Now That's What I Call Toxic!

by hilzoy

During the past year or so, I have sometimes wondered exactly how toxic all those toxic assets really are. It's hard to tell, since they differ from one another, and are not traded that often. However, the Financial Times (h/t) has some answers:

"In recent weeks, bankers at places such as JPMorgan Chase and Wachovia have been quietly sifting data trying to ascertain what has happened to those swathes of troubled CDO of ABS. [Ed.: collateralised debt obligations of asset-backed securities.]

The conclusions are stunning. From late 2005 to the middle of 2007, around $450bn of CDO of ABS were issued, of which about one third were created from risky mortgage-backed bonds (known as mezzanine CDO of ABS) and much of the rest from safer tranches (high grade CDO of ABS.)

Out of that pile, around $305bn of the CDOs are now in a formal state of default, with the CDOs underwritten by Merrill Lynch accounting for the biggest pile of defaulted assets, followed by UBS and Citi.

The real shocker, though, is what has happened after those defaults. JPMorgan estimates that $102bn of CDOs has already been liquidated. The average recovery rate for super-senior tranches of debt – or the stuff that was supposed to be so ultra safe that it always carried a triple A tag – has been 32 per cent for the high grade CDOs. With mezzanine CDO’s, though, recovery rates on those AAA assets have been a mere 5 per cent.

I dare say this might be an extreme case. The subprime loans extended in 2006 and 2007 have suffered particularly high default rates and the CDOs that have already been liquidated are presumably the very worst of the pack.

Even so, I would hazard a guess that this is easily the worst outcome for any assets that have ever carried a “triple A” stamp. No wonder so many investors are now so utterly cynical about anything that bankers or rating agencies might say these days."

If I'm reading this right, within four years of being issued, two thirds of these CDOs are in default, and their recovery rates are very, very low. That's just staggering. It's actually hard to understand how the banks managed to do this badly: you'd think they could have done better hiring people off the street and paying them to put all those nice little loan documents into piles at random, or tossing mortgages down the stairs and bundling them based on how they landed. They certainly didn't need to hire people with advanced math degrees and pay them seven- or eight-figure salaries to get these kinds of results.

And how about those ratings agencies? They would have done a better job using a Magic 8-Ball to rate the CDOs. ("Signs point to junk!") I have been hearing for years and years about how the financial services sector pays such exorbitant wages because the people who work there are so immensely talented that they are cheap at $50 million a year. I never particularly bought that line before. But I never imagined that all those Masters of the Universe would do quite this badly. If we had paid them $50 million a year to go far, far away and leave our financial system alone, it would have been a bargain.

The column ends with a very important observation:

"Those American officials who are implementing flashy new “stress tests” of banks would do well to take note."

Word.

Comments

[Debra]

Put this way, just what kind of a diploma do you need to play Russian roulette ?
I thought so.
With economic theory as the new religion, too many people have been bowing down and not asking pertinent questions.
A good entry level economic theory book allows you to conclude that predictions on stock exchange activity rely on a crystal ball.
And since I'm an idiot in economics and concluded this, it's rather troubling that so many more intelligent people didn't do so...

[Tony P.]

$450 billion, you say? That's about:

$1500 per man, woman, and child in the US;

$4500 per household in the US;

$450,000 for each of the richest 1% of households in the US;

$45,000,000 for each of the richest 0.01% of households in the US;

$1.13 billion for each of the Forbes 400.

I'm looking at that range of numbers with an eye to this question: who bought those CDOs? Who paid the $450 billion? Was it a few very wealthy clients of those Masters of the Universe? Or was it a vast number of us, through pension plans, credit union deposits, and so forth?

--TP

[Steve J.]

They certainly didn't need to hire people with advanced math degrees and pay them seven- or eight-figure salaries to get these kinds of results.

The Quants didn't make that kind of money, the (relatively) innumerate managers of the Quants did.

[TJ]

They certainly didn't need to hire people with advanced math degrees and pay them seven- or eight-figure salaries to get these kinds of results.

Yeah, they did. If you want to pull off the crime of the millenium, you hire Moriaritys, not average schlubs. The complexity is there for a reason; it obscures the fact that their assumptions were worthless.

Never assume massive incompetence when malevolence is simpler.

[cleek]

It's actually hard to understand how the banks managed to do this badly:

isn't this where the infamous formula came in? it told everyone that these tranches were safe because the different classes of assets of which they were comprised weren't correlated strongly enough. and the ratings agencies giddily agreed.

[John Thullen]

As d'd'd'd'dave might whine, eat the quants.

An unregulated, shadow banking system has collapsed.

The privation will not be measured in months or years, but in generations, like the Great Depression.

If I were a Republican economic strategist with their views on what caused us to get out of the Depression, I'd start finding a war candidate for 2012.

A third World War should do the trick. Maybe Bobby Jindal can declare war on the Indian sub-continent.

I calculate that an invasion of the west coast of India would kill off 50 or 60 million otherwise unemployable mortgage bankers in both countries.

The radiation alone will empty those foreclosed homes, saving local municipalities the sheriffing costs.

Hopefully, the enemy (for the Republican Party, that would be everyone) will bomb inner-city hospitals first, saving the rest of us from the burden of healthcare for the poor.

Wars, like Depressions, are part of freedom. They can be fun and they stimulate the movie industry so that out-of-work folks can work as extras.

We should do nothing to prevent or ameliorate them.

[MadRocketScientist]

One way it got so bad:
http://madrocketsci.blogspot.com/2009/02/why-understanding-math-is-so-important.html

[low-tech cyclist]

They certainly didn't need to hire people with advanced math degrees and pay them seven- or eight-figure salaries to get these kinds of results.

I have a PhD in math. I had no idea that such lucrative work was available for someone with my credential. I'd have been more than willing to work for half the salary these guys were getting, and I'm sure I could have produced better results.

[Alchemist]

Wired had a great article on the equation used to make these bad loans.... I'm not sure how to link here but you can google it easily. 5 minute read, stunning that 450 billion dollars were lost on a single oversimplified equation.

[Mr Blifil]

Incomptetence is an unrealistic explanation. This was fraud, pure and simple.

[That Settles It]

t's actually hard to understand how the banks managed to do this badly: you'd think they could have done better hiring people off the street and paying them to put all those nice little loan documents into piles at random, or tossing mortgages down the stairs and bundling them based on how they landed. They certainly didn't need to hire people with advanced math degrees and pay them seven- or eight-figure salaries to get these kinds of results.
=======

Oh, but they did. And they have to pay them seven- or eight-figure BONUSES to hold onto them (lest they leave in a fit of pique and all get jobs at Mickey D's?), so that they can what? keep on getting these kinds of results?

And with taxpayer bailout billions, too.

It's Midas in reverse, it's the gift that keeps on giving, and it's just plain good to be king.

The Internet Said It
I Believe It
And That Settles It

[Omega Centauri]

I'm having trouble fathoming this. If a CDO is backed by a house, even if the bank had to dynamite the house, and sell the recovered materials, and land, I can't imagine getting only $.05 on the dollar!

[ericblair]

Incomptetence is an unrealistic explanation. This was fraud, pure and simple.

I agree: it doesn't take a towering genius to think "hmm, these correlation coefficients are pretty important, I think we'd better do a lot more work in statistically characterizing them and run some sensitivity analyses to see what could go wrong". To just slam a number in there and dump it on the client is absurd, and should have gotten the analyst shitcanned post-haste if, of course, management wasn't in on the scam.

[John, Paul, George and Django]

who bought those CDOs? Who paid the $450 billion? Was it a few very wealthy clients of those Masters of the Universe? Or was it a vast number of us, through pension plans, credit union deposits, and so forth?

A lot of them were sold in Europe and Asia, who bought them on the basis that the US mortgage market was rock solid, and of course they were rated AAA.

That's why Iceland has gone bankrupt.

I'm no lawyer, but I'd say that Iceland would have a pretty good lawsuit against the SEC for its bogus "AAA" ratings.

[Troy]

I'm having trouble fathoming this. If a CDO is backed by a house, even if the bank had to dynamite the house,

Try packaging a collection of 2nd mortgages in Las Vegas, Florida, or California.

The 80/20 was a pretty recent innovation, ca. 2002-2003. The 20% junior liens from 2004-2007 are simply screwed in most areas.

[mdh]

"They certainly did need to hire people with advanced math degrees and pay them seven- or eight-figure salaries to get away with these kinds of results."

just my 2c.

[Eric Martin]

I'm no lawyer, but I'd say that Iceland would have a pretty good lawsuit against the SEC for its bogus "AAA" ratings.

Does the SEC rate such instruments?

[Nathan]

But I never imagined that all those Masters of the Universe would do quite this badly.

So I shouldn't invest with Grayskull of Eternia any more?

[cleek]

I'm not sure how to link here

here it is.

:)

[Mike Schilling]

Why did they need people with advanced math skills? It's a variant on the old joke: To err is human, to really %@#$ up requires a computer.

[hairshirthedonist]

cleek, thanks for the link. Great article. I assume this was what Alchemist was referring to at 9:58. This jumped out at me:

The CDS and CDO markets grew together, feeding on each other. At the end of 2001, there was $920 billion in credit default swaps outstanding. By the end of 2007, that number had skyrocketed to more than $62 trillion. The CDO market, which stood at $275 billion in 2000, grew to $4.7 trillion by 2006.

Compare the ratios of CDSs to CDOs. The periods discussed for each are off by 1 year, but still. There was ~3.35 times as much money in CDS in 2001 as CDOs in 2000 but 13 times as much for 2007/2006. What is the point of allowing people to gamble this way? I can understand off-loading risk if you're actually holding the debt instrument, but I can't understand letting third parties just place bets on mortgages or bonds that they don't have any direct interest in. Shouldn't people try to make money by actually doing something instead?

[Brian Schmidt]

I'm missing something - the default rate on the CDOs exceeds the default rate on the mortgages that back them. Even if you assume the underlying property has zero value (probably only true for second mortgages, and only in some areas) for the defaulting mortgages, what about the others that are still paying.

In what market is the top two-thirds of mortgages defaulting 68% of the time?

[Dave]

Eric, it was the ratings agencies and the hammer will fall on them soon enough IMO. It'll be damn near impossible to prove fraud unless some smoking gun arises. I suppose it's possible but no one's found it yet.

Tony P., institutions did buy them. Pension funds, mutual funds, money market funds, etc. were buying AAA-rated CDO tranches. They also lost big time when the securities were downgraded because by law, many of these investors are unable to invest in anything other than AAA-rated bonds. They had to sell in a market where everyone was selling. They took a bit of beating although not as bad if they were holding that paper today.

Ericblair, common sense would dictate that but since you wouldn't have had any historical data to support your claims, your investment banker clients would have disemboweled you, hung you out to dry, gone to your supervisors and said you had no idea what you were talking about and your bosses would have agreed and capitulated to them.

It's not that I don't agree with what you are saying but you have to consider the environment that a lot of this was being done in.

As for fraud, good luck. It's hard to see where anyone can make a case for the Wall Street firms or the ratings agencies unless someone can find internal correspondence suggesting otherwise. Yes, there was a lot of stupidity in play but I don't know how high someone can take fraud charges.

By the way, you guys ought to know that CDOs consisted largely of the first-loss positions of other pools of mortgages. They took BBB (and may lower) rated bonds from other pools, dumped them into the CDO. I'm still amazed why no one couldn't see the obvious with respect to how little protection the AAA-rated tranches had compared to other mortgage-backed securities.

[Dave]

Brian, it's a cash flow allocation issue. You can have a 1% default rate on the underlying mortgages and a 100% default rate (or loss of principal) on a CDO if the CDO's underlying collateral is that specific 1% of the cash flows being generated from the underlying mortgages.

[hairshirthedonist]

Looks like madrocketscientist was hip to the wired article as well. (Didn't want to leave anyone out.)

[cleek]

I'm missing something - the default rate on the CDOs exceeds the default rate on the mortgages that back them.

here's a detailed explanation of CDOs and their relatives.

short answer: a CDO isn't just a bunch of repackaged mortgages. they are very highly-leveraged instruments, built from other assets which themselves are leveraged from sliced and diced bundles of mortgages. once mortgages started going bad, all that re-leveraged leverage brought everything else down, too.

(OT: any news on moving to WordPress? been getting a lot of errors while trying to comment the past few days)

[Ginger Yellow]

"I'm having trouble fathoming this. If a CDO is backed by a house, even if the bank had to dynamite the house,"

"I'm missing something - the default rate on the CDOs exceeds the default rate on the mortgages that back them. Even if you assume the underlying property has zero value (probably only true for second mortgages, and only in some areas) for the defaulting mortgages, what about the others that are still paying. "

It's all about leverage and attachment points, folks. A straightforward subprime RMBS transaction, even a really bad one, will have losses of, say 40%. But because the transaction is tranched, the bottom of the capital structure eats those losses. And CDOs of ABS were bakced by the bottom of the capital structure (or at best the AA tranches) of lots of RMBS. So it's entirely possible for every underlying tranche in a CDO of ABS to have defaulted with no recoveries,wiping out the CDO entirely, even though other tranches of the same RMBS have substantial recoveries.

Incidentally, most of the time the super senior tranches where most of the losses have been made (because they were the bulk of the capital structure), were not actually sold. Instead the banks which arranged the deals kept them, sold the lower tranches, and bought protection from an insurance company to effectively remove them from their balance sheets (certainly from their capital calculations). Hence AIG's problems and the fall of the monolines, which in turn brought all these super senior tranches back into the picture.

[Anarch]

I have a PhD in math. I had no idea that such lucrative work was available for someone with my credential. I'd have been more than willing to work for half the salary these guys were getting, and I'm sure I could have produced better results.

I do too, I did know, and in fact a number of my friends went and did just that. I'll be honest, the grinding poverty and "lifestyle" of graduate school made me jealous of their wealth and success. And I'll be even more honest and admit that, for me, the one bright spark in all this tragedy is the thought that, finally, they're falling back to my level.

[Plus, those innumerate hedge fund managers? Many of them were Type-A assholes I knew in college, and you have no idea how gratifying their comeuppance is to me.]

Let me also add one other factor not mentioned in the Wired article: hubris. Last summer one of my friends left town to become a financial analyst on Wall Street, using sophisticated modelling techniques developed by his college roommate. At a party just before he left, I caught him earnestly explaining to another friend how (what we now know were the beginnings of) the financial troubles were a completely anomalous event; and about how those models were almost always right, and they had only failed because of an incredibly unlikely series of events that no-one could have predicted.

I wanted to punch him in the face.

Never mind the fact that those events were in fact predicted. This is the thing that the Wired article missed: they knew the model didn't work. They being the quants, they being the managers. They knew it didn't work; they just convinced themselves that the events in question could never happen, and if by some miracle they did, they convinced themselves they could survive them. So -- and this is the part that burns me more than I can say -- they never looked at what would happen if it failed.

First rule of being a mathematician: when someone hands you a theorem, figure out why it doesn't work. Second rule: if it's true, figure out why it can't be applied to a broader context. That's just what you do. That's how we're trained. But these yoyos decided that because the risk was epsilon, the penalty for failure was irrelevant. As if epsilon were zero and therefore the negative outcomes need never be calculated. As if they could not fail.

They believed they'd conquered risk.

And now we're all paying for their sins.

[Different Eric Blair]

The mortgage banker's association noted as long ago as 1997 that these subprimes were failing at the rate of 10% or more a year. And not just default, foreclosure. If the graph was correct, 16% were in foreclosure in 2003. The graph only listed percentages, not numbers, but its not hard to see where one eventually ran out of loans to keep the ponzi scheme going. All the while, moody & poor and all these other rating companies kept saying these were a good deal. And I will bet cash money that in the end it will be found out that those ratings could be bought.

[ThatLeftTurnInABQ]

I'm missing something - the default rate on the CDOs exceeds the default rate on the mortgages that back them.

Good links have already been posted by others above, so this is probably just piling on, but here's another explanation by Sam Jones at ftalphaville explaining the devil's alchemy shop that CDO issuers where playing in and how it blew up in their faces, based in part on a series of posts by Felix Salmon made late last year; here is a link to the derivatives topic where the latter can be found. I won't post links to all of them because the kitty doesn't like excessive linkage - scroll down to the late Nov - early Dec time frame and look for posts about CDO overcollateralization, synthetic bonds and super-senior tranches.

[ThatLeftTurnInABQ]

Less technical, but more colorful: Paul Volker Godwins the Quants.

One of the saddest days of my life was when my grandson – and he’s a particularly brilliant grandson – went to college. He was good at mathematics. And after he had been at college for a year or two I asked him what he wanted to do when he grew up. He said, “I want to be a financial engineer.” My heart sank. Why was he going to waste his life on this profession?

A year or so ago, my daughter had seen something in the paper, some disparaging remarks I had made about financial engineering. She sent it to my grandson, who normally didn’t communicate with me very much. He sent me an email, “Grandpa, don’t blame it on us! We were just following the orders we were getting from our bosses.” The only thing I could do was send him back an email, “I will not accept the Nuremberg excuse.”

Ouch. That's going to leave a mark.

H/T Paul Wilmott (which also contains a notably colorful smackdown)

[Ginger Yellow]

"The mortgage banker's association noted as long ago as 1997 that these subprimes were failing at the rate of 10% or more a year. And not just default, foreclosure. If the graph was correct, 16% were in foreclosure in 2003."

That's kind of missing the point. Nobody claimed subprime loans were as good as prime loans. But the idea was that you could tranche deals with enough credit enhancement that it was (supposedly) as unlikely for the the most senior tranche to default as the (differently enhanced) senior tranche an RMBS backed by prime loans. A foreclosure rate of 16%, back in 2003, meant actual losses of around 4%, because you could recover a decent amount in foreclosure. That's what's known as expected losses, and the bottom 4% or 5% of the capital structure would be known as equity and held by the originator or sold to specialist high risk investors. Only tranches above that 4% or 5% were sold
and would only suffer losses in a severe downturn.

Several things happened next which turned everything sour.

First, house prices started going up dramatically, year after year. Instead of suggesting there might be a bubble, people started to get overconfident, as they do in a bubble. This contributed to things like the mistaken assumptions about correlation mentioned in posts above.

Second, new, riskier products were devised such as Option ARMs which would be much riskier in a downturn than more conventional subprime loans, and on whose behaviour in a downturn there was little to no historical data.

Third, Wall Street stupidly and greedily invented CDOs of ABS to leverage the subprime RMBS and arbitrage the difference between the yield on underlying triple-B RMBS tranches and the higher rated tranches which made up the bulk of the capital structure of the CDO. CDO technology isn't necessarily bad, but it only works when the correlation between the underlying assets/sectors is well understood and accounted for. Correlation is notoriously hard to calculate/estimate, and pretty much everyone screwed up really badly here, primarily by including assets all from the same sector. A really bad idea.

Fourth, investors who didn't really understand securitisation started buying these deals in huge numbers, because they offered a chance to get highly rated bonds and high yields in a time when credit spreads were tumbling and yield was hard to find.

Fifth, the latter development created a huge boost in demand for subprime RMBS, which in turn allowed underwriting standards to slip and meant that subprime mortgages became a huge part of the overall US housing market.

Sixth, this macro shift left the US economy as a whole far more vulnerable to the underperformance of certain products, and helped push correlation to 1. The assumption had been that there was only weak correlation between different housing markets in the US. That was no longer the case, because badly underwritten subprime loans permeated the entire economy.

Sixth, when the bubble burst, as they all do eventually, people discovered all too late that actually subordinated tranches of RMBS are highly correlated, leveraging highly correlated assets will cause massive losses in a downturn, and securitisations need a lot of expertise to analyse which in most cases they didn't have.

Seventh, panic and chaos.

[Izak]

I think what is missing in this analysis is the derivitives. You have the CDO which is backed by a tangible asset (though so chopped and processed it is difficult to reconstitute which assets correspond to which CDO's), but you also have the Credit Default Swap that is the "insurance" against the CDO failing. Now the derivitives market is completely unregulated and CDS were sold to parties not involved in the original CDO. Additionally Credit Default Swaps were sold many times over for the same CDO's. And the CDS's are not backed by tangible assets but by leverage (debt).

What I still haven't gotten answered is: who made the smart bets on the Credit Default Swaps (i. e. who without owning any CDO's bet that the CDO's were crap) and are these people still collecting from institutions that should be bankrupt (i. e. AIG).

Any financial institutions that are taking bailout money should be defaulting or in a holding pattern on any CDS payments. Otherwise this is a transfer of taxpayer money to speculators.

I think the talk of executive bonuses is a smoke screen for the real scandal.

[rea]

I'm having trouble fathoming this. If a CDO is backed by a house, even if the bank had to dynamite the house, and sell the recovered materials, and land, I can't imagine getting only $.05 on the dollar!

I'm missing something - the default rate on the CDOs exceeds the default rate on the mortgages that back them.

Doubtless there are others here that understand this better than me, but from what I've read, the problem is essentially that they pyramided these things. In other words, not simply securities backed by mortgages, but securities backed by securities backed by securities backed by mortgages . . .

If I am misunderstanding, I'm sure someone will tell me.

[russell]

It's hard to see where anyone can make a case for the Wall Street firms or the ratings agencies unless someone can find internal correspondence suggesting otherwise.

Does this count?

Any request for loan level tapes is TOTALLY UNREASONABLE!!! Most investors don't have it and can't provide it. [W]e MUST produce a credit estimate. It is your responsibility to provide those credit estimates and your responsibility to devise some method for doing so.

"Loan level tapes" here means "documentation on the underlying loans".

From here, h/t Yves.

I keep wondering when the "are you freaking kidding me?!?" moment is going to kick in and we start sending these guys to jail.

I am, personally, so far doing OK. I've lost something like 40% of the value of my 401K but I'm not retiring for a while, so I'm ignoring it. My daily life is, very thankfully, still kind of fine.

If I had lost my house, or a more tangible form of savings, or if my wife or I were sick and couldn't get medical attention, I would seriously be looking for someone to beat up.

I'm a big guy, it would hurt.

I'm probably marginally more irascible than the typical American, but I bet not by much.

The "musn't upset the bankers" approach is not getting it done. The horsesh*t Geithner is peddling, where every week brings a fresh new plan to throw money down the toilet, is starting get old. And it's only been a month.

These people thought they could make a whole lot of money for themselves, and did not give one flying f*ck about the consequences for anyone else.

We owe them nothing.

"Stress testing" is going to be a BS dog and pony show. If we continue as we are, we are going to continue to, literally, pour good money after bad. It won't help, because the banks are fundamentally insolvent. Not all of them, but the ones we're throwing tens of billions of dollars at, yes, they are freaking broke.

The FDIC takes over insolvent banks at the rate of a couple a week. Next up should be Citi.

Take them over, give the management a nice, polite three months to clear out their offices, bring in some competent people whose goal in life is not to have the biggest d*ck on the planet, and clean house.

The sooner the better. I'm pissed, and I still have my job and my house. If I was one of the 600K folks who lost their jobs in the LAST WEEK, I'd be way more than pissed.

[Izak]

In other words, not simply securities backed by mortgages, but securities backed by securities backed by securities backed by mortgages . . .

I don't think it is securities backed by securities, but derivitives backed by leverage. The unregulated derivitives market is why the housing bubble is threatening to bring down the global financial market. Originally, derivitives were a financial instrument to limit risk. Say I purchase a mortgage and at the same time I purchase a Credit Default Swap against that mortgage as insurance. I pay a quarterly fee on the CDS as insurance that the mortgage doesn't default. This is well and good and makes sense. I have an interest in the mortgage performing, but I am hedging my bet with the CDS.

Now the bad part: the derivitives market is unregulated. I do not need to be an interested party in the mortgage to purchase a CDS on the mortgage. I can speculate that the mortgage will default (because I know the mortgage owner is using his credit card for his monthly payments and periodically re-financing to pay his credit card). The financial institution I enter into the CDS with does not need the assests on hand to cover the CDS in case the mortgage defaults, they can use leverage to issue the CDS.

The ability to speculate many times over is what has brought the housing bubble/mortgage crisis to OMFG proportions. This is the magnifying affect.

[JanieM]

nice, polite three months to clear out their offices

That takes politeness too far. Inform them quite civilly that they have three hours to clear out, and if they don't, their fancy stuff will be dumped out on the curb, or better yet auctioned off to help some of the folks who have lost their jobs.

And don't wait even the three hours, in fact don't wait three seconds to take away their access to company computers.

[ThatLeftTurnInABQ]

I keep wondering when the "are you freaking kidding me?!?" moment is going to kick in and we start sending these guys to jail.

The metaphorical explanation:

This will happen when the music finally stops and we find out that most of the chairs are already gone (they were stolen) and start fighting over who gets what little is left in the way of furniture (the ottoman and the umbrella stand) while the room fills up with floodwater.

Translation:

Right now we are papering over the true losses with borrowed money, propping up the zombie banks with funds which are only available because interest rates are still very, very low. That is why Joe6Pack isn't rioting in the streets. The reckoning will come when nobody wants to loan us any more money at dirt cheap rates.

Ancien Régime France provides a rather direct parallel with the parlous state of our finances. No need to worry though, because Calonne Geithner will fix everything.

[radish]

I don't know that the Big Math was really necessary, but I can certainly see how you'd need to pay any mathematician you came into contact with big money, to keep them quiet if nothing else. Because the inadequacy and eventual catastrophic failure of formulas like Mr. Li's is something that's very well known and taken for granted in every other discipline (besides economics) that deals with complexity.

The likelihood that this kind of approximating won't eventually go spectacularly wrong is somewhere near the likelihood that you'll be able to walk though a wall because quantum physics doesn't actually rule it out. Basically, these approximations are only useful for describing variables in a closed linear system at a particular point in time. They're guaranteed to go wrong when applied to only a subset of the observed variables, or to an open system, or to a nonlinear system. And for procedural reasons models tend to go wrong even if you resample and revise regularly. Though of course you don't know when they'll go wrong.

The simplest expression of this is the three-body problem you learn about in high school physics. You can estimate that the locations of three concentrically orbiting bodies are very tightly coupled at time t, based on an observation at time t, and you can observe that at time t+n body A is at location L. Your ability to predict the actual location of bodies B and C is still going to drop dramatically as n goes up, because the degree of coupling varies nonlinearly.

The coupling coefficient you calculated between A and B at time t was only true for time t, and might be completely different at time t+n.

Lots of people know about this, including the people who designed the models. Even behind a technically correct thicket of greek letters, nested brackets, and summarized references to other thickets, the fatal flaw behind the whole exercise would have been immediately obvious to anyone who started out in, say, astronomy.

[rea]

I don't think it is securities backed by securities, but derivitives backed by leverage.

So, in other words, I got the terminology wrong, but the part about pyramiding was right.

[laxel]

Ginger Y: "It's all about leverage and attachment points, folks."

I'm sure most of you have seen this animated illustration of the whole economic collapse (and many probably have an even more sophisticated understanding of it)... but this was a helpful way to visualize it.

Also, if I lost my job and can't pay my bills, can put in a request to have my credit score calculated by the same geniuses who gave AAA ratings to these CDO's? I'll wager small fish like me won't get breaks like that.

[Dave]

laxel,

The presentation is good but it doesn't really demonstrate how bad the toxic assets really are in the way that Ginger Yellow explained above. It's a bit muted in that respect but everything else seems to be fine and I wouldn't discourage anyone from watching it based on that criticism alone.

[cleek]

The "musn't upset the bankers" approach is not getting it done. The horsesh*t Geithner is peddling, where every week brings a fresh new plan to throw money down the toilet, is starting get old. And it's only been a month.

and how.

and while i could understand if this was happening because Geithner is a dumbass and really is trying to help out his Wall Street buddies, Obama doesn't seem like that kind of guy. that makes me worry that what we see as burning money by the boatload is really an effort to hold back something truly catastrophic that we don't know about... yet.

[bartkid]

>It's actually hard to understand how the banks managed to do this badly

No.
Cue teh google to find that John Dillinger quote about the easiest way to rob a bank is to run a bank.

[laxel]

Probably so Dave, just thought I'd link it for non-experts like myself... guess I could also link Billmon's take on why Big S***Pile is what it is, getting deeper into the toxic nature of it. I don't have the expertise to judge all the specifics, but have come to trust certain writers like Billmon, and many here, when it comes to the arcana of wonkery.

From Hilzoy's post: "you'd think they could have done better hiring people off the street and paying them to put all those nice little loan documents into piles at random, or tossing mortgages down the stairs and bundling them based on how they landed."

They should have got that stockpicking lizard who outperformed the market back during the dot com bubble... I hear he still has a strong following in Minnesota.

[ThatLeftTurnInABQ]

and while i could understand if this was happening because Geithner is a dumbass and really is trying to help out his Wall Street buddies, Obama doesn't seem like that kind of guy. that makes me worry that what we see as burning money by the boatload is really an effort to hold back something truly catastrophic that we don't know about... yet.

Caveat: This is nothing but pure speculation.

I think the US banks are being propped up, and eventually will be more or less recapitalized by the taxpayers, because they have major counterparties who are sovereign states. Obama and Geithner are sitting on a geopolitical powder keg. If the US banks are allowed to fail and bondholders take a big haircut, the governments of one or more nations in key parts of the world will suffer such huge losses that they will either resort to extreme nationalism (= war with somebody) to maintain their legitimacy, or risk being toppled by populist revolts and replaced by God-knows-what.

I'm thinking that this includes at least 3 states which have nuclear arsenals. And I don't mean North Korea.

[Izak]

Crisis of Credit Visualized was good as far as it went, but it didn't go into Credit Default Swaps in any detail at all. Suspiciously so. The major point with CDS's is that they are unregulated. They have magnified the issues with the housing bubble to huge proportions. And no one is investigating who is making money from the Credit Default Swaps (trust me, some one is).

[laxel]

"The major point with CDS's is that they are unregulated. They have magnified the issues with the housing bubble to huge proportions."

And far beyond housing, if I'm not mistaken. I mean, trillions in obligations for (quite literally) bets made on the failure of bonds, delinked from actual ownership of the bonds, passed around between around between banks who have no incentive or compulsion to reveal the magnitude of liabilities to other banks, shareholders, regulators, or the public. I don't think it's suspicious that the artist who did the video didn't mention this in full detail in a ten minute primer, after all, I think it took This American Life something like 2 hours to describe the issue.

Off topic:

"Yes, as through this world I've wandered
I've seen lots of funny men;
Some will rob you with a six-gun,
And some with a fountain pen.

And as through your life you travel,
Yes, as through your life you roam,
You won't never see an outlaw
Drive a family from their home."

[(The Original) Francis]

This is a great thread. I just have a few points to add.

1. CDSs are insurance on the CDOs. The seller of CDSs was AIG; the buyers were the big banks (Citi etc.) Citi had, over the years, paid so much out in salary and dividends that its CDS comprised a large part of its asset base. The rest was gone. Spent. Vanished into college education, expensive vacations and mortgage payments on Long Island houses.

(Note: AIG could, and did, sell the insurance to people who didn't hold the underlying asset. Instead, the counterparties were just betting on changes in the risk. This is one way that even more leverage came into the system -- people were borrowing money to bet with AIG on small changes in the price of insurance.)

2. So, if AIG failed to pay, Citi (and others) would all have been insolvent.

3. So, what we're doing is propping up the worldwide financial system two ways: first, by funding AIG's payoff of the CDSs, and second, by directly injecting new capital into the banks.

4. The alternatives to the slow-bleed approach are all scary. No one really knows what would happen to 401(k)s, pension funds, and the non-financial components of the economy if the shareholders (common and preferred) of the big banks were wiped out and the bondholders took a massive haircut. The financial system has been so important and so profitable for so long that the ownership of those enterprises is everywhere. Trying to deflate the bubble too quickly could have unintended consequences.

5. Regulatory capture. The only people who know enough about the worldwide financial system to manage the deflation of the bubble have spent their life in it. They have an inherent commitment to the status quo and are most afraid of the possibility of unintended consequences.

6. I'm a water lawyer, not a banker. All of the foregoing could be wrong.

[russell]

I'm thinking that this includes at least 3 states which have nuclear arsenals.

Yikes.

Well, OK then.

But can we at least prop them up *and* throw them in jail?

[ThatLeftTurnInABQ]

But can we at least prop them up *and* throw them in jail?

How about this: we weld the doors and windows shut on their Wall St. office buildings and turn them into high-rise minimum security prisons for white-collar criminals. They can stay at their desks (and we'll have a security checkpoint to allow for inbound pizza delivery and Chinese takeout) but they can't leave. Sort of like a Green Zone in lower Manhattan.

Given the hours that some of the traders work, they might not even notice the change for a least a couple of months.

[Tony P.]

6. I'm a water lawyer, not a banker. All of the foregoing could be wrong.

The question is, are you a gambler? I mean, if CDS were pure bets on CDOs, then the CDS market was nothing but a zero-sum gambling syndicate.

I know nothing of the niceties of gambling beyond what I've gleaned from The Sopranos, but I gather that a bookie who can't pay off on winning bets is in big trouble. Not from the cops, but from the aggrieved "winners". In the present case, that would be the people who bet correctly that CDOs would tank. I take it some of those people were owners of the CDOs hedging their positions; others were straight-up punters. Their problem boils down to having placed bets with bookies (like AIG?) who never had the cash to cover the action. So why are we not seeing mobs of bankers in suits, wearing brass knuckles and carrying baseball bats, working over other guys in suits and burying them under bridges in Jersey?

Less poetically: why are we not seeing more intramural violence among the Masters Of The Universe? Is it because both sides in this Mob War are playing us civilians for suckers?

--TP

ps, to TLT: you have an astonishingly good idea there, doctor.

[dmbeaster]

Indeed, a great thread.

To just add to the detail, at its core this is just an ordinary real estate bubble. It happened in the 80s also. It happened in Florida in the 20s. Two factors have, however, changed how this one behaves.

The first is the extent to which all of the exotic security instruments have leveraged the bet on ever increasing housing prices. The resulting financial instrument bubble is massively larger than the underlying real estate bubble, and that is the key to understanding why this situation is so out of control.

Second, the financial bubble enormously pumped up the real estate bubble in an evil feedback loop. It enabled massive amounts of money to fund housing purchases on ever easier terms, and skewed the lending incentive to create product no matter how nonsensical the loan. It became all about the origination fees and the bundling fees for securitized debt.

Housing prices predictably zoomed, which fueled the bubble, but to an extent never experienced before. Imagine if the 90s dot.com tech stock bubble was also fueled by a massively larger derivatives market that was speculating on the tech stock run-up.

[A.J.]

If the US banks are allowed to fail and bondholders take a big haircut, the governments of one or more nations in key parts of the world will suffer such huge losses that they will either resort to extreme nationalism (= war with somebody) to maintain their legitimacy, or risk being toppled by populist revolts and replaced by God-knows-what.

more like, replaced by Godwin-knows-what.

(Does Godwin's law apply when discussing the political dangers inherent in economic destabilization?)

[hilzoy]

TLT: "I think the US banks are being propped up, and eventually will be more or less recapitalized by the taxpayers, because they have major counterparties who are sovereign states."

That has been my best guess as well. I haven't written about it because it is pure, unadulterated speculation on my part. But it's my best guess.

[Hans]

I used to think the financial wizards were stupid. And then I realized two things.

Even if they take 5 years off from academia to double their salary, get laid off and go back, they still make out.

You have to be very smart to "mathematically prove" you have re-packaged $1M in net value as $10M.

[OCSteve]

TLT: Just to echo many others, I appreciate you walking us all through this stuff. I’ve learned a lot from you with regards to this mess.

[Ugh]

I think the US banks are being propped up, and eventually will be more or less recapitalized by the taxpayers, because they have major counterparties who are sovereign states.

What I hate most about the Obama administration is that they make me feel like a wild-eyed conspiracy theorist.

[Dave W.]

The March 9 edition of Time that just came out has an article on how the cascading tiers of CDOs increased the losses for one example CDO. Unfortunately, the online version seems to be lacking the graphic that shows how a 4.4% default rate cascades up to a 59% loss rate of the Jupiter High-Grade V bond, but the text of the article is available here.

Short version of the graphic: the original Mortgage bonds have a 4.4% loss rate (so far). The lowest tranch of these was used to create CDO No. 1, which has a 14% loss rate based on that 4.4%. Then the lowest tranch of CDO 1 was used to create CDO No. 2, which has a 36% loss rate, and a middle chunch of that creates CDO No. 3, the Jupiter V, with a loss rate of 59%. Because Jupiter mixed and matched chunks of 223 other bonds/CDOs, credit rating agencys considered the apparent diversity and gave 93% of it a AAA rating back in March 2007.

[jmt]

"Toxic Asset" is an oxymoron.

[CAMP]

Ugh, Manhattan is a very small island. Only Geithner, Paulson, Bernanke and Lloyd Blankfein were in the room on Sept 17 when the FED took title to AIG. Blankfein wasn't there out of public interest. It works like this: if I sell you a put on CITI at $40 I can only lose $40 but if I sell you a CDS on CITI I can lose an infinite amount. The thing goes up in price if the news gets worse. The market piles in. The rating agency ratings are now viewed as worthless so the CDS price is viewed as a "market call" on bank credit quality which makes every wholesale bank customer in the world nervous. So the daily bad news hype machine keeps pumping up the CDS. Planes? who should care really. AIG sold the CDS in a unit separate from their highly profitable insurance business. The AIG total loss was estimated to be $ 40BLN in September. Last count our loss is $ 160BLN and counting. It may also explain was Giethner was so irreplaceable.

[CAMP]

There are CDS markets now on the State of California bonds and the Country of Ireland bonds. These unregulated 'insurance' contracts are deadly to the writers. Good thing its the government who owns the writers now. But the buyers will make a great deal of money as they drive multi-billion dollar enterprises like GE and BAC into the ground. Charlie Gasbag on CNBC thinks the GS guys (Margaret Carlson calls them Government Sachs) are all geniuses and "God bless America." Santelli wants us to be angry with the guy down the street whose getting a deal from the government. I think Hank Paulson's tax forgiveness from the government was $ 48 mm the day he picked up his Treasury id card. Same with Kashkari and the 150 guys he's hired at Treasury.

[Michael Cain]

“Grandpa, don’t blame it on us! We were just following the orders we were getting from our bosses.” The only thing I could do was send him back an email, “I will not accept the Nuremberg excuse.”

Indeed. It's not the military, and it's not wartime, and failing to serve doesn't get you imprisoned or executed. At some point, you have to vote with your feet: "Here's my two weeks notice, I'll be pursuing other opportunities." I'm willing to grant exceptions in extreme circumstances, but every story like this I've seen can be reduced to "Yes, I knew my work was being used unethically, but the pay was just so damned good..."

[Bernard Yomtov]

Their problem boils down to having placed bets with bookies (like AIG?) who never had the cash to cover the action. So why are we not seeing mobs of bankers in suits, wearing brass knuckles and carrying baseball bats, working over other guys in suits and burying them under bridges in Jersey?

Because, as they say, dead guys are lousy payers. Better to hope you can figure out a way to squeeze the cash out. It's not like the deterrent effect is going to be valuable.

[laxel]

One thing I'd like clarification on, I thought credit default swaps were mainly centered in the financial arena (big bankers and hedgers) as opposed to the traditional insurance sector. Is this roughly correct?

I understand these are instruments that have no true home, but thought their most destructive use was not coming from an entity like AIG as much as it was in the hands of folks at Sachs and Lehman and Stearns... BTW, I mostly lurk on account of that axiom about opening the mouth to prove the fool.

[Ugh]

There are a couple big problems with the CDS market as it emerged. First off is that the people selling CDS's are essentially selling insurance. However, the sellers were not regulated as insurers, which is critical. There is a reason why insurers are heavily regulated, and why AIG's "insurance" business are still around and profitable, and it's to prevent people from selling insurance who don't have the ability to pay on the policies when they come due. AIG sold these policies and essentially assumed they would never have to pay off on them and so didn't bother to set aside funds for the purpose. (see this article)

Second, the people purchasing the CDSs were not required to have an insurable interest in the underlying securities, i.e., they were not required to actually own the assets they were buying insurance on. This is a basic requirement of insurance, whether it be home, auto, or life. The reason for this is simple: if I buy homeowners insurance on my house, I (generally) don't have any incentive to see my house burn down, given all that I will lose in the process and despite the fact that I will be compensated by the insurance. OTOH, if, say, someone who lives across town buys insurance on my house, they generallly would like to see my house burn down as they lose nothing and get a big payout. This insurable interest requirement does two things: it limits the amount of insurance being sold in the market, and it limits the incentives for people to bet that any one particular insurance policy to pay off.

These two things allowed AIG and many other players to sell an unlimited amount of insurance to anyone, and they did. And when things went bad the predictable happened: AIG was unable to payoff on the policies it wrote, thus placing all the people who purchased the policies in jeopardy of going bankrupt as well. Which is why the gov't is dumping cash into AIG, not so much to save it, but to save all of its counterparties (as noted in the linked article). And we're left holding the bag.

[Bernard Yomtov]

There is a reason why insurers are heavily regulated, and why AIG's "insurance" business are still around and profitable, and it's to prevent people from selling insurance who don't have the ability to pay on the policies when they come due.

Like CDS sellers, option sellers and futures traders are also essentially selling insurance. The exchanges on which these transactions take place have various well-established mechanisms to guard against what Ugh describes.

Whether those specific ideas could be used on a CDS exchange I don't know enough to say, but the lack of any constarints is just silly. Casinos know enough to limit the size of bets and the number of players who can bet on any one spin. AIG didn't think of that.

[Ugh]

Like CDS sellers, option sellers and futures traders are also essentially selling insurance. The exchanges on which these transactions take place have various well-established mechanisms to guard against what Ugh describes.

That's a good point, I've never really thought of a put option as insurance, though it does have similar characteristics to CDS's. As you note, the exchanges on which these things trade provide important safe-guards, including nearly instantaneous pricing as well as limited contracts (i.e., you can't buy a put on every single outstanding share of GE). Transparency is key.

The CDS market, OTOH, lacks all of these things, and was used in certain instances to re-create/duplicate bonds ad infinitum. Equity markets can do this as well, through total return equity swaps, where neither party owns the underlying equity.

[ral]

Insurance, the White Man's Burden

SEAGOON:
You mean you're offering me free of charge the deeds to the English channel?

GRYTPYPE:
He heard you Moriarty.

MORIARTY:
Do you accept the English channel then, le channel englais?

SEAGOON:
Yes. I only hope I can live up to it.

GRYTPYPE:
I'm sure you can Neddie. However, one slight formality Neddie. For your own protection of course, the jokal style of protection, you must insure it lad.

SEAGOON:
Insure it against what?

GRYTPYPE:
Fire Neddie

[Neddie buys the insurance and then says to the audience]

SEAGOON:
Dear listeners. These men think I'm a fool. Little do they know that the moment their backs are turned I'll be down to that channel, set fire to it and collect the forty-eight thousand nicker!

[Ginger Yellow]

"So why are we not seeing mobs of bankers in suits, wearing brass knuckles and carrying baseball bats, working over other guys in suits and burying them under bridges in Jersey? "

Because most of the bankers aren't exactly in a position to throw stones. And besides, as long as the government is propping up AIG, they're golden.

A central clearing counterparty would have, in theory, prevented many of the problems that we've seen in the CDS market and obviated the need to bail out AIG (or at least the apparent desire to do so at any cost). In a centrally cleared system, if your counterparty goes bankrupt, the capitalised central clearing house, usually owned and funded by market participants, picks up the tab. So, in theory, your CDS is solid protection against whatever credit event you're worried about. Now, personally I'm not convinced that a central clearing house would have had enough capital to weather this particular storm, but the situation would have been a lot better - for a start, if its capital were threatened, there would be an obvious, market based and broadly equitable way to shore it up without having to prop up failed institutions.

Now, obviously, that's not all that's needed - it's absurd that insurers didn't have to put capital against CDS, and even more absurd that they didn't in practice (there's a difference between economic capital and regulatory capital - that's what was being arbitraged by the banks).

[Bernard Yomtov]

GY,

I'm not convinced that a central clearing house would have had enough capital to weather this particular storm, but the situation would have been a lot better - for a start, if its capital were threatened, there would be an obvious, market based and broadly equitable way to shore it up without having to prop up failed institutions.

My understanding, though, is that all trades go through clearing house members, who thus have a pretty strong incentive to make sure traders are able to meet their obligations. Of course they also have margin requirements and the like to keep things from getting out of bounds. It seems as if the lack of good price information on CDO's would interfere with that.

[Dave]

As you note, the exchanges on which these things trade provide important safe-guards, including nearly instantaneous pricing as well as limited contracts (i.e., you can't buy a put on every single outstanding share of GE). Transparency is key.

Transparency especially with respect to counterparty risk.

Now, personally I'm not convinced that a central clearing house would have had enough capital to weather this particular storm

I know I'm speaking in a hypothetical context but in a situation where there would have been a central exchange, would the CDS market ever gotten to this size without raising any concern.

Based on BY is saying above, if the members of the exchange would have observed the investment banks writing a large quantity of CDS contracts on the CDOs they were issuing, would they not have been able to step in and address the issue before it got completely out of hand?

I know it's all backward looking and hypothetical but it still seems pretty important to me because I would guess that one of the steps taken moving forward is to require some kind of exchange for credit derivatives.

[ThatLeftTurnInABQ]

OCSteve,

Thanks for the kind comments. Today is going to be a busy day so I probably won't have time for any substantive exchanges, but I wanted to say thanks before this thread gets too stale. I've been trying to encourage more cross-pollination between the political blogs and the econ/finance blogs, and it feels like this is happening; if I can take some small credit for that then it has been time well spent.

The other thing I'll claim credit for is suggesting to Brick Oven Bill that his talents might be better employed in the snark-infested waters over at Balloon-Juice. He seems to have taken that suggestion to heart and the results have not been disappointing.

[CAMP]

But the game continues. You can hold a position in GE-CDS without owning GE or its bonds. In fact, you can have an active interest in GE failing. You can borrow the money to buy the CDS to take a position in GE failing, for leveraged returns. For instance, Bill O'Reilly can comment every night from his perch on FOX, that GE is badly managed. In fact, he does. Even as they pass along all the negative commentary about the banks, the presenters on CNBC don't seem to realize their own company is being played. The specs who own the CDS are busy making money at the expense of the taxpayer while driving multibillion dollar enterprises that took generations to build, into the ground. Could the Treasury close down this casino? Well, they could announce that CDS will only pay out if you demonstrate exposure to the underlying equity or bond. The scramble to buy back GE, BAC and so on would put quite a bid into the equity market. It would be the equivalent of requiring a company to borrow the shares before shorting them.

[Ginger Yellow]

"I know I'm speaking in a hypothetical context but in a situation where there would have been a central exchange, would the CDS market ever gotten to this size without raising any concern."

I doubt it would really have raised concern, given market psychology. And the whole market size thing is a bit of a red herring, if you ask me. The main problems were that a) these were largely unfunded, undercapitalised bets, and b) people didn't put nearly enough weight on counterparty risk. Now obviously there were other issues (eg there was a large backlog in settling CDS until recently), but those were the ones that made the market "toxic". It doesn't really matter that there could many times more CDS notional than underlyings outstanding. What mattered was that people weren't allocating capital to those CDS and that both the underlying asset and the CDS were massively mispriced on a risk adjusted basis.