Friday, September 6, 2013

Something fishy here... John Lott on part time jobs

Greg Mankiw points us to John Lott, who suggests that 96% of new jobs this year are part-time.

OOOOBBBBAAAAAMMMAAAAACCCCCAAAAAARRRRREEEE!!!!!!!

::shakes fist::

I don't have time to delve into this too deeply right now, but one thing immediately jumps out at me - he is comparing January to August numbers.

You know January. It's that month that starts a couple days after Christmas and the holiday shopping season. You may also be familiar with August. It's that month before all the kiddies go back to school.

Something tells me we should wait until January 2014 and do a year-on-year comparison instead of subtracting the nadir of part time employment from its zenith and comparing it to total employment figures that likely are just poking along.

These do not appear to be new hire figures. That was my second thought - since there is higher turnover in part time jobs you don't want to compare new hire statistics either (net job creation would be OK).

How much you wanna bet Casey Mulligan picks this one up?

UPDATE, 11:08 am: One thing I do notice is that they are seasonally adjusted. Does anyone know how seasonal adjustment works for things like part time employment? Is all seasonality in all series completely cleaned out or does it clean out some kind of aggregate seasonality (in which case we'd still expect quite a bit of fluctuation in specific labor markets - this ought to be nailed down.

6 comments:

  1. Unles I'm reading the tables incorrectly, and comparing August 2012 with August 2013 (in thousands):

    PT Emp 8/12: 26,899
    PT Emp 8/13: 27,250
    % Change in PT: +1.34%

    FT Emp 8/12: 115,275
    FT emp 8/13: 116,920
    % Change in FT: +1.43%

    Tot Emp 8/12: 142,164
    Tot Emp 8/13: 144,170
    % Change in FT: +1.41%

    PT Emp as a % of Tot Emp, 8/12: 18.9%
    PT Emp as a % of Tot Emp, 8/13: 18.9%
    PT employment growth as a & of Total employment growth:
    18.0%

    ReplyDelete
  2. Using seasonally-adjusted data, by the way.

    ReplyDelete
    Replies
    1. Thanks for doing the leg work. Assuming these all check out, another way of putting this is 2,006,000 new jobs over that year in total, 351,000 of which were PT, so 17.49% of the total.

      The constant PT percent of total employment suggests this is near typical I suppose?

      So whatever "seasonally adjusted" does to the below-headline numbers, it clearly does not completely account for the seasonality of part time jobs.

      Delete
  3. And year-over-year should wash the seasonality out, more-or-less. (And apparently, Lott did not just look at PT employment, but at a subset of it--see some of the comments at MR.) The man is the definition of hack.) (I left my comment above on Lott's blog; we'll see whether it gets "approved" and posted.)

    ReplyDelete
  4. Lott did post my comment, and another making a similar point.

    ReplyDelete
  5. Although Mark Thoma may have been correct in saying that the U.S. economy's unemployment rate went down for the wrong reasons, I somehow doubt John Lott's calculations. (For Mark Thoma's post, please see the link below.)

    http://economistsview.typepad.com/economistsview/2013/09/unemployment-rate-drops-for-wrong-reasons.html

    The U.S. economy is still in poor shape and undergoing a lackluster recovery, but if 96% of all the jobs added were really part-time, wouldn't that actually be a sign of even worse things to come?

    BTW "doc", thanks for doing those calculations.

    ReplyDelete

All anonymous comments will be deleted. Consistent pseudonyms are fine.