5.2 The analysis of credit and liquidity risks in securities clearance and settlement that was developed in Section 2 built upon the analysis of risks in netting and settlement systems that was developed in the Angell Report and the Lamfalussy Report. The Lamfalussy Report also contained a framework for evaluating the implications for central bank policy objectives of cross-border and multi-currency netting and settlement schemes for payment orders and foreign exchange contracts. This framework took the form of a set of minimum standards for such schemes (see Annex 4), along with a lengthy set of supporting explanations and analysis. Together the standards and the supporting analysis identify a set of critical issues Shat designers and operators of such schemes must address.

5.3 The decisions by central banks to set strict standards and to establish principles for cooperative central bank oversight to ensure compliance with the standards reflected a concern that multilateral netting systems for foreign exchange would create a concentration of risks that had the potential to spread systemic pressures throughout international markets quite rapidly. In particular, the absence of a DVP mechanism for settlement of foreign exchange contracts implies that the central counterparty in a multilateral foreign exchange netting system would have substantial credit exposures to participants. And, by the very nature of foreign exchange contracts, a credit or liquidity problem in one national market has the potential to spread quickly to other national markets.

5.4 Clearly, some of these key concerns about multilateral netting mechanisms for foreign exchange do not arise in all cases for securities settlement systems. Nonetheless, rather than building a new framework for the evaluation of securities settlement systems from the ground up, the Study Group saw a number of advantages in exploring whether the issues involved are sufficiently similar to those identified in the Lamfalussy Report that the Lamfalussy analysis can be applied to such systems. The remaining sub-sections of this section evaluate the relevance of the issues highlighted by each of the Lamfalussy standards point by point. At the beginning of each sub-section some of the key issues that arise when a securities settlement system is analysed in these terms are highlighted, including some issues not explicitly addressed in the Lamfalussy Report.

Legal Basis for the System

• Effectiveness of funds transfers, including netting of funds transfers.
• Effectiveness of securities transfers, including netting of securities transfers.
• Effectiveness of liens on securities.

5.5 Any netting or settlement system, whether for foreign exchange, securities or other financial interests, should have a well-founded legal basis. If a securities settlement system involves securities issues, markets, participants, custodians or settlement banks subject to the law of more than one jurisdiction, the legal enforceability of the system's rules and procedures in all relevant jurisdictions should be established. In terms of the specific analysis of the implications of this standard, the Lamfalussy Report focused on one legal issue: the effectiveness of the netting of sums of money owed in the event of bankruptcy of a system participant. In those securities settlement systems that provide for the netting of funds transfer instructions and rely on the effectiveness of the netting in designing their risk management systems (model 2 systems and some model 3 systems), this issue is clearly a critical one. Should the system or its participants come to rely on exposures produced by netting when the netting is not legally enforceable, they may face credit and liquidity exposures far larger than they realise (and can cope with) in the event of a participant default. In addressing this issue it is often important not only to assess whether any netting is effective, but also to determine at what stage in the processing cycle the netting takes effect. As noted earlier, in some model 3 systems net funds debit and credit balances are provisional until all net debit positions have been covered; if a net debit position is not covered, an unwind of transfers back to the original gross obligations is possible.

5.6 The legal issues encountered in the design and operation of a securities settlement system go well beyond those specifically analysed in the Lamfalussy Report, however. Indeed, the basic goal of DYP - establishing the strongest possible linkage between delivery and payment - requires a careful determination of whether (and when) a securities transfer is effective. Although not all of the additional legal issues necessarily give rise to systemic concerns, the Study Group has not attempted a thorough analysis of the issues and their implications. Rather, in this report the Study Group has merely attempted to identify some of the relevant legal issues. One group of issues that is clearly relevant relates to the effectiveness of transfers of securities in the settlement system. For example, does the transfer of a security on the books of the settlement system constitute a legal transfer in all relevant jurisdictions? If securities transfer instructions are netted, is the netting legally binding? Another significant set of legal issues arises when a securities settlement system looks to securities as collateral for credit extensions to participants. For example, is the lien effective in all relevant jurisdictions? Is it permissible for the participant to transfer rights to customer securities to the system? Finally, if the system provides a securities lending facility, are the terms of the lending agreement, including any lien on collateral, enforceable?

Understanding of Risks

• Division of responsibility for risk management between system operator and participants.
• Implications of loss-sharing agreements.
• Implications of unwinds.

5.7 If the financial risks that arise in securities clearance and settlement are to be adequately controlled, both the system operator and the participants must clearly understand those risks. Securities settlement systems are inherently multilateral arrangements. As noted in the Lamfalussy Report, multilateral arrangements necessarily involve a multiplicity of relationships that make possible the shifting of risks both among participants and between the system participants and the system operator. The most basic issue that multilateral arrangements, including securities settlement systems, must address is the division of risk management responsibility between the system operator and the participants. The Lamfalussy Report concluded that more than one approach to the division of responsibilities is feasible. What is critical is that both the system operator and the participants understand clearly where the responsibility rests.

5.8 The discussion of alternative models of securities settlement systems in the previous sections revealed that several different approaches to allocating credit and liquidity risks are in fact currently utilised. In some systems, completion of settlement is guaranteed, either by the system operator or by one or more guarantor banks. In such systems the risks and the responsibility for risk management clearly rest with the guarantor or guarantors. From the perspective of the participants, the primary source of risk is the failure of a guarantor. As noted above, to protect themselves and, where necessary, to make their guaranty credible guarantors have strong incentives to impose a variety of risk controls on participants. In some cases, these include loss-sharing arrangements that charge back losses to participants, either on the basis of their overall activity in the system or (less frequently) on the basis of their dealings with a defaulting participant. When losses are based on overall activity, participants may have difficulty gauging the magnitude of their contingent obligations, although the obligations are usually limited (to their contributions to a clearing fund or a multiple thereof).

5.9 As noted earlier, in some model 3 systems settlement is not guaranteed either by the system operator or by a third party. Rather, if a participant fails to cover a net debit position, some or all securities transfers are unwound (deleted) and the settlement obligations of the other participants are recalculated. In such a system the responsibility for risk management rests with the participants, who would bear the liquidity pressures (as well as any losses) arising from the failure of a counterparty to complete settlement. Participants may understand this clearly, and, therefore, have incentives to manage risks prudently. However, several factors may tend to blur their understanding and diminish their incentives. First, participants may assume that the system operator (or a central bank) will provide the necessary credit, directly or indirectly, to avoid a settlement failure. Secondly, incentives to manage risks associated with customer transfers may be limited if participants are able to pass on the risks to customers. Concerns about the consequences of unwinds may, nonetheless, be substantially limited if the system imposes risk controls that limit the likelihood and potential consequences of a settlement failure.

5.10 Regardless of the approach taken to risk management, both the system operator and the participants face serious difficulties m ascertaining their credit exposures if the system fails to establish a firm linkage between delivery and payment. Without such a linkage, the potential losses from the failure of a participant could clearly be quite large and cannot be precisely quantified, although an upper limit may be ascertainable. Even if the system guarantees settlement, the ability of the guarantors to absorb potential losses from a participant default would be difficult to gauge in such circumstances. Furthermore, as noted earlier, without a firm linkage, in times of stress many participants are likely to balk at completing deliveries or payments for fear of large credit losses. The potential liquidity demands that might emerge in that event would also be difficult to anticipate.

Incentives and Capabilities to Manage Risks

• Need to limit credit exposures produced by participants.
• Reliance on membership standards and collateral, rather than caps.
• Trade-offs between risk and efficiency associated with collateral requirements.

5.11 The need for securities settlement systems to have clearly defined procedures that provide appropriate incentives for risk management has already been discussed in the previous sub-section. System operators and participants also need to consider whether relevant parties (the system operator and/or the participants) have the capabilities to manage and contain the risks they bear. Lamfalussy standard III (and standard IV) provides specific criteria by which to judge whether a foreign exchange netting system adequately addresses this issue. With respect to credit exposures, this standard states that -limits- should be placed on the maximum level of credit exposure that can be produced by each participant. The analysis of securities settlement systems in the previous sections noted the importance of establishing safeguards to limit the potential losses from a participant's default. A more difficult issue is the circumstances under which various specific safeguards are necessary. As noted earlier, securities settlement systems tend to rely on membership standards and on collateral to limit credit exposures. While Lamfalussy standard III implies a need for a cap or quantitative limit on funds transfer activity, securities settlement systems seldom employ such a safeguard.

5.12 A key question, then, is whether membership standards and collateral procedures alone can adequately limit credit exposures in a securities settlement system. In assessing this question, it is important to remember that most securities settlement systems either achieve DVP (models 1 and 3) or introduce assured payment systems (model 2) intended to make principal risks negligible. Thus, in well-designed systems, counterparty credit exposures to participants are effectively limited to replacement cost exposures. By contrast, in settlement systems for foreign exchange contracts, the absence of a DVP mechanism creates cross-currency settlement exposures (Herstatt exposures) as well as replacement cost exposures. If only replacement cost exposures are involved, collateralisation may by itself adequately contain participant credit exposures. Whether this is so would depend on the quality of the collateral and how tightly the collateral controls are administered. If the collateral is high-quality, is conservatively valued (high -haircuts- are applied), and is revalued frequently, counterparty credit exposures can be reduced to minimal levels. On the other hand, if defaults by issuers of the securities are a significant possibility or if mechanisms are not imposed to ensure that sufficient high-quality collateral is available, significant counterparty credit exposures may still remain.

5.13 Furthermore, the decision whether to impose even a collateral requirement confronts designers of securities settlement systems with a short-run trade-off between risk and efficiency (as measured by the proportion of scheduled transfers that settle) and a longer-run trade-off between risk and the cost of maintaining the collateral necessary to allow a higher proportion of transfers to be completed. As noted earlier, relatively few systems monitor collateral values ex ante and make the completion of a securities transfer conditional on the availability of adequate collateral. Their reluctance to do so has in some cases reflected a concern that if sufficient collateral were not posted unacceptably high fail rates could result, with a possible adverse impact on market liquidity. In addition, if transfers blocked by the collateral controls were initiated by the central bank, implementation of central bank monetary operations could be frustrated. Such concerns in some systems may argue decisively against collateralisation, especially if settlement volumes are sufficiently low that replacement cost exposures are unlikely by themselves to create systemic problems. On the other hand, if replacement cost exposures are sufficiently large to be a potential source of systemic problems or if collateral positions can be readily adjusted, a collateral requirement would appear to deserve serious consideration.

Arrangements for Ensuring Timely Completion of Settlements

• Containment of systemic liquidity pressures.
• Assessment of potential pressures from a participant's failure to settle.
• Assessment of liquidity resources available to the system operator and participants.

5.l4 System operators and participants should also consider whether the potential liquidity pressures from the failure of a participant in a securities settlement system to cover a net funds debit position (or, equivalently, to cover a funds overdraft) can be adequately contained. As noted earlier, the achievement of DVP does not by itself ensure that the operation of a securities settlement system adequately limits systemic risk. In particular, even if principal risks are eliminated and replacement cost risks are contained (through collateralisation or otherwise), a failure to contain liquidity pressures may create systemic problems. Thus, from a central bank perspective, liquidity risk is an important concern, and it is important to assess the ability of a securities settlement system to contain it in the event of a participant's failure to settle.

5.15 Analysis of this issue requires assessments both of the potential liquidity shortfall from a participant's failure to settle and of the extent of liquidity resources available to the system operator or to participants. If the system imposes limits on participants' net funds debit positions (or overdrafts), potential liquidity shortfalls can readily be gauged. However, as noted earlier, relatively few securities settlement systems impose caps. In the absence of caps, an assessment of potential liquidity shortfalls can be based on historical data on net debit positions. However, in using historical data for this purpose, it is important to recognise that during market disturbances (like the 1987 stock market crashes) activity levels (and, therefore, net debit positions) may substantially exceed historical averages. Thus, an assessment of potential liquidity pressures in the absence of caps would generally require a simulation of the potential effects of peak levels of activity.

5.16 Once the potential liquidity shortfalls have been assessed, the next step is to assess whether adequate liquidity resources are available to contain the resultant liquidity pressures. In those systems in which completion of settlement is guaranteed, it is the liquidity resources of the guarantor that are relevant. With the exception of systems in which completion of settlement is guaranteed by a central bank, those liquidity resources are limited. In such systems, collateral requirements ensure the availability of a pool of assets to support liquidity needs, but without prearranged facilities the collateral may be difficult or impossible to sell or pledge in time to use the proceeds to complete timely settlement. Thus, access by the guarantor to central bank or private sector liquidity facilities would generally be necessary. In those model 3 systems that do not guarantee the completion of settlement, it is the liquidity resources of the participants that are relevant. Here again, the return of securities that had been provisionally transferred to the participant that failed to settle provides a potential source of liquidity, but selling or pledging those securities may not be possible in time to use the proceeds to meet other obligations. Thus, the participants' access to money markets and to central bank or private credit facilities is a critical element in the assessment of their ability to contain the resultant liquidity pressures. Because an unwind has the potential to create short positions in securities, the liquidity of securities lending markets must also be assessed.

5.17 Without further information about actual and potential levels of activity and the availability of liquidity resources, it is difficult to generalise about the capacity of securities settlement systems to contain potential systemic liquidity pressures or about the circumstances in which specific safeguards (for example, caps) are necessary. The Lamfalussy Report can be taken to imply that cross-border and multi-currency netting schemes for payment orders and foreign exchange should impose caps to limit such pressures. But underlying this conclusion was a presumption that such schemes were likely to reach a scale sufficiently large that the failure of a single participant to settle could by itself create a liquidity shortfall so large that it could not be contained by the system operator or by the participants and could, therefore, create systemic disturbances. In the case of securities settlement systems, however, such a presumption is not always warranted.

5.18 Whether a securities settlement system is a potential source of systemic disturbances depends on several factors. As noted above, the liquidity of money and securities markets is a critical consideration. This, in turn, depends importantly on the time of day that a settlement problem would become apparent. If it occurs at a time of day when markets are active, the loans of securities and funds needed to contain a disturbance may be readily available. Conversely, systemic disturbances may be inevitable if the settlement problem arises late in the day when the money and securities markets tend to be illiquid. Finally, the scale of operation of some securities settlement systems may suggest that they are unlikely to be the source of systemic disturbances, although an unusual concentration of settlement activity among a small number of participants could create systemic concerns even in a system whose overall scale of operation appears modest. In the absence of a strong presumption that systemic disturbances are possible, the arguments for caps should be considered on a case-by-case basis, bearing in mind the discussion above of the fundamental trade-off between efficiency and risk in securities settlement systems.

5.19 Lastly, it should be noted that while a securities settlement system should ensure the timely completion of securities and funds transfers associated with all transfer instructions accepted for execution by the system, it cannot ensure that all such instructions will be accepted. Rather, the acceptance of transfer instructions is always conditional on the availability of the securities in the seller's account; if the securities are not available, the trade will fail to settle. This reflects the fact that markets for securities and for lending individual securities tend to be much less liquid than money markets. The arrangement of automatic securities lending facilities can facilitate the timely completion of securities transfers, but no system can completely eliminate settlement failures resulting from the seller's inability to buy or borrow the securities needed. Because new securities cannot usually be created in order to complete settlement, there is no lender of last resort for securities. ³⁶

Admission Criteria

• Financial and managerial capacity of participants.
• Need to avoid use of membership standards intended solely to create competitive advantages for participants.

5.20 Securities settlement systems typically establish criteria for admission that address the financial and managerial capacity of a participant to satisfy its obligations and to manage the credit and liquidity risks that it bears. In some cases further criteria not directly related to financial or managerial capacity are also applied. For example, some systems operated by central banks limit participation to banks. Some might argue that such additional criteria are inconsistent with -fair and open access-. However, the Lamfalussy Report recognised that membership criteria can appropriately distinguish between financial institutions subject to effective supervision and those that are not and may also appropriately distinguish between institutions that have access to central bank credit facilities and those that do not. What the Lamfalussy Report indicated would be inappropriate is a membership criterion intended solely or primarily to create an arbitrary competitive advantage for participants. Although the Study Group has not examined membership criteria for securities settlement systems in detail, they generally appear to focus on factors relevant to the financial and operational capacity of participants.

Operational Reliability

5.21 The breakdown of a critical operational component of a securities settlement system can obviously create serious liquidity problems and, by delaying settlements, can increase replacement cost and credit exposures. Consequently, such systems should ensure that all hardware, software and communications facilities that support their operations have a high degree of reliability and integrity. Contingency plans should be established for the potential failure of each critical component, including the identification of backup facilities capable of completing the settlement process on the settlement day and performing the accounting and processing work necessary to prepare for the next settlement day. Because the Study Group has not focused on operational safeguards, this report cannot offer a firm opinion as to how securities settlement systems typically measure up to this requirement.

Conclusions

5.22 The discussion above indicates that issues addressed by the Lamfalussy Report are also relevant to securities settlement systems and that the standards are an effective means of highlighting issues concerning the design and operation of such systems that are relevant to central bank policy objectives regarding financial stability and the containment of systemic risk. However, in using the Lamfalussy Report as a framework for the evaluation of securities settlement systems, several limitations should be borne in mind. First, certain issues that are important in securities settlement systems are not directly addressed by the Lamfalussy Report. For example, the discussion of legal enforceability revealed a number of such issues. Secondly, the Study Group's analysis suggests that the potential for systemic risk in securities settlement systems may not in some cases be as great as in the multi-currency and crossborder netting schemes addressed by the Lamfalussy Report, which were viewed as having the potential to reach a very large scale and to link interbank money markets in several countries in ways that could transmit systemic liquidity pressures across national borders very quickly. Consequently, certain safeguards that the Lamfalussy Report concluded were essential for netting schemes may not in some cases be necessary for securities settlement systems.

5.23 These considerations argue against characterising the Lamfalussy standards as minimum standards that securities settlement systems must in all cases meet to address central bank concerns about systemic risks. Nor should a system that meets the standards necessarily be assumed to adequately address those concerns. Nonetheless, the standards and accompanying analysis provide a valuable analytical framework that, when applied to securities settlement systems, serves to heighten awareness of central bank concerns about their design and operation. Failure of a particular national system to meet one or more standards should prompt further analysis of whether the system is a potential source of systemic disturbances. If such further analysis suggests that the system is a potential source of systemic disturbances, the central bank would presumably seek enhancements to the system, working in concert with securities supervisors or other responsible authorities where necessary. Such enhancements could be implemented over time, allowing the costs of enhancements to be minimized by incorporating them into the broader overhauls that such systems must undertake periodically.

Footnote:

36. In at least one case, however, a settlement system for government securities provides for the creation and temporary lending of securities to facilitate settlements.