Risk Library
   Documents by Author
     Committees at the Bank for International...
       Clearing Arrangements for Exchange-Trade...
         6. Potential Weakness in Clearing Arrang...


Clearing Arrangements for Exchange-Traded Derivatives

6. Potential Weakness in Clearing Arrangements

The combination of risk management safeguards employed by exchange clearing houses has generally proved quite effective ­ in particular, there have been relatively few defaults by members of clearing houses in the G-10 countries, and none by settlement banks. Furthermore, in no case has the default of a clearing member significantly impaired the financial integrity of any of those clearing houses. Nonetheless, in several cases sharp movements in market prices have triggered events that raised doubts about the financial integrity of clearing houses, and in so doing, pointed to sources of vulnerability. For example, the collapse in US equity prices in 1987 triggered money settlements that were several orders of magnitude larger than clearing houses had previously experienced. Delays in completing these settlements triggered rumours about the financial health of the clearing organisations involved and may have exacerbated stock market volatility by widening the discounts at which some derivatives were trading relative to the cash markets. Another example is provided by the drop in Japanese equity values in early 1995, which triggered the collapse of Barings Plc and defaults by Barings entities that were direct participants in futures clearing arrangements in Japan and Singapore. Even in the Barings episode, the losses were apparently covered and obligations to other participants were met on schedule. Still, confidence in the financial integrity of the clearing houses was for a time shaken when unexpected delays were encountered in closing out or transferring positions and margin assets held by Barings and its clients.

More generally, analysis of the typical approaches to risk management suggests that clearing houses simply cannot be made fail-safe. With respect to the management of the risks of clearing member defaults, each of the safeguards typically relied upon by clearing houses has potential weaknesses and inherent limitations. To be sure, weaknesses in one safeguard may be compensated by strengths in other safeguards. For example, even if a clearing house's own resources are quite limited, there may not be cause for concern if it imposes very conservative margin requirements. Indeed, the failure of a single safeguard to perform as expected would generally not by itself imply credit losses or liquidity pressures that would be unmanageable. Nonetheless, certain events could simultaneously weaken several safeguards. One such event would be an extreme price movement that leads to an endogenous default of a clearing member, that is, a default that results from losses on house or client positions carried by the clearing member at the clearing house, rather than from losses from some other (exogenous) source. By hypothesis, such an event would produce credit losses that would not be covered by the defaulting participant's margins. It would also tend to strain money settlement arrangements, perhaps exposing hidden weaknesses. Another such event would be the failure of a clearing member that also is a settlement bank or liquidity provider. The remainder of this section discusses potential weaknesses relating to each of the typical safeguards and then considers the potential weaknesses in money settlement arrangements, weaknesses that can increase risks of losses and liquidity pressures from clearing member defaults or from settlement bank failures. It concludes with some further discussion of the events that appear to pose special dangers.

The effectiveness of membership requirements may be undermined by limitations of regulatory capital requirements and difficulties clearing houses may face in obtaining relevant information about the financial condition of clearing members. In recent years, both banking and securities regulators have devoted considerable attention and effort to strengthening capital requirements for market risks generally and especially for derivatives. The issues are quite complex, however, and the capital treatment of risks on open positions in options arguably remains less than fully satisfactory.54 In any event, like clearing houses, regulators typically do not seek to set capital requirements sufficiently high to protect regulated entities from losses from all conceivable market conditions. Furthermore, as noted above, even if capital requirements are effective in ensuring that losses are ultimately covered, they may not ensure that clearing members possess the liquidity needed to meet their obligations to the clearing house within the very tight time frames that are typically involved.

Even if compliance with regulatory capital requirements could ensure the solvency (or even the liquidity) of clearing members, reliable information on compliance is usually available only at discrete intervals, say monthly or quarterly. Given the leverage and liquidity provided by modern financial markets, the financial condition of clearing members can change radically between regulatory reporting dates. As discussed earlier, clearing houses typically recognise the potential for rapid deterioration in the financial condition of members and, for that reason, conduct ongoing surveillance of the condition of members. However, globalisation and financial innovation have made effective surveillance increasingly difficult. A clearing member's vulnerability to market price changes can only be evaluated on a portfolio basis, and the relevant portfolios often consist of related positions on different exchanges and in the OTC derivative and cash markets. A reliable analysis of the financial condition of clearing members requires information on all material positions, but information on positions in other markets that is both reliable and timely is often very difficult to obtain.

As noted earlier, both derivatives exchanges and their regulators have declared their intention to routinely share information on positions of common or related participants, including clearing members, but in many cases these intentions have not yet produced concrete agreements, in part because some clearing houses do not yet routinely collect information on large positions of clients of clearing firms.55 Obtaining reliable and timely information on positions that are held in the OTC markets poses even more formidable challenges.

In the case of margin requirements, clearing houses face some of the same market risk measurement difficulties that regulators face. However, the limited range of products traded on individual exchanges allows the development of valuation and risk measurement models tailored to those specific products. In this regard, clearing houses were, in fact, pioneers in the application of option pricing models to the measurement of risk. Also, the time horizon over which risk is appropriately measured (one day or perhaps somewhat longer to allow for market illiquidity) is far shorter than the relevant horizon for regulators,56 and risk measurement over longer horizons is inherently much more difficult. Still, the risk measurement problems facing clearing houses are challenging. Option pricing models are, after all, only models, and market prices often diverge from model values, especially when options are far out-of-the-money. Furthermore, as exchanges trade an ever wider range of products, difficulties in aggregating measures of risk for different underlying assets become more important.57

Another important potential weakness relating to margin requirements is the fact that many clearing houses have only limited control over intraday exposures to clearing members, and none has real-time controls. As a result, a defaulting member could have substantially increased its open positions between the end of day T-1 and the morning of T+1, when margin requirements typically must be settled. Because most clearing houses agree to act as counterparty to all matched trades and trades are typically matched throughout the trading day, the clearing house is exposed to losses on the newly established positions, but has collected no initial margins to support them. To be sure, improvements in recent years in the speed with which clearing houses can match trades and compute open positions have enhanced their capacity to monitor this source of risk. Indeed, screen-based trading systems may permit real-time monitoring of positions and exposures. Furthermore, the authority to conduct intraday margin calls offers a mechanism for managing the risk. Nonetheless, some clearing houses still measure their exposures to clearing members only at the end of the trading day. Moreover, even with real-time position and exposure monitoring and intraday margin authority, sharp intraday price movements could still produce significant intraday exposures to clearing members.

Clearing house risk management systems implicitly assume that in the event of default the defaulting member's open positions can be closed out or transferred quite promptly. However, as already discussed, market conditions may not permit a prompt close-out of house positions (or, where necessary, client positions). In addition, operational and legal impediments to prompt resolution may be encountered. A recent IOSCO report emphasised the importance of contingency planning for a default, but some clearing houses may not have adequate plans, in part because they have never experienced a default.58 Finally, the clearing house's efforts to promptly close out a defaulting member's positions and liquidate its margin collateral (or to transfer its client positions and margin assets) might in some cases be frustrated by legal regimes that seek to freeze the assets of an insolvent firm. If delays arise for any of these reasons, potential credit losses from a default increase.

Still another potential weakness is the possibility that supplemental clearing house financial resources are insufficient in amount or in liquidity to cover losses or liquidity pressures from clearing member defaults. The factors that produce a need for supplemental resources ­ the potential for extreme price movements, increases in open positions between margin settlements and delays in resolving defaults ­ are all extremely difficult to quantify. Moreover, clearing houses may not have procedures in place to even attempt to make timely assessments of the magnitude of these risks and their implications for the need for supplemental resources. Finally, clearing houses may be overestimating the liquidity of the resources they have. As already discussed, the value of margin assets of the defaulting member may not be realisable in time to be used to meet the clearing house's obligations. Likewise, if capital and reserves have been invested, delays could be encountered in realising their value; credit lines or bank guarantees may not require the banks to provide the funds until some time has elapsed; and assessments on members almost surely cannot be collected in time to help the clearing house meet its obligations on schedule.

Clearing houses may also be vulnerable to weaknesses in money settlement arrangements that can increase their credit risks vis-à-vis clearing members or private settlement banks. These potential weaknesses have two basic sources: (1) the use of payment systems that entail the risk of unwinds of provisional transfers late in the day,59 and (2) a lack of clarity regarding the obligations of various participants in the settlement process ­ clearing members, settlement banks and the clearing house ­ and how those obligations would be affected by the failure of a clearing member or private settlement bank.

Although central banks in the G-10 countries and in many other countries have implemented or are about to introduce real-time gross settlement (RTGS) payment systems in which funds transfers are final, that is, irrevocable and unconditional, when processed, in many countries today funds transfers are still netted and are provisional until all participants have covered their net debit balances. If any participant fails to settle its obligation, some or all of its provisional transfers of funds to other participants can be unwound. Those other participants could have difficulty coping with the resulting liquidity pressures, especially if the effects of the unwinds do not become apparent until late in the day, when money markets tend to be rather illiquid.

In the central bank settlement model, if the central bank payment system is a provisional net settlement system, the clearing house's exposures to a clearing member are not extinguished until the payment system achieves finality, which may be well after a provisional payment is received from a member. During the interval between the time a provisional transfer is received from a clearing member and the time it becomes final, exposures to the clearing member can increase as a result of changes in contract values and collateral values and new trades that increase its open positions. If the clearing member's failure only becomes apparent when the payment system attempts final settlement and if money markets are illiquid at that time, the unwinding of transfers to the clearing house from the failed clearing member may place substantial liquidity pressures on the clearing house.

In the private settlement bank model, the clearing house's exposures to its clearing members can be extinguished before the payment system achieves finality, provided that the settlement agreement specifies that transfers between the clearing members and the clearing house on the settlement bank's books are final. However, the clearing house's exposures to settlement banks generally cannot be eliminated until each settlement bank at which clearing members on net owe money to the clearing house has completed a final transfer to a settlement bank at which clearing members on net are owed money by the clearing house. Use of a deferred net settlement system increases the clearing house's credit risks vis-à-vis its settlement banks by lengthening the interval during which such surpluses are held by the settlement banks and, therefore, lengthening the interval during which the clearing house is vulnerable to loss should a settlement bank fail. Liquidity risks may also be substantial if an unwind of a transfer between settlement banks requires the clearing house to promptly cover the deficit of the settlement bank that received a provisional payment from the failed bank that was subsequently unwound.

Although a lack of clarity can be a weakness in any money settlement arrangement, it is perhaps especially likely to arise in the private settlement bank model because of its relative complexity. In either model, one question that must be answered to assess the risks to the clearing house from a clearing member's failure is when, and under what conditions, transfers on the books of settlement banks from clearing members' accounts to the clearing house's account become final. If the clearing house believes that a credit to its account is final when in fact it is provisional, it may seriously underestimate the potential losses and, especially, potential liquidity pressures from a clearing member's default. In the private settlement bank model, a proper assessment of the risks of settlement bank failures requires an understanding of when and under what conditions funds transfers between settlement banks become final. In this case, if a clearing house mistakes a provisional interbank transfer for a final transfer, it may seriously underestimate the potential losses and liquidity pressures from a settlement bank's failure.

Individually, the potential weaknesses in the various safeguards that have been identified might not jeopardise a clearing house's ability to cover losses and meet its obligations without delay. But perhaps the greatest danger is posed by events that might simultaneously weaken several safeguards. One such event would be an extreme price movement, larger than that allowed for in setting margin requirements, that leads to an endogenous default of a clearing member, that is, a default that results from losses on house or client positions carried by the clearing member at the clearing house, rather than from losses from some other (exogenous) source. Price movements not covered by margin leave the clearing house with uncollateralised credit exposures to one or more insolvent clearing members. As knowledge of the default spreads, the exchange markets might become less liquid, especially if doubts were to emerge regarding the clearing house's financial integrity,60 and the loss of liquidity would increase the time and resources needed to resolve the default. Finally, as was discussed in Section 2 and was illustrated in Exhibit 2, when prices on the exchange change dramatically, so too does the value of money settlements. This would increase the size of exposures to private settlement banks and would tend to bring to the surface any disagreements about the obligations of the various parties involved in the settlement process ­ settlement banks, clearing members and the clearing house.

Another event that could jeopardise a clearing house would be default by a settlement bank. This, too, could be an endogenous default if the settlement bank itself was also a clearing member and failed because of losses incurred on the exchange through proprietary trading or trading activities of its clients, including affiliated firms. Even if a settlement bank's default were not endogenous, it could be perilous if the bank was utilised by multiple clearing members which owed large amounts to the clearing house or if the settlement bank was also relied upon heavily as a guarantor of other clearing members or a provider of liquidity to the clearing house or many of its members.


54 A particular problem is how to capture the non­linear responses of option values to changes in the value of the underlying asset. In the recent amendments to the Basle Capital Accord, for example, the quantitative standards for use of the internal models approach to capital requirements for market risk require that "banks' models must capture the non­linear price characteristics of options positions". But in the near term, a complete simulation of the effects of non­linearities does not appear to be required. Instead, national authorities will rely on "stress tests" to reveal weaknesses in options modelling techniques and have the authority to require more capital if significant weaknesses are evident.

Interestingly, some securities supervisors have made significant progress in measuring market risks on exchange-traded options by incorporating into capital standards the results of simulations of option pricing models developed and maintained by the exchanges' clearing houses. As discussed earlier, many clearing houses perform full revaluations of options portfolios to determine margin requirements.

55 A recent report by the Technical Committee of the International Organization of Securities Commissions (1996b) discussed the purpose and structure of information-sharing agreements and surveyed the extent to which information on large exposures, defined as open positions sufficiently large that the clearing house would be placed at risk if the exchange member (clearing member or client) were to default, is currently available to market authorities (including clearing houses).

56 For regulators the relevant horizon is that over which they can reliably detect and resolve an insolvency of a regulated entity. This arguably should be measured in months, not days. For example, the Basle Committee's internal models approach to market risk measurement sets out the goal of measuring risk over a ten-day interval, and then applies a multiplier (a minimum of three) to the resulting estimate, in part to reflect the potential for losses to cumulate over a longer time horizon.

57 The basic issue is the degree of correlation between changes in the values of the various underlying assets. Correlation estimates tend to exhibit temporal instability, either because the correlations vary over time or perhaps because correlation estimates are subject to substantial error.

58 Technical Committee of the International Organization of Securities Commissions (1996b).

59 Clearing houses that accept securities as margin collateral may face similar vulnerabilities if the securities settlement systems used are deferred net settlement systems.

60 Fears about the clearing house's financial integrity could become self-fulfilling if, for example, clearing members become reluctant to meet their obligations to the clearing house.

Contact us * Risk Library * Documents by Author * Committees at the Bank for International Settlement (BIS) * Clearing Arrangements for Exchange-Traded Derivatives