Richard Brookstaber: Blowing up the Lab on Wall Street

Richard Bookstaber writes:

Blowing up the Lab on Wall Street -- Printout -- TIME: Looks like Wall Street's mad scientists have blown up the lab again... [by] the alchemy of creating collateralized debt obligations (CDOs) compounded by the enormous amount of leverage applied by big hedge funds.... Here's the recipe... you package a bunch of low-rated debt like subprime mortgages and then break the package into pieces, called tranches.... 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... wave their magic wand over these top tranches and declare them to be a golden AAA rated.... The problem is that CDOs were untested; there was not much history to suggest CDOs would behave the same way as AAA corporate bonds. After all, the last few stress-free years have not exactly provided much of a testing ground.... t's not the first time this has happened.... One August day nine years ago, Russian bonds defaulted. A surprising result of this default was the spectacular failure of Long-Term Capital Management (LTCM), a hedge fund in Greenwich, Conn. Surprising because LTCM had nary a penny in Russian bonds....

A number of hedge funds are failing; others are seeing returns plunge. Among these is Goldman Sachs's flagship Global Alpha Fund, which burned a quarter of its $10 billion value.... Global Alpha was not a player in subprime junk. Indeed, Global Alpha's problems have not come from mortgages at all, but from a portfolio of stocks.

Why does this happen?.... [H]igh leverage. For example... every dollar of investor capital claimed six dollars of positions. This is the dry kindling for a market firestorm. When things go bad for a highly leveraged hedge fund, it gets a margin call and has to sell... prices drop. The falling prices mean a further decline in the fund's collateral.... And so goes the downward cycle.... It can be difficult to sell the stuff that's causing the problem.... So if you can't sell what you want to sell, you sell what you can sell....

As the subprime crisis propagates, it doesn't matter that some instruments are fundamentally strong and others are weak. What matters is who owns what, who is under pressure and what else they own.... A world in which highly leveraged hedge funds share similar strategies makes it inevitable that what we are seeing now will occur again. And the more complex the strategies, the more surprising the linkages that will emerge.... These funds hired the best and the brightest, yet they became embroiled in crises largely of their own making. If it could happen to them, it will happen again. And we'll all share in the consequences. Again.

Blowing up the lab strikes me as OK--as long as it doesn't blow up the economy. If people who took large leveraged positions without sufficient capital find that they have to sell at fire sale prices to those who do have sufficient capital, that is not necessarily a bad thing. It is a bad thing only if the consequences hit the prices of underlying assets, and thus hit spending flows. And so far that hasn't happened.

Comments

[Hal]

"And so far that hasn't happened."

I tend to whistle in the dark too. The darker it is the louder I whistle.

[christofay]

Along with the Greenspan Put there is a Bernanke Call. Word of the Fed intervention leaked and some organizations bought. An insiders' no-risk call. Now if enough people get upset about this, there will be an investigation and a couple of middle-managers will get punished.

The object of the helicopter mission was to save the interconnected and over-leveredged web of hedge funds and investment banks. Whatever organizations did profit from the leak were just doing what they were supposed to do. It's tasting the dinner in the kitchen before it is placed on the table for all.

[Brett McDonnell]

In all the coverage, I have yet to see an instance of a AAA tranche of a CDO, or more generally an asset-backed security, getting hit by defaults in payments on the underlying loans. Has that happened? Everyone now seems to be assuming that the AAA ratings were overly optimistic, but is there yet any evidence for that?

[christofay]

Brett:

There is no bid in the CDO marketplace whether for the less than triple A or for the triple A portion of the CDO. The big four plus BS have stopped offering bids in order to prevent the triple A portion of the CDO getting a market price. This is in order to preserve the made-up model based value they assign to that stuff.

There is no market so there is no default.

The BS funds that blew up where in the higher rated portions of the CDOs. It's a mystery where the less than triple A parts went.

You're not being a cool hand luke, your info set is not complete.

[Gwailo]

"It's a mystery where the less than triple A parts went."

So if the sliced & diced AAA CDO tranches are looking even slightly shakey, doesn't that mean that the lowest tranches have now become pretty worthless? Who, I wonder, held or holds the lowest tranches into which the default risk -- and now the default losses -- got concentrated?

[christofay]

A lot of it was shoveled overseas, check the bail outs of the German banks. Some of it landed in Taiwan, where I live, for example; a bank or two and an insurance company or two have already written down their BS (BS? who names these places?) investments. Re-reference Michael Lewis' Liars Poker where he first blows up a German speaking banker for on-the-job-training. This is sure to help our "soft power" and continued support of the dollar via purchases of agency paper.

"It is a bad thing only if the consequences hit the prices of underlying assets, and thus hit spending flows." Capital One just shut down a mortgage unit and fired 1,900 people, so there won't be any new spending from those folks. That's the good thing of having a biomass of middle managers, they make easy firing targets when the crap hits the stock price.

The big 4 + BS just had a record setting bonus year. What was the bonus profile at the Bush supporting Capital One last year?

In fairness there should be a bail-in first from the big 4 + BS, hedge fund managers and their investors before the Bernanke Fed starts bailing. Rather than a well-time Friday bail out, Bernanke could have pulled an all nighter over the weekend, picture Robert De Niro with a baseball bat in the dinner scene in "The Untouchables."

Sample dialogue, "Now I don't want to see anyone pull a wuss move like BS during the LTCM melt-down. Before I start handing out the liquidity shots, I want to see who has faith in the American system." Fantasy, but fantasy is a frequent topic on this blog.

[Brett McDonnell]

christofay,

You reply "There is no market so there is no default." That strikes me as a non-sequitur, given my question. The tranches work by dividing up payments on the underlying loans to the various tranches. If there are defaults, the lower tranches lose out first. The higher tranches still receive their payments unless enough defaults occur in the underlying loans. My question was whether any AAA-rated bonds had actually had their scheduled payments reduced because of defaults in the underlying loans. I'm not denying that they have, I'm just saying that I can't tell from the coverage that I have seen whether or not that has occurred.

That is a quite different question from what prices are being paid in the secondary markets for the securities. Of course, if the market is working rationally, the two questions should be very closely related. But, it could be that market participants are unwilling to buy AAA bonds even if those bonds have not yet been hit by defaults. That could be happening either because investors are panicking or because they rationally foresee that in the future AAA bonds will be hit by defaults, even if that hasn't happened yet. My question is whether that is in fact what we are seeing. Thus, information on what the secondary market for the securities is doing is not directly relevant to my question.

[Craig Tindale]

Lol yeah whistling in the dark really loud while looking for a sleeping bear with a sharp stick that you knows in there with you somewhere .

true we dont know if AAA payments have been reduced but we do know from insiders that tranches within AAA are misrated and stink!

Brad Setser makes has a good go at it

http://www.rgemonitor.com/blog/setser/211007

[christofay]

Brett, sorry blowing steam through my ears.

You are correct. I have not seen news or Trip A tranches getting ripped on the payments.

this could only be a matter of time and not too many months in the future.

[quartz]

The triple A ratings depended on the subordinate tranches being there to absorb defaults. As defaults have risen higher than expected, and as the value of the collateral has flattened or fallen, the losses have eaten higher into the subordinate tranches than expected, leaving far less credit protection for the AAA tranches. Although it does not appear that those tranches have yet seen payments default, they are in fact now far less "protected" and riskier, no thus no longer AAA, and so worth less. It appears that in some cases defaults are heading to a point in which some of thes tranches can expect losses.

[Ginger Yellow]

Brett asks a very good question, and one I haven't seen a good answer to. There are two elements to any answer. First, interest payments on the CDOs are met out of interest payments on the underlying securities. So long as (enough of) the underlying securities don't default, then interest payments on the CDOs will continue to be met. So far as I know, no AAA rated CDO has failed to make payments. The other factor is what is known as rated overcollateralisation. As explained by quartz, a highly rated tranche will have some considerable protection beneath it. But when the rating of the underlying securities fall, the rated overcollateralisation falls, making it less likely that the CDO's principal can be repaid at maturity and resulting in a downgrade of the CDO.

There's nothing inherently wrong with CDOs. There is a legitimate question of whether or not CDOs backed entirely by one asset class (eg poorly rated subprime mortgage backed securities) should ever be granted triple-A ratings. As in any investment portfolio, diversification is safer than putting your eggs in one basked. But even the most skeptical observer would probably accept that there is some suitably high level of credit enhancement above which a AAA assessment is reasonable. The question is how much enhancement that is, and whether AAA ratings are supposed to stay AAA forever, or move down and up along with the credit cycle even if they don't default.

[nonotyetsure]

I am also wondering about how the ratings have changed on the higher rated tranches, and if that was really unanticipated. I have seen models that include the possibility of default and changes in default hazards in the future. Are the higher rated tranches over-colleralized enough? With stochastic default intensities, there surely is a chance of a downgrade throughout the life of the bond.

The question is whether the realized defaults are higher than in the models' realized paths. But I have seen little evidence about this---and how could you even tell? It's like a weather forecast of a 10% chance of rain tomorrow, and then indeed it rains tomorrow. With one observation, you cannot tell if 10% was a bad weather forecast or not. You probably need many observations to estimate low probability events, even in the best of statistical world. Of course, if you ignored that the same information could cause many mortages across different pools to have high realized default rates--that is, assumed that the risk in defaults could be diversified away--then you truly have a stupid model.

So, are the realized returns to the CDOs to the end of their lives really lower than expected? Were the initial spreads too small on the low rated tranches really too small? What about the high rated tranches? And something I always wondered about: What institutional friction caused there to be a gain in value from splitting up the cash flows in the first place?

I bet yes the returns will be too low---fraud by the lenders, so that the pool characteristics were not exactly as reported. But I have seen no hard evidence that meant that the bonds were mispriced up front. Yet.

I guess we do know that the bonds are priced too low right now, since they seem to be valueless in the marketplace...That's the liquidity issue.

Of course the leakage to the equity market--value stocks underperforming and growth stocks outperforming--was in very few quant equity models. And the value premium histocially was never terrible when the market dropped. Our 'Black Swan.' That's the part that we should really be worried about. Which 'positive alpha' strategy is next?

[tjallen]

Brett challenged whether any 'asset-backed security [are] getting hit by defaults in payments on the underlying loans' - forclosures are up, 3 and 6 month non-payments up, doesn't that bubble up? Non-payment is not just a signal, it is non-payment.

My impression is that over time the mortgages backing the securities became lower quality. Wouldn't that mean things are going to get worse?

The cross-ownerships that normally spread risk are now spreading the contagion.

Investors expect an ideal of friction-free buying and selling of investment instruments. How much of the value of the instrument is its liquidity, as opposed to the fundamental value of the underlying assets? Or, how much value does the investor place in his freedom of action, to instantly buy and sell financial instruments? Lots, apparently.

It used to take months to get your asking price on a house. Now it may take months to get your asking price on your mortgage-backed securities!

[Ginger Yellow]

"Brett challenged whether any 'asset-backed security [are] getting hit by defaults in payments on the underlying loans' - forclosures are up, 3 and 6 month non-payments up, doesn't that bubble up? Non-payment is not just a signal, it is non-payment."

Kind of. The thing is the non-payment is not-paid at the bottom of the capital structure. There is excess spread (the difference between the interest earned by the mortgages and that paid by the securities), cash reserves, equity, then rated securities starting at B or BB. You'd have to have non-payment on 10%+ of the portfolio for months on end before the AAA securities began to be affected, depending on the collateral quality. Moreover, these securities are (usually) structured as pass-through notes, meaning that they don't have fixed amortisation schedules, only an expected rate and a final maturity. So if a loan becomes delinquent, it just delays repayment without causing a default so long as there is enough cash coming in to pay the interest. You only get a default when either enough loans are completely are foreclosed on below par to eat into the principal on the relevant tranche of notes, or if you reach final maturity (which is often 30 years from now) without being able to pay off the notes. This is one of the reasons that ABS is less liquid than corporate or agency bonds. CDOs are a different affair but similar principles apply.