by Calculated Risk on 1/12/2020 02:03:00 PM

A few comments on the current headline jobs streak.  Also, the job streak if we adjust by temporary decennial Census hiring, and the possible impact of the annual revision (to be released with the January employment report).

The employment report has shown positive job growth for a record 111 months.

Headline Jobs, Top 10 Streaks
Year Ending	Streak, Months
20191	111
1990	48
2007	46
1979	45
1943	33
1986	33
2000	33
1967	29
1995	25
1974	24
1Streak still going!

However, if we adjust for Decennial Census hiring and firing (data here) the streak of consecutive positive jobs reports is actually 118 months long. It makes sense to adjust for the Census hiring and firing since that was preplanned and unrelated to the business cycle.

If the job streak continues into 2020, then the headline streak will probably end in June 2020 when a large number of temporary Census workers are let go.  But if we adjust for temporary Census hiring, then the streak might continue.   Of course the streak could end at any time.

The Census will hire approximately 500,000 temporary workers over the next 5 months (mostly in May).   This is why I wrote last year: How to Report the Monthly Employment Number excluding Temporary Census Hiring

Click on graph for larger image.

This graph shows the monthly change in payroll jobs, ex-Census (meaning the impact of the decennial Census temporary hires and layoffs is removed - mostly in 2010 - to show the underlying payroll changes).

Note that there were 3 thousand temporary Census hires in December, so this graph shows 142 thousand jobs added (not the headline number of 145 thousand).

The previous longest job streak was 48 months ending in 1990.  However, if we adjust for the 1990 Decennial Census, that streak was actually 45 months - making the streak ending in 2007 at 46 months the second longest.

Ex-Census Jobs, Top 10 Streaks
Year Ending	Streak, Months
20191	118
2007	46
1979	45
1990	45
1943	33
1986	33
2000	32
1967	29
1995	25
1974	24
1Streak still going!

Note: If you have questions about this adjustment, see this post (including my discussion with the BLS).

Another interesting question is will the annual benchmark revision, to be released with the January employment report, show that the job streak ended last year? This is possible. The preliminary estimate of the Benchmark revision "indicates a downward adjustment to March 2019 total nonfarm employment of -501,000". Usually the preliminary estimate is pretty close to the final estimate. These jobs are subtracted from the March 2019 total, and then wedged-back to March 2018 (reducing employment gains by about 42 thousand per month during the 12 month period).

Since the lowest month for job gains, during the 12 month adjustment period, was February 2019 with 56 thousand jobs added - it seems the job streak will still be intact. But maybe not. The annual adjustment includes "the effect of applying the rate of change measured by the sample to the new benchmark employment level, as well as updated net birth/death model forecasts and new seasonal adjustment factors". The combined effect of these adjustments could be a negative month during the adjustment period - something to watch for next month!