Good news for a Fed that was (probably) going to cut anyway: Jerome Powell’s favorite gauge of US employee compensation rose less than expected in Q3, according to a marquee BLS release on Thursday.
The Employment Cost Index printed 0.8% on the headline (0.7855% unrounded), softer than the 0.9% consensus expected.
As the figure below shows, the Q3 readout counted as another “coolest since inflation accelerated in earnest” print for a key input in the Fed’s decision calculus. The last time this measure printed below 0.8% was Q2 of 2021, and before that Q3 of 2020.
I put a lot of emphasis on ECI for a reason: It’s the most comprehensive measure of comp costs, and Powell cited it specifically in explaining his initial hawkish pivot in late 2021.
As BMO’s Vail Hartman wrote, of Thursday morning’s release, “This is a piece of information that help[s] lessen the degree of inflationary angst that’s returned to the forefront of the macro narrative.”
Note that the private industry series from the ECI release exhibits a pretty strong relationship with the quit rate from the JOLTS report.
As the figure above shows, 12-month pay growth for the private sector is likely to decelerate quite a bit more going forward if the quit rate — which slipped to 1.9 in Tuesday’s JOLTS release — is any guide.
So, coming quickly full circle, a key gauge of comp costs for the Fed came up benign, and when taken in conjunction with the signal from JOLTS and, less importantly but still worth a mention, cooler pay growth for “job switchers” in an otherwise scorching-hot ADP release, that suggests Powell can “safely” cut rates next week.
There are two caveats. First, PCE prices for September (i.e., the Fed’s preferred inflation metric) told a slightly different story and so did another very low initial claims print. I’ll get to those momentarily.
Second, all we’re really saying with ECI and JOLTS is that there’s no risk (or very minimal risk) of a wage-price spiral. That’s necessary to green-light a rate-cutting cycle, but it’s arguably not sufficient.
