Nearly all CDO of ABS was held in buy and hold portfolios. This was terrific until some wanted out, and until the impairments became obvious.

Deutsche Bank, Jan 2005, “Global CDO market 2004 overview”: `There remains only a limited two-way market for CDOs backed by structured product. The disconnect cannot exist indefinitely given how large the new issue market has become. We believe it is only a matter of time before holders of ABS CDOs will seek to opportunistically sell deals to either take profits or for other reasons.’

Again: the Fed is claiming abnormal conditions in a secondary market that barely if ever existed for CDO of ABS.

Not for synthetic CDOs.

“I can unequivocally say that purchasers were constantly assured that, especially for the AAA tranches, a secondary market did exist or the bank would make a market for them if necessary.”

True enough, but I’m not sure an arranging agreeing to make a market is the same as claiming there’s a liquid secondary market. I’d agree that people probably assumed there was more liquidity at triple-A than there really was, but outside of SIVs which needed them to be liquid and in some cases ignored the dangers to maintain their yield, I’m not sure how many investors really considered them trading assets (note, I’m not talking trading vs banking book here).

And yes, the swaps that the i-banks did with Cdos (often hedging the AAA tranches or backstopping lower-rated tranches) can most definitely have negative values (meaning that the Cdo owes the bank money but is not presently expected to have the ability to pay it). I have heard that some banks have tried to securitize their portfolios of swap receivables linked to Cdos and sell them to various investors. Now that would be a fun call – if they cannot value them in-house – how do they expect investors to do so? Oh right…they will be rated AAA.

I would also like to know the answer to the question whether the portfolio can have a negative value? It seems to me everyone assumes that there are no swaps in there, or no other instruments which can have negative value. Does anyone know?

Wink, Wink.

But doesn’t the whole pricing concept for CDOs assume liquid markets? In a reasonable model, wouldn’t the illiquidity problem alone reduce the value below that of the underlying assets?

Well, how are you defining liquid? I can see some definitions where that might be true, but those would emphasise the depth of the liquidity and the maturity of the investor base, not what most people think of as liquidity itself – the ability to quickly and efficiently enter or exit a position. By most conventional indicators of liquidity (availability of quotes, bid/offer spreads, dealer overhang) most structured products did have a liquid secondary market in the bull years. Even in Europe, where the secondary market was much less developed and there were no indices, billions of euros worth of structured products were traded every day. Clearly some sectors, especially CDOs and sub-investment grade tranches in all asset classes, were never liquid in secondary, but I don’t think anyone ever claimed they were.