CDS/CDO's Part V- The Rise of the G(r)eeks

In Part I we discussed how purchasers of corporate bonds were actually impliciltly involved in the credit options market. We went on to discuss how this was explained to the Italian banks and this was the beginnings of the cash CDO market. As time went by, and as the available assets became more expensive, some financial engineers suggested that instead of just talking about the embedded options, they should actually extract the options from the structure and start to trade just the option and forget about the "risk-free" component of a bond.

If an investor were being "paid" in the form of a yield premium to "sell" a put on a specific credit, it would be simple to insert an alternative buyer i.e. not the issuer, but rather a trader who had a view of the value of a specific credit. This idea was than expanded to allow traders to "sell" the option, as well as buy it.

A new terminology evolved so instead of buying or selling puts the credit default market spoke of buying and selling protection. If you "liked" a credit, you bought protection; if you didn't like it you sold protection. "Liking" a credit could be a relative value- the real driver was whether it was viewed as being rich or cheap.

This became the CDS market. The Geeks extrapolated the idea and created a structure which would allow them to "synthetically" create a CDO. They could get investors to buy protection on a group of corporate credits or "names" which created the same position for the investor as if they had sold puts, but in this case they actually purchased the protection, and of course paid a good price to do so.

The financial engineers were very good at statistics and they had created a correlation model which allowed them to present a list of 100 names to the investors, which in aggregate would be rated AAA, but in reality the trading desk would find a smaller pool of names which statistically replicated the 100 names, the aggregate cost of which was much lower than the price paid by the investor.

The rating agencies were statistically focused as well and their guidelines meant that there were on average 4 or 5 questionable names-i.e. poor credit risks given their ratings- which could be put into the pool of 100 names in order to get the yield on the pool higher, but were almost preprogrammed to create problems in the pool sooner or later due to the pyramid structure of the synthetic CDO cash flow.

This will be discussed tomorrow.