A dollar supported by real yields, awaiting CPI

Minutes of the 15-16 September FOMC meeting, published on Wednesday, gave the front end what it expected: the hike to 3.75-4.00% was unanimous, and “most participants” considered a further hike “likely appropriate by the end of the year”. The latest verified CME FedWatch readings, between 2 and 8 October, put the probability at 17% for 28 October and around 81% for December. The 2-year yield ended at 4.80%, down 3 bp over the week and 12 bp below its 28 September peak.

Our 8 October Daily linked rising long-term yields and the Fed's hawkish tone to energy inflation; the data contradicts that reading. The 10-year breakeven has barely changed since January (2.36% then, 2.33% now), while the real yield gained 101 bp to 2.91%, just below its September peak (2.93%). Investors are therefore not hedging against an inflation overshoot: they demand a higher real return to hold US debt. This support for the dollar should last longer than an inflation scare, but give way more quickly to bad news on growth.

Auctions showed where buyers were. The Treasury sold $39bn of 10-year notes at 5.300%, with a bid-to-cover ratio of 2.77 versus a 2.52 average over ten auctions, followed by $22bn of 30-year bonds at 5.618%, the highest since August 2000, with a bid-to-cover ratio of 2.54 and 72.3% indirect bidders. The 10-year yield, at 5.31% on 5 October, ended the week at 5.24%. The term premium attracted buyers at these levels, removing the dollar's long-end driver and leaving it dependent on the Fed's path.

That path rests on a weakening labour market. September employment added 29k jobs versus around 90k expected, with 60k of downward revisions to July and August; unemployment rose to 4.2%, and hourly earnings growth slowed to 3.0% year on year, the lowest in the dataset. October Michigan sentiment fell to 46.3, with current conditions at a record low, but households' 5-10-year inflation expectations rose to 3.5%. Hawkish FOMC members draw their argument from this figure more than from wages.

Wednesday's CPI release will distinguish between these two readings. Consensus expects headline inflation of +0.6% month on month and 3.6% year on year, with core at +0.2% and 2.5%; Thursday's retail sales are expected at +0.3% after +1.2%. A disappointment would meet speculative positioning already short 63k contracts in the euro, 91k in sterling and 79k in the Canadian dollar.

On 8 October, Donald Trump wrote that the United States would not attack Iran before the 3 November midterm elections, citing “productive discussions”. The naval blockade of Iranian exports remains in place, and conditions in the strait have not improved: UKMTO reported at least 12 incidents between 28 September and 5 October, IMF PortWatch counted just 4 transits on 4 October versus 85 a day before the crisis, and the EIA mentioned attacks on Saudi Arabia's East-West pipeline, the main bypass route.

The dataset puts EIA spot Brent at $120.03 on 2 October, up 33.77% over one month, and WTI at $98.07. Press reports put Brent futures around $103.5 on Friday; the dates and instruments differ, but such a gap suggests a premium on physical barrels whose exact size we cannot verify. The EIA raised its fourth-quarter Brent forecast by $14 to $105, assuming 4.5 million barrels a day of Middle Eastern production is shut in, while US crude inventories fell by 3.2 million barrels against expectations of a 1.7 million-barrel rise. OPEC+ kept November quotas unchanged on 4 October and will not meet again until 1 November.

Over one month, NZD lost 4.11% and the euro 3.83% against the dollar, while the S&P 500 gained 1.2% over the week to 7,811.54: the decline affects importers' currencies without accompanying equity stress.

The political timetable moves the binary risk to 3 November, after the FOMC, ECB and BoJ, which will decide in late October with expensive oil and a still-closed strait. The IEA quantifies what a diplomatic outcome would cost long-dollar positions: its 1.3 million-barrel-a-day deficit for 2026 would become a 4.6 million-barrel-a-day surplus in 2027 if Hormuz reopened. In our reading, the euro and yen, the currencies of importers, would benefit first.

September flash HICP surprised at 3.8% versus 3.6% expected, the highest since September 2023, with energy at +18.8% and services at 3.2%; core remains at 2.5%. The Reuters poll of 5-8 October captures consensus: 70 of 73 economists expect a hold at 2.50% on 29 October, 64 expect a 25 bp hike in December, and 58% of respondents see the peak at 2.75%. Activity does not justify euro weakness: German industrial production rose 2.0% month on month in August, and institutes raised their 2026 German growth forecast to 1.3%.

The rate differential widened despite the ECB. The 10-year Treasury-Bund spread reached around 182 bp in September, and the US 2-year yield, at 4.80%, exceeds the deposit rate by 230 bp. Two ECB hikes have failed to keep pace with the Fed.

France adds a premium that the ECB cannot offset. The Lecornu government's €43bn savings plan targets a deficit of 5.0% of GDP in 2027, with debt at 121.7% of GDP, interest costs of €91bn and borrowing of €340bn, up 10%. The OAT-Bund spread touched 154 bp intraday on 2 October, its widest since 2011, then 134 bp on 7 October, above Italian and Greek spreads and far from its 75.5 bp one-year average. November's budget vote and the April 2027 presidential election leave this premium in place for several months.

EUR/GBP isolates the euro-specific component. The pair touched 0.8465 on Wednesday and ended at 0.8476, down 0.32% over the week, even as the UK also faces long-term yields at record highs. Traders appear to favour the central bank that will raise rates first, penalising French sovereign risk more heavily than pressure on gilts.

Markets now price a BoE hike to 4.00% on 5 November with a probability of around 83-84%. September's 6-3 vote to hold at 3.75% paved the way, and Megan Greene said on Thursday it was “dangerous to assume markets will do the BoE's work”. Goldman Sachs now expects a November hike, while Citi expects one in the fourth quarter and another in the first quarter of 2027. The BoE bases this tightening on its forecast of CPI around 3.75% in the fourth quarter and above 4% in early 2027.

Observed inflation is less pressing: August CPI is 3.1% and core 2.6%, below a Bank Rate of 3.75%, with unemployment at 4.9%. Like the Fed, the BoE is preparing to tighten against imported inflation that it anticipates more than observes. With a hike priced for 5 November, sterling offers short-term carry that the euro does not offer before December.

The limit will come from the 28 October budget, John Healey's first for the Burnham government, published on the same day as OBR forecasts and the FOMC. The Chancellor rules out increases in income tax, VAT and employee National Insurance, while the 30-year gilt yield exceeded 6% on 1 October, its highest since 1998; Bloomberg Economics estimates that higher yields reduce £23.6bn of headroom by around £12bn. If gilt investors start treating the UK like France, a BoE hike would become a funding cost rather than carry. August GDP, expected at -0.1% on Thursday, will be the first test.