BoE researchers flag potential flaws in CCP default funds

Research published today by the Bank of England (BoE) tests the defences of central counterparties (CCPs), finding they are robust but not impenetrable.

The financial stability paper asks whether the ‘cover two' standard for CCPs is enough to ensure their safety - finding it to be a "prudent" framework, but with some weaknesses.

Cover two states that systemically important CCPs must have sufficient resourcesto survive the failure of their two largest members, under "extreme but plausible" circumstances. Both EU and US regulators have implemented versions of the standard.

CCPs often sit at the centre of a vast number of derivatives trades, making many of them ‘too-big-to-fail'. The paper's authors, BoE economists David Murphy and Paul Nahai-Williamson, quote former BoE deputy Paul Tucker in saying "it is an understatement that it would be a disaster if a clearing house failed".

CCPs are well aware of this risk, and as a result maintain numerous lines of defence. First, they only admit members that meet a certain standard. They also adopt a ‘defaulter pays' model, requiring members to post enough margin to, theoretically, cover all losses.

If the margin is not enough, the CCP's own capital is at risk, but that is supplemented by a default fund that all members pay into. The CCP can call on members for additional funds in a crisis, and in the extreme its own equity can be used to absorb losses. After that, it fails.

The probability of failure

Cover two aims to ensure the default fund is of suitable size to prevent failure, but Murphy and Nahai-Williamson note that it is an "unusual standard" as it pays no heed to the number of members a CCP has. They set out to estimate the likelihood of a CCP's failure if it holds the minimum level of funds, and test how this probability shifts in the presence of different distributions of risk.

The authors deliberately use a simple model to display their results "simply and starkly". They use data on credit default swaps in a ‘Gaussian copula' model, which gives an estimate of the probability of hypothetical clearing members A and B failing simultaneously.

The results imply that cover two is a "prudent standard" under those risk distributions that are "likely to be found in practice", the authors say. However, some distributions, particularly where risk is spread evenly throughout clearing members, give rise to "weaknesses" in the cover two standard.

Murphy and Nahai-Williamson suggest regulators ought to take action against such a possibility. First, "it would be sensible for CCPs and their supervisors to monitor the whole of the loss distribution," they say.

Second, they propose a new standard that could be applied in future revisions to the cover two standard. Regulators could set a "backstop" to cover two by demanding that default funds meet a minimum ratio of cover two to ‘cover all' – the amount needed to cover all members defaulting.

This would have the benefit that if risk becomes more evenly distributed among members, the backstop will become binding, preventing the fund from shrinking too far.

"One basic requirement for calibrating this percentage would be knowledge of the ratio of cover two to cover all, so a reasonable first step would be the disclosure of both measures by all systemically important CCPs," the authors say.