Conventional stress tests usually estimate the ability of a CCP to withstand non-payment by its two largest counterparties, but, the report indicated that the safeguarding via beneficial DTCC data and the ‘Cover 2’ standard, draw a “different conclusion when [they] take into account network contagion”.
In the report entitled ‘How safe are central counterparties in Derivatives Markets?’, economists and researchers, Mark Paddrik and Hobart Peyton Young, explained that the introduction of the network contagion model in a stress test highlights new vulnerability to central clearing models.
Paddrick and Young further indicate how the network contagion model coupled with the Cover 2 standard, increases the likelihood of a CCP defaulting during a stress test of similar magnitude to the Federal Reserve’s 2015 Comprehensive Capital Analysis and Review (CCAR) shock.
This shock was “specifically designed to subject the financial markets to a severe but plausible market stress”.
Such a shock triggers a sudden drop in the value of credit instruments, which translates into large and sudden VM payments on CDS contracts.
The report stated: “When we take account of these network spillover effects, we find that the CCP for this market is potentially more vulnerable to default than conventional stress tests would suggest.”
Paddrick and Young suggest their advantage in their investigation is mainly due to their “use of credit default swap data to estimate the direct and indirect impacts of a default by CCP counterparties in derivatives trades.”
The contagion model changes the game by tracing how payment delinquencies by some firms can increase as they travel through the network of credit default swap (CDS) exposures.
The report states that the “Cover 2 standard is typically applied to a scenario where the two defaulting members are assumed to be those with the largest net variation margin (VM) obligations to the CCP”.
“A novel feature of this model is the treatment of stress transmission, which depends on firms’ liquidity buffers and their risk management policies. We demonstrate the sensitivity of the model to different values for the shock and stress transmission parameters.”
“The model allows us to estimate the probability of a CCP failure relative to the probability of a member’s failure while making minimal assumptions about the degree of correlation among member failure rates.”