Christopher Mahon is head of Dynamic Real Return in Multi-Asset Investing at Columbia Threadneedle Investments.
In February, the Bank of England updated its estimates for the lifetime losses of the asset purchase facility, the vehicle through which it conducts quantitative easing.
Its prediction for APF losses had crept up: from £95bn in September, to £115bn at the year end. The scale of these figures — and the change, in a matter of months — dwarf the £14bn of savings chancellor Rachel Reeves announced in her spring statement.
Now, some might argue that comparing a lifetime figure to an annual budget is unfair. And the purported benefits of QE — supporting the economy during some of its darkest days — risk being overlooked if the focus lands too much on the bottom line. But these are big, unignorable numbers — so how should we think about the financial out-turns of the BoE’s QE programme?
Recap: How QE works
To recap the mechanics of these losses: under QE, the BoE borrowed cash to buy government bonds (and the bonds of select companies) to support the economy. Financially speaking, this appeared to work well while cash rates were zero and locking in a 1 per cent yield on a 30-year gilt looked attractive.
But fast forward to today, with cash rates at 4-5 per cent or higher, and the same 1 per cent yield locked in for 30 years isn’t quite such a good deal. In fact, it is immensely costly, as a result of the “negative carry” on those low-yielding bonds.
How the BoE losses compare
One way to think about these losses is to compare the BoE against other central banks. Each supported their economies more or less successfully through QE, so we can look at which central bank provided this help for the lowest cost, in the most efficient manner.
However, with each central bank using its own accounting methods — with varying degrees of transparency — this is no easy task.
What we do know is what type of bonds each central bank bought, the dates of the purchases and how the bonds have performed since. Combined with certain other approximations, we can create rough mark-to-market estimates of the profit and loss of each of the central bank’s government bond holdings.
It won’t be perfect, but this approach allows for a rough and ready like-for-like comparison. On this basis, our broad estimates of the cumulative losses since QE began are shown below:
The picture is much worse for the Bank of England than its peers, with losses nearly four times the Fed’s. Why is this?
Well, here’s a handy cheat sheet of the key choices the BoE made compared to the Fed:
The first two rows are the most significant. The ~4x worse losses on the BoE’s government bond holdings boil down to the BoE owning roughly twice as many gilts, and them being twice as rate sensitive as the Fed’s treasuries:
Maturity risk vs maturity mismatch
A big chunk of the losses come from the long maturity bonds that were more of a focus for the BoE. As an example, during QE in May 2020 the BoE bought a series of 2061 gilts at a price of £101. Under QT it has been selling the same securities at prices as low as £28, a loss of 73 per cent.
Arguably, these longer-dated maturities did not need to be bought at all. As an example: UK household borrowings are far shorter than in the US (think 5-year mortgages vs 30-year in the US). The transmission mechanism to the wider economy from buying a 30 or 50-year gilt was always fairly limited.
With the Fed choosing to focus QE only on certain parts of the curve, there was a good reason for the BoE to consider doing similarly, tailoring their focus to suit local conditions, rather than buying quasi-passively all the way along the curve.
The role of active QT
It is too late to salvage most of these losses. But, when it comes to reversing QE though quantitative tightening, the BoE again finds itself on its own. Only the BoE is actively selling their holdings into the market — no other major central bank is pursuing QT in this way. Others are simply opting for the passive method of allowing their bonds to mature without replacement. The Fed is still partially reinvesting coupons.
The BoE argues this difference makes little impact. Others take a different view.
Research published by NBER last year highlighted how the BoE’s QT programme has had a negative impact on long-term bond yields4.
Researchers found a ~70 bps impact on long term gilt yields, against a 15-20bps impact from QT in the US — with a good portion of the difference being driven by active QT:
Certainly, the underperformance in gilts versus treasuries since QT began is noteworthy. Our own view is a combination of active QT, a faster pace of QT, and no coupon reinvestment has been weighing on gilts and perhaps adding about 30-40bps to yields versus comparable markets:
How to reconcile the Bank’s benign view against these more pessimistic assessments? The BoE’s own research tends to focus on the lack of evidence of market dislocations. But prices can be pushed down by a large seller without disorder and without liquidity worsening.
The incurious MPC
Perhaps surprisingly given the scale of the losses — and unlike decisions on interest rates — decisions about QE typically see unanimous votes by the MPC. Few public differences are aired or orthodoxies challenged. There is little reflection on the strengths or weaknesses in the UK’s approach. And there is no real acknowledgment of the alternative configurations that were — and still are — possible.
The BoE has avoided allowing consideration of profits or losses to drive changes in policy. Deputy governor Sir Dave Ramsden stated: “The MPC does not take into account financial risk or profit when taking monetary policy decisions, including about the gilt portfolio”.
Well, maybe it should. After all, taxpayers have to fund these losses. And if other central banks can do it cheaper, maybe there are lessons to be learnt.
Even at this late hour, if we are right about active QT exacerbating the supply demand imbalance in the gilt market, cancelling the extra sales from active QT and returning to central bank norms could be worth an ongoing £1bn per year saving to the Treasury in lower interest costs.
Please, Governor, can we have no more?
https://www.ft.com/content/45a441c9-26e0-4de3-b985-faafc5968a49
