NYT Twists Data: Makes Great Personal Income News Appear Awful (UPDATE: Reporter Responds)

Try to imagine what the New York Times did with the following data:


Go ahead. Then go to the full post to find out.

Here it is (link requires free registration; the Times's figure of $55,714 differs from the table above because of rounding):



The New York Times has to work very hard to make the performance of the economy during the past few years look bad. This morning, David Cay Johnston did his part.

Bear with the technical stuff for a bit; the meat will arrive shortly.

Every year, the IRS publishes information about the tax returns it received in the second preceding year.

2005's tax-return data was released yesterday (data referred to here is not linked because it is in PDFs and Excel files; anyone who wants to see the underlying data can e-mail me). Among the stats the IRS produces is one with the confusing name of "Adjusted Gross Income Less Deficit." I will call it "Revised AGI" for this post. "Revised AGI" adds a long list of items back to taxpayers' reported Adjusted Gross Income (the number at the bottom of Page 1 of the long-form 1040) in an attempt to approximate taxpayers' total income, whether it is taxed or not.

There are problems with the data I will note shortly.

As to Johnston's article, he and the Times chose the specious "2000-2005" reportorial spin noted above, even though:

  • The more important news by far is that the real increase in Revised AGI in both 2004 and 2005 is greater than during either of the final two good years of the Clinton economy;
  • The real decreases in 2001 and 2002 occurred largely because of the bursting of the Clinton-Era dot-com bubble and the September 11 terrorist attacks. The dot-com bubble got mentioned by the Times in the context of quotes from White House sources (the better to make it look like excuse-making); Johnston and the Times made the September 11 attacks invisible.

An obvious omission: The IRS does not include the Earned Income Credit (EITC) in its compilation of Revised AGI, even though the EITC represents real money that either reduces other taxes or goes directly into beneficiaries' pockets. Total EITCs claimed increased from $19.5 billion in 2000 to $42.4 billion in 2005, largely because of the credit's expansion in the Bush tax legislation of 2001 and 2003. Spread over the annual average of roughly 132 million tax returns filed during that period, the $22.9 billion increase in the EITC ($42.4 bil minus $19.5 bil) amounts to over $170 per return.

Don't forget the tax-rate cuts: Now add the benefit of the Bush rate cut in the lowest bracket, which has benefited all but the very top few percent of taxpayers. Since its inception in 2001, that rate cut from 15% to 10% -- never mind the reduction in higher brackets -- has reduced taxes for most single filers by at least $350 per year, and for most joint filers by at least $700.

The EITC expansion, the cut in the lowest rate just noted, and the reductions in the rates applied in higher income brackets, when combined, surely more than make up for the $477 difference ($55,715 minus $55,238) between 2000's and 2005's Revised AGI amounts. So while average pre-tax income may have fallen, average after-tax income has risen -- even during the Times's artificially induced period of analysis.

Johnston's obviously agenda-driven bottom line (see MY RESPONSE below for a revision of this assessment) is this:

The fact that average incomes remained lower in 2005 than five years earlier helps explain why so many Americans report feeling economic stress despite overall growth in the economy.

The growth in real incomes in the two years noted, the continuation of that growth in 2006, and the positive impact of the Bush tax cuts all make Johston's contention about the sources of whatever economic stress may exist absurd.

A more accurate revision to Johnston's claim would be this: "The fact that Old Media won't report the success of the economy since the Bush tax cuts took hold helps explain why so many Americans report feeling economic stress despite overall growth in the economy."

Cross-posted at BizzyBlog.com.

UPDATE: Times reporter David Cay Johnston commented at BizzyBlog. His response is reproduced here:
You make interesting points — and also some unfair ones, especially leaping to the conclusion that I am dishonest.


The idea that in the most scrutinized news report in the world I could twist facts for some venal purpose is laughable. We fire reporters who do that and we should.

I have been active in exposing dishonest reporting since 1973. I am the only reporter whose expose of news manipulations and blackouts lead to a broadcast station being sold to avoid losing their licenses (six stations were sold). I have many other published articles on such issues over the decades in both newspapers and journalism publications.

Readers and reporters can reasonably disagree on what is significant and the choices in the limited space and time I had.

I did not get into this line of work more than 40 years ago to make things up or twist them. If I wanted to I would have become a novelist or a screenwriter.

I got into it to tell people things they did not know and would not know but for my work. That’s the joy of it. And throughout my career I have found things were not at all as I imagined, including in the world of tax that I have covered for more than a dozen years.

Given your jaundice it may surprise you to know that I was the reporter — the only reporter — to do the calculations revealing that the very highest income America got a much bigger tax cut from a law signed by Clinton than they did from Bush, reporting this several times in articles and with detailed graphics.

That is to say, I report the numbers the same way regardless of who is in the White House or any other position of power. My focus is on what happens after the politicians speak and enact laws, not on what they say.

When I write about one-year changes bloggers criticize me for not taking a longer look at the data. In this article I focused on the peak of the previous economic expansion (and compare it to results going back to 1945) and you accuse me of being dishonest for not cherry picking data from the late 90s.

It is reasonable to disagree about whether using the high point of the last economic expansion is the most informing measure. It is certainly common to use this measure. Making that choice is neither dishonest nor absurd.

Behind my article today are extensive spreadsheets I did analyzing the data by income group, components of income (wages, dividends, etc.) and other factors. I then took care to not cherry pick the data, but to select those data points that exemplified what careful checking and cross checking showed. There are problems with using this analysis because of what might be called income bracket creep, which cannot be backed out from the data I relied on.

EITC is not in the IRS data, which you use as a cudgel against my work. This money totals less than a half of one percent of all income (and is not market income). There are other forms of income not counted as well that you do not mention. For example, this administration and has said that there is widespread and significant understating of non-wage incomes at the top, as did the last.

That is to say, there is no perfect measure and my article identified the measure I used — income tax data.

Not in my report today was this finding from my analysis, which in light of your closing comments you may want to ponder:

Among the under $100,000 income group, comparing 2005 to 2004, the number of taxpayers rose 0.5%, total real AGI fell by -$73 billion, or -1.9 percent, and average AGI incomes declined by -$801 from $33,847 to $33,046.

In the year 2000 the average AGI (in 2005$) of those making under $100,000 was $35,286. That means, compared to 2005, that average incomes in this group are down -$2,240 or -6.3 percent from the peak year of the economic expansion.

The figures are not directly comparable because some people moved up and out of that bracket. However, in 2000 this group comprised 90.7% of all taxpayers and it is now 88% of all taxpayers.

But that is also not a large change in share of taxpayers.

I cite this to show the care I take to analyze the data fully before I wrote about it and to make sure anomalies are not treated as substance.

The fact is that average incomes remain below their peak and when you look at groups below $1 million of AGI, which 99.77 percent of all taxpayers in 2005, the declines are clear. It suggests that my reportorial observation about many people feeling economic stress (of which there is plenty of evidence from polls and other sources) is not absurd and does not lack a reasonable basis for mentioning.

The growth in incomes from 2002 to 2003 is statistically insignificant at 0.2 percent or $99, contrary to your assertions about how many years we have had growth in average incomes.

Whatever the data show is what I report. If, for example, the data showed that the incomes of those at the bottom were rising I would report that. Oh, gosh, I did report this morning that the number of Americans reporting very low incomes declined by more than 5 percent. That could have been the lede, but average incomes provides a broader measure of greater interest so I lead with that. Reasonable people can disagree without resorting to attacking the reporter’s integrity.


I, along with other BizzyBlog commenters, salute Mr. Johnston for responding. He also called me and was eager to explain various things to me. I can assure you that the comment above is his (the source IP addess for his comment made me wonder at first). I think the most important point to raise is the time-frame treatment. As one BizzyBlog commenter noted, mutual funds are required to show 1-, 3-, 5- and 10-year return data for a reason, i.e., complete context. Though Johnston could argue that the article’s graph covers the 1-3-5-year issue, I don’t think that cuts it, and I suspect that most readers here don’t either. By not specifically noting the stellar 4%-plus performances in each of the past two years, it’s almost impossible NOT to think there’s an agenda. Adding fuel to the fire, Johnston then doubles down by making the 5-year result the only factor worth mentioning as relevant to how consumers feel now — again, after two really good years of income growth. C’mon, David — How does that get into a non-agendized article? People in this “What have you done for me lately?” world typically aren’t thinking about where they are financially compared to four or five years ago (unless a certain president helps them do it :–>). The other point I want to make is that, contrary to his contention, I was not, and am not, accusing Johnston of dishonesty or lack of integrity. Those terms, or similar ones, do not appear in my post. Johnston says that any time you accuse a reporter of having an agenda (as I most definitely did), you are in effect accusing them of dishonesty and lack of integrity. I don’t agree. Let’s leave it at this — If David didn’t have an agenda when writing the article, I have little doubt that his editors felt like they had died and gone to heaven (or wherever it is Times editors think they’ll go after they die) when his article showed up.
Taxes New York Times Government & Press

Sponsored Links