Long Gilts Have Lost Both Of Their Sellers And The Curve Has Not Noticed

The headline from Threadneedle Street this week was that the Bank of England refused to follow the Federal Reserve. Rates held at 3.75 percent, three of nine votes for a rise to 4 percent, and a Fed that had gone the other way the day before. That is the story that ran. It is the wrong one.

The important decision on 16 September was not about Bank Rate at all. It was about the balance sheet, and its effect is this: the Bank of England has removed itself as a seller of long-dated gilts into the private market, permanently, and the gilt curve has priced perhaps a fifth of what that is worth.

Combine it with what the Debt Management Office has already done to its issuance mix and you arrive at a conclusion most people have not yet drawn. Britain has, within a single fiscal year, retired both of the institutions that were supplying long duration to the market. The 30-year gilt fell about 12 basis points on the news, to 5.7415 percent. It should have fallen considerably more, and over the coming quarters it will.

Read the announcement backwards

The framing the Bank chose was tightening. The MPC set out a multi-year path to reduce the stock of gilts held for monetary policy purposes to zero, through annual sales of around 20 billion pounds alongside maturing gilts, running to 2034. Zero is an austere-sounding number. Anyone skimming would file this under quantitative tightening, continued.

Now look at the composition, which is where the substance lives. Of the roughly 488 billion pounds of gilts in the Asset Purchase Facility, the Bank Executive has decided that 222 billion pounds of holdings maturing before 2035 will simply be held to maturity. A further 120 billion pounds of the longest-dated gilts will stay in the facility indefinitely, held to maturity for the purpose of indirectly backing banknote issuance. Neither tranche will be sold to anybody.

That leaves 146 billion pounds, in purchase-proceeds terms, of gilts maturing between 2035 and 2049. This is the only part of the stock the Bank still intends to shift. And it does not intend to shift it to you. The Bank has engaged with HM Treasury and the DMO on a model under which the MPC's decision would be implemented through sales from the facility directly to the Government, at market prices, rather than through the Bank's own auctions. Those auctions are paused in the meantime, with a review of progress before April 2027 so that any implementation can be folded into the DMO's annual financing remit.

Add it up. Not one pound of the Bank's gilt portfolio is now scheduled to be sold to a private investor. The path to zero is an accounting path, not a market one. Quantitative tightening by outright sale, in the United Kingdom, is finished, and it was announced in the vocabulary of its continuation.

The Debt Management Office got there first

This is the connection that is not being made, because the two institutions are covered by different people. The DMO's financing remit for 2026-27 came in at 252.1 billion pounds, slightly above expectations, but the composition mattered far more than the total. Planned long conventional gilt sales were cut to roughly 23 billion pounds, about 9 percent of issuance. The programme is tilted decisively towards short and medium conventional gilts.

The DMO has been candid about why. Its chief executive, Jessica Pulay, pointed in March to declining structural demand for long-dated gilts from domestic pension funds. That is the long tail of 2022. Liability-driven investment mandates have been recalibrating for four years, schemes have de-risked, and a great many have moved to buyout with insurers, whose appetite sits further in on the curve. The captive buyer of 40-year paper is not coming back.

So the sovereign has been shortening its issuance to avoid a thin market, and the central bank has now stopped pushing supply into the same thin market. Two separate institutions, two separate decision processes, one combined effect on the quantity of long duration that private investors will be asked to absorb between now and the end of the decade. That quantity is collapsing.

The objection, and why it does not hold

The serious counter-argument runs like this. Selling gilts to the Government does nothing for net supply. If the DMO buys 20 billion pounds a year of 2035 to 2049 paper from the Bank, it has to finance that purchase, and it finances it by issuing gilts. The market ends up holding the same amount of government debt either way. Moving the transaction from an auction to an intragovernmental transfer is theatre.

On the arithmetic of the cash, that is correct. On the arithmetic of duration, which is what actually prices the long end, it is not. The DMO chooses the maturity of what it issues, and it has already told us what it chooses. When it refinances a purchase of 20-year and 30-year paper, it will refinance it in line with a remit that allocates 9 percent to long conventionals. The transaction is therefore not supply-neutral. It is a duration swap, and it runs in the market's favour: roughly 146 billion pounds of long-dated risk retired from private hands over seven-odd years, replaced predominantly by short and medium paper that has a functioning buyer base.

Anyone who still thinks the long end is priced off the same fundamentals as the 10-year should look at what the curve has been doing. The 30-year gilt reached 5.89 percent on 1 September, its highest since March 1998, before easing back. The 10-year sits at 5.2169 percent after the decision. That is more than half a percentage point of curve between the two, and very little of it is a forecast about Bank Rate in 2050. It is a term premium demanded by a shrinking pool of buyers being handed paper nobody else wanted. Shrink the paper and you shrink the premium.

Why the hawkish dissent is the distraction

The three votes for 4 percent are being read as a signal about the path of policy. Look at what they were reacting to. CPI rose to 3.1 percent in the twelve months to August 2026, up from 2.9 percent in July. The largest upward contribution came from transport, and within transport from motor fuels, where annual inflation jumped to 23.0 percent from 15.5 percent. Petrol rose 9.1 pence in the month to 161.3 pence a litre, diesel 14.2 pence to 181.8 pence. That is Brent above 101 dollars a barrel arriving at the forecourt.

Underneath, core CPI held at 2.6 percent and services inflation was unchanged at 3.4 percent. The domestically generated component did not move. The MPC majority evidently understood this, which is why it held while the Fed raised its target range to 3.75 to 4.00 percent in a unanimous 12-0 vote, its first increase since 2023.

The dissent is a reaction to a petrol price. The balance sheet decision is a structural change to the plumbing of the gilt market that will outlast three MPCs. Markets spent the week pricing the first and skimming the second.

The price of this, which should be said plainly

None of the above is a compliment to the institutional arrangement. Selling gilts to the Government at market prices, to be refinanced by the Government in whatever maturity the Government prefers, is uncomfortably close to cancelling them with extra steps. The line between monetary policy and debt management, which the United Kingdom has guarded carefully since 1997, has just been drawn with a softer pencil. That is a real cost and a reasonable person could decide it outweighs the benefit.

But the question for an investor is not whether this is tidy constitutional practice. It is what it does to the price of 30-year gilts, and what it does is remove the marginal seller. The 30-year gilt yields 5.7415 percent against a 30-year Treasury at roughly 5.37 percent. Britain is being charged around 37 basis points more than the United States to borrow for three decades, at a moment when the United States is raising rates and Britain is not, and when Britain has just retired the supply that justified the discount.

What would change my mind

This argument is specific enough to be wrong in identifiable ways, and here are the three that would do it.

• The April 2027 review rejects the model. If the Bank and the Treasury cannot agree terms and APF auctions resume across the curve, the seller returns and the case dissolves. This is the largest single risk and it is genuine, because the arrangement has obvious critics.
• The Budget on 28 October puts long duration back. If the new Chancellor's fiscal arithmetic forces a materially larger remit and the DMO lifts the long conventional share well above 9 percent, the duration relief is cancelled at source.
• The term premium turns out not to be about supply. If long gilts fail to outperform Treasuries over the next two quarters while supply falls, then what the long end has been pricing is inflation risk and fiscal credibility rather than indigestion, and the mechanism I have described is not the one that matters.

Absent those, the trajectory is straightforward. The gilt curve between ten and thirty years should flatten, and long gilts should close ground on Treasuries. The Bank spent this week being praised or criticised for a decision on Bank Rate. It made a much larger one, and it made it in the technical annex.