Clearing house innovation of the year: Ice Clear Credit

Risk Awards 2020: Clearing house lures fund business with efficient new Monte Carlo methodology

Left to right: Stan Ivanov, Richard Jordan, Ian Springle, Erik Pieczkowski

Photo: Geraint Roberts

When it opened in March 2009 in the aftermath of the crisis, with the industry’s backing to clear credit default swaps (CDSs), Ice employed a stress-based approach to margining the contracts. But it didn’t want to. From the get-go, Ice had ambitions to run a full portfolio simulation methodology that would unlock deeper capital efficiencies for members and clients, and give them confidence to clear non-linear products.

That ambition marked the start of Ice’s new Monte Carlo margin framework, which it finally switched to on February 1 this year.

However, the clearing house decided to be patient, choosing instead to expand the service with additional products the market needed, before executing on its vision for a more risk-sensitive framework. While the stress-based methodology focused on modelling fewer, larger moves to set margin requirements, the Monte Carlo-based framework simulates some 20,000 potential scenarios, with variables changed each time to parse a portfolio’s entire risk.

The new framework draws on the scenarios generated using the Monte Carlo approach to jointly model credit spread and recovery rate changes together – indicating how a position is affected by a change in spreads between the CDS and its underlying, as well as the chance of extracting part of the debt obligation upon default.

This changes a major historic bugbear for the clearing house: the fact that offsets between single-name and index CDSs were not reflective of risk, and that hedging an index position with single-name CDS or vice versa would be punitive from a margin perspective. Under the new methodology, firms with portfolios constructed in this manner have seen margins drop between 10% and 12% on average.

But the new model’s risk sensitivity cuts both ways: while initial margins decreased on average by 2%–3% across the clearing house, more directional portfolios have seen increases of 7%–8%.

Ice sees this formulation as mutually beneficial: “It is absolutely critical that our members and buy-side participants have the confidence and knowledge that the new portfolio approach will appropriately reflect the risk they are taking on and actively managing,” says Stan Ivanov, president of Ice Clear Credit. “Providing the proper incentive to firms to maintain flatter risk profiles and hedge their positions is key.”

The clearing house is betting big on members and clients using more single-name CDSs in their portfolios – and the numbers appear to bear this out: 42 clients cleared a single name for the first time in the first three quarters of this year, with a year-to-date growth of 18% in client notional compared with the same period last year.

Complementing the framework is a new margin analytics platform. Dubbed Pace, it allows members and clients to quickly and easily determine how initial margin requirements for a portfolio will evolve as trades are added or removed, and market dynamics shift. Ice says the new platform has been a big factor in helping it grow its buy-side clearing offering.

Providing the proper incentive to firms to maintain flatter risk profiles and hedge their positions is key

Stan Ivanov, Ice Clear Credit

The other major reason for the move, however, was for the introduction of more non-linear instruments: Ice hopes to launch index options by the end of the second quarter of next year. The clearing house wanted to introduce the new framework to better account for strategies like straddles, strangles or butterflies, says Ivanov, which can be affected by big moves in implied volatility, despite there being no significant move in the underlying instrument.

“A stress-based approach utilises a fairly small set of scenarios, that typically includes some significant credit spread widening or tightening outcomes,” says Ivanov. “However, the set doesn’t incorporate many scenarios encapsulating more moderate market moves that may lead to increased risk in portfolios with non-linear instruments as CDS options. You could have a very well-controlled and well-hedged delta risk, but the implied volatility or gamma risk, for example, of an options strategy can be very significant. In these cases, one needs a wide range of well-constructed scenarios to probe the many potential market outcomes, something the Monte Carlo approach does very well.”

The move to the methodology was well telegraphed to the market. After 2012, regulators and members started receiving quarterly back-testing results indicating the difference in performance between the stress-based approach and the Monte Carlo framework. Buy-side participants received reports three months before the February 2019 launch date. The CFTC and SEC approved the model in just three months, in part due to their familiarity with the methodology after years of back-testing reports.

Members say the switch was handled smoothly, and praise the robustness of the new model.

“We spent a lot of time reviewing the Monte Carlo framework and making sure it was appropriately calibrated,” says a CDS expert at a large US bank. “We think a framework like this actually helps the clearing house better understand changing market conditions, and that’s where traditional models break down.”

Primary force

The new model comes at a time when Ice Clear Credit remains dominant across several segments of the CDS market: the clearing house is set to trump volumes from last year in iTraxx credit derivatives in the Apac region and growth in client single-name volumes.

Like all models, the Monte Carlo-based approach has its detractors, with some saying these models are difficult for some clients to understand or replicate. Ice’s competitor LCH uses a VAR-based methodology for its CDS business, having jettisoned a Monte Carlo method for this reason. Ice Clear Credit’s sister clearing house for futures and options in Europe is opting for a simpler historical simulation value-at-risk based approach.

However, Ice is convinced that for parsing risk in credit derivatives markets, the Monte Carlo-based approach is the most sensible way forward.

“The CDS market requires a greater set of scenarios that goes beyond what has been observed in the past,” says Ivanov. “If a reference entity has never defaulted or approached a near-default state, pure historical simulations would not capture the potential risk.”