Four Outlets Published Four Different Odds of a September Fed Hike Yesterday. The Two-Year Treasury Did Not Move at All

The most consequential number in global markets this week is not a bitcoin number. It is the probability that the Federal Reserve raises interest rates on 16 September — a decision that moves mortgages, savings rates, currencies and every risk asset on earth, bitcoin included. Fed funds currently sit in a range of 3.50% to 3.75%, and until Friday almost nobody expected them to move up.

Then Fed chair Kevin Warsh gave his first Jackson Hole keynote and said inflation data are “more concerning” than the labour market, citing PCE inflation at 3.7% against a 2% target, and noting that more than half of the goods and services the government tracks rose 3% or more over the past year, against roughly a third in the two decades before the pandemic. Markets repriced immediately. What they repriced to depends entirely on which outlet you read.

The same tool, four numbers, seventy-two hours

An 8.6-point spread, one underlying tool, and no contradiction between any of them. FedWatch is a live instrument: it recomputes from fed funds futures continuously, so a figure quoted at 06:42 UTC and a figure quoted eight hours later are simply two observations of a moving series. The failure is not in the numbers. It is that most of the coverage printed the probability without the timestamp, which converts a live reading into a claim about the world — and then readers encounter two such claims and conclude somebody is lying.

This desk has now made a version of this error twice in three weeks, both times with figures pulled at different depths or different hours, and the mechanical fix we committed to on 31 August applies here without modification: a moving number ships with the clock that produced it, in the same sentence. So: 66.1%, CME FedWatch, as quoted by CNBC on Monday 31 August 2026. By the time you read this it will be a different number.

The instrument that has to actually pay did something else

FedWatch is a derived probability. The two-year Treasury yield is a price at which real money changes hands, and it is the cleanest single read on where the market thinks policy is going over the next couple of years. Here is what it did, on US Treasury CMT par yields — the official series, published daily.

US Treasury daily CMT par yields, own pull, 1 September 2026.

The two-year did its repricing on Friday, the day of the speech: +14 basis points, from 4.20% to 4.34%. That is a real and substantial move and it is exactly what you would expect from a hawkish surprise. Then on Monday, the session in which the published hike probability went from the high fifties to 66.1%, the two-year did not move by a single basis point.

What moved on Monday was the long end — the ten-year up 2bp and the thirty-year up 3bp to 5.25%, a fresh high for this move. That is the opposite shape from what a rising near-term hike probability implies. A market genuinely raising its odds of September tightening lifts the front end and, if anything, flattens the back. Monday steepened.

One reconciliation note, because two correct numbers are circulating for Friday’s two-year. CNBC reported it “jumped more than 6 basis points higher at 4.298%” — a live market yield quoted intraday. Treasury’s CMT series puts 28 August at 4.34% — a par yield struck at the 3:30pm close. Both are right; they are different instruments measured at different moments, and the direction is identical.

The people who think the whole thing is overdone

A hike at 66% is a lean, not a lock. Jim Bianco of Bianco Research put it plainly on X, quoted by CoinDesk: “The next Fed meeting is a lean hike not a done deal.” The threshold at which the Fed tends to validate market pricing rather than surprise it is usually put somewhere between 60% and 70%; the 90%-plus readings that make a move a formality are not present here. ABN AMRO Investment Solutions and Brandywine Global Investment Management have taken similar positions.

The more interesting argument comes from Robin Brooks of the Brookings Institution, formerly chief economist at the IIF, who wrote that a September hike, if it happens, “will happen to anchor the 10-year yield and avoid a repeat of the bond market sell-off after July 29. Its purpose will therefore be the opposite of what a traditional hike aims to do.” On that reading the Fed would raise the policy rate specifically to keep long-term financing conditions loose — a credibility purchase rather than a tightening. If Brooks is right, the thirty-year at 5.25% and climbing is not a side effect of Monday’s repricing. It is the reason for the hike.

We have no view on which reading is correct and we are not going to pretend to one. What we can say is that these are two testable stories with opposite implications for the long end, and the thirty-year yield over the next three weeks will separate them without anyone needing to argue.

What bitcoin actually did

The reflexive claim is that rate hikes are bad for bitcoin: higher dollar yields raise the opportunity cost of holding a non-yielding asset, and a firmer dollar pulls capital toward Treasuries. The mechanism is real. The tape is less obedient than the mechanism.

Bitcoin fell roughly 3% on Friday to under $77,000 — the Bitstamp low was $76,876.88 on 28 August — which is a clean, attributable reaction to the speech. But it then closed August at $78,571.17, finishing the month +25.078% from the 1 August open of $62,817.96, the second-largest August in its recorded history behind 2017’s +65.78%. And it has opened September green: $78,979.03 on Bitstamp at 06:13 UTC on 1 September, +0.519%, with Binance BTCUSDT at $79,014.01, up 1.313% over 24 hours.

So the honest summary is: bitcoin took the hawkish surprise on the day, gave up about a third of a percent of the month to it, and then went back up. A 25% month that ends with the Fed more likely than not to tighten is not evidence that rates do not matter. It is evidence that in August they mattered less than whatever else was bidding — and the ETF tape says a good deal of what else was bidding was $3.54 billion of net ETF inflow, 87.30% of it through a single fund.

The two dates that decide it

Friday 4 September, 08:30 ET: the August employment report. Consensus is roughly +55,000 payrolls after July’s −23,000, with a forecast range running from +102,000 to −25,000 — a spread wide enough that the print will move the odds in one direction or the other regardless of what it says. Warsh explicitly ranked inflation above the labour market as his concern, which means a weak jobs number is a smaller obstacle to a hike than it would have been under a different chair.

Wednesday 16 September, 14:00 ET: the decision itself, at the close of the 15–16 September FOMC meeting, accompanied by a Summary of Economic Projections and a new dot plot. The dots will matter more than the decision. A 25bp hike delivered alongside dots showing no further increases is a very different event from the same hike alongside dots showing two more.

The marker

One falsifiable claim, as standing practice. H3: the US 30-year Treasury CMT par yield closes at or above 5.25% on the first business day after the FOMC decision — Thursday 17 September 2026 — settled on the US Treasury daily yield curve series. The bar is Monday’s reading of 5.25%, unchanged. This is a direct test of the Brooks thesis: if a hike is being delivered to anchor the long end, the thirty-year should not be making new highs after it. If it is at or above 5.25% on the 17th, the credibility purchase failed or was never the point.

Method: prices, funding, open interest, basis and on-chain figures in this article are pulled directly by Bitcoin Mastery at the timestamp stated — Bitstamp BTC/USD daily candles for closes and monthly returns, Binance BTCUSDT spot and perpetual for intraday and derivatives, Binance COIN-M quarterly contracts for basis, mempool.space for difficulty, hashrate and fees, alternative.me for the Fear & Greed series, Farside Investors’ all-data table for ETF flows and US Treasury CMT par yields for rates. Volume-weighted average prices are computed from Binance daily quote volume divided by base volume. Where a third-party figure is cited we name the source and its date; where two sources disagree we print both. Every streak or extreme figure is published with the first date of its series in the same sentence.

Disclaimer: This article is for informational purposes only and does not constitute investment advice. Cryptocurrencies are volatile and you can lose money. Nothing here is a recommendation to buy or sell any security, including MSTR, STRC or any exchange-traded fund. Do your own research and consult a licensed financial advisor before making investment decisions.