August 10, 2004
Quality of Jobs
Greg Mankiw (or someone good enough to spoof an originating address from inside the Executive Office of the President's class-C network: 198.137.240.x) emails: "Since I know you aim for honest, unbiased commentary on your blog, I know you will be interested in this new study from factcheck.org":
Kerry's Dubious Economics : He says new jobs are paying $9,000 less than the old ones. That's not a fact.
In his acceptance speech at the Democratic National Convention July 29 Kerry repeated a claim that the economy is creating jobs that pay $9,000 a year less than those they replace. He bases that on disputed analysis from a liberal think tank. In fact, economists disagree.... Even some Democratic economists say the economic numbers are simply too rough and contradictory.... And when Kerry said the "middle class is shrinking," he was referring to what happened in the recession of 2001 and the initially slow recovery of 2002. But the economy has picked up considerably in the 19 months since, so what was true then may be untrue when phrased in present tense....
Kerry bases his claim on an analysis of Bureau of Labor Statistics data by the Economic Policy Institute. But the EPI figures don't support what Kerry said, because they don't actually compare new jobs and old jobs -- only broad averages for entire industries. And as we reported July 9, other BLS numbers that compare occupation groups within industries tell a completely different story -- showing higher-paying groups growing faster than lower-paying groups....
I had not previously been impressed by factcheck.org: an organization that claims that "...Bush had plenty of reason to believe what he said..." about Iraq seeking uranium in the face of George Tenet's declaration that "...these 16 words should never have been included in the text written for the President..." has lots and lots of explaining to do. If Bush had plenty of reason to believe them, why shouldn't they have been included? If they should never have been included, then by what process could George W. Bush possibly have plenty of reason to believe them, and why have those in the loop on this process not be fired?
That said, factcheck.org is correct: Kerry, relying on the Economic Policy Institute, should be saying not that "new jobs are paying $9,000 a year less than old ones" but that "jobs in the industries that are growing pay an average of $35,410 a year, while jobs in the industries that are shrinking pay $44,570 a year." In that, factcheck.org is correct. The EPI's numbers are consistent with new jobs in growing industries paying more than the industry average (and also with new jobs in growing industries paying less than the industry average).
That's why it would be better to stick to the aggregate average earnings numbers--they are of much higher quality, and are also more relevant: they take account of what is happening to the quality of currently-existing jobs as well as of the quality of newly-created jobs, and what is happening to the quality of currently-existing jobs overwhelms the effect of newly-created jobs as far as the well-being of the workforce is concerned.
But if factcheck.org really wants to assert that:
A new set of numbers from the Bureau of Labor Statistics actually shows solid growth in employment in relatively higher -paying occupations including construction workers, health-care professionals, business managers, and teachers, and virtually no growth at all in relatively lower-paying occupations including office clerks and assembly-line workers. It's the most detailed breakdown yet -- looking at 154 different job and industry groupings. These statistics are a FactCheck.org exclusive -- supplied to us by BLS at our request and not previously published.
you should not believe them. Their numbers appear to be very sensitively dependent on classifying "Construction and Extraction Occupations in the Construction Industry"--a job/industry classification that pays higher than the median but lower than the mean--as a "high wage" segment of the labor force.
Looking at the data is almost always instructive:
The point farthest to the right on the scatter is "Service Occupations in the Leisure and Hospitality Industries" (a big minus on the wage scale); the point second farthest to the right is "Construction and Extractive Occupations in the Construction Industry" (neutral on the wage scale); the third, fourth, and fifth points to the right are "Service Occupations in Professional and Business Service Industries", "Professional Occupations in the Health and Education Industries", and "Management Occupations in Financial Activities Industries", respectfully (which are a minus, a plus, and a big plus on the wage scale). The point farthest to the left on the scatter is "Production Occupations in Manufacturing Industries" (neutral on the wage scale); the point second farthest to the left is "Office and Administrative Occupations in Manufacturing Industries"; third, fourth, and fifth are "Professional Occupations in Professional and Business Service Industries", "Office and Administrative Occupations in Professional and Business Service Industries", and "Office and Administrative Occupations in Financial Activities Industries", respectfully (a big plus, neutral, and neutral on the wage scale).
This is not the picture I expected to see upon reading factcheck.org's description of what their data showed.
Treat declarations by Annenberg's factcheck.org with suspicion.
Posted by DeLong at August 10, 2004 01:45 PM | TrackBack | | Other weblogs commenting on this postLeaving aside the actual data and how it can be spun, why is Mankiw pointing to an outside group for analysis? Shouldn't the CEA chair have access to lots of qualified researchers to do this kind of analysis? If not, why, and why does Mankiw put up with it? Has he been cast adrift, or does he not like the in-house numbers?
Posted by: Tom DC/VA on August 10, 2004 02:33 PMYou say that Kerry should be saying "jobs in the industries that are growing pay an average of $35,410 a year, while jobs in the industries that are shrinking pay $44,570 a year" instead of "new jobs are paying $9,000 a year less than old ones" but I guess I don't really see that much difference between the two statements. If jobs in industries that are shrinking pay around $9000 a year more than jobs in industries that are growing, then aren't the old jobs (jobs in industries that are shrinking) being replaced by new jobs (in industries that are growing) that pay $9000 a year less? It's not like people who lose a job in an industry that is shrinking don't get jobs in industries that are growing.
Posted by: Shane Wealti on August 10, 2004 02:57 PM"Treat declarations by Annenberg's factcheck.org with suspicion."
I'm pleased to be in 100% agreement with Professor DeLong. This is an absolutely atrocious example of erroneous claims and flawed logic:
http://www.factcheck.org/article.aspx?docID=231
Posted by: Patrick R. Sullivan on August 10, 2004 03:03 PMNotice: please ignore Pat Sullivan's obvious attempt to drift away from the economy (which happens to be the topic of this post... sorry if that's not what your party wants to play up right now. but what is it that it wants to play up?) We shouldn't reward disrespect for this blog, its writer and readers with any kind of interest, even negative. Period.
Posted by: Jean-Philippe Stijns on August 10, 2004 03:42 PMhttp://www.prospect.org/web/page.ww?section=root&name=ViewWeb&articleId=8316
The Fog Machine
By Jared Bernstein
In fact, fewer than 100 days before the presidential election, unemployment is stuck where it was when the recovery began two-and-a-half years ago. Real wages are down over the past few months. And many who have found new employers after losing their jobs during the recession or its jobless recovery are earning less than they used to....
Posted by: Anne on August 10, 2004 03:42 PM"virtually no growth at all in relatively lower-paying occupations including office clerks and assembly-line workers"
So how did they ignore the occupations in the bottom-right quadrant of that wonderfully done plot?
Maybe its semantics. "virtually no growth" LOL
Perhaps in "virtual" reality.
I must say I think Brad's being a bit harsh on factcheck - they're a bit contrarian-for-the-sake-of-it and such, but I'll defend them on both the charges here. Brad more or less admits that the "construction workers" thing is a fine point of pedantry; always fun to see a pedant hoist by their own petard, but nothing that makes you think they're arguing in bad faith. And on the Bush thing, I don't like this argument:
"If Bush had plenty of reason to believe them, why shouldn't they have been included? If they should never have been included, then by what process could George W. Bush possibly have plenty of reason to believe them, and why have those in the loop on this process not be fired? "
...because it looks too much like a short way with the actual defence of the Bush/Blair axis; that they had been told that the intelligence was solid (which would be a "good reason" to believe it) but that the spooks were dangerously, culpably and horrendously incompetent (which would mean that they "shouldn't have" been told).
I happen to regard this putative defence as laughable, but again it could be offered in good faith.
Posted by: dsquared on August 10, 2004 04:10 PMHa! Let's get you a job in a service occupation in the leisure and hospitality industries, Mr. Davies!
Posted by: Brad DeLong on August 10, 2004 05:30 PMWould the host or anyone else say what is wrong this this evaluation by S Wealti:
You say that Kerry should be saying "jobs in the industries that are growing pay an average of $35,410 a year, while jobs in the industries that are shrinking pay $44,570 a year" instead of "new jobs are paying $9,000 a year less than old ones" but ***I guess I don't really see that much difference between the two statements.***
(emphasis added)
Kerry could say the exactly correct thing and it seems to me the audience would have exactly the same reaction. Would they be making some kind of statistical or conceptual mistake?
But Prof D's remark makes me think that he may be in the leisure and hospitality industry himself. Or is the blog supposed to be in the education or journalism industry? Or all three. Is there a "blog owner" industry category yet, and if so, how are those folks doing?
Posted by: jml on August 10, 2004 05:42 PMUmm, well, I have been reminded that they *would* be making a conceptual mistake, or at least they would be making an assumption. It requires the assumption that the average new hire (or new fires) receives the average wage of the respective industry.
But my point is... how would you calculate the probabilities for your future if you also had lots of anecdotal evidence that the age distribution of the new hires was shifted up because many of them are older people who had been laid off? And were excepting jobs in the lower of the lower paid growing industries.
The anecdotal evidence may be biased, but we are not sure about that, but it is evidence to use, even if poor quality.
I can't find the age distribution of both displaced and new hires on the BLS internet site, but it must be there somewhere.
Posted by: jml on August 10, 2004 07:11 PMaccepting
However, I have read that there is evidence that one reason the unemployment rates is so low is that many young people are staying out of the labor force, and many older layoffs are leaving the labor force through early retirement or disability. So "excepting jobs" may not be too wrong. So that is what I was thinking when I typed that... yeah, that's the ticket.
But I meant "accepting jobs."
Posted by: jml on August 10, 2004 07:22 PMdsquared wrote, "I must say I think Brad's being a bit harsh on factcheck - they're a bit contrarian-for-the-sake-of-it and such, but I'll defend them on both the charges here."
Not really. I looked at them about a year ago. Some of their stuff was pretty good; some was awful. Overall, pathetic for a website with that title.
------------
From http://www.dailyhowler.com/dh020504.shtml :
"Meanwhile, partisans link to the stories they like. On Tuesday, Eric Alterman said that ABC was “lying for Bush” and summarized Robinson—thereby offering an account of the facts which strongly cut against Bush. Meanwhile, Andrew Sullivan—saying there’s “not much there”—linked to an absurdly Bush-friendly account produced by Brooks Jackson at factcheck.org, a brand-new Annenberg operation. How absurd is Jackson’s account? He lavishes time on those two Bush friends—friends who recall Bush saying he was pulling Guard duty. But incredibly, other evidence sleeps with the fishes. Incredibly, Jackson doesn’t mention the May 1973 report in which Bush’s superiors at Ellington air base said he’d been absent for the whole year. It’s simply astounding that this report was produced by a man who was, until recently, CNN’s ballyhooed fact-checking star. Annenberg should be embarrassed to have this laughable report on its site."
On USENET, Patrick R. Sullivan often goes by "susupply". Here's an amusing collection of ripostes to Patrick's assertions:
http://tinyurl.com/5tcrv
Professor-
This may seem a bit prissy, but when Greg Mankiw writes: "Since I know you aim for honest...." would it be allowable to edit his statement for correctness so that it read: "Because I know...."?
On the graph: make that a jpeg image - for a 1.1MB file it is quite crappy. (Unless you really don't want people on dialup reading you.)
Posted by: chris bond on August 10, 2004 09:18 PMQuoting your article: "That's why it would be better to stick to the aggregate average earnings numbers--they are of much higher quality, and are also more relevant..."
Actually, MEDIAN earnings are much more relevent to the average guy than AVERAGE earnings, the average numbers are skewed by the few earners with huge incomes. (If Bill Gates walks into a bar, the average income of a wage earner in the bar suddenly becomes both huge and meaningless.)
Jim
As long as Mankiw is pontificating on the subject of "honest, unbiased commentary," you might email him back and ask him why he's STILL trying to flog the job creation numbers in last month's household survey - even though every reputable economist (which, needless to say, doesn't include the loons at The National Review) knows the household survey isn't designed to measure job growth, and is extremely vulnerable to errors in the Census Bureau population growth data.
I mean, as long as he's being such a putz and all...
Posted by: Billmon on August 10, 2004 11:42 PMKerry's numbers are based on an EPI study. EPI responded to the Factcheck.org critique. http://tinyurl.com/4824m See also http://tinyurl.com/4jx7a
Posted by: Dave Meyer on August 11, 2004 08:36 AMWhat's Mankiw (if it's him) doing using his official, taxpayer funded internet connection for this?
Posted by: BJ on August 11, 2004 02:04 PMWhat's Mankiw (if it's him) doing using his official, taxpayer funded internet connection for this?
Posted by: BJ on August 11, 2004 02:08 PMWhat's Mankiw (if it's him) doing using his official, taxpayer funded internet connection for this?
Posted by: BJ on August 11, 2004 02:13 PM