Volatile bond markets and the case for more transparent quantitative tightening
Published here:
The debate around quantitative tightening can be technical and confusing. It need not be. Costas Milas and Christopher Mahon argue that QT deserves the same scenario testing as the rest of monetary policy. But policymakers would still face the challenge of setting the appropriate interest rate to navigate the economy through rising inflation and volatile bond markets.
Quantitative tightening (QT) is the process of shrinking a central bank’s balance sheets after the bond purchases made under quantitative easing (QE) – the extra stimulus given during the global financial crisis and the pandemic.
The Bank of England (BoE) has taken an unusually active approach to QT, selling bonds rather than relying solely on the passive alternative of holding to maturity. Critics have argued that selling at today’s high yields and low prices sidelines value for money.[1][2][3]
Whether that is true depends on your assumptions.
Two economists. Two opinions
Writing in the Times in June, Andrew Bailey, the governor of the Bank, noted that the financial outcome for taxpayers would be the same under any type of exit plan – whether active or passive QT.[4]
The Bank recently announced a wide ranging set of changes to the policy of QT. So it is interesting to see David Aikman, the director of the National Institute of Economic and Social Research, showing that this change of heart could in fact save taxpayers as much as £1 billion a year.[5][6]
Can both be right?
The difference lies mainly on how yield curves are treated. Mr Bailey’s argument assumes that future interest rates follow the curve exactly. On this basis, Bank Rate would average over the next three decades the same as the 30-year gilt yield , currently just under 6 per cent.[7][8][9]
Mr Aikman assumes something different – that curves do not just reflect a mathematically precise view about future interest rates. Curves also contain term premium – an insurance-like cost demanded by wary investors for the uncertainty over such a long horizon. Models that are often used put that term premium today rather high, especially at the longer end of the curve.
Selling the bonds means paying up. Selling less – or adjusting the type of bonds to be sold as in the new BoE plan – can avoid these costs.
Framing the QT debate around concepts like term premia is unlikely to help the Bank in its communications strategy. There is another way.
How to choose a QT plan
In September 2026 the BoE’s Monetary Policy Committee (MPC) announced that it hopes to sell around £146 billion ($192 billion) of gilts from the Asset Purchase Facility (APF) to the UK Debt Management Office (DMO), at the same time while retiring gilts that reach maturity. The £146 billion figure replaces the gilts sold to the market and will likely take place at a fixed pace of £20 billion per year until 2034. The MPC also decided to pause all gilt sales until April 2027 while the plan is coordinated with the DMO.[10][11]
So how did the Bank choose £20 billion per year as right pace?
We know that the MPC considered faster, slower and more flexible paths. But their final choice was influenced by a desire to be predictable .[12]
At times, the Bank’s pursuit of predictability has drifted into inflexibility. The QE programme announced in November 2020 committed it to bond purchases until December 2021. Yet by May 2021 inflation was already above target and rising fast. Still, the QE juggernaut rolled on, buying around £70 billion of bonds while inflation was running hot, not cold. A more agile approach might have limited both the inflation overshoot and the eventual QE losses.[13][14]
Given the new programme stretches even further out – a horizon of eight years (until 2034) – it would improve the debate to see how the trade-offs were weighed.
At heart these are break-even questions. Leaving gilts in the APF means continuing to finance them through reserves remunerated at Bank Rate – currently 3.75 per cent. Transferring that financing to the DMO replaces this floating cost with a fixed cost that will likely follow the maturity profile of newly issued government debt – currently over 5 per cent.
Mortgage holders will know the dilemma: take the risk at a lower floating rate or lock-in a higher-but-more certain cost. That is a useful reminder that QE exit plans are not financially equal but do involve trade-offs.
The Bank already uses scenarios in its Monetary Policy Reports to examine how inflation and monetary policy might evolve under different assumptions. It could apply the same discipline to QT by combining those with assumptions for the DMO’s likely issuance profile. The simulations can be extended into the future, influenced by the BoE’s own estimates for the so-called neutral rate (namely the interest rate prevailing when the British economy produces at full capacity and inflation settles at its 2 per cent target).[15][16]
While neutral rate estimates fall into a wide band of between 2 per cent and 4 per cent, that can be used to our advantage. The adverse scenario considered by the BoE can be extended, with cash interest rates close to the higher 4 per cent estimates for neutral. The milder scenario of the BoE could also be extended with cash rates close to the 2 per cent estimates for neutral. The figures do not have to be precise – these are approximations for the average cash rates over longer periods of time.
The results could show how alternative balance-sheet strategies perform cost-wise if interest rates follow different paths. If the scenarios favoured a gentler unwind, sales could be slowed. If faster withdrawal looked prudent, they could be accelerated. And the type of “passive” approaches used elsewhere should remain in the comparison. For example, the ECB and the Fed are not selling their holdings, or conducting a BoE/DMO style debt swap, but simply holding their bonds to maturity. This type of approach can be more openly weighed for its own cost and benefit.
The point is not to replace judgment with a spreadsheet. It is to give policymakers a clearer basis for debate, and to make the reasoning behind any judgment visible.
A framework for better choices
The Bank has rejected allowing cost considerations to influence the BoE balance sheet. That has not worked out well. At £164 billion, the losses estimated by the Office for Budget Responsibility are significant. On a like-for-like basis the cost from the UK’s variant of QE dwarf those of other central banks. Making some effort to figure out the pound signs attached to the options open to the MPC is better than none at all.[17][18][19][20]
It is also an opportunity to show a forthcoming parliamentary inquiry that the Bank is properly kicking the tyres on the alternatives. As former MPC member Sushil Wadhwani has pointed out , weighing the expected level of loss is a viable tool to select from otherwise similar options.[21][22]
Scenario testing balance-sheet decisions is hardly exotic. Australia’s framework explicitly requires the Reserve Bank of Australia to examine exit paths under different scenarios , as well as to consider costs and benefits. It would be helpful if the upcoming UK remit letter nudged the Bank do the same. Even without that step, the MPC would be behind other central banks if it refused to look at unintended consequences.[23][24][25]
The QT debate can be improved – though interest rate decisions remain difficult
The BoE’s latest decision on QT is an improvement. Although the decision initially brought some relief to gilt markets, government bonds faced renewed pressure from the on-going war in Iran which has raised inflation expectations. Indeed, Bank of England’s policymakers now expect UK inflation to slightly exceed 4 per cent in early 2027.[26]
With borrowing costs rising and mortgages costs exceeding 6 per cent, financial conditions remain tight. Had the BoE’s continued with its active sales regardless, it could have added further pressure on yields (of 0.20 to 0.30 percentage points on the Bank’s figures, or 0.37 percentage points and 0.70 percentage points according to recent academic studies).[27][28]
The latest decision has brought a bit of breathing space, but what comes next remains tricky: keep interest rates on hold and your risk having to raise interest rates more steeply afterwards to avoid runaway inflation.
Rate-setting will never be an exact science. Interest-rate decisions will always require judgement. That is precisely why QT deserves the same scenario testing as the rest of monetary policy.
This article gives the views of the author, not the position of LSE Business Review or the London School of Economics. You are agreeing with our comment policy when you leave a comment.