How RTGS inadvertently killed system liquidity

    Twitter
    Facebook
    LinkedIn

Print this page
11
8 hours ago

By: Izabella Kaminska

Long before blockchain was considered a transformational ledger technology, something far less hyped was being treated as equally revolutionary: the transition to real-time gross settlement systems (RTGS).

RTGS might sound like a meaningless jargon acronym that you shouldn’t care about, but it’s probably one the most important and least-talked about technologies in the world. It is the tech that underpins central bank clearing and the settlement of large value payments by allowing the transfer of funds between banks in real time.

As a result, it is the anchor of the world’s value systems and the proverbial beating heart of the financial flows which lubricate the global economy.

But it is also a tech that fundamentally changed the way central banks managed system liquidity and its cost.

To really determine whether concepts like central bank digital currencies (CBDCs) are a logical evolution, it’s thus essential to step back in time as far back as the 80s when the big fintech discussions of the day were centred not around distributed ledgers or cryptocurrencies but RTGS systems.

Killing off Herstatt risk

By the 80s central bankers had become acutely aware that the netting systems of the day -- hailing centuries back to the days of “town clearing” which saw couriers physically run cheques to city hubs for end-of-day settlement rounds -- were struggling to keep up with the gigantic intraday credit exposures that were accumulating in the banking system due to expanded international trade and commercial activity.

The first clue something was up came following the collapse of the German lender Herstatt in 1974 due to FX trading-related losses, an incident that mentally scarred a generation of central bankers. The trauma arose not because the bank’s risk management practices were exceptionally bad, but rather because the size of the bank’s intraday FX (and de facto payments) settlement exposures and its ultimate interlinkage with the US-dollar clearing system had been grossly under appreciated by the market. The domino effect that came to threaten the banking system in the wake of the bank’s collapse soon became known as “Herstatt risk” -- a scenario where well-managed banks become threatened by the actions of a singular bad actor because their unsettled exposures vis-a-vis the collapsed bank are so huge.

And so it was that by the early 90s, most central banks had been sold on the idea that the only way to control Herstatt risk was to do away with netted settlements entirely. The theory was simple: if banks couldn’t accumulate extended levels of debt between each other throughout the day because they were settling on-the-go, they couldn’t threaten other banks if they collapsed the way Herstatt did.

The Fed in that regard was both a trailblazer and a laggard. It led because it alone had established a de facto real-time settlement system for large value transactions as early as the 1920s (utilising Morse code). By the time Herstatt came to pass that system, called Fedwire, had even progressed to a computerised format.

Nonetheless, severe restrictions still impacted Fedwire’s usability. First off, its operational hours were strictly limited, reducing its capacity to deal with timezone discrepancies in the FX market. Second off, Fedwire was expensive to use relative to netting systems -- encouraging banks to still depend on netted alternatives wherever possible. (This remains the case today for all RTGS systems due to the additional liquidity pressures they impose on banks -- but more on that later). Last, the Fed’s early foray into RTGS was meaningless unless other central banks processing the corresponding legs of FX transactions were also on similar systems.

What’s important is that by the time of Herstatt, banks processing large USD FX transactions were still inclined to use netted alternatives like the Clearing House Interbank Payments System (CHIPS) -- something that eventually increased the exposures significantly.

The impetus for all the world’s top central banks to adopt real-time gross settlement technologies for large value transactions and expand their operational hours so they would overlap between them* thus became clear.

It’s the liquidity stupid!

Except! Even then it was understood that it was impossible for the financial sector to have its cake and eat it. Yes, Herstatt risk could in theory be entirely eliminated with RTGS, but banks would have to pre-fund all their transactions, something that stood to impose huge liquidity costs on banks.

Under the legacy netting system banks had simply provided each other with zero-cost intraday credit on the basis they were operating with a small cadre of trusted counterparties.

Under RTGS this could no longer be the case. Banks needed funding not credit because without such funds in situ, real-time settlement could not be contractually achieved. That pre-funding need, however, would heighten the system’s sensitivity to logjams imposed by single institutions, and with it threaten total system gridlock.

To prevent this, central banks quickly realised they would have to stand ready to provide additional liquidity whenever the system required it. But this, they appreciated, would have implications on how they conducted monetary policy both in terms of liquidity provision and rate transmission.

Writing in 1992, then Bank of England governor Robin Leigh-Pemberton highlighted the issues thusly (our emphasis):

    Precisely because it has been no more than a by-product of existing systems, the liquidity -- in other words the interbank credit -- which has enabled them to operate efficiently has not only been uncontrollable and therefore potentially unlimited, it has also been provided free of charge. That being so, the introduction of a system depending on explicit transactions for liquidity is bound to entail new, explicit costs. Instead of a hidden creation of liquidity, banks participating in a real-time gross settlement system need to ensure that there are adequate funds on their central bank accounts, or to raise funds in the money markets, or to make use of central bank credit facilities; each of which is bound to involve a cost, whether in the form of a financial charge or the loss of an earning opportunity.

No central bank rolling out RTGS approached the problem the same way.

In the case of Bank of England, as the this 1996 BoE working paper notes, it was recognised that the unfeasibility (in terms of opportunity cost) of banks holding sufficient settlement balances to cover their largest net outflows, would see banks become dependent on financing at least part of their net outflows by borrowing intraday funds. It was also recognised there were only two possible sources for these funds, the central bank or the interbank market.

The BoE’s long-term hope was that an interbank market for intraday would eventually develop. But in the interim the central bank would have to bridge the gap.

When the Bank launched RTGS in 1995 it did so alongside a policy that provided intraday liquidity free-of-charge via repos to combat gridlock risk. This liquidity was theoretically unlimited, with the only constraint the minimum quality of the collateral accepted by the Bank.

Some central bank practitioners worried shifting the BoE’s regime to one that effectively gave away free intraday liquidity could impact the way monetary policy was transmitted to the real economy. But the counter argument was that the zero cost was necessary to lessen the burden and cost to banks (and society in general) of transferring the BoE to an RTGS system. Besides, it was also anticipated, monetary policy could still be enacted through overnight rates and beyond.

The ECB’s model was even more generous in the conditions it offered to banks -- not least because the euro-system was designed as an RTGS one from inception and the ECB always had the objective of making the transition to the euro as painless as possible. Like the BoE it too aimed to provide intraday liquidity on-tap but did so against an even wider range of collateral. Unlike the BoE, it did not aim to penalise players for rolling this liquidity on overnight.

Importantly both of these systems (as well as many others that were launched by other central banks around the same time) differed to the key components of the much more longstanding Fed model.

Over in the US, the objective was always to minimise the amount of excess central bank balances banks held intraday. While banks were free to drawdown their required reserves throughout the day at no cost to meet such funding needs, any additional intraday liquidity from 1994 onwards was only available through an overdraft credit facility at the Fed at a charge. Unlike the European intraday liquidity support mechanism this credit facility was available on an uncollateralised basis but very specifically only to creditworthy institutions. Unlike the discount window for overnight funding, drawing credit was not stigmatised. It was, however, costly and capped.

Appreciating the trends in the wider central banking world and the growing demand for intraday liquidity by some institutions, the Fed, however, decided to make additional liquidity available on a collateralised basis from 2008 onwards.

Here’s how the Fed’s intraday credit facility was utilised in the run-up to the crisis and how that use crashed in the aftermath of QE:

 So why is any of this important or even related to central bank digital currencies?

There are three key points that emanate from all this.

The first, as a report by RTGS experts Peter Allsopp, Bruce Summers and John Veale for the World Bank in 2009 spelt out, is that the lack of harmonisation between central banks in terms of their RTGS approaches has led to a number of practical problems for central banks, users and other beneficiaries. One relates to the inconsistent terms on which central bank credit is available, especially against collateral, at times of stress.

It is FT Alphaville’s contention that there is an intimate and under appreciated relationship between system liquidity needs (and costs) and the RTGS technology central banks use to provide settlement services because these systems do not function smoothly unless commercial banks maintain sufficiently large reserves to cover payment settlement risk.

Second, RTGS’ sensitivity to logjams in the context of expensive intraday liquidity makes the system particularly vulnerable to the hoarding of receipts by banks, which imposes even greater liquidity demands across the system. An efficient RTGS is one in which available liquidity provided by incoming payments is promptly recirculated among the participants. But in that context there is always a risk -- just in terms of a general unevenness in payment flows - that such liquidity ends up pooling at one institution, creating distribution issues.

But as the report also notes, an overnight loan that emerges from such an imbalance is practically speaking an intraday loan that has not been repaid by the close of business. In that sense, it can be argued, central banks which injected overnight funds to support their interbank markets from 2008 onwards were de facto harmonising their intraday and overnight credit provision policies by extending the generous terms of the former for longer periods.

Third, the same report noted that access to central bank settlement services had been a prominent consideration in the transition from planned and socialist economies to modern market economies, as was the way these systems should be ultimately funded.. Notably, it was determined that peak efficiency would emerge from wide accessibility for institutions offering deposit money accounts as well as general non-subsidisation of the system. (Banks in general are charged for the use of RTGS on a transaction and membership basis.)

Currently, there is a major inconsistency across central banks on this front, which leads to the possibility some cbanks may be inadvertently importing risk associated with subsidisation as well as varying policies towards non-bank accessibility into systems not geared up for such risks.

The introduction of a central bank digital currency at one institution but not at others -- CBDC equalling the broadest extension of central bank settlement services in terms of accessibility and payment size -- would only heighten such inconsistency risk. It would also introduce a new level of uncertainty about intraday liquidity demand and how exactly to responsibly provide it, especially to the central bank itself.

Last, if CBDCs were to be rolled on RTGS systems (or even blockchain equivalent ones) they would likely require significant subsidisation to be affordable for everyday users given the increased processing and liquidity burdens that would come with it.

*The instatement of major central bank RTGS systems with overlapping working hours was an effective pre-requisite for the development of a functional continuous linked settlement (CLS), the ultimate mechanism for killing off Herstatt risk.

The development of payment-versus-payment (PVP) systems also helped to curb such risks, but as the World Bank report also notes, RTGS systems can be considered “a single point of failure across the entire financial system, in that participants face liquidity impacts from all of their delivery-versus-payment (DVP) and (PVP) links, as well as settlement links to clearing houses and other elements of financial markets infrastructures, sometimes in multiple countries and currencies simultaneously”.

Related links:
A Market for Intra-day Funds: Does it Have Implications for Monetary Policy? - Bank of England
The development of a UK real-time gross settlement system - Bank of England
REAL-TIME GROSS SETTLEMENT SYSTEMS - BIS
The Fedwire® Funds Service Assessment of Compliance with the Core Principles for Systemically Important Payment Systems - Federal Reserve
Large-Value Payment Systems: What Have We Learned? - Wayne Angell comments at 12th payment systems international conference 1992.
THE EVOLUTION OF REAL-TIME GROSS SETTLEMENT - Word Bank
Developments in wholesale payments systems - Bank of England
Intraday Liquidity Management: A Tale of Games Banks Play - New York Fed
A story about a liquidity regime shift - FT Alphaville