I Do Not Understand This

Felix Salmon writes:

Market Movers by Felix Salmon: Did Anyone Other Than Citigroup Have Liquidity Puts? Why hasn't this "liquidity put" thing gotten greater play? I never made it down to the 11th paragraph of Carol Loomis's interview with Bob Rubin, where she introduces the concept more than 900 words into her article. Floyd Norris, today, does a bit better, taking less than 400 words to get to them. A gold star, then, should go to Peter Cohan of BloggingStocks, who read the Loomis article, realized what he was looking at, and promoted the liquidity puts to headline status back on Monday.

Liquidity puts are a big thing, and indeed it seems that they were more or less singlehandedly responsible for the downfall of Chuck Prince at Citi. Basically, Citi told the world – and kidded itself – that it had sold billions of dollars in CDOs to investors. In reality, however, those CDOs had "liquidity puts" attached, which essentially transformed the CDO "sales" into glorified (or debased) repos. Any time that the investor found the CDO difficult to sell – and CDOs are always difficult to sell – he had the option to put the CDO back to Citi at par. And that's exactly what happened; it was those return-to-sender CDOs which were written down the same weekend Prince resigned.

I do not understand this. Which CDOs? What obligations, exactly, did Citigroup assume? Is this something I already know about under another name? It's not as though Loomis or Norris are comprehensive in their explanations...

Comments

[Ken Houghton]

Best guess is that they sold a bundled asset: the CDO and a Credit Default Swap.

It may not be so much a "liquidity put" as a "credit put"--the security is no longer investment-grade, so it gets taken by the seller of the swap.

That's without looking at either of the articles, or any details provided except what you posted, but it seems the way to bet.

[wkw]

it doesn't say *which* CDOs, but that doesn't seem as important. most likely, as the market got flooded with CDOs, Citi attached the put option in order to make their CDOs more attractive to investors. since they obviously mis-calculated the systemic risk, they also mis-calculated their potential liabilities if the CDOs went south and they had to buy them back. hence, the write-downs.

either that, or i'm missing something.

[c.l. ball]

_Fortune_ define the liquidity this way: "The put allowed any buyer of these CDOs who ran into financing problems to sell them back - at original value - to Citi. The likelihood of the put being exercised, however, was regarded as extremely remote because the CDOs were structured to be high-grade entities called "super-senior.""

I am unclear whether Citi is in trouble as the asset manager of the CDOs or is on the hook for the puts as the issuer of CDOs. Fortune says that the CDOs themselves were off their balance sheets, but the puts are the problem.

[Colin Danby]

The super-senior is that part that mystifies me, if they really were super-senior. Historically, across time and place, most mortgages get paid. So even with widespread defaults, the upper tranches should be pretty good. Is it just that the particular pool of mortgages that was being tranched was especially bad? Or is this a liquidity freezeup for perfectly good assets?

[christofay]

I pulled up to the gas station this morning, along with the premium, there was a new grade, super senior. It's made from the really old dinosaurs. It's modeled so that you get 0 to 60 faster, less knocking in the engine, and weirdly enough, greater gas mileage. Gas mileage, who cares about that? I wanted to get an under 30-year-old in the seat next to my 48 yr old body.

The station franchiser is a friend of mine, I asked him how they do it, and he doesn't know.

[gorobei]

Sounds more like a tear-up type thing to me - for a lot of exotic products, the sellers agree to be the buyer of last resort. The seller provides daily, or whatever, prices, and the buyers can tear-up the contract at that price. To reduce credit risk, you can even cash settle at that price. It's a put only in the sense that any offer to buy is a put.

As for "par," I can't believe Citi promised a floor price of the original price. I'm guessing either par was set at 80% or whatever of the original price, or was defined as the daily quote.

Financial geeks in the audience may now speculate on ways to prevent the issuer from low-balling the daily price. Feel free to suggest new derivative products on the daily prices.

[Colin Danby]

Sounds like the problem is lousy pools.

This piece, (linked to by the calculated risk blog)

http://www.dallasfed.org/research/eclett/2007/el0711.html

is nice work. Over the years I've seen a lot of good stuff from the Dallas Fed.

[William Newman]

I would be interested to learn more about the definitions involved too. A pity, then, that there doesn't seem to be a chance in hell of figuring it out from what the bank told their stockholders.

As far as I understand it, it's "It's complicated to estimate the value of these puts, which we defined them ourselves for the purpose of convincing quants to buy securities which had the puts bundled in. We're so confident the value is so negligible that we didn't need to tell our stockholders about them at all, much less give them enough details to estimate the exposure themselves."

Maybe it just goes without saying in this sophisticated crowd, but Marc Andreessen ran this under the explicit title "if this is true, someone's going to jail." (skipping lightly over the sometimes-large gap between "should be" and "is")

[jck]

"Is this something I already know about under another name?"
Yes, backstop credit line.
Basically a synthetic cdo will have a trigger in case of rating downgrades in the underlying portfolio and this in turns may create demand for additional collateral to cover the [short protection] cds synthetic portfolio.That will be met by converting a fraction of the super senior swap [unfunded] into a funded cp super senior.
If the cp market is frozen, the backstop line activates.

[lonesome moderate]

I'm not an economists or a finance quant or anything like that, so my apologies for my rudimentary terminology Isn't this just a money-back guarantee?

[wood turtle]

All I can say is, thank goodness not very many people read blogs.

[wood turtle]

They has been pwned.

[Ginger Yellow]

Colin, the problem is that these aren't mortgage backed securities. They're CDOs backed by mortgage backed securities. If the collateral is mezzanine RMBS, then you only have to have relatively low losses (across a wide range of mortgage pools) to completely wipe out the whole capital structure of the mezzanine CDO. Because the correlation between mortgage backed securities has been so high, mezzzanine CDOs were basically binary investments. Either the whole capital structure performs, or (almost) all of it doesn't.

What none of the articles mentions is that if you apply Citigroup's own structured credit research methodology for estimating CDO write downs at other banks to Citi's own exposures, it seems they don't actually plan to write the liquidity put exposure down much at all. This seems to be because the puts were for CDOs issued before 2006, which have much better collateral. Citi's research piece argues for minimal write downs on super senior exposures of 2005 and earlier vintage. You can account for $8.6bn of write downs using Citi's methodology (which admittedly uses a very broad brush) on the non-liquidity put exposure, which fits in the range of $8bn to $11bn Citi gave.

The really interesting question is whether other banks were still using these liquidity puts in 2006 and/or 2007, because those exposures would be subject to much greater write downs, and the banks should have known better by then.