Investment bank retention of subprime risk post 2005 took two main forms. They often kept the residual risk in the underlying loan portfolios (through their ownership of the lenders or through warehouses) and they kept super-senior tranches of CDOs of RMBS. There are complications like Citi's infamous put options, which meant that a lot of risk they had seemed to have transferred ended up coming back on balance sheet when it mattered, but that was the bulk of it.

Previously, AIG and the monolines would write CDS on the super-senior stuff, supposedly removing the risk. When AIG stopped doing so, the banks tried to find other counterparties that seemed to make economic or regulatory capital sense, but where they couldn't they typically figured the risk was minimal so just kept it themselves.

I'd like you to point out where I have defended securitization and disintermediation, much the less "consistenlty". Describing how it operated and why it took hold is not tantamount to defending it. You make a big leap and attribute it to me.

I've been hugely critical of the more extreme versions, CDOs, have said CDS should be abolished, and have pointed out ad nauseum how mortgage securitizations impede doing the right thing, namely one off mortgage mods when the borrower could be viable at a loss that is lower to the bank /trust than foreclosure.

I'm a a great admirer of your blog. Well Done.

Your query about the dynamics of risk and responsibility around AIG and Wall St raises a different question for me.

You have consistently defended securitisation and disintermediation through this debacle. Yet with all your expertise, contacts, resources and time aplenty, you are unable to identify the breakdown of responisbility that underpinned the crisis.

This is not a criticism of you. It is part of an observation that securitisation is inherently opaque and overly complex. It has opacity and irreposibilty risk written into every split in the lending chain.

Is it not time for you to reconsider its rescue in favour of traditional, readily regulatable banking vis-a-vis Glass-Steagall?

David Smith
http://www.themarketstate.blogspot.com

When you pick apart the point of whether the banks actually took the risks that AIG previously had taken in 2006, I did not perceive that as a statement of fact, rather the perspecitive of the AIG traders.

In reality, we know that a lot of paper was sold far and wide, to chumps who wanted to buy high yielding bonds. But the banks had a pipeline, and obviously got caught with a large bag of dreck too.

I thought the piece brought a lot of new perspective to the conversation, namely that Goldman made AIG post collateral on a downgrade. I also get the sense that far from dancing until the music stopped, players wanted to stop but were unable.

Haven't had time to read Michael Lewis's article, but the logic outlined in your post seems quite rational. From my understanding, AIG and the monolines were writing guarantees on the AAA tranches of the subprime securitisations. The mezz tranches were largely stuffed into CDOs for resecuritisation. Because the AAA tranches had credit enhancement in the forms of both subordination of lower-rated tranches and credit guarantees from AIG, they were easier to sell to investors. When AIG stopped providing the guarantees, the AAA tranches could only rely on subordination, and from all accounts, became harder to sell, so the banks retained the AAA tranches from 2006 onward. So, the banks took the risks because they held on to the AAA tranches without CDSs from the insurers. By saying "getting the deal done", it just means that they structured the deal, but they maintained large exposures.

AAA tranche investors are most concerened with correlation risk, where a large number of underlying assets default at the same time. This is what happened to subprime, where geographical diversification did not lower correlation. In any event, I believe all of AIG's subprime exposure was all to the AAA tranches.

But I dunno, maybe Mikey the Lewis will pop up here and 'splain himself.

The first one was the other monolines such as MBIA, AMBAC or CIFG. They may have been well too happy to snatch some business from AIG-FP.

The second one was the "leveraged super senior" structure that was the craze then. Essentially, it meant that hedge funds or HNWI were able to write super senior protection by just putting a small amount of cash as collateral. When the fast market movements generated collateral calls and forced unwind of the positions, banks didn't find anyone to take the position again, and ended up with it.

Caveat : I was not in this particular business, so this is just an educated guess.