1The bond market repriced to a hiking cycle, and the argument about why it moved is unresolved
The move is not in dispute. The ten-year Treasury yield closed at 4.95% on 10 September, up from 4.80% two sessions earlier. The two-year went 4.39% to 4.56% over the same stretch, so the front end led. The gap between them narrowed from 0.41 to 0.33 points by Friday. Joseph Wang, on Fed Guy, puts the thirty-year at about 5.35%.
What moved it is disputed, and the two people who spoke this weekend disagree completely.
Wang's answer is oil, and he will not entertain another. He walks through the mechanism: Houthi forces took the Yemeni port of Mocha and reached Dhubab on the Bab el-Mandeb Strait, which is the release valve Saudi Arabia has been using to route barrels around a partly closed Strait of Hormuz. Drones launched from Iraq shut the Saudi East-West pipeline the same week. Refined products followed crude higher — US diesel set a record on Friday, per Reuters. Higher energy costs feed global inflation, central banks respond, and yields rise everywhere. He notes the European Central Bank hiked for the second time this year and is forecast to hike again. Of the competing explanations — the deficit, Treasury supply, AI hyperscalers crowding out the bond market — he says the people selling them "are actually propagandists for the war."
Bob Pozen, on Excess Returns the same weekend, gives one of those competing explanations as his own closing argument. Asked what he believes that his peers do not, the former president of Fidelity says bonds are a bad deal. His reasoning is two-part: bonds do badly against stocks in inflationary times, and "are we in times in which the treasuries are going to have to become a much bigger supply? Yes, because we're running all these deficits." That is the supply story Wang dismisses, stated by someone who ran a trillion dollars.
They reach the same trade from opposite premises, which is why the distinction matters rather than being a debating point. Wang names his own test: war on, yields up; war off, yields down. If he is right, a settlement in the Gulf pulls yields back down. If Pozen is right, the deficits are still there on the other side of a ceasefire.
One detail neither explanation covers. Friday's hot inflation print should have hit stocks, and instead the S&P 500 rose 0.86%, the Dow 0.98% and the Nasdaq 0.96%. Wang says so plainly: "that was really weird to me." He guesses at unwound hedges without claiming to know.
2A September increase is priced at 79% on both venues that trade it, not the 90% in circulation
August core inflation is what did it. Core CPI rose 0.29% on the month against a 0.2% forecast and 2.45% on the year; the headline index rose 0.40% on the month and 3.35% on the year. Governor Waller had said publicly that a core print hot enough to carry into core PCE would move him to a September increase. It was, and the market read his own stated threshold back at him.
Wang puts the resulting odds at "basically 90%." Both venues that actually trade the contract put it at 79%.
Polymarket prices a 25 basis point increase on 16 September at 79.5%, against $141m of cumulative volume and roughly $1.05m traded in the last day; no change sits at 20.5%. Kalshi, which was recorded as unavailable for pricing in recent weeks, quotes the same outcome at 79% bid against 80% offered on $11.7m of volume. Two venues, independently, eleven points below the figure quoted on air.
The four-hike claim needs its window stated to be judged at all. Wang says the market went from pricing two increases to four in a single week, and that this is the real story. Polymarket's calendar-2026 contract puts four increases this year at 0.2% and concentrates 82% of the probability on one or two. Those are not the same question — only three meetings remain in 2026, so a fourth increase inside the year would need a half-point move — which means Wang's path is mostly a 2027 question and the contract cannot test it. The rungs are also thin, $200 to $1,300 traded in a day, so they are weak evidence even within their own window. What the deeper market does support is the direction: an increase at some point in 2026 prices at 89% on $9.2m, up 17.5 points on the week.
A separate contract expects the decision to be contested. Zero dissents prices at 31%, three or more at 28%, on $36k of volume that is too thin to lean on. The hike itself is close to fully priced, so the surprise available next week is in the dot plot and the vote, not the decision.
3The Hormuz deal being negotiated would manage the strait without reopening it
The most violent repricing of the week was not in rates. An Iran-Oman agreement to manage the Strait of Hormuz went from a long shot to the base case: 53% by 30 September, up 37 points on the week, and 68% by 31 October, up 40 points.
A reader could take that for de-escalation. The same market says otherwise. The contract on Hormuz traffic returning to normal by 31 December prices at 19.5% on $11.5m — and it fell 7 points on the week, while the agreement ladder was rising 40. The market is pricing a deal that does not restore the flow.
Reuters reported on Friday what that means in plain terms: the Iran-Oman understanding "does not provide for immediate reopening" of the strait, per Tasnim, and an Iranian official said no signed deal is expected out of Monday's Oman meeting. The near-dated rung moved accordingly — a deal by 14 September fell 11.5 points in the last day to 29%, which is the market marking that briefing.
The physical evidence supports the pessimistic half. Hormuz shipping traffic has fallen to single digits. Saudi oil supply is at a more-than-three-decade low, per the IEA. Tanker rates set record highs. On Fed Guy, Wang adds that strategic reserves have been drawing down for months with no published operational floor, and that Chinese buying has returned. Chevron's chief executive, quoted by Reuters, made the same point about depleted buffers leading to higher prices.
Escalation risk at the second chokepoint is real but not imminent. Bab el-Mandeb being effectively closed by 30 September prices at 5.3%, having fallen 4 points on Friday; by 31 December it is 21.5%, up 5.5 points on the week. The week added closure risk at the longer horizons and Friday took some of it back.
That combination explains an otherwise odd tape. Crude ran about 8% on the week to just above $100 and then fell 2.4% on Friday, with down 2.79% and down 2.20%. Energy equities did not follow: closed up 0.13% at a 52-week high and up 0.32%. A deal that caps the escalation premium without returning barrels deflates the spike and leaves the constraint, and the constraint is what equity holders own.
4Private credit's ratings are set where no published methodology has to stand up
Pozen's hour on Excess Returns is mostly a career interview, and the part worth reading is a disclosure argument that has no price attached to it.
Insurance-company holdings of private credit have gone from a small share two decades ago to as much as $2 trillion. Much of the growth is what he calls affiliated: private equity firms buy insurers — Apollo first, then others including — and place debt they originated, or debt of their own portfolio companies, into the insurer's book. He is careful that the maturity match can be defensible on its own terms, since insurers have long liabilities and private credit is illiquid and higher-yielding.
The problem is how that debt is rated. Public bonds usually carry two ratings with published methodology and assumptions anyone can attack. A large share of private credit is rated instead by small private raters, which he says produces rating shopping. He cites two academic studies, one from Columbia and one from Imperial College, finding the private ratings systematically too generous; correcting them, on his account, would require insurers to hold $400bn to $500bn more capital each year. Layered on top is the absence of a national insurance regulator, so fifty state regimes add jurisdiction shopping to rating shopping.
He points at a live case rather than a hypothetical. Regulators are examining roughly $20bn of private credit held by insurers controlled by Mark Walter and Guggenheim that appears to have been affiliated, some of which he says may have financed the purchases of the Los Angeles Dodgers and the Los Angeles Lakers. Disclosure of affiliation to regulators has always been the rule; the question is whether it happened. "We don't need a blowup for this to happen," he says of reform.
Two smaller observations from the same conversation are worth keeping. On private equity entering 401(k) plans, he is against it as a standalone option and wants it inside target-date funds with a sponsor cap of 10-20%, because a participant facing a medical emergency or a required minimum distribution has no liquidity. He also flags an accounting rule under which a fund buying another fund's portfolio at a 5% discount may immediately mark it up to par — "instant markups," which he wants changed. And on the index complex he notes that , Vanguard, , Fidelity and together hold more than 30% of many companies' shares, with the market-weighted S&P 500 making that self-reinforcing. His worry there is specific and new: the AI data-centre financing has become "a bit circular," with one company committing to take the power and another guaranteeing the lease.
His closing advice is the one place the two videos touch. Bonds, he says, counterbalanced stocks in only 10 of the last 60 years, and in 2022 both fell about 18%. He runs 90% stocks and 10% cash himself. Whether that holds depends entirely on the argument in the first section: it is a very different bet if the ten-year is at 5% because of a war than if it is there because of the deficits.
Monday's Oman meeting is the first test of the deal the market has already bought
The Oman meeting on 14 September is where the 29% rung resolves, and an Iranian official has already said not to expect a signature. Reuters carried a fresh report of an attack on Hormuz shipping early Sunday, which cuts the other way.
The Federal Reserve decides on 16 September with the dot plot attached, and at 79% priced the decision itself is not the event — the projected path is. reports on 16 September and after the close the same day, both read as demand indicators, with on 15 September.
