My Photo
Note: Jeff does not accept guest blog posts on A Dash of Insight.

For inquiries regarding advertising and republication, contact [email protected]

Follow Jeff on Twitter!

Enter your email address:

Delivered by FeedBurner


  • Seeking Alpha
    Seeking Alpha Certified
  • AllTopSites
    Alltop, all the top stories
  • iStockAnalyst
Talk Markets
Forexpros Contributor
Copyright 2005-2014
All Rights Reserved

« Weighing the Week Ahead: Will Disappointing Earnings End the Stock Market Rally? | Main | Weighing the Week Ahead: Time for a Year-End Rally? »

October 23, 2013



Chris -- The Dieli chart uses the seasonally adjusted series -- -as it should.



Eric -- Seasonal adjustments do not "smooth" data. That would be a moving average.

Seasonal adjustments identify a regular and repeated pattern and remove it, thereby exposing the reality of what occurred. Over the course of a year the seasonal adjustments net to zero.

The process for seasonal adjustment of government data is thoroughly documented and professional. There is no bias involved --either on the part of the government or Dr. Dieli in his analysis.

On the contrary, those with an agenda often choose unadjusted data. They rely upon the fact that most people do not understand how seasonal adjustments are done, and think that government statisticians are at the beck and call of the President.

So the seasonal adjustment did not "produce" a 5 sigma event. BTW, the margin of error in the household survey is on the order of 450K.

My interest is in making sound investment decisions. The political debate is distracting people from that mission.


Patrick R. Sullivan

'It is pretty obvious from this chart that the part-time story is based upon the recession, not anything that has happened after 2009.'

I guess that depends on your definition of 'not anything'. If this was a normal recovery the two lines should have reversed (back to where they were in 2007).

Eric Kuehl

Jeff, while I agree one should not put political considerations into their process, one also needs to investigate the statistics when a 5 sigma event such as the current month's data shift implies. If you look past the seasonally adjusted data and look at the raw data you will see that the reason for the huge data shift came not from some magical appearance of full time jobs and loss of part time jobs but a change in the seasonal adjustment. One can not declare their viewpoint victorious until rigorous analysis is applied, IMO. Since the adjustment formula is not known then one can only take an informed guess as to the major shift...are you doing so without some bias? The total full time jobs DECLINED 560k and part time jobs INCREASED 702K. This coincides with the increase in multiple job holders of 176K. Do you believe the raw data that confirms the recent trend or do you believe the seasonally adjusted data that produced a 5 sigma event in the data? (I thought seasonal adjustments were to smooth the data...just saying.) Here is the raw data:

Chris of Stumptown

Excellent commentary. One question: Doug Short's chart is seasonally adjusted numbers. Dielli's data doesn't specify. Do you know whether his is as well?

It'd be interesting to see Short's chart with a similar unadjusted chart.

Douglas Lee

I am missing the point. The part time work we care about are those who want to work more but have had their hours cut or cannot find full time work. The number of those part-timers increased an insignificant amount in September. August was down some from July, but this number has really fluctuated in a narrow range all year -- 7918 in Jan and 7926 in Sept. Those who choose to work part time are not a problem.


Guerrilla Economics; Providing an insight and counterbalance to the noise.

See ya at the track.


Great post Jeff, and very informative!

The comments to this entry are closed.