Source item

Quoted context

These are the quoted author’s words.

ADP Shows A Rebound But The Seasonal Adjustment May Be Doing More Work Than The Labor Market The September 15 ADP NER Pulse improved to 16,250 jobs per week over the 4 weeks ending August 29. That is almost double the July 25 low of 8,250, but still about 33% below the June 13 reading of 24,250 and only back to the same level as July 4. More importantly, this is a preliminary, seasonally adjusted 4 week moving average not the number of jobs created in the latest week. ADP also uses a 2 week lag so the release contains nothing from September 5 or September 12. The Rebound May Be Partly Statistical My base case is that some of this improvement is real, but some may reflect seasonal adjustment interacting with a labor market that already weakened earlier in the cycle. Imagine an industry normally hires 40,000 seasonal workers and later lays off 40,000. If demand weakens and it only hires 10,000 this year, it may later only lay off 10,000. The raw workforce still falls. But if the seasonal model expects the normal 40,000 decline, losing only 10,000 can mathematically appear as 30,000 seasonally adjusted job growth. That is the danger near turning points. If the normal hiring upswing never happened, the normal seasonal layoffs cannot happen either. A model calibrated to the historical hire then fire pattern can mistake fewer layoffs than usual for genuine underlying improvement. ADP then smooths those adjusted weekly estimates again through a 4 week average. So the headline is built from the raw employment change then the seasonal adjustment, then the 4 week smoothing. That does not make the 16,250 false. It means the raw weekly data and seasonal factors would be needed to know how much of the rebound represents actual new hiring. August Already Showed Weak Breadth The August monthly report produced just 38,000 private jobs and the composition was narrow. • Education and health services added 45,000 • Leisure and hospitality added 16,000 • Manufacturing lost 17,000 • Professional and business services lost 16,000 • Trade, transportation and utilities lost 5,000 • Information lost 4,000 • Natural resources and mining lost 5,000 5 of 10 major industries contracted. Education and health alone created more jobs than the entire private economy. Excluding that category employment fell roughly 7,000. Excluding education and health plus leisure and hospitality, the rest of the private economy lost roughly 23,000 jobs. Large establishments added 34,000 jobs, nearly 90% of the headline gain. Small establishments added only 3,000 net while medium sized establishments added nothing. What Is Probably Driving The Improvement The Pulse gives no industry or employer size detail, so we cannot know. But based on August’s composition I would expect education and health services to remain the strongest contributor followed by leisure and hospitality, construction and large employers. I would be much more skeptical that manufacturing, business services, information or trade and transportation suddenly became major engines of hiring. There is also a possible World Cup effect. Weak post event employment weeks may now be rolling out of the 4 week average, allowing less negative late August weeks to lift the headline without a genuine surge in gross hiring. My Take The rebound is real within ADP’s adjusted estimate. But August already proved that positive payroll growth can coexist with deteriorating breadth. The September Pulse only tells us the national aggregate improved afterward. It does not tell us whether the weak sectors recovered, whether the improvement came from fewer separations rather than more hiring, or whether seasonal adjustment amplified the rebound. Until manufacturing, business services and smaller employers improve I’d treat this as a narrow late cycle rebound, with seasonal adjustment possibly exaggerating some of the strength.

onechancefreedm ↗