ES -- DX/CL -- isee -- cboe put/call -- specialist/public short ratio -- trinq -- trin -- aaii bull ratio -- abx -- cmbx -- cdx -- vxo p&f -- SPX volatility curve -- VIX:VXO skew -- commodity screen -- cot -- conference board

Friday, October 02, 2009

 

BLS acknowledges birth/death model misstating job creation


bloomberg:

The U.S. economic slump earlier this year was so severe it short-circuited the government’s model for calculating payrolls, raising the risk that today’s jobs report may be too optimistic.

About 824,000 more jobs may be subtracted from the payroll count for the 12 months through last March when the figures are officially revised early next year, a Labor Department report showed today. The revision would be the biggest since the government started adjusting the numbers in 1991.

The bulk of the miss occurred in the calculations for the first quarter of this year, the Labor Department said. The economy shrank at a 6.4 percent annual pace in the first three months of 2009, the worst performance since 1982.

The figures raise the possibility that the government’s calculations continue to miss the mark.

“We are probably still underestimating job losses,” said John Silvia, chief economist at Wells Fargo Securities LLC in Charlotte, North Carolina. “There could be another 30,000 to 40,000” that the data isn’t picking up.

That would mean the loss of jobs for September could turn out to be as high as 300,000, rather than the 263,000 reported today by the Labor Department. Today’s report also showed the jobless rate climbed to 9.8 percent last month, a 26-year high.

The potential revision for the year through last March would mean that the economy lost 5.6 million jobs for the period instead of the 4.8 million now on the books.


that would be a 16% underestimation of job losses for the period ending march 2009.

From April 2008 through December, the tax records showed the Labor Department’s payrolls figures overestimated payrolls by about 150,000, said Chris Manning, the national benchmark branch chief at the Bureau of Labor Statistics. That implies the estimates missed the mark by about 675,000 in the first quarter of this year, which currently shows a 2.1 million drop in payrolls.

“In this period of steep job losses, the birth/death model didn’t work as well as it usually does,” Manning said in an interview. “To the extent that there was an overstatement in the birth/death model, that is likely to still be there.”

The model added about 184,000 jobs to the payroll total last quarter compared with a 135,000 increase in the same period in 2008, before the financial crisis deepened with the collapse of Lehman Brothers Inc.

“This birth/death model is still assuming that we are getting new jobs from new-business creations,” David Rosenberg, chief economist at Gluskin Sheff & Associates Inc. in Toronto, said in an interview.

“These additions are coming somewhere from ‘Alice in Wonderland,’” he said....


the birth/death adjustment has been a bone of contention for quite a while, with a dash of insight among others advocating the model and barry ritholtz disparaging it. while i joined in criticizing the adjustment, a year ago i believed oldprof to have been proven out. the voiding in his perceived vindication, however, appears to be material as ritholtz's essential criticism -- that the adjustment would fail, as an extrapolation of past conditions, to properly state job creation during turns of trend -- looks spot on.

Labels: ,



Barry (and many others) have criticized the BLS approach to measuring job creation for many years. Up until Q109, the modeling process worked very well.

I thought that q408 would be the "stress test" for this approach.

I'll write something that goes a bit deeper, and I trust that you will keep reading with an open mind, as you always do.

Thanks,

Jeff

 
------ ------- ------
of course, prof -- you are always thought provoking and worth reading! these are extraordinary times and it seems few models have dealt well with them. here's looking forward to your analysis!

 
------ ------- ------
Trying to measure the number of jobs to the nearest 10,000 in a country with 160 million or so working people is an impossible task--by the time you can finally figure out what the count was in any particular month it is far past the time that the number was of any particular use.

I think that the real time (OK, one day delayed) withholding tax receipts data is a much better barometer of how things are going in the labor market. I assume that it is correct. . .

 
------ ------- ------
What? We don't believe that 100K jobs a month were created in companies too small to count the past 2 years? Gasp ;)

http://www.fundmymutualfund.com/2008/01/monthly-jobs-report-birthdeath-model.html

 
------ ------- ------
Trader Mark -- Instead of gasping, you could actually look at some data. The last published data from the Business Dynamics Series (which draws on the same state employment records used in this article) showed that during Q408 there were about 1.3 million new jobs at newly opening businesses (and therefore not in the survey) and 5.3 million at existing businesses.

So that would be closer to 100K jobs per week, not per month.

You may now breathe again.

http://oldprof.typepad.com/a_dash_of_insight/2009/08/the-most-important-data-you-missed-last-week.html

 
------ ------- ------

Post a Comment

Hide comments


This page is powered by Blogger. Isn't yours?