By Nick Philpott in association with Stablecoin Standard

Few people realise what happens behind the scenes when they make a payment. On the surface is a modern-looking interface for sending money to someone else’s account, remitting money overseas or buying something online. In the UK consumers have a rare glimpse behind the scenes when they buy a house and must pay a “CHAPS” charge, which is usually around £30. This small sum can feel like a final insult after paying thousands of pounds to estate agents and lawyers for a house purchase.

CHAPS, the Clearing House Automated Payment System, is one of the largest high-value payment systems in the world, and interfaces with the UK’s Real Time Gross Settlement System (RTGS) for GBP. These payments include transfers between corporations via their banks, property purchases and the settlement of transactions in the multi-trillion dollar-a-day foreign exchange market. Access to CHAPS is restricted to its direct participants, many of whom are large banks. Members need to deposit funds to be able to make payments to one another, giving rise to capital and liquidity risk.

What is less well known about CHAPS is that it, together with the UK’s RTGS service, was offline for six hours on Monday 14 August 2023. This followed another critical RTGS outage in 2014, also affecting the CHAPS system. Such outages are not confined to the UK, with multiple instances across the world since the Financial Crisis:

1. TARGET2 (2018 and 2020): the Eurosystem’s real-time gross settlement system, known as TARGET2, experienced a major outage in November 2018. The system was unavailable for several hours, impacting cross-border payments and settlements within the Eurozone. TARGET2 had five outages in 2020 (the report is here).
   
2. FEDWIRE (2019 and 2021): The Federal Reserve’s real-time gross settlement system, Fedwire, encountered problems in 2019 and 2021. The outages impacted various financial institutions relying on the system for interbank transfers and settlements.
   
3. Reserve Bank of Australia (2022): the Reserve Bank Information and Transfer System (RITS) experienced an outage that impacted 800,000 transactions over the course of four hours. The cause was a planned change to the software that manages its virtual servers.

RTGS are centralised databases. For example, Fedwire is a series of Oracle databases, which even has a publicly available user manual. In this sense, wholesale payments are not Internet-based. A payment involves moving data within a central RTGS database. FX transactions are more complicated since settlement involves moving data in two different RTGS databases in two different countries.

Arguably Fedwire is the oldest RTGS, dating back to the 1970s. By 1985 three countries had adopted RTGS, growing to about 90 covering all the major currencies by 2005. This growth in adoption was encouraged by the Committee on Payment and Settlement Systems (CPSS), part of the Bank for International Settlements. The intention behind this expansion was to reduce or even eliminate credit risk between large institutions that had become systemically important.

This systemic importance and the risk it represented had been illustrated to many central bankers by the failure of Herstatt Bank in 1974. Herstatt was closed at the end of the German business day before it could make many of its USD payments. At the time, its crucial role in the USD clearing system had been significantly underestimated. Continental Illinois Bank’s failure in 1984 emphasised the point. Illinois acted as a correspondent bank for at least 180 other banks, making it a systemic risk to the USD payment system. Regulators intervened with a rescue package, deeming it “too big to fail”.

Before the arrival of RTGS, banks had netted payments on their own balance sheets rather than settle them gross. The Bank for International Settlement’s Committee on Payment and Settlement Systems (CPSS) (now the Committee on Payments and Market Infrastructures (CPMI)) and its G10 central bank members were not comfortable with this, preferring that assets used for large value settlement should be claims on central banks rather than on commercial banks. However, while the move away from bilateral netting reduced counterparty risk between the banks, it created liquidity risk, since commercial banks needed to fund their gross payment needs in an RTGS at the central banks, which in turn meant they sometimes needed to sell assets to release cash.

The move to RTGS also created a single point of failure for the entire financial infrastructure of a country or even a group of countries should a currency union such as the euro or the West African franc be involved. What the adoption of RTGS had done was reduce credit risk but increase both capital and liquidity risk as well as operational risk in regarding the reliability of the RTGS.

The criticality of an RTGS to its economy created a need to maintain confidence in the system and, by extension, the central bank(s) that operate them, as a 2014 SWIFT report on resiliency illustrates:

‘As it happens, several RTGS systems have experienced failure already. However, only the major instances have reached the public domain, and then only because an outage proved impossible to conceal.’

It is noteworthy that there is very little information in the public domain on RTGS failures. Only when they are so critical that they ‘proved impossible to conceal’ have they come to light and, as was the case with CHAPS, RITS and TARGET2, the subject of an independent review.

In contrast, cryptoasset platform failures are public by their very nature. When settlement takes place on the blockchain it is a form of gross settlement that needs to be fully funded. However, not all cryptoasset transactions are settled gross since some counterparties allow each another discretionary risk limits, which serves as a form of bilateral settlement netting. Active traders can trade during the day and then net settle at the end of the day. Considering the high availability and responsiveness of cryptoasset platforms, counterparties have the discretion to conduct additional settlement cycles during the day (including weekends, unlike RTGS). In some ways, what this creates is a hybrid system of both gross settlement as well as the netting practices that were phased out by the advent of RTGS.

Stablecoins, a form of cryptoasset, are representations of fiat currency that sit on the same blockchain platforms as tokens such as Ethereum, Lumens or Matic. These are Internet-native tokens insofar as the chain on which they sit can be hosted and or downloaded by anyone who has access to the open Internet. These are relatively new systems, with Ethereum being one of the oldest, having been created in 2015.

The relative novelty and complexity of these systems makes prudential regulators understandably nervous. In December 2022 the Basel Committee on Banking Supervision published rules on the Prudential Treatment of Cryptoasset Exposures. This is designed to ensure that banks and other regulated entities do not expose themselves to potentially significant risks by engaging in activity with cryptoassets. The rules include an infrastructure risk add-on that regulators can use based on any observed weaknesses in the blockchain infrastructure on which certain cryptoassets sit. In other words, if there are failures in the blockchain platforms that cryptoassets and stablecoins use, then banks that engage in activities with those assets may in turn need to add to their risk-weighted assets. This add-on started at a flat 2.5% during the consultation phase for the new rules but has since been amended to allow regulators to exercise discretion.

If one were to take the prudential rules that are being proposed for cryptoasset platforms and apply them to RTGS fiat payment platforms, then there is an argument that an infrastructure risk add-on should be applied to anyone engaging in USD, EUR, GBP and AUD fiat payments (as well as all the currencies whose outages ‘proved possible to conceal’ to paraphrase the World Bank).

This article illustrates that neither cryptoasset systems nor RTGS systems are perfect. Both can suffer outages, as was seen with Ethereum in May 2023 and Bitcoin in 2010 and 2013. What is unhelpful and potentially dangerous is pretending that one system is infallible by concealing failure while penalising the transparent system with infrastructure risk capital charges. A varied ecosystem that has the capacity to thrive when under stress or attack rather than just be resilient (i.e., able to recover quickly) or robust (i.e., able to withstand failure) should be the ultimate goal. Those designing market structure should avoid being wedded to a one-size-fits-all approach, since that will ultimately lead to a fragile system with a single potentially catastrophic point of failure.

Industry bodies like the Stablecoin Standard and digital asset ecosystem providers like Zodia have as part of their mandate a desire to engage with stakeholders so that the market structure is designed in a robust manner across legacy and digital channels.

Interested to learn more? Contact us.