October 5, 2012 
For those that know what PAI is, the first part of this post can be skipped, the rest of the post is my understanding of why PAI is applied to OTC portfolios, and not Futures. This arose from a question at a training course I delivered, and was interesting to understand.
What is PAI?
Imagine two banks, with an Interest Rate Swap, cleared through a CCP such as SwapClear. At the beginning of the life of the swap the PV is close to zero, so worth little to either party. Over the life of the trade the value of the floating leg will vary leading to an NPV to one of the parties. The change in this NPV from day to day is what Variation Margin is, calculated and moved by a CCP on a daily basis.
Imagine you are the unlucky party where the NPV is against you (out of the money), the CCP will require you to pay the NPV into the CCP, which they will transfer onwards to the other party.
- You: -NPV (and suffer a funding cost)
- Counterparty: +NPV (and invest to earn interest)
In this case the other party enjoys the benefit of depositing the NPV and earning interest on it, very nice. You on the other hand, have to fund the NPV payment which costs you money.
PAI is used to level this uneven playing field by transferring the benefit of receiving the NPV back to the payer of NPV (VM) like this:
- You: -NPV (and suffer a funding cost) : Receive PAI
- Counterparty: +NPV (and invest to earn interest) : Pays PAI
PAI is the overnight cost of funding the VM, debited from the receiver and transferred to the payer. The rate at which PAI is calculated is chosen by the CCP so may not match your own cost of funds.
So the net effect is this:
- You: Fund the VM/NPV but receive an amount of PAI to compensate
- Counterparty: Receive VM/NPV but are debited PAI which offsets the interest you might earn.
The question
Why is PAI included in the CCP model for OTC products, and not for futures?
Legal aspects of Futures
A clue can be found in this blog post by Noah Melnick at Linklaters, which describes how trades such as futures are treated at an exchange / CCP.
For a Futures trade, the rules of the CCP define that each day the closing price of the contract is used to calculate VM, and the profit or loss is debited from your margin account daily. The exchange regards this payment as full and final settlement of the contract, as should your counterparty default, you will receive no further payment on this contract. This means a Futures trade is regarded as “final” or “matured” on each and every day, until the expiry date of the contract, in which case the VM payment is a settled amount, and not a margin asset.
In this case you are free to re-use any VM payment, as in the example above, without the PAI adjustment.
OTC PAI treatment
For an OTC trade, the VM payment represents the value of future cash flows, which remain unsettled until each coupon date, and final maturity, so you regard VM as an amount of margin which you would receive back should the value of the contract move in your favour. The difference in default, is that in some CCPs for OTC products, the portfolio of trades owned by the defaulting party (for direct members) are auctioned, such that eventually a new party takes ownership, and the contract continues unchanged until maturity.
This means the VM payment for an OTC trade is characterised as a margin asset, which you would expect not to lose should your counterparty default, hence the application of PAI, equivalent to receiving interest on your margin asset, is applied.
A further aspect is equivalence of an OTC IRS in the cleared and bilateral world. Should PAI be ignored by a CCP, the payer of VM would feel the penalty of clearing, which they wouldn’t outside a CCP. PAI evens this out to some degree.
More work is being done by ISDA to promote the use of the Standard CSA to align the costs of VM with the currency and accrual basis of the Swaps, using OIS rates instead of the LIBOR curve.
Noah
Read the blog post from Noah who covers an unresolved question on why or how VM should be segregated under US law, but of course it isn’t held by a CCP but passed through between parties.
Update May 16th 2013 from Jon Skinner
- PAI also causes the convexity factor to be added to swap curve building / pricing methologies.
- Banks treat cleared VM as collateral not settlement in the futures model. However, CME treats VM as settlement even for swaps.
- Deliverable Swap Futures are another variant and the VM in that case is supposed to be settlement i.e. realized P & L. They do have PAI but the VM is clearly not collateral.
Amir Khwaja
Ben Larah
Michael Beaton
corinaki24
Jon Skinner
Mihir Vaidya
Maria Leontiou
Martin Schuhmayer
samihodgson
Bill Hodgson
Tom Riesack
Makes perfect sense! Excellent post. Somewhat related, an underappreciated subtlety of moving from uncleared to cleared environment is where collateral movements are still under title of payer who may recieve interest on asset, where in cleared environment this benefit disappears since the receiving end can withdraw VM – correct me if I am in mistaken.
VM is not normally held by the CCP so the receiver is free to do as they please. Securities are indeed held at the CCP, with full title transfer so removing the right to coupons.
I think that’s what you’re getting at?
Yes that was what I was getting at. Also, my thought was for uncleared postings of additional collateral, posting party may still have title. Moving to cleared world, posting additional means VM (cash) and is immediately available to receiver – thanks.
I thought the point of PAI was “sterilize” the pricing effect of variation margin compared with uncleared swaps. The problem is not that the payer of VM is disadvantaged relative to the receiver. When trading at market, there is no reason to assume that either party is more likely to suffer. The problem is that the economic costs and benefits of daily settlment of VM are asymmetrical. Fixed rate payers receive VM when rates rise, at the same time when investing that VM is more profitable. Conversely, they only have to pay VM when rates fall, and it is thus cheaper to fund. This asymmetry means that, without PAI, cleared swaps would have to trade cheaper than uncleared (just ask Jeffries). This is the case with futures, which trade at a discount to expected forward rates because of the convexity effect introduced by variation margin.
Good points Geoff, thanks.
Agree with Geoff, if you did not have PAI, the cleared and uncleared swap would not be economically equivalent, and the convexity correction introduced by the futures style margining would be large for long dated swaps. The case of Jefferies and IDCG is described here in this bloomberg article, http://www.bloomberg.com/news/2011-09-19/jefferies-group-sues-international-derivatives-clearing-group-on-rate-swap.html
I agree completely and was going to add a link to the Jeffries case. The reason for this post was to illuminate why PAI isn’t used at a typical futures exchange, all your input on why it is needed for OTCs is spot on. Thanks all.
Hello, I have a few questions. Firstly, I do no understand the argument that just because I have paid VM on a future yesterday the position would be closed. I would still be liable to fund VM the next day and pay it if the market would move against me. So, why should I not demand PAI either? Secondly, I understand the argument why PAI would need to be included to ensure equivalence to the OTC market. However, looking at the point raised by me before, how should the CME Swap future with pysical delivery be margined? I would say always with PAI (and on all futures for that matter), but given it is a future and the general argument made earlier I am confused. .
and how would one, should one treat the new CME Swap futures?
I found a CME PDF on their Swap Future, it’s here: http://www.cmegroup.com/trading/interest-rates/files/dsf-overview.pdf
Page 6 says there is no PAI on the Future, the reason being, as described above, that whilst you are party to a futures contract, the close-out period is immediate, so each VM movement is regarded as a realised profit or loss on the contract, and not an on-going margin amount.
On default of your counterparty to a Futures contract, the CCP will close out the contract same-day, there will be no more VM movements on that contract. This from a legal view, is your final settlement, therefore the daily VM is seen as just that, not as a long term piece of collateral.
In the OTC CCP market, the VM is preserved in a default of a direct member, and is returned to you as the value swings in your favour. The auction process for an OTC contract means the VM continues uninterrupted, and therefore PAI is paid and received, as legally the VM is seen as a collateral asset, rather than a P or L payment.
That’s the best re-iteration I can do, check the CME PDF above and see if it helps.
Bill.
hm, thanks Bill. No sure whether I fully agree with your default argument but I suppose I will need to accept it. What would happen if you were long 50 yards of this 30 year swap future or any future. I doubt it that the exchange could handle that in a day. Who would put any price in the order book? It seems to me more a question of convention. Perhaps Futures exchanges just did not see the funding implications that were associated with swaps for a long time or thought people would day trade only?
Is the CME Swap Future a 30 year contract? Or is that the underlying IRS has a 30 year duration? I believe it’s the latter, therefore the margin on a short term future is much lower, as the CCP only has to margin for the price changes over a relatively short time.
it is a 6 month contract expiring into a 30 year swap…. I doubt that they could ever liquidate a large position 2 daysbefore expiry…. funnily enough, once transferred into a cleared swap they would be forced to calculate the initial margin for a 5 day liquidation period…
Excellent back and forth on this issue – I am attending the swap futures lunch and learn on 10/23/12 and will be bringing questions. https://swapfutures.eventbrite.com/
Reblogged this on Carl A R Weir's Blog and commented:
PIA for IRS’s.