Inside Warsh's Jackson Hole Speech
The data came in line — so why did the odds jump?
① At the Jackson Hole symposium on August 28, 2026, Fed Chair Kevin Warsh delivered his debut keynote, and the September rate-hike probability jumped from 35% to 57% in a single day.
② Yet the data point widely expected to move that number — July PCE — came in in line with consensus. What moved the odds wasn't a new number; it was the Chair's tone.
③ Even among stocks that beat earnings, the reaction split — richly valued chip stocks sold off while cash-rich software names rallied. A sign the market started re-litigating "quality of earnings" the moment rates were back in play.
🎤 What happened — Warsh's Jackson Hole debut
The Jackson Hole Economic Symposium is the Fed's annual policy gathering held every August in Wyoming, where the Chair's keynote has effectively served as a preview of the next FOMC meeting. On August 28, 2026, just past his 100th day in office, Chair Kevin Warsh delivered his first keynote there.
The message was unambiguous. Warsh declared that "the Fed's foremost task right now must be prices." While saying he was "deeply impressed" by the broad strength of the US economy, he repeated concern that inflation's "underlying trend" had not improved, and left the door open to further tightening with the phrase "there is more work to do."
| Metric | Value | Vs. Fed target |
|---|---|---|
| July PCE (YoY) | 3.7% | Well above the 2% target |
| July core PCE (YoY) | 3.3% | |
| July CPI (YoY) | 3.4% | |
| Unemployment rate | 4.1% |
Notably, the speech itself contained no specific hike timeline or firm decision. Warsh distanced himself from traditional forward guidance, saying "I stand here today committed to a discipline, not to a decision." The market interpreted it as hawkish anyway — why, is what §5 works through.
📊 The market reaction, in numbers
The reaction was immediate and broad-based. Indices had been up as much as 0.5% at the open, but every bit of that gain evaporated by the close.
| Asset | Move | Reading |
|---|---|---|
| September hike odds (CME FedWatch) | 35% → 57% | The debate shifted from "cut vs. hold" to "hold vs. hike" |
| Kalshi prediction market | Up to ~48% | Lower than FedWatch, same direction |
| 2-year Treasury yield | +8bp | The maturity most sensitive to policy rates moved first |
| 10-year Treasury yield | +5bp, closed 4.72% | Bear flattening — the front end drove the hawkish repricing |
| Dollar Index (DXY) | 99.68, +0.53% | Higher hike odds fed straight into dollar strength |
| Gold | $4,505/oz, -2.27% | Rising real yields plus a stronger dollar reversed a recent rally |
| S&P 500 | -0.25% | +0.5% at the open → -0.25% at close, reversing in a single session |
| Nasdaq Composite | -0.52% | Tech (XLK) was the weakest sector, -1.55% |
Gold's reaction stands out. A drop of over 2% in a traditional safe haven signals the market read this hawkish repricing as rising real yields, not rising uncertainty. Higher real yields make holding a zero-yield asset like gold relatively less attractive.
💸 Why chips fell hardest — earnings quality is back on trial
What stands out most is the timing colliding with earnings season. The day before, on August 27, Nvidia jumped 7.63% overnight, breaking a four-quarter pattern of falling after beats. One day after the Jackson Hole remarks, a large share of that gain was given back.
| Stock | Aug 28 move | Note |
|---|---|---|
| Nvidia (NVDA) | -4.57% | Closed $217.55 — gave back much of the prior day's rally |
| Arm Holdings | -6.33% | Closed $239.05 |
| Intel | -2.85% | — |
| AMD | -2.33% | — |
| Broadcom | -0.74% | — |
| Marvell Technology | -9%+ | EPS $0.94 (beat $0.93), revenue +37% — fell anyway |
Marvell is the clearest example. Adjusted EPS of $0.94 beat the $0.93 consensus, and revenue grew 37% year over year — yet the stock dropped more than 9%. Analysts pointed to "a steep run-up going in" and "caution on richly valued names amid rate-hike concerns." As covered in how to read an analyst report, what mattered wasn't the beat itself but how much of it was already priced in.
The same day, ServiceNow (NOW) +4.54%, Amazon +3.97%, Microsoft +1.68%, and Alphabet +1.74% held up well among large software and platform names. Elastic beat adjusted EPS of $0.70 against a $0.58 consensus and raised guidance, jumping more than 18%.
The split between chips (high-P/E assets that lean heavily on future capex and growth expectations) and large software with already-realized cash flow reflects a simple principle: higher rates discount far-future earnings more heavily. marketbrief's editor note summed up the day: "Even good earnings offer no shelter in front of a hawkish Fed."
🔍 Was this scenario expected? Yes — but the trigger wasn't
This wasn't entirely out of the blue. The CPI/PPI analysis published August 14 already laid out three scenarios, and "Scenario 2 — hike" came with an explicit condition.
"The key is the 8/26 PCE print. If this month's core PPI surprise flows through to a core PCE reading above consensus, the three FOMC members who already voted for a hike in July gain leverage, and September hike odds could jump back above 50%."
Here's the paradox worth flagging. The 8/26 PCE print was never a surprise. Core PCE came in at 3.3% year over year, in line with consensus, and all three indices traded in a narrow band within 0.3% that day. The pathway Scenario 2 required — a PPI surprise flowing through to PCE — never happened.
And yet, two days later, hike odds landed exactly in the "above 50%" range (57%) that scenario anticipated. The trigger wasn't the data — it was the speech itself.
"A number surprised, so the odds jumped" and "the number was unchanged, but a speech moved the odds anyway" are entirely different signals. In the first case, a single upcoming data point could reverse things. In the second, the Chair's own inclination in reading the data becomes an ongoing variable that keeps shaking the market regardless of any one print. Given Warsh has repeated "inflation is still too high" since taking office, this Jackson Hole reads less as a new concern than as a reaffirmation and amplification of an existing one.
📅 What's next — through September 4 and the September FOMC
Two major data points remain before the September FOMC (15–16).
| Date | Event | Why it matters |
|---|---|---|
| 2026-09-04 (Fri) | August Jobs Report (NFP) | Consensus +50K (prior -23K). For Scenario 3 (the cut tail risk) to gain traction, a sharp cooling signal is needed — the first major data point after Jackson Hole |
| 2026-09-10 (Wed) | August PPI | Whether July's PPI surprise carried into August |
| 2026-09-11 (Thu) | August CPI | The last CPI before the FOMC — does core CPI keep cooling |
| 2026-09-15–16 | September FOMC + dot plot | The actual call among hike/hold/cut, plus a hint of the 2027 path |
The September 4 jobs report is the next swing factor. At Jackson Hole, Warsh stressed inflation concerns while also saying he was "deeply impressed" by the economy's overall strength — if the jobs data comes in meaningfully weak, the Fed's two mandates, prices and employment, start to conflict. Should that happen, the hike odds that reached 57% here could retreat. This piece's conclusion isn't "a hike is coming" — it's that the direction stays open until the next data point.
🛡️ What this means for your portfolio
The principles from the August 14 CPI/PPI piece — favor cash-flow-strong companies and dividend payers, re-check valuations, buy in tranches — still hold. A few things worth adding now, after living through the Jackson Hole reversal.
| Check | Detail |
|---|---|
| Exposure to high-multiple growth | As long as a hike scenario stays live, high-P/E, no-dividend names sensitive to the discount rate can keep getting shaken — chips were this quarter's example. |
| Re-check earnings quality | "Good earnings" alone isn't enough; check how much was already priced in and how certain the cash flow is — see what hides behind the price target. |
| Currency buffer | The won-dollar rate actually fell 7.82 won to 1,375.67 that day, cushioning won-denominated losses even as the S&P fell. But a sustained hawkish stance could push the dollar higher again — a double-edged factor. |
| Cash cushion into 9/4 and 9/15–16 | Keep some dry powder for volatility ahead of the next major data points. |
Check how much your own portfolio leans on the high-multiple growth and chip names shaken this session at the portfolio check. For a broader read on rate sensitivity by sector, sector rotation walks through eleven sectors' data.
The core fact from this Jackson Hole isn't "a hike is now confirmed" — it's that the odds moved on words, not on data. The next data point (the Sept 4 jobs report) could reverse this move or confirm it. This is a moment to prepare for volatility until that print, not to call a direction.
After the September 4 jobs surprise and the September 11 core-CPI surprise, hike odds climbed into the 90s%, and on September 16 the Fed raised its rate 25bp to 3.75–4.00% (unanimous, 12-0). At the press conference, Chairman Warsh reaffirmed the same hawkish stance from Jackson Hole with a new phrase — "a timelier return" — and the Dow dropped 631 points right after he said it. The full dot plot and a timeline of his remarks are covered in Inside the September FOMC.
※ Written August 29, 2026. Based on Chair Warsh's Jackson Hole speech (August 28, 2026, per the Federal Reserve's official transcript) and marketbrief daily briefing market data (August 26 and 28, 2026). Rate-hike probabilities are point-in-time snapshots from CME FedWatch and the Kalshi prediction market and will keep changing as new data lands. The September FOMC's actual decision may differ from the scenarios discussed here. This content is informational and is not a recommendation to buy or sell any security or asset. Investment decisions and their consequences rest with the investor.
※ This report is provided for informational and educational purposes only and does not constitute a recommendation to buy or sell any security.
New research, when it lands
Subscribe and the next deep dive comes to you, along with the daily market brief — free, unsubscribe anytime.