No-one knew what was in them -- the value was asserted by the ratings agencies and they were purchased based on expected rate-of-return. Even the people who went short on the market ended up using heuristics: there is a wonderful passage in The Great Short which describes a character who tells his broker that he will short anything bought by one of the people he has just met.
Once the extent of fraud became clear the entire market dumped down to 35 cents on the dollar. The fact that transparency was an issue is also apparent right here, since nowhere near 65% of consumer mortgages failed. But no-one would buy because no-one could tell if any particular CDO was backed by anyone who was still capable of making payments.
With Ethereum and smart-contract based mortgage systems this completely disappears, since the blockchain broadcasts every single payment made in real-time and it is possible for anyone to simply look and see if whether defaults are rising or falling among the mortgages that underpin their collateralized securities.
The fact that CDOs were trading 35cents on the dollar is largely a part of default EXPECTATIONS resetting up. Smart contracts cannot expedite this process of resetting expectations. Ethereum (which, by the way, is too slow to run CDO pricing models) can only broadcast what the latest market price is. Ethereum helps price dissemination no more than a Bloomberg terminal can.
Yes, exactly! But take the next step and ask what sets EXPECTATIONS about the worth of a security in a smart-contract environment? It stops being generalized market fear and it starts being actual DATA about performance.
> Ethereum (which, by the way, is too slow to run CDO pricing models)
You don't run the monitoring software ON the blockchain (you could but it would be inefficient). You run that off-chain using blockchain analysis software.