US labor productivity rose the most in almost three years in Q2, data out Thursday showed.
Market participants tend to ignore this release, or if that’s too strong, it’s fair to suggest it often gets short shrift. That’s too bad. Because it’s relevant. And eminently so in the current macro environment.
The 3.7% jump for Q2 blew away estimates, and came courtesy of a 2.4% increase in output and a 1.3% decline in hours worked. The latter represented the first drop since 2020.
Measured against Q2 of 2022, productivity rose 1.3%. It was the first four-quarter increase since late 2021.
Given the composition of the decline in hours worked (it was entirely attributable to a decrease in average weekly hours, as employment was unchanged), it’s worth asking if we should be using different metrics to assess labor market slack.
It’s possible (indeed it looks more likely by the week), that in the era of labor hoarding, metrics like the headline NFP print, jobless claims and the unemployment rate may not tell us as much as they once did.
In any case, unit labor costs rose 1.6% in Q2, Thursday’s data showed. That was well below consensus, and Q1’s 4.2% increase was revised lower to 3.3%.
As a reminder, ULC is just the ratio of hourly comp and productivity. If productivity is rising, ULC will be lower, all else equal. Put differently: Increases in productivity help blunt the impact to businesses of higher labor bills.
Hourly compensation rose 5.5% in Q2. Notably, the 2.7% increase in real comp was the first in nearly a year.
When taken in conjunction with last week’s cooler-than-expected update on the Employment Cost Index, these figures were a good news story for the Fed. Generally speaking, they help make the soft landing case.
Higher productivity, positive real (i.e., inflation-adjusted) hourly compensation, lower unit labor costs and a drop in hours worked that doesn’t entail a drop in employment, together constitute a favorable conjuncture.




Reading between the lines, I wonder if the increase in productivity is, in part, due to a return to “work at the office”, as “working from home” gets phased out/minimized?
I work more at home than I did at the office. Furthermore, I wonder how productivity is measured for software engineering as it’s not very cut and dry so to speak.
My lay impression is that productivity is GDP divided by hours worked, as measured and estimated by the BEA and BLS. I think these metrics look at sales of final goods and services, inter-industry sales of intermediate goods and services, compensation, taxes paid, operating surplus, etc. Page 3 of the third link has some sort-of-understandable examples. They aren’t trying to count lines of code produced 🙂
https://www.bls.gov/productivity/technical-notes/output-measures-major-sector-productivity.pdf
https://www.bea.gov/system/files/papers/P2006-8.pdf
https://www.bea.gov/sites/default/files/methodologies/industry_primer.pdf#page=11