David Kelly, Managing Director
Last year Einar Aas, a Norwegian trader in a small corner of the power market racked up losses he could not cover, leaving Nasdaq AB, the central counterparty (“CCP”), to sell off the position leading to a trading loss of €114m. The trade was a spread between Nordic and German power that moves typically in tandem, but thanks to heavy rain in hydro-central Norway triggered a bifurcation.
The size of the losses dwarfed what the risk modellers of the CCP thought was possible that held only €7mm of margin against these trades.
To cover the €107m shortfall Nasdaq told its members to replenish the default fund that was only €166m. Members are asking questions around why they have no transparency or say in the management of other positions but are liable when things go wrong whereas the CCP avoid losses even though it is their risk process that has failed.
The Nasdaq spokesperson suggesting the market move was a “specific situation”. Given the west coast of Norway is affected by the Gulfstream, with annual precipitation in Bergen of 2200mm, “special” looks a stretch. The statement does indicate that the assumptions of market liquidity were optimistic given the Power market is dominated by generators who understand physical delivery at short notice. Their explanation shows a naivety in the assumption that the correlation between similar by different contracts is stable and can quickly return to mean.
The introduction of mandatory clearing in the past decade has concentrated the risk of default to a small number of clearers with limited financial resources. No CCP has a default fund that can cope with a scenario of more than one significant member defaulting. To cap it off, CCP is companies in competition to provide a service that paradoxically relies on members for their capital for their continuity. One way to gain market share, in particular in these small markets is to offer more attractive initial margin rates than the competition or provide generous netting offsets (50% in the above example) to entice participation.
Banks apply a 2% haircut against their exposure to CCPs in line with Basel prescription which works fine for the rates products that are the bulk of outstanding derivatives and are thus liquid. For the less liquid markets, banks have started to ask whether they are measuring the risk correctly and are wondering what the optimal strategy for CCP membership.
Unlike measuring the risk of bi-lateral trades, the information available is limited. Given banks have no choice but to use CCPs, the question of risk measurement is around which membership to invest in and which to divest.
To start measuring the risk of outlay due to member default, we start with the auction and each bank’s participation. During default proceedings, the CCP aims to offload the portfolio, but the likelihood is that the positions are illiquid reflected in the price available from the participants. The bank could find themselves benefiting from taking a position on where there is “blood on the street” bounce or wishing they had not and end up “catching a falling knife”. Either way, the bank needs to know that taking on any position in a fast market creates heightened PL volatility.
The CCP uses the Default Fund contribution of the defaulted player first before using the funds of the other players to cover the losses from the auction. The CCP then tells the membership to replenish the Default Fund. Not all members participate in the auction, so the CCP has to compensate those that do as they can easily suffer further losses and also have to replenish the default fund. On the other hand, there are games to be played around where to price the auction portfolio.
The risk-taking activities of a defaulting member have put considerable strain on the resilience of the CCP. The CCP relies on the network of members to step up in terms of taking on risky positions as well as default fund replenishment. If one of the remaining members also indulged in the same risk misadventure, then the potential for an additional member default increases, leading to a death spiral scenario on a systemic piece of financial infrastructure.
“The Quant Foundry CCP credit model”.
The model defines likely stress scenarios based on the CCP product suite. The model uses the Quant Foundry Stress Indicator to highlight bubble-type market events that can lead to a member default.
The network model uses public information of banks and CCP balance sheet information – illiquid assets, liquid assets, margin, debt, equity to provide a estimate that such default events could come to pass.
The member defaulting scenario then goes through the Quant Foundry Network Resilience model where the user can define their commitments in the CCP. Commitments include whether they are involved in the auction sale, how the CCP compensates them and allocates losses and overall market size.
The model provides the user with a distribution of potential losses based on the complex interplay of market movements, member default, follow-on market events post-auction and default fund replenishment. The model also provides an assessment of network resilience of the CCP members to avoid a second default.
Bank users can take a percentile point from the distribution of potential losses and provide insight into the relative value and potential downside of having membership across several CCPs.
The Quant Foundry CCP Credit Model is currently in prototype form and is on our integration pipeline onto the KRM 22 Analytics suite for their Risk Cockpit platform.