June 02, 2005
Why Oh Why Can't We Have a Better Press Corps? (Danny Okrent Jumps the Shark Once Again Edition)
Paul Krugman writes:
My thingie and Okrent's reply are up:
Something you can quote me on:
"Okrent is lying to cover his mistake when he accused me of blending data from the household and establishment surveys. He now claims that he was only referring to my estimate of how many payroll jobs the economy needs to add per month [to keep labor market conditions from deteriorating], which for some reason he thinks is based on the household survey.
"But that's not what he said to me: he claimed that the basic numbers I gave on job growth were mix-and-match. In fact, in our correspondence, when I said that it was all payroll data, he declared that "your insistence that you relied only on one set of numbers is very puzzling. I don't see how the math works any other way; maybe you could further enlighten me."
"In other words, he screwed up completely..."
And here, in the extended entry, are Krugman and Okrent, with a few of my annotations:
The New York Times: Public Editor's Web Journal (Forum/Message Board): Daniel Okrent, in his May 22 farewell column as the first public editor of The New York Times, criticized Paul Krugman, an Op-Ed columnist for the newspaper. Prof. Krugman, who disputed the validity of Mr. Okrent's comments in the public editor's regular reader-letters column in The Times on Sunday, elaborated in a longer e-mail message for this Web Journal -- with the understanding that Mr. Okrent's response would be posted simultaneously.
* * *
Krugman Lays Out Why He Believes Okrent Was Wrong
When I asked Daniel Okrent for the specifics behind his final attack, he offered two examples of what he claimed was improper use of numbers. This was the first time I heard from him, or anyone else, about either alleged problem.
Let me start with the example that, I think, sheds most light on what is going on: Mr. Okrent's claim that I engaged in "blending, without explanation, numbers from the household survey and the establishment survey -- apples and oranges -- apparently in order to make a more vivid political point about Bush (5/25/04)."
He's referring to two different surveys conducted by the Bureau of Labor Statistics, which provide alternative estimates of employment. Some people play games by mixing and matching numbers from the two surveys, and Mr. Okrent has apparently spent the past year firmly believing (without having checked with me) that I did the same thing, to score political points. But I didn't. All the numbers in my 5/25/04 column came from the establishment survey.
Moreover, I not only played fair with my readers, I urged them to check the data for themselves. Here's what I wrote in the column:
"Go to the Bureau of Labor Statistics Web site at stats.bls.gov. Click on 'U.S. economy at a glance,' then on the green dinosaur next to 'Change in payroll employment' for a 10-year chart of monthly job gains and losses."
If Mr. Okrent had done that, he would have seen for himself that what I said about job growth was true.
For his other example, Mr. Okrent criticized me for "asserting that the 40 percent unemployed out of work for more than 15 weeks was a 20-year record" (2/10/04, 3/12/04) without acknowledging that the comparison only applies back to the redesign of the CPS questionnaire. See Polivka and Miller, "The CPS After Redesign," on the BLS Web site.
This sounds like another accusation that I blended two sources of data, without telling readers. In fact, all I did was use the Bureau of Labor Statistics data series on long-term unemployment, which is available on the BLS Web site, where there is no indication given to the public of any problem with comparisons between different time periods. Lou Uchitelle did the same thing in an article published in the New York Times business section, "The New Profile of the Long-Term Unemployed", two days after Mr. Okrent's blast. That article made the same point that I did in the columns Mr. Okrent criticized: long-term unemployment is unusually high.
After Mr. Okrent directed me to Polivka and Miller, I checked it out; it's a 1995 research paper which suggested that the 1994 redesign of the Current Population Survey questionnaire might have raised estimates of long-term unemployment. It wasn't an official statement that pre-1994 comparisons are improper, and the BLS didn't consider the questions raised in that paper serious enough to warrant a warning for consumers of its data. Like most such consumers, I don't go hunting for research papers suggesting possible problems with the numbers unless the BLS says there's reason to be concerned otherwise, it would be impossible to get any work done. Let me also say that the issue is pretty trivial: adjusting the data might put long-term unemployment at a 10-year rather than 20-year high, but it's unarguably very high by historical standards.
To summarize: when I asked Mr. Okrent for evidence of my malfeasance, he provided one example in which his description of what I did was simply wrong, and another in which he accused me of pulling a fast one on readers, when all I did was use official data in a standard way.
In correspondence with Mr. Okrent, I pointed out that his specific attacks -- especially the blatantly wrong characterization of my 5/25/04 column -- were unfair. I asked him to do what he would have expected me to do, and admit that he had been in error. He refused.
Let me repeat that Mr. Okrent never raised these issues as public editor. He now says that he didn't because he "experienced your best-defense-is-a-good-offense approach, and found it futile to deal with it."
Maybe a description of some of my experiences with him will give some sample of what he found difficult to deal with.
On 6/8/04, I made a numerical mistake, reading from the wrong line in a table of tax rates during the Reagan years. Although the mistake didn't change the column's conclusions, I reluctantly issued a correction. But I forgot to use the word "correction," which I hear got Mr. Okrent upset.
Mr. Okrent questioned my assertion (10/12/04) that Congressional Budget Office estimates show tax cuts were responsible for two-thirds of the fiscal 2004 deficit. I explained that in each of its budget projections the CBO estimates how much of the change from its previous projection is due to changes in tax law, and that the Center on Budget and Policy Priorities adds these numbers up to calculate the CBO's implied estimate of the overall cost of tax cuts since 2000. I provided Mr. Okrent with the data used for that calculation.
Mr. Okrent challenged my assertion (5/9/05) that the Bush Social Security "progressive indexing" plan would impose its largest percentage reductions in retirement income on middle-income workers.
I explained that the term "retirement income" normally refers to income from all sources, not just Social Security benefits (the Social Security Administration says on its Web site that "you should not count only on Social Security for your retirement income.") I supplied him with a study (pdf) that used Social Security Administration data to show that because high-income workers depend much less than middle-income workers on Social Security, they would have smaller percentage cuts in overall retirement income than middle-income workers. This was similar to a point I made, using different data, a week earlier (5/1/05), so I was surprised that Mr. Okrent even raised the issue.
If Mr. Okrent was unsatisfied with my explanations in these and other cases, it was his right to demand a fuller explanation, and, if he was still unsatisfied, to say something specific in his column.
I hope we aren't going to get into an extended period in which Mr. Okrent, who failed to air his concerns when that was his job, then failed even in private to provide examples that bear any resemblance to what he accused me of doing, keeps throwing out new accusations.
* * *
For a man who makes his living offering strong opinions, Paul Krugman seems peculiarly reluctant to grant the same privilege to others. And for a man who leads with his chin twice a week, he acts awfully surprised when someone takes a pop at it.
Because only a fool or a supply-sider would eagerly engage in a debate on economics with Prof. Krugman, I'll try to eschew argument and stick to facts -- or, at least, the sort of statements that he himself represents as purely factual:
1. I offered him only three examples of "shaping, slicing and selectively citing" (for some reason, he's left one out of his rebuttal) Note: the example Krugman left out is an Okrent complaint that is not about numbers at all--Okrent's complaint Krugman called a study by Jagadeesh Gokhale and Kent Smetters a "Treasury Study" rather than a "study by Treasury Department economists." But the study was much more than a mere academic study by Treasury underling economists. It was a study commissioned by ex-Treasury Secretary Paul O'Neill and reviewed by then-OMB Director Mitch Daniels and ex-NEC head Larry Lindsey--but that's too long to get into a 700 word column
CNN: The Financial Times reported Thursday that the Gokhale-Smetters study was commissioned by Paul O'Neill when he was treasury secretary, and Smetters told the paper that White House advisers Lawrence Lindsey and Mitch Daniels read and were 'very engaged' with it. The Treasury Department Thursday denied having anything to do with the study, which is likely to be published by the AEI in July, and Gokhale said it was meant only to be a 'talk piece.'
because I was at home when he began bombarding me with outraged demands for retraction and apology; I'd completed my tenure as public editor the preceding week, and did not have any files with me. When I had the chance to consult some of my reader mail later in the week, some of his greatest mis-hits immediately came to the fore. I'll get to a few of those in point No. 5, below.
2. This was the first he heard from me on these specific issues partly because I learned early on in this job that Prof. Krugman would likely be more willing to contribute to the Frist for President campaign than to acknowledge the possibility of error. When he says he agreed "reluctantly" to one correction, he gives new meaning to the word "reluctantly"; I can't come up with an adverb sufficient to encompass his general attitude toward substantive criticism. But I laid off for so long because I also believe that columnists are entitled by their mandate to engage in the unfair use of statistics, the misleading representation of opposing positions, and the conscious withholding of contrary data. But because they're entitled doesn't mean I or you have to like it, or think it's good for the newspaper.
3. The mixing of household and establishment numbers in his 5/25/04 column: Missing from the BLS chart he cites is any number that even resembles the 140,000 new jobs each month needed to keep up with the growing population a statistic he cites in the column, and upon which he seems to have based some of his computations. To my knowledge, that number only appeared in the household survey.
Note: to my--certain--knowledge, that number appears in neither the household nor the establishment survey: it's an estimate of the current trend growth rate of payroll employment driven by rising population. I have no idea who could have told Okrent it came from the household survey, or why.
4. The Polivka-Miller paper: On the substance, readers can come to their own conclusions by examining the report themselves, particularly the chart and related narrative addressing Duration of Unemployment on page 23 (pdf).
Note: Polivka and Miller's numbers imply that the 1994 CPS survey redesign raised the reported average duration of unemployment by a week. Unemployment duration is reported at 19.6 weeks today. It averaged 15.4 weeks in the 1984-1993 decade before the survey redesign, and 14.2 weeks in the decade before that. It's not quantitatively important.
On Prof. Krugman's defense of his unfamiliarity with it, he's effectively saying, "If I didn't know about it, it must not be important." This is a polemicist's dodge; no self-respecting journalist would ever make such an argument.
5. Some other examples of Krugmania that popped out of my copious files:
His 1/27/04 assertion that the cost of unemployment insurance "automatically" adds to the federal deficit. This two-fer misrepresents a pair of facts: that unemployment insurance is largely borne by the states, and that major federal contributions to the states come about only because of an act of Congress, which is hardly automatic.
His 2/3/04 assertion that tax proposals offered by Democrats would help the 77 pecent of taxpayers in the 15 percent bracket or less. The most recent generally accepted figures available at the time indicated that the number was actually 64 percent. Note: I believe that 77% of *all* taxpayers are in the 15% bracket or less; 64% of those who pay *income* taxes to the Treasury are in the 15% bracket or less; there are a bunch of people who pay taxes but not income taxes.
A very recent example that nonetheless escaped my memory until Prof. Krugman generously reminded me of it in his letter: His 5/9/05 column on progressive indexing. The column itself (without the ex post facto explanation) suggestively conflates "retirement income" and "social security benefits" without sufficient explanation, but with plenty of apparent point-making.Believe me -- I could go on, as could a number of readers more sophisticated about economic matters than I am. (Among these are several who, like me, generally align themselves politically with Prof. Krugman, but feel he does himself and his cause no good when he heeds the roaring approval of his acolytes and dismisses his critics as ideologically motivated.) But I don't want to engage in an extended debate any more than Prof. Krugman says he does. If he replies to this statement, as I imagine he will, I'll let him have what he always insists on keeping for himself: the last word.
I hate to do this to a decent man like my successor, Barney Calame, but I'm hereby turning the Krugman beat over to him.
Posted by DeLong at June 2, 2005 11:23 AM