Markets are pricing in rate cuts. The UK Treasury just told them they’re wrong.
When an entity that borrows billions starts predicting higher interest costs, you don’t shrug it off. You read the ledger.
Here’s the setup: A market brief hit the wires claiming HM Treasury expects the Bank of England to hike at least once in 2026. Not 2025. Not a cut. A hike. The exact source is unclear, but the signal is sharp. This isn’t a central banker’s vague guidance—it’s the fiscal arm publishing a forward rate forecast. That’s rare. That’s deliberate.
Context: The Policy Divergence
The consensus going into H2 2025 was that the BoE would begin easing by mid-2026 at the latest. Soft landing narratives dominated. Gilt yields were compressing, sterling was drifting, and rate-sensitive sectors were pricing in relief. Then this.
A Treasury prediction of a 2026 rate hike flips the timeline. It implies the BoE will need to tighten after the current hiking cycle is supposedly over. The underlying assumption: inflation is stickier than the market thinks—likely driven by services and wage growth that refuse to roll over. The Treasury’s message is that they expect the BoE to fire one more bullet, even if it means higher debt servicing costs for the government. That’s a statement of priority: kill inflation, tolerate fiscal pain.
Core: Order Flow and the Repricing Machine
When I built my Python scripts to scan Deribit’s options flow for volatility mismatches, I learned that the market’s biggest moves come from resetting expectations. This is that moment for sterling fixed income.
Let’s quantify the adjustment. If the market had fully priced a 25bp cut in 2026, and now must price a 25bp hike, the net shift in rate expectations is 50bp. That repricing hits the 2-year to 5-year Gilt corridor hardest. Expect a bear flattening of the curve: short-end yields rise faster than long-end as the rate path is pulled forward.
The first signals will appear in the sterling OIS curve. Look for the 1-year forward 1-year rate to trade above 4.00% (assuming current base rate is 5.00%). If that happens, the market is validating the Treasury’s call. If it stalls, the news is noise.
From my own audit of macro cycles: during the Terra collapse, I learned that the majority of traders anchor to the consensus narrative until the first domino falls. Here, the Treasury is the domino. The question is whether the BoE will validate it in the February 2026 Monetary Policy Report.
Contrarian: The Real Bet Is Not on Rates
The obvious trade is long GBP, short Gilts. But the contrarian angle is simpler: the Treasury’s prediction is a political hedge, not an economic forecast.
Think about it. Why would a Treasury that borrows at floating rates signal higher future borrowing costs? Because they know the BoE is going to start QE again in a downturn, and they need to build credibility now. By forecasting a hike in 2026, they are saying to the market: “We are not the weak party. We will not tolerate a wage-price spiral.” It’s a preemptive credibility injection.
The real risk is that the Treasury is wrong. Suppose UK GDP dips in Q4 2025 and core inflation collapses to 1.5%. Then this prediction becomes an anchor that forces the BoE into a policy error—keeping rates too tight for too long. The contrarian play is to fade the initial GBP spike and instead buy protection on long-end Gilts via receivership swaptions. If the data turns, the Treasury’s signal becomes a trap.
Takeaway: The Levels That Matter
Watch the 10-year Gilt yield at 4.20%. If it breaks above 4.35% on this narrative, the repricing is real. If it stays below 4.10%, the market is not convinced.
For sterling, a close above 1.28 against the dollar would confirm the hawkish rerating. Below 1.2750, the market smells bluff.
The final thought: The Treasury just showed their hand. But in this game, the best trades are the ones where you wait for the second signal—the BoE’s actual vote count. Until then, treat this as a coded message, not a trade ticket.
When the code bleeds, the ledger keeps the truth.
Arbitrage is just violence disguised as math.
black box
