Friday, April 20, 2012

How to lie with statistics

Along with David Leonhardt's interesting article "Taxmageddon," last weekend's  New York Times Sunday Review included this pair of graphs. These belong high up in the pantheon of "How to lie with statistics" (one of my favorite books) examples. 

These graphs are paired left and right in the original.  (I made them big and split them up so you could see them. They're even clearer on the Times' website)  On the left, is this graph:

Right next to it, is this one

(The graphs had little to do with the article, so I presume they are the work of the Times staff, not Leonhardt.)

It's damning, right? The rich got huge tax cuts (top graph) and so made a ton of money courtesy of the government (+528% change in income, numbers to the right of the first graph). The rich are also feeding at the trough of tax breaks (bottom graph). Outrage!


Now wait a minute here...The top graph is a tax rate, the percentage of income paid, while the bottom graph is total dollars. To say this is comparing apples and oranges is an insult to fruits.

In fact, wealthier people pay nearly all Federal income taxes. So it's not surprising that they benefit more in dollar terms from tax deductions -- except credits, which is money the government pays you even if you pay no taxes. 

If we expressed the bottom graph as a percentage of taxes, or a percentage of income -- the same units as the top graph -- you'd see a dramatic reversal of the implication. Since the lower percentiles have so much less income and pay so much less taxes, the graph would suggest those with less income get the largest (percent) benefits.


The top graph is even more misleading, at least for the Times' goal which is to back a raise in Federal income taxes for the wealthy.

Where does this 60 and 71% tax rate in the 1960s come from? The basic fact of the Federal taxation is that it raises about 20% of GDP despite wild variation in the statutory tax rates. In 1960 Federal tax receipts (NIPA table 3.2) divided by national income (NIPA 1.12) were 93.9/2013.9 = 19.8%. In 2004 this ratio was 2013.9/10534.0 = 19.1%.

Statutory tax rates in the 1960s were as much as 90% marginal rates on the highest incomes. (Remember George Harrison's "Taxman?" "One for you, 19 for me." He wasn't kidding.) But the tax code was so shot full of loopholes that the Federal government didn't collect nearly that fraction of income from anyone.

So where does the 71% come from? At least the Times gives their source, so you can go back and see what the heck the number means. These are estimates by Emmanuel Saez and Thomas Piketty of total Federal taxes -- individual income, corporate income, payroll (Social Security, etc.), and estate taxes -- divided by an estimate of income, which excludes Government transfers.  (The paper is here and a longer working paper version here.)

To what extent is this the statutory rate and to what extent is it actual money paid? I'm still tracking this down, but it appears to be some of each. For example, "We use the TAXSIM calculator developed at the National Bureau of Economic Research ... to compute federal individual income taxes." That seems to imply this is taxes the NBER thinks they should have paid, not what people actually paid.   But they do have individual level IRS data, so in theory know what people actually paid. On the other hand, it doesn't add up: Total tax recepits are 20% of income. So how can everybody's rate come down yet the total rate stay stuck at 20%?  How can the rate of everybody who has any money in 1960 be above 20%, yet the average is still 20%? 

But let's not get in to the depths of the sausage factory, as  it does not matter for the point here. (And my head starts to hurt anytime I delve in to the details of this kind of calculation.) 

The important point, for the Times is that graph has basically nothing to do with Federal income taxes. All of the action comes from Saez and Piketty's assigment of corporate taxes and estate taxes. They assume all corporate taxes are paid by stockholders and bondholders.  This is conceptually right -- it is not true that "corporations" bear any tax burden. Someone is paying, through higher prices, lower salaries, or lower returns to investors. Saez and Piketty assume it's all the latter.

Here is Saez and Piketty's breakdown of how taxes changed between 1960 and 2004 (source):

The actual individual income tax line has not changed much at all, other than to fall slightly for all income groups. Almost all of the Times' fabled taxes the rich were happily paying in 1960 comes from Saez and Piketty's assignment of corporate taxes to wealthy people and their calculations of estate taxes! (Estate taxes are notorious for the games the rich pay to avoid them.)

(It also appears to me that Saez and Piketty are a bit off here: If you charge corporate income tax against the rich, don't you have to divide that tax by an income measure that includes corporate income? In general, you have to divide taxes by pre-tax income not post-tax income. Dividing corporate taxes by individual income, and not including corporate income, can produce "rates" above 100%. On the other hand, if their "income" number attributes all corporate income as individual income to the wealthy, then the distribution of income is grossly overstated. Ok, we're not going in to the sausage factory, maybe for another post someday, but I'm still scratching my head.)   

As Piketty and Saez put the matter:
The larger progressivity in 1960 is not mainly due to the individual income tax. The average individual income tax rate in 1960 reached an average rate of 31 percent at the very top, only slightly above the 25 percent average rate at the very top in 2004. Within the 1960 version of the individual income tax, lower rates on realized capital gains, as well as deductions for interest payments and charitable contributions, reduced dramatically what otherwise looked like an extremely progressive tax schedule, with a top marginal tax rate on individual income of 91 percent.

The greater progressivity of federal taxes in 1960, in contrast to 2004, stems from the corporate income tax and the estate tax. The corporate tax collected about 6.5 percent of total personal income in 1960 and only around 2.5 percent of total income today. Because capital income is very concentrated, it generated a substantial burden on top income groups. The estate tax has also decreased from 0.8 percent of total personal income in 1960 to about 0.35 percent of total income today. As a result, the burden of the estate tax relative to income has declined very sharply since 1960 in the top income groups
Note the tiny percentages of total income involved. These are not going to balance anyone's budget.
Now, when we talk about the "Buffet rule," that is about raising the individual Federal income tax rate. If we calculate Warren Buffet's taxes this way -- including all corporate taxes paid by all Berkshire Hathaway companies (and why not property taxes, business taxes, business contributions to social insurance, and all other business-paid taxes), Buffet's tax rates would be correctly measured, and a lot more than his secretary's tax rate! 

This assignment of corporate taxes takes us into the dark territory of who bears the burden of taxes rather than who actually pays them. Saez and Piketty are assuming that rates of return on investments are reduced because corporations pay taxes, so rich people get less return than they would otherwise. Hence, it's "like" paying more taxes. Ok, that's how economists think about things, but why stop here? Who really bears the burden of the much larger wedge between what employers pay and what employees get? And all the other taxes,  that distorts prices and wages all over the place. And while we're at it, why not "who bears the burden of regulation?" through higher prices or lower returns?

Bottom line: It may be fine for Saez and Piketty's purpose, but I doubt any New York Times reader had the faintest idea they were looking at a graph that primarily said "rich people were hurt by taxes in the 1960s not because they actually paid more taxes but because we assume corporate income taxes drove down the rates of return on their investments!"

And, in case you think this all means  we should go back to the days of "Mad Men" taxation, Saez and Piketty warn:
The surge in top incomes since the 1970s has been driven in large part by a steep increase in the labor income component, due in large part to the explosion of executive compensation. As a result, labor income now represents a substantial fraction of income at the top. This change in composition is important to keep in mind, because the corporate and estate taxes that had such a strong effect on creating progressivity in the 1960s would have relatively little effect on labor income.
In sum, this graph has nothing to do with the main point -- establishing facts about who pays Federal income taxes.  It would be great if our national discussion were to broaden up and consider all taxes -- yes, proper attribution of corporate taxes (all corporate taxes); along with estate, excise, state and local income taxes, sales taxes, property taxes, and so on and so forth. And proper attribution of the burden of taxes. But it isn't.


Now, look at the nefarious pairing of the decline in (statutory) tax rate with the change in income on the right hand side of the top graph. We cut rich people's taxes and look how they got richer!

Here, the Times got too clever by half. The cause and effect insinuation here is actually a supply sider's dream, if you can read and add. The insinuation is, the rich got richer because they got to keep all that income that they're not paying to the government. Even that doesn't add up: a 528% rise is much more than (1-0.34)/(1-0.71) = 2.28 = 128% rise in after-tax income.

But the tabulated rise is in pretax income. (At least the labels are honest.) As tax rates came down, people went out and made an enormous amount more income in the first place.

A 528% increase in income is a lot. 71% x $100 = $71.00.  34% x $100 x (1+5.28) =  $213.52. So, using the New York Times' numbers, we would infer that lowering the tax rate on the top earners corresponded to tripling the tax revenue earned from that group! The rich are, apparently, paying much more in taxes than before.

If you take the Times' numbers seriously, Art Laffer's wildest dreams came true.


Update: Abel Winn notes it's worse than I said:

 The top graph’s y-axis is scaled according to position in the income distribution, while the bottom graph’s y-axis is scaled according to position in the distribution of taxpayers. Since only about half of income earners pay income taxes, being in the top x% of the income distribution means that one is in about the top 2x% of taxpayers. So when we see massive benefits going to the top 20% of taxpayers, that means the tax code was benefiting the top 40% of income earners. But that doesn’t fit very nicely into the 99% rhetoric that we’ve been hearing so much of late, and that the NYT graphs appear to be backing.