BLS Overstated Job Creation by 14.38% in 2007

For those of you who have been wondering, here is the data on the 2007 Benchmark Revision.

The good news is when looking at the total employment situation, believe it or not, BLS was pretty accurate. Their models of employment came within 0.2% of the actually measurement of jobs.

Where they fell down was accurately measuring NFP job creation. Based upon their own benchmark revisions, BLS overstated new job creation by a whopping 19.95% 14.38%.

Nearly every month, the subsequent revisions were rather large:
>

2007 NFP Benchmark Revisions

Bls_revisions_212008_2

>

I suspect this says less about BLS, and more about the difficulties in
trying to accurately measure new job creation in near real time.

>

 

Source:
Benchmark Article
Daniel Jackson
BLS, February 1 2008
http://www.bls.gov/web/cesbmart.htm

PDF version
http://www.bls.gov/ces/cesbmart.pdf

>

 

 

Print Friendly, PDF & Email

What's been said:

Discussions found on the web:
  1. michael schumacher commented on Feb 1

    the huge revisions are a benefit to this administration because it decided to switch to a model that could show whatever they want WHENEVER they want.

    Coincidence that the NFP #’s started to fall off a cliff and then POOF we now have a model that doesn’t allow that to happen.

    job creation , in this country, is the big myth. ever since someone decided that China was a cheap alternative…say 20 or so years ago.

    Ciao
    MS

  2. tyoung commented on Feb 1

    Sept. should be positive/green.

  3. fede commented on Feb 1

    check the difference
    i think it´s -191K instead of -265K

  4. Harley Evans commented on Feb 1

    Something wrong with this article – the September input is backasswards? how does it look when corrected, does that change your story?

    ~~~

    BR: Good catch — I’ll fix above.

    There’s an errant negative sign in the Septmber data series. That changes the overstatement from 19.95% to 14.38%

  5. Steve commented on Feb 1

    fede,

    Heh…it really IS hard to count jobs.

  6. Neal commented on Feb 1

    95K per month–not even close to equilibrium.

  7. Vermont Trader.. commented on Feb 1

    Why would you trust the revisions any more than original numbers?

    Action in market the last two days feels like a big player is getting liquidated or blowing up. Watch youself everyone.

  8. kk commented on Feb 1

    Bernanke had a reason to lower aggressively after all.

    Charlie Gasparino looked like his dog died this AM with the monoline demise looking more unlikely by the day. Last night on Kudlow he gave the monoline death spiral the old college try, and it was pretty entertaining. These days Charlie is a must watch.

  9. Steve Barry commented on Feb 1

    Market could crash here…I have a funny feeling…look at gold tank on no news. This rally is ridiculous. If Dow 12000 breaks, it will hit 9800 very fast.

    Look at Dow chart…right shoulder top right here. Aug and Nov lows are right around here too and are serious overhead resistance. Ominous.

    head and shoulder

  10. scorpio commented on Feb 1

    Barry: i agree and thanks to your post got me off the pot and out of cash. good call so far.

  11. Vermont Trader.. commented on Feb 1

    From John Hussman’s most recent post..

    http://www.hussmanfunds.com

    – “there is one particular scenario that would be ominous in my view. That would be if we see a relatively uninterrupted series of declines that breaks cleanly through the August and November lows, followed by a one-day advance of 200-400 Dow points. That’s a script that markets tend to follow pre-crash. Though it’s not a strong expectation or forecast, it’s something worth monitoring, because we’ve started to see the pattern of abrupt jumps and declines at 10-minute intervals that is often a hallmark of nervous markets.”

    From my perspective, that’s largely the script we observed last week. Short-term market movements are demonstrating a form of chaotic instability that has generally indicated growing contagion rather than independence among investors. The most important difference between current market conditions and prior crash events is the behavior of interest rates. For example, rates were rising persistently before both 1929 and 1987. The clear downtrend of interest rates may turn out to be a saving grace here, given that the market’s most spectacular losses featured hostile rate trends. Still, the best interest rate action has been in Treasuries, while credit spreads have been pushing to new highs, so the favorable trends in Treasury yields are partly a symptom of growing default risks in other areas.”

  12. Ross commented on Feb 1

    The arbs don’t seem to be too confident about the YHOO bid.

    Leaving 12 to 15% under the bid is a very bad sign……..If this puppy breaks it will be bad karma. Just a thought.

  13. michael schumacher commented on Feb 1

    steve-

    I agree with the H&S…it has been setting up since fall last year. The real issue is if it will be “allowed” to happen.

    Fed just said it would do $60 billion TAF’s every two weeks for an “indefinate period of time”……

    That is not a sign that all is well….I wonder if that amount will show up verbatim in the fractional reserve requirements that seem to be populated by the exact auction amount. See Shedlock’s post about this:

    http://www.minyanville.com/articles/MER-C-jpm-bac-MBI-WB/index/a/15712

    I am a bit surprised that more people are not aware of this…..not really sure if it means much but….it should.

    How do you cause a bank run???…….LOL
    let’s see if it matters…..

    Ciao
    MS

  14. Steve Barry commented on Feb 1

    With the move in gold, I have a feeling dollar has made a major bottom. Time for other currencies to be debased to catch up…this will further cripple S&P earnings due to lower foreign revenue.

  15. Marcus Aurelius commented on Feb 1

    Gold is dropping like a rock. Somebody is dumping (I wonder who?). If the market absorbs this supply quickly (as in this isn’t a general sell-off), we’re in some deep doodoo.

  16. michael schumacher commented on Feb 1

    repost:

    on Gold:

    My guess is that some bank (most likely Spain) is dumping yet again. They actually back up the printed money with a real asset…funny that…..like for like.

    Ciao
    MS

  17. michael schumacher commented on Feb 1

    being the first of the month I would say ECB’s are getting capitalized via Gold sales.

    Ciao
    MS

  18. Justin commented on Feb 1

    What we need to do is create an Federal Independent Statistical Department. Just like the FED – lol!!!

  19. Paul in NYC commented on Feb 1

    Want to share here what I just read on another board:

    Okay, let me help you all understand what the Fed’s moves mean for mortgage rates.

    First, a little background. 799 FICO score, millions in liquidity, looking at a 2BR in the $1.5M range, solid mid-6 figure income.

    I spoke to Chase today and the jumbo rate they quoted was 7% for 30y fixed, 0 points. Conforming (I can pay an outsized down payment to get me there) is 6.25%.

    What the fuck, right?

    When I asked about the Fed’s moves, Chase’s response was that mortgage rates would actually RISE.

    Yes, you heard me right – Chase said that rates WILL GO UP TOMORROW as a result of the Fed. Why?

    Because the yield on the 10 year jumped. Because inflation expectations have gone up.

    What the Fed does has little to zero effect on mortgage rates for the rest of us. In fact, it hurts mortgage rates. It does, however, have a beneficial effect on credit card rates, auto loans and student debt.

    So, the folks… who claim that all is well because of the Fed simply do not have a grasp of the facts. At all.

    So, it seems like there will be a lot less home buying goin’ on in at least the near future if this is any indication.

  20. Justin commented on Feb 1

    Barry, why don’t you make it interesting…all you gurus should start calling each other out! You guys play this game of “I’m right, but I don’t want to step on your toes.” If one has to crash and burn, to tell the truth, well then lite the fire! Can’t believe we are not at 2001 levels, when the economy is so much worse off. I just got back from a long drive and there were so many for sale signs, my head is swimming. I just turned 50 and can’t remember anything like this. Perhaps there is a happy story to be told when it is all over. These Fuckers on tv, pump, pump, and pump – Kudlow should have goldielocke’s stuffed down his throat – I’ve got his goldie right here.

  21. michael schumacher commented on Feb 1

    how many times are we going to see
    Moody’s, Fitch put out PR that says they “might” downgrade the mono’s.

    I “might” get a million dollars…..

    Hell one of them has been “assessing” that very thing for over two months now…..

    And still no call outs at all except Ackman, who got really quiet all of a sudden..

    Ciao
    MS

  22. Pool Shark commented on Feb 1

    Stuart,

    Fascinating stuff! Makes you wonder how much the bullion banks have been losing recently, being forced to buy gold in the open market to cover their lease agreements.

    “Staring into the abyss” indeed.

  23. dblwyo commented on Feb 1

    OT – a fair and balanced post on Barry’s part. The revisions weren’t major though the data challenges make things noisy. Which means you ought to be looking at trends and patterns, especially YOY changes (which you can see in his charts these days). The key point is that the slowdown in employment growth which started circa Jan06 continues and is accelerating. FWIW some charts looking at NFP & Unemployment (which is often clearer on trends) are posted here: http://tinyurl.com/229g4a

    When you look at that data a couple of things. First the differences in the old vs revised from JAN79 to now are about -375K jobs. Not significant IMHO. BUT…the economy as a whole is -2.9 million jobs in the hole when you use 150K/month as the breakeven point. Now that’s significant.

    OffTopic – this does indeed look like a nervous market. Thanks for the Hussman post. Excellent guy who I wished there was more time to follow. It looks like Mr. Market is debating whether or not have that bounce let along “recover” from the “correction”. As various pieces of the credit breakdown keep rippling outward (bond insurers, corporate debt, housing walk-aways, increasing delinquency rates)these stresses will just increase.

  24. Jmea commented on Feb 1

    Statistics, Statistics, and, then there are lies.

    Obfuscating the nation into a war, Guantanamo, the ingeniously named “Patriot Act”, down to the commandeering of the school curriculum (“intelligent design”) and national statistics (CPI).. and, now, this revision of the employment data !

    Goebbels will envy this record!!

    But, then again, I have an open mind. This may be all just coincidence. Especially,if tyoung has his facts right that a 14% revision of the data is within the historical norms of BLS revisions!!

  25. Michael Donnelly commented on Feb 1

    Barry, you have to remove November and December of 2007 and add in Nov and Dec of 2006.

    Neither Nov and Dec of 2007 are final, they are still in the 2nd and 3rd revision stage and certainly will be subject to big changes next Feb of 2009.

    And for the record? Nov of 2006 was revised down by 14,000 jobs and Dec of 2006 down by 27,000 jobs (old monthly change over new change)

  26. LFC commented on Feb 1

    FMI, I’ve often heard the figure cited that we need to add 150,000 jobs per month to keep up with population growth. Can anybody tell me if there’s an actual source for that figure, or if it’s just based upon some percentage of average population growth?

    Thx.

  27. Michael Donnelly commented on Feb 1

    LFC

    That number is just the %chg that payroll needs to keep up with population growth. It’s also an old number. You’d need to take a look at population growth rates in the 18-65 age group and then apply that percentage growth rate to payroll jobs to back out the latest number but I’d bet 150,000 jobs per month is pretty close

  28. Johnny Vee commented on Feb 1

    This small market rally feels a little strange–Like we are in the eye of a hurricane. Once it passes, the back half of this sh!t storm is going to be severe.

  29. LFC commented on Feb 1

    Thanks, MD. I was wondering if that was the case, or if there was a tracked “official” number anyplace.

  30. dblwyo commented on Feb 1

    LFC (& Michael) – actually the 150K number is labor force growth (a subset of pop but not a big deal) + productivity growth. It’s been pretty well analyzed though sources escape me at the moment. We had a big debate on CalculatedRisk last year and there were some strong arguments for it going down to 120-140K give or take slowing pop/lbr force growth plus slowing productivity. The BLS has a very handy tool for dloading many of these states and the STL Fed FREDII site keeps a subset that is very easy to dload. You can check the math by looking at Pop growth and looking up some productivity stats. Anyway that’s how it was explained to me FWIW :)

  31. Steve C commented on Feb 1

    Currently, (3PM, EST) the DJIA is bouncing against its resistance level for the 3rd time in 2 days. Will it break out above, or fail for good? Is Ben B watching?

  32. dblwyo commented on Feb 1

    LFC (& Micheal): intriguing. Went and did some fiddling around. The compound rate of growth of the Labor Force is about .0114% monthly, with an annul average of 1.6%, so with a LF of about 153M looks like you’d need 174K jobs/month. Add on a labor productivity of 1.5-2.0% and wow we really have a problem. No wonder real wages haven’t been keeping up and the middle class (me at least) don’t feel so optimistic. By the way Labor productivity from 1948 to now was, over the period, downtrending from about 2.5% to 2.0%. Similarly for 60-now. BUT from 80 to now it was flat at 1.5% from 80-96, increased to 2.0% 97-00 and is showing a slight decrease since about ’06 or so. Can you spell re-engineering (downsizing) ? :)

  33. Michael Donnelly commented on Feb 1

    dblwyo,

    I’ve gotta disagree with you on this stuff. The way I described it is the way we did it at Global Insight.

    If you use Labor Force you’re going to get the wrong number because folks can and do step in and out of the labor force very much depending on what the business cycle is doing.

    Using Productivity just makes zero sense to me at all. All that does is tell you how much more STUFF we can churn out, not how many JOBS are needed to maintain the unemployment rate.

  34. Michael Donnelly commented on Feb 1

    Over the past 24 months population has grown on average by 0.89% y/y, so to get payroll to grow by that same 0.89% y/y you would need 102,000 new jobs every month.

    Population growth was faster during the 1990s, hitting 1.3% at times so you’d need 150,000 jobs.

    Of course all this assumes that labor force participation remains the same and assumes the 18-65 age group is growing the same as overall population.

    but enough, or Barry will tell me to GYOFB

  35. Pat Gorup commented on Feb 1

    14% was probably in their ballpark. Hey, that’s about how much inflation is understated do you think there’s a correlation?

  36. dblwyo commented on Feb 1

    Michael – Pop vs LbrFrce ? Oh well. Conceeded. I was playing back of the envelope with the data I had.

    On productivity though if fewer hours will produce the same amount of goods I need less labor. To absorb that excess labor I either need a)lower wages, b)increased output or c)more job opportunities. Or all of the above.

    Anyway the 150K figure of merit is also one I’ve heard the Fed use and for the Pop + Prod growth reason. Believe Mankiw ahs made the same point among others.

    But a fun debate all the same. And usefully OT for Barry’s post IMHO :)

  37. blam commented on Feb 2

    The standard error is on the order of +- 500,000, so the month-to-month is meaningless.

    All in all, pretty good data. If I was a gambler, the serial correlation of downward revisions suggests that the job growth trend is below historical expectations.

  38. MikeW commented on Feb 2

    Two questions.
    Why does the Market pay so much attention to Non-Farm Payrolls when its methodology is, as we discuss here every month, so fanciful?

    Why wouldn’t the ADP employment report be more accurate and trustworthy? As the biggest payroll processor, they should have access to a big enough sample of very accurate data, and they shouldn’t have the political agenda that an executive branch agency of the Federal Government might have.

  39. kio commented on Feb 3

    Effect of increasing labor force on developed economies could be not only strictly positive. It is well known that increasing labor force participation rate, LFPR, results is decreasing productivity, P.
    I have compiled a simple figure
    (link to the figure
    http://inequalityusa.blogspot.com/2008/02/effect-of-labor-force-growth-on.html )

    illustrating this observation for the USA. For the sake of better representation, the measured growth rate of labor force participation rate, dLFPRm/ LFPRm, has been converted according to the following relationship:
    dLFPRc/LFPRc= -2.0dLFPRm/LFPRm + 0.022

    Effectively, dLFPRc/LFPRc, is a scaled inverted and shifted dLFPRm/LFPRm.
    In addition, I have smoothed the original data series using leading MA(3), i.e. leading 3-year moving average. It is clear from the figure that any increase in dLFPR/LFPR resulted in a doubled reduction in the productivity growth rate.

    More interestingly, cross-country comparison shows that relative (i.e. ratio) productivity changes could be completely explained by corresponding changes in participation rates. In other words, when productivity grows in one country relative to another country, corresponding LFPR move in opposite direction.

    http://inequalityusa.blogspot.com/2007/10/some-myths-about-slowdown-of.html

Posted Under