The Irrelevance of Universal Basic Income

The Modigliani Miller Theorem says that a firm's financing policy is irrelevant. It's wrong of course, but it's a good place to start thinking about firms' financing policies.

It would be presumptuous to talk about an Irrelevance "Theorem" for Basic Income. The math is trivial, and the economics is obvious. (And I hope this is not at all original.) But it might be an equally good place to start thinking about Basic Income.

There is only one assumption needed for the Irrelevance "Theorem": the only thing that matters to individuals and to the government is Net taxes (taxes minus transfer payments).

Let there be an initial net tax function Ni = N(Income of individual i, other stuff). It is possible to define a new net tax function M( ) - C  = N( ), where C is some constant. Interpret C as "Universal Basic Income". Done. An outside observer, who observed only the net taxes each individual paid, would be unable to distinguish between a system with and without a Universal Basic Income.

If we want to argue for or against UBI, we must argue that the assumption is false. For example:

1. Framing Matters. It makes a difference to individual behaviour whether they frame a (say) 50% marginal net tax rate as a 50% clawback of welfare or as a 50% income tax with no clawback. Since the presumably educated chattering classes often can't see through the framing (or I wouldn't need to write this post), this is not implausible.
2. Taxes are an individual obligation and government right, and transfer payments are an individual right and government obligation. If the default option is that neither government nor individual pays the other anything, it might make a difference who has the obligation and who has the right to do what. Sort of like where the Coase Theorem doesn't work because of "transactions costs".
3. ?

Comments

[Sean Hunt]

I believe you want M() = N() - C, no?

I agree with this post, and I think that subpoint 3 is ease of administration. The costs of administering different schemes vary (although by how much, who knows).

[Alex Godofsky]

Nick I think you are missing several points:

3. Replacing all the various clawbacks with taxes makes it clear just how high the MTRs are on the lowest income brackets. The goal is to fix those extremely high MTRs

4. Using a UBI to replace in-kind welfare payments can create value out of nothing.

[Luis Enrique]

I am writing this without having read your stuff on this properly, so expect missing point or something already written, but: at that margin when somebody living on welfare decides to get a job, the situation (A) under the existing regime and the situation (B) under UBI look very different. And yes, we could design the tax system using negative income taxes or whatever so that when moving from receiving a conventional means-tested (non-universal) welfare benefits to employment it looks like B, but we don't have that system. So it seems to me you are talking at cross purposes somewhat - UBI supporters are saying: we want to move from status quo to UBI, you are saying we could achieve the same thing using existing system but changing tax schedule, without calling it UBI.

[Nick Rowe]

Sean: I never trust my own arithmetic, so neither should anyone else! I confess I had it the other way round originally, but I think I've got it right now.

Alex: Isn't your 3 just a variant of my 1? (One framing may be clearer, for policymakers, than another.)

On your 4: the distinction between $ vs in-kind seems to me to be orthogonal to the tax vs transfer distinction. "We give you food, but you have to give us food back if your income rises".

Luis: "So it seems to me you are talking at cross purposes somewhat - UBI supporters are saying: we want to move from status quo to UBI, you are saying we could achieve the same thing using existing system but changing tax schedule, without calling it UBI."

OK.

[Majromax]

#3) Jurisdiction matters. Taxes are predominantly collected by higher-level governments, but benefits are paid in large part by lower-level governments. This is obviously economically irrelevant, but if something like this comes down to implementation there's going to be arguments over who cuts the cheques.

#4) Program proliferation matters. High clawback rates can be due to individually high rates, or they can be due to the combination of overlapping programs, each with a moderate clawback rate. Even determining the true marginal rate structure at low incomes is a nontrivial task currently, making computing your "net tax function" more of a thought experiment.

#5) Universality matters. Right now, we often choose to target benefits to those society deems most worthy, such as giving families with young children preference for subsidized housing. As an extreme, in some US States it's essentially impossible for a single, non-disabled, working-age male to receive government benefits of any kind. A UBI system would at least smooth these differences out, if not eliminate them entirely.

#5b) Income matters. If we wrap up existing programs into a "net tax function," then we're stuck with some odd taxes. Many cash welfare systems, for example, withhold benefits if a recipient has too many assets, whereas the classical tax system is indifferent to money locked in a safe.

#6) Cash matters. Many government benefits are currently distributed in-kind, such as US food stamps or subsidized housing rent vouchers. UBI eliminates many of these programs in favour of a cash payment. Whether this is good or bad depends on your assumptions about whether the poor as a group are wise (or at least rational) managers of money and whether they are particularly vulnerable to financial exploitation. Currently, many policymakers think of in-kind benefits as a salutary feature of the welfare system, see for example the great effort taken in the US to prevent food stamp recipients from selling on their benefits for cash.

Notwithstanding the above you do make a good point, but I also fear you're tilting at a windmill. Simply looking at the existing social safety net via the "net tax function" is enough of a change of perspective that I think policymakers would balk.

It's a shame, too, because for all the discussion about how terrible high MTRs are for the rich, there's never any corresponding discussion about those effective marginal rates starting at 50% for, say, recipients of Ontario Works.

[Miguel Navascues]

So, you are saying that the only important matters is income net of taxes. Is that right? I agree, but I don't see the need to any arithmetic. That make me dubious about the message. ;-€

[Harold]

Hi Nick, I've been reading you since I was 13!

One thing: Isn't the premise false because it matters who gets the basic income and pays the tax? If everyone gets the basic income and the tax is progressive, then that means net taxes are less for the poor and pretty much the same for the rich. The poor have a higher MPC and if we assume that the savings rate increases with income, they also save more.

[Lord]

I agree with Luis a negative income tax is a UBI but that isn't something we currently have. I would say those who support a UBI should support more progressive income taxes but often they just want less spending on the poor, redistributing what is spent on them upward.

[Ralph Musgrave]

I can’t see much wrong with the Modigliani Miller theory, at least so far as it affects bank capital ratios. Reasons are thus.

Bank shareholders demand a return for two reasons: first the “abstinence from consumption” point which is common to all saver / investors. Second, shareholders demand a return for the risk they take.

As to bank depositors, they demand a return for the above first reason, so there’s no difference between shareholders and depositors there. And as to the risk point, that’s covered by deposit insurance in the case of depositors. But if the risk of the investment (shares or deposits) becoming valueless in any one year is say one in a hundred, then shareholders will charge 1% for the risk. But so too will the deposit insurer!!

Ergo funding banks via equity rather than deposits has no effect on bank funding costs.

I’m sure some clever dick can think of SPECIFIC circumstances where MM does not hold EXACTLY. But basically the MM is clever and more or less valid, far as I can see.

[Alex Godofsky]

Alex: Isn't your 3 just a variant of my 1? (One framing may be clearer, for policymakers, than another.)

Yes, except that the immediate next step that's supposed to be included with a UBI is then actually fixing the ridiculous MTRs.

The pathology that has repeatedly occurred with means-tested welfare systems is that the people who design them do not understand that phaseouts are taxes. They just. don't. get it. So you get rid of the phaseouts... which is definitionally the same as just having a UBI.

On your 4: the distinction between $ vs in-kind seems to me to be orthogonal to the tax vs transfer distinction. "We give you food, but you have to give us food back if your income rises".

Yes, but politically making the case of "replace these in-kind benefits with money" is easier if you can attach it to a second good idea. UBI replacing existing welfare systems solves multiple problems at once.

[Tel]

It is possible to define a new net tax function M( ) - C = N( ), where C is some constant

The first derivative with respect to my leisure time changes significantly. If I earn $10 per hour (after tax) under UBI or earn $20 per hour (same job, smaller tax) without UBI then my decision making on the margin is different.

[Brett Reynolds]

Isn't cashflow another issue, income taxes being typically withheld when earned and refunded at year end? This obvioulsy doesn't have to be the case, and UBI doesn't have to be regular, but those seem to be the factory settings. Those whose income is sporadic will often take on high-interest loans to get through the lean times, and having a regular paycheck would go some ways to innoculating them against the need for such expenses.

[Hugo André]

With UBI, gross taxes have to rise. Simple example w/o UBI: A man earns €1 000 000 a year but because the income tax is 50% he only gets €500 000 after taxes. Simple example w/ UBI: The same man earns €1 000 000 and gets €50 000 in Basic Income. In order to pay for it the government levies a tax of 55% so that the man receives the same income (€500 000) net of tax. Of course, this tax increase isn't restricted to the income tax.

This is a problem because of the distortionary effects of taxation. Moreover, the decreasing nature of the laffers curve is to a large extent caused by people trying harder to avoid taxes. (Yes, in theory high taxes also causes high earners to work less but the evidence for this is fairly weak). Instead of wages they buy things through the company, store their assets in tax havens, make their transactions in a country with a lower sales tax, get a better tax advisor, etc. All of these actions have costs but as gross taxes rise the net benefit of evasive actions goes up. Such actions force tax up even higher and are by themselves also a cost to society. The point is that high earners will always receive their basic income (as long as they remain citizens) whereas taxes can be avoided.

Hopefully I didn't leave anything out that makes my whole argument redundant.

[Avon Barksdale]

Nick,

Please read the logic of MM more closely. MM is about the irrelevance of how the risk in the firm is split because investors can always trade claims. Investors will not pay a premium to have the firm trade for them, whether the investors are heterogeneous or not. The take away message of MM here is that investors do have preferences over cash flows, but not securities – they do not need the firm to organize the match between preferences and cash flows.

Now, can I trade my government transfers? Suppose my net tax function is positive, but I receive all my transfers in the form state provided yoga lessons, but I don't like yoga? Unless there is an exchange for yoga lessons, MM will not apply.

[Nick Rowe]

Hugo Andre: a positive marginal tax rate is not distortionary. A positive marginal *net* tax rate is distortionary. (E.g. a 50% clawback of benefits is as distortionary as a 50% tax.)

Brett: fair point. The exact timing of taxes and transfers may matter, if some people are borrowing-constrained.

Majro: I think your 3 and 4 are closely related, and possibly important. But it's not obvious how UBI eliminates those problems. I've been vaguely thinking about this, but haven't got my thoughts clear yet.

Harold: "Hi Nick, I've been reading you since I was 13!"

That's a terribly misspent youth! ;-)

Avon: I think the cash vs in-kind transfers (or taxes, like the draft) is orthogonal to the BI question.

[Avon Barksdale]

Nick,

The point of UBI is to replace the inefficiencies of the in-kind transfers and the distortions from high marginal tax rates on the poor. If you want to argue against UBI, based on efficient conditioning or tagging, political feasibility, etc., sure, those arguments all make sense and have merit. But MM as an argument against UBI does not. MM is a technical theorem about dividing a pie; that in complete markets, investors only pay attention to real assets and not how they are packaged. Government transfers are far from a complete market. In fact, by design (e.g., health care) the government prevents a market in its transfers. MM does not apply here.

Be careful with saying that the Modigliani Miller Theorem is “wrong”. It's not. It's a theorem, it has a proof. What you mean to say is that in real world situations, some or all of the ifs that make the theorem work don't hold. Sometimes the violation is extreme, sometimes it's very small. Assuming the Earth is a perfect sphere is good approximation in many cases, but if you want GPS precision, you need to worry about oblateness. Same goes for MM – use it to avoid fallacies while organizing your thoughts so that you can get a handle on why it is that financial decisions really do matter.

[Mike Sax]

Off topic, I wonder if you have any thoughts on Brexit? It seems to me there are no real good reasons for it. The UK is thankfully not on the euro so being part of the EU they actually get the best of both worlds.

If Brexit does happen it could even lead to the split up of the UK-as the Scotts are totally against it.

But that's me. Curious about your thoughts.

[Ryan]

I must be missing something obvious, because I'm thoroughly confused.

Take the simple case of a flat tax rate (t) on all income (i): N = t*i. How do I define the new net tax function? Isn't the idea that the new net tax function gives the same result for all i?

[Nick Rowe]

Ryan: here's an equally simple and slightly more realistic example:

Suppose the current system has $10,000 welfare, with a 50% clawback rate. And all income over the basic exemption of $20,000 is taxed at 50%.

The new system has a UBI of $10,000, and a 50% tax on all income.

The new and old systems are identical in terms of the Net Tax schedule.

If N=T-R, observing N does not tell us T or R.

Mike: What are countries/nations, and why do they exist? Economics has little to say on the subject, which is not on our radar. Countries/nations are black boxes. Which is why economists are befuddled by Brexit/Trump/whatever.

[Mike Sax]

See Nick, that sort of befuddlement, that sort of agnosticism by economists I just find fascinating.

To me it's always seemed obvious that politics is the main event-as this is where policy comes from, how it's decided, as interesting as economic theory might be.

But for economists politics, nations, it's all this enigmatic thing.

I always try to imagine what it must be like to see the world that way. But I never can.

[Miguel Madeira]

I think that a "UBI" is nothing more than a tax function [M(i)] with three properties: M(0)<0; dM/di>=0; dM^2/d^2i>=0

- The tax is negative for an income of zero

- The marginal tax rate is always zero or positive

- The 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

[Nathanael]

This piece makes no sense at all.

Miguel Maderia is largely correct about what a UBI is. I would add that a proper UBI guarantees that following the application of taxation, everyone has a "reasonable" (defined however you like) amount of money to live on -- that they have enough to live on and are out of the "under duress" category.

Our current tax system in the US penalizes the poor quite substantially; many are left basically in a "jump! jump for your supper!" situation; and the true marginal tax rate decreases with income in many brackets. So it's not a UBI system.

[Nathanael]

"Suppose the current system has $10,000 welfare, with a 50% clawback rate."
Yeah, so here's the problem: what system are you talking about here?

A UBI system like Alaska's Permanent Fund Dividend?

Because I know of literally *no other system* which hands out $10,000 welfare to everyone, or even $5000. Your "current system" which you describe as "slightly more realistic" is nonexistent. All existing welfare systems are complex hodgepodges of qualifications where you have to "prove your worthiness" and some people don't qualify.

Perhaps you're missing the importance of the word "universal" in "Universal Basic Income"?