Here's the recipe for a CDO: you package a bunch of low-rated debt like subprime mortgages and then break the package into pieces, called tranches. Then, you pay to play. Some of the pieces are the first in line to get hit by any defaults, so they offer relatively high yields; others are last to get hit, with correspondingly lower yields. The alchemy begins when rating agencies such as Standard & Poor's and Fitch Ratings wave their magic wand over these top tranches and declare them to be a golden AAA rated. Top shelf. If you want to own AAA debt, CDOs have been about the only place to go; hardly any corporation can muster the credit worthiness to garner an AAA rating anymore. Here's where the potion gets its poison potential. Some individual parts of CDOs are about as base as bonds can be — some are not even investment grade. The assumption has been that even if the toxic waste bonds really stink, the quality tranches can keep the CDO above water. And life goes on.
The problem is that CDOs were untested; there was not much history to suggest CDOs would behave the same way as AAA corporate bonds.
Nouriel is characteristically more blunt. He concludes a recent post on the securitization of subprime lending by noting:
That “toxic waste” of unpriceable and uncertain junk and zombie corpses is now emerging in the most unlikely places in the financial markets.
One of those unlikely places is the money market. Perhaps not directly -- but it sure seems like a lot of financial firms issued commercial paper to finance what look like mini-credit hedge funds. Blogger RIIP:
Banks, investment banks, hedge funds, and PE firms have been issuing commercial paper as funding for carry trades. They put in capital of $100, borrow $2,000 in CP, and invest the proceeds in higher yielding stuff: CDOs, MBS (including sub-prime) and other ABS. So, sports fans, this is why we are all rushing to call our Schwab brokers and blowing out our cash sweep accounts and other money market funds as though they were leveraged Egyptian equity funds. The (ex-)masters of the universe were using the commercial paper market – formerly a way for quaint old fashioned “companies” to get short term funding – as financing for leveraged carry trades. Now some of these conduits (also known as SIVs, structured investment vehicles) may be related to the bank’s operations (e.g., a bank makes mortgage loans and then sells them to the off-balance sheet conduit) but .. a lot of these vehicles … look suspiciously like off-balance sheet credit hedge funds.
Money market funds effectively financed highly leveraged financial structures that bought higher-yielding, but illiquid CDOs -- and in come cases in commercial paper issued directly by a CDO.
Right now, though, no one wants to own this debt. Maturing commercial paper isn't consistently being rolled over. Various investment vehicles -- most recently Sachsen's Ormond quay -- have had to turn to their parent banks for funding and Sachsen's case, it in turn had to turn to a consortium of German banks for funding. Other funds are looking to raise cash as well.
The money market funds don't want to buy the paper of the conduits, and no increasingly investors don't want money market funds either. The result: Treasury bill yields have absolutely collapsed, and the 3m bill-3m libor spread has truely blown out in both the US and Europe. Bloomberg.
``I've never seen it like this before,'' said Jim Galluzzo, who began trading short-maturity Treasuries 20 years ago and now trades bills at RBS Greenwich Capital in Greenwich, Connecticut. ``Bills right now are trading like dot-coms.''
``We had clients asking to be pulled out of money market funds and wanting to get into Treasuries,'' said Henley Smith, fixed-income manager in New York at Castleton Partners, which oversees about $150 million in bonds. ``People are buying T-bills because you know exactly what's in it.'' (Emphasis added)
Call it the downside of complex financial engineering. That engineering took some risks off the banks balance sheet (literally in some cases), but it also means that no one quite knows where the subprime losses are. And there is a suspicion that some of those losses are hiding in funds that haven't offered adequate compensation for the risk.
A few months ago a lot of subprime debt could be packaged into a security that was worth more than the sum of its parts (with a bit of help from the credit rating agencies. And this process was widely lauded.
The IMF argued that the United States unique skill at creating innovative fixed income “product” was pulling in the capital needed to finance the US current account deficit.
The Fed argued that financial innovation allowed the banks to sell risks that they previously might have held on their balance sheet -- though it is also worth noting that the banks themselves were big buyers of MBS as well. Risks were divided and then sold to those best able to manage them.
There were, of course, notes of caution from the Fed. Some warned that many new instruments had not been tested by a real downturn – and hinted that credit spreads might not be commensurate with the risks. But I think it Fed and certainly Greenspan generally applauded these shifts, even if -- as the prescient Gillian Tett noted -- this new financial technology helped make the markets more opaque.
But apparently the Fed's thinking is starting to shift.
Splicing and dicing turned a lot of illiquid loans into a lot of illiquid securities. Some CDOs now seem to be as illiquid as a portfolio of plain mortgages would have been in the pre-securitization days.
More importantly, the Fed now seems to think that the wide dispersion of subprime risk – and the fact that it is now embedded in a broad swath of “structures” – is one reason why the market is now so illiquid. Krishna Guha and Eoin Callan of the FT report:
They [policy makers] believe that markets are paralysed by lack of information as to the ultimate size and distribution of losses – which has contributed to a sudden drying up of liquidity in the three-month interbank and commercial paper markets.
The information problem has two components. First, investors do not know where the losses from subprime – which Ben Bernanke, the US Federal Reserve chairman, suggested last month could be up to $100bn – lie.
Second, they have lost confidence in their ability to value complex structured credit products that include some exposure to subprime bundled up with exposure to other underlying assets. ...
In principle at least, investors can overcome the problem of not knowing where subprime losses lie by investing in a diversified pool of credits. A few of these investments may turn sour, but the portfolio as a whole should not.
However, the uncertainty over where losses lie may be compounded by what economists call information asymmetries and adverse selection. Banks or other entities sitting on large losses may seek funding even at unattractive rates in current market conditions. Those in better shape may hold back, hoping for better times.
If this is the case, it would be unwise to lend to even a broad range of institutions now coming to market.
In other words, no one wants to buy what others want to sell out of fear that those who are selling are selling for a good reason. The lemon problem in the used car market also seems to apply to the creation of the rocket scientsists. No one wants to pay for prime and get a bit of subprime …
While putting a lot of different loans together previously created something that could be sold for more than the sum of its parts, now that structure is likely worth less than the sum of its parts.
The Fed’s solution?
Take the cash flows from CDOS that contain a bit of subprime, which previously had been mixed together with a bunch of other cash flows in order to try to reduce the correlation between the various cash flows, and repackage them yet again – so all the subprime cash flows are once again bundled together.
Ultimately the complexity problem too is solvable.
The financial institutions that created these products in the first place will break up the products into separate income streams investors can understand and can price: subprime, non-subprime mortgages, auto loans, and credit card receipts. " Emphasis added
If the income streams that were bundled together in an effort to reduce correlation and, as a result, created higher rated debt are broken apart, those who want to bet that subprime is trading at too big a discount would be able to do so – and those who don’t want to buy subprime wouldn’t run the risk of inadvertenly buying a structure that includes subprime debt. At least that seems to be the Fed's current thinking.
But I cannot but note the irony of trying to use a new round of financial engineering to reverse the outcome of the past round of financial engineering …
Maybe it would have been easier just to have sold a bunch of subprime mortgages packaged together as a (probably not terribly highly rated) security from the start. And if there wasn’t enough demand for subprime backed mortgage backed securities in their pure form, well, maybe a bit less credit would have been extended.
Do read last week's Guha and Callan piece.
It hints at a potentially important shift in how the Fed is thinking about financial engineering -- as well as the Fed's thinking about how rate cuts would work. The Fed recognizes that rate cuts won't solve information problems. But they could help insulate the economy from the fallout from the financial sector's current difficulties ...