February 2, 2018

January 2018 Employment Situation Summary (020218): +200K SA Jobs Added, Jobless Rate Still 4.1 Percent; SA Figure Vastly Overstates Underlying Strength

Filed under: Economy,Taxes & Government — Tom @ 7:05 am


Not Seasonally Adjusted Job Losses:

January is a month when there are huge job losses on the ground before they are seasonally adjusted:


So the question isn’t how many jobs will be added; it’s how many will be lost — obviously, the fewer the better. Really good results in today’s release for total nonfarm and the private sector would be -2,700K or fewer and -2,200K or fewer, respectively.

I would also hope we’ll see an upward revision to December’s strangely low 148K jobs added.

Finally, in the Household Survey used to determine the unemployment rate, January is the month when population controls are adjusted, affecting most of that survey’s key figures, and forcing all related changes into January instead of spreading them over previous months. So it may be important to separate the “real” January effect from the population control effect, which BLS will do. Last year, doing so showed that there was an unusually high increase in the civilian labor force and the number of people employed, and a significant decrease in the “not in labor force” figure.

The report will be here at 8:30 (full report with tables will be here).

HERE IT IS: Better than expected, but the ADP-BLS gap continues to increase —

Total nonfarm payroll employment increased by 200,000 in January, and the unemployment rate was unchanged at 4.1 percent, the U.S. Bureau of Labor Statistics reported today. Employment continued to trend up in construction, food services and drinking places, health care, and manufacturing.

In January, the unemployment rate was 4.1 percent for the fourth consecutive month. The number of unemployed persons, at 6.7 million, changed little over the month.

… Among the major worker groups, the unemployment rate for Blacks increased to 7.7 percent in January, and the rate for Whites edged down to 3.5 percent. The jobless rates for adult men (3.9 percent), adult women (3.6 percent), teenagers (13.9 percent), Asians (3.0 percent), and Hispanics (5.0 percent) showed little change.

The number of long-term unemployed (those jobless for 27 weeks or more) was little changed at 1.4 million in January and accounted for 21.5 percent of the unemployed.

The civilian labor force and total employment, as measured by the household survey, changed little in January (after accounting for the annual adjustments to the population controls). The labor force participation rate was 62.7 percent for the fourth consecutive month and the employment-population ratio was 60.1 percent for the third month in a row.

Establishment Survey Data

Total nonfarm payroll employment rose by 200,000 in January. Employment continued to trend up in construction, food services and drinking places, health care, and manufacturing.

Construction added 36,000 jobs in January, with most of the increase occurring among specialty trade contractors (+26,000). Employment in residential building construction continued to trend up over the month (+5,000). Over the year, construction employment has increased by 226,000.

Employment in food services and drinking places continued to trend up in January (+31,000). The industry has added 255,000 jobs over the past 12 months.

Employment in health care continued to trend up in January (+21,000), with a gain of 13,000 in hospitals. In 2017, health care added an average of 24,000 jobs per month.

In January, employment in manufacturing remained on an upward trend (+15,000). Durable goods industries added 18,000 jobs. Manufacturing has added 186,000 jobs over the past 12 months.

Employment in other major industries, including mining, wholesale trade, retail trade, transportation and warehousing, information, financial activities, professional and business services, and government, changed little over the month.

The average workweek for all employees on private nonfarm payrolls declined by 0.2 hour to 34.3 hours in January. In manufacturing, the workweek declined by 0.2 hour to 40.6 hours, while overtime remained at 3.5 hours. The average workweek for production and nonsupervisory employees on private nonfarm payrolls edged down by 0.1 hour to 33.6 hours.

In January, average hourly earnings for all employees on private nonfarm payrolls rose by 9 cents to $26.74, following an 11-cent gain in December. Over the year, average hourly earnings have risen by 75 cents, or 2.9 percent. Average hourly earnings of private-sector production and nonsupervisory employees increased by 3 cents to $22.34 in January.

The change in total nonfarm payroll employment for November was revised down from +252,000 to +216,000, and the change for December was revised up from +148,000 to +160,000. With these revisions, employment gains in November and December combined were 24,000 less than previously reported. (Monthly revisions result from additional reports received from businesses and government agencies since the last published estimates and from the recalculation of seasonal factors. The annual benchmark process also contributed to the November and December revisions.) After revisions, job gains have averaged 192,000 over the last 3 months.

More analysis after digging.


On the Household Survey side, the BLS’s population-control adjustments saw a 318,000 increase in the number of people employed at the end of 2017 than was estimated a month ago:


This stands in stark contrast to last year’s population-control adjustments which ended up showing 487,000 fewer people working at the end of 2016.

NOT SEASONALLY ADJUSTED JANUARY 2018 RESULTS: The raw declines in payroll jobs were 3.085 million for total nonfarm and 2.593 million in the private sector. Both declines are the largest since the recession year of 2009.

As seen below, Team Trump got a huge break in the seasonal adjustments from what one could reasonably have expected based on previous years’ results:


Based on the raw data, it would have been reasonable to expect seasonally adjusted job LOSSES.

Anyone claiming today’s results are genuinely strong needs to understand this, and change their minds — because they’re not.

More thoughts shortly …

UPDATE (references are to seasonally adjusted data unless otherwise indicated):

  • The participation metrics haven’t moved up much since the 2016 election. It’s only limited solace that they were expected by many to continue to decline.
  • Black unemployment jumped to 7.7 percent. The BLS claims that black unemployment increased by 186K, while it decreased by 294K for everyone else (because the overall net decrease is 108K). That’s pretty hard to accept.
  • Full-time employment increased by 293K, and part-timers increased by 14K, which makes the reported reduction in average hours worked a little hard to take, both short-term (down by 0.2 hours in Jan,) and long-term (down 0.1 hours from a year ago, despite 2.391 million more full-times and 92,000 fewer part-timers)
  • Last month’s 9 cents per hour increase in average hourly pay was revised up to 11 cents. Add that to the 9 cents reported this month, and you have the BLS’s first inkling of genuine wage growth in a long time. The press is turning this two-month improvement into an obsession over looming inflation. Give me a break, especially given that weekly pay dropped a bit because of the (questionable) decline in the average work week.

Overall — This was not a good month, though the hourly pay improvement was heartening.



  1. Quick thoughts:

    - Participation is continuing to flatline (LFPR NSA 62.3%/SA 62.7%, latter unchanged from December, is 0.2 points NSA/0.2 points SA below January 2017), while those who want to work are finding it (emp-pop NSA 59.5%/SA 60.1%, latter unchanged from December, is 0.3 points NSA/0.2 points SA above January 2017).

    - The NSA job change was UGLY-with-a-capital-Y. -3.085M non-farm/-2.593M private should not have come close to a SA +200K/+196K.

    Comment by steveeg — February 2, 2018 @ 8:58 am

  2. Agree completely on NSA vs. SA, and just posted my calcs.

    This then REALLY begs the question of why ADP looks so much better, because if you considered today’s result to really be -30K SA, ADP’s result Wednesday was 264,000 higher, in ONE month (though ADP never discloses NSA, so there’s no way to get an answer).

    Comment by Tom — February 2, 2018 @ 9:32 am

  3. Maybe in the rush between you two to talk down the Trump economy you ignore that the raw data could be wrong, misleading or BLS BS.

    Comment by zf — February 2, 2018 @ 10:38 am

  4. Well, the raw data is subject to revision, which we all know, but usually that doesn’t change things by more than 75K-100K on the outside — and that could go either way.

    I have consistently pointed out that ADP is showing higher figures month after month, a problem which hasn’t been resolved in now-13 months, and that BLS may be sandbagging news that is in reality better than they’re reporting.

    But when I run the raw numbers through the BizzyBlog Jobs Report Reality Check, a name I’ll use from here on out, and see huge differences indicating that the seasonal adjustments are screwy and don’t reflect the underlying raw numbers, I’m not going to ignore them — especially with raw results that are the worst Jan. since the recession, AND the second-worst Jan. in almost 80 years of monthly jobs reporting (before considering labor force size).

    UPDATE: Further, and I didn’t expect this, the BLS’s Birth-Death Model, which is part of the total raw job loss figures, estimated fewer job losses in Jan. 2018 (-198) than in Jan. 2017 (-247):

    2018: https://www.bls.gov/web/empsit/cesbd.htm
    2017: https://www.bls.gov/web/empsit/cesbdhst.htm

    This means that BLS is at least recognizing that net business formations are going up during the Trump era (though both Jan. figures are losses because Jan. has so many post-year-end let-gos). Maybe there’s more new business formation happening than BLS is picking up, and if so, that would largely explain why ADP may end up being more right than BLS.

    UPDATE 2, 4:30 P.M.: Also, let’s say the raw numbers are wrong, and that actual job losses were 250K less. That would roughly translate into 250K being added to the SA number, which would now be 450K. The problem that the seasonally converted amount doesn’t reflect the underlying reality would still be there.

    UPDATE 3, Feb. 4: I should also note that the markets’ bad day on Friday would seem to indicate that investors weren’t fooled by the topline seasonally adjusted figures.

    Comment by Tom — February 2, 2018 @ 2:10 pm

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.