Friday, August 01, 2008

And 15 once we reach Alderaan...

Today Ambac is out with news that they too have reached an agreement to terminate a CDS contract with Citigroup. The CDS had $1.4 billion notional and was settled for a payment of $850 million. On Monday, Security Capital reached a similar agreement with Merrill Lynch on 8 CDS, canceling $3.7 billion in CDS for a payment of $500 million.

Remember that a CDS is nothing more than an exchange of risk. One party agrees to pay a periodic fee, while the other party agrees to make up any lost cash flow on a referenced security.
So let's say you enter into a plain vanilla CDS contract referencing Kraft. Let's say its $10 million notional and the deal spread is 76bps (about what Kraft is right now.) Now let's say that the spread on Kraft CDS widens to 100bps. Your contract at 76bps has some positive cash value (about $110,000 on $10 million notional), because of the movement in spreads.

Why? Because you currently own protection on Kraft and only have to pay $19,000/quarter for that protection (76bps * 10 million / 4). Anyone who wants to buy protection now has to pay $25,000/quarter (100bps *10 million /4). Through the magic of Bloomberg, we can calculate what your contract is worth: about $110,000. This means that someone would be willing to pay up front $110,000 to "buy" your lower protection payments.

Now let's say Ambac was the counter-party on your Kraft CDS, and Ambac wants out of the trade. Maybe you are happy to monetize your profits so when they offer you $110,000 you agree.

The media headlines would say that Ambac had exited their $10 million in Kraft risk in exchange for a payment of $110,000.

Now let's look at the deal cut by SCA and Merrill Lynch. Since CDS on the ABS CDOs in question don't trade regularly, we can't tell what the "fair value" actually is. Safe to say that the deal between SCA and Merrill was in part a distressed exchange. If we look at the AAA tranches of the ABX index, the prices range from $88 to $44. It is my contention that ABS CDOs, even if they include large non-RMBS exposure, can't possibly be worth more than the ABX, assuming we're controlling for vintage.

We know that SCA paid Merrill approximately 13.5% of the CDS notional value. That'd imply a dollar price on the CDOs of $86.5. Its not quite that simple, but its close. Anyway, if the price range of the ABX is $88 to $44, the right value of these CDO positions is no better than $60. Probably more like $30. So you'd say that Merrill's acceptance of $500 million is anywhere from 1/3 to 1/5 of what they'd have received from a well capitalized counter-party. Admittedly, I'm completely wagging these numbers, but it's safe to say this was a distressed exchange.


GS751 said...

good post, i just look at CDS as put options for bonds, but a bit more complex, am I wrong or right here?

Sivaram Velauthapillai said...

I don't understand this point. Would you mind elaborating:

"It is my contention that ABS CDOs, even if they include large non-RMBS exposure, can't possibly be worth more than the ABX, assuming we're controlling for vintage."

Is it not possible for the ABX index to be mispriced relative to the actual assets (say the actual CDOs) due to heavy hedging (i.e. shorting of ABX)? Or am I misunderstanding your point?


PNL4LYFE said...

When these numbers came out, the thing that struck we was that CDS was being unwound at over 80c on the dollar versus CDOs being sold at 22c (ignoring the financing component). It made me wonder whether this was due to

1) extremely distressed prices on both legs of the trade.

2) extremely poorly matched hedge

3) a little of both

4) something else.

On the CDS side, I think they were being carried on the books at about twice what they were actually unwound for or roughly 70. I think the CDOs were marked at around 40 for Q2. This leads me to believe that it's probably a little of both, but I was hoping something here might have a little more info or better analysis.

asphaltjesus said...

Thanks for the post clarifying the valuation issue. It has helped me immensely.

pnl4lyfe brought up another question which may be obvious to some, but not me. How do these transaction get marked to market? I'm not asking a GAAP-specific question.

From my very simple understanding of the transactions, it seems to me there are a number of opportunities to "enhance" the accounting/book/? value of both parties regardless of whether or not wealth was created. Another way to say it that might make more sense: Selling the risk seems like it could be used to inflate the numbers that Wall Street uses to establish value and performance metrics on both sides of the sale.

It's totally possible I'm still not thinking about this the right way. If that's the case, please be patient and give me some more reading material references in order for me to answer my own questions.

PNL4LYFE said...

My limited knowledge of the accounting issues is from interest rate derivatives; specifically, certain OTC derivatives can use accrual accounting rather than MTM if it qualifies as a hedge to specific assets or liabilities. The rules are complicated, but I think the issue here is that the CDSs MER purchased from SCA qualified as a hedge on the ABS CDOs that they owned. Under normal circumstances, this would allow both pieces to avoid MTM.

One way this can create 'phantom' PNL is that the two sides of the transaction account for it differently. If the CDS buyer uses accrual and the seller uses MTM, the seller would show a gain if the spread tightens while the buyer would not show the offsetting loss. However, in this case, it would have been the opposite; SCA would take losses as the spread widened while MER would not have realized gains. This example is probably moot though since MER has been taking writedowns on the CDOs for several quarters along with writeups on the CDS. I'm not sure what event triggered the change in treatment; probably "extraordinary change in value" or something like that.

arb-it said...

The CDS that your writing about are on CDOs, most likely High Grade ABS CDOs. These were highly levered instruments that had collateral with an average of about Aa3, and a size of 1-3 billion dollars. The leverage was a minimum of 100 times (so a $1.5B security would have a $15M equity piece). The structure would include a super senior, a junior senior (both Aaa), Aa, A, and possibly a Baa tranches along with the equity. The majority of the structure was in the super senior. Using our 1.5B structure as an example, the Super Senior would be about 1.2B. This would be "wrapped" by a monoline like Ambac or MBIA. Wrapping is almost the same as a CDS. The provider gets a fee for ensuring payments to the note holder.

There are some added wrinkles. The wrap provider has to approve the documents and the collateral. And during the life of the deal, if certain conditions are not met, the provider can reassign the deal to another manger or even take it over and liquidate it.

The ABX index is basket of 20 names all initially rated at the stated rating. The Aaa index includes 20 Aaa rated (last pay Aaa, but still Aaa) HEL bonds. Currently the basket is rated somewhere around Aa3 or even A1, but when the index was created, they were Aaa. There is another set of indeces called the TABX that is the tranched index that corresponds very closely to CDOs, but nobody trades it and it is hardly ever used.

So why did SCA get off the hook with ML? or did they? SCA only needs to pay on an event of default on the Super Senior. The bond is most likely paying and will continue to pay for a while. Why? because there is still some subordination and because the cash traps in the CDO have most likely failed the Super Senior is getting all of the cashflow that the collateral generates (minus fees, expenses and swap payments).

ML most likely owns a considerable portion of these CDOs, including the Super Senior and most likely some of the other tranches. It is in their interest to get as much of their money back as quickly as possible. SCA is getting 8-10bp on the portion they've wrapped and don't want to pay. By getting rid of the contract, the CDO gets a lump sum payment now and and extra 8-10bp on 80% of the notional of the deal, which will help pay down the notes of the deal and they have a remote possibility of getting some of the other notes that they own that were not wrapped paid off.

I agree that it was a distressed exchange, but SCA had a lot of legal rights in these transactions and it seems that they got paid pretty well for signing them away.

Megan said...

Pretty effective info, thanks so much for this article.