I have to admit that I guffawed a bit when I saw Jan Hatzius, chief economist for Goldman Sachs, answer a question on CNBC yesterday as to whether the Fed will hike in July. He answered that his firm pegs 35% odds on a move in either July or September, not trivial but hardly a base case, but it was his reference as to the Fed needing to see if the next round of employment data validate the poor May report, or whether the last release was some sort of isolated one-off.
This seems to now be a bit of a consensus — that the May jobs number was an anomaly.
My question is how can that be? How can it be treated as some sort of one-off?
Have economists lost sight of the fact that between October and March, aggregate hours worked rose at two-and-a-half times the pace the growth in real nonfarm business output. Did that make any sense?
Yet back then, nobody seemed to be commenting on the contracting productivity backdrop and what the implications would be for job creation. Well, we’re now seeing the answer.
In other words, the employment slowdown is not a one-month wonder, it is purely fundamental, and a pattern of erosion is now taking hold.
This wasn’t a one-off as employment has been weakening for three months running. Take the average of the past three months — it is down to a mere +116,000.
I remember the days when a trend like this would have economists wondering if a recession was around the corner.
Goods-producing employment has contracted for four months in a row.
So this is not just one data-point but a visible weakening trend — this critical cyclically sensitive segment of the economy has contracted now for four months in a row and the cumulative damage is 77,000 jobs or a -1.2% annual rate.
Private service sector job gains have throttled back big-time — from +222,000 in February to +167,000 in March to +130,000 in April to +25,000 in May (ratified by ISM’s non-manufacturing survey as the jobs index sagged to 49.7 in May from 53 in April, tied for the second weakest reading of the past five years; is this supposed to be a one-off data point too?).
Temp-agency employment shrunk 21,000 in May, down now in four of the past five months and by a cumulative 64,000 which is a losing streak we have not seen since August 2009. Again, this seems like a trend, not a one-off.
Full-time employment declined 59,000 on top of a 316,000 plunge in April. And the Household survey, when put on a comparable footing to the Establishment (payroll) Survey, the “population and payroll concept adjusted” employment measure slid 105,000 after a 293,000 falloff the month before.
Employment for adult males contracted by 82,000 and that was not a one-off number — it followed a decline of 97,000 in March.
And the losses were in the highest income cohort, but this is also the area on the age strata most vulnerable these days to corporate cost-cutting — the 45-to-54 year cohort (plunging 105,000 in April following a 51,000 turndown in March).
The Fed’s Labor Market Conditions Index (LCMI; the compilation of Janet Yellen’s “dashboard” indicators) contains 19 different ingredients and it rang in with a -4.8 reading in May — the data are reported as the change in the index as it is the direction that is important, not the level. So the April change was the worst we have seen since May 2009, when the economy was still in recession.
The LCMI has declined for five months running and has slowed now for seven months running. Again, that is not “one number”; that is a visible trend.
And the Job Opening & Labor Turnover Survey (JOLTS) data were disappointing, as the number of new hires shrank 198,000 on top of a 220,000 March slide — that 418,000 two-month plunge was the fourth steepest on record. The number of new hires is all the way down to an eight-month low.
The decline was broadly based across construction, retail trade, professional & business services, the once-hot leisure & hospitality sector, and government. Again, not a one-off.
And the hiring rate, in a new and descending pattern, has gone from 3.8% in February to 3.7% in March to 3.5% in April to stand at its lowest level since August 2014.