• US job growth remains highly resilient
  • Wage growth is accelerating
  • Hours worked point to soft GDP unless productivity growth remains strong
  • Markets scaled back Fed cut pricing until ISM-services disappointed...
  • ...but ISM’s data quality may be questioned by its misleading jobs signal
 
  • Nonfarm payrolls, m/m 000s // UR %, Dec, SA:
  • Actual: 216 / 3.7
  • Scotia: 200 / 3.8
  • Consensus: 175 / 3.8
  • Prior: 173 / 3.7 (unchanged)

The US created 216k jobs last month which was only modestly above expectations especially considering confidence bands around the estimates. Wages accelerated and have posted the strongest back-to-back gains since July.

Markets initially responded by scaling back cut pricing even further (chart 1) until ISM-services disappointed at 10amET (50.6 from 52.7). As a caution toward ISM’s reliability, note that it said employment in the services sector contracted last month with an employment subindex that fell below 50 to 43.3 from 50.7, despite the gains noted below.

Chart 1: Market Pricing of Fed Fund Rate

Wage growth landed at 5.4% m/m at a seasonally adjusted and annualized rate (chart 2). Wage growth by this measure is accelerating from 3.2% in October to 4.3% in November and now 5.4%. This trend is why the year-over-year rate pushed higher to 4.1%.

Chart 2: US Hourly Earnings

By sector, job gains had significant breadth (chart 3). The private sector led the way with 164k more jobs created while government added 52k mostly at the state and local levels.

Chart 3: December Changes in US Non-Farm Payroll Employment

By sector, goods added 22k jobs and services add 142k. Within goods, construction was up 17k and manufacturing was up a little (6K). Within services, education and health led the way with a 74k gain followed by leisure and hospitality (+40k) and then retail trade (+17k), IT (14k) and professional/business services (13k) despite a 33k drop in temp help. The financial sector was flat (2k).

Aggregate hours worked fell 0.2% m/m but were up by a mild 0.7% q/q SAAR in Q4. There is no momentum baked in to Q1 hours worked based solely upon the hand-off effect as we await data on Q1 (chart 4). At the margin, unless productivity continues to surge, the waning trend in hours worked points toward a slowing GDP growth rate.

Chart 4: US Total Hours Worked

The unemployment rate held steady at 3.7% because it is derived from the household survey that registered a 683k drop in jobs offset by a 676k drop in the size of the labour force that cancelled each other out (chart 5).

Chart 5: US Unemployment Rate

So what is the household survey telling us? Always be cautious given that there is a lot more statistical noise to the household survey than payrolls. The 90% confidence interval for the change in unemployment in the household survey is +/-300k versus the 90% confidence interval for changes in nonfarm payrolls of +/-130k. The household survey is extraordinarily volatile as evidenced by the fact that just the prior month registered a gain of 586k jobs. Another difference is that the household survey also captures unincorporated businesses lacking formal payrolls, but again, the noise factor makes it hard to tell what’s really going on. Regardless, over time, the nonfarm (establishment) and household surveys tend to align (chart 6).

Chart 6: US Employment Levels