« Like Apples to Manatees | Main | Smoking: Getting Worse »

August 30, 2006

Why The EITC Works

This'll be a little wonky, but I'll try and make it short. Every once in awhile, you hear conservatives argue that the Earned Income Tax Credit (EITC) can't lift folks out of poverty because it's not counted in the Federal Poverty Rate. Tim Worstall, thinking he had a "gotcha," made this argument in the comments to the "Apples and Manatees" thread:

Please Ezra, learn something about the statistics you are commenting upon. The calculation of the Federal Poverty Rate (that 12.6% that everyone is commenting upon at present) cannot be affected in any way whatsoever by the expansion or even the abolition of the EITC. Because the EITC is not included in the calculation of the Federal Poverty Rate.

Condescension will getcha everywhere, Tim. This argument, made routinely by conservatives seeking to ascribe welfare reform panacea status, is ridiculous. The EITC is a refundable tax credit that's calculated as a percentage of your income. In other words, it makes your job pay more. Which makes low wage jobs more attractive economic prospects. Which spurs more people to take them. Which lowers the poverty rate (surely conservatives agree compensated labor lifts folks out of poverty).

The reason Clinton's expansion of the EITC mattered for the poverty rate, in other words, is that it compelled folks whose available employment options were below their reservation wage to take the job anyway, secure in the knowledge that their EITC credit would make up the difference. Sure enough, research shows that the EITC did indeed boost labor participation, particularly among single mothers. And while the EITC isn't factored into the poverty rate, the salary from the job they took is -- which lowers the poverty rate. But hey, don't take it from me: Ronald Reagan famously called the EITC the "the best anti-poverty, the best pro-family, the best job creation measure to come out of Congress." Speaking of which, it's about time we pumped that sucker up again.

August 30, 2006 | Permalink

Comments

These are the posts that get me to read your blog - debunking.

All hail debunking - it is greatly appreciated.

Chuck

Posted by: Chuck | Aug 30, 2006 10:58:43 AM

Reservation wage, feh. That's lame marginalism. It does not necessarily explain an LFP effect. Reservation earnings is more like it -- an income effect. With the EITC, if you work your income can be as much as $2K or $4K (w/one or two kids, respectively) higher. Note that the reverse of your positive wage effect -- the implicit marginal tax rate for the EITC phase-out range -- does not seem to have much impact on hours of work.

I do think your argument about higher LFP reducing the poverty rate is right.

More important than any of this was the tight labor market, which opened up jobs to those ordinarily written out of the labor market. People don't need incentives to work. If employers are wanting them, they will work. The EITC makes it pay more, which is a good thing.

Posted by: Miracle Max | Aug 30, 2006 11:02:01 AM

The reason Clinton's expansion of the EITC mattered for the poverty rate, in other words, is that it compelled folks whose available employment options were below their reservation wage to take the job anyway, secure in the knowledge that their EITC credit would make up the difference.
The flip side of that, of course, is that employers would be able to push wages below the reservation wage, secure in the knowledge that Uncle Sam would be bribing workers with the EITC credit to take the job.

In my mind, it is far from clear that the EITC is all that and a bag of chips. It introduces inefficiencies and perverse incentives. In the long run, I suspect that it improves the lot of the poor rather less than one would hope given what a big program it is. According to Andrew Leigh of Australian National University, increasing the EITC has the unfortunate consequence of depressing the wage of low skilled workers -- "A 10 percent rise in the generosity of the credit reduces hourly wages for high school dropouts by 4 percent, and reduces wages for those with only a high school diploma by 2 percent. The EITC has no effect on the wages of college graduates. Although the EITC has a much larger effect on the labor force participation of workers with children than those without, the wage effect appears to be similar for workers with and without children. This
suggests that what matters is the average EITC rate in a labor market, not an employee’s own EITC rate. This net downward adjustment in wages indicates that the EITC has less impact on reducing income inequality than if hourly wages were unaffected by the credit."

Posted by: Battlepanda | Aug 30, 2006 11:27:38 AM

Fish. Barrel. Gun. Hilarity ensues.

Nicely done, Ezra.

Posted by: Tom Hilton | Aug 30, 2006 11:50:23 AM

BP does a great job emphasizing the problems of the EITC. I think on the whole the EITC is a solid, expensive costly program. While it does lower wages thanks to the subsidy, on the hole it boosts after-tax income. Think of it as a subsidy to low-income workers and companies that hire them. And of course you have the EITC sharp phase-out which makes it expensive to work past $15K and strong disincentives to marry and combine incomes. Max doesn't believe the phase-out rate has a disincentive effect but I disagree.

But as to Worstall's pt: If you include taxes and the EITC in the poverty rates, child poverty drops pretty dramatically. Just look at the alterntaive Census tabulations that show poverty measures, in particular the child poverty rate, decline dramatically.

Posted by: hederman | Aug 30, 2006 11:53:14 AM

..umm overall you may be right Ezra. But I think that the effect is on allowing people to keep their jobs, not in spurring people take ones they would ordinarily have turned their nose to.

I know of noone thinking of a minimum wage job that sits down and does their taxes to determine if they'll take it. Generally they need a job, they get one, and they sink or swim when tax time comes.

Posted by: david b | Aug 30, 2006 12:13:07 PM

My reading of the evidence is that there is little impact on behavior in the phase-out range. I would add that if there is, so what? We are not talking about people leaving the workforce and becoming bums. We are talking about an incremental adjustment in work hours for families with children who are still working after the reduction. Not necessarily a bad effect.

As to the subsidy effect, my reading of the evidence on that count is that it is not significant either (though probably not zero). I have written about why, but it is too much for this little box.

I will have to investigate Leigh's work, with which I am not familiar.

Is that you, Rea?

Posted by: Miracle Max | Aug 30, 2006 12:13:36 PM

David -- click on the link that says "research shows." There is evidence, as one would expect, of people taking jobs they wouldn't otherwise take because the incentives are better. If you're a single mother, the EITC is really very generous, and a position that once wouldn't have merited time away from your kids may now be enough.

Posted by: Ezra | Aug 30, 2006 12:33:49 PM

I am inclined to be a big fan of the EITC, but Mark Schmitt convinced me to think twice about it.

The Everyone Has Turbo Tax Fallacy

"This complexity also sucks away a huge amount of the value of the tax credits that have become the main way to support low-income working people. The complexity drives people to tax preparation services, which not only charge their own fees (average, $120), but also encourage people to take Refund Anticipation Loans, which as far as usury goes, makes the credit card industry look like UNICEF.

As I was writing this I started to look into Refund Anticipation Loans, which I knew were not a good thing, but I didn't know the half of it. According to the National Consumer Law Center, 12.15 million taxpayers took these 7-14 day loans in 2003, paying effective interest rates that ranged from 70% to over 700%. Half of the people who took these loans were recipients of the Earned Income Tax Credit. Almost 80% had household incomes below $35,000. A majority didn't realize it was a loan. Tax preparation fees combined with RAL fees average about $250. When you consider the big deal that was made about a $600 child credit, you can see how much is lost through this scam. "

Posted by: NickS | Aug 30, 2006 12:36:59 PM

It is Max.

As I recall, Eissa and a few others found a negative impact on the labor for married women in the flat or phase-out range. So you could say that the EITC encourages stay at home moms! I'd rather figure a way to increase the phase-out for married couples so that the steep marriage penalty wouldn't be there and help give married couples the option of working more and staying home. I know extending the phase-out range or increasing the credit is easy answer but you have to figure a) cost and b) how far up the income scale do we want to extend the EITC.

Eissa and Hoynes found very, very slight lower earnings but much higher family income. That's why I think the trade-off is worthwhile.

Posted by: hederman | Aug 30, 2006 12:37:18 PM

But as to Worstall's pt: If you include taxes and the EITC in the poverty rates, child poverty drops pretty dramatically. Just look at the alterntaive Census tabulations that show poverty measures, in particular the child poverty rate, decline dramatically.

I never understood the Republican view on this. If you add in the various government programs that help poor people, suddenly they're not technically poor any more, so those benefits should be taken away?

Seems to me that kinda shows that government programs to help the poor actually work, which is almost exactly what the Republicans want to prove false.

Posted by: paperwight | Aug 30, 2006 1:40:16 PM

Ezra, Ezra,
When we calculate the number of people in poverty (the Federal Poverty Rate, that 12.6% number that everyone is going on about now) we do not include any income that people get from the EITC. (As we do not include food stamps, housing vouchers or Medicaid).

I agree absolutely that the EITC (and the other programs) do indeed lift people out of poverty. I think the EITC is a great program, I’d love to see it expanded (I’m not, you might want to note, a conservative, I’m a liberal, even if of the Classical sort), in the absence of a Citizen’s Basic Income or a pure negative income tax I think it’s the best thing we have to alleviate poverty.

But, and here is my point, my "gotcha" if you will, the EITC does not have an effect on the number described as being below the Federal Poverty Line, because we don’t include it when we calculate that number.

I’m not, in any way, arguing against the EITC. Just about the way in which the number in poverty is calculated without taking account of all the alleviation attempts that are made. As hederman says above, there are better, adjusted, measures of poverty to use and the debate would be moved forward immensely if we actually started using them, not the current flawed one.

As a guide to why it’s so misleading, let’s construct a case. This is purely for illustration purposes and is not meant to reflect actual welfare rates.

Imagine that the fed pov rate was $14k for a three person household (mother, two kids). Before welfare reform, income to that household was 15k. This is made up of cash trasfers (which welfare is and was) and thus this household is not below the poverty rate and is thus not classified as poor.
Welfare reform kicks in. The mother goes back out to work, makes min wage on 2,000 hours a year, $11,500. The EITC tops this up to 15k (please, remember, all these numbers are imaginary).
Income to the household is exactly the same. But that household has now moved from not being in poverty to being in poverty simply because of the very odd way that poverty is defined, including cash transfers to households but not transfers through the tax system (like the EITC).
(Again, just to make it clear, these are very fake numbers purely to illustrate the point).

Your original statement:

"Putting aside the question of whether the decrease in poverty was due to welfare reform or the expansion of the EITC and the MASSIVE economic growth of the late-90's (poverty, by the way, increased every year between 2000 and 2004),"

You are using, as your definition of poverty, the number below the Federal Poverty Line. As above, the EITC is not included in the calculation of this with the extremely perverse effect that we could double, triple, the EITC and make no difference at all to your preferred measure of poverty.

All I want is that poverty is measued post-tax, post-benefits, in the way that other countries do. That way we can work out how much more we need to do.

Posted by: Tim Worstall | Aug 30, 2006 1:50:20 PM

Yeah, I really fail to see how Ezra's post debunks Tim's point.

Posted by: spike | Aug 30, 2006 2:10:17 PM

There are a growing number of programs that provide free tax preparation services to lower-wage workers during the tax season. You can visit the National Community Tax Coalition's website for more info: http://www.tax-coalition.org.

Many of the organizations that offer tax preparation services are also involved in efforts to increase the availability of non-predatory financial services in low-income areas.

Posted by: Erin | Aug 30, 2006 2:50:12 PM

But hederman (comment#1, not #2), the poverty rate also tends to undercount living expenses as well.

How much of the rollback of Bush's tax cuts for the rich would be used up by replacing the phase-out with a cap? That would eliminate the problem of shifting the poverty trap effect higher up the income ladder.

Posted by: BruceMcF | Aug 30, 2006 3:09:52 PM

Posted by: Tim Worstall | Aug 30, 2006 10:50:20 AM ... But, and here is my point, my "gotcha" if you will, the EITC does not have an effect on the number described as being below the Federal Poverty Line, because we don’t include it when we calculate that number.

Uh, yes, it does, if (1) as a result of the EITC there are more people who seek work and (2) there is more work to be had.

So under Clinton, it did, because both (1) and (2) are met. It is arguable that under the present slack labor markets (certainly the slowest rate of job growth during a recovery in my lifetime), under the Bush regime, it may not.

It seems to me that the New Right* critique has a point, if we are foolish enough to allow the New Right run the economy.

(* "Conservative"? talk about needing truth in advertising!)

Posted by: BruceMcF | Aug 30, 2006 3:14:17 PM

The point is that the EITC may reduce poverty, but it doesn't lower the number of people measured to be in poverty, based on the current criteria for measuring poverty. Which was all Tim was saying, and which hasn't been addressed in any response.

Posted by: spike | Aug 30, 2006 3:17:22 PM

It is odd to see Tim, a conservatarian, ignore Ezra's point about a behavioral, "supply-side" impact on LFP, hence pre-tax income, hence the poverty rate, from a change in the returns to labor, on the margin if you like.

There's room for more than one measure of poverty. It would be nice for measures to be consistent. The current measure is not truly pre-fisc because it includes some transfer payments and not others, as Tim says. Even so, if its limits are kept in mind it can still be informative. With no change in transfer payments included in income for the purpose of calculating a poverty rate, a change in the rate -- up or down -- is of interest. Since 2001, for instance, with no evident decrease in transfer payments, the rise in the rate is a legitimate rock to throw at "welfare reform."

Posted by: Miracle Max | Aug 30, 2006 4:31:24 PM

Tim, Tim,

I have to say, I fail to see how I'm being unclear here. The expansion of the EITC is widely ackonwledged to have lowered poverty by pushing folks -- particularly single mothers -- into the workforce. When people have jobs, they're less likely to be impoverished. Even Michael Cannon at Cato admits the point: "The EITC does not directly affect the poverty rate, but it does affect the decision to earn other income that does."

We both agree that the EITC doesn't count in the poverty rates. But that doesn't mean its expansion didn't lift people out of poverty by convincing folks who otherwise wouldn't have entered the workforce to find themselves a job in order to become eligible. Do you deny that?

Posted by: Ezra | Aug 30, 2006 4:31:29 PM

Right, the EITC does not directly affect the poverty rate. My point, exactly. So we should not talk about the effect of the EITC upon the poverty rate.

"But that doesn't mean its expansion didn't lift people out of poverty by convincing folks who otherwise wouldn't have entered the workforce to find themselves a job in order to become eligible. Do you deny that?"

I say that this could in fact lead to an increase in the measured poverty rate. Because people are moving from a form of assistance, cash transfers, which is included in the measurement of the poverty rate, to transfers through the tax system, which is not included in the calculation of the poverty rate.

Note "could" please.

To see what I’m worried about, take a look at sites like Think Progess. Out come the numbers on the federal poverty rate, everyone starts screaming that this shows the current economy is sh*t etc etc. People crawling all over each other to point out that "the number of people living in poverty" hasn’t changed in the third or fourth year of growth.

But, as we both agree, that isn’t true. It’s the number of people who need help to not live in poverty that is static, unless we look at the post-tax, post-benefit numbers, we don’t actually know how many are living in poverty. Is there poverty? Yes. Should it be alleviated? Yes. How do we alleviate it? The EITC. Great, so can we please use a measurement of poverty that includes the EITC so that we can measure how much poverty there is and how much we should expand the EITC?

As for welfare reform, mentioned by Miracle Max. An awful lot of people seem to have forgotten the theoretical underpinning of it. Much of that theory was first laid out by Richard Layard at the LSE (I had it rammed down my throat as an undergraduate) in the 1980s. The aim, the purpose, was to shift the Phillips Curve and thus reduce NAIRU.

Now, whether it was welfare reform alone that did this (I tend to think not alone but wouldn’t want to try and prove this) it is indisputable that NAIRU has indeed fallen: we have strong growth, an unemployment rate that a decade or two ago would have been considered below the sustainable full employment level without wage inflation, and yet wages are not going up (as all too many people are noting).

That was the original aim of welfare reform and it does seem to have worked, doesn’t it?

Posted by: Tim Worstall | Aug 31, 2006 4:06:12 AM

Tim's ignoring all but the immediate effects of EITC. It's true that if, for example, you abolish EITC on Monday, 4 September, the poverty rate on Tuesday 5 September won't change, because EITC isn't counted as income from the point of view of calculating whether someone is in poverty or not.

In this limited respect he is right when he says "You are using, as your definition of poverty, the number below the Federal Poverty Line. As above, the EITC is not included in the calculation of this with the extremely perverse effect that we could double, triple, the EITC and make no difference at all to your preferred measure of poverty." Double the EITC on Monday and, sure enough, on Tuesday the poverty rate won't have shifted.

He's right, but misleading.

As far as I understand Ezra's argument, the abolition of EITC on Monday 4 September would mean, six months later, that fewer people would be in work, and this (because low-paid jobs tend to lift people on to higher-paid jobs and thus out of poverty) would increase the poverty rate in March 2007. People wouldn't be "pushed into the workforce". The EITC isn't just a way of making living below the line less painful; it's also a way of making it easier to get above the line.

Posted by: ajay | Aug 31, 2006 5:58:30 AM

Aargh. Italics off?

Posted by: ajay | Aug 31, 2006 5:59:06 AM


Posted by: spike | Aug 30, 2006 12:17:22 PM The point is that the EITC may reduce poverty, but it doesn't lower the number of people measured to be in poverty, based on the current criteria for measuring poverty. Which was all Tim was saying, and which hasn't been addressed in any response.

No, the EITC can reduce poverty in two different ways, one through the payout to those who are working in any event, second, through an increase in work which leads to more people working their way out of poverty.

Of course if we allow the New Right to run the economy, there will be slack job growth and the real minimum wage will continue to decline ... both of which will substantially weaken the second channel.

Suppose that the poverty level is (in the currency of a hypothetical country), H$100 per week for a family of four. Family of four A is making H$95/wk, and after the EITC is put into place, adds H$10/wk to that.

OK, Familiy of four B was making H$95/wk before the EITC was put into place, and after the EITC is put into place, feel that they can afford to take on more work, and end up with H$105/wk in work and H$10/wk in EITC.

The status of Family A WRT the official definition has not changed, and the status of Family B WRT the official definition has changed. If these are the two most common types of change, the poverty rate will decline.

As with any program to make someone more employable, whether eliminating poverty traps from means-tested entitlements phasing out or in training them to perform a job they are more interested in performing, demand for their work is also necessary for the increased employment. Supply-side labor policies are largely wasted unless there is demand to take up that increased supply.

Posted by: BruceMcF | Aug 31, 2006 8:06:25 AM

This has always been one of my pet peeves. A negative tax....the ability to get more out of the tax system than you put in.

Wouldn't it be better to have the appropriate welfare bureaus that are designed to give away the government's money oversee this?

Posted by: Fred Jones | Sep 1, 2006 5:51:34 PM

The comments to this entry are closed.