In July of 2011, Reuters ran an exclusive entitled "Deutsche's Firing of Top Trader Sparks Probe". The story revolved around Alex Bernand who, prior to being hired by Deutsche Bank (NYSE:DB) in 2006 as Global Head of Credit Correlation, was the architect of Bank of America's structured credit business. Before Deutsche, Bernand was a "rising star in the derivatives world" and authored "The Bank of America Guide To Advanced Correlation Products" in 2004. At Deutsche, Bernand managed a book of derivatives from London affectionately called the "exotics book" and reported initially to Boaz Weinstein, a terribly ironic factoid given Weinstein's role in harpooning another London derivatives trader early in 2012.
Bernand was fired by Deutsche in October of 2009 after a complaint by Matthew Simpson, a former risk manager at the firm who, at the time of the complaint, was working at the bank's New York credit correlation desk. Simpson alleged "substantial anomalies" in the firm's credit default swap portfolio, allegations which, according to Reuters, led Deutsche to "the discovery of improper trading in one of Bernand's personal accounts." In all likelihood, it wasn't Bernand's personal trading accounts that Simpson was concerned about. In fact, Simpson's claims -- which eventually found their way to the SEC in the form of a Sarbanes-Oxley whistleblower case -- revolve around the routine mismarking of credit positions in a deliberate attempt to inflate profits and bonuses, according to the Financial Times.
Enter Eric Ben-Artzi, the PhD mathematician and former quantitative analyst at Goldman hired by Deutsche in 2010 to model risk. On November 4, 2011 Ben-Artzi filed the SEC whistleblower complaint at the heart of last week's controversy surrounding Deutsche Bank and its crisis-era derivatives book.
A common refrain this past week was that the trades at issue "are complicated"; that valuing those positions was a "very complex" exercise. The implication here is that the average investor shouldn't concern himself with such matters let alone opine on the merits of Ben-Artzi's claims; these things after all, are best left to the experts. The problem with this is that Ben-Artzi is the expert.
Ben-Artzi modeled "gap risk" (the point of contention underlying the whole fiasco) at Goldman and based on this experience determined that Deutsche's failure to account for the gap option in its book of leveraged super senior securities (effectively the bank was treating the securities as if they weren't leveraged) allowed the bank to hide some $10 billion in paper losses between 2008 and 2010.
Allow me to demystify this trade a bit and explain why Ben-Artzi is probably right because contrary to what you might have read, this is not rocket science. Sometime around 2005 or 2006, Deutsche Bank bought credit protection from Canadian pension funds then turned around and sold credit protection on some of the constituent companies in the CDX. In the process, Deutsche pocketed the difference between the insurance premium they paid the Canadian pension funds and the premiums they received from the buyers of the protection. So essentially, Deutsche was short credit in terms of its deal with the pension funds and long credit via the protection it sold to hedge funds.
The problem is that the trades with the Canadians weren't fully collateralized -- these were leveraged trades. Specifically, the Canadians were posting collateral worth only 10% of the notional value of the trades. None of this is really a problem as long as some once in a generation crisis doesn't cause credit spreads to suddenly spike; that is, if the waters remain calm, Deutsche sits back and collects the spread on the premiums while claiming its overall position is neutral (sort of).
However, should spreads move dramatically higher (a deterioration in credit), Deutsche only holds collateral representing 10% of the notional value of its short credit position. Meanwhile it has sold protection (gone long credit), presumably for a comparable notional amount, and is on the hook for the payouts at a time when credit spreads are suddenly moving against the long position.
So here is the question: what happens if the Canadians simply say "Hey Deutsche, take that 10% and keep it, we're backing out" while the hedge funds come calling for their profits on their bought protection? The difference in the uncollectible amount from the Canadians and what would theoretically have to be paid out to the hedge funds is the 'gap' -- well, technically the 'gap option' is the amount Deutsche was short via the difference in the collateral posted by the pension funds and the notional amount of the bought protection, but you get the idea. An excellent article on DealBreaker has more on this.
This whole enterprise becomes even more absurd when you consider the fact that as the gap option risk increases, the value of the bought protection rises (the more dicey the situation becomes in terms of credit market stress, both the value of the protection and the probability of the writer walking away increases) so if you, as the buyer of the protection, fail to account for the gap option while simultaneously marking to market the bought protection, the void between theoretical losses and reported earnings grows:
"When credit spreads deteriorate, [Ben-Artzi] knew banks should not just book the mark-to-market profit from the increased value of their protection but also the gap option: the mark-to-market losses associated with the counterparty walking away."
Ben-Artzi determined, based on the model developed by Goldman, that the gap option risk was in this case worth some $10.4 billion, enough which, if recognized, might have caused the bank's tier one capital to fall below the government-mandated 8% causing Deutsche to require a government bailout.
You can see why Ben-Artzi found it unconscionable that Deutsche would not account for this in their treatment of the leveraged trades. It's not a matter of protecting your interests, it's a matter of presenting an honest assessment of your financial condition to shareholders.
Speaking further to the absurdity of Deutsche not including the gap risk in its models, some literature seems to suggest that even a pricing model which accounts for the gap option is insufficient to convey the risk associated with leveraged super senior structures:
"...the standard approach of pricing an LSS transaction as equal to the equivalent un-leveraged protection value less some 'gap risk' is inherently inconsistent and could lead to some unpleasant surprises for LSS issuers." (my note: 'issuers' here refers to protection buyers, i.e. Deutsche Bank)
Even more astonishing is what the bank apparently did to hedge against this faulty hedge -- yes, that's right, this is yet another example of a hedge of a hedge gone wrong. From the Financial Times:
"A person close to Deutsche Bank says the bank hedged its exposure by betting that the S&P 500 stock index would drop."
That's it?! You have tens of billions in theoretical losses on a complex credit derivatives book to protect against and you bought yourself some S&P 500 puts?! This action was apparently taken after the firm determined that modeling the gap risk was "economically unfeasible." In other words: "Yeah we modeled it and it turned out the losses would be huge, so we scrapped that model." This is a modified form of data mining or, in statistical parlance, "a fishing expedition."
Of course the best part of this entire saga is that the SEC's chief of enforcement is Robert Khuzami. As noted by NakedCapitalism, this is a hopelessly ridiculous situation for the SEC as Khuzami
"...was general counsel for the Americas for Deutsche from 2004-2009, so this behavior took place on his watch. Although he has recused himself from this probe, that's inadequate. Recusal does not work when the people working on the matter in the end have the party with the conflict as their boss."
In sum, I would encourage readers to be skeptical of analysis which seeks to exonerate Deutsche on the grounds that marking its book to market would have caused the firm to collapse. Here is an example of this type of argumentation from Felix Salmon:
"...it's fair to say that if you have a broad economic crisis and there's not much liquidity in the credit market, then if you assiduously marked every asset to market, the entire banking system would be insolvent."
"Deutsche was not selling its super-senior portfolio during the crisis, it was holding on to it. Should it have marked the value of that portfolio down by $12 billion on the grounds that mark-to-market rules required it to do so? I have no idea. But here's a certainty: seeing Deutsche Bank take a $12 billion writedown at the height of the crisis would have been almost as bad for the system as a whole as seeing Lehman go bankrupt. The time for kitchen-sink writedowns is when you can afford them, not when you can't." (emphasis mine)
In my opinion, this sort of analysis is egregious. You do not get to choose when to take writedowns based on whether or not you think your balance sheet can withstand the blow. Besides being outrageously counterintuitive, this flies in the face of accrual based accounting and the matching principle. Furthermore, how can investors be expected to make informed choices when the companies they are evaluating are hiding losses behind the scenes because now doesn't seem like a convenient time to recognize them?
There are two takeaways here. First, as I have noted before, the average investor is at a hopeless disadvantage when attempting to evaluate the merits of an investment in a major bank. Second, investors should steer clear of Deutsche Bank's shares. Even if you don't believe the firestorm surrounding this issue will have a meaningful effect on the bank, no one should voluntarily take an ownership stake in a company devoid of transparency on positions as large as that in question.