"Well, What Do You Want It To Be?"
Today I will follow a debt trail, from loan origination all the way to its ultimate existence as part of a credit derivative product. I will use a sup-prime mortgage loan as an example, but any debt obligation will do. Keep the question of the title in mind, it will make sense in the end.
Let's start two years ago with Ron and Ronda White, a couple in their early 30's with a combined income of $60.000 who have their eyes set on a $300.000 house to call home. They have saved only $5.000 to put down, which barely covers the closing costs. Their mortgage broker talks them into a $250.000 first mortgage ARM with an initial 2-year teaser rate of 2% rising to prime+1% thereafter and a $50.000 second, 30-year fixed at a whopping 10.5%. Despite the obvious problems apparent right from the start, such loans were made to hundreds of thousands of people. But no matter...
The two loans were immediately sold to investment bank XYZ who pooled them with other loans (creating Residential Mortgage Backed Security, or RMBS) and placed them inside a CDO. Using recent default data, the financial engineer employed by XYZ took 90% of the White's outstanding mortgage amount and placed it in CDO Tranch A, the supposedly safest portion rated AAA and paying 0.10% more than other AAA straight corporate bonds. The rest was apportioned 7% to Tranch B rated BBB, paying 1.5% more than equivalent bonds and the remaining 3% to Tranche C, also known as the "equity" tranche, which was unrated and paying 10% above Treasury bonds. In case of default, Tranche C gets hit first until it is exhausted, then Tranche B and, finally, Tranche A. This is a "cascade" or "waterfall" pattern, common to all such collateralized products.
Notice how 100% of a loan package that could be described as CCC has been turned into 90% AAA, 7% BBB and 3% NR. In plain terms, the "engineer" is betting that no more than ~3% of the total principal and interest will be lost, including recoveries from selling foreclosed real estate.
Call this the First Derivative - depending on conditions, the market prices of the CDO Tranches will vary significantly more than straight corporate or government bonds.
Those CDO Tranches are then sold as follows, typically:
Tranche A to a pension fund attracted by the slight yield premium on a AAA bond.
Tranche B to a fixed income mutual fund.
Tranche C to a hedge fund attracted by the high yield - or is retained by XYZ.
So far this has been a plain vanilla process, the only question mark being how high or low the "engineer" places the assumptions for defaults and recoveries.
The next step in debt "derivativization" is the issuance and trading of Credit Default Swaps on the first two Tranches of the CDO. It can be done by XYZ, another bank, a hedge fund or all of them - there is no limit. These swaps guarantee payment in case of default events by the CDO, itself a conduit for Mr. and Mrs. Smith's mortgages. These CDS's require an up front payment and subsequent semi-annual ones, usually for up to five years. One can think of them as tradeable insurance policies. Naturally, Tranche A carries a much smaller insurance premium than Tranche B, given the respective AAA and BBB ratings.
Call this the Second Derivative - the prices of CDS's will certainly vary more than the prices of the underlying Tranches and much more than straight bonds.
These CDS's generate income to the seller, who assumes the risk of making the buyer whole if the CDO Tranches experiences payment shortages. Who sells this insurance?
CDS on Tranche A may generate 0.15% annually and is typically sold by a bank or a pension fund attracted by the income generated by insuring a AAA credit.
CDS on Tranche B may generate 1.30% annually, commonly sold by hedge funds.
Who buys the stuff? It would seem pointless for the CDO owner to buy protection for the bonds he already owns, but it does happen for portfolio hedging purposes or even as a way to "trade" the underlying CDO's without actually selling or buying the actual bonds.
But there are more buyers than just hedgers, as we shall see below.
Another investment bank, it could even be XYZ itself, buys a bunch of such CDS's and creates another CDO, also with tranches, ratings, etc. Remember, all you need for a CDO is a stream of regular payments to slice and dice into tranches.
We have now reached the Third Derivative stage: the potential volatility of such a product can be orders of magnitude greater than a simple bond. This is a CDO made up of CDS on another CDO made up of RMBS's - the alphabet soup is thickening fast. The credit leverage, i.e. what happens to the price of this type of product for a given rise in loan defaults in the, by now very distant, mortgage is very, very high.
Should an "investor" actually provide cash to purchase the above Third Derivative CDO we have what is known as a "funded" CDO. But this is becoming increasingly uncommon, because there is yet another derivation that can be performed on Mr. and Mrs. White's nortgage.
Another investment bank (or the original XYZ) can construct an "unfunded" CDO from the CDS's in step three, an instrument that just pays out or demands payment from its owners on a quarterly basis, depending on the shrinking or widening of the CDS spreads. This "synthetic" CDO owns nothing - not even the CDS; it just uses them to mark to market the said CDS spreads and to thus calculate the quarterly payments.
We are up to the Fourth Derivative. Sorry, but I have run out of superlatives to describe the leverage, volatility and credit risk sensitivity of these constructs. And yet, les apprentices sorciers who cook them up think it "innovative financial engineering". Like any fourth derivation, the price of such instruments is completely unrecognizable versus the original mortgage.
Thus the title of today's post, explained best by an anecdote:
Bank XYZ is looking for a dealer for its derivatives desk. Candidate A walks in the door and the desk manager asks, "What is 2 and 2?" --- "Four, sir" answers the job candidate. "Next", says the manager.
Candidate B is asked the same question: "Depends on if you mean 2 plus 2, or 2 minus 2", answers B ..."Next"...
Candidate C comes in and before he answers the question he looks at the manager and asks: "I just want to be absolutely sure - the job is for the credit derivatives desk, right?"
"Certainly", answers the desk manager.
"In that case, 2 and 2 is whatever YOU want it to be", says C.
"Hired, when can you start?"
P.S. Now, let's say that Mr. and Mrs. White default on their loan and, along with them, another 6% of the mortgages default, too - way above what the "engineer" had assumed. What will happen to the prices of the above RMBS, CDO, CDS, Synthetic CDO's? Aren't we lucky we have trader C to tell us - or aren't we?
P.P.S. If it was not made clear, unlimited CDS's can be written on a particular debt obligation, including the CDO in step one. Say a "first derivative" CDO has $100 million outstanding. The CDS's issued against, however, may amount to many times that - as I said, there is no limit. Therefore, there is no limit to the third or fourth derivative CDO's that may be issued, themselves backed by those multiple CDS's. This situation can easily create a viral contagion negative effect, as one "sick" original CDO can infect many others through the CDS market. Ouch.
