The Bureau of Labor Statistics releases the Employment Cost Index four times a year. The release lands at 8:30 ET on the last Friday of January, April, July, and October, covering the prior calendar quarter. It is the smallest release on the wage complex by column-inch of coverage, and the largest by weight in the actual Fed reaction function. This piece lays out why the ECI carries the wages read the FOMC prices against, how the composition-corrected structure differs from average hourly earnings, and what the front-end rates desk does with the print inside a decision cycle.
Two wages lines, two different questions
Average hourly earnings is the line the desk reads on the first Friday of every month, inside the Employment Situation Summary at 8:30 ET. It reports the change in average hourly earnings for all employees on private nonfarm payrolls, month-over-month and year-over-year, to a tenth of a percent. The consensus figure is the median of the Bloomberg or Reuters survey published the Monday of the release week. The desk carries AHE as one of the four lines inside the NFP framework, alongside headline payrolls, the U-rate, and the two-month revision.
The Employment Cost Index answers a different question. AHE asks: what did the average worker earn per hour in the reference period? ECI asks: what did the price of labor do in the reference period, holding the composition of the workforce fixed? Those are different questions because the composition of the workforce moves every month.
The composition problem shows up in two places. The first is industry mix. If a hurricane closes a set of low-wage restaurants for a month, the average hourly earnings across the remaining workforce goes up. Nobody got a raise. The second is occupation mix inside an industry. If a firm hires ten new junior analysts and promotes two senior analysts to management, the AHE at that firm goes down. Nobody took a pay cut. The BLS built the ECI to strip both of these effects out.
The ECI holds industry and occupation shares fixed at the level of the 2018 Occupational Employment and Wage Statistics survey. When a new quarter’s data comes in, the BLS reweights the underlying wage observations to the fixed 2018 basket. The resulting index moves only when the wages paid to a given occupation in a given industry actually move. The FOMC prices against that index because the FOMC cares about the price of labor, not the average paycheck.
What the ECI actually reports
The ECI release publishes six headline series. Three cover total compensation, wages and salaries, and benefit costs for civilian workers (private plus state and local government). Three cover the same three cuts for private industry workers only. Each series reports the quarter-over-quarter change and the year-over-year change to a tenth of a percent.
The FOMC watches two of the six as primary lines. The first is private wages and salaries, quarter-over-quarter, seasonally adjusted. This is the cleanest read on what the market price of labor did in the reference quarter, before the pension and health benefit calculations run through. The second is total compensation for private industry workers, year-over-year. This is the trend anchor. The desk cross-checks the state-and-local components as a labor-market-tightness cross-read, but the primary read is on the private side.
The historical corridor on private wages and salaries runs at about 0.6 to 1.5 percent quarter-over-quarter in a healthy expansion, with the 2021 to 2022 cycle running above the top of that range on a sustained basis. The year-over-year on total compensation runs a percentage point or two above the CPI-adjusted trend the Fed treats as consistent with the two-percent inflation target, which is why the ECI year-over-year is the line that shows up in the FOMC statement’s discussion of wage pressures. When Fed governors give a speech that mentions the two and a half to three percent corridor as the level of wage growth consistent with two percent inflation, they are quoting the ECI year-over-year, not the AHE year-over-year.
The composition wedge is the read
The most useful thing the ECI print does is create a wedge against the monthly AHE trend from the same quarter. Take the three AHE prints inside a quarter, average them to a quarterly rate, and compare that rate to the ECI wages and salaries quarter-over-quarter for the same period. The difference is the composition effect the AHE picked up.
When the ECI runs hotter than the AHE quarter, composition was pulling AHE down while the underlying price of labor was running firm. That resolves hawkish for the Fed read, because the Fed is looking at the composition-adjusted line. When the ECI runs cooler than the AHE quarter, composition was pushing AHE up while the underlying price of labor was running soft. That resolves dovish for the Fed read, and it usually shows up in the FOMC statement as an acknowledgment that wage pressures are easing even though the monthly AHE trend does not obviously reflect it.
The two most common composition wedges in the post-2020 cycle came from leisure and hospitality reshuffling and from remote-work-driven occupation mix inside the professional services complex. Both created AHE reads that overstated or understated the underlying labor-cost trend, and both got corrected out at the end of the quarter by the ECI print.
What the desk does with the release
The 8:30 ET print on ECI Friday is a small event by front-end volume. The two-year on-the-run typically moves two to four basis points on the release, an order of magnitude smaller than the eight to twelve basis point moves the NFP framework can generate. But the ECI print sits inside the Fed’s decision paragraph in a way the AHE print does not. The desk reads the ECI release as the last read the FOMC gets on wages inside the pre-meeting data window, and the reprice on the release runs through the next-meeting probability strip, not the on-the-day yield.
The Q2 print, released on the last Friday in July, lands inside the pre-September-meeting window. The Q3 print, released on the last Friday in October, lands inside the pre-December-meeting window. Both are the terminal wages read for the meeting that follows. That timing is the reason the ECI release is the wages line the desk carries into the meeting, and the reason a soft ECI print in late July has done more historically to move the September probability than any single monthly AHE print in July or August has.
The Fed reads ECI because ECI answers the question the Fed is trying to answer. The desk reads ECI because the Fed reads ECI. Everyone else reads AHE.