Lying with Statistics at the New York Times

Russell Roberts at Cafe Hayek examines how the New York Times presented the "facts" in a recent article on immigration's impact on wages and prices. Russell finds that the NYT's "multimedia" presentation (by which it apparently means: graphs) more or less pulls its techniques from Darrell Huff's classic text How to Lie With Statistics.

Specifically: when the NYT wants to establish the point that a certain set of percentage changes are a Big Deal, they graph the percentages themselves as Big Fat Scary Bars. On the other hand, when it's desired to demonstrate that a different percentage change is Not A Big Deal At All, they graph the percentage as a small chunk of a very big pie. There are other problems with the article as well; Russell concludes: "What a dishonest article. The Times should be ashamed."

I commented similarly (here and here) about an NYT article last month claiming to show that Vermont was in dire demographic shape due to young people moving out. The statistics quoted to support this were uniformly jiggled and presented to paint Vermont as in a uniquely bad situation. For example, the article claimed Vermont was the "second-oldest state;" this turned out only to be true by a single measure: median age of population living in households. By median age of the entire population, Vermont fell to fifth place. And if you (reasonably enough) instead measure the "oldness" of a state by percentage of population over 65, Vermont is (as of the 2000 Census) in 31st place.

The Times also pointed with alarm to the factoid that Vermont's over-65 population was projected to be 30.4% of the adult population in the year 2030. Note they've restricted their sample to make that number higher than it would be otherwise; if they look at the entire population, that percentage shrinks to 24.4%. Also, unlike the previous statistic, the NYT makes no effort to compare this number to other states. Why? Well, because eight other states are projected to have a higher percentage of over-65 adults in 2030. Pointing this out would have been consistent, and honest, but would have weakened the thesis of the article.

All that is prelude to the latest example, today's article from David Cay Johnston, headlined "Big Gain for Rich Seen in Tax Cuts for Investments". The article is apparently authored in "close cooperation" with the reliably lefty think tank Citizens for Tax Justice. (Their two-page PDF press release "Tax Cuts on Capital Gains & Dividends Doubled Bush Income Tax Cuts for the Wealthiest in 2003", also released today, is here; it helpfully advises "For additional coverage of this issue, see the New York Times, April 5, 2006, page A1". Gee, I wonder how they knew that?)

The article is full of Big Scary Facts, presented to make them as Scary as possible. Consider, for example:

Among taxpayers with incomes greater than $10 million, the amount by which their investment tax bill was reduced averaged about $500,000 in 2003, and total tax savings, which included the two Bush tax cuts on compensation, nearly doubled, to slightly more than $1 million.

Gasp! But some things are worth pointing out:

  • The accompanying graphic makes clear that we're talking about (apparently exactly) 6,126 tax returns. It also hypothesizes that these returns saw an average of $1,019,369 "total tax cut". Let's do what the NYT doesn't: multiply these two numbers together. We're talking about $6.2 billion dollars. Compared to (say) the annual deficit, this is relatively small potatoes. Compared to the total amount of money collected by the IRS, it's even smaller potatoes. But of course, since the NYT doesn't even calculate that number, it can't very well compare it to anything or put it in context.

  • Also in hiding is any notion of the actual amount of tax collected from those rich bastards. You can figure it out, though: those 6,126 tax returns each average $5,780,926 sent to the IRS, totalling about $35.4 billion. The great outrage that the NYT/CTJ alliance is trying to hype is (apparently) that this number isn't (35.4+6.2 =) $41.6 billion instead.

  • The average over-$10M filer's Adjusted Gross Income is $25,975,532, working out to an effective average tax rate of about 22%; absent the hypothetical "cut", the rate would bump up to 26% instead. Even if we totally buy the NYT/CTJ numbers, it's obvious that a 26%-vs-22% tax rate for a few thousand tax returns isn't going to send anyone to the class-war barricades. Once again, such a calculation waters down the thesis, so it isn't mentioned.

  • It's doubtful, however, whether even that $6.2B number is actually real: it's based on CTJ's taxation "model". In calculating the "cuts", does the model assume that those 6,126 tax returns would have exposed the same amount of dividend and capital gains income to tax at the higher pre-2003 rates? I don't want to get all supply-side on you, but that seems unlikely; folks making over $10M are notoriously flexible and nimble at arranging their finances to minimize their tax burden. It's not hard to imagine that the reason they're "saving" so much on the tax cuts is that they chose to take dividends and capital gains income in response to the lower rates. There's no indication whether either the CTJ or the NYT did even a wild-assed guess at the magnitude of that effect.

The NYT accompanies the article with a "multimedia" (again, this means "graph") section, of which I've included a snippet. It's ordered by increasing income bracket, and the bottom line refers to those over-$10M filers. [scary bars] Ooh, look, those scary bars! And the big one must mean "the rich" are getting away with something, right?

Well, it's even less meaningful than Russ's example. In this case, it means that almost half the (again, hypothesized) "cuts" for the over-$10M bunch come from cuts in investment tax rates. If the total "cut" came 2/3 from investment rate cuts and 1/3 from "regular" cuts, then they'd draw a +200 bar instead!

It's a calculation, in other words, that gives you a Damn Big Scary Bar, which is why they did it.

Another paragraph of note:

Because of the tax cuts, even the merely rich, making hundreds of thousands of dollars a year, are falling behind the very wealthiest, particularly because another provision, the alternative minimum tax, now costs many of them thousands and even tens of thousands of dollars a year in lost deductions.

First, the use of "falling behind" is just fascinating: a staple phrase of class-warfare rhetoric, but Johnston can't resist employing it even in this wildly inappropriate context. Note that he's referring here to the "merely rich" (making $500K-$10M) paying income taxes at a 25-26% average rate vs. the "very wealthiest" (>$10M) paying at an average 22%. That's "falling behind"! Most everyone in my neck of the woods would be more than willing to "fall behind" while pulling in over $500K per year. And I somehow think that those "merely rich" aren't especially resentful of the "very wealthiest" either; instead they're trying to figure out how to be in that bracket next year.

Second, it's worth pointing out that the Alternative Minimum Tax us a legacy of a previous round of legislating resentment of "the rich" via the tax code. (Jonathan Rauch has a good resentment-free article on it here.)

I could go on; Johnston's article is a target-rich environment. But the point is clear; the New York Times habitually jiggles and cherry-picks numbers to support its arguments, and displays them misleadingly.


Last Modified 2012-10-24 4:17 AM EDT