The Wage Channel Is the Weakest Link in the Case for a September Rate Increase

by Nora Halloran · September 2, 2026

Senior Analyst, US Market Current

There are two ways to read a 38,000 private payroll gain. The first is that the labour market is soft enough that raising rates would be gratuitous. The second is that it is soft for reasons rates cannot fix, and that the inflation problem sits somewhere else entirely. The second reading is the one the hawks are making, and the pay figures ADP published on Wednesday morning are the part of the record that sits least comfortably with it.

ADP said median base pay for job-stayers rose 3.0 percent from a year earlier in August, and median base pay for job-changers rose 4.7 percent. On a gross basis, which captures bonuses, commissions and tips, the two figures were 4.4 percent and 7.3 percent. The Bureau of Labor Statistics, on its own separate establishment survey, had average hourly earnings up 3.2 percent over the twelve months to July, in the Employment Situation released on August 7.

None of those numbers describe compensation running away from the Federal Reserve. The measure that covers the largest share of the workforce, the pay of people who did not switch employers, is growing at a rate that would be consistent with the inflation target were productivity growth anywhere near normal. The 4.7 percent for switchers is a scarcity premium paid to a shrinking pool of people who are actually moving jobs, not a broad wage impulse.

The hawkish case is an energy and credibility case

That is roughly what the more hawkish commentary concedes when it is set out in full. Marketplace, reporting on August 31, quoted Dan North of Allianz Trade saying there had been no progress on inflation and that the economy was back where it stood before the war started, and that without a rate increase the Fed would have a hard time reaching its 2 percent target, which he described as still some distance off and very sticky. The argument turns on the level of measured inflation, not on the pace of wages.

The credibility strand runs alongside it. Phil Camporeale of J.P. Morgan Wealth Management, writing on August 5, argued that a slower-than-expected normalisation of supply chains around the Strait of Hormuz, combined with markets questioning the Fed's inflation-fighting resolve after the July meeting, had lowered the bar for an increase in September. Both halves of that sentence are about something other than the labour market.

There is dissent. Marketplace also cited Jay Hatfield of Infrastructure Capital Advisors, who considers headline inflation overstated and expects core measures to fall despite elevated oil, and who regards an increase as ill-advised at present. Erasmus Kersting of Villanova University was cited making the more procedural point that the data does not clearly compel a move in either direction.

What the labour data actually shows

Underneath the ADP headline, the August gain was narrow to the point of being a single sector. Education and health services added 45,000 jobs; the entire service-providing category added 48,000. Professional and business services, which is where cyclical white-collar demand tends to show up first, cut 16,000. Manufacturing cut 17,000. Establishments with 20 to 49 employees cut 17,000 between them, and the West lost 8,000 jobs on net.

The official series tells a similar story with worse arithmetic. July nonfarm payrolls fell 23,000 against an average monthly gain of 34,000 over the preceding year, and the same release revised May down by 66,000 and June down by 37,000, a combined 103,000. An unemployment rate holding at 4.1 percent through that is doing so partly because people are leaving the count: the participation rate stood at 61.4 percent in July, and the BLS release states that since January the participation rate has declined by 0.7 percentage point and the employment-population ratio by 0.5 percentage point.

That last figure deserves more weight than it usually gets. A falling participation rate flatters the unemployment rate mechanically, and a policy framework that treats 4.1 percent as evidence of a tight labour market will read a shrinking labour force as strength. It is the same number arriving by a different route.

Friday, and then the sixteenth

Market pricing has moved toward tightening. Marketplace reported on August 31, citing CME FedWatch data as of that day, that futures implied roughly a 66 percent chance of a quarter-point increase in September and about 50 percent for December. That pricing predates Wednesday's ADP release. Consensus for Friday's official August report, as set out by TOPONE Markets on August 31, sits at roughly 50,000 to 55,000 payrolls with the unemployment rate unchanged at 4.1 percent and average hourly earnings around 3.0 percent year over year; TOPONE does not name the compiler behind every figure in that table.

The federal funds target range has been 3.50 to 3.75 percent since December 2025, and the committee announces its decision on Wednesday, September 16, according to TOPONE Markets' August 31 preview. Between now and then the Fed publishes its Beige Book, scheduled for today on its own release calendar and not yet out as of midday, and the BLS publishes the August Employment Situation on Friday, September 4, at 8:30 a.m. Eastern. Neither has been released yet, and nothing in this article anticipates what they will contain.

What can be said is where the tension sits. If Friday's report matches consensus and average hourly earnings come in near 3 percent, the committee will be weighing a rate increase against a wage series that is not obviously contributing to the problem, in an economy where the last two official payroll prints were negative or barely positive after revision. That is a defensible decision to make on energy and expectations grounds. It is a harder one to make on the labour market.