A default position among what passes for finance cognoscenti in the blogosphere is to argue that media stories pointing up bank improprieties are making a mountain out of a molehill. The form of the argument is usually, “If you only understood XYZ technical issue, this is not such a big deal.” Now that isn’t to say that position is wrong; we’ve more than occasionally made just that type of argument. But if you are going to go that route, it’s incumbent on you to take account of the relevant background; otherwise, whether you intend to or not, your argument can wind up being the equivalent of “Look, over there!”
We’ve seen this type of diversion-as-argumentation take place on the brewing Deutsche Bank scandal over losses that three separate whistleblowers allege that that bank hid from investors during the crisis. The debate has focuses on one position that accounted for the overwhelming majority of the misvaluation, involving so called leveraged super senior trades. In particular, Matt Levine of Dealbreaker and Felix Salmon have both contended that all the bank needed to do was wait and its positions would have worked out. Felix explicitly and Levine implicitly analogize a position carried in the bank’s trading books as a long-term bank asset. Um, trading book positions and loans are chalk and cheese, witness the old Wall Street saying: “an investment is a trade that didn’t work out.”*
There are two problems with that argument. First, it’s a classic example of hindsight bias. Tell me who during the crisis would have assigned a 100% probability to that view, because that’s what this amounts to. It runs afoul of Keynes’ observation: that markets can remain irrational longer than you can stay solvent.
Second, Levine and Salmon say, to use Levine’s turn of phrase, that all the German bank did was ignore the losses until they went away. That is a misrepresentation of what actually happened. Levine and Salmon airbrush out what has been called “the biggest structured credit restructuring in history” and “the largest restructuring in Canadian history.” If the restructuring had failed, Deutsche most assuredly would have had major losses on its hands, not just from being forced to unwind the trades at an unfavorable time, but also from litigation, which is not factored into the whistleblowers’ estimates.
And the restructuring failing was not a hypothetical risk. There was not assurance it would get done, and analysts with no dog in the fight were of the view at the time that it was not likely to be completed successfully.
A short overview: three former employees have charged the German bank with misvaluing various positions during the crisis, allegedly totaling $12 billion at the peak, when by comparison, the banks had roughly $45 billion of equity. The largest single misvaluation involves the biggest risk position in the bank’s trading book, that of a leveraged super senior trade, a type of synthetic CDO that was funded by Canadian commercial paper buyers.
To make things very simple (you don’t need to get deep into the transaction details to understand the restructuring risks) the LSS trade was a CDO. The asset side of the CDO was corporate credit default swaps. For conventional CDOs, the liability side consists of investors who when times were good would take the payments made on CDS (like insurance premiums, remember CDS were tantamount to insurance) and if the deal came a cropper, would make payouts. Conventional deals would have 100% of the possible losses covered (at least in theory) by various investors, usually tranched by payment priority, from equity (highest income/greatest risk of loss) to multiple AAA classes. Even though a deal where the assets were all CDS in theory needed no cash, most had a cash reserve which meant the investors in the highest risk categories paid money, as opposed to merely offering to fund losses. That was so that the higher-tranche investors had some assurance that the first losses could readily be paid.
By contrast the leveraged super senior was “leveraged” because it only had a portion of the potential losses covered by investors. First, it dispensed with insuring all of the default loss; the attachment point was usually 10% to 15%, meaning the first 10% or 15% of any losses CDS weren’t apportioned to investors in the LSS trade. The deal was then “levered” in Deutsche’s deals, 10 to 1. So if the notional value being insured was $150 million, investors would pay only $15 million and that would pay the losses immediately above the attachment point. That meant if the losses exceeded the amount the investors put up, Deutsche would be on the hook (Deutsche internal documents consistently describe these deals as non-recourse, with the non-recourse in bold. Levine spends a lot of time intimating DB thought the counterparties might and maybe did kick in more. The DB discussions are crystal clear on this point, that the only use for the various triggers was not to get more money out of the counterparties, but to liquidate the trades to contain losses.)
The other wrinkle was that the CDS were 5 to 10 years in maturity, while the investors were Canadians commercial paper buyers. Every 30 to 90 days, depending on how the deal was set up, the CP would be “rolled”, meaning the current CP investors would be paid off and new investors would buy new CP. Now since it was possible that there might be some awful event like 9/11, these deals usually had a liquidity guarantee from the bank. Deutsche’s annual report for 2007 confirms it was a liquidity provider on these LSS trades. (Deutsche indicates that its deals had both extendible CP and conventional CP; I’ve not yet seen any breakdown of this aspect of the structure, so if any readers know how Deutsche’s liabilities on these deals were configured, that would be helpful).
What happened was that investors had bought these deals blind. From 2009 article in the Corporate Studies and Law Report:
According to IIROC, very often the information memoranda did not disclose the underlying asset class composition, the asset and liquidity providers, the role of the sponsor, the issuing and paying agent or the distribution agents. In a June 2007 article in the Bank of Canada Financial System Report, the Bank of Canada said that “[t]he fact that securitization is a complicated process involving many participants would seem to argue for a high degree of disclosure. But the market is relatively opaque.”
Would transparency have solved the problem, however? The market was so complex that transparency might have simply confused the issue. When making investment decisions, dealers and investors relied almost exclusively on the credit ratings and the knowledge that the programs had liquidity support. There was no detailed public disclosure of asset class composition by trust issuer to differentiate the asset securitization strategy amongst ABCP [asset backed commercial paper] trusts.
So investors had no way of differentiating among ABCP trusts in Canada, and even if they’d been told the CDS ones were better designed and much safer than the subprime ones that were blowing up, no one was likely to have given that much credence anywhow. When investors are leery of risky, telling them that something is “less risky” is not good enough.
The ABCP market in Canada froze on August 13, 2007. Investors in the US had been fleeing structured investment vehicles another type of special purpose vehicle funded by commercial papers, that often contained large exposures to subprime mortgages. Notice that in Canada when the market halted, there was no way for a Canadian investor in ABCP to know what he’d gotten into.
In theory, the liquidity providers should have stepped up and paid off the investors in the maturing commercial paper. In practice, those liquidity clauses were cleverly drafted so they could shirk that duty. While in the US, the Fed ultimately rescued the ABCP market, the Canadians took the unprecedented step of using bankruptcy laws to restructure an entire market. These negotiations were called the Montreal Accord, since talks started there. The negotiations were fraught; the initial 60 day standstill was extended more than once, until the deal was filed with the courts, just the way a US bankruptcy must be approved by a judge. And remember, just as Chapter 11s are not assured a judicial wave-through, neither was this deal.
Consider this take as of November 21 from David Chang of Clarity Financial Strategy:
When the conduits entered into swap agreements, their counterparties – typically foreign banks – were essentially given first claim on the assets of the conduits in the event of a problem. If the value of the assets backing up their investment deteriorated, they could place a margin call forcing the conduit (or its sponsor) to add more collateral. Conduits would do this by tapping the capital markets for more money, which would have been possible in normal times. But of course now the markets are not functioning at all and the plan is to morph the commercial paper into 10-year notes with floating interest rates.
Under the circumstances it’s a fine idea except for those pesky counterparty agreements in the leveraged conduits: They don’t just go away for free.
The only logical trade would have been to ask the counterparties to drop the right to demand more collateral in exchange for not having to provide liquidity, but as it happens the liquidity agreements were so weak that there was effectively no risk of them ever being triggered. The banks weren’t going to trade something that had very little risk to them for something that exposed them to great risk…Therefore, there doesn’t appear to be any motivation for the foreign banks to play ball, and given that more than half the commercial paper is leveraged, it’s hard to see the Montreal Accord succeeding.
Ah, but the investors did have something they could give that the banks VERY much wanted, a release from liability. The Montreal Accord contained mutual releases save for very narrow types of fraud. This was the biggest bone of contention, since some investors wanted to sue the banks to kingdom come. The accord was approved by lower court, the malcontents appealed over the inclusion of the liability waivers in the deal. The appeals court upheld the original deal, the Supreme Court declined review, and the Plan Implementation Order became effective in January 2009.
This isn’t our only bone to pick with the defenders of the Deutsche books-fudging. We thought it made sense to start with the restructuring not only to illustrate how critical aspects of this history are being airbrushed out but also to remind readers how uncertain this period was. There were serious doubts at the time as to whether official interventions, much the less private pacts, would succeed. Remember, Paulson, Bernanke and Geithner thought they had a deal cinched for Lehman on Sept. 14, 2008 when it turns out they didn’t.
* Levine contradicts himself by saying accounting in unimportant (huh? internal P&Ls, cash flows, all that are accounting. So do you want banks just to pay staff, rent, fly around in planes, book trades and hope that they have enough money sitting around when the bills come due?) and then saying that if DB had understood the risks of the LSS trade, like Goldman, they probably would have done very little of it.
This piece is cross-posted from Naked Capitalism with permission.