Friday, January 04, 2013
I don't get Table A (the Household survey). What it is currently showing is that we haven't added any new jobs since October. There is a reported loss of 23K jobs since October, but that's too small to be meaningful. The quality of jobs has risen - if you look at table A-9, you see that in 2012 we finally got a good increase in full-time jobs and that it continued through December.
The establishment data is much more positive, and here the construction jobs finally showed up. They are there, as ADP showed. However the pace of jobs added is too low (esp. compared with last year at this time). I'm confused as to how we managed to get a cut in temp employment? This whole thing just feels off.
I've got to go through this thoroughly. Tentatively, some of the weakness showing in the Household survey is a result of SA adjustments and expansion in hours worked in retail rather than expansion of employees - which says something about prior weakness. The number of unemployed individuals went up by 164K, but the participation rate did not change (63.6). Nonetheless, given the expansion in the unemployment rate for women 20 and older (7.0 > 7.3), I'd say that unemployment was generated by apps for holiday jobs.
The above interpretation appears somewhat supported by detail from the establishment survey, which records a loss of 11.3K jobs in retail. Again, that didn't happen - that's from seasonal adjustment factors that anticipated more seasonal hiring than occurred, but the difference was probably made up in more hours for the workers already there.
The number of marginally attached and discouraged workers both went up, and are both higher than they were a year ago. That is perturbing. Early unemployment rose in all categories, which is a bit puzzling, because Sandy should have come right back out. The unemployment rate for women who maintain families rose to 11.3% from 10.7% in the prior month. That rate is choppy, but somewhat intimidating. The unemployment rate for full-time workers rose from 8.1 to 8.3%.
I guess we'll wait until next month to see how this shakes out. Claims are remarkably distorted so they aren't really reliable at this time, and changes in seasonal hiring patterns can cause quite a bit of distortion.
I actually think that ADP showed a "truer" reflection of the US economy this month, but that seems to be an inane opinion of mine. I expected much better on the household survey this month, and about 175K on the establishment survey, so reality just took a dump on my head.
Sandy ought to be creating extra jobs now, and the pace of that increase should be quickening. It does seem to show up in specialty construction this month, but not nearly enough. '
I'd like to retreat to the M_O_M cave and insist that the Household survey is really FUBAR, but the underlying details say it may not be. Teen unemployment rose over the past year and the unemployment rate for those over 25 with some college or an associate degree did too. Since last year the employment growth trend has markedly worsened. Any year in which the Sept-Dec employment numbers don't improve for this group is highly suspect, and this year they did not. Since emerging from recession in 2009, each year we've seen a growth of employment for this group from September to December. The recent trajectory is:
2008: lost about 300K jobsI find that even more surprising because the amnesty thing should have boosted jobs in this group.. This doesn't mean that we are in recession, but it does imply that employment growth this year will be somewhat disappointing at best.
2009: gained about 60K jobs
2010: gained about 260K jobs
2011: gained over 400K jobs
2012: lost more than 240K jobs
For the population as a whole, in 2011 Sept-Dec SA employment rose by 762K. In 2012 it rose 331K, but all of that was in Sept-Oct. And more - since October we have lost jobs. The huge bump in Sept SA employment from August could explain some of this, but not all.
*From the non-seasonally adjusted Establishment Data jobs -2,000.
*From the non-seasonally adjusted Household Survey jobs -100,000.
*From the seasonally adjusted Household Survey jobs -97,000.
*This compares a year ago and then the change from November to December. This I believe is the only valid way to look at the numbers.
Just looking at the change from the previous month the way the headline jobs number does:
From the non-seasonally adjusted Establishment Data jobs -103,000
From the non-seasonally adjusted Household Survey jobs -489,000**
From the seasonally adjusted Household Survey jobs +28,000
** not a mistake on my part – you can see they didn’t like the big negative number and seasonally adjusted it to a positive 28k. Same for the headline number going from -103,000 to a +168,000 for an adjustment of plus 271,000 jobs thank for very much, BLS Bureau of Lie Statistics.
Conclusion. The bigger established companies reported stable employment. The smaller companies not so good.
What to expect. January in the real world always shows a really big drop in jobs. I am sure seasonally adjusted we will see job growth. All the job growth each year happens February through June. Then the number of people employed stays more or less the same the rest of the year. Again, this is in the real world, not the dreamland of seasonally adjusted data. If January’s drop in employment is less than previous years, 2013 is off to a good start.
Links to this post: