Display:
I don´t think so?

Here is a nice explanation.


Let's jump back 18 months. Huge Investment Banks (HIB, for short) encouraged mortgage banks to make home loans, often providing the capital, and then would purchase these loans and package them into large securities called Residential Mortgage-Backed Securities (RMBS). They would take loans from different mortgage banks and different regions and generally grouped them together according to their initial quality, such as prime, ALT-A and the now-infamous subprime mortgages. They also grouped together second lien loans, which were generally made to get 100% or cash-out financing as home owners borrowed against the equity in their homes.

Typically, a RMBS would be sliced into anywhere from 5 to 15 different pieces called tranches. Ratings agencies would then give them a series of ratings on the various tranches, and who actually had a hand in saying what the size of each tranche could be.

The top or senior-level tranche had the rights to get paid back first in the event of a problem with some of the underlying loans. That tranche was typically rated AAA. Then the next tranche would be rated AA and so on down to junk level. The lowest level was called the equity level and would take the first losses. For that risk, they also got residual funds if everyone paid. The lower levels paid very high yields for the risk they took.

Since it was hard to sell some of the lower levels of these securities, HIBs would take a lot of the lower level tranches and put them into another security called a Collateralized Debt Obligation (CDO). And yes, they sliced them up into tranches and went to the rating agencies and got them rated. The highest tranche was typically again AAA. Through the alchemy of finance, HIBs took subprime mortgages and turned 96% of them into AAA bonds.

He probably gets to 96% because the HIBs didn´t stop with the simple CDO. They also created "CDO squared" (CDO^2). Buying the lower tranches of different CDOs and slicing them again into new tranches. I´ve even read about "CDO cubed" (CDO^3).
And all of them rely on the performance of the underlying mortgages.

The problem for ratings though was:


The ratings agencies used data supplied by the investment banks on what the likely default rates would be. It was like taking an open book test where you get to write the questions. And since home values had only gone up, default rates were low. Of course, the data was from an era when bankers lent money expecting to get paid back.

The situation of course has changed.

Mike Shedlock describes one mortgage pool originated in 2007:


This cesspool from May of 2007, was 92.6% originally rated AAA, even though loans had full doc only 11% of the time. In one year, the pool was 29.07% 60-day delinquent or worse. Of that, 13.87% is in foreclosure and 6.21% is bank owned real estate.

Mind you, if you only own the highest tranche you´re probably still pretty well protected. There are just two problems here as well:

  1. As I understand it, some funds (pension funds?) are only allowed to own "AAA" rated securities. If a CDO - even only the lower tranches - performs a lot worse than the models, the rating agency can downgrade every tranche of the CDO. Which means that the fund now has to get rid of its tranche, typically at a loss.
  2. If you want to really value a CDO, you have to take a close look at the underlying loans/mortgages. That´s probably time consuming, expensive and you´ll need experts for it.

Both of it means that almost everyone is now shying away from buying such securities. Unless they´re sold at steep discounts. Like Merrill Lynch in May 2008 (?) selling $30 billion of such securities at 22 cents on the dollar.
by Detlef (Detlef1961_at_yahoo_dot_de) on Wed Sep 24th, 2008 at 12:24:55 PM EST
[ Parent ]
mmm, differential calculus spewing.

Never my strong suit.

Diversity is the key to economic and political evolution.

by Cat on Wed Sep 24th, 2008 at 06:59:39 PM EST
[ Parent ]
... there is not a heavy systemic risk, because the rating was based datasets when fewer mortgages were as heavily exposed to systemic risks and the systemic risks were not as severe.

You mark to model with an optimistic model that assumes that the risks you can't track account for are a risk of 0, and you are in serious trouble is there is a systemic downturn.

And of course, that is the nature of systemic risk. Mixing together junior derivative tranches of different pools in different parts of the country will do fine to diversify the stochastic risks, and the senior-most tranches will be relieved of almost all stochastic risk.

But if there is a systemic risk that they are all exposed to that they are all ignoring, then the entire pool of junior tranches are at risk of being completely wiped out, and being the senior-most tranche from a pool of junior tranches from pools of mortgages means that almost all the risk exposure is systemic risk. So the higher the grade in the second-level pool, the worst the overvaluing if exposure to systemic risk is systematically underestimated.

And then the junior tranches of those pools of junior tranches are garbage and are pooled together to provide more senior tranches were very little stochastic risk, but the understated systemic risk exposure rising exponentially.

And regulated institutions have that crap in their balance sheets as if they were bona fide AAA grade securities ... that is, institutions that the law still decides after fifty years of financial deregulation cannot be permitted to play high risk games with their funds.


I've been accused of being a Marxist, yet while Harpo's my favourite, it's Groucho I'm always quoting. Odd, that.

by BruceMcF (agila61 at netscape dot net) on Wed Sep 24th, 2008 at 08:01:34 PM EST
[ Parent ]

Display:
Login
. Make a new account
. Reset password
Occasional Series