David R. Henderson  

Saez You: Income Distribution without Key Components of Income

PRINT
Bartender, Cashier, Cook, Jani... Economics and The Rapture...

When you leave out major sources of income, surprise, surprise, you estimate that income is lower than it really is.

What about the "other 99 percent," whose income supposedly rose by only 0.4 percent from 2009 to 2012? Piketty and Saez compare real incomes at different income levels without including Social Security, unemployment and disability benefits, food stamps, Medicaid, etc. Government transfers totaled $2.3 trillion in 2012, up 24.6 percent in real terms from 2007 and up 68 percent since 2000. Because Piketty and Saez estimate only pre-tax, pre-transfer income, they also ignore $149 billion in Treasury checks to lower-income families from refundable tax credits. They'll also ignore huge Obamacare subsidies next year.

This is from Alan Reynolds, "The Truth About the 1 Percent."

Reynolds cites this recent paper by Emmanuel Saez, an economics professor at UC Berkeley. And, sure enough, Saez confirms that he left out huge components of income of the bottom 99 percent:

We define income as the sum of all income components reported on tax returns (wages and salaries, pensions received, profits from businesses, capital income such as dividends, interest, or rents, and realized capital gains) before individual income taxes. We exclude government transfers such as Social Security retirement benefits or unemployment compensation benefits from our income definition. Non-taxable fringe benefits such as employer provided health insurance is also excluded from our income definition. Therefore, our income measure is defined as cash market income before individual income taxes.

This is ironic in light of the measures that many critics of inequality advocate. As I wrote in 2008:
Interestingly, because of the way incomes are often measured, many of the policies that various redistribution advocates propose would, in fact, increase measured inequality. These policies include: 1) increasing marginal tax rates on high incomes and/or 2) increasing the Earned Income Tax Credit.

The first policy--increasing marginal tax rates--would increase measured inequality as long as the supply curve of high income earning labor is even slightly upward-sloping. [DRH note: I no longer think this is a necessary result; it depends on both labor supply elasticity and labor demand elasticity.] The reason is that an increase in the marginal tax rate would discourage work. This reduction in the supply of labor would drive up the before-tax pay of the highest earners. All other things equal, their after-tax pay would decrease and after-tax inequality would fall--but measured inequality would rise. One can easily imagine advocates of such a policy promoting even higher marginal tax rates on high earners on the grounds of "we didn't increase taxes enough to have an effect." Of course, the ironic result would be a further increase in measured inequality.

The second policy, increasing the Earned Income Tax Credit (EITC) by raising the upper end of the income range over which people qualify, could have a similar effect to increasing marginal tax rates. Those at the new upper end would likely cut their work hours--and, therefore, their pay--to get the marginal subsidy for not working. Assuming this effect were not enough to cause a noticeable effect on wage rates, before-tax incomes of the newly qualifying EITC recipients would fall. Of course, they would be better off in income terms, but if the measure of income does not include the tax credit, measured inequality would increase. Just as in the case of increases in marginal tax rates, one can imagine advocates of an EITC increase calling for further increases on the grounds that the previous increases were not large enough.


Comments and Sharing





COMMENTS (17 to date)
Bostonian writes:

There is a verb missing in the sentence

"I no longer this is a necessary result; it depends on both labor supply elasticity and labor demand elasticity"

Eric writes:

Having attended many seminars in my department on this subject, this seems like the rule rather than the exception when income inequality or income distribution is considered. When income inequality without transfers is the norm, this "problem" can never be solved by government transfer payments. In fact, it makes the "problem" worse as low-income individuals lose incentives to earn non-transfer income, the only kind that counts.

Eric writes:

Oops. Should have read the whole thing. You already made this point in 2008 :-)

David R. Henderson writes:

@Bostonian,
Thanks. It’s fixed.
@Eric,
No problem. BTW, notice the converse also. If we reduced the size of these income transfer programs, Saez’s measure of inequality would fall. So he would favor such a reduction, right? :-)

Eric writes:

@David R. Henderson

I almost mentioned the converse as well in my original comment, but I wasn't 100% confident it was true. I think you're right that elasticities matter but I suspect that for people who charge hourly or by project (as I do sometimes) it would require a higher nominal rate to get me to do it (although it might also create situations where the client and I cannot come to a mutually acceptable agreement). I suspect the first effect would win but I am not sure.

The fix you recommend is so easy and so commonly needed (I wear out my arm in seminars asking if transfer payments and taxes are included) that it feels as though the omission is deliberate, even though I suspect it isn't.

Eric Falkenstein writes:

An acquaintance of mine who works with the poor noted there are a couple of strong incentives in the welfare system that discourage marriage. If so, that's another big driver of inequality that's the result of trying to help the classic 'widows and orphans' group.

Craig Richardson writes:

However, a point must be made- if my health insurance benefit rises by $100 is this synonymous with my pre tax income rising by $100?

No it is not. I am healthy and the rising health insurance costs have more to do with rising obestity of many Americans, huge investments in prolonging the last 6 months of life, and so on. This is not the same as more money in my pocket.

Adding up benefits and dollars earned is not appropriate since they are not fungible.

Moreover, the big point is that even when you add all these benefits in.. somehow counting them the same as dollars, there is very little change in middle class households (who get none of the welfare benefits anyway).

Russ Roberts even points this out in his latest post at Cafe Hayek. He notes that the big Wall St. bailouts have created a huge shift in wealth towards the 1 percent at everyone else's expense.

Since 1973 median incomes have flattened out- Tyler Cowen wrote a book on this, and the numbers should not be denied or somehow whitewashed. Reynolds' work shows very modest upward changes when the benefits are included.

They are troubling to those who believe in upward mobility as a great motivator for entrepreneurship and the dynamic aspect of our economy.

John Voorheis writes:

The last bit should be testable empirically, although identification will be tricky. You can back out pre-tax, pre-transfer income inequality (and, if you want to get a broader measure of income, you can include employer contribution to health insurance, etc.) and post-tax, post-transfer (and, if you wish, post-in-kind-assistance) incomes from the CPS (or, with a great deal more effort, the ACS).

So pick an intervention, and you should be able to see a divergence between pre-tax, pre-transfer inequality and post-tax, post-transfer inequality.

Andrew_FL writes:

@Craig Richardson:

Russ Roberts even points this out in his latest post at Cafe Hayek. He notes that the big Wall St. bailouts have created a huge shift in wealth towards the 1 percent at everyone else's expense.

This is an interesting way of characterizing the statement:

I have written many times that the subsidies to Wall Street and its employees have distorted economic outcomes for a handful of folks relative to the rest of us. So I want to start by pointing out that there is at least one source of inequality that is destructive to democracy and capitalism.

But there are many other sources of inequality that have nothing to do with our economic system or with public policy.

"A handful of individuals" is hardly "a huge shift in wealth towards the 1 percent."

I don't think you read his post very well but were just looking for a gotcha that proves your preconceptions. You certainly didn't read this:

http://invisibleheart.com/HalfFull10.pdf

Which shows that real median income of all families grew 24% from 1972 to 2004, of households with the wife working by 41%, and of single female heads of household by 29%, and, let's see how income was defined:

For each person in the sample 15 years old and over, the CPS asks questions on the amount of money income received in the preceding calendar year from each of the following sources:

1. Earnings

2. Unemployment compensation

3. Workers’ compensation

4. Social security

5. Supplemental security income

6. Public assistance

7. Veterans’ payments

8. Survivor benefits

9. Disability benefits

10. Pension or retirement income

11. Interest

12. Dividends

13. Rents, royalties, and estates and trusts

14. Educational assistance

15. Alimony

16. Child support

17. Financial assistance from outside of the household

18. Other income

From http://www.census.gov/population/www/cps/cpsdef.html

So it includes lots of stuff beyond wages and salaries. But it also doesn’t include benefits—employer contributions to health care or retirement, for example, except as it gets converted into income. And it is also corrected for inflation by an imperfect price index that overstates inflation and understates real growth.

Which establishes that much of what you are saying is just mistaken.

Craig Richardson writes:

Andrew:

You can ask Russ what he meant by a "handful" but I suspect he meant the small fraction (top 1%). I don't have preconceptions- I am looking at the data. In lengthy interviews on EconTalk, Russ has indeed pointed out that no matter how you slice it, all these adjustments don't add up to a lot.

Let's not kick up our heels about median income rising 24% from 1972 to 2004- a 32 year span. So that's let's see, about 0.75% a year, roughly speaking. And you leave off the last 10 years..

Not sure why you are now going to the CPS definition, as David's original post highlighted that employer provided health benefits were not included. I was responding to that.

I am looking at the data, as libertarian Tyler Cowen did in The Great Stagnation, and seeing some troubling trends.

MingoV writes:

@Craig Richardson: "However, a point must be made- if my health insurance benefit rises by $100 is this synonymous with my pre tax income rising by $100?"

Yes. The fact that your benefits don't increase but the insurance premiums cost more is irrelevant. You still got an extra $100 in pretax benefits, and that $100 is added to your total compensation.

Andrew_FL writes:

@Craig Richardson-

I suspect he meant the small fraction (top 1%).

Preposterous. There is no way 3.1 million people would ever be described by anyone as a "handful."

Let's not kick up our heels about median income rising 24% from 1972 to 2004- a 32 year span. So that's let's see, about 0.75% a year, roughly speaking. And you leave off the last 10 years..

You said it was zero. And you still didn't read, because you would know 24% vastly understates the growth due to changes in family structure. The last ten years are left off because that is the data included in Russ's analysis. You can go to his data sources and figure out what has happened since then if you want.

Not sure why you are now going to the CPS definition, as David's original post highlighted that employer provided health benefits were not included. I was responding to that.

Oh gee, I don't know, maybe because you wanted to assert that health benefits shouldn't count as income. My point was even if we grant such a bizarre claim, it just isn't true that there has been no median income growth.

I am looking at the data, as libertarian Tyler Cowen did in The Great Stagnation, and seeing some troubling trends.
I don't have preconceptions- I am looking at the data.

And yet somehow, you ignore all data that contradict what you believe and handwave with bizarre arguments why it shouldn't count against the stagnation thesis.

Craig Richardson writes:

@MingoV

Perhaps on an accounting ledger they are the same, but try taking your pretax health benefits to the movies.

If Obamacare raises your premiums from $500 to $1000, and your salary drops by $500 to pay for it, are you saying you are indifferent between the two?

Really?

Craig Richardson writes:

[Comment removed for rudeness.--Econlib Ed.]

Andrew_FL writes:

@Craig Richardson-

What benefit does it provide to you to know my last name? Do you intend to have my daring to disagree with you have a negative impact on my life in the real world? Does the fact that you don't know it prevent you from forming arguments against me?

Seems to me this is the hostile and inflammatory gesture. Seems to me that calling someone uncivilized for correcting incorrect statements is hostile and inflammatory, too.

If I were really being so horribly rude to you, I'd count on moderation to disallow me to so offend the sensibilities of the community here.

But let me be even more measured in my comments, and clear:

As far as I can tell, your claims are wrong. I presented data that contradicts those claims. You dismissed that data-even though it met the conditions you seemed to indicate were necessary for you to pay attention to that data. I personally find that strange, and I am sorry if you are offended that I find that strange.

Russ Roberts writes:

Craig Richardson,

I have never said on EconTalk that these adjustments amount to very little. I guess it depends on which adjustments you are talking about. I actually think that looking at compensation vs. wages makes a big difference. I also thing that these demographic effects are very large. Work on poverty rates shows enormous impact of family structure changes on measured poverty.

steve writes:

Interesting. Intended or not its a feature. The purpose of excluding transfers from income in the first place is to inflate the number of reported poor. Implementing policies that further inflate the number of poor isn't going to be seen as a problem.

Comments for this entry have been closed
Return to top