Nine days ago the market was positioned for a September rate hike. On Friday morning one payroll print flipped it to pricing cuts, and the whole move happened inside a Fed that has deliberately stopped telling anyone what it plans to do.
That is the actual trade in the July employment report, not the 23,000 job losses. Kevin Warsh removed forward guidance from the FOMC statement, and Friday demonstrated the cost: when the committee refuses to describe its reaction function, a single monthly number, one with an error bar wide enough to drive a truck through, becomes the entire reaction function.
The Repricing
The 10-year Treasury yield fell five basis points to 4.63% on the release, with the front end moving harder. The dollar softened, the yen caught a bid, and equities slipped rather than rallied on the dovish read, which tells you the market took this as a growth scare and not a liquidity gift. Traders went from pricing no chance of a half-point cut at the September 16-17 meeting to pricing roughly a 17% chance of one, with at least two cuts now expected by year end.
Compare that to where positioning sat at the end of July. The FOMC held at 3.50% to 3.75% on July 29 with three dissents, and the consensus going in was that the next move was more likely up than down. Nine days and one data release later the curve is priced for easing. That is not a market digesting news. That is a market with nothing to anchor to.
Nobody Should Be Trading This Number This Hard
Here is what makes the reaction genuinely uncomfortable rather than merely volatile.
Payrolls came in at negative 23,000 against a Dow Jones consensus of 83,000. In the same release, the Bureau of Labor Statistics cut May by 66,000 and June by 37,000, a combined 103,000 revision. June, which printed as 57,000 five weeks ago and which we covered at the time as a cooling but functioning labor market, is now 20,000. The revision to two months of history was more than four times the size of the headline everyone traded.
Then there is the denominator problem underneath all of it. Breakeven payroll growth, the monthly job creation needed just to hold unemployment flat, has collapsed. Federal Reserve Board staff estimated in April that labor force growth is now near zero, driven by an aging population and collapsed net immigration, which drags breakeven employment growth down with it. Researchers at the Kansas City Fed reached the same conclusion, and the St. Louis Fed’s own estimate range for 2026 spans 15,000 to 87,000 jobs a month, which is a polite way of saying nobody knows.
Sit with that range. If breakeven might be 15,000 and might be 87,000, then a print of negative 23,000 is somewhere between a modest undershoot and a genuine contraction, and the release itself does not tell you which. The unemployment rate did not help either: it fell to 4.1%, but participation dropped to 61.4%, the lowest since February 2021, so the improvement is arithmetic rather than information.
Markets are pricing 50 basis points of Fed policy off an indicator whose signal-to-noise ratio has quietly degraded to the point where the revision is bigger than the number.
What Warsh Built and What It Costs
Warsh’s argument for killing forward guidance is defensible on its own terms. He has said the Fed spends too much energy telling markets what it will do instead of explaining the conditions under which it would act, and there is a real case that a decade of guidance turned the FOMC into a hostage of its own dot plot. We wrote about the guidance vacuum he created at the July meeting and the risk was legible then.
The bill arrived Friday. Guidance was never only communication. It was volatility suppression. It gave the market a framework so that any single data point could be weighed against a stated reaction function rather than treated as the reaction function. Remove it and you do not get a market that thinks harder about conditions. You get a market that overreacts to whatever number lands next, because the next number is all it has.
For anyone actually running money, the practical consequence is that macro hedging just got more expensive. Rate volatility is now the base case into every BLS release, every CPI print, every JOLTS report, and the pattern will repeat until the committee either restores some description of its reaction function or the data starts behaving.
Who Wears It
The corporate read is narrower than the macro read, and worse.
The job losses were concentrated in local government education, down 50,000, and retail, down 19,000, heading into the back-to-school window. That is not an AI-displacement story or a tech-layoff story. That is municipal budgets and consumer-facing payrolls contracting at the same time, which is a demand signal, and it lands on staffing firms, discount retail, and anything selling into households whose second income just disappeared. Health care kept hiring, as it has for two years, which mostly tells you where the rest of the economy is not creating jobs.
The hires rate has fallen for three straight years and now sits at 2013 levels. Companies are not firing. They are simply not opening requisitions, and a hiring freeze does not show up in initial claims, which is exactly why the labor market can look stable in the weekly data right up until it does not.
If September brings a cut, the transmission lag means it reaches corporate hiring plans somewhere around next spring. The businesses cutting headcount now are not doing it because money is expensive. They are doing it because they cannot see demand, and no cut fixes visibility.