1.1 Background

In September 1992 the Bank for International Settlements published a report entitled Delivery Versus Payment in Securities Settlement Systems, which had been prepared by the Committee on Payment and Settlement Systems of the central banks of the Group of Ten countries. That report (referred to here as the DVP Report) defined and analysed the types and sources of risk in securities settlements and clarified the meaning and implications of delivery versus payment (DVP). Building on that framework, it identified common approaches to DVP in the G-10 countries and evaluated the implications of the various approaches for central bank policy objectives concerning the stability of payment systems and financial markets and the containment of systemic risk.

In the course of preparing the DVP Report, some preliminary analysis of the risks associated with settlements of cross-border securities trades was conducted. This analysis suggested that issues arise in a cross-border context that were not addressed adequately in the DVP Report. In particular, the DVP Report focused on the settlement of trades between two direct participants in a central securities depository (CSD). However, in a cross-border trade one counterparty or both typically settle through one or more intermediaries - a local custodian bank, a global custodian or one of the international CSDs (ICSDs - Cedel and Euroclear). The involvement of such intermediaries complicates the analysis of risks in cross-border securities settlements and, if the intermediaries not only hold securities but also settle trades on their own books under rules and procedures that differ from those established by the local CSD, the risks can be fundamentally different from those faced by direct participants in the local CSD.

At a meeting in December 1992 the Committee set up a new study group and asked it to examine the implications for central bank policy objectives of the expansion of cross-border trading and the concentration of securities settlement activity in certain intermediaries. Specific objectives of the study included: (a) the development of a clearer understanding of the alternative channels through which settlements of cross-border trades can be effected; (b) an analysis of the risks associated with the use of the various channels; and (c) an assessment of whether cross-border settlement arrangements have important implications for central bank policy objectives relating to the containment of systemic risk and the oversight of payment and settlement systems.

The remainder of this section of the report summarises the results of the study group's work. Section 2 addresses certain preliminary issues, including the definition of key terms (cross-border trade, cross-border settlement) and the availability of data on the volumes of cross-border trades and cross-border settlements. Section 3 describes the alternative channels through which cross-border trades can be settled and discusses the utilisation of the various channels by different types of traders. Section 4 analyses the risks associated with the use of each of the major channels. Section 5 considers the implications of cross-border settlement arrangements for central bank policy objectives. A glossary is provided at Annex 1. Annex 2 describes the growing markets for repurchase agreements (repos) and securities loans and discusses the implications for securities settlements. Annex 3 reviews the legal issues that arise in cross-border settlements. Annexes 4 and 5 examine the "bridge" linking the Cedel and Euroclear systems and the link between Euroclear and the German CSD respectively. Annex 6 contains a bibliography.

1.2 Preliminary issues: definitions and data

In this study, a cross-border trade is defined as a trade between counterparties located in different countries. In most cases one counterparty is located in the country in which the security is issued, while the other counterparty is located in another country. A cross-border settlement is defined as a securities settlement that takes place in a country other than the country in which one trade counterparty or both are located. Again, in most cases the settlement takes place in the country of issue of the securities, but one counterparty or both are located outside the country of issue. It is important to recognise, however, that there are frequent exceptions to these typical patterns. For example, various European government securities are traded very actively by securities dealers in London, and large volumes of trades in such securities are settled in Belgium (through Euroclear) and in Luxembourg (through Cedel).

Comprehensive data on cross-border trading or cross-border settlements simply do not exist. Where data on cross-border trading of bonds and equities are available, they indicate that the growth of such transactions has far outpaced the growth of economic activity in recent years. The best data available on cross-border settlements are those compiled by the ICSDs. However, they are far from comprehensive; for example, the ICSDs settle very few trades in equities and currently do not settle any trades in some government bonds that are heavily traded internationally. Nevertheless, these data document enormous growth in cross-border settlements of trades in certain European government securities, notably German, Dutch, French, Danish and Spanish issues. This growth has reportedly been spurred by the increasing use of sophisticated trading and financing strategies in European fixed-income markets. In particular, the ICSDs indicate that a large and growing proportion of such settlements is accounted for by the use of repos for financing and hedging positions in European government securities and related derivative instruments. The rollover of short-term repos financing longer-term positions and the dynamic adjustment of hedges have the potential to push settlement volumes even higher as such techniques come to be used more widely.

1.3 Alternative channels for effecting settlements

The study group's investigation of the channels through which settlements of cross-border trades are effected confirms that non-residents seldom participate directly in local CSDs, although local branches or subsidiaries of non-resident firms sometimes do. Instead, traditionally non-residents have normally settled their trades through a local agent (a custodian that is a direct participant in the local CSD). Local agents continue to play an important role in cross-border settlements, but today additional intermediaries are frequently involved. Many institutional investors now utilise the services of a global custodian that settles trades and safekeeps securities in many countries through a network of sub-custodians (local agents, including its own local branches). Internationally active securities dealers often settle trades through the ICSDs, which in recent years have established direct or indirect (through local agents) links to CSDs in many countries. Use of these alternative channels is not mutually exclusive; some of the most active dealers utilise both an ICSD and a local agent for settling trades in certain securities. Finally, numerous CSD-to-CSD links have been established in recent years. However, with very few exceptions, these links have not been heavily used to effect cross-border settlements.

1.4 Risks in cross-border settlements

The analysis of risks in domestic settlements that was presented in the DVP Report served as the starting-point for the study group's analysis of risks in cross-border settlements. The DVP Report identified and analysed the risks involved in domestic settlements, including principal risk, replacement cost risk and liquidity risk. The Report concluded that the largest single source of risk in securities settlements is principal risk, that is, the risk that a seller of a security could deliver but not receive payment or that a buyer could make payment but not receive delivery. A DVP mechanism was defined as a link between a securities transfer (delivery) system and a funds transfer (payment) system that eliminates principal risk. The Report identified several models for achieving DVP. However, it cautioned that no securities settlement system can eliminate replacement cost risk - the risk of the loss of an unrealised gain on an unsettled contract because of the counterparty's default. Nor can it eliminate liquidity risk - the risk that the counterparty will not settle its obligation when due, but on some unspecified date thereafter. The degree of replacement cost risk or liquidity risk in a system depends critically on the risk controls employed by the CSD. If a CSD does not impose adequate controls, replacement cost losses or liquidity pressure arising from a default by a CSD participant could cause systemic problems.

As far as they go, the analysis and conclusions in the DVP Report are equally applicable to cross-border settlements. In fact, the only inherent differences between risks in cross-border settlements and in domestic settlements are differences in legal risks and the potential for foreign exchange settlement risks to arise in a cross-border context. By definition, cross-border settlements involve multiple legal jurisdictions, thereby raising issues relating to choice of law and conflict of laws that complicate the analysis of legal risks and can introduce new sources of risk. In addition, if money settlements in foreign currencies are funded through foreign exchange transactions, non-resident counterparties face foreign exchange settlement risks.

However, as was acknowledged in the DVP Report, the analysis it presented was limited in certain respects, and those limitations generally are far more significant in a cross-border context than in a domestic context. By far the most important limitation is that the DVP Report focused heavily on the design and operation of an individual CSD and the implications for risks to its direct participants. The Report paid relatively little attention to the effects on settlement risks of the involvement of other intermediaries. To be sure, other intermediaries, especially custodian banks, often play a role in domestic settlements, but, as already noted, in cross-border settlements the involvement of other intermediaries is pervasive.

This involvement increases the importance of several issues that were given scant attention in the DVP Report. The most basic issue is that when a non-resident (or any other party) holds its securities through an intermediary, it is exposed to custody risk, that is, the potential loss of the securities in the event that the intermediary becomes insolvent, acts negligently or commits fraud. The most important factors determining the degree of custody risk are the accounting practices and safekeeping procedures used by the intermediary. The key to protecting the non-resident's interest is often the separation (segregation) of its assets from those of the intermediary and any other parties for which it holds securities.

Another issue that increases in importance when intermediaries other than CSDs are involved is the settlement of so-called back-to-back trades and the opportunity costs and liquidity risks that arise if such trades cannot be settled efficiently. A back-to-back trade is a pair of transactions that requires a counterparty to receive and redeliver the same securities on the same day. Securities dealers frequently need to settle such back-to-back trades. In providing liquidity to markets, dealers often buy and sell the same security for the same settlement date. In addition, they often rely on repos to finance their securities inventories and reverse repos and securities loans to meet delivery obligations, including those created by short sales. As will be discussed in greater detail below, the settlement of back-to-back trades by dealers (resident or non-resident) that are not direct participants in the local CSD poses difficulties in some settlement systems. In those systems dealers are often forced to pre-position securities or borrow securities to meet delivery obligations. These requirements can add significantly to intermediation costs and, therefore, may significantly reduce secondary market liquidity. Securities that cannot be delivered out cannot be used to obtain secured financing. As a result, dealers incur higher financing costs and may also be exposed to greater liquidity risks, because unsecured financing tends to be less reliable, especially when financial markets are under stress.

A third issue that is quite important in a cross-border context but that was not addressed in the DVP Report is the risks associated with cross-system settlements, that is, settlements effected through links between securities transfer systems. These include the direct and indirect links that have been established between CSDs and ICSDs. Such cross-system settlements often involve significant inefficiencies that derive from the need for the transfer systems to exchange information on whether the two counterparties have the securities and funds (or access to credit) necessary to complete settlement. In particular, the settlement of back-to-back trades in which one or both settlements are cross-system settlements is often not possible, so that dealers are obliged to pre-position or borrow securities to complete such settlements.

Special problems can arise in cross-system settlements when one or both transfer systems are what the DVP Report termed model 3 DVP systems, that is, systems that make provisional transfers of securities that are not final until money settlement is completed later in the day (or on the following day in the case of systems that process instructions for settlement on S during the evening of S-1). In such systems, if a participant fails to cover its money settlement obligation, transfers involving that participant, including transfers of securities from that participant to participants in other settlement systems, may be unwound. At a minimum, the unwinding of such transfers would adversely affect counterparties of the defaulting participant. However, depending on how losses are borne or allocated by the system that received the provisional transfers, others among its participants that were not counterparties to the defaulting participant in the other system could also be adversely affected.

The relative importance of these issues - custody risk, back-to-back settlements and cross-system settlements - depends on the intermediary that a counterparty uses to hold its securities and settle its trades, on the services (especially credit services) that the intermediary provides to the counterparty, and on the counterparty's trading and financing strategies. In the case of a non-resident counterparty that settles its trades through one of the channels identified above, the risks associated with such cross-border settlements often differ significantly from the risks to a direct participant in the local CSD that were the focus of the DVP Report.

A local agent typically holds securities and settles trades for non-residents through an account it maintains for its customers at the local CSD. Trades involving non-residents that participate "indirectly" in the CSD through a local agent settle according to the same rules as any other trades settled by the CSD. Consequently, the settlement risks faced by a non-resident using a local agent are in many respects identical to those faced by a direct participant in the local CSD. For example, if the local CSD does not achieve DVP, the non-resident is subject to principal risk.

Nonetheless, in some other respects a non-resident settling through a local agent faces costs and risks that differ from those faced by direct participants. Quite clearly, the custody risk incurred by the non-resident is greater because it holds its securities indirectly through the local agent. In addition, a non-resident that uses a local agent may need to maintain larger balances of cash and securities than a direct participant would to settle the same set of transactions. The DVP Report emphasised that nearly all CSDs explicitly or implicitly extend substantial amounts of intraday credit to their direct participants to enable them to economise on holdings of cash balances and thereby to reduce opportunity costs and cash deposit risks (the credit risks associated with holding cash balances with an intermediary). Non-residents must look to the local agent for such intraday loans of funds. If such intraday credit is not available, a non-resident would need to hold larger cash balances than a direct participant, implying higher opportunity costs and greater cash deposit risk.

As noted earlier, the settlement of back-to-back trades by dealers that are not direct participants in the local CSD poses difficulties in some settlement systems. Specifically, in systems in which instructions to transfer securities are processed in a single batch cycle, the instruction to deliver out the security must be input prior to the processing cycle and, therefore, prior to receipt of the security. This often poses a dilemma for a local agent settling back-to-back trades through a single account for multiple customers (individual sub-accounts for customers are often not available). If the dealer fails to receive the security from its counterparty, its delivery instruction may, nonetheless, be completed using other securities in the local agent's account at the CSD.

In such systems, back-to-back trades by a non-resident dealer can be settled only if the local agent provides the dealer with what is, in effect, an intraday securities loan. If the local agent enters the delivery instruction and the securities are not received from the dealer's counterparty, the intraday loan becomes an overnight loan. In such circumstances, if the local agent does not own the securities itself, it would need to borrow them, either externally or internally (from another of its customers), in order to avoid a shortfall in its custodial holdings. Local agents can reduce the need for such borrowings in several ways, however. First, the likelihood of failed deliveries can be limited by prematching settlement instructions with the seller or its local agent prior to transmission to the CSD. Second, if a local agent can attract a critical mass of securities dealers as customers, it can settle back-to-back trades between those customers on its own books. Such a local agent is effectively operating as a securities transfer system, and the costs and risks associated with settlements on its books can differ substantially from those involved in settling directly through the local CSD.

Many institutional investors use global custodians rather than local agents to settle their cross-border trades. The risks associated with settlement through a global custodian are in many respects similar to those associated with settlement through a local agent. A global custodian settles the non-resident's trades in the local market through a local agent acting as its sub-custodian. Thus, in this case, too, the non-resident's trades would typically be settled in the local CSD, effectively subject to the local CSD's rules. As in the case of use of a local agent, the non-resident faces custody risk, and the further tiering of securities holdings may exacerbate custody risk and certainly makes such risk more difficult to assess. On the other hand, by providing a single standardised gateway to multiple markets, the use of a global custodian may reduce operational risks. In addition, global custodians typically provide other services to their customers that are designed to reduce uncertainty about money settlement obligations arising from failed deliveries of securities and delayed income payments. These services (contractual settlement date accounting and contractual income collection) usually involve provisional debits and credits to customers' cash accounts that can subsequently be reversed by the custodian if the anticipated securities or funds are not received within an interval that the custodian establishes. These services reduce liquidity risks and cash deposit risks on a day-to-day basis. However, if a customer is not given prior notice of reversals of provisional credits, the objective of increasing certainty about cash flows is compromised, and the customer could face substantial liquidity risks. In addition, if a customer misunderstands the provisional nature of these credits and debits, it could underestimate its credit exposures to counterparties and securities issuers.

Use of an ICSD can change fundamentally the costs and risks associated with securities settlement. The ICSDs settle the majority of their participants' trades on their own books. Such internal settlements are effected under the ICSD's own rules and operating procedures, which often differ significantly from the rules and procedures in the local markets to which they are linked. In terms of the taxonomy developed in the DVP Report, the ICSDs are best classified as model 1 DVP systems. Because DVP is achieved, internal settlements do not involve principal risk, even if the local CSD in the country in which the securities are issued does not achieve DVP. Replacement cost risks in internal settlements currently tend to be higher because the settlement interval for trades between ICSD participants is often longer than the settlement interval in local markets, implying larger credit exposures on unsettled contracts. However, with effect from 1st June 1995, such trades will be settled on the third business day after the trade date (T+3), the same settlement interval as in many local markets.

The costs associated with settling trades on the books of the ICSDs are often lower than the costs of settling through a local agent. On the securities side, same-day turnaround of internal receipts for internal delivery is always possible, whereas, as noted above, in some local markets dealers that are not direct participants in the CSD may be required to pre-position securities for delivery or to borrow securities, adding significantly to settlement costs. On the cash side, the fact that the ICSDs run their processing cycles during the night and report cash balances early in the European business day facilitates participants' efforts to minimise holdings of cash balances.

Like local agents, the ICSDs reduce the participants' opportunity costs by extending intraday loans of funds. Because of differences in operating hours between the ICSDs and the various national payment systems, the exposures incurred by the ICSDs as a result of these credit extensions are of longer duration than the credit exposures that CSDs and local agents typically incur. These exposures are typically strictly limited and are collateralised by participants' securities holdings. Nonetheless, the exposures are substantial, and the choice of law and conflict of laws issues that arise in cross-border settlements can create ambiguities about the effectiveness of the liens involved.

ICSD participants enter into trades and financing transactions not only with other ICSD participants but also with counterparties that settle their trades in the various local markets. Settlement of these trades with local market participants is effected using the links that the ICSDs have developed to the local markets. The costs and risks involved in such settlements are heavily influenced by practices in the local markets. For example, principal risk exists if the local CSD does not achieve DVP. In fact, the costs and risks arising in such cross-system settlements often exceed the costs and risks associated with domestic settlements in the local markets because of inefficiencies in the links between the ICSDs and the local markets. As noted earlier, the settlement of back-to-back trades across such links is often not possible. Like local agents, the ICSDs provide what are, in effect, intraday securities loans to allow such trades to settle without imposing substantial opportunity costs on their participants.

In the case of the link between the two ICSDs (the "bridge"), the inefficiencies associated with cross-system settlements (between Cedel participants and Euroclear participants) and the consequent costs and risks have largely been eliminated by the introduction in 1993 of multiple daily processing (and settlement) cycles and exchanges of information on completed settlements. To strengthen their local market links, which have become increasingly important because repo trades are quite often settled in the local market, the ICSDs have supported the introduction by local CSDs of multiple processing cycles - at a minimum, one cycle before the night-time processing cycles at the ICSDs and one cycle after. This allows ICSD participants to complete same-day turnarounds of securities, regardless of from whom they are received (the local market or an ICSD participant) or to whom they are delivered. However, as noted earlier, a serious weakness in some of these links to local markets is that the transfers from the local market that result from the evening processing cycle are not final until money settlement occurs the next day. The potential for unwinds of provisional transfers creates significant credit and liquidity interdependencies between the systems - disruptions from a settlement failure in the local market would promptly be transmitted to the ICSDs and their participants.

Even if CSD-to-CSD links are not vulnerable to unwinds of provisional transfers, such links create significant operational interdependencies between and among CSDs (and ICSDs) and can also create credit and liquidity interdependencies. An operational problem at one CSD would result in a failure to complete deliveries between their participants, which could affect the completion of deliveries at other CSDs, including CSDs not directly linked to the CSD with the operational problem. Credit and liquidity interdependencies are created when one CSD provides another CSD with a cash account and settles trades between its own participants and participants in the other CSD by debiting or crediting the other CSD's cash account. The CSD using the cash account is exposed to cash deposit risk and liquidity risk, while the CSD providing the account is exposed to credit and liquidity risks if (as often is the case) it permits overdrafts or debit balances.

1.5 Implications for central bank policy objectives

Central banks have an interest in the design and operation of securities settlement systems because of their implications for central bank policy objectives relating to financial stability and the containment of systemic risk and to the effectiveness of central bank oversight of payment and settlement systems. Central banks have broad responsibilities for the stability of the financial system as a whole. In particular, as lenders of last resort, they are usually at the centre of efforts to contain threats to financial stability. These responsibilities require central banks to identify sources of systemic risk and to consider how such risk can be diminished. The failure of a large trader or settlement intermediary to meet its obligations could produce liquidity pressures or credit losses on a scale sufficient to threaten the stability of the financial system as a whole. In particular, a disturbance in a securities settlement system could spill over to money markets or to payment systems. Indeed, the potential for such spillovers is likely to have increased in recent years because of the increased importance of repos as money market instruments, the increased use of securities collateral to control risks in payment systems, and the rapid growth of securities settlement volumes.

Central banks are especially concerned about disturbances to payment and securities settlement systems and to money markets because such systems and markets are relied upon as vehicles for the execution and transmission of monetary policy. Because of these concerns, central banks oversee developments in their domestic money markets and payment systems. In the case of settlement systems for government securities, most of the G-10 central banks actually operate the home country CSD as part of their role as fiscal agent for the Government. Some of the G-10 central banks also play a role in the oversight of privately operated securities settlement systems, although in other cases such oversight is the responsibility of securities supervisors or is shared by central banks and securities supervisors.

Systemic risks in domestic securities settlements were analysed in the DVP Report. However, as already noted, the DVP Report focused heavily on the management by individual CSDs of their credit and liquidity exposures to their direct participants. In particular, the Report emphasised that nearly all CSDs extend substantial amounts of intraday or overnight credit to their participants and that the degree of systemic risk in a securities settlement system depends critically on the risk controls that a CSD imposes to limit potential losses and liquidity pressures in the event that participants fail to repay such credit extensions. The DVP Report paid relatively little attention to other intermediaries involved in the settlement process, including those that hold securities for counterparties that are not direct participants in the CSD and, in some cases, settle trades between such counterparties on their own books. In cases in which settlement activity and settlement risks tend to be concentrated in these other intermediaries, this is a significant shortcoming. A failure by such an intermediary to meet its obligations could be a significant source of systemic risk, even if the CSD that operates the core settlement system continues to meet its obligations to its direct participants.

The study group has concluded that the most important implication of the expansion of cross-border settlement activity for central bank policy objectives relating to the containment of systemic risk is that a central bank should be concerned not only about how the home country CSD manages its risks, but also about how the other intermediaries that play a central role in cross-border settlements of trades in home country securities manage their risks and about how the effects of financial or operational problems at such intermediaries would be contained. To a greater degree than is typically the case in domestic settlements, risks in cross-border settlements are concentrated in such intermediaries, especially in the ICSDs and in local agents that settle trades for international securities dealers. Furthermore, differences between the operating hours of the ICSDs and the operating hours and settlement practices (especially finality rules) of national payment systems and local CSDs require the ICSDs to make credit extensions to their participants that are of unusually large size and long duration, in order to reduce the opportunity costs of maintaining cash and securities balances to meet settlement obligations. Finally, the links that have been developed to effect cross-system settlements between the two ICSDs and between the ICSDs and local markets create significant operational interdependencies among settlement systems and can also create significant credit and liquidity interdependencies. As noted earlier, the possibility in some links that provisional transfers of securities between systems may subsequently be unwound is a particularly disturbing example of such interdependencies.

Should a disturbance to a cross-border settlement arrangement occur, central banks and other authorities could face special difficulties in containing it. In the first place, the authorities in the home country might not become aware of the problem promptly, especially if the disturbance originated with a counterparty or intermediary located outside the home country. Once a problem became apparent, its resolution would be likely to require coordination among several authorities in the home country and probably in other jurisdictions as well. Their efforts could be complicated by choice of law and conflict of laws problems that would create uncertainty about the finality of securities transfers, ownership rights or rights to collateral. Limited access by non-resident counterparties or intermediaries to liquidity in the home country currency could also complicate crisis management.

The critical role in cross-border settlement arrangements played by intermediaries other than the local CSD poses challenges to central bank oversight of domestic interbank markets and payment and settlement systems. The most basic challenge stems from the lack of transparency in cross-border settlement arrangements, which, as just noted, could handicap the central bank in responding to disturbances. When the critical intermediaries are located outside the home country, additional challenges arise. In such circumstances, the financial and operational problems of the non-resident intermediaries are a potential source of systemic disturbances to home country money markets and payment and settlement systems, but the central bank and other home country authorities may have only limited influence over those intermediaries' design and operation. Finally, as already noted, containment of the consequences of a serious financial or operational problem at a foreign intermediary would require a high degree of cooperation and coordination among central banks and other relevant authorities in multiple jurisdictions.