UBS Payrolls Preview
Payrolls +190k, a 0.1 pt drop in the unemployment rate, on-trend earnings Employment growth likely continued to expand in July (UBSe 190k, consensus 180k), albeit not as quickly as in June. A number of distinct accelerations lifted payrolls to their 287k rise in June: -Returning Verizon strikers (worth an estimated +35k) -Above-trend hiring in arts/entertainment/recreation (+20k) -Probably unsustainable surges in food manufacturing, child care, and hotels (each worth an incremental 10k). More generally, the figure was probably boosted by payback for exaggerated weakness in May and by the five-week gap between May and June employment surveys, which we estimate contributed around 30k to payrolls. (Both of these effects are intermingled with industry swings in payrolls in June.) We doubt that the underlying pace of private payrolls is higher than the 158k per month averaged year to date. However, more limited retooling shutdowns than are typical of the beginning of July probably pushed July payrolls slightly above that pace. We also allow for a bit less weakness in weather-sensitive industries; a warm winter probably pulled some hiring forward into Q4-Q1, resulting in less seasonal hiring in Q2.
Recent data are somewhat at odds but on balance consistent with our forecast The ADP estimate of private payrolls (179k) appears consistent with our July estimate, but ADP errors relative to the official BLS estimate suggest caution in putting too much weight on that. (In June, the initial BLS estimate exceeded the ADP estimate by 93k.) But more generally, the ADP figures corroborate the gradually slowing trend in payrolls since the start of the year. Jobless claims and consumer sentiment also suggest continued labor market improvement, with still low claims and further gains in households' assessments of the labor market. (See charts.)
However, ISM employment indexes have been somewhat weaker, and employment indicators from the household-survey have been a lot weaker. Our all-economy ISM employment index, which combines the manufacturing and nonmanufacturing measures, has looked softer than payrolls since early this year but only after appearing stronger than payrolls before then. (Recent ISM employment index levels historically would be consistent with a trend in payrolls of only about +100k per month.) The payroll-equivalent employment measure from the household employment survey (the underlying survey for the unemployment rate) has been falling outright in recent months, with declines since Q1 averaging 172k per month. The household-survey data tend to be much more volatile than the establishment-survey data (payrolls)—for example, they averaged 329k per month in H215 versus payrolls' 237k. We expect a reacceleration in the household-survey measure in coming months—one reason to look for renewed decline in the unemployment rate.
We've been surprised by payrolls volatility, but not by broad slowing The slowing in employment growth this year roughly matches our expectations. Private payrolls averaged 221k per month in 2015, 181k per month in Q1, and 135k per month in Q2. Our credit/employment models, which relate credit conditions to the following year's trend in private payrolls, signalled a 150k per month trend in private in 2016—very close to the year-to-date pace of 158k. The shortfall in hiring in May looks anomalous, and both it and the June pace were exaggerated. But smoothing through that volatility yields a gradually slowing trend in hiring consistent with our model.