What is money, anyway? The everyday view equates the concept with pieces of polymer in our wallets, coins in our pockets, or numbers on our mobile banking apps. But, as anyone who has had the fortune of sitting next to a banking scholar at a dinner party would know, those three things are – in some fairly fundamental ways – different. There are even differences between paper money and coin, the most recognisable forms of government-issued money, since they owe their legal characteristics to different laws and, consequently, come into existence in different ways.
After we put currency to one side and bank money to another, a third, far more mysterious, kind of ‘money’ remains: central bank reserves (also referred to as central bank money). In a recent article in the Review, Will Bateman and Jason Allen attempt to shed some light on the legal nature of central bank reserves. In doing so, they deliver a novel – and much-needed – exploration of how these reserves are seemingly created out of nowhere by central banks. They contest the widely held conventional view that reserves are simply liabilities of the banks which issue them (‘the liability view’). Such a description, they argue, raises several issues. Of those, the most significant is a descriptive deficiency arising from the idea that the liability view only focuses on a ‘deposit pathway’ for reserve creation. This focus, they assert, is misplaced since the deposit pathway has, in the context of modern central banking, a marginal role in reserve creation. Rather, the vast majority of reserves are created through a central bank’s open market operations with its counterparties, rather than the deposit of currency by reserve account holders. Accordingly, they propose a dual-track account of central bank creation, which asserts the separate existence of a ‘transaction pathway’, separate and parallel to the deposit pathway, which denotes the legal mechanism(s) through which reserves are created when a central bank transacts with its (predominantly) private-sector counterparties.
At surface level, their approach is descriptively attractive. The neat separation of deposit-created and transaction-created reserves seem to capture the breadth of ways that reserves come into existence in modern monetary systems. More broadly, it makes explicit the legal framework of an immensely significant part of the monetary-financial system’s plumbing. Given that importance, it is surprising that it has taken until now for scholars to present a consolidated account of the law which underlies this immensely important piece of economic infrastructure.
Nonetheless, I will suggest that the ‘bifurcated view’ is mistaken and needlessly complicates matters. In reality, the transaction pathway that they envisage is only practically, and not fundamentally, different to the deposit pathway. Ascribing fundamental differences between the two mystifies the operations of already-opaque institutions, and diverts focus away from the true legal source of the ‘money-ness’ of central bank reserves – the ability of central banks to incur transferable liabilities, which are treated by market participants as capable of satisfying debts at par. This orthodox view,[1] I will argue, accurately reflects the conceptual, legal, and practical reality of reserves.
I. Against Bifurcation
Monetary authority is the power to create money. The most recognisable form of legal tender is, of course, currency. Issuing currency involves the exercise of will to transform the juridical status of a physical object. Take the Coinage Act 1971. Section 9 prohibits the making or issue of ‘piece[s] of gold, silver, copper, bronze, or any other metal or mixed metal… as a coin or token for money’ except under the authority of the Treasury. At the same time, section 2 states that coins of various metals, if they meet specifications under the Act and relevant delegated legislation, ‘shall be legal tender … for payment’. In a similar vein, section 1 of the Currency and Bank Notes Act 1954 stipulates that ‘The Bank of England may issue notes of such denominations as the Treasury may approve’ and that ‘All bank notes issued under this section shall be legal tender in England and Wales’.
The effect of these provisions is twofold. First, it prohibits anything except official currency from being legal tender. Second, they create the power to confer the status of legal tender onto physical objects. When that power is exercised, what was originally raw material – i.e. pieces of metal, paper, or polymer – subsequently takes on the legal characteristic that a recipient is under a duty to accept it as discharging a debt owed to her by a debtor. ‘Moneyness’ in this sense is less to do with the physical characteristics of an object and more to do with how the law requires it to be treated by others. Money is, as traditionally described, ‘the unit of account, the medium of exchange, and the mode of payment used in a society’.[2]
Nevertheless, our concern here is not with the monetary authority to issue currency. The ‘money’ issued or destroyed by central banks on a day-to-day basis is a different juridical construct entirely. Like currency (and commercial bank money), it is a monetary unit that is used to settle payments between counterparties. Unlike currency, it has no physical manifestation and purely exists on the records of central banks. And, since only certain kinds of institutions (like large commercial banks, broker-dealers, countries, and international financial institutions) are allowed to have accounts at central banks, the vast majority of us will never directly interact with reserves – much less possess any.
But, for our purposes, the most important difference – at least in relation to the US Dollar, Sterling, and Euro systems – is the absence of express statutory authorisation which empowers central banks to make reserves. That is, there are no ‘Reserve Acts’ analogous to the UK’s Coinage and Bank Notes Acts. In this sense, Bateman and Allen are entirely correct to note that reserves’ status as money is ‘best characterised as implied or incidental to central banks’ express powers to transact in financial markets’ (at 403, emphasis in original).
Where I disagree with their account is the source of those implications and the conclusions drawn from them. In their analysis, the law of reserve creation has three ‘elements’ (at 412ff). First, the ‘deposit pathway’: commercial banks can make accounts at a central bank and make deposits into them. Second, the ‘transaction pathway’: central banks can purchase financial assets (principally government or, to a lesser extent, corporate debt) from their counterparties, or straightforwardly lend to those counterparties by making loans. Third, reserve balances can be transferred across central bank account holders to settle debts with finality. The first two elements mean that a central bank can create claims against itself. The third element makes those claims money-like. Importantly, on Bateman and Allen’s account the deposit and transaction pathways are set out as being legally distinct. The former is asserted to owe its existence to institutions’ ability to open reserve accounts and make deposits into them. By contrast, the transaction pathway is said to be distinct because it has nothing to do with ‘commercial banks making any kind of deposit’ into their reserve accounts. Instead, creating reserves through it involves the central bank undertaking transactions to buy assets from, or extend credit to, counterparties.
On closer inspection, however, no such distinction exists. Conceptually, there is no difference between how £1000 of reserves are created when a clearing bank deposits with the Bank of England £1000 in currency, sells to it £1000 worth of a government payment obligation (i.e. gilts), or sells to it £1000 worth of its future payment obligation (i.e. a loan). Of course, there is a factual difference because currency, unlike financial assets, is definitionally valued at par. There are no uniform – much less legal – criteria for how central banks should value financial assets. Since a central bank must determine how much it should pay for those assets, there is perhaps an impression that there is something different happening when a central bank purchases, say, government securities compared to when it accepts deposits. Yet, a central bank purchasing an asset still fulfils its side of the contract by conferring to its counterparty the asset’s agreed-upon value. When it does so, there is no difference if it pays in currency and immediately accepts that currency as a deposit, or if it simply credits its counterparty’s account for the same amount. Conceived in this way, the transaction pathway is strictly speaking a subset of the deposit pathway. And, viewed from the other direction, the deposit pathway is also the transaction pathway insofar as a central bank’s receipt of a currency deposit in exchange for a certain amount of reserves is a purchase of that currency.
Going further, we can dispute the suggestion that deposit-placing acts are central to commercial bank deposits in a way which makes them characteristically distinct to central bank reserves. This is because even ordinary bank deposits are not solely or primarily created through the deposit of currency. Quantitatively, this is obvious, given that the amount of currency in circulation is only a small subset of the overall amount of money. For the Sterling system, as of April 2023, £93 billion of currency is dwarfed by the £2.2 trillion of ‘broad money’,[3] which includes banknotes, coin, deposits, certificates of deposit, repurchase agreements, and some securities.[4] This discrepancy exists because the majority of money exists because of private credit creation, i.e. lending. When a commercial bank makes a loan, it does not advance currency. Instead, it simply credits the value of the loan to the borrower’s account. In this way, the bank does not transfer money which it previously had, but instead creates new monetary value which is given to the borrower in exchange for the latter’s obligation to repay at a later date. At the risk of repetition, we should observe that this – as noted above – is directly analogous to what central banks do when they purchase government or corporate debt. If reserves are not ‘deposits’ because central banks typically create reserves without receiving currency deposits, then neither are commercial bank deposits. Accordingly, a conceptual distinction between balance-sheet money created through deposits and those created through transactions is, at best, artificial.
- For Unity
In affirming the orthodoxy, we can point to a single source which explains the ‘legal parentage’ of central bank reserves: central banks’ legal personhood, including their ability to incur liabilities to other legal persons. In this sense, Bateman and Allen were not far off when they delved into the history of the Bank of England and ascertained the central importance of the Bank’s Royal Charter from 1695 and the Bank of England Act 1694.
With respect, however, they seem to have missed a crucial subtlety. There is a difference between possessing a legal power to buy or sell assets, and the power to incur specific kinds of obligations, like indebtedness. The former, which is what Bateman and Allen identify across the three monetary systems they explore, can exist without the latter. It may be the case that a legal person or body corporate has the power to transact with the wider world only in a limited set of ways. This was, of course, the crux of the ‘local authority swaps litigation’. Those cases began with Hazell v Hammersmith and Fulham LBC [1992] 2 AC 1, where the House of Lords determined that, under the law in force at the time, local authorities did not possess the legal power to enter into certain derivative transactions, with the effect that those putative transactions were void.
So, Bateman and Allen are right to note that the importance of the Royal Charter and the Act giving the Bank of England the ability to ‘purchase and acquire … sell, grant, demise, alien and dispose of’ various legal entitlements is important in its ability to create reserves through transactions. But, the ability to ‘create the monetary units necessary to carry out those transactions’ (at 416) is in no way a necessary implication of those powers since, to point out the obvious, the Bank may transact using currency or other financial assets.
What this shows is that the question ‘What legal power(s) does a central bank exercise to create reserves?’ is imprecise and leads to an overdetermined answer. There is a distinction between a particular form of monetary authority, and the different legal conditions which, if met, permit its exercise in any particular instance. Ascertaining the former is an issue of general importance which requires conceptual clarification, as doing so sheds light on the institutional nature of central banks. The latter – which is what the authors direct themselves to – is a more positivist inquiry; its answers are potentially idiosyncratic, reflecting what they correctly refer to as the ‘path dependency of monetary law’ (at 420ff).
III. For ‘Liability’
The authors also use the bifurcation of reserve-related monetary authority to support a separate, but connected, argument: that reserves should not be characterised as liabilities of the central bank. As with their earlier argument, this contrasts with the conventional view that reserves are ‘just another form of government liability’,[5] meaning that the ‘owner’ of reserves possesses an obligation owed by the central bank. This view analogises central bank reserves with commercial bank deposits. In practice, this ‘accounting view’, as I will refer to it, is the norm.
They present three facts as being at odds with the accounting view: (1) that reserves do not entitle their holder to payment by the central bank; (2) that reserves are ‘created by central banks crediting reserve accounts’ rather than through deposits; and (3) that the role of reserve requirements, as a part of macro-prudential regulation, has waned (at 421). The implication is that the truth of these facts makes a difference to the validity of the accounting view’s ‘balance sheet treatment’ of reserves.
None of these points is persuasive. The second point has already been dealt with above. On the third, it is not clear what difference the existence of reserve requirements makes to the nature of reserves in general, and their characterisation as deposit liabilities in particular. At best, one can discern the idea that such requirements create a duty on those subject to it which can only be met through the possession of reserves. Through that, reserves may be seen as a central bank ‘liability’ insofar as possessing them allows banks to meet their duty to possess sufficient regulatory reserves. Still, this is a highly strained interpretation which conflates two senses of ‘liability’: the deposit liability to pay or transfer a monetary obligation on request, and a regulatory liability (or, more precisely, obligation) to maintain adequate reserves. At best, this interpretation has only incidental bearing on whether it is accurate to describe reserves as balance sheet liabilities.
Only the first point, therefore, poses a conceptual challenge to the accounting view. It is entirely proper to ask what – if reserves are liabilities – are central banks liable to do about them. In response, we can affirm the analogy with commercial bank deposit liabilities by noting two things. First, it is not obvious that the conceptualisation of commercial bank deposits as liabilities hinges solely on the bank’s obligation to redeem balances, on demand, in currency. For example, until recently, the United States Federal Reserve System’s Regulation D restricted the frequency that an savings account-holder could make withdrawals. More generally, and across the world, there are routine daily limits on how much currency one can withdraw from ATMs, beyond which one must go to bank branches to withdraw more. Even then, beyond certain amounts, branches may also require notification to ensure they have available the cash to be withdrawn. Importantly, these limits are much smaller than the analogous limits on bank account transfers. So, it is not clear how much the conception of commercial bank deposits as ‘liabilities’ has to do with the ability to exchange balances with currency.
Second, the analogy with commercial bank deposits is only reinforced when we consider the nature of modern fractional reserve banking. In contemporary practice, the binding constraint on commercial banks’ deposit liabilities (which they can increase by issuing loans) is not the amount of currency they have on hand, but the amount of reserves they can use to meet outflows. As recent events have shown, depositors redeem their right against their bank’s liability not by requesting currency but, in effect, by requesting a transfer of reserves from one bank to another. In a sense, then, Bateman and Allen get it exactly right in pointing out that ‘in the context of commercial banking, a “deposit” functions as money because it is treated as a transferrable claim for payment against the bank at which the account is held’ (at 421–22). Their error is over-emphasising ‘payment’ and understating ‘transferrable’. If, instead, transferability were emphasised, the analogy with commercial bank money would be more apparent.
For these reasons, a descriptive distinction between central bank reserves and commercial bank deposit liabilities made with reference to the absence of an entitlement to ‘payment’ in currency is not persuasive. The ‘intellectual durability’ of the ‘deposit pathway’ is not owed to some misapprehension by central bankers or practitioners, but to the fact that it is correct – at least in the analogy it makes between reserves and commercial bank deposits.
Finally, we should note the authors’ puzzling dismissal of viewing interest on excess reserves (‘IOER’) (and, presumably, any remuneration for the maintenance of reserve balances more generally) as ‘interest on debt obligations owed by central banks to commercial banks’ as ‘highly artificial’ (at 422). It is not clear what is meant by their replacement explanation that IOER is ‘that central banks are simply paying additional sums to financial participants in volumes which are referable to the quantity of reserves issued via QE programmes’ (at 422). To be sure, there is a strong practical correlation between the payment of IOER and reserve creation under QE given their simultaneous use as policy mechanisms in recent years. But, there is no conceptual, legal, or practical requirement that central banks only pay interest on reserve balances at the same time as making outright purchases of financial assets. In fact, for both the Bank of England and the Federal Reserve System, the idea to remunerate the holding of reserves emerged before the introduction of large-scale quantitative easing programmes after the Great Financial Crisis. So, perhaps it is not ‘the idea of loan agreement between a commercial bank (the creditor) and a central bank (the debtor) in their characterisation of a reserve account’ which is at variance with ‘financial reality’ (at 422).
- Conclusion
To conclude, we should ask: why does getting this right matter? As a description of the conditions for the exercise of monetary authority, it is arguable that Bateman and Allen’s account is more useful in practice than the conceptual response I have articulated above. Internalising the idea that reserves are created in exchange for both currency and financial assets feels like a more valuable insight than having a coherent theoretical understanding of what reserves are.
However, I suggest that at stake is conceptual clarity which is necessary to limit the possibility of confusion further down the line. At the very least, we should have a comprehensive and accurate understanding of how central banks exercise their incredible ability to affect real-life outcomes. Given their legal and material power, ‘close enough’ is not good enough. In public law terms, there is a rule of law valence to discussions in this area. We must know what powers are involved in reserve creation because those capacities are an essential part of a modern state’s governing apparatus.
More practically, getting the concept of reserve creation right impacts the cogency of reasoning and conclusions reliant on it. It is possible to be unduly concerned with the importance of what I have been referring to as the conditions for the exercise of monetary authority and as a result to mistake this for the monetary authority itself. Ultimately, those conditions are contingent on a particular, perhaps technical, set of choices about how money can be created at any given time in any given place. Those conditions are subsequent – not prior – to a polity’s conception of monetary authority, and therefore cannot characterise its fundamental features. Of course, understanding those contingencies is important. We should inspect how the ways we take monetary authority to be exercisable reflect back onto how we conceptualise monetary authority in the first place.
But it would be wrong to conclude that those (practical) modes are constitutive of monetary authority more broadly. It appears that the authors make this error. They seem to take the view that the transaction and deposit pathways are the only ways that central banks can issue reserves. In their opinion, the central banks they survey (at 428):
… can, at present, only execute their mandated objectives … through financial market transactions that involve commercial banks (and other private entities) as willing counterparties. It is only through such transactions that central banks are able to create reserves …
This leads to a final concern with their account. The insistence on the specificity and exclusivity of the pathways leads to a restrictive view of central banks’ legal capacities. Naturally, it can only be true that ‘central bank mandates are limited by central banks’ legal powers’ (at 428, emphasis in original) and that an expansion of their mandate ‘does not per se entail an extension of the central bank’s legal power and capacity’ (at 428). The authors imply that this might be problematic: it means that a central bank pursuing ‘QE for the people’ or ‘helicopter money’ policies ‘might hit against legal constraints’ and ‘falter due to an absence of legal power on the part of the central bank’ (at 428).
But, if I am correct, this is a red herring, and the problem is precisely the opposite of what they suggest. Thinking from the ground up, we should be less concerned with whether central banks currently do not possess the legal powers to implement innovative approaches to monetary policy or payments systems. If those innovations are prudent and desirable but the relevant institutions do not have the legal powers to implement them, then their governing documents should be amended to give them those powers. But a desire for those institutions to possess those (definitionally extensive) powers in advance of any democratic decision to make use of them surely puts things the wrong way around. It would be naive to minimise the challenges that independent central banks pose to principles of democratic governance, like transparency and accountability, in favour of political expediency. When it comes to matters of high policy (rather than the more technical aspects of financial stability), those challenges mean that we should err on the side of a more restrictive posture when evaluating central banks’ legal powers. Therefore, the animating worry should not be about whether they have the power to implement ‘QE for the people’. It should be about how they had the power to, say, implement QE in the first place, and whether there has been sufficient legal and democratic oversight over those enormously consequential programmes. These, I believe, are the true ‘deep institutional issues’ that this area of monetary law implicates – not whether central bankers think they can accept deposits from ordinary people.
Cite as: Shukri Shahizam, ‘Defending the Orthodoxy: Central Bank Money as Liabilities: A Reply to Will Bateman and Jason Allen, “The Law of Central Bank Reserve Creation” (2022) 85(2) MLR 401’ [2023] MLRForum 001 (available from: http://www.modernlawreview.co.uk/shahizam-bateman-allen).
