CPMI-IOSCO Public Disclosure by European CCPs

European CCPs have recently started to publish lots of quantitative data covering their Default Fund, Initial Margin,  Collateral, Credit Risk, Liquidity Risk and other Financial Disclosures. This has been done in compliance with the guidance provided by CPMI-IOSCO with the public disclosure aiming to further increase the transparency of financial markets and CCPs in particular.

The data is made publicly available on each CCPs’ websites. A good resource is the website of the European Association of Clearing Houses (EACH), which provides links to the data.

European CCPs

Disclosure is made on a quarterly basis with a quarterly lag, so currently data as of 30 Sep 2015 is available for the following European CCPs:

• BME Clearing S.A. (Spain)
• CC&G (Cassa di Compensazione e Garanzia, Italy)
• CME Clearing Europe
• Eurex Clearing
• ICE Clear Europe
• LCH.Clearnet Ltd
• LCH.Clearnet S.A.
• LME Clear
• Nasdaq Clearing
• NCC (National Clearing Centre, Moscow Exchange Group)

Public Quantitative Disclosure

Now the obvious benefit in the public disclosure is being able to compare European CCPs on a like-for-like basis for specific quantitative measures. That is what I planned to do when I started researching this blog. However as there are approximately 200 fields of quantitative data, each with a lengthy description, that is going to be a tall ask to do in one article, both for me to distill from the detail and for you the reader to digest.

So for today, I am going to focus on some of the most interesting numbers disclosed and use one of the largest CCPs, LCH.Clearnet SwapClear’s disclosures, to illustrate the value of the numbers.

Lets get going.

Default Fund (4.1)

The first interesting set of numbers are for the Default Fund, which for SwapClear consists of:

• Pre-funded – Own Capital Before (Member resources) of £34.7 million ($50m)
• Pre-funded – Aggregate Participant Contributions of £2.66 billion ($3.8b)
• Committed – Aggregate Participant Commitments to address a default of £2.66 billion ($3.8b)

Big numbers indeed.

Initial Margin (6.1)

Lets now look at Total Initial Margin (IM) Required to get an insight into the size of positions held by members.

Again for SwapClear:

• IM for House Net is    £28.8b
• IM for Client Gross is £16.3b
• IM for Client Net is     £  0.6b
• IM Total is                    £45.7b

Even larger numbers and interesting to see how significant the Client IM is at 37% of the total.

To get a sense of the IM numbers lets do a quick comparison to the Eurex Clearing numbers for Fixed Income. This includes the Bund Futures and Options franchise and has a Total IM of €5.1b. Adding in Eurex Fixed Income/OTC of €0.8b, gives a Eurex Clearing Fixed Income Total IM of €6billion.

Converting the SwapClear total into Euros gives €59b, making the SwapClear IM ten times larger than the Eurex Clearing number.

This illustrates the difference between the amount of risk cleared and outstanding at these CCPs and reflects the difference between Swap portfolios with a huge number of outstanding trades and long term exposure (higher margin) versus Futures with a huge number of daily trades but a lower number of outstanding positions and shorter term exposure (less margin).

Variation Margin (6.6)

A couple of interesting metrics on Variation Margin (VM) at SwapClear:

• Average Total VM paid to the CCP by participants each business day in the quarter was £2.1 billion
• Maximum Total VM paid to the CCP on any given business day over the period was £6.5 billion
• Maximum Aggregate IM call on any given business day over the period was £738 million

So large amounts can move back and forth between SwapClear and its members with the maximum VM move over three times the average move, reflecting large market moves days. Given the volatility we have seen in 2016, I imagine this multiple will be higher when we get numbers for 1Q 2016.

Next looking at the disclosures in the Back testing of initial margin we see:

• Number of times over the past twelve months that margin coverage held against any account fell below the actual marked-to-market exposure of that member account is 263.

So more than one a day.

While that might sound alarming, the next two disclosures quantify the size of the deficits:

• Where breaches of initial margin coverage have occurred, size of uncovered exposure:
  • Peak Size of £141 million
  • Average Size of £4 million

Not much to be concerned about there.

Position, Margin and Default Fund Concentration (18.2, 18.3, 18.4)

What about concentration of members, as we always hear anecdotally that the Swap market is dominated by the five largest dealers, so lets check for SwapClear:

• Percentage of open positions, including both house and client, in aggregate, held by
  • the largest five clearing members, the average is 28% and the peak is 29% over the quarter
  • the largest ten clearing members, the average is 47% and the peak is 49% over the quarter
• Percentage of initial margin, including both house and client, in aggregate, posted by
  • the largest five clearing members, the average is 25% and peak is 27% over the quarter
  • the largest ten clearing members, the average is 38% and peak is 40% over the quarter
• Percentage of participant contributions to the default fund contributed by
  • the largest five clearing members in aggregate is 16%
  • the largest ten clearing members in aggregate is 28%

Meaning the Top 5 concentration is approx. 25% and the Top 10 approx. 40%. Neither is as high as I would have expected even for a total membership of 98 participants. Definitely a good thing in terms of lowering the concentration risk.

It is interesting that the Default Fund contributions for the largest members is lower than their IM percentage; 16% to 25% for the top 5. Often the Default Fund contribution is a direct function of the IM, but in this case it appears to be influenced by additional factors, which reduce the percentage attributed to the largest members.

To understand this, I would need to research LCH.Clearnet’s methodology for determining Default Fund contributions, but that is a task for a different day. Onward.

Client Clearing (19.1, 14.1)

What about Client Clearing? For SwapClear key measures include:

• Number of clients is 11,402! (Surely this is client accounts and not client firms)
• Number of direct members that clear for clients is 50
• Percentage of client transactions attributable to
  • the top five clearing members is an average of 86%
  • the top ten clearing members is an average of 97%

So in this case we do have a very high concentration of client clearing in the top five members.

And for the Client margin requirements:

• 10% is in Individually segregated accounts
• 11% is in Omnibus client only accounts
• 79% is in legally separated but operationally co-mingled (LSOC) accounts

Meaning Client collateral is held secure from Member accounts.

Credit Risk (4.4)

Lets see what we can find under the Credit Risk disclosures.

Lots of disclosures here, so let me just show a few interesting ones for SwapClear:

1. The estimated largest aggregate stress loss (in excess of initial margin) that would be caused by the default of any single participant and its affiliates (including transactions cleared for indirect participants) in extreme but plausible market conditions:
   • Peak day amount in the previous 12 months was £1.1 billion and
   • Mean average was £835 million
2. The number of business days, on which the above amount exceeded actual pre-funded default resources (in excess of initial margin) was Zero.
3. The actual largest aggregate credit exposure (in excess of initial margin) to any single participant and its affiliates (including transactions cleared for indirect participants):
   • Peak day amount in the previous 12 months was £3 million and
   • Mean average over the previous 12 months was £63k.
4. The estimated largest aggregate stress loss (in excess of initial margin) that would be caused by the default of any two participants and their affiliates (including transactions cleared for indirect participants) in extreme but plausible market conditions:
   • Peak day amount in the previous 12 months was £1.98 billion and
   • Mean average was £1.49 billion.
5. Number of business days, on which the above amount exceeded actual pre-funded default resources (in excess of initial margin) was Zero.

The key one here is number 4, which shows that even with the default of the largest two members in extreme market conditions, the largest aggregate loss averages £1.49b and peaks at £1.98b, both less than the Pre-funded Member Default Fund Contributions of £2.66 billion, let alone the additional member commitment of the same amount again. How protected can you get? Sounds safe enough from a Credit Risk perspective to me.

The End

Thats it for today.

I think the public disclosure of these quantitative numbers is great.

It provides transparency to understand the resources of a CCP.

A new insight into quarter-on-quarter changes.

And a comparison between CCPs on a like for like basis.

I plan to cover some of the other European CCPs in future articles.

As well as look at quarterly changes and compare CCPs.