http://biz.yahoo.com/ap/081109/aig_bailout.html

NEW YORK (AP) — American International Group Inc. late Sunday was reportedly near a deal for a revised bailout package from the U.S. government that would make borrowing terms easier for the troubled insurer.
A proposed $123 billion bailout package would be replaced with a new $150 billion package, according to the Wall Street Journal.

Details of the arrangement could be announced as early as Monday, when AIG is scheduled to report its third-quarter results, the Journal said. The plan reportedly would replace an $85 billion two-year loan with a $60 billion five-year loan at a lower interest rate.

The government also reportedly would inject $40 billion into AIG in exchange for preferred stock.

AIG representatives were not immediately available for comment.

———————

Unreal.

In the case of the synthetics, the risk of the entire portfolio is tranched into several different derivative swaps which assume the losses from the underlying portfolio according to a set of priorities. The middle tranches were credit enhanced by the most junior tranches and the senior most tranches were credit enhanced by the middle and junior tranches. The risk of the underlying portfolio, that was to be transferred through the tranches of risk in the synthetic (Merrilly BUYING PROTECTION in the form of SCDO tranches), could be assumed by Merrill (the assumption of risk here is actually a hedge for the risk already sold through the SCDO protecton) in either cash or derivative form. In the case of cash form, the underlying bond is bought and financed with repo and any interest rate risk is delta hedged with swaps. The result is a replicated strategy for a default swap. IN the case of derivative form, Merrill simply sold protection on the reference portfolio on which it (partly) bought protection in tranched form.

Now, in Merrill’s case, they fell in love with the “fees” and league table status of the deals they were printing even though they retained lots of “tail risk” in the form of super senior tranches. Risk models assigned a very low probability of loss and also a very low VAR to these tranches and hence the notional amount that could be retained against a modest amount of risk capital was quite large. It is only once the underlying assumptions deviate substantially from the initial assumptions (in the adverse direction, ie market implied default rates increase) do they begin to require substantial amounts of risk capital. At that point, they would have realised that they had a problem but they would also have realised that their positions were to big to place – the act of placing that much risk would drive the losses on the retained risk to unacceptable levels.

In Merrill’s case, they did this with CDO-squareds in which the underlying were (already leveraged) mezzanine (or middle) tranches of CDOs backed by home equity or subprime mortgages. The models underestimated the volatility of the underlying and hence the loss distributions assumed to “haircut” or determine the amoutn of required risk capital for the super senior tranches greatly underestimated the risk of loss on these tranches and certainly underestimated the mark to market volatility of these tranches.

Risk management was clearly lacking. The notional sizes involved in the retained super senior tranches was in the tens of billions. Even if one believed the risk was minimal, the nominals should be limited by the prospect of that assessment being wrong since these tranches by their very nature provide a small income compensation for low-probability but HIGH SEVERITY potential events.

I have written too much.

Hayek

Hybrid CDOs have supersenior tranches too. I suspect that GM is not distinguishing between hybrid and purely synthetic CDOs. It’s my understanding that most CDOs issued after 2005 were hybrid.

I don’t think they really wanted the senior tranches. I think they were stuck with them.

They were booking record profits in 2007. Who was buying record amounts of CDOs in 2007?

You are selling ‘insurance” on iffy mortgage bonds in 2007 (the default swaps) and no longer paying AIG to cover it. You also have a certain amount of cash flow coming in from the underlying mortgages.

If the housing market bottomed out (and their have always been some thinking it would) they would have a great 2007 and never look back.

Obviously that is not what happened.

My only question is: who bought the junk tranches then? Since CDOs where often made up of some pretty junky pieces to begin with, I am curious who would be buying the junk of the junk.

I suspect that they gave it away, and just made assumptions that made the accounting look good. But that would be a diff ult trick to pull off.

What will happen when Ford, GM and Chrysler become extinct?