Stress testing central counterparties for resolution planning

The first-ever resolution plan for a Canadian financial market infrastructure

The Payment Clearing and Settlement Act requires that the Bank of Canada develop a resolution plan for each domestic designated financial market infrastructure (FMI).1 In early 2024, the Bank completed its first resolution plan for a designated Canadian FMI. This plan applies to the Canadian Derivatives Clearing Service (CDCS), which is operated by the Canadian Derivatives Clearing Corporation (CDCC).2 Other designated FMIs within the scope of the Canadian resolution regime include CDSX, Lynx, the Automated Clearing Settlement System (ACSS), Interac e-Transfer^® and the Interac Inter-Member Network.3

To develop a credible resolution strategy, it is essential to have a sense of the magnitude of potential losses that the Bank will likely have to temporarily manage if an FMI were to fail. For this purpose, we apply the methodology based on the extreme value theory (EVT) for stress-testing central counterparties (CCPs) that was developed in Raykov (2022) to estimate the amount of default losses for various scenarios of CDCC’s failure.

In this note, we provide an overview of this methodology, which can help estimate financial resources that may be needed to successfully resolve CDCC.

Why financial market infrastructures need resolution plans

FMIs are systems for making payments and for clearing and settling financial transactions. Because they are crucial to the financial system, the services they provide must continue even in periods of extreme stress. Therefore, the Bank designates and oversees FMIs that pose systemic risk to the financial system. These FMIs must have robust risk management practices in place. As a result, it is unlikely that a designated FMI would fail. Nevertheless, the Bank, as resolution authority for Canadian FMIs, needs to be well prepared and have the tools to respond if necessary. For this reason, the Bank has a resolution regime for domestic designated FMIs with objectives aimed to:

• promote financial stability
• maintain the continuity of FMIs’ critical services
• minimize the exposure of public funds to loss

One element of the framework is resolution plans that describe on a strategic level how to respond to the unlikely failure of a designated domestic FMI. They are customized to the FMIs that they are prepared for and include sections on the risk profile of the system, resolution strategy, financial resources assessment and operational plan (Bank of Canada 2020). These plans must be updated at least once every 12-month period.4

When an FMI is in resolution, the Bank might need to temporarily provide financial resources to stabilize the FMI and to minimize any disruption in the provision of its critical clearing and settlement functions. To this end, the Payment Clearing and Settlement Act gives the Bank the power to take temporary control of a failing FMI to limit the impact of its failure on Canada’s financial system and economy. The Bank will recover the cost of resolution from the FMI and its members afterward to mitigate any moral hazard and meet its statutory obligation to minimize loss of public funds. The financial resources section of the resolution plan helps the Bank determine the amount of funds that would be temporarily needed to support resolution.

Overview of CDCC’s credit risk and default management resources

CDCC is a systemically important CCP that provides a central clearing service for:

• exchange-traded derivative products (futures and options) traded on the Montréal Exchange
• over-the-counter Canadian fixed-income transactions (both cash and repurchase agreements) and equity options

Both the exchange-traded derivatives and fixed-income clearing services are critical to the stability of the Canadian financial system and would need to be supported by the Bank during resolution. Currently, the clearing services that CDCC provides to the Canadian financial market have no substitute.

CDCC serves more than 30 clearing members, including major financial institutions and brokers in Canada as well as their clients worldwide.5 This includes the six Canadian systemically important banks and their broker-dealer subsidiaries and the Canadian investment dealer subsidiaries of some large foreign banks.6

As a CCP, CDCC reduces counterparty risk for its clearing members by becoming the counterparty in each contract through a process known as novation.7 Novation replaces the original contract between the buyer and seller with two new contracts: one between the CCP and the buyer and the other between the CCP and the seller. As a result, if a clearing member were to default, CDCC must meet that member’s obligations until its positions are closed out. Because losses can be incurred during the management of the default, CDCC is exposed to credit risk and must maintain pre-funded financial resources to absorb such losses.8

The pre-funded financial resources that CDCC can use to cover credit losses if a clearing member were to default are as follows.

• Initial margin: Collateral collected from each clearing member so that, if the member defaults, potential losses in the value of that member’s positions can be covered with a high degree of confidence (e.g., 99% or higher). Figure 1 shows that only the defaulting member’s initial margin is used to cover losses when a default occurs.
• Clearing fund: A mutualized fund with cash collected from CDCC’s members to account for extreme-but-plausible market events. Figure 1 shows that, in the case of a default, the losses are first covered using the defaulting member’s contributions and then CDCC’s default risk capital. Once those are exhausted, the survivors’ clearing fund contributions are used.
  Because CDCC is designated as systemically important in Canada, its clearing fund is resized regularly to meet the Cover 1 regulatory standard, by maintaining sufficient pre-funded resources to cover the default of a single clearing member (and its affiliates) representing the largest stressed aggregate credit exposure for CDCC.9
• CDCC default risk capital: CDCC’s own capital reserves (skin-in-the-game).

CDCC’s default management framework (Figure 1) shows the order in which CDCC must draw on these financial resources. If residual losses remain after applying all the resources in the default waterfall, CDCC may initiate the recovery process as part of its default management framework. During this process, CDCC can use various recovery tools (e.g., reduced amounts distribution, cash calls, voluntary contract tear-ups) for loss allocation. If the Bank judges that exhausting the default waterfall and recovery actions results in financial stability risk, it may consider entering CDCC into resolution even before CDCC uses all the tools in its default management framework.

Extreme value theory: Estimating the magnitude of extreme price shocks

The main source of stress for a CCP that could trigger its resolution is a large amount of unmet payment obligations when multiple members default at the same time. For such scenarios to occur, the CCP needs to experience sufficient financial stress arising from a combination of sufficiently large open positions and price movements.

To generate sufficiently large price movements in our analysis, we estimate counterfactual price shocks beyond the extreme-but-plausible range with EVT and apply them to historical member positions at CDCC to calculate the resulting payment obligations.

EVT is a statistical method used to quantify the severity of extremely rare events. It is based on the idea that the tails of heavy-tailed distributions are well approximated by power functions using a tail index, α (Figure 2). EVT is well suited for our purpose because the returns on main derivatives contracts traded on CDCC feature heavy tails, with higher probabilities of extreme outcomes than the normal distribution.

EVT offers significant advantages relative to conventional stress testing approaches. Conventional stress testing is based on historical or hypothetical scenarios. However, we cannot use a historical scenario to assess the resources needed for CCP resolution because there is no prior case of CCP resolution in Canadian history.10 Further, hypothetical stress scenarios inherently have no historical frequency, which makes it difficult to put them in context or defend one hypothetical scenario as being more plausible than another. Unlike conventional stress testing methods, EVT allows us to obtain valuable information about the tail’s behaviour even outside the range of observed data. This allows us to gauge both the severity and the likelihood of extreme events even if they have not yet occurred.

For our EVT stress test, we estimate extreme price movements with very low likelihoods (once in 50 years and once in 100 years) for the three most-traded futures contracts on the Montréal Exchange, which account for over 90% of the open interest in the futures trades before 2023:11

• BAX: three-month Canadian bankers’ acceptance futures
• CGB: 10-year Government of Canada bond futures
• SXF: S&P/TSX 60 index standard futures—Canadian large cap stock market index

As expected, the EVT price shocks of one in 50 years and one in 100 years are much larger than the lowest returns observed in the sample data (Table 1). For example, the EVT price shocks for the SXF are -23.32% and -30.49%, around two to three times larger than the lowest observed return of -10.20%. Because any market stress that could potentially trigger a resolution of CDCC would likely be worse than anything we have seen in recent history, these substantially larger price shocks compared with the historical lows are considered appropriate for our resolution scenarios.

Counterfactual analysis: Estimating losses for default scenarios

CDCC is required to have sufficient pre-funded resources to cover the loss resulting from the default of a single clearing member and its affiliates with the largest credit exposure under stressed market conditions. For this reason, any resolution scenario due to default must entail multiple members defaulting at the same time. We therefore focus our analysis on the scenarios of two or three members with the largest exposures defaulting on the same day.

To estimate the magnitude of loss from multiple defaults, we first conduct a series of counterfactual analyses by applying the EVT price shocks reported in Table 1 to the historical positions data provided by CDCC. We then compare these losses with CDCC’s pre-funded resources, including the initial margin of the defaulters as well as CDCC’s clearing fund and its default risk capital.

We use two measures to assess:

• the loss to be absorbed by survivors
• the uncovered or residual loss after exhausting all pre-funded resources

Measure 1: Loss to be absorbed by survivors (survivors-pay)

This is calculated as the loss minus defaulters’ pre-funded resources. This measure refers to the magnitude of default losses the surviving members need to absorb. CDCC can cover part of this loss using the pre-funded resources (its own default risk capital and survivors’ contributions to the clearing fund) as part of its default management framework (Figure 1).

Chart 1 shows the amount of loss to be absorbed by surviving members in the quarterly worst-case scenarios under four cases:

• two defaults under a 1-in-50-year price shock
• three defaults under a 1-in-50-year price shock
• two defaults under a 1-in-100-year price shock
• three defaults under a 1-in-100-year price shock

For resolution planning, we focus on the worst-case scenarios under each of the four cases:

• lower bound: highest loss in the two defaults under a 1-in-50-year price shock
• upper bound: highest loss in the three defaults under a 1-in-100-year price shock

Chart 1: Counterfactual credit loss estimates in various resolution scenarios

Chart 1: Counterfactual credit loss estimates in various resolution scenarios

Stress-test loss after applying defaulters’ margin and their clearing fund, quarterly maximum

Note: The amount of simulated losses and the identity of the clearing members are omitted intentionally to maintain confidentiality.
Sources: Canadian Derivatives Clearing Corporation and Bank of Canada calculations

Measure 2: Uncovered or residual loss after exhausting all pre-funded resources

The uncovered or residual loss is calculated as the loss minus all pre-funded resources, including those of surviving members and CDCC. The estimate of the residual loss indicates the amount of funds that CDCC will need to recover from survivors using its recovery tools that are not pre-funded. If we assume that this could risk broader contagion by imposing stress on survivors who may already be under stress given the market context of multiple defaults (e.g., increased margin requirements) and the Bank steps in as resolution authority, then this amount would be the loss that the Bank would need to cover temporarily until it can be recovered from survivors later in the resolution process once the stress has stabilized.

Chart 2 shows the quarterly worst-case losses to be covered by surviving members (bars) and the pre-funded resources of the surviving members on the same day (markers). If the loss (bar) is greater than the pre-funded resources available (marker) in the corresponding period (e.g., the first quarter of 2022), the pre-funded resources do not cover the loss due to default, and the Bank may have to cover the shortfall temporarily if it steps in to avoid potential for contagion.

Conclusion

Developing a credible resolution plan for a designated FMI requires information on the amount of losses that the resolution authority may need to temporarily manage if the FMI becomes non-viable. Because a failure of an FMI is extremely rare, the stress-testing methodology must be able to capture the loss distribution at the extreme ends effectively, even if the historical data do not include any such events. We suggest the EVT methodology, which is particularly well suited for such application and demonstrate its application for CDCC. We plan to refine our EVT-based stress-testing methodology in the future. We also plan to explore other methods that perform well when dealing with extreme scenarios.

Acknowledgements

We would like to thank Natasha Khan, Isabel Zheng and Alexandre Ruest for their valuable comments and support for the project.

References

Bank of Canada. 2020. Guideline Related to the Bank of Canada’s Role as FMI Resolution Authority.

Faruqui, U., W. Huang and E. Takáts. 2018. “Clearing Risks in OTC Derivatives Markets: The CCP-Bank Nexus.” BIS Quarterly Review (December).

Financial Stability Board (FSB). 2017. Guidance on Central Counterparty Resolution and Resolution Planning.

Financial Stability Board (FSB). 2024. Key Attributes of Effective Resolution Regimes for Financial Institutions.

Raykov, R. 2022. “Systemic Risk and Collateral Adequacy: Evidence from the Futures Market,” Journal of Financial and Quantitative Analysis, 57 (3): 1142–1173.

Rehlon A. and D. Nixon. 2013. “Central Counterparties: What are They, Why Do They Matter and How Does the Bank Supervise Them?” Bank of England Quarterly Bulletin (Q2: 147–156.

Endnotes