A blog on financial markets and their regulation
Who creates CDOs today?
December 4, 2008Posted by on
While many think that CDOs and other complex instruments
contributed to the crisis, the US government has been very
enthusiastic in its use of these structures. In fact, if the behaviour
of the US government is any indication, CDOs seem to be part of the
solution. Let me give a couple of examples.
One of the best is the bail out of Citigroup. If one looks at the
term sheet for this transaction, it is clearly a CDO structure:
- First loss tranche (10%) retained by Citigroup. Using typical CDO
numbers, this would be all the pieces rated A and below. But if we
assume that valuations are generous, then this would be just the
equity piece or may be also a bit of the mezannine piece.
- Second loss tranche (1.6%) taken 90% by US Treasury (TARP) and 10%
retained by Citigroup. Depending on one’s valuation assumptions, this
could be the BBB piece, the A piece or the AA piece. This is a thin
tranche implying high loss given default.
- Third loss tranche (3.3%) taken 90% by the FDIC and 10% retained by
Citigroup. Depending on one’s valuation assumptions, this could be the
A piece or the AA piece.
- Senior piece (85%) taken 90% by the Federal Reserve and 10%
retained by Citigroup. Under good old CDO norms, this is the correct
attachment point for a AAA tranche and if you are very gullible, you
could consider this a AAA piece. Since the Fed took this piece, you
would assume that either they were gullible, or that some valuers
there are willing to turn a Nelson’s eye towards this
Now why is such a complex structure required when most of the
parties involved are arms of the government itself and every
reasonable person would agree that all pieces are clearly at
significant risk? Clearly, the government is now abusing CDOs for the
same reasons that Wall Street abused it – to fool oneself or to
As a second example, consider the terms of the
restructuring of the loans to the AIG announced by the New York Fed
yesterday. There is an LLC set up to buy CDOs and this has a complex
- There is an equity tranche of a little under 15% contributed by
AIG and the remaining 85% (senior tranche) is provided by the New York
Fed. Note once again that the attachment point of the senior tranche
is once again the magic number of 15% of the old days.
- There is a complex payment waterfall structure in which the Fed
receives principal before it receives the interest (Libor + 100 basis
points). To me this looks like make believe accounting where the Fed
would be able to claim that there was no loss of principal and only
the interest was lost.
- The profits are shared between AIG (33%)and the Fed (67%). Since
AIG provided all the equity, one can think of this as an implied IO
(Interest Only) tranche that is owned by the Fed.
- The term sheet glosses over the fact that the only source of cash
for AIG to provide the equity piece of $5 billion is borrowing from
the various schemes of the Fed and the Treasury. So the Fed is now
playing the role of a prime broker lending to a hedge fund that picks
up the equity piece of a CDO. What can the Fed do when there are so few
prime brokers still left?
My last example has nothing to do with CDOs and is not a recently
designed instrument, but the consequences of some needless complexity
in the design is showing up only now. Long ago the US Treasury
introduced inflation indexed bonds (TIPS) whose coupons and
redemptions are indexed to the consumer price index. This is close to
a real risk free bond and is a useful asset class for many
investors. It is also useful in creating a market implied estimate of
expected inflation (simply subtract the TIPS yield from the nominal
yield of an ordinary government bond).
The US Treasury however fouled up this simple and elegant
instrument by adding a totally unnecessary complexity when it stated
that the redemption will not drop below par even if inflation over the
term of the bond turns out to be negative. This adds a European put
option exercisable at par at maturity to the instrument. Under normal
conditions, this option is far out of the money and can be ignored. If
one wanted to be more accurate, one could assume a volatility for the
inflation rate, value this put and compute the option adjusted spread
(OAS) for the TIPS.
The problem is that these are not normal times. Some people believe
that the risk neutral distribution of the CPI at maturity is bimodal
with one peak at 80% of current levels and another at 140% of current
levels. Black Scholes valuation is hardly appropriate and nobody knows
what the true value is. What we do know is that the yields on two TIPS
with the same residual maturity have vastly different yields (a spread
of 200 basis points) depending on when they were issued and therefore
how much of inflation adjustment is already impounded in the
principal. Mankiw blog has an excellent discussion
on this issue. Econbrowser also discusses
this and I have drawn on ndk’s comments in that post for the
bimodal distribution mentioned above.
There is an old joke which asks what is the difference between the
godfather and the investment banker. The answer is supposed to be that
the godfather makes you an offer that you cannot refuse while the
investment banker makes you an offer that you cannot understand. The
US government is now very clearly in the business of making offers to
the rich and well connected that the tax payer cannot understand.