Template:Csa Eligible Collateral summ

{{{{{1}}}|Eligible Collateral}} — with or without the (VM) suffix, and finding it tremendously irritating as he does, JC will largely go without — is what is, or is not, acceptable to post under a CSA. In the olden days, this was all a bit fraught and accounted for much of the practical complication of the CSA, and which (in the JC’s eyes) made it a sort of primordial smart contract, long before there were such things as smart contracts. It looks like you can choose what you like, right, but CFTC rules in practice restrict it to cash in certain currencies. This is broadly analogous to what happens in Europe under EMIR and the other major regulatory regimes imposing margin requirements.

First, a terminology check. Bless them the NY CSA and the English law CSA use different labels for essentially the same things: Eligible Credit support (British) is the same idea as Eligible Collateral (American). This appears to be a matter of pure preference and you would like to think the ’squads could have duked it out in the mid-Atlantic in 1995 to something but WE ARE WHERE WE ARE.

In any case, an ancient CSA governs a deterministic, but nonetheless highly complicated flow of securities and currencies back and forth between counterparties daily, based on the {{{{{1}}}|Exposure}} of their portfolios. (Modern VM CSAs are simpler, as they are cash only).

In the ancient CSAs this is all very contractual and cash and a variety of types of corporate and government bonds might be acceptable; in the modern CSAs even though it looks like you can choose what you like, right, regulators’ rules, market conventions mean in practice variation margin is limited to cash in certain currencies. This makes a lot of the erstwhile complication and Ballschmerzen of CSA operations goes away.

Oh, go on: let’s have a table:

Comparison of Eligible Collateral Types Across CSAs

CSA Year	Margin	Gov. Law	Holder	Eligible Collateral
1994	IM and VM	NY	Other party	Cash, Government Securities, Investment Grade Corporate Debt
1995	IM and VM	English	Other party	Cash, Government Securities, Investment Grade Corporate Debt, Equities
2016	Reg VM and optional IM	NY	Other party	Cash only (in practice)
2016	Reg VM and optional IM	English	Other party	Cash only (in practice)
2018	Reg IM only	NY	Custodian	Cash, government bonds, corporate bonds
2018	Reg IM only	English	Custodian	Cash, government bonds, corporate bonds

You may wonder why regulatory VM is cash and Regulatory IM is usually securities and why you have to post Reg IM to a third party — if you are asking that, then for a half a pint a week you can subscribe and find out.

In any case, who has to pay whom on any day, and what, is derivative — I know, right — of a load of extraneous factors besides that {{{{{1}}}|Exposure}} as a bare number: the “shape” of the derivatives book, the state of the market, the composition and market valuations of the {{{{{1}}}Credit Support Balance}} already posted. Even a small firm will have multiple brokerage counterparties; a swap dealer will have hundreds of thousands. There is simply no way of manually monitoring and accurately performing an arrangement as complicated as a Credit Support Annex. Only computers can do it. Hence, it is rather like a smart contract, and not just superficially, but practically: the “contract” as a meaningful thing is the operational reality, not the document. It lives in the configurations the parties’ operations teams punch into their respective collateral management applications.

If a number is mis-punched, or the application does not or cannot reflect what is stipulated on paper in the Credit Support Annex, no one will ever know. At least, not until it is too late.

If your counterparty is a smoking ruin and you find out the {{{{{1}}}Credit Support Balance}} you are holding comprises of illiquid pref shares listed on the Manila bourse, while the CSA prescribes only G7 on-the-run government securities, then guess what.

Initial margin

For reasons best known to themselves there are curious rules in the US relating to initial margin which lead counterparties to use different ISDA CSA forms on different occasions. The details and variegation at play here — are you an SBSD counterparty? Are you subject to EMIR too? Is the AANA threshold met? — are all too ghastly for the JC to presently get into, other than to say the tiny peripheral benefits you get from gaming the fractal edges of the system of rules are most likely not worth it and it is probably better to have a single, maximally conservative, theory of the game: all that being well and good, of course, if your counterparties have a different view of the world, or a different idea what is the maximally conservative theory of the game.

This is what happens if you try to standardise and centralise a market that was designed to be unstandardised and decentralised, but still.

The dark side of margin

On the subject of pained exegeses which no-one will heed, the JC also has an impassioned essay about why bilateral variation margin may have a destabilising effect on the broader financial system, potentially weakening swap dealers’ liquidity and risk positions — a risk which was admirably demonstrated during the Archegos fiasco.

But no-one listens to the JC, so you should treat this as extended universe fan fiction only.