On 2 September the Bank of England was lending British banks £208.7 billion against government-bond collateral. That is £124.8 billion through the weekly short-term repo and £83.9 billion through the six-month indexed long-term repo, set against a total reserves stock of £643.7 billion. Close to a third of the central bank money in the UK banking system now exists because a commercial bank pledged collateral and asked for it.
None of that is a malfunction. It is the design. What the numbers also show, and what the Monetary Policy Committee will vote on again on 17 September, is that a year of quantitative tightening has changed the composition of the Bank’s balance sheet without meaningfully changing its size.
The stock that would not fall
In March, Victoria Saporta, the Bank’s executive director for markets, set out where the transition had reached. The most recent survey of banks put the preferred minimum range of reserves, the level below which the system starts to feel scarcity, at £365 billion to £515 billion. Reserves at that point stood at around £640 billion. “Repos now supply around a quarter of reserves,” she wrote.
Six months later reserves stand at £643.7 billion. They have not fallen. The repo share has risen to roughly 32 per cent.
That is the whole story in two numbers. Quantitative tightening drains reserves by taking gilts off the Bank’s books. The Term Funding Scheme, which peaked at £193 billion in 2021 and is down to £41.8 billion, drains them as banks repay. Both ran through the period. Reserves are flat, because the banking system borrowed back roughly what the two of them removed.
The records were never a stress signal
The weekly repo has produced a run of headlines. Allotments hit £75.65 billion in August 2025, £100.9 billion in January 2026, and £122.9 billion by early June. Each was reported as evidence that liquidity was tightening.
Read the facility’s terms and a more precise reading emerges. The short-term repo is full allotment against Level A collateral, meaning any eligible bank that turns up with qualifying gilts gets as much cash as it asks for. It is priced at Bank Rate. Reserves are remunerated at Bank Rate. In pure carry terms, borrowing reserves through the window costs a bank close to nothing.
What a record allotment measures, then, is willingness rather than scarcity. Banks are not paying up for liquidity. They are choosing to hold reserves rather than gilts, at a price that makes the swap almost free, and the Bank is happy to accommodate them because that is precisely the demand-driven system it has spent four years building. More than 30 firms bid in a typical short-term repo. Around 80 use the indexed long-term facility.
The signal in the numbers is not distress. It is that the distribution of reserves, not the aggregate, is what now determines whether the system feels comfortable, and the aggregate is the number the framework still reports.
What the September vote is actually about
The MPC set the current envelope a year ago, voting 7-2 to reduce the stock of gilts held in the Asset Purchase Facility by £70 billion over the twelve months to September 2026, down from £100 billion the year before. About £21 billion of that was to come from active sales. The rest was maturities.
The split matters more than the headline. In the first quarter of 2026, gilt redemptions took £19.8 billion off the pile. In the second, five sale operations removed £6.1 billion. Active selling, the part that draws gilt-market commentary and Treasury Committee questions, is the smaller lever by a wide margin. The Bank’s loan to the Asset Purchase Facility stood at £521.8 billion on 2 September, against a path the MPC set toward roughly £488 billion.
The Bank’s own Market Participants Survey, published on 31 July, showed respondents expecting the envelope to fall again, to £50 billion for the year to September 2027, with sales weighted 43.3 per cent to the three to seven year bucket, 41.1 per cent to seven to twenty years and 15.6 per cent to longer maturities.
On the framework’s own arithmetic there is room. Reserves at £643.7 billion sit about £129 billion above the top of the preferred minimum range. On the composition, the transition is already most of the way done. The committee is choosing a pace for something that has, in practice, already happened.
Tenor, not size
The change worth watching is what the balance sheet is made of. A gilt portfolio is an asset the Bank owns outright, with a maturity profile measured in years and no counterparty. A repo book is a loan that has to be rolled. The short-term facility resets every week.
That shifts the point of failure. Under the old arrangement, reserves existed whether or not banks did anything. Under the new one, they exist because firms turn up on Thursday with collateral and the operational capacity to use it. Saporta’s message was blunt about the implication: “in this evolving liquidity environment, resilience hinges on firms maintaining the operational readiness to repo confidently.” Firms, she wrote, must remain “ready to repo”.
It also changes what the Bank is. Over the QT period it has sold gilts into the market and then lent reserves against gilts pledged back to it. The institution is becoming a secured lender to the banking system on a scale it has not run since before quantitative easing, and it is doing so at a moment when the rate decision itself is contested. Bank Rate has been held at 3.75 per cent through the summer, and the July meeting split 6-3, with three members preferring a rise to 4 per cent after a 7-2 hold in June. The debate on rates is about whether to go higher; the debate on the balance sheet is about what shape it settles in.
Three markers will tell you more than the September envelope. Whether the repo share of reserves passes 40 per cent. Whether the indexed long-term facility, which takes wider collateral for six months, grows faster than the weekly window, because that would signal genuine term funding need rather than convenience. And whether the Bank revises its preferred minimum range again, having already cut it from a £375 billion to £540 billion range down to £365 billion to £515 billion.
The size of the balance sheet is the number in the vote. The tenor is the number that matters.
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.