Template:M intro isda Party A and Party B

In this episode of the JC’s series of unfeasibly deep explorations of superficially odd things in the ISDA metaverse, consider the bilateral nature of the ISDA Master Agreement and its curious designators: “Party A” and “Party B”, and that curious descriptor of both of them: “counterparty”.

These set the ISDA apart; give it a sort of otherworldly aloofness; a sense almost of social justice. Other banking and broking transactions use labels which help you orient who, in the power structure, is who: a loan has a “Lender” (always the bank) and “Borrower” always the punter. A brokerage has “Broker” (master) and “Customer” (servant).

But not the ISDA Master Agreement. From the outside its framers — the First Men — opted for the more gnomic, interchangeable “Party A” and “Party B”.

Why? Well, we learn it from our first encounter of an ISDA Schedule. bilaterality.

Bilaterality

A belief in even-handedness gripped the ones whose deep magic forged the runes from which the First Swap was born.

For most finance contracts imply some sort of dominance and subservience: a large institutional “have” indulging a small commercial “have-not” with debt finance for the privilege of which the larger “have” extracts excruciating covenants and enjoys a preferred place in the queue for repayment among the have-not’s many scrapping creditors.

But swaps, as the First Men saw them, are not like that.

“A swap contract,” they intoned, “is an exchange among peers. It is an equal-opportunity sort of thing; Biblically righteous in that, under its awnings, one be neither lender nor borrower, but an honest rival for the favour of the Lady Fortune, however capricious may she be.

“We are equals. Rivals. Counterparties”. Covenants, privileges of credit support and so on may flow either way. They may flow both ways. In our time of regulatory margin, they usually do.

And, to be sure, swaps are different from loans and brokerage arrangements. They start off “at market” where all is square. Either party may be long or short, fixed or floating. At the moment the trade is struck, the world infused with glorious possibility. One fellow’s fortunes may rise or fall relative to the other’s and, as a result, she may owe (“out-of-the-money”) or be owed (“in-the-money”). And swaps, too, are professional instruments. Moms and pops, Belgian dentists and the like may buy bonds, but they din’t, and never have, entered ISDA Master Agreements.

Now the ISDA Master Agreement itself never uses the terms “Party A” or “Party B”. Being genuinely bilateral, it never has to. The labels are arbitrary assignations that apply at trade level. Thus, they only appear in the Schedule and in Confirmations, to be clear who is who on a given trade: who is paying the fixed rate and who the floating; which thresholds, maxima, minima, covenants, details, agents and terms apply to which counterparty. This much is necessarily different. Nothing beyond: the ISDA Master Agreement assumes you already know who is who, having agreed it in the Schedule.

So we agree: for this relationship we will call you “Party B”, and me “Party A”.

These colourless and generic terms hark from a time where, we presume, the idea of “find and replace all” in an electronic document seemed some kind of devilish black magic. Some kind of Tipp-Ex-denying subterfuge.

But anyway. These generic labels still lead to practical difficulties. A dealer with ten thousand counterparties in its portfolio wants to be “Party A” every time, just for peace of mind and literary continuity when perusing its collection of Schedules, as we know dealers on occasion are minded to do. If, here and there, a dealer must be “Party B”, this can lead to anxious moments should one misread such a Schedule and infer its infinite IM Threshold applies to the other guy, when really, as it ought, it applies to you. Frights like this are, in their way, quite energising. You quickly get over them when you realise it is your error of construal, not the negotiator’s of articulation.

Less energising are actual errors: as a group, negotiators are redoubtable, admirable creatures but, like all of us fallible and prone to oversight: they may, by lowly force of habit, forget to invert the “Party” labels when inserting the boilerplate PPF Event rider for that one time in a thousand when the firm is not “Party A”. It is easily done, and just the sort of thing a four-eyes check will also miss: If it does, no-one will never know — unless and until it is too late.

Is it bilateral though?

But there is a better objection: for all our automatic protestations to the contrary, the ISDA is not really a bilateral contract, and it is often financing contract, in economic effect even if not in formal structure. Where there is a customer gaining exposure to a risk and a dealer providing delta-hedged exposure to that risk, a swap is a sort of “synthetic loan”.

We should not let ourselves forget: beyond the cramped star system of inter-dealer relationships, there is a boundless universe where one party is a “dealer” and the other a “customer”. This is the great preponderance of all ISDA arrangements. The customer and a dealer roles are different. The difference does not depend on who is “long” and who “short”, or who is the fixed rate payer and who is the floating rate payer. Hence the expressions “sell-side” — the dealers, who sell exposure — and “buy-side” — their customers, who buy it.

For the buy-side, the object of trading a swap, or making any investment, is to change its market exposure: to get into a positions it did not have before. This sounds obvious. But, being a bilateral contract, its corollary ought to be that the sell side ought to be changing its position, too. But it is not. A dealer “sells” a swap to earn a commission without changing its market exposure.

Now, a swap is a principal obligation, so being a party to one necessarily changes the dealer’s market exposure, so the dealer must then “delta hedge” that position away, by taking on an equal and offsetting position in the same asset somewhere else: by buying the underlying asset, or matching off its exposure under “long” position against another exposure under a “short” position in the same underlier.

There are plenty of ways to delta hedge, but the basic economic principle is this: on the asset side of the swap the equity leg) the dealer has not changed its market position. It has not, over all, made an investment. It has not borrowed anything.

Provided the dealer knows what it is about, its main risk in running a swap portfolio not market risk — there should not be any — but counterparty risk. Should a counterparty fail, suddenly the dealer might have a lot of market risk. Hence, having adequate collateral from its customers, to cover the eventuality that they should fail, is very important.

Swaps are usually synthetic loans

But how does this make a swap into a synthetic loan? It is best illustrated by comparing a swap with an actual loan. Take this scenario:

Hackthorn Capital Partners owns USD10m of AUM. It wishes to buy USD10m of Cryptöagle. It can either: sell its existing AUM and use the proceeds to buy Cryptöagle, or keep its existing portfolio and borrow USD10m.

Sale
If it sells its existing portfolio outright, the position is as follows:

Sold: USD10m.

Borrowed: Zero.

Amount owed: Zero.

Bought: 10m Cryptöagle @ USD1 per share.

Amount due: total return on 10m Cryptöagle.

Loan
If it keeps its existing portfolio and borrows, the position is as follows:

Sold: Zero.

Borrowed: USD10m.

Amount owed: floating rate on USD10m.

Bought 10m of Cryptöagle @ USD1 per share.

Amount due: total return on 10m existing portfolio and 10m Cryptöagle.

Note the cashflows in the loan scenario:

During loan, Hackthorn pays floating rate on USD10m and is exposed to the market price of Cryptöagle.

On termination of the loan Hackthorn sells Cryptöagle. If sale proceeds exceed loan repayment, Hackthorn repays the loan and keeps the difference. If sale proceeds are less than loan repayment, Hackthorn must finance the shortfall from its existing portfolio, thereby booking a loss.
Therefore, Hackthorn’s net exposure is USD10m - Cryptöagle spot price.

These are the same cash flows you would expect under a delta-one equity derivative:

During swap, Hackthorn pays floating rate on USD10m and dealer pays total return on Cryptöagle. On termination, if the swap termination amount is negative, Hackthorn pays it to dealer. If it is positive, dealer pays Hackthorn.
The swap termination payment is USD10m - Cryptöagle spot price.

Like a loan, the equity swap gives Hackthorn exposure to Cryptöagle whilst keeping its existing portfolio, which Hackthorn uses to fund cashflows on its new capital asset. This is a form of leverage. The floating rate Hackthorn pays is implied funding. The dealer will only accept this if it is satisfied Hackthorn has enough capital to finance its swap payments and settle any differences at termination. This is the same risk calculation a bank lender would make.

But, hang on: this is a bilateral swap arrangement, so isn’t the same also true of the dealer? Isn’t the dealer paying a rate to get exposure to the synthetic cashflow of an asset — the floating rate — in the same way, so is, in a sense “borrowing” by paying its total return? Is not a “short” equity derivative, for a dealer, exactly the same as a “long” equity derivative for a customer?

Generally not, because in providing these swap exposures to its customers, the dealer is not changing its own market position. It delta-hedges. At the same moment it puts on a swap, it executes an offsetting hedge. The dealer’s net position on its derivative book will generally be flat. You don’t need to borrow money to take no position.

On the case for one-way margin

In recent years — ironically, just as the “dealer” vs “customer” dynamic has become more pronounced — the global regulatory-industrial complex, still fighting last decade’s war, has forged rules which overlook this. Notably, the coordinated worldwide approach to bilateral regulatory margin. As swap positions move in and out of the market, swap counterparties must post each other the cash value of the net market movements each day. This is a little like closing positions out at the end of each day and settling up, with a key difference: you don’t close out your positions. The parties are still on risk.

Well — one of them is. The customer: the one who initiated the trade, to put itself into a market position of some sort. The other party, remember, is delta-hedged. It didn’t initiate the trade, but accommodated it, on the precise grounds that its market position would not change, and its credit position, against the customer, would be satisfactory.

Requiring margin from a customer who is net out-of-the-money makes sense: if the customer fails, the dealer’s corresponding hedges are defeated and it will be left with an open market exposure to all of the customer’s positions. This is exactly what the dealer wants to avoid. It is meant to be flat. So, daily variation margin mitigates the dealer’s market risk to date; initial margin covers it for the forward market risk while it closes out its hedge portfolio against the defaulted client. As long as the dealer is covered, market disruption is minimised, and the dealer’s own solvency — which, due to its interconnection with the rest of the market may well present a systemic risk — is not threatened.

Requiring a dealer to post margin to its customer against the customer’s net in-the-money position makes no sense whatsoever:

Firstly, dealers are delta-hedged against every customer. They are not nursing losing positions against profitable clients. Dealers are not “the other side of the trade”: generally, they will be happy when their customers realise profits on their swaps. They do not lose money: they take another commission on the unwind, and are standing by to accept new business.

Secondly, dealers and banks are capitalised and regulated for systemic risk. There are already constraints on how they must operate, and how much capital they must hold against the contingency of portfolio losses. This capital is, in part, a function of the risk the banks have to their customers. That risk is greatly exaggerated if their clients have to post cash reflecting paper gains their clients have made but not yet realised.

For several reasons:

Banks are Banks are not losing value on the customer positions: as per the above, they are delta-hedged. Banks have no market exposure unless their clients go bust.

Banks generally go bust because their clients go bust. They don’t go bust by themselves—I know, I know, Silicon Valley Bank did, but it is an honourable exception that proves the rule.

If we are worried about bank solvency, then forcing the bank to cash settle unrealised gains on derivatives portfolios is a bad idea.

“settling to market” every day with an important distinction: you don0147

Banks are independently capital regulated for solvency.

Swap counterparties are sophisticated professionals with the tools and resources to monitor credit exposure to their brokers. (We take it that the financial weapons of mass destruction that these sophisticates truck in require more expertise than does weighing up the likely failure of a regulated financial institution).

It is a much better discipline for sophisticated counterparts to manage their credit exposure — spread it around, so to speak, than to require their banks to send hard cash out the door as collateral to clients with outsized exposures.

Every dollar these banks pay away reduces the capital buffer the bank has available for everyone else. It also provides the customer with free money on an unrealised mark-to-market position. This is like paying out while the roulette wheel is still spinning, in the expectation of where the ball might land. That is a loan: if the customer doubles down and loses, you don’t get your money back.

Daily mark to market moves are mainly noise. Yet this is what we collateralise. The signal emerges over a prolonged duration. Over the short run posted collateral can, as we know a system effect: if I double down on an illiquid position, it will tend to rise, and I will get more margin, and — this is the story of Archegos.

The increased systemic exposure of banks failing — which is what the margin regs were designed to address — is not caused by the banks themselves, but by their client exposures. Client exposures in turn are a function of client failures, which are in turn a function of leverage

Their failure shouldn't, typically, be a systemic risk unless their unusually size, interconnectedness or unintended system effects make them systematically important, in which case they should be regulated if they are systemically important, and made to hold capital, and (b) they have the market position and bargaining power to negotiate margin terms.