Comments on: Did ABN Amro hedge its ACA exposure? http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/ A slice of lime in the soda Sun, 26 Oct 2014 19:05:02 +0000 hourly 1 http://wordpress.org/?v=4.2.5 By: Greycap http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/comment-page-1/#comment-14105 Tue, 27 Apr 2010 16:46:16 +0000 http://blogs.reuters.com/felix-salmon/?p=3579#comment-14105 Apologies, Sandrew, I was guilty of some careless reading and thought you were saying the protection ABN bought amounted to a CCDS. Agreed, the protection they wrote to Goldman was equivalent to a CCDS.

I am also guilty of some careless writing: I should make clear that a CVA hedge is a hedge of expectations, not risk. The fact that any given hedge “fails” is therefore not evidence in itself that it was wrong. Nonetheless, I would be surprised if it were right in this case because the incompleteness and illiquidity of the reference markets makes it pretty hard to come up with a unique and verifiable expectation in a pricing measure.

]]>
By: Sandrew http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/comment-page-1/#comment-14100 Tue, 27 Apr 2010 14:43:20 +0000 http://blogs.reuters.com/felix-salmon/?p=3579#comment-14100 Greycap, maybe I have this wrong, but I don’t think our interpretations are that far off.

I think it’s apparent from the record that ABN’s $909M notional supersenior position amounted to a contingent CDS. The contingency part (the reference swap) is the MtM of the $909M supersenior (50-100) tranche of the $1.8B Abacus portfolio. The CDS part (the reference obligor) is ACA default risk. Whether it was structured as a CCDS we don’t yet know, but that’s moot. For example, it might simply be the case that ABN is literally an intermediary between Goldman and ACA on the supersenior tranche trades, where with the former ABN agreed to ACA’s no-collateral terms and with the latter ABN agreed to Goldman’s demand for (bilateral?) collateral posting. Please correct me if you see it differently.

We’re speculating as to the why’s and how’s of ABN’s hedge of this position. I agree that, whatever the form, ABN’s net supersenior position is functionally equivalent to retention of ACA counterparty risk. So it would make a lot of sense that ABN’s CVA desk would be tasked with determining how much of a hedge to put on to cover their mean negative exposure to ACA on the supersenior trade. Good observation on your part.

Another way to think about this trade is to go back to the motivations and roles of the parties. Goldman was basically acting as a broker between Paulson and ACA. Goldman had a limited appetite to retain any jump risk, so they demanded that ACA agree to pony-up collateral if/when the MtM on the supersenior moved against them. ACA said no. ABN decided it had the appetite for the jump risk and agreed to intermediate. 17bps is the price ABN charged for retaining ACA counterparty risk. If they hedged this with a CCDS (whether to Goldman or anyone else), then ABN would have been acting as a broker of counterparty risk, and that wouldn’t really make sense. Rather, ABN is acting as a market-maker. They take on the ACA counterparty risk freely, and dynamically hedge it (likely out of the CVA desk) as they do other bits of counterparty risk. Given the speed with which the subprime market tanked, the illiquidity of the underlying Abacus tranche, and the wrong-way risk (correlation b/w ACA and Abacus tranche), it’d be no surprise if ABN’s hedge failed.

Of course, I’m at a loss to explain the bizarro rating migration term, which makes me suspicious that I’m possibly way off base on this whole mess.

@dukey: Take what I say with a dose of salt. There’s a good chance I’m an idiot.

]]>
By: Greycap http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/comment-page-1/#comment-14098 Tue, 27 Apr 2010 14:05:18 +0000 http://blogs.reuters.com/felix-salmon/?p=3579#comment-14098 Just realized: I kept writing “MtM” when I meant “PFE”.

]]>
By: Greycap http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/comment-page-1/#comment-14097 Tue, 27 Apr 2010 13:38:05 +0000 http://blogs.reuters.com/felix-salmon/?p=3579#comment-14097 Unusually, I don’t think I can agree with Sandrew here, unless I have completely the wrong end of the stick.

First, this buying 27mm protection on ACA is a “sun rises in the east” story. A modern iBank dynamically hedges all of its OTC counterparty exposure as measured by bilateral credit valuation adjustement (CVA). The way this works is that there is a central CVA desk that essentially sells credit insurance to all of the other trading desks in every asset category: FX, IR, equity, energy, credit, whatever. There is a kind of internal transfer pricing scheme for this insurance that may be reflected in the prices made by the business desks. So what probably happened is that ABN crunched the numbers and found that the GS/ACA deal moved their ACA CVA by 27mm. In that case, that they bought protection from GS would have been a coincidence and completely different traders involved from the Abacus deals.

Note that CVA hedging is a dynamic process; it must be continually updated not only as new positions arrive but as the MtM of current positions changes. CVA traders usually try to hedge some of their MtM sensitivities to other asset classes, but that would not have been practical here because of the illiquid underlying. The way you get nailed for 800+mm of loss when you are “hedged” is if the MtM changes quickly in tandem with the price of credit protection.

Why the weirdo “credit rating” reference terms? I dunno – did ACA have any bond issues that could have served as reference securities?

Finally, what is described by Felix was not a contingent CDS (sorry, I haven’t had time to look at any of the primary docs.) The “contingent” part means that the amount of protection is contingent on the MtM of some reference portfolio (all before recovery rates are applied.) But here the 27mm notional was fixed. If it *had* been a contingent CDS sold by GS to ABN – that really wouldn’t have made sense!

]]>
By: Sandrew http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/comment-page-1/#comment-14095 Tue, 27 Apr 2010 12:22:04 +0000 http://blogs.reuters.com/felix-salmon/?p=3579#comment-14095 “they wrote protection on ACA *credit rating*”

Huh. What is that called, a rating migration swap? That’s bizarre. Ignore my previous comment; I have no idea what ABN was thinking.

]]>
By: STORYBURNcom_0 http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/comment-page-1/#comment-14092 Tue, 27 Apr 2010 11:19:14 +0000 http://blogs.reuters.com/felix-salmon/?p=3579#comment-14092 ABN AMRO is going to end up looking foolish here, as will the SEC

]]>
By: alea http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/comment-page-1/#comment-14090 Tue, 27 Apr 2010 09:48:32 +0000 http://blogs.reuters.com/felix-salmon/?p=3579#comment-14090 “If Goldman was happy to write credit protection on ACA…”
they weren’t and didn’t, they wrote protection on ACA *credit rating* (not my definition of a plain vanilla CDS btw, as Sandrew wrote above)

“I’m sure that Goldman charged ABN Amro more than 17bp for ACA protection”
when you buy insurance or in this case credit enhancement, you don’t charge, you pay, so GS paid ABN AMRO 17bps to enhance ACA’s credit.

Which all goes to show that ABN AMRO, was hardly negative on ACA as they “hedged” only for a credit downgrade and not for a normal credit event like failure to pay or bankruptcy and didn’t bother to hedge the reference porfolio at all.

]]>
By: hillemarc http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/comment-page-1/#comment-14089 Tue, 27 Apr 2010 09:20:53 +0000 http://blogs.reuters.com/felix-salmon/?p=3579#comment-14089 Please forgive my non-financial naïveté, but if anyone could de-confuse me on the following I’d be grateful:

1. Who OWNED the underlying portfolio of assets that Paulson/ACA selected and which subsequently defaulted? If they were chosen from GS’s books, mustn’t GS have taken a massive hit too (equal to that of ACA)when they tanked? Wouldn’t they have done better keeping the Abacus credit protection from ACA for themselves, rather than selling it in to Paulson? (Perhaps they hedged these assets with other investors outside of the Abacus structure?)

2. If they didn’t own these assets, where/how did Paulson find them?

3. Still if they didn’t own the assets, did ACA and IKB know that the CDSs making up Abacus were thus “naked” from the start? Perhaps they thought they were “covered” CDSs, i.e. that they weren’t entering into a straight bet with a short investor, but simply providing credit insurance to an investor who had no intrinsic interest in seeing the underlying assets default?

Thanks for the help!

]]>
By: longandshort http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/comment-page-1/#comment-14086 Tue, 27 Apr 2010 08:31:06 +0000 http://blogs.reuters.com/felix-salmon/?p=3579#comment-14086 Only CDS that ABN could have bought to hedge the structure would be a CDS on ACA. Buying a CDS on another CDO would just mean paying for basis risk.

And it would be a nifty if GS did sell a CDS on ACA to ABN which GS in turn purchased from esteemed, revered client.

]]>
By: HBC http://blogs.reuters.com/felix-salmon/2010/04/26/did-abn-amro-hedge-its-aca-exposure/comment-page-1/#comment-14085 Tue, 27 Apr 2010 07:43:34 +0000 http://blogs.reuters.com/felix-salmon/?p=3579#comment-14085 Another trip down the GS cayman rabbit hole in pursuit of balance-sheet logic? You can speculate all day where those elusive supersenior losses might actually have arisen but chances are you’d have better luck finding WMDs in Baghdad.

But you’re right that it’ll get clearer when Europe does the heavy lifting of a real prosecution. By that I mean a sophisticated prosecution: criminal and civil, with a twist of human rights violations thrown in. I wouldn’t speculate on this not happening.

]]>