Britain's proposed child benefit taxback is inefficient

Today the BBC reports that child benefit is to be axed for higher-rate taxpayers.

Some background: ideally, a person's tax liabilities reflect his or her ability to pay taxes. Most tax systems recognize that having children reduces ability to pay, so provide parents with some kind of tax relief.

In the late 1970s, the British government changed its "child tax allowance" - a tax deduction generally claimed by the higher income parent - into a child benefit, received by the mother. The policy provided greater benefits to lower income parents, who did not earn enough to benefit from the child tax allowance, and put income in the hands of mothers. It made a difference: Shelley Lundberg, Robert Pollak and Terry Wales found that families began spending more money on goods for children after child benefits were introduced.

The latest announcement means that "any couples where one parent earns about £44,000" will have to repay some or all of their child benefit. 

The policy can be criticized on fairness grounds - yes, people earning, say, £60,000, are comfortably off. But within that income group, people supporting children have less ability to pay tax than people without such responsibilities. There is an equity argument for recognizing that lesser ability to pay.

And as Nick Rowe and I argued in "The efficiency case for universality" , taxing back child benefit is inefficient, as well. Here is a non-technical summary of our argument. It is based on the concept of an "effective marginal tax rate":

The first thing to realize is that people care about the total tax/benefit package they receive from the government. That is, they care about their “effective taxes,” not their taxes or benefits alone. If I pay $10,000 in taxes and receive benefits worth $3,000, my effective taxes are $7,000. Both tax policy and social policy involve giving benefits and taking income away. The two are intrinsically linked, and must be considered simultaneously.

...what distorts economic decisions is the marginal tax rate. A person who loses $40 in benefits for every $100 of extra income they earn faces exactly the same disincentives as one who pays an extra $40 in taxes. Both face a 40 per cent effective marginal tax rate. The “clawback” of benefits raises effective marginal tax rates, creates disincentive effects and distorts economic decisions in exactly the same way taxation does.

Recognizing that taxing back benefits is the same as raising tax rates is the key to understanding why benefit clawbacks are inefficient:

We define universality to mean that the same benefits are paid to rich and poor alike, and are not taxed back from the rich. In a tax/benefit system with universality, therefore, individuals with the same market income face the same effective marginal tax rate.

With targeted benefits, on the other hand, people receiving benefits face clawbacks, and therefore pay a higher effective marginal tax rate than people with similar incomes who do not receive benefits. Which tax/benefit system is best: one with universality, or one with targeted benefits?

A government aiming to maximize social welfare must balance efficiency considerations, which argue for low tax rates, with equity considerations, which argue for the higher tax rates needed to finance redistributive social programs. We do not know — we suspect no one knows — what an optimal tax rate schedule would look like. But it seems very unlikely that either efficiency or equity requires radically different tax schedules for Canadians with and without dependent children.

And that is why Britain's proposed child benefit taxback is inefficient. If more revenue is needed - it is predicted that this change will save £1 billion a year -  it could be raised efficiently by increasing taxes on all people earning over £44,000 by a small percentage. The efficiency impacts of such a small change would be negligible. Raising the tax rates on just some of those taxpayers 5 or 10 percent is much more likely to distort people's work choices.

And that is the efficiency case for universality.

Comments

[Nick Rowe]

Well-spotted! So the UK used to have it right, right? You got a fixed amount of child benefit that didn't depend on income and didn't get taxed or otherwise clawed back as income increased? And now they've screwed it up, like we have in Canada.

Hope there are some UK bloggers reading this. Your title should attract their attention.

[Leftoutside]

Hello, a UK blogger here, reading this.

Thanks for writing this, I shall pilfer this and disseminate it on my side of the pond. Today I have started my masters programme so I shall pilfer this excellent discussion a little later rather than immediately.

[Mark Dowling]

When it comes to income traps and targeted benefits I worry far more about non-monetary benefits which are completely withdrawn above a certain income level rather than even some kind of transition benefit in kind sum assigned, which is causing havoc in Ireland among part-time workers whose bosses can't get them to work one minute above 20 hours (not even a dollar amount - time worked!) because at that point they lose their government medical card.

@Nick - I've posted this to the discussion on Irish Economy. A similar child benefit is payable there and given the budgetary situation it's possible there will be a grab for it.

[Guillaume]

... But this is not done for economic reasons, it's all politics. Here's the political case: since they no longer benefit from the tax credit, high income earners will start to resent it. You therefore create a constituency that's opposed to a redistributive measure. Since you set the threshold at a nominal level, a growing number of people fall in that category as time passes by and soon enough you find yourself with a large number of people wanting to abolish that program altogether. That's the past 30 years of politics redux. Expect much more of it in the years to come.

[Frances Woolley]

Leftoutside, Mark Dowling, excellent, pilfer away.

Guillaume. True, so true. Sad. But true.

[Determinant]

I believe Britain has a very flat income tax structure, with only two brackets. The government may not be inclined to introduce another.

Second, general tax increases are political nightmares.

[Nick Rowe]

Determinant: but, if they clawback child benefits from the rich, they have, in effect, just introduced a third tax bracket. Why not, say, introduce a special tax on all left-handed people who earn more than L44k instead? It would make as much sense.

[Neil]

People who live in London and earn £44,000 have less ability to pay tax than someone with the same income living in Lower Penponds. Should they receive a tax benefit for it?

I understand the efficiency issues, since a childcare benefit exists, and presumably a city living allowance does not. But fairness? I'm much less convinced of that.

[Nick Rowe]

Neil: suppose we decide that a family earning $20k with 1 kid needs $5k more after tax income than a family earning $20k with no kids, for fairness. Wouldn't we also say that a family earning $80k with 1 kid needs $5k more than a family earning $80 with no kids, for fairness, too? (Some might argue even more than $5k, but I can't see how it could be $0.)

[Frances Woolley]

Neil, the cost of living issue is a real one.

In Canada, the same tax back rate applies everywhere in the country. But living costs and incomes are considerably higher in big cities such as Vancouver or Toronto than in smaller centers such as Shediac or Moose Jaw. What this means is that, in large cities, child benefits are taxed back from families where both parents are working full-time just trying to scrape enough together to pay the mortgage, whereas a small town family enjoying the same standard of living will keep more of their benefits.

So differences in costs of living (and incomes) decrease the equity of taxbacks even further.

I imagine you're thinking that it would make sense to have a "London allowance" in the tax system - just as I got a London allowance on my scholarship as a grad student? There's some merit in that proposal, especially since it would decrease incentives for people to buy cheaper houses miles out of town and take environmentally damaging commutes. There are some income tax systems that recognize living costs to some extent, e.g. Canada has a special allowance for people who live in the Far North. But I don't know of any large-scale attempts to recognize differences in living costs.

But given that, in most political systems based on the UK parliamentary model, rural ridings have more voters per riding than urban ridings, I don't think "tax exemptions for Londoners" is a platform likely to get adopted by anyone seeking election.

[Neil]

Nick: I'm not convinced that "fairness" is, or should be, a goal of the child tax credit. My guess is the policy goal is to alleviate child poverty. And to buy some votes. A child in an $80k family is undoubtedly worse off than a child in an $85k family. But that $5k isn't the difference between going to school hungry and, well, not. Since there is no poverty to alleviate, the higher income family does not require the benefit.

If fairness is the policy goal, then why is having children a deciding factor? Why is choosing to spend your money by having children worthy of a government handout, while many other lifestyle choices are not? There are so many different situations that taxpayers may find themselves in that make them less able to pay taxes than someone else with a comparable income, creating a policy to deal with each one is unwieldy.

As for vote buying, someone making $80k a year is probably less likely to switch their vote for $5k than someone earning $20k. Vote buying amongst the upper middle class can get expensive.

[Mark Dowling]

@Neil - I'm sure municipalities in rural Ontario would love to have even lower welfare rolls by telling their indigent that they'll get more cash in hand by moving to Toronto...

[Mark Dowling]

@Frances: " in large cities, child benefits are taxed back from families where both parents are working full-time just trying to scrape enough together to pay the mortgage, whereas a small town family enjoying the same standard of living will keep more of their benefits. "

I'm sure Ian Brodie (see post "why we can't have nice things") would have no problem with that - after all, there's no seats to be had rewarding cityfolk.

[Nick Rowe]

Neil: OK, if fairness is not the issue in your eyes, I would have to fall back on the efficiency argument for universality. Efficiency requires the same marginal tax rates across different people (unless you can show that elasticities of labour supply differ).

[Mandos]

... But this is not done for economic reasons, it's all politics. Here's the political case: since they no longer benefit from the tax credit, high income earners will start to resent it. You therefore create a constituency that's opposed to a redistributive measure. Since you set the threshold at a nominal level, a growing number of people fall in that category as time passes by and soon enough you find yourself with a large number of people wanting to abolish that program altogether. That's the past 30 years of politics redux. Expect much more of it in the years to come.

This, by the way, is one of the reasons for cheap or low tuition costs at universities.

[K]

What a good idea.  I live in a big house in a downtown neighbourhood of a big city.  It's *really* expensive.  I want a commensurate tax credit in the name of "fairness." Seriously!?

How about if I don't like it, my government protects my freedom of mobility.

[Andrew F]

Frances: I'd worry about the second order effects of a 'city allowance'. Much like mortgage interest deductibility, I think it would translate into higher home prices and unchanged levels of affordability. It would be a one time transfer of wealth to current real estate owners in cities, much like the milk quota benefits no one presently, but did benefit the farmers when the scheme was first created.

[Andrew F]

Isn't the other argument for subsidizing children that some countries are not reproducing at the replacement rate?

[Frances Woolley]

K: "What a good idea. I live in a big house in a downtown neighbourhood of a big city. It's *really* expensive. I want a commensurate tax credit in the name of "fairness." Seriously!?"

Before I digress: the point of the universality argument - that it makes no sense to arbitrarily impose higher marginal tax rates on some people than others - has nothing to do with where people choose to live.

What the universality argument says is this: *if* there's a tax credit for living in expensive places, then that tax credit should be available to everyone who lives in expensive places. Doing otherwise imposes a higher marginal tax rate on people who live in expensive places for no particular reason.

The digression: There is an efficiency argument for recognizing differences in living costs. The existing Canadian tax/benefit system makes low pay/low cost of living lifestyle choices (growing your own vegetables) more attractive than they would be otherwise, and high pay/high cost of living lifestyle choices (starting a new company) less attractive - even if the high pay/high cost of living lifestyle choice represents a more productive, higher value use of labour.

[Frances Woolley]

Andrew F - let me tell you where I'm coming from with the cost of living thing: a few years ago I crunched some numbers. At that time, an *average income* two parent family in Toronto had an income above the level where they would be eligible for *any* child tax benefit. I see that as a problem. Like you, I don't see the lack of special recognition of the high cost of living in West Vancouver as a problem. And it's beside the point of this post anyways.

Right now the tax system recognizes differential living costs in all sorts of ways - e.g. the non-taxation of per diem payments, Northern residents allowances, etc. And that's fine. What Nick and I argue doesn't make sense is to allow low-income people to have tax free per diem payments (say), but make high income people pay tax on per diem payments, thereby imposing (arbitrarily) a higher tax on people collecting per diems.

Re your point on subsidizing kids - yup. That would tend to argue for child benefits that increase with income, though, wouldn't it, because what high income family would have a child just to collect a few hundred dollars in child benefits?

[K]

Francis:

I completely agree with the universality arguments.  Dead weight losses, efficiency, fairness, etc. 

But on the digression:  you're making an assumption that the current system, outside of taxes/benefits, is neutral. Ie. that there are no structural, unearned benefits such as state protected monopolies (corporate, IP, unions, professional societies, bank charters), overconsumption of negative externalities (natural resources, the environment, land, seignorage, deposit insurance), access to justice, power, etc that accrue to wealthier members of society. The game is played within a particular context of rules that is in no way inherently efficient, and overall the returns to effort appear more than worth it.

As far as incentivizing the production of children... Paying rich people seems like the least efficient possible use of funds...?

[Nick Rowe]

K: Sure. But you need to do more than just say those other distortions exist. You need to argue that they go in a particular direction. If you assumed (say) that an extra dollar of income for a family with kids produced a greater marginal negative externality than for a family without kids, then it would be efficient to impose a higher net marginal tax rate on those with kids. But if you made the opposite assumption, you should have a lower marginal tax rate, and so pay higher child benefits to the rich. A good starting point, or benchmark ass umption, is that all those other distortions are orthoganol to the kids-income parallelogram..

[Tim Worstall]

I think there's a little confusion here. It's not a "taxback". It's that the benefit won't get paid in the first place.

And yes, it is entirely political, but not quite for the reason given above.

A couple of months back they decided to limit housing benefit. Cue howls of outrage and then the limit sunk in. £400 a week. Yes, week, not month. The general comment changed rapidly to "You what? People were getting *more* than that before? The limit now is only going to be £20,000 a year? More than about half the country actually earns?"

So with stopping child benefit for the top 15% of households in the income distribution. "You what? We were taxing dustmen to pay for music lessons for the kids of the rich?"

Have a look at this pasting a Guardian journalist is getting in hte comments.

http://www.guardian.co.uk/commentisfree/2010/oct/04/osborne-child-benfit-war-families

Similarly with the recent announcement that total benefits to any household will be capped at about the median household income. While the chatterati are holding their noses and weeping at the ferocity of it all a good 90% of the country (entirely made up number for rhetorical reasons) is cheering them on.

This may or may not be good economics (and I'm normally in favour of economics over politics) by By Jove, this is incredibly good politics. Start slicing away at the bits of the welfare state that the average man in the street regards as hopelessly over generous.

"You what? They get more than I bloody earn just to spend on rent? Sod 'em" rinse and repeat....

[K]

Nick:  I completely agree.  I was only commenting on [email protected]:54PM [argh! Misspelled your name above.  I'm sorry!] "digression" - that we should recognize that the existing benefit/tax system results in an inefficiency that works against greater income/consumption.  I understood it to mean that she was arguing for a less progressive tax system through cost of living adjustments.  I might not have understood though...

[Tim Worstall]

Apologies, having got myself up to speed on the method they're suggesting of doing it you were right and me wrong. They are proposing to tax it back.

Loons they are for that part of it.

[Frances Woolley]

K: "I understood it to mean that she was arguing for a less progressive tax system through cost of living adjustments."

Overall, I tend to lean towards greater progressivity. Within an overall progressive tax structure, there may be good reasons to recognize differences in people's ability to pay taxes due to differences in personal circumstances, e.g. the costs of living with a disability, medical expenses, the costs of caring for elderly relatives, the costs of being a student, or - yes- the costs of living someplace particularly expensive - ever tried to get a glass of water in Iqaluit?

Tim: for me, this is just deja vu all over again. The child benefit is part of the tax system, it's a recognition of differences in ability to pay taxes. It's paid out in cash to mothers because that way it's more likely to get spent on children. Get rid of it, and you have the problems described in my post above.

Why can't otherwise intelligent people - like those you describe in your post, and I know you're right in your appraisal of the issue because we've been through this same debate in Canada - see this?

Though as a political observer, I'm actually surprised that Britain has retained the universal benefit as long as it has - I suspect this may be due to the lingering influences of Fabian thought, and the intellectual legacy of the family allowance movement, led by people like Eleanor Rathbone.

[K]

Frances:  I agree with almost everything you said.  In particular, I have no issue with subsidizing people who live in Iqaluit, *if* we consider it especially beneficial that some people live in Iqaluit (e.g. they stake a claim to our coast line).  But I don't see any justification for a subsidy just because it's expensive to live there or in West Van - that is just a regressive form of taxation.

[Neil]

Nick:You said "if fairness is not the issue in your eyes, I would have to fall back on the efficiency argument for universality. Efficiency requires the same marginal tax rates across different people (unless you can show that elasticities of labour supply differ)."

Which is reasonable. But you are a supporter of giving money to poor people, right? How do you give money to poor people without at some point having to claw back those handouts, creating a higher marginal rate? Efficiency is a laudable goal, but is it practical in all cases?

[K]

Neil:  Just give it to everybody.

[Larry Willmore]

Neil, Elaborating on K's comment, universal benefits solves this problem. We do not 'claw back' handouts of free schooling and free medical care when these are given to everyone, and financed from general taxation. The same rule applies to cash transfers, such as child benefits and old age pensions: with universality, there is no claw back.

[Nick Rowe]

Neil: Here's another way of looking at it: yes, you have to "clawback" the transfers you give to poor people, by having positive marginal tax rates, but the marginal clawback rate should be independent of whether or not you have kids.

[Larry Willmore]

Nick: That is a nice way of putting it.

[Nick Rowe]

Thanks Larry!