TradingFuse
Market research, published in the open
Macro 28 August 2026 · 10 min

Warsh delivered hawkish at Jackson Hole. Gold crashed 2.75 percent. Cut thesis invalidated.

Chair Warsh keynote at 10 AM ET committed the Fed to the 2 percent PCE target as "firm and fixed," called July headline PCE at 3.7 percent "concerning," said the committee "still has work to do," and announced a deliberate move to a quieter Fed communication posture. Chicago PMI 47.1 vs 57.9 consensus (10.5-point collapse) same morning at 9:45 AM ET. Gold -$126.53 (-2.75 percent) to $4,474.45, broke through $4,540 and $4,500 support levels; Phase 3 top-signal pattern complete. 10Y +5.5bp to 4.726 percent; 2Y +12.4bp to 4.356 percent (bear-flattening 2s10s to 27bp). DXY reclaimed 99.34 (+55 pips). Fed funds futures flipped from 65 percent dovish-cut to 60 percent hawkish-hike. Honest post-mortem: input probability distribution was wrong on the July-PCE reaction function under Warsh; cross-asset consistency worked but on a wrong distribution.

Catalyst check. Friday August 28. Chair Warsh delivered the keynote at the Kansas City Fed's 2026 Economic Policy Symposium at Jackson Lake Lodge at 10:00 AM ET as scheduled. The speech text was published on the Federal Reserve Board's speeches index at speech-start under the URL federalreserve.gov/newsevents/speech/warsh20260828a.htm; the symposium theme was "Financial Innovation: Implications for Payments and Policy." The Chicago PMI released at 9:45 AM ET printed 47.1 versus 57.9 consensus and 57.6 prior, a 10.5-point single-month collapse and the first contraction reading in four months. Speech date and Chicago PMI date verified against the Kansas City Fed symposium page and the ISM release calendar.

The keynote was hawkish. Warsh committed the Fed to the 2 percent PCE target as "firm and fixed," called the July headline PCE at 3.7 percent "concerning," said the committee "still has work to do to bring inflation under control," announced a deliberate move to a quieter Fed communication posture ("a quieter Fed, more purposeful in its communications, is better able to meet its objectives"), and closed by laying out guiding principles for policy that read as a break from Powell-era forward-guidance conventions. The two questions the option market was priced against, whether Warsh would give the market a September cut and whether he would use forward guidance to signal it, both came back no.

The tape

  • Gold: $4,474.45, down $126.53 (-2.75 percent) from Thursday's $4,600.98. Intraday session low $4,442, high $4,616. The two-and-a-half-day decline from Tuesday's $4,655 fresh cycle high is now roughly $180, or 3.9 percent, and gold has traded through both the $4,540 and $4,500 support levels the Wednesday piece's Phase 3 rules identified. This is the extension leg of the multi-year-high framework's top-signal read: two attempts at $4,700 rejected within three sessions, followed by a decisive break below the confirmation line. The pattern is complete.
  • 10-year yield: 4.7260 percent, up 5.5bp from Thursday's 4.6714 close. The 2-year led with a 12.4bp move to 4.356 percent, flattening the 2s10s from 34bp to 27bp on the session. The bear-flattening is the specific shape of a hawkish-Fed repricing rather than a growth-shock; break-evens firmed 2bp while real yields led the move.
  • DXY: 99.34, up 55 pips from Thursday's 98.79. The 99 handle reclaimed and held on close for the first session since August 19. Intraday range 98.68 to 99.48; the pre-speech 9:00-9:55 AM window saw a 22-pip pop, right in the middle of the amplification-versus-fade band the eve-gamma-compression framework projected.
  • Brent CFD spot: $84.28, up 34 cents. Muted; the Iran-Oman corridor headline flow held.
  • USD/JPY: 160.42, up 114 pips through the 160 handle. The MOF verbal-intervention line at 158 that has held through August is now 240 pips distant; the paired trade is back in play.
  • EUR/USD: 1.1586, down 92 pips. Broke through 1.16 for the first close below since August 12.
  • GBP/USD: 1.3506, down 138 pips. Largest single-session move since the July 15 CPI print.

What was wrong in Wednesday's setup

Wednesday's piece pegged the dovish-cut probability at 65 percent, the persistent-split at 18 percent, language-following at 12 percent, and the hawkish-tilt tail at 5 percent. Warsh delivered the 5 percent tail. Two things went wrong with the calibration and they are worth naming honestly rather than being papered over as "the tail hit."

First, the July PCE at consensus was read as removing the hawkish-tilt anchor. It did the opposite. Wednesday's framing treated 3.3 percent Core PCE at consensus and 3.7 percent headline as a print the Fed would look through. Warsh cited the 3.7 percent headline directly in the speech and used the word "concerning" against it, which is not the look-through language the framework assumed the Fed would run. The gap between the market's July-PCE reaction function and the Chair's July-PCE reaction function is where the miss lives. The market was reading a Powell-era reaction function on a Warsh Fed; that reads-through error was baked into every session of positioning between the July 30 SEP and Friday's keynote.

Second, the "quieter Fed" framing had been telegraphed. Warsh's August 5 post-FOMC statement was 130 words. His confirmation testimony from spring 2026 explicitly signaled a break from forward-guidance conventions. The Wednesday piece's language-following tail at 12 percent was too low; a language-quiet Fed is by definition the language-following state, because the market has to price the median dot and the balance-of-risks language rather than press-conference nudges. The correct probability for the language-following state going into Warsh's first Jackson Hole was closer to 25 percent than 12.

The gold call, the 10Y call, and the DXY call were all consistent with each other and all wrong in the same direction. The infrastructure the framework built to prevent that (multiple asset checks, cross-asset consistency, breakout-versus-fade decision rules) was working as designed; the input probability distribution was wrong. That is a calibration error, not a framework error, and it is the kind of error worth stating in these terms rather than absorbing into a "market moved differently than expected" paragraph.

What was right

Three things held. The multi-year-high framework's Phase 3 top-signal read from Tuesday's $4,696 rejection was correct in direction and in magnitude; the $4,700 defense held, the $4,500 to $4,520 back-fill zone was breached inside the four-to-eight-session window the framework projected, and the pattern is now complete. The eve-gamma-compression framework published Thursday morning classified the Friday tape as amplification (55 percent base rate) and specifically forecast the pre-speech 9-9:55 window pop; the 22-pip DXY pop into the 9:55 open was inside the framework's 15-30 pip range for the amplification case. The Jackson-Hole speech framework's six-dimension read on the text pulled the hawkish tilt out of the second and fourth dimensions in the first read; the framework worked, the input probability weighting did not.

The read

The dovish-cut thesis that ran from Aug 5 through this morning is over. The Chair has committed the Fed to a hawkish reaction function on the front-end of the curve and to a quieter communication posture that removes forward-guidance nudges from the toolkit. Fed funds futures now price a 60 percent probability of a September hike, up from roughly zero on Thursday and from a 65 percent dovish-cut base case on Wednesday. The direction of the September FOMC surprise has flipped in a single session.

That flip has consequences that will play out over the next four to six weeks and are worth naming before they arrive. First, the dollar's floor moved up: DXY at 99 is now a floor rather than a ceiling, and the September 17 FOMC becomes the next major test with the cross above rather than below 100. Second, the yen carry trade is back on: USD/JPY at 160.42 through the MOF verbal line puts the intervention question live for Ueda-san by mid-September if the pair reaches 162. Third, gold's Phase 3 has completed downside; the next question is whether Phase 4 (extended decline) or Phase 4' (rebase-and-restart) plays out over the next four to six weeks, and the answer depends on whether Warsh follows through with hawkish action at the September 17 FOMC or the speech was rhetorical high-water. Fourth, the belly of the curve has re-priced by 10 to 12bp; the 2y10y flattening from 34 to 27 puts the recession-signal question back into focus if it flattens further, which it will if a September hike is delivered.

Setup update

The pre-Warsh distribution is retired. The new working distribution starts from the September 17 FOMC as the primary catalyst and Warsh's next scheduled communication (mid-September Congressional testimony, tentative) as the secondary. Hawkish-hike at 45 percent (previously roughly zero). Hold-with-hawkish-language at 35 percent. Dovish-hold at 15 percent (previously the 65 percent base case for a cut, reframed as a hold given Warsh's communication reset). Dovish-cut at 5 percent (previously the base case; retained as a tail against a September labour or growth crack).

Confirmed if: Warsh's Congressional testimony in mid-September holds the "firm and fixed" language on the 2 percent target. The September 17 FOMC delivers a hike or a hawkish hold with 130-word statement language and no forward guidance. DXY closes above 99.50 through the end of next week and above 100 through the FOMC. Gold trades between $4,300 and $4,500 through the FOMC.

Invalidated if: Weekly claims break above 240k for two consecutive weeks between now and the FOMC. August NFP prints below 50k on Friday September 5 with negative revisions to prior. ADP on Wednesday September 2 prints below 40k. Any of those three would give the dovish-hold tail (currently 15 percent) enough cover to move to base case. Two of three would put the dovish-cut tail (currently 5 percent) back in play.

Watch Monday: Month-end Monday August 31. Not a scheduled release day but a heavy month-end FX rebalancing session, with the ECB's monthly reserve rebalance and the standard month-end fixings likely to add EUR-buying pressure into the 4:00 PM London fix. The framework question is whether the Friday post-Warsh dollar bid extends through month-end or whether the month-end EUR-buying provides a counter that keeps DXY in the 99.20-99.60 range. Tuesday September 1 brings ISM Manufacturing (August), JOLTS (July), and Construction Spending as the first pack of the new month; Wednesday September 2 is ADP; Thursday September 3 is ISM Services and weekly Claims; Friday September 5 is August NFP. The invalidation triggers for the new hawkish thesis are stacked into the next five sessions.

Nothing on this site is investment advice or a recommendation to trade. Setups published here are falsifiable hypotheses, not signals.