Wider than the war
Map the war and the Fed out to 2028 and the 10-year yield centres near 5.1%. History says the road around that middle is three points wide.
On Thursday 1 October the 10-year US Treasury yield touched 5.34%, its highest since 2002. It closed the week at 5.28%. A fortnight ago, in The Cushion, I showed that move was the price of money, not an inflation scare.
Then Friday’s jobs report landed: just 29,000 jobs added in September, unemployment up to 4.2%. Bonds rallied, then gave it all back. A cooling job market against a war-driven oil price: that is the tug-of-war behind this note.
The question here is simple. Where does the 10-year sit two years from now, and does the rest of the G7 go with it?
- OfficialStraight from the source: central banks, the Treasury, the Fed’s yield history.
- DerivedArithmetic or simulation on other inputs.
- JudgmentMy call: the paths, their odds, and where each one lands.
The roadsThree roads, then seven paths.
The war sets the oil price, and oil sets the inflation backdrop, so the map starts there. As in The Cushion: a deal at 35%, a grinding stalemate at 45%, escalation at 20%. Each road then forks on what US jobs do and how the Fed answers. The Fed hiked in September to 3.75–4.00% and sees rates near 3.9% in 2028; Friday’s jobs number is already testing that.
The likeliest path, about one in five, is today carried forward: sticky inflation, one more hike, the 10-year around 5.25%. A deal plus cooling jobs points toward 4%. Escalation plus a Fed that keeps hiking points toward 6%.
The widthTwo years is a long time.
Those centres are only half the answer. Since 1962, starting from a 10-year between 4.5% and 6%, the middle half of two-year moves ran from about 0.9 points lower to 0.9 points higher. One time in five it moved more than 1.5 points.
So each path is spread around its centre by that much, and the result is the chart at the top: a middle near 5.1%, and eight times in ten a 10-year somewhere between 3.6% and 6.6%.
The uncomfortable part: the war and the Fed explain only about 30% of where the 10-year ends up. The rest is what nobody can name yet: a surprise recession, a fiscal shock, a buyers’ strike.
The war tells you where the middle is. History tells you how wide the road is.
The methodHow the odds were built.
Path odds (judgment). War odds times the odds of each job-market and Fed response.
Centres (judgment). The Fed rate in each world plus a term premium, smaller in calm paths, bigger in stressed ones.
Term-premium bump (judgment). An extra 0.40–0.75 points, likelier when things go badly: 15% in a deal, 35% in a stalemate, 45% in escalation.
Width (calibrated). Each path spread as a bell curve about 0.94 points wide, so the whole map moves as much as two-year yields have historically. A million random draws give the odds.
| Path | War | Then | Odds | 10-yr centre | Bump |
|---|---|---|---|---|---|
| A1 Soft landing | 35% | 60% | 21% | 4.50% | 15% |
| A2 Jobs recession | 35% | 40% | 14% | 4.00% | 15% |
| B1 Sticky inflation | 45% | 50% | 22.5% | 5.25% | 35% |
| B2 Stagflation-lite | 45% | 30% | 13.5% | 5.50% | 35% |
| B3 Fed cuts into inflation | 45% | 20% | 9% | 5.38% | 35% |
| C1 Fed hikes into the shock | 20% | 35% | 7% | 6.00% | 45% |
| C2 Demand destruction | 20% | 65% | 13% | 4.63% | 45% |
| 10-year in Oct 2028 | This map | History from 4.5–6% | Since 1990 |
|---|---|---|---|
| Below 4.75% | 38% | 33% | 58% |
| 4.75 to 5.5% | 25% | 18% | 22% |
| Above 5.5% | 37% | 49% | 20% |
The extra half-pointThe risk the Fed doesn’t control.
The 10-year sits about 1.4 points above the Fed’s rate. Part of that is where traders think the Fed is going. The rest is the term premium: extra pay for lending to Washington for ten years. It can rise without the Fed doing a thing.
The usual triggers: a midterm result on 3 November that points to bigger deficits, weak Treasury auctions, a fight over the Fed’s independence, or Japanese money staying home. They tend to arrive with bad news, which is why the bump is three times likelier in escalation than in a deal.
The rest of the G7Oil matters more than the Fed.
Every G7 long rate sits more than a point above its central bank’s rate. The ECB hiked to 2.50% on 10 September, the Bank of Japan to 1.25%, the Bank of England held at 3.75%, and Canada sits alone at 2.25%.
Most of these markets take their first cue from Treasuries, then add their own story. For most of them, that story is oil.
Canada moves most like the US, roughly three-quarters as much, and yields more than a point less. If the US slows, Canadian bonds should rally harder. If US inflation sticks, they follow, but not as far.
Britain is the most exposed to oil. A quarter of its debt is linked to inflation, and 30-year gilts are at their highest since 1998. Escalation hurts gilts more than Treasuries; a deal helps them more.
Germany sits in between: euro-area inflation is 3.3%, and record borrowing of €525.5 billion means Bunds won’t rally as hard as Treasuries in a deal.
Japan follows its own central bank, but it matters to everyone. With its own 10-year near 3%, Japanese investors have less reason to buy foreign bonds. If that money stays home, Treasuries and gilts lose a big buyer at once.
The other sideWhere this could be wrong.
History isn’t this war. The width comes from 64 years that include the 1970s and a 30-year bond rally. The real road could be narrower, or wider.
I’m gloomier on jobs than the Fed. My paths average a Fed rate near 3.6% in 2028, below its own 3.9%. Two strong jobs reports would tilt the map higher.
The market has its own odds. Treasury options price all of this. Where they disagree with mine, the market is the one taking money for it.
Five tellsWhich way the road is leaning.
Weight shifts toward escalation and the higher half of the map.
Points to the lower half, whatever oil does.
The market doubting the Fed. Long yields stay high even as rates fall.
The term-premium bump switching on.
The Fed’s answer to the jobs data, then what Washington plans to borrow.
The middle is a guess. The width is the forecast.
The 10-year centres near 5.1% in two years, and eight times in ten lands between 3.6% and 6.6%. The rest of the G7 mostly follows, with oil deciding by how much. Don’t ask where the middle is. Ask which way the road is leaning.
The receiptWhat’s real, what’s noise, and what we still don’t know.
What’s real
- The 10-year closed the week of 2 October at 5.28%.
- The Fed, the ECB and the Bank of Japan all raised rates in September.
- US hiring slowed to 29,000 in September.
What’s noise
- “Yields will settle around 5.1%.” That’s the middle of a road three points wide.
- “Global yields just follow Treasuries.” Outside Canada, oil matters more.
What we still don’t know
- Whether the war ends in a deal, and when.
- How the Fed reads the jobs data on 28 October.
- What Washington borrows after 3 November.