Small, speculative clearing members – are they worth the risk?

Central counterparties (CCPs) had long being perceived as nearly risk-free entities. The risk mitigants they employ, such as initial margin and variation margin, are designed to absorb losses in the event of a member default. In addition, each member contributes to the default fund, a safety net triggered when margin levels are not sufficient to cover the losses. Expectations are that the default fund would be affected only if multiple significant adverse events happen simultaneously.

However, in September 2018, a 17-standard deviation move in European power market prices set off a chain of events that led from the default of a relatively small clearing member – power trader Einar Aas – to the loss of two-thirds of Nasdaq Clearing’s default fund, equal to €107 million ($119.7 million).

The incident taught a lesson to CCPs, clearing members and regulators: existing models fail to capture the risk that a small clearing member can pose for all other clearing members – even clearing members that don’t have direct trading exposure to the small counterparty.

An explanation and a solution to this problem have been proposed by Leif Andersen, global head of the quant strategies group, and Andrew Dickinson, lead of the CCP analytics group, both at Bank of America. They realised that a CCP’s risk model needs to take into account the wrong-way risk that arises from the positive correlation between losses on the cleared portfolio of a member and that member’s own default.

The correlation can be close to one when the size of the portfolio cleared through the CCP is a significant portion of the total wealth of the clearing member. If the portfolio suffers a sudden large loss, the CCP will ask the member to post a large margin, but if the member has most of its assets invested in the cleared portfolio, it will be unable to fulfil the request and almost automatically find itself in default, leaving other clearing members to deal with the losses.

A small member with an unhedged position disproportionately large compared to its wealth, Andersen explains, “introduces as much credit risk as all the other members combined, just because of this wrong-way risk”.

The risk affects all clearing members – even the largest ones – because their contributions to the CCP’s default fund are at stake from a member default, as was evident during the Aas episode.

For typical clearing members – large financial institutions – a cleared portfolio with a given CCP is only a small fraction of their total assets. And in the case of banks, their trades, executed on behalf of clients, are normally hedged. An adverse market move is unlikely to affect their financial solidity.

“When we wrote the first paper we made some assumptions on who are the members,” Andersen recalls, referring to his earlier work with Dickinson in which they developed a model to calculate a bank’s exposure to a CCP. “Most of the members we were looking at were well-diversified large financials, but the attention was drawn to the fact that this is not always the case.” The discovery surprised the authors: “There’s much more heterogeneity, to the point where an individual trading from a remote fishing village can be an actual member of a CCP. Maybe I shouldn’t have been surprised, but this was an eye-opener for me,” Andersen adds.

A model that quantifies the counterparty credit risk to a CCP and includes that generated by a member with such a disproportionately outsized position, allows a bank to take decisions on its own membership of the CCP. If a CCP accepts a member with some element of wrong-way risk, a bank may want to avoid running greater credit risk and decide not to trade with the CCP altogether. The argument is clear: “If you allow members to trade outsized positions then you can contaminate the entire default fund,” Dickinson says.

There’s much more heterogeneity, to the point where an individual trading from a remote fishing village can be an actual member of a CCP

Leif Andersen, Bank of America

From a CCP’s perspective, equivalent considerations can be made. To keep the overall credit risk at a given level, Dickinson says CCPs have a few options: “they can limit the size of the positions…limit the membership or increase initial margin for those members that pose the greatest risk.”

The model is an extension of the earlier framework the duo has developed. The clearing member’s exposure to the CCP is constituted chiefly by the risk of default of another member and the cost of funding of margin and contributions to the default fund. The authors model them using existing methodologies for credit valuation adjustments and margin valuation adjustments.

A key ingredient of the updated model, in order to be able to capture wrong-way risk and the large market moves that generate it, is to use a statistical distribution that allows for fat tails. Andersen and Dickinson’s model uses a Student-t distribution which can be adjusted to capture any thickness on the tails, important for credit events derived by moves as large as those in the power market and even more.

The events at Nasdaq Clearing have focused the attention of both industry and regulators on CCP credit tail events, and one can expect further scrutiny of these risks in future. “With our new framework, market participants now have a practical tool to analyse and manage CCP tail risks,” Andersen notes.