Template:Csa Events of Default summ: Difference between revisions

From The Jolly Contrarian
Jump to navigation Jump to search
Created page with "{{subst:Csa grace periods}}"
 
No edit summary
 
(13 intermediate revisions by the same user not shown)
Line 1: Line 1:
Those of ISDA credit support arrangements that are standalone {{isdaprov|Credit Support Document}}s<ref>In other words the {{nycsa}}, {{nyvmcsa}} and the {{imcsd}}, but ''not'' the {{csa} and the {{vmcsa}}.</ref> bake in a two {{nyvmcsaprov|Local Business Day}} [[grace period]]. This is one Local Business Day ''longer'' than the standard one LBD [[grace period]] baked into the {{2002ma}}, and one Local Business Day ''shorter'' than the standard 3 LBD [[grace period]] baked into the {{1992ma}}. Which is nice.  
====Grace periods====
Those of ISDA credit support arrangements that are standalone {{isdaprov|Credit Support Document}}s the {{nycsa}}, {{nyvmcsa}} and the {{imcsd}}, but ''not'' the {{csa}} and the {{vmcsa}} bake in their own two {{{{{1}}}|Local Business Day}} [[grace period]] into their own customised Events of Default.  


The English law CSAs, by the way, don’t have this problem, since they are deemed to be {{isdaprov|Transaction}}s under the {{isdama}} itself, and therefore inherit the {{isdama}}’s Section {{isdaprov|5(a)(i)}} [[grace period]]s. Which, you would have thought, ''has'' to be the more preferable arrangement. But anyway, Americans (and those of you Brits with {{imcsd}}s, the same goes for you, thanks to Section {{imcsdprov|7}} I am afraid) '''your [[grace period]]s might well not match'''.
The {{1995csa}} and {{2016csa}}, being Transactions, don’t need to and inherit the [[grace period]]s for the {{isdaprov|Events of Default}} in the {{isdama}}.
 
That can, at the limit, create a dissonance for Americans and IM posters, as your CSA grace periods do not automatically track your ISDA ones. Two {{{{{1}}}|Local Business Day}}s is an {{isdaprov|LBD}} ''longer'' than the standard one LBD [[grace period]] in the {{2002ma}}, and an {{isdaprov|LBD}} ''shorter'' than the standard 3 {{isdaprov|LBD}} [[grace period]] baked into the {{1992ma}}.
 
Which is nice. This is not a problem for the English law Credit Support Annexes that are {{isdaprov|Transactions}}.
 
=====Including “comprehensive” Termination Events=====
Consider expanding of the {{{{{1}}}|Default}} provision to include “{{isdaprov|Termination Event}}s where all {{isdaprov|Transactions}} are {{isdaprov|Affected Transaction}}s”. This is as per Section 3.2 of the 2001 ISDA Margin Provisions which recommend that Paragraph {{{{{1}}}|6}} of the [[CSA]] should apply where all {{isdaprov|Transactions}} are closed out following an {{isdaprov|Event of Default}} or “{{{{{1}}}|Specified Condition}}” — the latter of which is defined to include the {{isdaprov|Termination Events}} listed under the {{isdama}}.  All {{isdaprov|Transaction}}s would likely be {{isdaprov|Affected Transactions}} should a {{isdaprov|Credit Event Upon Merger}} or {{isdaprov|Additional Termination Event}} occur.
 
The {{vmcsaprov|default}} paragraph explains how you value the {{vmcsa}} itself — being a {{isdaprov|Transaction}} in its own right, of course — when closing out an {{isdama}}. The basic gist is that you treat the {{vmcsaprov|Credit Support Balance}} as of the {{isdaprov|Early Termination Date}} — being the total value of the {{vmcsaprov|Credit Support}} you have ponied up at any time — as an {{isdaprov|Unpaid Amount}}, rather than treating is as a contingent return obligation, the present value of which would go into the {{isdaprov|Close-Out Amount}} (or {{isdaprov|Loss}}, or {{isdaprov|Market Quotation}}, if you still labour under an antediluvian {{1992ma}}).
 
====Why do {{ttcsa}}s reference “Unpaid Amounts”?====
A {{ttcsa}} is technically a {{isdaprov|Transaction}} under the {{isdama}} in its own right — that is deep [[ISDA]] lore — but it is still a ''weird'' {{isdaprov|Transaction}}, and the standard “[[replacement cost]]” of valuation method doesn’t work brilliantly: rather than having defined payments upfront, each of which can be valued and discounted back to now to reveal a [[present value]], payment obligations under a {{ttcsa}} depend on the aggregate discounted cashflows of all the ''other'' {{isdaprov|Transaction}}s under the {{isdama}}. So good luck determining, in the abstract, the [[replacement cost]] of something like that.
 
The good news is you don’t have to: the {{{{{1}}}|Credit Support Balance}} isn’t calculated by reference to its own discounted future cashflows: rather, it is just the inverse of the aggregate present value of all the other {{isdaprov|Transactions}} under the {{isdama}}. So the “[[replacement cost]]” on any day is just the prevailing value of the {{{{{1}}}|Credit Support Balance}}. It’s easier to treat that as an {{isdaprov|Unpaid Amount}} (none of this tedious mucking about with replacement costs and so on).
 
But that means you have to deem the {{isdaprov|Close-Out Amount}} (or {{isdaprov|Market Quotation}}/{{isdaprov|Loss}}, if under a {{1992ma}}) as zero.<ref>
 
Spoddy point: “{{isdaprov|Loss}}” in the {{1992ma}} ''includes'' the “{{isdaprov|Unpaid Amount}}” concept in the definition: “{{isdaprov|Loss}} includes losses and costs (or gains) in respect of any payment or delivery required to have been made (assuming satisfaction of each applicable condition precedent) on or before the relevant {{isdaprov|Early Termination Date}} and not made...”  </ref>

Latest revision as of 13:30, 14 August 2024

Grace periods

Those of ISDA credit support arrangements that are standalone Credit Support Documents — the 1994 NY CSA, 2016 NY Law VM CSA and the 2018 English law IM CSD, but not the 1995 CSA and the 2016 VM CSA — bake in their own two {{{{{1}}}|Local Business Day}} grace period into their own customised Events of Default.

The 1995 CSA and 2016 VM CSA, being Transactions, don’t need to and inherit the grace periods for the Events of Default in the ISDA Master Agreement.

That can, at the limit, create a dissonance for Americans and IM posters, as your CSA grace periods do not automatically track your ISDA ones. Two {{{{{1}}}|Local Business Day}}s is an LBD longer than the standard one LBD grace period in the 2002 ISDA, and an LBD shorter than the standard 3 LBD grace period baked into the 1992 ISDA.

Which is nice. This is not a problem for the English law Credit Support Annexes that are Transactions.

Including “comprehensive” Termination Events

Consider expanding of the {{{{{1}}}|Default}} provision to include “Termination Events where all Transactions are Affected Transactions”. This is as per Section 3.2 of the 2001 ISDA Margin Provisions which recommend that Paragraph {{{{{1}}}|6}} of the CSA should apply where all Transactions are closed out following an Event of Default or “{{{{{1}}}|Specified Condition}}” — the latter of which is defined to include the Termination Events listed under the ISDA Master Agreement. All Transactions would likely be Affected Transactions should a Credit Event Upon Merger or Additional Termination Event occur.

The default paragraph explains how you value the 2016 VM CSA itself — being a Transaction in its own right, of course — when closing out an ISDA Master Agreement. The basic gist is that you treat the Credit Support Balance as of the Early Termination Date — being the total value of the Credit Support you have ponied up at any time — as an Unpaid Amount, rather than treating is as a contingent return obligation, the present value of which would go into the Close-Out Amount (or Loss, or Market Quotation, if you still labour under an antediluvian 1992 ISDA).

Why do title transfer CSAs reference “Unpaid Amounts”?

A title transfer CSA is technically a Transaction under the ISDA Master Agreement in its own right — that is deep ISDA lore — but it is still a weird Transaction, and the standard “replacement cost” of valuation method doesn’t work brilliantly: rather than having defined payments upfront, each of which can be valued and discounted back to now to reveal a present value, payment obligations under a title transfer CSA depend on the aggregate discounted cashflows of all the other Transactions under the ISDA Master Agreement. So good luck determining, in the abstract, the replacement cost of something like that.

The good news is you don’t have to: the {{{{{1}}}|Credit Support Balance}} isn’t calculated by reference to its own discounted future cashflows: rather, it is just the inverse of the aggregate present value of all the other Transactions under the ISDA Master Agreement. So the “replacement cost” on any day is just the prevailing value of the {{{{{1}}}|Credit Support Balance}}. It’s easier to treat that as an Unpaid Amount (none of this tedious mucking about with replacement costs and so on).

But that means you have to deem the Close-Out Amount (or Market Quotation/Loss, if under a 1992 ISDA) as zero.[1]

  1. Spoddy point: “Loss” in the 1992 ISDA includes the “Unpaid Amount” concept in the definition: “Loss includes losses and costs (or gains) in respect of any payment or delivery required to have been made (assuming satisfaction of each applicable condition precedent) on or before the relevant Early Termination Date and not made...”