The complimentary insight below was written yesterday by Hedgeye Macro analyst Christian Drake.
BLS published the preliminary annual benchmark revision for the national employment data this morning. The estimate revisions occur on an annual basis and cover the April to March period (so, April 2018- March 2019 in the present instance). The latest revision totaled (negative) -500K … which means, on average, private sector employment was approximately -43K less per month than originally reported.
The BLS couches the size of the revision in terms of the total employed base. Total Private Sector employment as of March 2019 was 128.29 Mn, so the -514K revision equates to a revision of -0.3%. Not too bad. But framing the revision in this way may be somewhat misleading. A more meaningful contextualization would be to frame the adjustment in relation to the number of jobs added over the revision period, as originally reported.
The existing/original estimate showed we added 2.4 million private sector jobs over the 12-month reference period. A negative -514K revision on a base of 2.4 million means over 21% of the jobs reportedly gained during the period never actually occurred. More than 1 in 5 jobs being completely illusory is obviously more material than that connoted by the official -0.3% contextualization.
The revisions won’t be incorporated into the official data until the final benchmark revision is issued in Feb 2020, but the downward revision is notable, particularly given that it occurred concurrent with the peak in the U.S. economic cycle and appears to be further corroborated by the current deceleration in aggregate hours growth and aggregate private sector income growth.
In short, another data patch in the macro analytical mosaic …. but one reported on a severe lag and incorporated on a timeline delay so large as to preclude any practical utility. It does, however, provide further confirmatory evidence of the practical utility of our Quantamental process to front-run (the delayed reporting of) macro reality.