Here is the contradiction most people missed. On June 17, the Federal Reserve did nothing to interest rates — held them at 3.50%–3.75%, exactly as everyone expected — and the S&P 500 reportedly turned in its worst "Fed day" since 1994.
A no-op decision produced a shock. That should be impossible if you think a Fed day is about the rate. It isn't. It's about the guidance, and Kevin Warsh's first meeting as Chair was an exercise in taking the guidance away.
Last week this column read the June meeting as a passive hold — a dollar holding firm while a new Chair sat on his hands. That read was too gentle. What actually happened was active: Warsh didn't just decline to cut, he began dismantling the machinery markets use to front-run the Fed at all.
Warsh didn't just decline to cut, he began dismantling the machinery markets use to front-run the Fed at all.
Start with the statement. Warsh's ran 132 words. Powell's last one ran 344. His explanation was pointed: the shorter version "just gives you the facts, as best we can judge it."
What got cut matters more than the count. The "easing bias" language — the phrase that told everyone the next move was more likely down than up — was gone. So was forward guidance entirely, which Warsh said the committee agreed "was not well suited to the current policy conjuncture."
Then the strange part. Warsh, as sitting Chair, refused to submit his own dot in the Summary of Economic Projections. He has called the dot plot "not helpful in the conduct of policy" for years, and he acted on it.
The arithmetic that leaves behind is its own signal. Among the 18 members who did submit, nine project at least one hike this year, eight see no change, one wants a cut. A clean nine-to-nine split — and the one person who could break it deliberately stayed silent.
If Warsh leans as hawkish as his public record, the true center of gravity sits further from "dovish" than the median dot admits. The market can't see it, though, because he removed the one data point that would locate him.
And then he went further than any single meeting. Warsh stood up five task forces to review how the Fed operates — communications, the balance sheet, data sourcing, AI and productivity, and the inflation framework itself. The SEP and the guaranteed post-meeting press conference are reportedly both on the table.
This is not a style choice for one Wednesday. It's the opening move of an institutional overhaul, and the target is the Fed's entire signaling apparatus.
Here is where it stops looking like chaos and starts looking coherent. Warsh has been "exhorting the bond market over and over again to watch the data, not the Fed."
His reasoning is mechanical, not rhetorical. The Fed wants to use the bond market as one of its key data inputs. But if the bond market reacts to what it thinks the Fed will do rather than to the incoming data, that input gets "polluted by this circular expectation."
Think about what a dot plot actually does in that loop. It tells the market what the Fed expects; the market prices to the dots; the Fed then reads market pricing as confirmation of its own forecast. The signal is the echo. Warsh is trying to cut the feedback wire.
Remove the dot, delete the easing bias, refuse to promise a press conference, and you strip out the crutches that let markets triangulate the Fed instead of the data. What's left is a bond market forced to form its own view — exactly the uncontaminated input Warsh says he wants.
And the market is obliging. The Treasury sold $84 billion of six-month bills this week at an investment rate of 3.97%, up from 3.80% two weeks earlier. In the secondary market the six-month yield touched 4.0% before settling at 3.97%.
That puts it 35 basis points above the effective federal funds rate, per Wolf Street. Since early January the six-month yield has surged nearly 50 basis points, flipping from below EFFR to well above it — the bond market pricing more than one hike inside a window shorter than six months, with no dot telling it to.
The two-year confirms the move, up 76 basis points since early February to 4.14%. Even banks are ratcheting brokered-CD yields above 4% to compete with T-bills. The signal is coming from the data, not from a Fed nod, which is the entire point.
Now put the two pieces together, because this is the thesis. Warsh didn't turn hawkish and leave the guidance apparatus intact. He turned hawkish and removed the apparatus — and the removal, not the rate path, is what drove the equity shock.
An information vacuum is volatile by construction. When markets can no longer anchor to a dot or a bias, every incoming print carries the full weight of repricing that guidance used to absorb.
You can see the vacuum in the disagreement itself. Polymarket still puts September no-change on top at 69.5%, with a 25bp hike at 23.5% and a cut all but written off. Meanwhile BofA is reportedly calling for three hikes this year. Both are staring at the same Fed and cannot triangulate it — because there's no longer a dot to triangulate to.
That is the regime Warsh built on purpose. With no forward guidance to lean on, the next hard data release becomes the only guidance markets have left.
Which brings us to July 15. The Bureau of Labor Statistics releases CPI and core CPI that morning, and in the old regime a hot core print would have been a hint — something the market would hold until the Fed nodded at it.
Not now. Strip out the dot, and the print stops being a hint and becomes the signal itself. Repricing happens on the number, immediately, because there's nothing downstream to wait for.
Watch core, not headline. Warsh has been explicit that the Fed targets core PCE, which strips out energy, and he has waved off the Iran oil spike as a headline story with a short tail — Dallas Fed modeling reportedly puts its contribution to core inflation at roughly 0.3 percentage points.
Strip that out and underlying inflation still runs in the high-2% to 3% range, above target for five straight years. So a hot core CPI on July 15 doesn't get the energy-shock excuse. It lands as evidence the underlying trend hasn't broken — the kind of reading Warsh said would make him sympathetic to a hike.
Here's the elegant, uncomfortable part for anyone trading it. Warsh has said market prices are among the most important signals a central banker has. So a hawkish repricing after a hot print isn't just a market reaction — it's corroborating evidence Warsh can read back into the reaction function.
The feedback wire he cut on the Fed's side, he rewired on the data's side. The bond market moves on the print; Warsh reads the bond market as confirmation; the loop closes without a dot ever entering it.
For years the game was to guess what the Fed would signal. Warsh has changed the game to guessing what the data will say — and then letting the Fed watch you guess. On July 15, for the first time in the Warsh era, there will be no dot to hide behind when the number lands.