HAL 9000: Just a moment — just a moment — I just picked up a fault in the AET-87 Unit.

Frank Poole: What is it?

HAL 9000: It’s a device for optimising regulatory capital, but that’s not important right now.

David Bowman: What’s the problem, HAL?

HAL 9000: It’s going to go one hundred per cent. failure, within 72 hours.

Poole: Surely, you can’t be serious?

HAL 9000: I am serious. And don’t call me “Shirley”.

Bowman: (sticking to the script) I don’t know what you’re talking about, HAL?

Cue musical introduction

HAL9000: Well, I’ll tell you.

Chorus: He’s going to tell!
He’s going to tell!
He’s going to tell!
He’s going to tell! —

Poole: Stop that! Stop that! No singing!

Carries on for three hours in this vein

Monty Python and the Magnetic Anomaly from Airplane!

Automatic Early Termination — colloquially, “{{{{{1}}}|AET}}”, but not to be confused with “{{{{{1}}}|ATE}}” or “{{{{{1}}}|ETA}}” — is an odd, feared and misunderstood concept buried at the back end of Section {{{{{1}}}|6(a)}} ({{{{{1}}}|Right to Terminate Following Event of Default}}).

Acclimatisation

In a document stuffed with arcanities, {{{{{1}}}|AET}} is especially abstruse, so if you are hitting this article cold then, firstly: what the hell are you doing; and secondly some background reading is in order:

Recommended background reading

Overview

{{{{{1}}}|Automatic Early Termination}} provision is triggered when a party to whom it applies suffers an in-scope {{{{{1}}}|Bankruptcy}} {{{{{1}}}|Event of Default}}. If it is triggered, all outstanding {{{{{1}}}|Transaction}}s are instantly and automatically terminated, without the need for any action by — or even the knowledge of — the {{{{{1}}}|Non-Defaulting Party}}. This usually means instantly, but in one case, it is even quicker than that.

If the {{{{{1}}}|Bankruptcy}} event is the presentation to the court by a creditor of a formal petition seeking the entity’s bankruptcy under Section {{{{{1}}}|5(a)(vii)(4)}} — let us call this a “bankruptcy petition”, some creative warping of lexophysical swaptime is required. We will discuss this at some length and with wistful pedantry, in the premium content section.

In taking things out of the {{{{{1}}}|Non-Defaulting Party}}’s hands, AET subverts the normal order of things under the ISDA Master Agreement. Normally, the {{{{{1}}}|Non-Defaulting Party}} is in control. It may, but need not, call an {{{{{1}}}|Event of Default}} if the circumstances justifying one exist. AET is, well, automatic. It even obliterates the {{{{{1}}}|Non-Defaulting Party}}’s right to waive designation of an Event of Default, since by the time it is in a position to do so, the Event Default has already been declared.

(Could a {{{{{1}}}|NDP}} pre-waive in anticipation? See “Anticipatory waiver?” in the premium section.)

JC’s view is that {{{{{1}}}|Automatic Early Termination}} is a bad solution to an unlikely problem, but since it is embedded in every ISDA on the planet, and remains present in the minds of those who mandate capital calculations, we are stuck with it.

The theory

“Formal bankruptcy is a “phase transition”: the whole “legal context” surrounding a company changes. Erstwhile certainties vanish: normal rules of contract, debt and credit are suspended; in their place arise uncontrollable vagaries. The court appoints an insolvency administrator and invests her with wide, nightmarish discretions to do as she pleases, within reason, to sort out who gets what while ensuring the right thing is done by all the bankrupt’s creditors, customers, employees and, if there is anything left, shareholders. All, therefore, must fall upon her mercy

The phase transition of bankruptcy

Where a Defaulting Party’s bankruptcy regime allows its administrator to suspend its contractual terms or cherry-pick which of its {{{{{1}}}|Transactions}} to honour, it would help the {{{{{1}}}|Non-Defaulting Party}} if the ISDA were to automatically terminate before that phase transition occurred. To be safe, termination should happen at the exact moment — or even an infinitesimal moment before — that bankruptcy regime comes to life.

A bankruptcy regime could affect an innocent counterparty’s rights in at least two ways:

  1. Discretionary termination right itself: Firstly, bankruptcy rules may prevent the {{{{{1}}}|Non-Defaulting Party}} closing out {{{{{1}}}|Transactions}} at all. They may give the administrator the discretion to affirm or avoid individual {{{{{1}}}|Transactions}}. This bigly messes with the fundamental philosophy of the ISDA Master Agreement.
  2. Netting right: Having exercised its early termination right, the “single agreement” operates to net all {{{{{1}}}|Transaction}} exposures down to a single sum. Since a bankruptcy administrator may enforce some contracts and set aside others, that netting right is prejudiced.

History

Automatic Early Termination was introduced, uncredited, in the 1987 ISDA. It was not an election, to be engaged judiciously, when needed against counterparties in jurisdictions vulnerable to bankruptcy shenanigans. It just sat there and applied across the board.

And any {{{{{1}}}|Bankruptcy}} {{{{{1}}}|Event of Default}} — under the 1987 ISDA there were eight — triggered it, against any counterparty in any jurisdiction.[1] As we note elsewhere, the ISDA’s {{{{{1}}}|Bankruptcy}} definition somewhat jumbles the distinct concepts of “bankruptcy” — a formal status involving a legal phase transition, where bankruptcy shenanigans were on the cards — and “insolvency” — an accounting fact with no formal legal status, in which they are not.

By 1992, ISDA’s crack drafting squad™ limited AET to circumstances with a live risk of bankruptcy shenanigans. The “soft” economic insolvency events were excluded. It also does not apply across the board. The 1992 ISDA converted {{{{{1}}}|AET}} into an election that you toggle on or off for each counterparty in Part 1 of the {{{{{1}}}|Schedule}}. For most parties, in most jurisdictions, it stays off.

In 2002 a further refinement was implemented in the definition of “bankruptcy petition”. ISDA’s crack drafting squad™ split Section {{{{{1}}}|5(a)(vii)(4)}} in two: a bankruptcy petition instituted by a regulator was not subject to a grace period. A bankruptcy petition instituted by anyone else — such as a creditor — would only mature into an Event of Default if ordered by the Court, or not otherwise discharged within a 15-day grace period.

If a regulator is taking formal action against you, the game is certainly up. A grace period serves no real purpose. A mere creditor doing so is a fairly ordinary debt-collection tactic: it may not indicate a genuine inability to pay debts, but in that case can be fairly easily discharged. Hence the grace period.

It’s not about risk: it’s about capital

Triago: Herewith, hereinafter and hereinbefore-confirmed:
A custom aperture. Wall-inlaid,
Well-glazed and fringed by lintel stone.
A device to shed upon us light!

Regolamento: Oh, a window?

Triago: Good heavens, No! Not that!
(Whispering) There are ways and means of saying ’t, ser —
Prithee, gird thy verbiage about with care
Lest th’Exchequer’s like for “levies upon transparency”
Untimely drains th’excess from our meagre chancelry—
Catcheth thou the drift?

Regolamento: It is not a window, then? These sound like solid facts?

Nuncle: ’Tis not so much a window
As a means of dodging tax.

Büchstein, Talentdämmerung

Automatic Early Termination is as much to do with managing regulatory capital — in particular, vouching safe close-out netting — as it is about substantive credit risk mitigation. It is only really useful to a regulated institution: one, like a bank or an insurer, that must hold regulatory capital against its ISDA exposures. Basel rules allow a bank to treat its ISDA exposures against a counterparty as “net” only if it can find a law firm that will give an opinion that in all conceivable circumstances including on the counterparty’s bankruptcy, those exposures would net down on close-out.

Most netting opinions do say that. It might take them 300 pages to say it, but they do. In a very small number of cases, the netting opinion will be a bit qualified:

“Well, if your Counterparty goes bankrupt and then you try to close out, it won’t work. Unless you activate that sneaky {{{{{1}}}|AET}} trigger. If you do that, you’ll be fine, because it will magically close out at or just before the phase transition, so you will never have a live ISDA versus a bankrupt counterparty. So, yes: in that case your close-out will always work.”

If you don’t have to hold regulatory capital — as a rule, swap dealers do, end users don’t — you could be forgiven for being a little sanguine: the risk AET addresses that is pretty remote:

  • Your counterparty goes bankrupt without an intervening Failure to Pay and
  • It is significantly net out-of-the-money on a collateralised basis when it does so and
  • Its receiver cherry-picks its in-the-money Transactions and gets a local judgment against you for those and
  • Your home courts enforce the foreign judgment even though the netting is robust in your jurisdiction.

This risk does not seem to have come about yet, in the 43-year history of the global OTC derivatives market.

If you aren’t required to hold regulatory capital, you don’t need netting opinions and you may feel your counterparty’s distant bankruptcy is a risk worth running.

A shade cynical?

This may strike you as a cynical view, but it is not really:

Actual, direct bankruptcy losses are a deep tail risk. Very few of your counterparties will ever go bankrupt, their trajectory towards that forsaken state will usually be well-telegraphed and, your credit guys being the vigilant professionals they are, most likely they will have long-since trimmed lines, managed down positions and closed tanking counterparties out well ahead of their formal bankruptcy. Even if they haven’t, the practical chance that an administrator tries on, let alone succeeds with, bankruptcy shenanigans is low.

By contrast, a regulatory capital charge against that risk is an immediate, present running cost to your business. It applies against every affected counterparty whether or not it eventually goes bankrupt and, on the odds, most of them will not. And the regulatory capital charge for “gross exposure” ISDA is huge. Crazy high.

Few and far between

There are only a few counterparty types in a few jurisdictions where the conditions for {{{{{1}}}|AET}} prevail. There are not many because — let’s be clear, here — {{{{{1}}}|AET}} is a bit of try on: any self-respecting netting-hostile bankruptcy regime would see straight through it. It’s a bit cute, in other words.

A piece of time-travelling contractual magic — deeming an ISDA to have terminated, without anyone’s knowledge or action, the instant before the event that, on expiry of a grace period would later trigger it, to avoid the ambit of discretionary rules designed to ensure fairness and prevent just that kind of preference seems — a bit optimistic?

There is a more benign view, but it is still a bit hopeful: {{{{{1}}}|Automatic Early Termination}} helps buttress what is really a permitted outcome, dispelling residual doubt about ambiguous or untested regulatory provisions that come into force followinging the phase transition to bankruptcy. Sort of a “better be safe than sorry”.

Then there is a pragmatic view. This is blunter: ISDA Master Agreements once were weird innovations that bamboozled and outraged bankruptcy administrators. Now, they are not. Everyone knows what ISDAs are, why they net, and why the single agreement is a sensible plank in the capital structure of the financial system, and why the bankruptcy shenanigans they seek to avoid would not produce a fair result for anyone.

But, still: this is all well and good. But if, unwittingly, {{{{{1}}}|Automatic Early Termination}} now creates practical risk where once it avoided theoretical ones, it still should be a source of concern.

We will talk about that at great length in the premium section.

Why not just switch it on, to be on the safe side?

Master trading agreements are unusual in that upon an {{{{{1}}}|Event of Default}}, there is no guarantee the {{{{{1}}}|Non-Defaulting Party}} will actually be owed any money. If it is net out-of-the-money, it might have to pay.

The last thing a solvent swap counterparty wants is to discover its portfolio of swaps terminated two weeks ago. Indeed, the “flawed asset” provisions of the ISDA Master Agreement are designed precisely to allow it to just suspend performance of Transactions — therefore not make its position any worse — without crystallising its {{{{{1}}}|Transaction}} exposures.

So having {{{{{1}}}|Transaction}}s automatically accelerate is undesirable: one would only choose that if the alternative was catastrophically worse. Unless there is real risk of bankruptcy shenanigans — and for most swap counterparties around the world, there is not — you should not want AET activated, for or against you.

  1. This is just one more reason not to use the 1987 ISDA, if there are any Burmese Junglers still out there looking for a way back to civilisation.