Interbank funds transfer systems are arrangements through which funds transfers are made between banks for their own account or on behalf of their customers. Of such systems, large­value funds transfer systems are usually distinguished from retail funds transfer systems that handle a large volume of payments of relatively low value in such forms as cheques, giro credit transfers, automated clearing house transactions and electronic funds transfers at the point of sale. The average size of transfers through large­value funds transfer systems is substantial and the transfers are typically more time­critical, not least because many of the payments are in settlement of financial market transactions. The report focuses on these large­value systems.

The processing of funds transfers involves two key elements. The first of these is the transfer of information between the payer and payee banks. A funds transfer is initiated by the transmission of a payment order or message requesting the transfer of funds to the payee. In principle, the payment messages may be credit transfers or debit transfers, although in practice virtually all modern large­value funds transfer systems are credit transfer systems in which both payment messages and funds move from the bank of the payer (the sending bank) to the bank of the payee (the receiving bank). The payment messages are processed according to predefined rules and operating procedures. Processing may include procedures such as identification, reconciliation and confirmation of payment messages. The transmission and processing of payment messages in large­value funds transfer systems is typically automated (i.e. electronic).

The second key element is settlement - that is, the actual transfer of funds between the payer's bank and the payee's bank. Settlement discharges the obligation of the payer bank to the payee bank in respect of the transfer. Settlement that is irrevocable and unconditional is described as final settlement. In general, the settlement of interbank funds transfers can be based on the transfer of balances on the books of a central bank (i.e. central bank money) or commercial banks (i.e. commercial bank money). In practice, settlement in the vast majority of large­value funds transfer systems takes place in central bank funds. Although the rules and operating procedures of a system and the legal environment generally may allow for differing concepts of finality, it is typically understood that, where settlement is made by the transfer of central bank money, final settlement occurs when the final (i.e. irrevocable and unconditional) transfer of value has been recorded on the books of the central bank. The report focuses on the settlement finality of the central bank transfers in this sense.

Settlement characteristics and the types of interbank funds transfer system. Interbank funds transfer systems can be classified in several ways. Among other things, differences in the way settlement takes place provide a useful framework for the discussions in the later sections of this report. A common distinction in this respect is to divide systems into net settlement systems and gross settlement systems. In a net settlement system, the settlement of funds transfers occurs on a net basis according to the rules and procedures of the system. A participating bank's net position is calculated, on either a bilateral or a multilateral basis, as the sum of the value of all the transfers it has received up to a particular point in time minus the sum of the value of all the transfers it has sent. The net position at the settlement time, which can be a net credit or debit position, is called the net settlement position. Net settlement systems for large­value funds transfers in the G­10 countries are now primarily multilateral (rather than bilateral) net settlement systems in which each (settling) participant settles its multilateral net settlement position. In a gross settlement system, on the other hand, the settlement of funds occurs on a transaction­by­transaction basis, that is, without netting debits against credits.

Interbank funds transfer systems can also be classified according to the timing (and frequency) of settlement. Systems can in principle be grouped into two types, designated­time (or deferred) settlement systems and real­time (or continuous) settlement systems, depending on whether they settle at pre­specified points in time or on a continuous basis. In this report, these two types are more narrowly defined in terms of the timing of final settlement. One type of system is thus a designated­time (or deferred) settlement system, in which final settlement occurs at one or more discrete, pre­specified settlement times during the processing day. Designated­time settlement systems in which final settlement takes place only once, at the end of the processing day, are called end­of­day settlement systems. Currently, net settlement systems for large­value transfers are typically end­of­day net settlement systems that settle the net settlement positions by means of transfers of central bank money from net debtors to net creditors. In some countries, there are systems in which the final settlement of transfers occurs at the end of the processing day without netting the credit and debit positions - on a transaction­by­transaction basis or on the basis of the aggregate credit and aggregate debit position of each bank. Such systems are often called end­of­day gross settlement systems. On the other hand, a real­time (or continuous) settlement system is defined as a system that can effect final settlement on a continuous basis during the processing day. RTGS systems, as defined below, fall into this category. Table 2 summarises the main possibilities.

It is worth stressing here that the distinction between different systems such as RTGS and designated­time net settlement (DNS) systems concerns the form of settlement, not the form of transmission and processing. Like RTGS systems, many net settlement systems transmit and process payment messages in real time on a transaction­by­transaction basis, but they settle, by definition, on a net basis at discrete intervals.

An important concept that is often used in connection with the timing of finality is intraday finality or an intraday final transfer capability. The Noël Report defined this concept as the ability to initiate - and to receive timely confirmation of - transfers between accounts at the central bank that become final within a brief period of time. The Study Group believes that this definition is useful and in practice is sufficient when discussing the "intraday" nature of finality relative to end­of­day finality. However, the Group also recognised that the phrase "within a brief period of time" is not a precise one. Therefore, care needs to be taken when considering the extent to which systems in which final settlements occur at discrete but very frequent intervals of time during the day can provide some form of intraday finality similar to systems involving continuous settlement (i.e. real­time intraday finality).

In describing the settlement characteristics of net settlement systems, the concept of certainty of settlement is sometimes used. This concept is not related to the timing of final settlement per se but, as described in the Noël Report, refers to the certainty that the system will be able to effect final settlement when the netting cycle and the associated settlement procedures have been completed. Certainty of settlement relates to a multilateral netting system's ability to meet Lamfalussy Standard IV, namely that such systems should, at a minimum, be capable of ensuring the timely completion of daily settlements in the event of an inability to settle by the participant with the largest single net debit position. Multilateral net settlement systems that are secured in compliance with all the Lamfalussy standards, including Standard IV, and that can therefore assure settlement in the event of any single participant failure are often called Lamfalussy­compliant systems. Stronger forms of certainty of settlement arise when a net settlement system is capable of ensuring settlement in the event of more than one participant failing. These systems are often called Lamfalussy­plus systems. A particular category of Lamfalussy­plus systems concerns those that can assure settlement in all circumstances, that is, regardless of how many participants fail.

Central bank systems and private sector systems. Interbank funds transfer systems are sometimes classified according to whether they are central bank systems or private sector systems. The distinction typically depends on who owns and operates the systems (rather than on the identity of the settlement agent). At present, it is possible to identify two "typical" types of large­value funds transfer system: (a) central bank systems owned and operated by the central bank (or its affiliated entities) in which the central bank also provides settlement, and (b) private sector systems owned and operated by a private sector group (e.g. a banking association or clearing house), where the main operational role of the central bank is to act as the settlement agent. In the G­10 countries, for example, RTGS systems often belong to the former category and many DNS systems belong to the latter. Nonetheless, a number of DNS systems are owned and operated by the central bank, while in some cases RTGS systems are owned and operated by a private sector group (see Section II.1). Moreover, there are several DNS and RTGS systems in which ownership and operation are shared between the private sector and the central bank. 2. Types and sources of payment system risk

Earlier CPSS reports identified the major types of payment system risk. Credit risk and liquidity risk are two basic risks to which participants in payment and settlement systems may be exposed. Credit risk, which is often associated with the default of a counterparty, is the risk that a counterparty will not meet an obligation for full value, either when due or at any time thereafter. It generally includes both the risk of loss of unrealised gains on unsettled contracts with the defaulting counterparty (replacement cost risk) and, more importantly, the risk of loss of the whole value of the transaction (principal risk). Liquidity risk refers to the risk that a counterparty will not settle an obligation for full value when due but at some unspecified time thereafter. This could adversely affect the expected liquidity position of the payee. The delay may force the payee to cover its cash­flow shortage by funding at short notice from other sources, which may result in a financial loss due to higher financing costs or to damage to its reputation. In more extreme cases it may be unable to cover its cash­flow shortage at any price, in which case it may be unable to meet its obligations to others. Settlement risk may be used to refer to the risk that the completion or settlement of individual transactions or, more typically, settlement of the interbank funds transfer system as a whole, will not take place as expected. Settlement risk comprises both credit and liquidity risks.

Two major sources of these risks are (a) a time­lag between the execution of the transaction and its final completion and (b) a time­lag between the completion of the two legs of the transaction (i.e. any lag between payment leg and delivery leg). Within large­value funds transfer systems the first type of lag, which takes the form of a settlement lag between the initiation of payment messages and their final settlement, can be a major source of settlement risk. Settlement lags create the possibility that sending banks could fail in the meantime or at least not be able to settle their obligations when due.

Settlement lags can result in credit risk if the two functions of an interbank funds transfer system discussed in the previous subsection (namely the transmission of information about the payment and the settlement of the payment) do not occur simultaneously, so that settlement takes place after the information has been provided. As long as final settlement has not occurred, any payment activity undertaken on the basis of "unsettled" payment messages remains conditional and results in risk. For example, because of competitive pressures or system rules resulting from customers' demand for a fast intraday payment service, receiving banks may credit funds to their customers on the basis of the receipt of incoming payment messages and before final settlement; receiving banks are then exposed to principal risk because, if settlement does not occur (e.g. because of the default of a bank) and they do not receive the anticipated funds at the settlement time, they may not be able to retrieve these transfers. Settlement lags may also result in liquidity risk. Until settlement is completed, a bank may not be certain what funds it will receive through the payment system and thus it may not be sure whether or not its liquidity is adequate. If, in planning its liquidity needs, a bank overestimates the funds it will receive when settlement takes place, then it may face a shortfall. Indeed, if the shortfall occurred close to the end of the day, a bank could have significant difficulty in raising the liquidity it needed from an alternative source. See Box 1 for a description of settlement lags as possible sources of interbank risk.

BOX 1 SETTLEMENT LAGS AS SOURCES OF INTERBANK RISK
The time axis below describes the stages a payment instruction may go through during the lag between being entered into an interbank funds transfer system (at time t1) and being settled (at time t4). This lag is often referred to as a settlement lag.
t1 Payment message entered into system	t2 Partial information available to receiving bank or full information available only onrequest	t3 Full information sent automatically to receiving bank	t4 Interbank settlement takes place
In some systems the payment message is sent automatically to the receiving bank before interbank settlement takes place. This intermediate point is designated here as time t3. As described in the main text, between t3 and t4 there is a possibility that the receiving bank will act on this information in a way that exposes it to credit and liquidity risk. For example, even though the receiving bank cannot yet be certain that it will receive the funds from the sending bank (it will not know this until settlement has taken place at t4), it may nevertheless credit the funds to the account of its customer (the beneficiary of the payment) or it may commit itself to lending the funds in the interbank market. In other words, it anticipates the interbank settlement. In some cases there may also be an earlier point, between t1 and t3, at which the receiving bank receives partial information about the payment being processed, or at which it can receive the full information but only on request. This point is designated t2. In the first stage, between t1 and t2, no information on the incoming payment is provided by the system. However, the receiving bank may be aware of the incoming funds from other sources (for example, the sending bank may have passed the information directly to the receiving bank). The extent to which each of these portions of settlement lag may contribute to settlement risk depends on the characteristics of the individual payment system as well as the practices of banks in debiting and crediting payment transfers on their own books. The topic is discussed further in subsection II.4.2.

The risk implications of the timing of delivery of information about funds transfers preceding the timing of settlement have become increasingly important as various types of large­value funds transfer system have come to operate on the basis of real­time processing (i.e. information is transmitted to receiving banks in real time), while settlement may be delayed (either because the system is a DNS system or because, in an RTGS system, liquidity constraints may delay settlement at least temporarily).

The second type of lag, sometimes referred to as asynchronous settlement, is the largest source of principal risk in the settlement of foreign exchange and securities transactions, or, more generally, in exchange­for­value systems. This is the risk that the seller of an asset could deliver but not receive payment or that the buyer of an asset could make payment but not receive delivery, which could entail a loss equal to the full principal value of the assets involved. The DVP Report concluded that a delivery­versus­payment system, which ensures that the delivery occurs if and only if payment occurs, would provide a mechanism for eliminating such principal risk. Since the payment leg in an exchange­for­value system is supported by interbank funds transfer systems, the settlement characteristics of interbank funds transfer systems as described above have an important influence on how a DVP mechanism could be constructed for an exchange­for­value system.

Systemic risk. As overseers of payment systems, central banks are particularly concerned with systemic risk. This is the risk that the failure of one participant to meet its required obligations when due may cause other participants to fail to meet their obligations when due. Such a failure could trigger broader financial difficulties that might, in extreme cases, threaten the stability of payment systems and even the real economy. By their very nature networks, interbank payment and settlement systems are potentially a key institutional channel for the propagation of systemic crises. Central banks have a particular interest in limiting systemic risk in large­value funds transfer systems because aggregate exposures tend to increase with the aggregate value of transactions and potential risks in large­value transfer systems are therefore often significantly higher than those in retail funds transfer systems.

The vulnerability of a system to systemic risk depends on a number of factors. As analysed in the Lamfalussy Report, the size and duration of participants' credit and liquidity exposures in the interbank settlement process are basic factors affecting the potential for systemic risk. As these exposures last for longer and become larger, the likelihood that some participants may be unable to meet their obligations increases, and any participant's failure to settle its obligations is more likely to affect the financial condition of others in a more serious manner. Interbank funds transfer systems in which large intraday exposures tend to accumulate between participants therefore have a higher potential for systemic risk.

The potential for systemic risk may also be related to the propensity of systems to give rise to unexpected or sudden settlement obligations for participants. Such propensity may depend on the extent to which payment and settlement are conditional. A specific case concerns the unwinding of funds transfer obligations in the event of settlement problems. In some systems, if one participant is unable to settle its multilateral net (debit) settlement position and the position cannot be covered in another way, the system will recalculate a new set of net settlement positions for each of the remaining participants by deleting some or all of the (provisional) transfers made to and from the defaulting participant. However, in systems where positions are not controlled, this recalculation could lead to an unexpected and sizable change in the remaining participants' settlement obligations. For instance, if bilateral "net credits" due from the defaulting participant are no longer available, the remaining participants are likely to find that their net settlement obligations have increased. This could in turn seriously affect the financial position of these participants, which might result in further knock­on effects.