Cherry-Picking Income Distribution Statistics

Greg Mankiw is critical of Parade magazine’s “Annual Salary Survey.”? It’s a series of pictures of real people, who also ‘fess up to what they earned last year.? But this being a glossy magazine, they add in a?sample of celebrities.? In consequence:

about 14?percent of the people in Parade‘s sample earn more than $1 million a year.? In the real world, the actual percentage is about 0.2?percent.

But Greg is concerned that this sends the wrong message:

There is a common perception in some circles that we can solve all our fiscal problems if only we were willing to tax the rich some more. Yet, in reality, there are not enough rich for this to work. By presenting such a skewed cross-section of incomes, Parade inadvertently feeds an all-too-common misperception.

Well, yes and no.? If we are interested in thinking about the potential taxes the rich can pay, Mankiw’s 0.2?percent is incredibly misleading.? The issue here isn’t how many?people are rich, but rather how many?dollars are earned by the rich. In the spread, each picture is shown as if each person were equally able to pay more taxes.? But that surely isn’t true.? If Parade were trying to give a sense of the capacity of each person to pay taxes, they would show much, much larger photos of the rich, and proportionately smaller photos of the rest of us.?Relative to Facebook CEO?Mark Zuckerberg, the photos of most of us would barely register a couple of pixels. For obvious reasons, they didn’t choose to do this.? But this inadvertently feeds an all-too-common misperception that because there aren’t a lot of rich people, they don’t earn a large share of total income.? They do.

What proportion of income is earned by the rich?? Let’s turn to?Emmanuel Saez‘s compilation of income tax statistics.? The latest data are for 2007, and for simplicity’s sake we’ll examine the broadest measure of income.? The richest 0.5 percent of families all earned more than $632,000, and received 19.3 percent of all income. Or alternatively, we can focus on the richest 0.1 percent of families-who all earned more than $2 million, and collectively earned an average income of $7.1?million.? This sliver of the community-the folks Greg worries about-received 12.3 percent of all income.

The lesson?? Families earning more than $1?million probably do represent close to 14?percent of total income, and maybe more.? By arguing that only 0.2?percent of families are this rich, Mankiw risks distracting his readers from the fact that increasing the taxes paid by the rich can be a big part of the solution to our fiscal woes.


The Federal Government will spend about 25% of GDP this year. According to your numbers, if we tax the richest 0.5% of the country for all their income we will still have a Federal budget deficit of 6% to say nothing of state and local governments.
Rather than splitting society into distinct groups, one who funds the government and a second who consumes goernment largess, lets make sure everyone pays something. Then we will all be in the same boat and have a stake in the efficient operation of the government.
Your approach of taxing the rich just creates a tyranny of the middle and lower income populations.


Bruce: I don't think anyone is arguing that we should tax 100% of the top 0.5% of the country, or that the bottom 99% should pay no taxes.

Regardless, currently, all American citizens pay taxes on their income; even those who don't pay "income" taxes do pay a higher percentage of their salary on payroll taxes. For example, a self-employed person making $100k pays $12.4k in Social Security taxes, equal to 12.4% of income, compared to a person making $10million who also pays $12.4k, even though as a share of income it is only 0.12%. In that sense, a large part of our tax code, OASI, disproportionately favors the rich.


How, exactly, does "the rich" earn this income?

By draining the productivity of the middle class and NOT paying for it.

Mankiw, as common with all hyperpartisan Republicans, conveniently forgets that America has functioned quite well with high tax rates on the already-super-rich. Go do some research on the history of the top personal tax rates in America.. you'll be surprised just how high it was. Then compare the top tax rate with GDP growth and you'll be even more surprised. The economy actually grows FASTER with higher taxes than when the rich have lower taxes!

You might even conclude that the Republican fantasy-land of "cut taxes and the economy prospers" is patently false.

The only result of cutting taxes on the rich has been.. a wealth transfer to those same people from the middle class who received no such tax cut. But don't bother demanding results from the supply-siders! That would be ridiculous.

Poor statistics, bad science, and elitist arrogance? These are a few of Mankiw's favorite things.



flat tax 15% - first $35k exempted from income.

problem solved. I'm a freakin genius.

Paul '52

When Ayn Rand wrote "Atlas Shrugged" the top 1% of us earned about 8.5% of income. Top rates were 91% on ordinary income, 45% on capital gains.

Now the top 1% of us - 3.1 million people - are in households taking in about 23.5% of income. Top rates are 35% & 15% respectively.

And the bottom 40 % of us - 124 million people - are in households taking in about 17% of income. So they aren't paying Income tax (though they pay many other kinds of taxes).

Please, please, please learn to use some rational perspective. No one's talking about going back to the days of 91% - 45%,

even though the economy grew faster then than now!!!

And if we went back to the day when the top 1% was taking in 8.5%, perhaps it would be easier to look to the bottom 40% for a share.

Every time you hear "the top payers' share of income tax has doubled!" just consider that their share of income has tripled, thus the "burden" has decreased.

Atlas ain't gonna shrug. Not under these conditions.


Joe D

Bruce @1: "Let's make sure everyone pays something."

Payroll taxes are capped. They aren't refundable. They are tremendously regressive. Trust me, the working "poor" (say, the bottom quartile) are definitely paying.

State income taxes are generally flat, sometimes with an exemption, so they're not quite as bad (although my state, FL, tries to make up for its lack of income tax with a sales tax).

For someone who earns $7.1M annually, what's the marginal utility of that last $100K (1.4% of income)? Compare that to the marginal utility of $500 for someone making $35,500, and you see why there must be some method of making income tax progressive.

Blaise Pascal


Nice strawman you've built there; I bet that it's real easy to knock down.

Justin is pointing out that Mankiw's approach projects the appearance that everyone has an equal ability to pay,and thus ignoring the vast inequities of income distribution. Nowhere did he say that the rich should bear the entire burden; rather he said that the rich have a very disproportionate share of the overall income. At most, he is suggesting that the richest 0.5% of families should be paying at least 19.3% of the tax burden.


This is a tangent, but:

I wish that when you filed your income taxes, you got a report back that told you where you fell in the distribution -- are you at the median? Are you in the top 5%?

For extra credit, you could offer comparisons according to the whole country, the state, family size, and maybe gender and age. This would let me discover, e.g., that my household income is in the top 10% of the country, but only the top 20% of my (expensive) home state.

The reason I think this matters is because most of the self-identified middle class (about 95% of Americans) have no idea how their income lines up with their community, and so many below-median people try to maintain an unrealistic standard of living, or they oppose taxes on "the middle class", even though they wouldn't pay a dime.

Also, I suspect that if this were well-constructed and properly publicized, then people would stop tolerating politicians' idiotic claims that adults with no children and an annual income of a quarter million dollars are "middle class families".


Jules Siegel

Taxes aside, in terms of power, what percentage of the nation's assets to the top 0.5% own or control. The very wealthy control much greater assets than they actually own due to many legal techniques for restricting shareholder rights through different kinds of stock, as well as through preferential credit arrangements with lenders.

I've read that figure could go as high as 95%. Any takers on a better estimate?


@ htb - turbo dax does this, if I recall. it gives you a breakdown based on age, general population and maybe state as well.


Love those pretty pictures in parade magazine, but not many Joe Blows making minimum wage in a fast food joint or just above minimum at call centers.


A problem with this type of analysis which is not addressed in Mankiw's post or Wolfer's post here is that tax policy is not starting at zero. Less than 50% of wage earners pay income tax (I understand they pay other taxes, which is beside the point of this argument). As a result, it is misleading to suggest that the rich are less of the population without providing context as to the amount of income they earn proportionately, as Mankiw did. Likewise, is is also misleading to only suggest they earn a disproportionate amount of the income without also indicating they they also pay a dispropotionate amount of the taxes *already*. Clearly that fact limits the upside of financing further spending by taxing the rich, something both Mankiw and Wolfers should highlight.


I think the only way out is to grow the pie. Hauser's Law tells us that tax revenue as a share of GDP remains around 19% regardless of marginal taxes rates for the highest income earners (see chart attached).

If you shrink the pie by making us less competitive in the world (through high marginal taxes, and crippling regulation... etc.) tax receipts will suffer along with the programs they support.

The focus needs to be on shared burdens among all citizens to deliver a functional government (no more no less)... I also think we need to be less fixated on monetary taxes as the only thing a responsible citizen can bring to the table.


Follow up to Jules Siegel:

Generally, the top 0.5% have smart accountants and lawyers that also set up proper asset protection and estate planning vehicles. I would say that ownership in their name, as individuals, is extreemly low compated to contol.


This seems like a chicken-egg argument. Do governments tax the rich more when GDP increases or does GDP increase when governments tax the rich more?


Some of the back of the envelope calculations in the press arguing that higher taxes for the rich can solve fiscal problems assume that behavior remains constant: hours worked don't change, retirement age doesn't change, effort excerted at work doesn't change, entrepreneurial activity doesn't change. It assumes the rich and all potential rich people will just observe the hike in taxes and pay the. In reality, I would expect people to respond to tax changes suggested by reducing many of these growth and wealth generating activities ultimately hurting the US growth prospects and the ability to generate the wealth needed to pay for all these entitlements.


3 thoughts:

- when the top tax rate was 91%, how much energy was expended to engineer wealth into carried value, unrealized gains, non-taxable "perks" or other forms for avoiding income tax?

- Atlas *will* shrug. I make $200k, my wife makes $100k... and we are darn close to letting my wife stay home, firing the nanny and sending the kids to public school.

- Every argument written above is focused on the *collective* interest to grow tax income. What about an *individual's right* to enjoy the profits of their labor? Doesn't that even enter the picture? Indeed, this is the silly little idea that 18th century and 21st century 'tea parties' are about.


Another issue people misunderstand is the way in which our tax system is tiered. If the highest tax bracket is 35%, that does not mean people in that bracket pay 35% of their income to tax. Not at all! Rather, only money above that bracket level is taxed at 35%.

Here is a hypothetical example:

Suppose a tax system with tax brackets at $10,000 intervals which each bracket being taxed 5% more than then previous one.

Someone earning $10,000 would be taxed 5%.
Someone earning $20,000 would be taxed 5% on the first $10,000 and 10% on the second $10,000, for a marginal tax rate of 7.5%.
Someone earning $30,000 would be taxed 5% on the first $10K, 10% on the second $10K, and 15% on the third $10K, for a marginal tax rate of 10%.

As you can see, falling into an upper bracket doesn't mean your actual tax rate is that rate. This is such a basic and fundamental misunderstanding of our tax system that most people don't even know how to properly assess how effective it is. That multi-millionaire still pays no tax on the first few thousand he makes and pays 35% on "only" the top portion of his salary (which still may well be a majority but does not equate to 35% of the entire multi-million dollar salary).



Mankiw makes a bunch of money from textbook sales. Ergo, he's concerned about the rich.

Dan Heath

Your "lesson" is fine as far as it goes, but you obscure Mankiw's larger point.

Using your numbers and assuming 134,000 families, total income in 2007 was about $7.7 trillion and the top .5 percent earned $1.5 trillion.

Suppose the effective tax rate on these individuals is increased 10 percent. Assuming that would not cause them to alter their behavior, that would yield $150 billion in new taxes, a nice piece of change, but hardly enough to constitute "a big part of the solution to our financial woes" since annual budget deficits are projected by the CBO to total about $900 billion a year for the next decade.

Increase their effective tax rate 20 percent, and the tax yield is $300 billion, still only about one-third of the projected annual deficit. A 20 percent increase would mean a 35 to 40 percent marginal tax rate capital gains rate and over 60 percent rate on ordinary income. Add in state and local taxes and in many states the total marginal rate would approach 65 percent.

Can rates be that high without affecting incentives and altering behavior?

You would do well to skip down a couple of posts on Mankiw's blog to "The Enormity of the Fiscal Gap", which points out that marginal rates for families making more than $250,000 a year and individuals making more than $200,000 would have to increase to 77 percent to close the fiscal gap to 3 percent of GDP.

Despite your legerdemain, Mankiw's point holds. We can't solve the budget crisis merely by taxing the rich.