Template:M summ Equity Derivatives 12.7

Section 12.7(c) and the curious question of the multiple Determination Agents

Now, here is a funny thing.

In its ever-unquenched thirst to cater for every conceivable eventuality, however inconceivable, the ’squad devoted themselves in Section 12.7(c)(ii) to the contingency that there might be two Determining Parties appointed to hash out the Cancellation Amount that applies for a Transaction. This, we think, imagines some kind of collective co-calculation agent regime where the parties each have their own Man In Havana making independent calculations with the intention of split whatever difference there may be.

But alas, having had the energy to contemplate this vanishingly remote scenario, our ninja friends didn’t have the diligence left to write out what should happen properly, and as a result, what Section 12.7(c)(ii) tells the co-Determining Parties to do doesn’t make any sense.

In fairness, simply having co-Determining Parties doesn’t make any sense, but that won’t do as an excuse: if you insist on contemplating something stupid, you should at least work it through properly, so stupid parties who fall into the trap of thinking you knew what you were doing when you drafted the option, and who therefore select it, don’t get themselves into bother later, which we regret to say they will, if they select two Determining Parties. Not many do, but it is not unheard of.

But first things first.

Why it makes no sense to have two Determining Parties

Cast your mind back to the reason we have a Determining Party, and not just the Calculation Agent, in the first place. It is specifically for calculating Cancellation Amounts when nixing a trade that has suffered a catastrophic Additional Disruption Event:

...determination of a Cancellation Amount is inextricably related to the hedge and — especially where there is a disrupted market – this is best to be calculated by the one whose problem it is to unwind that hedge: namely, the Hedging Party. In theory (though almost never in practice) the Hedging Party might not be the Calculation Agent.

Now, remember what is going on here: we have a client, going long or short some Equity underlier without actually having to buy it, and a swap dealer, providing that exposure by using its deep connections into the world’s equities and futures markets. Customer, that is to say, and service-provider. The swap dealer has no skin in the game: it has no stake in the performance of the underlier. It will be hedging delta-one, usually by buying (or short-selling) the underlier outright. Other than to the extent that price keeps its customer happy and returning for more business, it is indifferent to the price of the underlier, as long as it can pass it on to its client. It has every incentive to get the best price it can: that is the commercial imperative.

Now, if the Transaction has been disrupted so badly it is to be cancelled, this means is hard to get a price in the underlier, That, in turn, means it is hard to liquidate the hedge. Whose problem is that? The Hedging Party’s. It went out and bought the hedge, in fair times, and now it has to sell it, while times are foul.

To be clear this is no idle intellectual speculation: liquidating a hedge is not simply looking at some fantastical model dreamt up by the most delusional quant on the trading floor, arriving at some mad price that will ruin the client for nothing. No. The Hedging Party is actually long the risk. It will have to crystallise real liability, using money from its own pocket to flatten out that risk. The amount it pays away is exactly what it will expect its client to suffer. That is the deal.

Running a synthetic equity business is mainly boring, often fraught, and on the odd occasion terrifying job. Calculating Cancellation Amounts upon market disruption counts, at least, as fraught. If the Hedging Party can’t match its hedge executions with its client pricing, it will lose money. It occasionally escapes the buyside community, but this is not the idea. The dealer is does not expect to profit, or lose, from movements in the underlier: the point is to pass those onto the client, who has specifically signed up for them.

Accordingly, the dealer will be most unamused if a client asks it to consider an alternative price someone else has come up with to value its own hedge liquidation. This is like saying, to a football fan, “look, I know Crystal Palace lost to Scunthorpe in extra time at the weekend, but my mate is a football expert, and he says Palace were dead unlucky, hit the crossbar a couple of times, and that Scunthorpe goal should have been called off-side, so why don’t we call this 4:0 to Palace?”

The dealer will, tersely, say, “Look, I know you have a great relationship with Wickliffe Hampton and everything, but I could not care a row of buttons where they sees the notional value of my hedge. They don’t have to sell it. I do. We are where we are. The price I see is X. Now if Wickliffe Hampton is prepared to make me a firm bid at Y to buy my actual hedge, from me, then I am listening. Otherwise, it is X.”

But it is Determining Party, not Hedging Party

The truly stubborn buy-side agitant will persist: “if this is really about who is the actual Hedging Party, why doesn’t it just say “Hedging Party”? Why do we need a new definition of Determining Party? It isn’t used for anything else. Your beloved ’squad doesn’t make these things up for the sake of it, after all.”

The galling fact is that we cannot explain this. Our best guess is yet more unnecessary over-elaboration on ISDA’s crack drafting squad™’s part — it may not have made a difference (and might have avoided this very article) had the ’squad not created an extra label, but we are where we are. It is possible, we suppose that the dealer hedges with a swap, and the actual price discovery happens away from the Hedging Party along a chain somewhere). But that isn’t a very satisfactory answer.

Why, if you must insist on having two Determining Parties, this clause doesn’t work

This is the clincher. the co-Determining Party language doesn’t work. Say we have, with heavy heart, acquiesced, and agreed two Determining Parties. Let’s further say there has been an Extraordinary Event, such that a Cancellation Amount is required. We reach for Section 12.7(c). Each Determining Party does its thing: the swap dealer’s sees a price of 102, and the client’s agent sees 104.

Logic should say the number we are after is 103, yes? The average. But, thought easily could have said that, that is not what Section 12.7(c) does say.

“...an amount will be payable equal to one-half of the difference between the Cancellation Amount of the party with the higher Cancellation Amount (“X”) and the Cancellation Amount of the party with the lower Cancellation Amount (“Y”) and Y shall pay it to X.”

The difference between X and Y (104 - 102) is two. Half of that difference is one. If it is meant to yield a consensus single Cancellation Amount, this, to put not too fine a point on it, is wildly, obviously, and patently wrong.

“Equity vs Equity” swaps, maybe? Don’t think so.

What could be going on? One thought we had — and it is a feeble, half-thought, so pay it little mind other than to dismiss it — is that this is meant to address a Transaction where both parties are hedging separate equity risks: rather than the usual equity swap, which pays an Equity Amount return against a Floating Amount return, the Transaction is structured as two offsetting Equity Amounts. This would at least justify there being two Determining Parties, and it would also justify them using the difference between the values, rather than their average — but not half the difference between the values. And even here, the concept doesn’t work: a Cancellation Amount is crafted the total termination amount for the whole Transaction, not just a valuation of the Equity Amount leg. Nor are such swaps otherwise contemplated in the 2002 ISDA Equity Derivatives Definitions, seeing as there is a much easier way of achieving long exposure to one underlier and short exposure top another: just enter two vanilla Transactions, one long and one short.