ADP’s upside was (again) a surprise (324k vs. Bloomberg consensus 189k). I’m not sure we should learn too much from this development.
figure 1: Private nonfarm payrolls from ADP (teal), Bloomberg consensus (tan squares), sa implied level of 000 from BLS (blue bold), Bloomberg consensus (red squares), by adding consensus change To calculate the level released in June. Sources BLS, ADP (from FRED, ADP, Bloomberg) and authors’ calculations.
It’s hard to see how big of a surprise this is. This is the same information in m/m variation (m/m log difference looks similar).
figure 2: Month-on-month change in private non-farm payrolls, from ADP (cyan), Bloomberg consensus (tan squares), from BLS (blue bold), Bloomberg consensus (red squares), in 000s sa source BLS, ADP via FRED, ADP, Bloomberg and author’s calculations.
My suspicion about extrapolating too much from this ADP number is that, seasonally adjusted The numbers are clearly weak (Goldman Sachs points to distortions in the seasonal adjustment process). Using the estimated relationship for June, Big BLS numbers expected. Instead, we get a number (149K) that is weaker than the consensus number (218K).
