Alan L Tyree

Regulating the payment system - Part 3 - Financial stability

Alan L Tyree*

1  Payment system risks

Every day, financial institutions in Australia receive payment instructions totalling more than $100 billion. Until last year, each of these payment instructions resulted in a short term debt between financial institutions, a debt that was not extinguished until the following morning at 9am. Until last year, these short term debts were treated the same as any other debt of the financial institution.1

These short term obligations are massive in comparison with the normal borrowings of an organisation. There is also a sense that they should be treated differently from the normal borrowings. The payment system operates by providing a system for the circulation of institutional liabilities. These large short term obligations arise almost by a legal accident from the operation of the system. It would be possible to imagine a payment system where the liability of Financial Institution A to a debtor D could be transformed into a liability of Financial Institution B to D’s creditor C without ever giving rise to new liabilities between A and B. But that is not our system.

The legal structure of the payment system was, and is, quite different from the way that financiers perceived the system to work. Obligations arising from payment instructions were “netted” between the participants in the payment system so that the actual payments of the resulting debts involved relatively small transfers when the debts were settled the following morning. Although concerns were expressed from time to time about the legal effectiveness of netting, the concerns did not become acute until the Privy Council decision in British Eagle International Airlines Ltd v Compagnie Nationale Air France2 where it was held that a multi-lateral netting agreement, similar in all respects to the agreement governing the settling of payment system obligations, was void as a preference. The Privy Council held that a private agreement between the parties could not have the effect of prejudicing other creditors. See [] for a full discussion of the British Eagle case.

The other serious problem with payment system settlement arrangements was the so-called “zero hour rule”. This rule deems an insolvency to begin at the “zero hour” of the day on which the insolvency occurs. The result is that payments made any time after midnight may be treated as revocable. In particular, the payment made at the 9am settlement might be deemed to be void. The resulting problem of “unwinding” payments would be a nightmare.

Although these problems were understood for some time, little importance was attached because of the size of the participants in the payment system. The only significant players were the large banks, and it was long assumed3 that it was to all intents and purposes impossible that one of them should fail to settle.

The expansion of the payment system to include more participants has changed that perception. The development of electronic payment mechanisms in the 1980s made it increasingly difficult for the banks to exclude smaller financial institutions and larger retailers from the payment system. With increased participation came the realisation that failure of a payment system participant was not impossible, and regulators began to seriously examine the consequences of such a failure.

These consequences go under the general name of “settlement risks”. Some particular settlement risks have more colourful names. “Contagion risk” is the risk that a “neighbour” of the failed institution, that is, one which is owed money by, or which owes money to, the failed institution may itself be unable to meet its settlement obligations. “Systemic risk” is the risk that the contagion becomes so widespread that all major financial institutions are brought down and civilisation as we know it comes to an end.

Settlement risks are real and very dangerous, and management of the risks has become a major endeavour for financial regulators. Risk management currently depends upon:

2  Limiting Access

When cheques were the only effective method of making third party payments from deposit accounts, management of settlement risk took the exclusive form of ensuring that the participants in the system could not fail. Since cheques, by definition, could only be drawn on a “bank”, banks were the only major participants in the payment system.

In recent years, the government has never shared the view of the banks that being a bank is itself a guarantee against failure. Various mechanisms of control have been used over the years to ensure that banks remain solvent. These have included restraints on lending, both in amount and for purpose, and requiring banks to maintain deposits with the central bank. More recent approaches have focussed on “prudential” controls which require the institution, in effect, to hold certain forms of assets.

As noted above, the development of electronic forms of payment broke the monopoly of the cheque as the only effective payment mechanism. This development also made it increasingly difficult for the banks to maintain a monopoly on the payment system. Participation by smaller institutions became inevitable, both technically and politically.

Following recommendations of the Wallis Report, legislation has implemented a uniform system of regulation for participants in the payment system. Most payment system participants will be engaging in “banking business” within the meaning of s5(1) of the Banking Act 1959.4 As such they must be Authorised Deposit Taking Institutions (ADIs).

ADIs are subject to prudential supervision by the Australian Prudential Regulatory Authority (APRA). APRA adopted the prudential regulation guidelines established by the Reserve Bank and in effect at the time when prudential regulation responsibility passed from the Reserve Bank to APRA. Credit Unions and Building Societies came under the authority of APRA from 1 July, 1999, and are deemed to have received authority to carry on banking business.5

It might be possible for an organisation to be a participant in the payment system without being an ADI. Should this become a problem, there is a further possible restriction on entry.6 The Reserve Bank, acting through the Payment Systems Board, may “designate” a payment system and may then give directions concerning access regimes and standards to be complied with by the participants.7

“Purchased payment facilities” are subject to special considerations which will be the subject of the next article in this series.

3  Controlling exposure risks

Prudential regulation attempts to ensure that the participants can meet their exposure risks. The second limb of regulation is to limit those exposure risks. There are four methods currently used to limit exposure risks:

Exposure risks could also be limited by putting an absolute limit on exposure amounts for each institution. Other methods involve shifting exposure risks from one institution to another. This may be done using guarantees or a market in exposure risks.8

3.1  Real time gross settlement

The ultimate form of settlement risk reduction is to have settlement occur at the same time as the receiving institution becomes liable to the ultimate payee. This is not feasible for “small” payments, but with the assistance of computer technology it is practicable for “large” payments. Such a system of settlement is known as “Real Time Gross Settlement” (RTGS). What is “small” and “large” will change with time as the technology becomes faster and cheaper.

In Australia, the RTGS system is built on the Reserve Bank’s RITS (“Reserve Bank Information and Transfer System”).9 There are several “feeders” into RITS, but for present purposes the most important is the SWIFT PDS.10 Banks may initiate payment instructions from SWIFT terminals. Transactions are settled individually as they occur using Reserve Bank Exchange Settlement Accounts (“ESA”).

While RTGS reduces settlement exposures, it introduces a problem for funds management. If there is a time-critical payment to be made, then the paying institution must ensure that funds are available to complete it at the critical time. RITS assists in this by allowing paying institutions to assign priority levels to payment instructions. Each transaction may be identified as “deferred”, “active” or “priority”. Active transactions will be processed immediately unless processing would cause the ESA to fall below a level predetermined by the institution. Priority transactions are processed using the full amount of the ESA regardless of any predetermined limits.

Liquidity is also assisted by the operation of the RTGS system. The system searches for “off-setting” transactions which will result in simultaneous settlement between two of the participants. This feature results in diminishing the demands on each of the participants. The system also incorporates an automatic or manual system for converting certain securities into ESA funds provided that they are repurchased at the end of the day.

RTGS has been very successful in Australia since its introduction in June 1998. According to [, page 14], some 90are now made using the RTGS system.

3.2  Payment Systems and Netting Act

Although RTGS has succeeded in removing much of the settlement risk from the payment system, there is still a substantial number of “small” payments that are not settled on an individual basis. In addition, of course, Australians still write some 4 million cheques each day, each cheque representing a payment order that is cleared and then settled under a deferred net settlement scheme. The daily value of the cheque clearing is over $14 bn per day: see [, Table 13, p 48]. In addition, “consumer” electronic funds transfers, direct debits and credits are also netted and settled the next morning.

Section 10 of the Payment Systems and Netting Act 1998 provides the legal underpinnings for an “approved” multilateral netting arrangement. By s12, the Reserve Bank may “approve” a multilateral netting arrangement provided it is satisfied, inter alia that systemic disruption in the financial system could result if a participant went into external administration and the arrangement were not approved under this section.

The Reserve Bank has not yet made any approvals under s12 of the Act.

3.3  The zero hour rule

As noted above, the “zero hour” rule has the potential to “unwind” a settlement. For certain “approved” systems, the rule is abolished by the provisions of the Payment Systems and Netting Act 1998. A “participant” in an RTGS system who goes into external administration is deemed, for the purposes of settlement payments made through the RTGS system, to have gone into administration on the following day.11

The Reserve Bank is given the authority to approve an RTGS system under s9 of the Act if it is satisfied, inter alia, that systemic disruption in the financial system could result if a participant went into external administration and the system were not approved under this section.

The Bank has given approval under section 9 to RITS and to the Austraclear System FINTRACS.

3.4  Turnback of cheques

The Cheques Act 1986 has been amended to permit the “turnback” of certain cheques drawn on a “failed institution”. The problem addressed by the new sections was the uncertainty about the rights to “unwind” certain cheques if the drawee failed. It was thought that the deposit institution might become liable to the depositor of the cheque at some time before it received value for the deposit.

Section 70A provides that a cheque is deemed to be dishonoured if it is lodged for collection with an institution other than the drawee institution and the drawee institution becomes a “failed financial institution”

The deemed dishonour is taken to occur at the time when the drawee institution becomes a failed financial institution. That time is specified by s70A(2). The circumstances include the ordinary circumstances of insolvency and the situation where APRA appoints an investigator under section 13A of the Banking Act 1959 and determines that the institution is to be treated as a failed financial institution for the purposes of the Turnback provisions.

The Act also defines the circumstances under which a “cheque has not been settled” for the purposes of s70A. A cheque had not been settled if and only if:

The Reserve Bank is given the power to “recognise” a settlement system by s70A(4). As of the time of writing, no systems have been “recognised” by the Bank.


Consultant, Mallesons Stephen Jaques, Sydney; formerly Landerer Professor of Information Technology and Law, University of Sydney.
“Deposits” held by banks were treated differently by virtue of the then s16, Banking Act 1959; see now s13A(3).
[1975] 2 All ER 390; [1975] 1 WLR 758
Wrongly, as the events of the early 90s showed.
In the next article in this series, we will examine the potential for participation in payment systems without carrying on “banking business”, at least as defined in the Act.
Financial Sector Reform (Amendments and Transitional Provisions) Act 1998 and the Regulations under that Act in 1999.
This is slightly speculative. The definition of “access” in the Act is wide enough to include “exclusion”, but the Act generally seems to be contemplating the situation where the Bank would order a system to accept a participant that is otherwise excluded.
See s10, Payment Systems (Regulation) Act 1998 which provides an “overview” of the Act.
See the PRESS system developed in Australia and then abandoned in favour of RTGS.
There is a complete description of the RTGS system in []
SWIFT is the Society for Worldwide Interbank Financial Telecommunications. The PDS is the “Payment Delivery System”, established in 1997 and controlled by APCA.
Section 6, Payment Systems and Netting Act 1998.