Wholesale payments systems are the glue that ties the banking system together. These systems allow banks to make monetary transfers to each other – either on their own account, or on the behalf of clients – quickly and more importantly, safely. The payments system allows bank clients to transfer money electronically to one another without worrying about the exact mechanism for the transfer.
The failure of a wholesale payments system is a frightening prospect for a modern economy. Commerce would be impaired, and entities would likely fail as they would be unable to receive inflows that they need to meet credit obligations. (One potential counterexample to these statements would be to look at the strike in Irish banks in 1970s. However, it is not clear that a messy failure of a wholesale payments system would be worked around as easily as that strike, as there was no credit event alongside the closure, and alternative payment systems existed. There is a reference discussing this topic below.) As such, the main payments systems are effectively backstopped by central banks and are watched very carefully by the banks that use the system. If the payments system is working properly, its behaviour is rather easy to understand, it is just a means of shipping “money” from one bank to another. (I put “money” in quotation marks on the basis that what it represents may not match the reader’s intuition.)
As a result of the technology sector attempting to intrude onto banking systems, there is an interest in payments systems that allow monetary transfers directly between non-banks. So far, such systems are handling relatively small retail-sized transactions and are therefore less critical for the overall functioning of the capitalist system. Since they are a mechanism to bypass the banking system, they do not fit within a primer discussing banks.
Simplest Visualisation of a Payments System
The simplest way to visualise a payments system is to treat it as transfers between banks within the balance sheet of the central bank. That is, the central bank is a bank for private banks, and they hold accounts there. The proper name for the account balance for a private bank at the central bank is a settlement balance, but “reserves” is what one typically sees. (This is because balances were held to meet bank reserve requirements, see Section TK.)
If Bank A needs to send Bank B $1 million, Bank A sends instructions to the central bank to make the transfer. Bank A’s balance sheet shrinks (smaller deposit at the central bank), and Bank B’s balance sheet expands (a larger deposit). However, this is just an exchange of the ownership of deposits at the central bank, and so aggregate bank balance sheet totals are unaffected. (Aggregated balance sheets are the result of adding up all the balance sheet entries of private banks, which has the effect of obscuring what is happening to individual private banks. Such aggregate data are the public information that is published by statistical agencies or central banks.)
If the central bank buys something from a bank client for $1 million, it will transmit money to the client via increasing the settlement balance of the client’s bank by $1 million, and the bank correspondingly increases the deposit balance of the client. If the central bank is buying something not considered money (e.g., gold in a gold standard, or a government bond), this action will increase money supply measures. This is referred to as “monetisation” (turning a non-monetary asset held by the private sector into money held by the private sector), which is a word that features prominently in crank complaints about central banks.
What is Money?
One perennial question – that often has dubious proposed answers – is “what is money?” In modern economies, the previous discussion gives us the operational answer – it is something that exchanges at par against a settlement balance at the central bank.
The monetary system consists of transferring central bank balances between private banks and the central government (which also banks at the central bank), and these transfers are used to support exchanges in riskier layers that are build upon this base. The first risky layer are balances at private banks, and on top of the banks come the money markets.
A key property of money is that it is a unit of measurement (typically referred to as a unit of account). There are several instruments that qualify as “money” in Canada – banknotes, deposits at private banks, deposits at the Bank of Canada – which are all denominated in the same unit (Canadian dollars). Other monetary units are also used – American dollar accounts at Canadian banks are somewhat common – but the key is that Canadian dollar measurements are what are used in most transactions and contracts. (There is a popular tendency to explain the ubiquity of the local currency in contracts as being the result of legal tender laws, but the problem with that theory is that those laws are very rarely litigated. Legal tender laws prevent people from repaying debts by giving instruments that are not legal tender, but it is hard to find people with a low enough sense of business ethics who are willing to attempt such an act.) Gold (for example) is largely useless in commerce in Canada since the payment systems use Canadian dollars are the unit of account for balancing flows back and forth, you cannot ship ounces of gold in a system that pipes Canadian dollars electronically.
Crypto-currencies have failed as a form of money because the brain trust that designed them made them a speculative vehicle that is supposed to appreciate (assuming the existence of greater fools), which meant that users want to hoard the crypto-balances and not send the balances to other people through payment systems as a means of operating a business. (A capitalist is someone who pays workers money in order to engage in productive activities that generate more money and is not someone who hoards money.)
Overdrafts
There is no need for a bank to have a balance more than a transfer out – it might be allowed to run a negative balance on an intraday basis. That is, there might be the equivalent of an intraday overdraft facility. All that matters is that it returns its balance to the required level by the end of the day.
Such facilities allow a banking system to function without banks holding reserves at the end of the day. This corresponds to how the Canadian banking system (among others) operated ahead of the COVID pandemic, and a simplified framework of government finance that I describe in my book Understanding Government Finance.
Such a possibility does not correspond to pseudo-scientific theories about money and money velocity that are remarkably popular. Many economists managed to convince themselves that there is some kind of physical limit on the speed of monetary transfers (“the velocity of money”). The alleged implication of this theory is that if the central bank creates more money, this will allow more transactions in the economy. (This then leads to the Quantity Theory of Money.) In the real world, electronic transactions represent the bulk of monetary transfers, and the “velocity” of those transactions is the speed of propagation of electromagnetic waves in the communication channels (wires, co-axial/fibre optic cables). The theoretical limit for transaction volume dwarfs actual transactions. The private sector has a desired level of transaction flows, and the central bank adding more monetary balances does not affect the ability to realise those flows.
Payment Systems as Stand-Alone Entities
It would be rather cumbersome for banks to call up their customer service representative at the central bank every time they wish to make an interbank transfer. As such, payment systems are set up as independent entities (generally either explicitly or implicitly backed by the central bank) that manage the transaction process. The system could be viewed as “monetary pipes” that ship money between the members of the system (mainly banks, but some systems include non-banks). The money is sent into the “pipe” by one member, and it is supposed to pop out of the pipe at the recipient member. So long as nobody goes bankrupt, nobody needs to worry about the plumbing. At the end of the day, the net transactions on the pipes will be reflected by changed settlement balances at the central bank. The issue is that bankruptcies cannot be assumed not to happen, and so the plumbing must be designed to be robust to failure.
The details of payments systems vary by jurisdiction. For example, the United Kingdom has a few systems designed for different purposes. The discussion below is based on the discussion in the page “A brief introduction to the Real-Time Gross Settlement system and CHAPS” by the Bank of England (BoE, link below, accessed in 2024). There are two systems highlighted here, showing the diversity of needs.
CHAPS, which is a real-time gross settlement system (RTGS). The implication of being a RTGS is that the bank transmitting the money to have the balance on hand, and it is immediately transferred (without netting against other transactions). It is used for high-value transactions for which failure would be awkward. The BoE website states that the CHAPS system represents 0.5% of the transaction volume while being 92% of the value of transactions. The value of transaction flows is best visualised as a percentage of GDP. “From March 2021 to end-February 2022, the CHAPS system settled over £87 trillion of payments, equating to an average of over £344 billion each working day, or equivalent to the GDP of the UK every six working days.” The average payment value was £1.8 million.
Most transactions use the BoE’s net settlement service. The banks exchange information on individual payments, and then transactions are settled at the end of the regular settlement cycle (e.g., daily). The operator totals up all transactions, and the members settle their net balance with the system at the end of the cycle. The BoE offers a pre-funding service that limits banks’ exposure to settlement failure to amounts posted in prefunded accounts at the BoE.
Aside: Cheques (Checks)
Cheques (or checks in American English) were important historically but are largely disappearing outside of the United States. Since they will show up in histories, I will give a brief description of their operation for the benefit of younger readers.
A cheque is a paper debt instrument that instructs a bank to make a payment to the beneficiary of the cheque that delivers it to bank. (In Canada, banks currently allow clients to deposit cheques via taking a photo of it in banking software.) Cheques are pre-printed with identifying information of the payee to allow easy processing. (Cheques do not have to be pre-printed in this manner, but the story that a cheque written on the side of cow was acceptable was a story about the fictional character Albert Haddock by the English writer A.P. Herbert.) As a paper credit instrument, they could be endorsed by the payee and transferred some other entity, making them a negotiable instrument.
Cheques would be sent by banks to a clearing house, which would determine which banks owe other banks money (and how much) after netting. The cheques would be returned to the payee’s bank, which would use the information to deduct the amount of the cheque from the payee’s account.
Unlike electronic payments, there is no way of validating that the payee has an account balance that supports the cheque amount. Cheques that exceed the chequing account balance will be rejected foe non-sufficient funds (“NSF cheque”), and the payee will be hit with a cash penalty for “bouncing” the cheque. (Certified cheques are ones where the bank verifies that the account has the funds to meet the cheque payment and are used for things like settling a house purchase.)
The delay in clearing cheques means that can be used as a means of obtaining temporary financing – called the “float.” Although not normally too exciting, financial innovators in Kuwait managed to cause a spectacular stock bubble financed by post-dated cheques (the The Souk al-Manakh Crash, reference below).
Concluding Remarks
This section is certainly not a comprehensive review of the operation of payments systems. Instead, the focus on what people who are interested in the economy – and not building payments systems – need to know. So long as the payments system does not fail, it is just a method for banks to ship balance sheet entries back and forth during the day. When we look at end-of-day balance sheet data (which are the only data publicly available), all we see are banks shipping settlement balances held at the central bank amongst themselves. This explains why “bank reserves” feature so prominently in discussions of banking. That said, there is no strict relationship between the volume of transactions and the end-of-day settlement balances, so simple mechanistic stories about “more reserves means more transactions” represents misleading folklore.
References and Further Reading
One discussion of the Irish bank strike: Krueger, Malte. “Money and Credit: Lessons of the Irish bank strike of 1970.” Credit and Capital Markets–Kredit und Kapital 4 (2018): 645-667. URL: https://www.econstor.eu/bitstream/10419/293854/1/ccm.51.4.645.pdf
One example internet source for the claim about writing a cheque on the side of the cow was from Snopes.com. URL: https://www.snopes.com/fact-check/check-on-side-of-cow/
“A brief introduction to the Real-Time Gross Settlement system and CHAPS” Bank of England, accessed in August 2024. URL: https://www.bankofengland.co.uk/payment-and-settlement/a-brief-introduction-to-the-real-time-gross-settlement-system-and-chaps Note that payments systems are continuously evolved, and if the reader is interested in payments systems, it is probably best to find a current primer at a central bank’s website.
Craig, Ben R. 2019. “The Souk al-Manakh Crash.” Federal Reserve Bank of Cleveland, Economic Commentary 2019-20. https://doi.org/10.26509/frbc-ec-201920
Neil W recommends nationalise payment system:
ReplyDeletehttps://web.archive.org/web/20161106173840/http://www.3spoken.co.uk/2013/05/making-banks-work.html
the transaction system is a massive cost overhead to banks and they hate it. Every bank has a unique system, they are fundamental incompatible and that means the duplication between banks is colossal. I've done quite a bit of work on bank transaction systems over the years and the scale dis-economies are quite spectacular. And so they follow management fads - like outsourcing and offshoring - that do little more that shuffle the costs around the business. The recent failure of the RBS transaction system is a case in point.
The banks can do things together. APACS (now UKPA) and LINK show what can be done with mutual co-operation. That needs to be taken one step further, with the transaction system incorporated into a similar body.
There are lots of ways of designing a mutual transaction system. But at its core is one concept - that transactions operate on the balance sheet of the central bank, not the individual banks. So you would have a Transaction Department at the Bank of England (alongside the Issue and Banking Departments) and current and savings account ultimately represent liabilities on that balance sheet.
The functional aspects are less important - existing bank accounts could be held in trust by the current banks, run as separate subsidiaries companies and a myriad of different other options. But the key point is that the operational entity is acting as agent and the legal ownership and responsibility is always at the central bank. That makes anything recorded in the transaction system exactly the same as holding cash. You have a receipt for liabilities at the central bank.
However that makes the individual banks short of deposits and balancing liabilites. The replacement on the individual bank's balance sheets is of course an overdraft from the central bank - as mentioned in the section on lending. Existing banks would then have to get the match funding to free themselves from the central bank lending restrictions, conform to requirements or just enter run-off.
The transaction system is like the road or rail infrastructure and is a common good required by all. Inevitably the state will have to fund its existence - because there is no money in running it. I see the state providing a 'white box' system that anybody authorised can put a marketing veneer on. Done correctly it would mean that you can literally operate your bank accounts through any of the competing front ends. Account numbers would stay the same whoever you are notionally with.
one casualty is interest on deposits. To have interest on deposits in a private system there has to be income from somewhere else to pay that interest. Therefore in this system it becomes a line item of government spending - likely via interest bearing accounts for individuals at National Savings. Paying interest on deposits in this way is then really just the same as paying coupons on Gilts. Gilts, of course, would cease to be issued under any rational government.
Managing payments for customers is a core competency of banks. If they don't want to keep doing that, they can convert themselves into a non-bank financial institution (and see how long they last). There's no need for the government to intervene given that banks around the world have already managed to get customer payment systems to work.
Delete