specific level, but at whatever is left over from the LVT after all other spending. That would give the mass of voters a good direct incentive to watch out for any pork in the political process.

Tracy W writes:

@Matthias: fiscal problems are because we have limited resources and unlimited desires. I don't see any reason that a LVT will make that aspect of human nature go away. You refer to growth in the tax base, but Western countries are already incredibly wealthy by 19th century standards, and yet we still have fiscal problems.

LVT may be very desirable for numerous reasons, but my opposition to a UBI is based on that it either requires cutting the income of the poorest drastically, raising taxes drastically, or, depending on how you define "drastic", both. LVT doesn't change that.

Matthias Görgens writes:

Tracy,

I guess it's a good time to take a break. Our remaining differences are empirical. That is good.

If LVT can't raise enough revenue, I totally agree with your assessment of UBI.

If LVT can raise as much revenue as I hope, without deadweight loss, ie without requiring capital and labour to require paying any taxes or pay more to land lords than they already do: I hope you would agree that a UBI would be feasible, if that's what voters wanted. (There might be better spending opportunities.)

Australia has specific data on land rent available. For them it's about a quarter of GDP that stays with land owners. A quarter of Austealian GDP per capita would finance a decent payout.

Funny enough, the observation that there are big problems despite lots of wealth was core to Henry George's book, so much he put it in the title: Progress and Poverty. "This classic work is an enquiry into the cause of industrial depressions and the persistence of poverty amid advancing wealth." (From goodreads.)

If you want some more argument against UBI, please add to your reportoire that a big chunk of the payments would just go to already rich landowners, of there's no introduction of LVT at the same time.

I am reasonably neutral on UBI, with the exception that I would like to see any means testing by income that anyone might want to do, folded into the assessment of existing income taxes. (If only just to avoid the stupidly high effective marginal tax rates some poor people face.)

Benji writes:

An LVT is merely the way we equally share the scarcity value derived from resources, supplied for free by nature/god.

As rents are best device we have for allocating scarce resources, the LVT which taxes rental values, is optimally fair and efficient.

So a UBI, paid for out of an LVT would be a good thing, in and of itself. It is not a benefit, but a right, without with humans are not equal or free.

Tracy W writes:

@Matthias: if you are proposing that the Australian government eliminate all its spending on health, education, the police, etc, to fund a UBI, then you run into one of my objections: that a UBI means cutting the income (or, to be more precise, the living standards) of the poorest drastically.

Tom G writes:

We should have a voluntary National Service, which offers everybody a Universal Basic Job - UBJ.

Everybody can "do something", and everybody should be doing something. Even if it is to take care of newborns or elderly with Alzheimer's.

Even the disabled can do something, tho it might be limited.

People need work in order to earn for themselves self-respect. Society also needs people to work.

This UBJ should replace all other means-tested gov't social safety nets.

This can be combined with Choose Your Boss and other schemes, as state or local gov't orgs try out different ways to minimize the gov't net cost of paying the almost fixed UBJ cost, while getting the greatest social benefit.

POST A COMMENT




Return to top