Forget Tiering! Give the BoE a (Proper) Balance Sheet

According to @Toby Nangle (FT, 11 June 2024, link), we now need to have a view on BoE reserve tiering:

This (...) question has broken into the mainstream agenda pretty fast. And if you don’t have a ready answer to this debate du jour, read on.

Really? My advice is if you discuss tiering at your dinner parties, you should get new friends. It is niche and misses the point. The BoE needs to emancipate and get a proper Balance sheet.

Context: tight public finances and APF losses

While the question has been discussed ad nauseam in trading circles ever since the ECB introduced such a framework, it has taken a new dimension in the UK when BoE’s QE profits turned into losses. According to @Chris Giles (FT), the BoE currently makes around GBP 23bn annual losses due to the gap between what it earns on its bond holdings (£13bn) and what it pays on corresponding reserves (£37bn).

Hence, the Reform Party argues, we should stop paying interest on reserves. This would save 30-40bn a year and allow the government to fund new spending.

I’m surprised that they haven’t written it on a bus yet (link). This proposal is merely a tax on banks and eventually on banking. It lacks panache: why not go for the “platinum coin proposal” (link): The government asks the Royal Mint to make a very special coin, from a very special material, worth £40bn. Like other coins, it’s fairly inexpensive to produce, but this specific coin is very very very heavy, making it practically useless for trade. It’s only good to store value somewhere in a vault. The government sells the coin to the BoE, and gets £40bn in credits on its account. Voila!

The long game: emancipate the Old Lady

I believe that the discussion about BoE’s £20-30bn of losses and how tiering would mitigate this, is a distraction. The BoE remains structurally profit-making, and annual losses happen from time to time. A technical loss, be it as large as £30bn, is just part of implementing monetary policy, which as we will see below, remains structurally profit-making. Tiering is not a great fix to temporary losses: It comes as a trade-off between the cost of policy and the control over policy rate.

Hence, while the question on the cost of running monetary policy is a very interesting one, tackling it through the lens of an annual loss is the wrong thing to do. As I will show below, the BoE remains structurally profit-making, even without tiering, and a way out of the current controversy is to consider setting it up with a proper balance sheet, P&L account, capital and all the shebang.

Currently, the BoE is more of a passthrough vehicle and while it has an operational balance sheet, published in its annual accounts, the APF is held off-balance sheet through a subsidiary called the Bank of England Asset Purchase Facility Fund Ltd. The APF P&L is fully equalized by the Treasury who mops up the profits and offsets the losses as they materialize. The Bank’s balance sheet will show this in “other loans and advances”. The Bank doesn’t have any meaningful loss-absorbing capacity because it is not expected to make any losses on the risk it carries on its balance sheet (which mostly boils down to collateralized funding operations).

Currently, the BoE is more of a passthrough vehicle. It doesn't own the APF risk. Why should it bother?

Some other central banks, such as the ECB, hold all the risk it takes on its balance sheet, especially financial risks stemming from asset holding and FX reserves, as well as residual risks from collateralized funding operations. In case of profits, the ECB can choose to retain or distribute. In case of losses, the ECB will try to offset these using its financial buffers or call on its stakeholders (the NCBs). By and large, the NCBs do the same, except that their final stakeholder is national treasuries.

The Bank knows this very well. In a recent working paper (Central bank profit distribution and recapitalisation, J. Long and P. Fisher, wp 1069, April 2024), BoE Staff review how other central banks handle their balance sheets:

Abstract (shortened): “Central banks retain a portion of their net profits as reserves and distribute the remainder to their finance ministry (…) This paper reports the findings of a survey of central bank profit distribution and recapitalisation arrangements across 70 jurisdictions and examines the range of features present, such as revaluation accounts and requirements for capital injections. The findings help establish the importance of a robust framework for managing central bank profit distribution and recapitalisation. The presence of such a framework should allow central banks to retain more of their profits and access external resources when capital is low, and to function as an income-generating asset for the government when capital is high, therefore ensuring both an appropriate use of public funds and the presence of a credible and financially independent central bank that stands ready to act when needed.”

A stylized balance sheet

To understand a Central Bank’s P&L, let’s review how central banks generate revenue. There are five main sources of revenue (see Long and Fisher 2024 for details):

• Seigniorage: This is the historical source of revenue generated “through being the monopoly issuer of physical domestic currency”. This is basically the difference between the cost of printing/minting the currency and the face value.
• Domestic market operations and net interest from asset holdings: Central banks provide liquidity to commercial banks and charge interest on these amounts. Central banks may also purchase assets and earn the spread relative to liabilities issued to fund these purchases.
• Gold and FX reserves: Some central banks hold Gold and FX reserves generating valuation profits and/or interest.
• Unremunerated reserves: Some Central banks do not remunerate (all) reserves and earn the spread between interest charged on deposits and the amount of corresponding lending operations.
• Regulatory levy: Commercial banks are requested to contribute to the cost of being regulated.

Apart from the regulatory levy, which is basically an annual fee, the other sources of revenue are derived from financial instruments held directly by the central bank and can be worked out from the stylized balance sheet presented in figure 1.

 Figure 1: A stylised central bank balance sheet

 Post-QE, the two main components of a Central Bank's Balance Sheet are:

• On the asset side, huge monetary operations, mostly bond holdings acquired through QE but also some liquidity lending operations.
• On the liability side, commercial bank reserves which are the counterpart of liquidity injections through asset purchases or liquidity provision.

BoE Asset Purchase Facility (APF): Profits then Losses

Since the BoE began QE in 2009, profits and losses arising from the BoE’s APF can be sequenced as follows:

• Initially, the BoE earned the positive spread between bond yields (i.e. 2% on average for the 10y interest rate between 2009 and 2022) and the cost of funding (i.e. the policy rate, 0.5% on average over the same period). see Figure 2.
• According to the BoE, “between 2009 and 2022, the APF’s activities generated positive net cash flows from the APF to HMT, peaking at a cumulative £123.8bn at end-September 2022” (link).
• Since 2022, the BoE tightened policy: the spread between the policy rate and the average yield of the APF portfolio (around 2%) turned negative; cash flows reversed. Losses raised both from the return differential as well as through valuation effects on bonds that the Bank was now actively selling back into the market, below its purchase price.
• Figure 3 shows estimates by the BoE of annual and cumulated cash flows related to the APF. It shows, under a stylized rate scenario that the program might break even around 2026, and generate cumulated losses thereafter, assuming the Bank maintains its current pace of “active” QE until the portfolio is run down.

Figure 2 Interest rates

There are two things worth highlighting at this stage:

• Firstly, a sudden and material rate hiking cycle is the worst possible configuration for a hold-to-maturity bond portfolio: as the funding costs shoot up, the average yield of the portfolio is sticky because only a small fraction of the bonds mature and are replaced by higher-yielding new bonds.
• Secondly, the decision to implement QT quickly after the end of QE, further limited the possibility to lock in higher yields (there has been hardly any turnover in the APF through reinvestment of maturing bonds), while also materializing valuation losses (because nominal yields are now higher, bonds are sold at a lower price than they were purchased).

I do not believe these losses could have been avoided. A slower hiking cycle would have limited the negative carry but would have been inconsistent with monetary policy objectives, eventually leading to a larger policy rate overshoot. A delayed QT, allowing for more portfolio turnover and a higher effective yield would also have been inconsistent with the required tightening insofar that abundant liquidity provision would have worked against monetary tightening. Finally, tiering (here it comes), while lowering the cost of reserve remuneration would also have contributed to lowering short-term market rates i.e. deliver monetary easing while the BoE wanted to tighten. Tiering unfortunately, comes with a structural tradeoff between the cost of policy and the control over policy rate.

Figure 3: APF cash flows (actual and projected)

The BoE will return to profit

To avoid opening a new front, the BoE has omitted in Figure 3 what I believe is the most likely scenario: what happens if the Central Bank maintains a large bond portfolio by continuously reinvesting maturing bonds?

In a must-read speech(link), A.Hauser explains why the Bank’s balance will likely remain on average much larger in the future, than it has ever been in the past. According to a recent commercial banks assessment, the ‘Preferred Minimum Range of Reserves’ is about £335-495bn, with theoretical estimates much usually larger than this. As a reminder, the APF currently stands at just around d £700bn.

 Hauser’s speech suggests that by 2026, at the current pace of QT, the liquidity provision by the BoE may already be within the preferred range of reserves. From there, the BoE may not be able to reduce its balance sheet further without risking a liquidity squeeze and spiraling short term rates. Coincidently, 2026 is also about the moment when APF’s cumulated cash flows break even.

What happens then? The Bank either stops QT and sits on a large bond portfolio, or runs it down further but provide fresh liquidity through lending programs, or a mix of both. Lending operations are by and large neutral from a revenue perspective (cost and remuneration at policy rate) while bond holding earn the (positive) term premia, over the long run.

That’s probably the best moment, to give the Bank of England a proper independent balance sheet including loss absorbing capacities, much alike how eurosystem NCBs and the ECB are set up.

In my view, an independent balance sheet is critical to central bank’s independence. Loss absorption capacity mitigates the reliance of central banks on their stakeholder by preventing frequent requests for capital injection, which is actually what the BoE is doing when it offloads APF losses on the treasury (i.e. those £23bn from earlier). By limiting the financial reliance vis-à-vis its stakeholder, the central bank is also incentivized to act independently. Experience shows that Central Banks, risk averse by nature, internalize the cost of their policy to ensure that they remain structurally profit making institutions.

Loss absorption capacities, would avoid the nasty “look through” effect of seeing the Treasury (i.e. taxpayers) footing the perceived losses of QE. it also make the central bank more aware of the risks it carries on its balance sheet.

 Initially, the BoE would be allowed to retain profits, and stash them into a capital account (see stylized Balance Sheet) as it builds up its balance sheet. The size of these buffers and the contours of the profit distribution rule is to be determined by P&L modelling, in agreement with the Treasury. If the buffers are well calibrated, recapitalization needs will remain exceptional (or nonexistent if you’re fine with the Bank running a negative capital position).

 Because the Bank has no mandate to hoard profits forever, the Treasury will continue to enjoy an average stream of positive cash flows. In turn, such a less volatile and more reliable stream of income will facilitate budget implementation as well as debt management.

#BankOfEngland #EconomicPolicy #MonetaryPolicy #CentralBanking #FinancialIndependence #QE #BalanceSheet #BoE #Finance #Economics #eurosystem #ECB