2.2 Using these diverse home-currency payments systems to settle cross-border and multi-currency transactions can be cumbersome and may entail considerable risk. For instance, when the hours of operation of two payments systems do not overlap, it is technically impossible to arrange for the simultaneous settlement of both sides of a foreign exchange transaction. Even when operating hours do overlap, local payments arrangements often make it difficult to control and coordinate the timing of payments in several currencies. More generally, differences and uncertainty in the timing of finality in each country present other obstacles to the effective management of the risks that arise in cross-border and multi-currency settlements.

2.3 Home-currency payments system hours and arrangements and the timing of finality thus make it difficult, if not impossible, to achieve settlements involving two or more currencies (i.e. multi-currency settlements) in which final transfers in one currency occur if and only if final transfers in the other currency or currencies also take place. In the absence of such a delivery-versus-payment (DVP) process for settling obligations in multiple currencies, significant credit risks can be present. These risks include the potential loss of principal (often called "principal" or "Herstatt" risk) that would arise if transfers in one currency became final while associated transfers in another currency did not take place. This is a significant risk to international market participants since the loss of principal in settling, for instance, a foreign exchange trade would dwarf any gain or loss that might have accrued to the counterparties to the original transaction.

2.4 Liquidity risks can also arise in the absence of a multi-currency DVP settlement process. For instance, a fear of incurring principal risk might lead some market participants to refuse to honour their obligations in earlier-settling currencies out of concern that a "suspect" counterparty would not be able to settle its associated obligations in later-settling currencies. The sudden interruption of a significant level of expected payment flows could cause serious liquidity problems for the counterparty and, hence, for other market participants that expect to receive payments from the counterparty. This liquidity risk might also spread to other payments systems in the same or other countries if concern about loss of principa1 during the settlement process became widespread

Summary of options

2.5 The Working Group examined a range of central bank payment and settlement services that might reduce these risks and increase the efficiency of cross-border and multi-currency settlements. The options that were considered by the Working Group included: (1) modifying or making available certain home-currency payment and settlement services; (2) extending the operating hours of home-currency large-value funds transfer systems; (3) establishing cross-border operational links between these payments systems; and (4) developing multi-currency payment and settlement services. These central bank service options were evaluated in circumstances where they would facilitate the settlement both of individual transactions and of a stream of transactions between two or more counterparties that have been netted through the operation of a private sector netting scheme.

2.6 Home-currency payment and settlement services. Certain home-currency payment and settlement services might be modified or made available to increase the level of support for international settlements. In particular, where they do not currently exist, settlement accounts and intraday final transfer capabilities could be made available by central banks to settle home-currency obligations related to cross-border and multi-currency transactions. An intraday final transfer capability is defined as the ability to initiate - and to receive timely confirmation of - transfers between accounts at the central bank of issue that become final within a brief period of time. It is important to recognise that the availability of these services would not be sufficient to permit the simultaneous settlement of obligations in all currencies. Hence these home-currency services could not, per se, eliminate the credit and liquidity risks that exist in the absence of multi-currency DVP capabilities.

2.7 Recognising that they cannot eliminate the risks associated with non-DVP settlement, intraday final transfers and settlement accounts nonetheless significantly improve the ability of market a participants to manage and control these and other settlement risks. Accurate information about when obligations in each currency are finally discharged would enable individual institutions and clearing houses to quantify and control more precisely and efficiently than they can at present the level and duration of exposures that may be incurred in the process of settling both individual and netted transactions in two or more currencies. It is apparent that the capability in each country's domestic large-value payments system to effect final transfers at any time of the business day between accounts at the central bank of issue is the foundation stone upon which risk reduction measures can be developed in respect of a range of domestic and cross-border transactions.

2.8 Operating hours of home-currency payments systems. The operating hours of home-currency large-value funds transfer systems could be extended to increase the level of support for international settlements. At one end of the range of possibilities, a modest lengthening of the operating hours of an individual payments system would reduce its current gap (or increase its current overlap) with the operating hours of payments systems in other countries. Towards the other end of the spectrum (which, in the extreme, could involve the large-value funds transfer systems in a number of countries operating round-the-clock), a major extension of the hours of several key home-currency payments systems could result in an operational overlap of most major currencies. Combined with the availability of final transfers over those systems, such an overlap could create the technical ability to conduct on the same value date a DVP settlement of all relevant currencies in which counterparties would be assured that payments in one currency would be made if and only if payments in all relevant currencies are made. This would help support the potential elimination of the credit and liquidity risks associated with the current lack of multi-currency DVP capabilities. However, without directly linking payments made over different large-value funds transfer systems, this assurance would have to come from private sector procedures that would use the available home-currency payment and settlement services during the extended hours of operation.

2.9 Cross-border links between payments systems. Another possible option might be the establishment of bilateral or multilateral cross-border links between large-value funds transfer systems in conjunction, where necessary, with an extension of their operating hours to increase the level of support for international settlements. In particular, direct operational and informational links could be created that would give participating central banks the joint capability to monitor, control and execute simultaneously final transfers over their respective home-currency payments systems. With such cross-border connections, central banks could directly provide the private sector with DVP settlement services for currencies with overlapping payments system operating hours.

2.10 Multi-currency payment and settlement services. Another possible option might involve the joint offering of multi-currency payment and settlement services. Multi-currency accounts and settlement facilities might be provided by the central banks of issue through a "common agent". Specifically, a central bank controlled common agent could accept deposits in multiple currencies and facilitate final transfers between these ac counts . A variant of thi s arrangement would involve one or more central banks acting as the common agent in providing multi-currency services. In both cases, an important issue is whether arrangements would be in place to provide assurances that sufficient liquidity would be available to complete settlement in the relevant currencies.

2.11 The purpose of jointly offering multi-currency services would be similar to that of creating an overlap in the operating hours (with or without direct operational and informational links) of the major large-value funds transfer systems: to provide the private sector with the technical ability to achieve DVP in the settlement of multi-currency obligations. With multi-currency services this would be accomplished by effecting settlement over operational accounts in each currency held either at the common agent or, in the case of the variant discussed above, at one or more individual central banks.

Summary of analysis

2.12 The Working Group's analysis of the various central bank service options indicates that each has potentially significant advantages and disadvantages. Where such facilities do not currently exist, the ability to initiate transfers between accounts at the central bank of issue that become final within a brief period of time when the local money market is open and liquid would add a capability that could substantially reduce the risks in the settlement of cross-border and multi-currency transactions. Such an intraday final transfer facility would have the effect of supporting the discharge of all home-currency settlement obligations, whether related to foreign exchange, cross-border or purely domestic transactions; it would thereby improve the overall safety and efficiency of the home-currency payments system. Furthermore, intraday finality is a necessary building-block for implementing the other potential central bank service options described in this report.

2.13 The benefits of intraday finality, however, also carry certain costs. To realise the desired risk reduction and improved efficiency conferred by intraday finality, private sector entities would need to develop and implement payment arrangements that use this capability in a safe and sound manner. In addition, for those central banks that currently do not offer account holders this capability, the implementation and operational costs of doing so could be major considerations. Furthermore, even though intraday finality might foster substantial reductions in risk, it cannot, by itself, eliminate the risks that exist in the absence of multi-currency DVP settlements. Nevertheless, intraday final transfer capabilities substantially improve the ability of the private sector to manage, and indeed reduce, the current risks that arise in cross-border and multi-currency settlements.

2.14 A modest lengthening of the operating hours of an individual payments system that creates or expands an operational overlap with payments systems for other currencies would increase the opportunity to conduct DVP settlements among those currencies. In the extreme, a major extension of the hours of certain home-currency payments systems (with or without direct links between them) or the development of central bank multi-currency services would create the technical ability to conduct a DVP settlement of all G-10 currencies on the same value date and thus could support the elimination of the current credit and liquidity risks associated with non-DVP settlements. In particular, multi-currency DVP settlements would eliminate the potential loss of principal that currently arises during the settlement of multi-currency transactions. Furthermore, by assuring market participants that they would not incur principal risk in the settlement process, a multi-currency DVP mechanism would encourage them to honour their settlement obligations even at times of market stress.

2.15 However, while facilitating the elimination of these important risks, multi-currency DVP settlements may also raise a number of important policy issues and concerns. One concern stems from the settlement linkages that would be created under a multi-currency DVP settlement mechanism. By virtue of the interdependencies created by directly linking the settlement of two or more currencies through a DVP mechanism operated by either central banks or the private sector, a disruption in the settlement of one currency would disrupt the settlement of other currencies. Such disruptions could result from public or private sector operational problems, or from private sector liquidity shortages. Multi-currency DVP settlements that rely on "off-hour" money markets for funding in one or more major currencies would be susceptible to liquidity problems if these off-hour markets were not sufficiently deep and liquid to provide a reliable source of funds. It is possible, or even probable, that money markets would eventually evolve into an adequate source of liquidity to support off-hour settlements in each currency. Until then, however, implementing this capability could involve a trade-off between different sources of systemic risk.

2.16 Another issue is the high degree of coordination that might be needed (on a regular basis and when dealing with settlement problems) to support multi-currency DVP settlements. Enhanced central bank coordination has potential advantages and disadvantages. On the one hand, growing interdependencies among the world's financial markets have increased the benefit to individual central banks of more accurate and timely information flows, especially at times of financial stress. In this regard, the operational and informational links that might be created in conjunction with multi-currency DVP settlements, as well as the formalised relationships that would accompany the establishment of a central bank controlled common agent, would provide mechanisms for enhanced central bank coordination.

2.17 On the other hand, the interdependencies created by directly linking the settlements of two or more currencies through DVP procedures might also reduce the ability of individual central banks to respond to liquidity problems in their home currencies. By definition, multi-currency DVP settlement means that payments in one currency will be made if and only if payments in all currencies will be made. It is possible, therefore, that home-currency settlement payments could be delayed or disrupted because of settlement problems in other linked currencies. In such circumstances, the ultimate resolution of the situation may depend more on the liquidity of money markets for other currencies and on the action of other central banks than on home-currency liquidity provision. More generally, because of these linkages, the presence of multi-currency DVP settlements may constrain the ability of each central bank to respond in a relatively independent manner to home-currency settlement problems based on local market considerations.

2.18 Multi-currency DVP settlements, which would be specifically created or supported by an extension of payments system operating hours, cross-border payments system links or multi-currency central bank services, might have undesirable implications for the implementation of monetary policy. For instance, depending on the specific time chosen for a multi-currency settlement, daily fluctuations in the size of settlement obligations might influence the volatility of the aggregate demand for central bank balances as used for monetary policy purposes in some countries. The magnitude and duration of this impact on the monetary forecasting capabilities of the affected central banks of issue would depend on the size, distribution, and predictability of settlement obligations among the banks that participate in such multi-currency settlements. Although this greater uncertainty could potentially reduce the ability of the affected central banks of issue to influence targeted domestic interest rates, in practice it might not be difficult for central banks to deal with the situation by making modest adjustments to their current monetary policy operating procedures. In addition, any central bank lending to support multi-currency DVP settlements also has potential monetary policy implications. Depending on the timing of such lending in relation to the timing of domestic reserve monitoring and open market operations, it might be difficult for some, although by no means all, central banks of issue to offset quickly the impact of settlement-related lending on domestic interest rates and exchange rates.

Implications

2.19 Central banks face an array of service options that could increase the level of support for international settlements. Some of these services - such as intraday finality, settlement accounts and other home-currency services - are either currently available or could be offered quickly by at least some central banks, while others - such as those that would create or support the development of multi-currency DVP mechanisms - might take time to develop. Indeed, the development of sound multi-currency DVP settlements would require certain preconditions: design of the organisational and operational mechanisms that could facilitate multi-currency DVP settlements; solution of the technical and legal problems concerning the interlinkage of domestic systems that could accompany a major extension of operating hours or the creation of a common agent; meeting the needs of central banks regarding coordination; and the development of off-hour money markets that could adequately support multi-currency settlements. From a practical point of view, home-currency final transfers and settlement accounts likely could be developed - and would need to be offered - before the establishment of multi-currency DVP capabilities. Accordingly, central bank services are likely to develop over time.

2.20 Many central banks are currently considering important changes to their home-currency payments systems. Where not already in place, the development of an intraday final transfer capability could, in addition to any domestic goals, also contribute substantially to the reduction of international settlement risks and would provide a necessary building-block for other potential central bank services described in this report. Furthermore, some home-currency payment netting systems are considering ways to strengthen their current risk control mechanisms in a manner consistent with the Lamfalussy minimum standards for netting schemes. Such improvements could also increase the level of support for settling multi-currency transactions.

2.21 Some central banks are also considering moderate extensions of their local payments system operating hours to support domestic markets and settlements. At the same time, many pairs of payments systems already have overlapping hours. Thus, notwithstanding other potential obstacles to achieving multi-currency DVP (e.g. local payment arrangements, cut-off times for international correspondent payment orders and the timing of finality), expanding the operating hours of home-currency payments systems that have intraday finality increases the opportunities for the private sector to eliminate or reduce the duration of settlement exposures by more closely aligning the timing of payments between two or more banks in two or more currencies. Expanded operating hours would also increase the scope for two or more central banks, should they choose to do so, to provide directly DVP services for their respective currencies of issue by linking their home-currency payments systems, if home-currency money markets develop into a reliable source of funds for settlement requirements. It is likely that different responses may occur in different countries depending, for example, on whether operating hours are increased at the beginning or the end of the current business day. In any event, extensions of local payments system hours may provide useful information about the desire of the private sector to achieve multi-currency DVP settlements and about the likely development of off-hour money markets.

2.22 The Working Group's analysis also identified some implications of central bank services for multilateral multi-currency netting schemes, should private sector proposals for such arrangements become more fully developed. At the outset, the Working Group noted that the Lamfalussy Report concluded that soundly constructed and managed multilateral netting schemes may usefully contribute to the control and reduction of Herstatt risk in the foreign exchange markets. Soundly constructed netting schemes would be expected to manage prudently the settlement risks for all currencies - including non-G-10 currencies - that are included in such schemes. As to central bank settlement, the establishment of settlement accounts for netting schemes, with the possibility of final intraday settlements, might be considered. Such developments, were they to occur, would not be expected to preclude other changes in central bank services identified by the Working Group that might help reduce risk and increase efficiency in foreign exchange settlements, both for netted and non-netted transactions. Indeed, over time, a number of changes might be made to central bank payment services in conjunction with, or perhaps as a catalyst for, market developments.

2.23 Section 3 analyses existing central bank payment and settlement services in the G-10 countries. Following a review of important concepts and a discussion of current institutional arrangements, Section 3 then describes multi-currency settlement risks and key features of home-currency payments systems that are essential to reducing these risks. Section 4 describes potential central bank payment and settlement services that could help reduce risk and increase efficiency in cross-border and multi-currency settlements. Section 5 presents the Working Group's analysis of the advantages and disadvantages of these service options in the context of seven broad policy criteria.