Ed Harrison of Credit Writedowns alerted us to the impact on collateralized debt obligations, which have been crowded out of the headlines as a source of worry by other credit market train wrecks, like the failure of Libor to show much movement despite heroic central bank interventions.
Note that the Iceland-related damage applies to a subset of CDOs, the the so-called synthetic CDOs. Synthetic CDOs bundled cashflows from premium payments on credit default swaps and then tranched them, like CDOs that (more conventionally) held tranches from other securitizations, such as pieces of subprime securitizations (they also could hold whole loans, but they tended to be a small part of the mix) Even in this sector, there was a great deal of heterogeniety in the structures and underlying assets. Many were called “trades” and never intended to be resold.
Many had also had tremendous embedded leverage, This hit mainly the lower tranches, but even with the top tranche, if you crossed a magic threshold in loss expectations, you often see very sudden, dramatic decay in value.
From Bloomberg:
Iceland’s collapsed banks pose a “substantial” risk to collateralized debt obligations that made bets on corporate debt, according to Standard & Poor’s.
Kaupthing Bank hf, Landsbanki Islands hf and Glitnir Bank hf were included in 376 CDOs worldwide, S&P said. Another 297 made bets on two of the three banks. The CDOs packaged credit-default swaps that pay investors if there is a default, and the government’s placement of the banks into receivership triggered a settlement of the contracts.
Because the so-called synthetic CDOs also bet on Lehman Brothers Holdings Inc., which filed for bankruptcy on Sept. 15, and Washington Mutual Inc., the bankrupt holding company of the largest U.S. lender to fail, the “impact of these exposures is likely to be significant,” S&P said in the statement yesterday…
The cost of hedging against default by the Icelandic government has soared to 948 basis points, according to CMA Datavision prices for credit-default swaps. That means it costs 948,000 euros a year to insure 10 million euros of debt for five years. It compares with 118 basis points for the Czech Republic and 238 basis points for Morocco.
Sellers of credit-default swap protection must pay the buyer face value in exchange for the underlying securities or the cash equivalent after a bankruptcy filing.
Many of the deals also will lose payments and loss cushions from contracts linked to Fannie Mae and Freddie Mac, the mortgage-finance companies seized by the U.S. government last month. The takeovers caused a technical default on the credit swaps.
The CDOs sell notes to investors that are repaid using the proceeds of credit-default swap premiums. Credit-default swaps are financial instruments based on bonds and loans that are used to speculate on a company’s ability to repay debt.
The cost of protecting corporate bonds from default rose today on investor concern a global recession will sap earnings and companies’ ability to repay their debt.
14 comments
Comments are closed.
“Synthetic CDOs bundled cashflows from premium payments on credit default swaps and then tranched them, like CDOs that (more conventionally) held tranches from other securitizations, such as pieces of subprime securitizations (they also could hold whole loans, but they tended to be a small part of the mix) Even in this sector, there was a great deal of heterogeniety in the structures and underlying assets.”
Wow! Great explanation.
Back to yesterday’s discussion of “innocent fraud”… could anyone really, really understand the risks associated with those products?
If you really want to understand CDOs, pick up a copy of The Trillion Dollar Meltdown. Excellent history and explanation of the whole mess.
I am struck by the full implication of the S&P statement.
The failure of Iceland's banks poses a substantial risk to the CDO market.
And, given the failures of major players in the CDO market to date, the “impact of these exposures is likely to be significant,''.
Hold on to your hats here.
This is definitely the money story of the day.
This is where the house of cards comes out from behind the curtain.
@Ron – I read somewhere that they run Monte Carlo simulations to get as close as they can to the true risk.
They can only get a statistical approx of risk.
If Iceland poses a threat, then just how big is the threat coming from the Wall Street giants? Any back of envelope figures?
They’re large in relation to Iceland, but according to Bloomberg their deposits are 9 times Iceland’s GDP or 19 billion, or 180 billion. That’s smaller than WAMU and less than 1/4 the size of Lehman, so while it would be bad, the Icelandic bank failures shouldn’t threaten catastrophe.
It is said there are beautiful women in Iceland, but as for the banks, that is a matter which needs more analysis:
FYI: Kaupthing is a European bank that operates in ten countries, including all the Nordic countries, Luxembourg, Switzerland, the UK and the US. In addition, the bank operates a retail franchise in Iceland, where it is headquartered.
Based on Kaupthing’s market capitalisation of €9.2 billion as of 10 August 2007, the bank is currently the seventh largest in the Nordic region. Kaupthing’s shares are listed on the OMX Nordic Exchange in Iceland and Sweden.
Both Kaupthing and NIBC have similar business models targeted at providing the full range of financial solutions to SMEs and affluent individuals, and the two businesses are complementary in terms of strategy and geographic spread. This is a great deal for Kaupthing and we all look forward to entering the exciting new markets.”
This is like shooting retarded fish in a barrel, so, is there a point to the stupidity of financial managers around the globe????
NIBC Bank profit after tax from continuing operations in the first nine months of 2007 13% higher at EUR 204 million
The Hague, 16 November 2007
Investment Management is responsible for developing asset management activities for NIBC’s Collateralised Debt Obligations (CDO) platforms and funds for both credit fixed income and private equity investments. It also manages NIBC’sminority interests in general partners of a number of third-party managed private equity funds. Investment Managementwas created to separate the asset management activities from the bank’s other activities
> http://www.nibc.com/press/Documents/Financial%20Press%20Releases/071116_NIBC_Q3_results_2007_final.pdf
FairEconomist,
I'm not reassured. Synthetic CDOs are pretty dodgy paper. I'm amazed this stuff hasn't been written to zero already, but from what S&P hints at, the Iceland matter is going to have a serious impact on these trades. And size of balance sheet pre-bankruptcy isn't necessarily a reliable guide. Lehman had far fewer CDS per dollar of cash bonds that Delphi when it filed for bankruptcy. I have no idea where Iceland sits on this matter.
And even if true, saying that the Iceland CDS unwind is smaller than Lehman isn't reassuring either. That's like saying pneumonia is less serious than congestive heart failure. But we aren't talking one versus another, we are talking one AFTER another. And pneumonia after congestive heart failure is a completely different kettle of fish than pneumonia in an otherwise healthy person.
Thanks for posting this story and the one on Russia – even with 3 nationalities I was being pretty parochial – I just assumed that other countries will do the same shenanigans that this country has done to prevent catastrophic credit events due to the bank failures ( e.g. 79.9 ( not 80%) control of AIG; FNM, FRE conservatorship with full guarantees ) – ok LEH went to a proper “credit event” but still…
Thanks for the reminder that a worldwide CDS/CDO shoe still has to fall and that IT might fall with a bigger thud than the local ones.
-K
Yves, you (or perhaps the Bloomberg author) don’t seem to have grasped the essence of the Synthetic CDO.
1A. The issuing trust/spv of a cash CDO physically OWNS the pool of underlying assets, the cashflow of which gets pressed thru the tranches to produce stratified returns.
1B. But in a synthetic CDO the pool never gets bought. Instead it’s a “notional” (imaginary) pool. It exists only as a list of instruments stapled to the back of your offering memo: the so-called “reference portfolio.”
A synthetic deal then proceeds thru time with an eye on the actual performance and mishaps of the instruments in the “reference portfolio.” But without ever owning them.
(It’s something like Rotisserie Baseball — where the actual performances of real baseball players get translated into the imaginary world of the Rotisserie league, and there produce results.)
Synthetic CDOs, then, are “naked” in the sense of (x) a Credit Default Swap where the subject security has not been warehoused — or (y) a short stock sale where the shares have not been properly borrowed.
For a richer picture of a synthetic cdo at work, see the third example here:
http://www.newcombat.net/article_whats_a_cdo.html
The example was taken from life — one of the Bear Stearns High Grade hedge funds.
1C. All the synthetic CDOs I worked on were — like the Bear example — entered into to hedge a specific securities portfolio. They were problem solvers for the owner of that specific portfolio.
Ie, they were NOT huge complex “naked” bets against a pie-in-the-sky reference portfolio by a third-party gambler.
It’s POSSIBLE synthetic CDOs might be used to place a wide-ranging “naked” bet — as so many CDS were used to bet nakedly against a single corporate issuer.
But my imagination protests that it would be an odd circumstance that would lead a client to pay a bank and the rating agencies millions in fees to concoct and close such a CDO — unless (again) that client has a bigger problem that the CDO helps to solve.
It sounds from the Bloomy piece that the synthetic CDOs were (again like the Bear example) hedges for somebody holding a lot of the Iceland bank bonds.
Ie the reference portfolio was the universe of CDS on the Iceland banks. The synthetic CDO then would be just a super CDS.
Thus, the Issuer of those CDOs would be, via a separate complex CDS built into the deal (as in the Bear example), buying protection, and a third party (most likely a “specialty financing” subsidiary of the investment bank that did the deal) would be the protection seller.
The buyer of the CDO tranches would thus be the net protection buyer.
Perhaps the Icelandic banks themselves, or the Treasury or central bank (if they have one).
2. Also au contraire: the type of instrument in the pool has nothing to do with whether the CDO is cash or synthetic.
Corporate bonds, ABS, CDO tranches, CDS, LBO loans — all might or might not be in any given cash or synthetic cdo pool.
3. You touch on secondary market trading.
To be gross to be brief: No secondary market worthy of the name existed for CDO tranches. They could be dumped if needed, dealing thru investment banks, but the dumper would expect to take a haircut.
In good part this is because CDOs were were problem-solvers for a particular client and tended to have idiosyncratic problem-solving mechanics.
In contrast, CMOs — where the pools consisted purely of mortgage bonds or long-term commercial leases — were both voluminous and more generic, more interchangeable, and thus the secondary mkt here was rather healthy.
The Iceland synthetic CDOs in the bloomy piece seem to have been “pure” — referencing the universe of CDS on the Iceland banks. But it’s hard to imagine the volume here worldwide would support active secondary trading. But maybe I’m wrong.
For a discussion of these secondary markets differences, see:
http://newcombat.net/Conversation/credit-crisis-alphabet-soup-phantom-market-value/
Krakpotkin, I suspect the level of “gambling” with synthetic CDOs may be higher than you realize. In the direct CDS market at least, the volume of protection purchased has been far in excess of the underlying notional value, confirming the existence of a large speculative presence. I would be very surprised if the synthetic CDO were any different than its underlying.
Perhaps large portions of the synthetic CDOs were constructed to meet specific demands, as you indicate, but that raises the question of why the purchasers didn’t simply buy direct CDSs. It is the tranche nature of the CDO, synthetic or otherwise, that reveals its inherent purpose: the tranches apportion the risk, and hence the returns, among the purchasers. But an entity looking for protection will simply buy straight protection; only a speculator or an investor is interested in the risk versus return.
While the secondary market for synthetic CDOs may be weak, the very existence of the tranche structure argues for the presence of speculators and investors; and if speculators and investors are involved, then the result of the CDS triggers may be more than simple payouts, but may have dramatic effects on the balance sheets of entities who, it will be discovered, had no actual exposure to the reference securities.
Dear Peripheral Visionary,
Re your argument that the very nature of the CDO tranche structure nearly determines its use for speculation, I refer (again) to the third example here …
http://www.newcombat.net/article_whats_a_cdo.html
… (headed in bold “A More Interesting Problem”) — where ALL the tranches of a syn CDO were bought by Bear Stearns Asset Management — to hedge the entire portfolio of one of its High Grade hedge funds — which HG portfolio precisely matched the notional reference pool.
Thus: No speculators. Indeed: No third-party investors of any sort.
The economic reality of the deal was BSAM hedging its hedge fund, with the investment bank (CSFB) that did the CDO acting as protection seller thru a subsidiary.
As to why BSAM didn’t choose to use a thousand CDS instead of the syn CDO, I can only guess:
— Because it would have been more costly?
— Because counterparties could not be found to write the protection on some of the (arcane?) securities in the HG portfolio?
Re the latter case:
— the BROADNESS of the CDO approach might have been attractive to the protection seller.
However that may be, take note that the tranche structure of the approach chosen was rather irrelevant, given that BSAM was buying every dime of every tranche.
That is: The tranche structure was merely an effect of the form of the chosen instrument and — crucial to note — of rating agency involvement in blessing the deal.
That is: It’s the rating agency’s chief function as risk evaluator (once it tweaks the detail mechanics of the deal) to assign the $$$ amount of each tranche (revising shortly before closing the estimates that the bank made early on as the deal was being crafted).
Is that point clear? (I fear perhaps not.)
In the example at hand, the fact that the solution to BSAM’s problem took the form of a tranched CDO was somewhat sumbebekos — “accidental.”
BSAM wanted to hedge a huge arcane portfolio. CSFB offered (I imagine) to sell that protection using the pooled form (ie, a CDO rather than a zillion CDS) — and this meant the rating agencies would, as usual, be divvying up the pie into tranches AS THE MEANS toward the end of rating the pool’s risk.
I doubt the agencies knew that BSAM was the only “investor” in the tranches. (Even CSFB’s lawyers didn’t know until an email accident spilled the beans a few days before closing.)
The tranched form of the CDO, then, was convenient and familiar — easy to do — but accidental to the economic reality of the deal.
And all the syn CDOs in our shop were similarly motivated.
Perhaps there were other shops out there using the same form to place speculative bets for Big Rollers (willing to pay several millions in fees for the pleasure). But I’ve never had my hands on such a deal.
OOPS — I made a journalistic error in the fourth-to-last sentence in my post above.
It should have read:
I doubt the agencies knew that BSAM — which was acting as Portfolio Manager for the syn CDO — actually owned the Reference Portfolio in one of its HG hedge funds. (Even CSFB’s lawyers didn’t know until an email accident spilled the beans a few days before closing.)