
June 30, 2025 (NO COMMENTS)
Complex problems often require novel thinking, yet a recent proposal for an industry-wide emergency collateral framework may have pushed creativity a step too far.
In its whitepaper on collateral liquidity efficiency in the derivatives market, the International Swaps and Derivatives Association’s Future Leaders in Derivatives (IFLD) group made 16 recommendations. Among them, a proposed stress event protocol that would temporarily expand the types of collateral that could be posted under a credit support annex (CSA) has stirred intense debate, even among the members of the 31-strong group that penned the report.
“What you’re doing is, you’re saying, ‘Well basically it doesn’t matter what the quality of collateral is anymore, just put it there’,” says a member of the group who had reservations about the specific proposal.
Other members countered that in some scenarios even high-quality assets can be hard to use as collateral and the ability to post other types of collateral – even for a short time – could keep the system moving.
Despite the split in thinking, the group included its most novel recommendation “to see how the market would take it,” says the member.
Feedback so far suggests the concept may be flawed.
Eleven industry sources who spoke to Risk.net flagged fundamental issues which would prevent the proposal coming to fruition, though many welcomed what they saw as a useful thought exercise.
“Simply from a documentation point of view, your entire capital calculation is premised on this and you’re saying, ‘I’m just going to throw that away when things start getting tough’. Every regulator in the world is going to load you with more capital,” says a financial markets consultant who has worked on collateral matters.
“I don’t think it’s wrong for them to bring it up. At least people have a debate,” this person adds.
The proposed emergency collateral framework would allow consenting parties to exchange broader forms of collateral, including money market funds, bank guarantees, surety bonds, corporate commercial paper, or emission certificates, under pre-agreed conditions.
Some say that simply announcing such an event may be self-fulfilling.
“If you have a mechanism that says, ‘Now you should be scared’, then everybody will be scared in that specific scenario. So, it’s a very delicate signalling issue that you must put into that design,” says Nicki Rasmussen, head of XVA at Danske Bank.
One of the top concerns is how a stress event would be determined and activated.
The whitepaper suggests it could be trigged by a designated calculation agent, a public authority, or an industry panel, similar to Isda’s credit determinations committee, which rules on credit events.
Industry participants are sceptical there would be appetite to sit on such a panel and point to the recent review of the CDS committees’ governance and structure.
“I think the difficulty is parallel to the credit events world for credit derivatives, where there’s got to be some committee body that exists that’s impartial,” says a source at a collateral vendor. “That old credit event world has always been a little bit of a hot potato,” he adds.
Equally critical is determining exactly which factors would be taken into consideration.
The whitepaper suggests both objective and subjective criteria would be considered. Objective criteria might include existing regulatory mechanisms being triggered, for example an official request by the European Securities and Markets Authority for the European Commission to temporarily loosen clearing requirements due to market disorder. Subjective criteria might include spikes in market volatility across multiple asset classes.
A trading head at one European asset manager says by the time adverse pricing is visible in the system, it may already be too late.
“Before you observe something in pricing, you already see cracks in the system,” says the trading head. “You could have some kind of governing body that says, ‘There are cracks in the system, we should set up emergency measures’. That has to be an independent or functioning governing body instead of 10 people calling up their regulator or their bank,” he adds.
Some warn that even if a stress event is identified in one part of the market, it could give some an unfair edge.
“If a stress event gets called for a certain type of event that only impacts a certain type of market participant – for example an FCM that has targeted energy trading firms or LDI fund clients – do they suddenly get an advantage over other competitors?” asks Tom Griffiths, head of product at margining software vendor Cassini Systems.
“They could have won some risky business historically, and now, because of this protocol, it suddenly becomes far less risky because of the relaxed rules under a stress event.”
Introducing an element of optionality into collateral eligibility could create additional complexity when modelling funding and collateral valuation adjustments for derivatives pricing, according to Danske Bank’s Rasmussen.
“I think we’d prefer to put into the document and say that while you post this stress collateral you pay an additional X basis points on the cash equivalent. That could be one way to solve for the trade-off between the risk we run for accepting the additional collateral types and the cost for the client,” says Rasmussen.
“For most banks, I suspect that would be the pricing model, that you only pay if you need to use the facility.”
He adds that such an expansion would typically require internal credit committee approval. This would need to consider one-way risk, for instance if a counterparty predominantly trades specific assets which suddenly come under stress, they may prefer not to receive those as collateral.
“It’s very hard to devise something where you’d say, ‘I would always accept this as long as this regulator or official authority calls the whistle to say we are now in a stress scenario,” he says.
Cassini’s Griffiths warns of the operational uplift associated with adhering to the protocol, noting it is “beyond the current capabilities” of many of their clients, which include asset managers, pension funds insurers and hedge funds.
Déjà vu
Recent bouts of extreme market volatility, such as the 2020 Covid pandemic, the 2021 collapse of Archegos and the 2022 gilt crisis have heightened liquidity risk concerns, having exposed difficulties in being able to post eligible margin quickly.
Temporary expansion of CSAs has been used sporadically to ease collateral flows in these scenarios.
During the gilt crisis, pension funds rushed to dealer counterparties to temporarily amend CSAs to include corporate bonds on top of gilts and cash. While this process typically takes months, pension funds managed to make those changes in a couple of days.
After the global financial crisis many banks focused on narrowing down and standardising the lists of eligible collateral accepted under their CSAs.
At least one bank is understood to have put in place templates allowing select clients to extend the list of eligible collateral to a small range of pre-agreed assets for up to 90 days by telephoning the bank’s collateral management head.
In response to the energy crisis, European regulators expanded the pool of eligible collateral acceptable at clearing houses to include both public guarantees from European Economic Area governments and uncollateralised bank guarantees posted by non-financial companies. Ice clear Europe also started accepting emission certificates in 2022.
In the US, CME accepts uncollateralised commercial bank guarantees under certain strict conditions from non-financial counterparties.
Isda is understood to have discussed the whitepaper proposals with members of its margin and collateral working group.
The white paper is the fourth report from the IFLD programme, which aims to identify rising talent from firms and steer innovation in the derivatives market. Membership is open to any individual working at an Isda member organisation. The group works to create and share content that tackles key challenges in the industry. Prior whitepapers have focused on energy security, net zero and generative AI.