Template:M intro isda ISDA purpose

In which the into the point of an ISDA Master Agreement. What it does, why you need one, and why you can’t just crack on and trade swaps without one.

This might seem like pocket-calculator stuff to you, my seasoned veterans, but it never hurts to stop and ponder the ostensibly bleeding obvious. Going back to basics is bracing for the spirit.

The JC encountered his first Aïessdiyé a good thirty years ago now — shout out to GFF, still at it, down in the southern wilds — and is still discovering new things about it every week.

On becoming a shibboleth

Through bad habits and inattention, humans tend to work around the easances we once made to our “built environment”.

So it is with the ISDA Master Agreement. What started as the shortest route to market can, through acquiescent disregard, become a shibboleth: a hindrance on the road to transaction.

Once precisely an easance — an artefact for quickly tidying up and dispensing with formalities it would be laborious to repeat for every trade — the ISDA became a mountain of its own. Sure, you only need to climb it, from the bottom, once — but that has become a three-month operation. Nor do you scale an ISDA master agreement the way the way Alex Honnold scales El Capitan, brave and alone, an aeronaut of the spirit. You must take the entire modernist machinery of your institution with you. KYC. AML. Compliance. Credit. The docs team. And, of course, dear old Legal.

Once, the aggravation of a “long-form confirmation” was the mischief the ISDA should to solve. Some miss the good old days. Where bank Legal departments have not legislated outright against them — most have, long since — the temptation now is to ask, “Must we have an ISDA? Would not a long-form confirmation do?

There are other good reasons for a master agreement, as we will see, but none necessitates all the bureaucratic machinery that has grown around the ISDA. This is how the military-industrial complex of agency operates: it shapeshifts to create work to occupy the available rent.

The three aims of an ISDA

The ISDA Master Agreement is a framework under which two “counterparties” can transact over-the-counter derivatives — mainly, but not only, swaps. Besides its original appeal as an easance, the ISDA has three main aims: it is a relationship contract; a credit risk management tool, and — for those who need it — a capital optimisation tool.

Relationship contract

A relationship contract, yesterday.

Firstly, the Master Agreement is a relationship contract: an agreement sealing a pact of amity and good dealing between two strangers in the jungle. It is, of sorts, a peace treaty: it establishes basic terms between the parties upon which they may deal, reciting their aspirations, outlining their cultural idiosyncrasies and behavioural red lines, and dealing with housekeeping matters like contact details, account numbers and authorised agents, and generally gathering up all the manifold dreary details needed to ease the experience of transacting with each other.

The Master Agreement does not itself create any obligations or liabilities. It does not commit anyone to any Transaction. While no Transactions are on foot, the ISDA doesn’t do anything: it just provides an architecture: walls within which parties may safely play; a roof beneath which they may comfortably dance if they both feel like it. If they don’t, no one says they have to. Therefore, curiously, it does not provide at all for termination without fault on notice. You can’t, unilaterally, terminate an ISDA at all. Even total close-out doesn’t terminate the master agreement.

If your relationship ceases — should your pact break down — the ISDA just lies there, fallow: a seed in the desert awaiting rain. This happens to be the premise of Muriel Repartee’s Z-Grade Fi-Fi schlock horror ISDA: Dawn of the Dead.

In that an ISDA is painful to put in place — if it only takes a couple of months you have done well — it is also a commitment signal. It shows you care enough to engage in the painstaking process of working up “strong docs”.

If, once the ISDA is inked, counterparties do decide to dance, they agree economic terms of a Transaction and sign a Confirmation that sets out those terms and inherits the remaining terms of their ISDA.

Risk management

Secondly, once the parties have decided to dance, the ISDA is a tail risk management tool. Most tail risks of trading derivatives, though not all, are credit-related.

The ISDA lets each manage and reduce its credit exposure to the other party as a result of all this derivatives trading. These include Credit Support and Close-out rights.

Credit Support may comprise margin “posted” by the counterparty against its exposure, or guarantees, keep-wells, letters of credit and so on provided by third parties on its behalf.

Events of Default and Termination Events entitle you to close out Transactions early, should your counterparty not perform (or should its creditworthiness deteriorate in oblique ways your credit department believes increase its risk of not performing). Most of these events address credit deterioration, but not all: some deal with other externalities — change in law, force majeure, adverse tax — that don’t directly affect either party’s credit position.

Expected events and tail events

We can, in any case, distinguish between “expected events” and “tail events”.

“Expected events” are welcome: we assume these explicitly, by entering the Transaction. They are the economic events that may happen to the instruments underlying our Transaction. Reference Entity Credit Events, rate rises, rate falls, option triggers and so on. If the underliers do not behave as we hoped, we have no complaint: that was the bargain we struck.

Expected events tend to be particularised, delimited and finely detailed. By necessity: they directly affect what you pay or receive under the Transaction, so their articulation must be exact, and room for debate minimal.

Oddly, documentation of expected risks in Confirmations is generally left to Muggles in the trading and operations departments, not Legal. Then again, conventions for trade terms been tested to destruction over thirty years. The rate of outright dispute on the basic meaning of trade terms is generally low, and resolution is typically quick and pragmatic. Lessons are learned: trade terms winnow themselves down to the genuinely critical over time. The drafting self-improves in that evolutionary process.

That Darwinian effect is far less pronounced for “tail events”. These are the externalities: things you don’t expect, but are resigned to, that might get in the way of you enjoying the financial risk and reward of the expected events.

These tend to be easy to foresee in general but impossible in particular, precisely because they are rare, unwanted and, by nature, present themselves when and where no-one is looking: if your counterparty blows up or is nationalised, embargoed or sanctioned. If the law changes, making the Transaction illegal, difficult to perform or less valuable. If the Great King of Terror descends from the skies in a flaming chariot, etc.

Articulation of tail risks is usually done by Legal or — under Legal’s supervision — the negotiation team. The language is baroque, sweeping and infrequently tested. When it is tested, by the courts, after times of crisis — it often turns out not to work as expected.

Division of labour

In any case, there is a functional division of labour between the Master Agreement, under which you minimise and mitigate unlikely risks in general, and the Confirmation, under which you assume and allocate likely risks in particular.

So: the Confirmation deals with what you think will happen and the Master Agreement deals with what you think won’t. The Confirmation is GPS navigation; the Master Agreement is seatbelts, airbags and those neat inflatable slides that turn into life-rafts when a plane crash-lands on water: something you’re glad you have, but hope you never have to use, even though a part of you feels it might be quite fun to have a go.

We are loss-averse: we spend more time and resources than is rational preparing for apocalyptic risks. The dividend of all this conceptual haggling, if done well, is a mythical contractual utopia, beloved of senior credit officers but poorly understood by anyone else, of “strong docs”. Anyway, I digress.

Capital optimisation

Thirdly, the ISDA is carefully designed to yield a specific regulatory capital treatment for regulated financial institutions. Regulatory capital is a big topic, well beyond the scope of this article.

Ninja point: it may look like it, but capital optimisation and credit risk management are not the same. Quite the converse: capital management addresses the period until a counterparty blows up; credit risk management addresses what happens afterwards. Capital rules define the amount of spare cash — “capital” — a dealer must keep uninvested to ride out the losses it suffers upon a counterparty default.

The tools of capital management, therefore, are designed to make that amount as small as possible. They address an “expected event”, the event being “the amount of money the bank must hold as capital each day”. They don’t address the tail event that the bank holds capital against itself. They assume it will happen.

The capital calculation has a daily direct impact on the bank’s capital position, its leverage ratio, therefore how much capital the bank can put at risk. In that very narrow sense, it is a “cost-of-doing-business” management tool, not, strictly, a risk management tool.

Close-out netting

The principal tool for managing the capital cost of a swap master agreement is close-out netting.

Derivatives are odd contracts. They are inherently levered and extremely volatile. They have large notional values, but, usually, lower mark-to-market values. An “at-market” swap starts with zero exposure either way and can fluctuate in either direction.

Compare this with a traditional loan, which starts with an exposure equal to its principal amount — the lender goes “in-the-money” for the full principal size of the instrument, and borrower “out-of-the-money” — and the “mark-to-market value” of the loan — I know, this terminology is wrong in so many ways — then fluctuates narrowly around the amount borrowed, to account for accrued interest, changing interest rates, and the borrower’s changing credit profile, until it is all repaid, in one go, at maturity. There is no scenario in which the lender owes the borrower. The loan covenants reflect this.

Given that the “amount originally borrowed” under a swap is, nominally, zero, the raw market exposure of a portfolio of swaps can be huge compared with that original capital commitment. The total of all your out-of-the-money positions, which you can assume you must perform, versus all your in-the-money positions, for which you will rank as an unsecured creditor.

This is the beauty of the Single Agreement: if you can offset your out-of-the-money positions against your in-the-money positions, and be an unsecured creditor only for the difference between them things look a lot rosier from a regulatory capital perspective. Indeed, where variation margin is exchanged, the parties’ net mark-to-market exposure resets to zero every day.

The regulatory “standard of proof” for “netting” exposures in this way is, however, huge: banks must obtain and keep up-to-date a battery of gruesomely-detailed external legal opinions — one for each counterparty type in each jurisdiction that the bank trades against, for each type of master agreement — that the netting contract actually will — not just should — work in all relevant jurisdictions: the bank’s, the counterparties, its branches, and the location of any assets. We have more to say about netting opinions elsewhere.

For now, suffice to say the degree of certainty each netting opinion needs to express is extremely high. The risk of netting failure may have been material when swaps were an exotic new technology in 1987, poorly understood beyond a narrow range of specialists in London and New York. But things have changed. Swaps nowadays are well understood. The risk of “netting failure” in most jurisdictions is minimal. (To be sure, swaps are still risky instruments — “financial weapons of mass destruction”, as Warren Buffet has it — but this has little to do with netting.)

Credit versus capital treatment?

No such opinions are needed to set credit lines, however. So, is there room for an arbitrage between credit and capital treatment then? While they address different aspects of a trading losses, both are predicated on the same event: the counterparty’s failure when, overall, it owes the bank money. (Credit addresses how much money the bank stands to lose; capital how much of a reserve the bank should hold back to cover for it.)

There was once a time when the credit team might take a more lenient view for credit risk management purposes than the treasury team could for regulatory capital purposes. “I might not, technically, be able to recognise netting from a capital perspective, but from a credit perspective, I am prepared to take a view and assume it will be okay.” This is an argument with some abstract appeal in before a financial crisis: it doesn’t hold up quite so well when you are being grilled about it by congress in the aftermath.

As the global financial crisis wore on, this sort of cavalier “IBGYBG” attitude gave way. (It probably didn’t make a lot of practical difference, at the end of the day: you can make your credit line as big as you like, but if you have to gross your potential exposures for capital purposes when pricing your trades you won’t be competitive in the market in the first place).

Conclusion

Of these three purposes, the one that occupies the most attention during the negotiation process is credit risk management. Achieving “net” treatment is in both parties’ interests, and the mechanism for achieving it is inviolate — the famous “Single Agreement” provision at Section 1(c) of the ISDA, sometimes the definition of Bankruptcy and for Germanic types, designating Automatic Early Termination. Likewise, no-one haggles too hard about your address for notices, who your process agent is. whether or not you are a multi-branch party, or who does what when it comes to portfolio reconciliation and dispute resolution on trade reporting.

But when it comes to Additional Termination Events, expect a firefight.