August 21, 2007

SOURCE DATA UPDATE POST: New York Times Twists Data to Make Great Personal Income News Appear Awful

Okay, this has been fun so far.

Reader interest dictates that I need to show more detail on where these items came from and demonstrate that they parallel Mr. Johnston’s numbers in his article. So here’s the table at the original post with more intermediate steps shown:


As to where the data came from:

  • Each year’s inflation came from this Bureau of Labor Statistics table using the “Dec.-Dec.” column for each year. It is probably more correct to use the change in each year’s average value, but Dec.-Dec. recreated Johnston’s 2000 figure in his article within $1 (actually 90¢), and alternative measurements didn’t get anywhere near that close.
  • Each year’s revised AGI came from downloading Excel sheets at this IRS stats page. I used the Table 1.1s. 2005 and 2000′s numbers in Johnston’s Times article both tied in. If you download the files, you’ll see what you’re looking for in these cells: 1998, 1999, and 2000 – F12; 2001 – F13; 2002 – F12; 2003, 2004 and 2005 – F10.
  • The rest is math.

Other notes:

  • Near the end of his comment at the original post today, Mr. Johnston claims that 2003′s real income increase is $99, or about 0.2%. I believe the above chart showing a $348 increase amounting to 0.7% is correct. But if I am mistaken, that would appear to make 2004′s real income increase a more-impressive 4.9% instead of 4.4%.
  • In that comment, Johnston also makes a number of statements about changes within income groups that I’m not going to try to get into, partially because I don’t have the time, but more importantly because more current, more relevant, quintile-sorted, and (I believe) more credible and consistent Census Bureau data relating to income for 2006 should be published shortly, possibly even by the end of August (Update — Definitely by the end of August, per this Census Bureau press release).
  • In light of the impending census data, the BizzyBlog commenter who wrote the following criticism of the Times appears to have a valid point — “Considering that the 2006 data has a significant potential to exceed the 2000 data, the window of opportunity to write this article was quickly passing. Got to seize the moment I guess.”
  • Over at Tom Maguire’s place, Johnston writes that “…. nearly half of taxpayers report incomes of less than $30,000 and two-thirds report incomes of less than $50,000. I used these specifics to give perspective.” I believe he is referring to TRUE Adjusted Gross Income (the number reported at the bottom of Page 1 of the long-form 1040). If so he is mixing True AGI apples with Revised AGI oranges (which the IRS strangely calls “Adjusted Gross Income Less Deficit”). Since Revised AGI is really an attempt by the IRS to estimate total taxpayer income, regardless of whether it’s taxed, Johnston’s comment leaves what I believe is a misleading impression that there are millions of relatively low income folks bringing down the average, and not that many really, really rich folks bringing up the average. If you want to see what “Adjusted Gross Income Less Deficit” really means, look at the left half of Page 14 at this 165-page, 1.4mb PDF.
  • This second Johnston comment at Maguire’s place really does make me think that he believes he’s working with True AGI in the article he wrote. That is not the case; anyone looking at the spreadsheets at the IRS page linked above will see that each figure is in a column called “Adjusted Gross Income Less Deficit” (what I called “Revised AGI” above).
  • I do agree with those commenters, like this one, who have said that quoting a source like Citizens for Tax Justice without getting a countervailing opinion from a center-right think tank is weak, especially when CTJ is so obviously playing the partisan card by referring to “trickle down.” Putting the Administration, which necessarily has to be restrained in what it says, against a liberal think tank without any “expert” bolstering the Administration’s position isn’t fair, or balanced. But it IS typical.
  • Finally a mea minima culpa — the original and NB post data for 1998 was changed by $5 from $43,402 to $43,407 because yours truly’s eyes are getting old and tired. The inflated number increased by $6. It didn’t do anything to the percentage change in real income, so there’s no reason to call attention to the tiny changes at the original or NB posts.

1 Comment

  1. can the NYT be trusted to report just the facts ? I think that the media are trying to direct the news not report the news

    Comment by James crawford — August 22, 2007 @ 11:27 am

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.