Slower US payrolls likely a one-off as weather casts cloud on data

So much for the hard data catching up with the soft. This month’s employment report showed a surprisingly weak increase in non-farm payrolls of just 98,000 compared with prior expectations of a 180,000 gain.

The downside surprise comes after a strong ADP private payrolls report earlier in the week and at a time when many were looking for actual figures on activity to catch up with the robust surveys.

Although this suggests some potential for downgrading the outlook we would note that the payroll numbers are notoriously erratic month-to-month and it is better to focus on the three-month trend which remains at 178,000 after strong readings in January and February (albeit slightly revised). In fact, until this release, payrolls had been one of the pieces of hard data which was more in line with the buoyant surveys.

Seasonal factors probably played a role. Warm weather will have boosted construction employment in January and February, artificially raising the seasonally-adjusted figures and thus limiting the bounce in March. And job growth in construction did slow significantly.

More worryingly, weakness was also apparent in the retail sector which shed jobs for the second month running. Although employment in this sector is under pressure from the increase in internet shopping, this would add to cyclical concerns over the outlook for consumer spending as real wages come under pressure from rising inflation. The latter, as measured by the CPI, has now caught up with average hourly wage growth with both running at 2.7%. There was little sign of an acceleration in pay growth in this report despite the unemployment rate falling to 4.5%, the lowest for ten years.

Putting these worries to one side, the drop in unemployment also hints at weather effects. The unemployment figures are calculated from the separate household survey which showed a bumper job gain of 472,000 on the month. This suggests that weather may have played a broader role in depressing the headline payroll figure as the latter is more of a one-week snapshot, while the household survey counts people as employed if they worked during the month even if they did not during the survey week.

All of which indicates a bounce back in payrolls next month, rather than the start of a downturn, with the headline number moving back in line with the more buoyant soft data.