The Daily Brief

The consensus that shifted this week was the Governor's. Speaking to lawmakers on Tuesday, Andrew Bailey pushed back on the idea that another increase is simply a matter of timing, saying decisions would follow the economic and geopolitical evidence rather than run ahead of it. Within a day, Megan Greene, who voted to hike in July, warned that a prolonged oil-price shock could make inflation expectations persistent. That is not a committee walking together toward December, it is a committee arguing about what kind of shock this is, and the argument got harder overnight when Brent cleared $102 and UK natural gas held its highest level since late 2022.

The curve has not been rewritten, but the cost of waiting has been repriced. Yesterday's session left the ten-year gilt around 5.19%, and the overnight marks have it at 5.27%, roughly eight basis points higher and close to a nineteen-year high. That is a meaningful break, because it lifts the pressure out of the very long end, where it has sat since the summer, and moves it into the part of the curve that prices the Bank rather than the Chancellor. A market comfortable that the 17 September hold is safe does not pay eight more basis points for ten-year money overnight.

The long end is what makes it expensive rather than merely awkward. Tuesday's thirty-year auction cleared at 5.8168%, the costliest long-dated gilt the Debt Management Office has sold since it was founded in 1998, and the arithmetic that follows is now widely put at roughly £13bn of remaining fiscal headroom into the 28 October budget. The Office for Budget Responsibility locks its forecast window next month, which converts a market level into a policy commitment: every week yields hold here, a larger share of the budget gets written by the gilt market rather than by the Treasury.

Sterling declined the whole story. Cable finished yesterday's session at 1.3554, up 0.10%, roughly a third of a percent firmer over four weeks and broadly flat over twelve months, which is a currency handed a hawkish repricing and refusing to take the money. UK equities are indicated sharply lower this morning, the FTSE marked around 10,670 and about 1.3% below yesterday's close, with France down 1.9% and Spain 1.5%. The pound is not trading the rate path because the rate path is insurance against an imported price rather than a return on domestic strength, and the equity market is pricing precisely that cost side.

At 1.3554 cable sits 0.8% below the 1.3661 ceiling it has not cleared all year, 2.1% above the 1.3279 floor of the range held since late July, and effectively on the 1.3543 quarter estimate with a twelve-month path at 1.3812. The interest-rate leg leans mildly in sterling's favour and is well cushioned by a hike already fully priced for December. The risk is not the MPC. It is that another auction like Tuesday's, or an OBR window that locks headroom near zero, converts a fiscal-credibility story into a funding story, and that is the route by which 1.3400 comes back into range faster than the rate differential alone would justify.

The auction went well. Treasury sold $39bn of ten-year notes at 4.834%, the highest yield a ten-year has carried since 2007, and demand ran at 2.71 times the amount on offer, the strongest cover in years, with indirect bidders taking 79.2% of the issue. The paper cleared more than a basis point through where it traded into the sale. On the specific test flagged in yesterday's brief, whether anyone wants duration at these levels, the answer came back comfortably yes.

The yield rose anyway. The ten-year is at 4.845% this morning after touching 4.853% intraday, its highest since November 2023, because Treasury put a figure on its long-dated buyback and the figure disappointed: $6bn of ten to twenty-year paper, triple the size of the last long-dated operation, against a market positioned for considerably more. The two events are connected in an uncomfortable way. Part of why the auction cleared so strongly is that bidders knew they could sell into today's buyback window, which means demand that looked like conviction was partly a funding trade. Appetite passed the test; supply management did not.

The dollar is collecting on none of it. The index sits at 98.739, marginally lower again and still at its weakest in four months, with Fed pricing unchanged at roughly 60% for a hike next week. The reason is structural rather than domestic: the ECB decides at 13:15 BST today and the Bank of Japan on 17 and 18 September, and the yen near 153.35 is holding close to a seven-month high. Between them the euro and the yen are 71% of the DXY basket, so the majority of the index is currently repricing foreign policy rather than American policy, and a currency can lose ground on a rate differential that narrows from the other side.

Today's two prints land in the same minute and pull against each other. Producer prices at 13:30 BST are forecast at 5.3% on the year against 4.7%, with the core at 4.6% against 4.2%, a set economists read as implying a 0.3% monthly core PCE against 0.2% in July. Jobless claims at the same moment carry a 205,000 consensus against 206,000, still within touching distance of the near sixty-year low set in July. A labour market that will not crack, sitting alongside a producer price series about to absorb $100-plus crude, is the specific combination that keeps next week's decision genuinely live rather than pre-committed.

At 98.739 the index is 0.32% under the 99.055 quarter estimate, roughly 1% weaker on the month, 1% firmer on the year, and about 4% below June's thirteen-month peak, with the twelve-month path at 97.44. The setup is unusual for a currency days away from a hike: the yield leg is at a three-year high while the currency is at a four-month low, which is what happens when the rate advantage erodes abroad rather than at home. PPI this afternoon and CPI tomorrow, with consensus at 3.4%, decide whether the index reclaims 99 or the four-month low becomes a floor that gets tested rather than defended.

At midday the domestic economy files the report that Tuesday's GDP print made unavoidable. Manufacturing production for July lands at 12:00 BST, forecast at 0.2% on the month and minus 0.2% on the year against minus 1.7%, with mining production earlier in the morning. These are the two industries named as the largest drags on the second-quarter contraction, mining down 3.0% and manufacturing 1.8%, so the July numbers are the first monthly evidence on whether the disruption was one bad quarter or the beginning of a trend. The mechanism the national accounts exposed is worth holding on to: output fell while metal prices were strong, because the accounts measure volume, and rail and port capacity cap how much ore actually reaches market.

The currency, meanwhile, has already handed back what it took. USD/ZAR touched the low 15.99 area in yesterday's session, its strongest in six months and within a whisker of the 15.93 low set in late August, and it has since traded back up to 16.0374, a quarter of a percent weaker overnight. That is the second failed attempt at 15.93 in a fortnight. A level rejected twice inside two weeks stops being a waypoint and starts being a floor.

What rejected it is the barrel. Brent at $102.73 is up around 15% on the month, and for a net energy importer with a fuel-levy pass-through that reaches the consumer index within weeks, that is a direct hit to the terms of trade that the metals complex has to work harder to offset. Gold at roughly $4,414 and silver at $67.5 both recovered overnight from the sell-off that funded Tuesday and Wednesday's move into crude, and copper is holding near record highs, but the export side is now running to stand still. The domestic bond market has drawn the same conclusion, with the ten-year at 8.845%.

That leaves the 23 September meeting in a materially worse position than a week ago. August inflation prints the same morning as the decision, forecast to rise to 4.7% from 4.3%, and the fuel component now has a much higher barrel sitting behind it. A 25 basis point increase to 7.25% is priced, and Lesetja Kganyago has signalled the committee can respond cautiously to the latest shock while holding the 3% target. Hiking into an imported price, against a contracting quarter and a long end already charging more for duration, is the least comfortable version of that call, and the currency is no longer supplying the cushion it did on Monday.

At 16.0374 the rand is 0.7% weaker than both the 15.93 six-month low and the 15.9281 quarter estimate, and 5.5% stronger than the 16.98 top of the band traded since early August, with the twelve-month path at 15.3654. The asymmetry has flipped inside forty-eight hours. A week ago the question was whether the rand could clear 15.93; it is now whether 16.00 holds as support, and the answer rests entirely on metals and a soft dollar rather than on anything the production prints will say this morning. Tomorrow's US CPI and the 23 September decision are the two events that resolve it.

At 13:15 BST the ECB becomes the first of three major central banks inside eight days to set a policy rate against $100-plus oil. A quarter point on both the main rate, to 2.65%, and the deposit rate, to 2.50%, is 99% priced, and all 65 economists in the most recent Reuters survey expect it. Which is exactly why the decision is not the event. The updated staff projections and the 13:45 BST press conference are, because that is where the ECB has to state what it thinks the shock does next.

The disagreement underneath is unusually clean. 91% of surveyed economists expect the deposit rate to finish 2026 at 2.50%, and 78% expect it still to be there through the middle of 2027. The market prices a second hike this year and 3.1% by late 2027. Roughly 60 basis points of daylight, and every one of them is a judgement about whether energy stays in the price level or moves into wages. Euro-area inflation at 3.3% in August against 2.9% in July is why the question is live; core running well below headline is the reason economists think it resolves itself.

History is in the room. If the economists are right, this becomes the ECB's shortest tightening cycle since 2011, when it raised twice into an oil spike and later treated the episode as a policy error, and that precedent is doing real work in the arguments for stopping here. The counter-argument sits in the forward curve. Brent is at $102.73, up around 15% on the month and roughly 55% on the year, at its highest since May, with European gas at fresh three-and-a-half-year highs and no visible ceiling while strikes on export infrastructure continue. A central bank that assumes the shock fades is publishing an energy forecast, not a monetary one.

The cross-asset tape is reading cost rather than growth, and it is splitting by market. Germany is the outlier, the DAX marked at 25,597 and close to unchanged, while France is indicated about 1.9% lower and Spain 1.5%, the ASX closed roughly 2% down and the Nikkei fell for a third session. Bund yields at 3.44% are eight basis points higher and near fifteen-year highs going into the decision, French ten-years at 4.33%, and the US ten-year at 4.845%. Gold and silver, sold on Tuesday and Wednesday to fund the move into crude, have turned back up. That reversal is the tell: the market spent two days hedging the barrel and is now hedging the policy response to it.

At $102.73 Brent is 5.5% above the $97.41 quarter-end estimate, against a twelve-month path of $113.59 and an all-time high of $147.50 from July 2008. The level has already done what yesterday's brief said $100 would, which is reopen the assumptions: today's ECB projections, the Bank of England on 17 September and the Fed next week all have to publish a view built on a number that moved after their books closed. The asymmetry is no longer about whether the premium unwinds. It is that three institutions will set policy off three different assumptions about a price none of them controls, and the divergence that creates is what EUR/USD and GBP/USD trade on for the next eight days.