Getting into the weeds of active QT and UK fiscal rules
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Over on MainFT is a column about the intersection between UK fiscal rules and the Bank of England’s active QT. Bottom line: the way the UK deals with central bank losses is weirdly hair-shirted, causing political problems, and should be changed.
There’s a line in the piece saying that the Bank’s Monetary Policy Committee has a direct hand in the determination of around £10bn of fiscal headroom. This post is mostly about how to get to this number.
Such minutiae may sound tedious to muggles. But you, dear readers, are wizards.
Some context. You’ll be aware that after many years of sending over dollops of cash each quarter to HM Treasury, the flow has — since October 2022 — been going the other way.
This is partly because Bank Rate (at which reserves that finance the Bank of England’s Asset Purchase Facility’s bond portfolio are compensated) is much higher than the stream of coupon payments that flow in. This delivers an “APF interest loss”.
But this is also partly because: a) the APF went and bought a bunch of bonds at prices higher than par, delivering a valuation loss at redemption; b) because the prices of bonds being actively sold by the Bank as part of their QT programme are much lower today than they were when the Bank bought them. Together, these deliver an “APF valuation loss”.
What’s with the scare quotes? It turns out that if you’re going to start talking about how much fiscal room the chancellor does or doesn’t have as a result of QT, it matters which of these APF losses you’re talking about.
Fiscal rule Britannia
In the UK, we have become used to referring to the amount of room a chancellor has to tax or spend before the OBR warns of a projected breach of fiscal rules as his (or potentially soon, for the first time, her) “headroom”.
In case you are a senior Treasury minister who needs reminding, a quick recap. There are three fiscal rules: a debt rule, a deficit rule, and a welfare cap.
For the purposes of these rules:
— ‘Debt’ means the Public Sector Net Debt ex Bank of England, aka PSND ex-BoE.
— ‘Deficit’ means the Public Service Borrowing Requirement, aka the PSBR.
The Charter for Budget Responsibility specifies that the Treasury’s mandate is:
to have public sector net debt (excluding the Bank of England) as a percentage of GDP falling by the fifth year of the rolling forecast period
UK policy wonks understand this as meaning that PSND ex-BoE in year five must be lower than PSND ex-BoE in year four of the forecast (as a percentage of GDP) if the rule is to be passed.
And as James Bowler, the Treasury’s top civil servant, wrote to Harriet Baldwin, chair of the Treasury select committee:
Losses incurred in the APF impact both PSNB and PSND ex BoE, and therefore the Government’s borrowing and debt fiscal rules.
Aha!
How do they impact the debt calculation specifically? Bowler’s words, our emphasis:
PSND ex BoE is impacted as the APF calls on the Treasury to cover losses via the indemnity. These include both losses from interest [” APF interest losses”] and the difference between original purchase prices and the eventual sale or redemption price [” APF valuation losses”]
So despite being an “ex-Bank of England” measure, as the Bank crystallises valuation losses over the next few years through active QT, the PSND ex BoE (aka debt) will increase.
Putting this together, we can just compare the Year Five number in the OBR’s central scenario, in which the Bank sells down its QE portfolio of bonds at an assumed £48 billion per annum pace to Year Five numbers in a scenario in which the Bank ceases active QT today.
Some debt rule numbers
The chart below shows how the OBR represents the different sorts of losses emanating from the APF portfolio in their central scenario (£48 billion a year of active gilts sales each year):
The idea that the Bank might actively sell £48bn of gilts each year might sound far-fetched to some analysts. But the OBR has to get a number from somewhere and rather than making a judgment call has simply taken an average of active sales to the year ending September 2023 and plans for the year ending September 2024. And, because of the UK’s fiscal framework, the OBR’s judgement calls end up becoming constraints.
The OBR doesn’t actually split valuation losses into redemption and sale categories. We did that ourselves by calculating the valuation losses due (assuming that bonds are not sold before maturity) on the Bank’s schedule of APF bond maturities. The reason for this split will become apparent soon.
This fiscal year, the APF is forecast to make a total draw of around £35bn on the public accounts. In 2028-29 (the all-important fifth year of the OBR’s forecast horizon from a fiscal rule perspective) the draw is projected to be £23bn.
These are BIG numbers!
If we want to make a wild guess as to how these numbers might look if active bond sales were scrapped, how would we go about it?
First up, we’d get rid of the active sale loss (light blue bars in the chart above). Can’t have active sale losses if there aren’t active sales. We’d retain the APF valuation losses on redemption (red bars). Because bonds bought above par will eventually mature at par, and this will crystallise book losses.
Then, we’d need to augment the dark blue bars showing interest losses. Stopping active bond sales would mean that the APF portfolio of bonds is bigger for longer. And if Bank Rate is above the average yield of the APF (which, let’s face it, it’s likely to be), this means that the interest losses will be higher. How much higher? We don’t know, so we’ll just scale up the OBR’s existing forecast APF interest loss in proportion to the augmented APF portfolio in a no-active sale scenario.
Doing all this gets us this chart:
If active QT stops, and our shaky assumptions hold up (Ed: they won’t) we’re talking about a draw on public accounts of £10 billion in the 2028-29 fiscal year. This is around £13bn less than the £23 billion loss in the OBR’s central scenario in which active bond sales continue.
To finance the cheque that HM Treasury is obliged to send to the Bank under the terms of the APF indemnity, taxes that would need to be raised, spending that would need to be cut, or gilts that would need to be issued. Cumulatively, over the next five years these differences under our two scenarios sum to £57 billion.
Some deficit rule numbers
Falling debt is just fiscal rule number one. What about the deficit rule? Again, James Bowler:
The impact on PSNB [deficit rule] is recorded through the interest losses that occur when the interest received on holdings of gilts and other assets is less than the interest paid on central bank reserves [” APF interest losses”].
That’s right. No mention of “APF valuation losses”. While the debt rule takes both APF interest loss and APF valuation losses into account, the borrowing rule looks only at the APF interest losses.
It may strike you as weird that a requirement for the public sector to borrow (or tax/cut spending) might be excluded from the Public Sector Borrowing Requirement. Not doing so handily stops lumpy valuation losses from driving the judgement around deficit compliance, which sounds maybe sensible.
So along with the mechanical fall in projected debt that might accompany cessation of QT, we would expect to see a mechanical rise in the projection of future interest losses. Given that these interest losses are captured in the PSNB — which is the metric that is referenced by the deficit rule — we would expect that this would mean that the amount of wriggle room the chancellor has under the deficit part of the fiscal rules would mechanically shrink.
To put some numbers to this, we reckon that it’s easy to see the OBR’s estimate of APF interest losses of £2.7 billion in 2028-29 rise to maybe £4.7 billion if active QT stops in September.
Are we there yet?
So halting QT would possibly mean the debt rule saying there is £13bn more headroom, and the deficit rule saying there is £2bn less headroom. Given that the OBR reckons that there is only £8.9bn of wriggle room on the debt rule but £56.8bn of wriggle room on the deficit rule, the debt target is the one that tends to get the focus. (The third fiscal rule — the welfare cap — is already broken, and has been for some time.)
The IMF picked up the importance of QT losses to UK fiscal rules in their recent Article IV recommendations. They reckon (our emphasis):
the profits/losses should be included in the debt definition used for the fiscal rule, as is currently the case, but there would be a case to exclude the profits/losses from any annually-applying deficit rule.
We guess they mean that interest losses should be excluded from the deficit rule (as valuation losses already are). Everyone is entitled to their own opinion, but this one looks odd. Let’s just say that we appreciate the Fund directing focus to the whole area.
But are we guilty of missing the wood for the trees? Is any of this really fiscal headroom?
Ceci n’est pas un <<fiscal headroom>>
On one level, sure — the draw on public accounts would be, hypothetically, £13bn lower in year five if active QT stopped. This sounds a lot like fiscal space. Especially to someone working in the gilt market.
But the calculation or estimation of fiscal space is contested, to say the least.
The IMF describes it as:
the room for undertaking discretionary fiscal policy relative to existing plans without endangering market access and debt sustainability
Faster QT increases “valuation losses” but also reduces “interest losses”. It *might* change the overall lifetime draw on the public accounts depending on the course of Bank Rate and the prices of bonds. But using today’s prices and expectations of Bank Rate the pace of QT doesn’t make much difference to the lifetime draw on the public accounts.
We know this because the Bank keeps on saying so. Here’s a chart they posted in December of the cumulative cash flows to/ from the APF under different unwind scenarios, so is a little out of date, but you get the message:
The yellow diamond on the right shows the Bank’s estimation of the lifetime loss if the Bank shrunk the APF by £100 billion between October 2023 and September 2024 as being indistinguishably close to the green diamond showing the same under a £50 billion shrinkage rate. The ‘Maintain Portfolio’ dot does look somewhat higher, but there might be reasons why it’s unideal for the Bank to run a huge balance sheet in perpetuity.
Looked at through this lens, it’s not at all obvious that the pace of QT — pulling losses forward or pushing them back in time — should matter very much to fiscal headroom.
Is this post nearly done? I mean, how much fiscal space is there out there?
We’re not really sure. It seems like a big and important question. And it’s one that a number of folks have had a stab at.
This chart, using data from a 2018 HM Treasury paper, has a go at compiling different estimates around the potential limits, thresholds and targets for government debt (augmented with latest level of gross debt from the ONS):
So there’s a range of views.
It’s reasonable that a fiscal rule should aim for ‘debt shouldn’t be spiralling into the deep blue yonder’ (if you want to impress economists at parties, call this, ‘satisfying the transversality condition’). Structuring rules to achieve this ambition without giving screwy results along the way looks like it has been difficult.
Comments