The FDIC’s Rosy, Theoretical, Misleading Lehman Resolution Counterfactual (or Why TBTF is Still TBTF)

The FDIC has released a document that purports to show how it could have successfully resolved Lehman Brothers using its new Title II resolution authority granted under Dodd Frank.

All I can say is that this is an interesting piece of creative writing. The Lehman counterfactual rests on a series of assumptions, which as I will discuss shortly, look pretty questionable. The most charitable assessment one can make comes from a famous exchange between two technologists. Trygve Reenskaug says: “In theory, practice is simple.” Alexandre Boily asks: “But, is it simple to practice theory?”.

But some longstanding Administration cheerleaders have jumped on the bandwagon, arguing that “pundits” have asserted “without evidence or analysis” that the resolution authority can’t work. That’s pretty amusing, given that Shiela Bair herself concedes, per the Financial Times, that the resolution authority will not work on a major international bank with retail and investment banking operations:

But for the new regime to work, financial groups would need to be “resolvable” – capable of being wound down. That required upfront change, Ms Bair said, so that an investment bank could be hived off more easily from a depository bank. Like the UK’s banking commission, she believes separate liquidity and capital could provide part of the answer….

Big US institutions identified as “systemically important” will have to prepare living wills to describe how they could be broken up in the event of collapse. Mr Johnson and Ms Bair believe some are too complicated and will need to be simplified.

So the Lehman report, if you take it at face value, is applicable to Morgan Stanley and maybe Goldman (presumably, the FDIC already had sufficient tools to resolve even a very large, domestic, commercial banks like a Fifth Third or even Wells Fargo). The FDIC demand for ring-fencing or better yet a breakup is presumably directed at JP Morgan, Ciitgroup, and Bank of America.

As an aside, it’s funny how all critics are lumped into the “no evidence” camp, when yours truly, Satyajit Das, who has worked on every major international financial bankruptcy in the last 20 years, and Josh Rosner have offered specific evidence and analysis. But then again, I don’t consider myself to be a pundit, so perhaps I am not included in this broadside.

Let’s look at why the FDIC’s counterfactual does not add up:
Assumes everything is governed by a single, uniform legal framework. This is fantasy for a global institution. For example, many if not most derivative contracts with counterparties outside the US would be governed by English law and be exclusive of non-exclusive English jurisdiction clauses. A counterparty may not be bound by a US court, let alone the procedure proposed. Satyajit Das offered these comments via e-mail:

The entire document proceeds on the premise that everything is liquid and trades and can be valued. This is unlikely in practice.

There is no recognition of how “set-off” clauses and laws work in different jurisdictions. Also there is no recognition of how collateral held against trades would work.

Foreign jurisdictions and institutions are increasingly becoming alarmed at the US approach to bankruptcy including:

(i) an attitude that US bankruptcy laws apply globally with far reaching consequences.
(ii) in Lehman, the court has made decisions which are viewed as “home town” ruling designed to favour US creditors, including not recognising foreign court decisions under normal “comity” doctrines, such as in the now infamous “flip” clause litigation.
(iii) many foreign countries are proceeding to implement a strategy of “subsidiarisation” – in effect, if you operate in my country you will need to set up a separately capiltalised and “ring fenced” entity to avoid precisely the problems set out above. India has done this and in my view China and a whole lot of other emerging market countries will do this. This will reduce the importance of US financial institutions and markets, which post SOX have lost ground in any case.

Egregious underestimation of Lehman losses. This part is just plain dishonest. The FDIC report reveals a major blind spot: it focuses solely on asset side mis-marks, and ignores the liability side.

Earth to base: Lehman was so desperate to dress up its balance sheet that it was overpricing highly visible transactions like SunCal and Archstone to create an illusion of solvency. Mismarking liabilities, particularly derivatives, is far less less obvious and was certain to have taken place. And the last sighting from the bankruptcy trustee Alvarez & Marsal shows losses vastly in excess of what the FDIC reports. This was our summary of the latest sighting in the Financial Times:

Take the creditor claims of $250 to $350 billion. Against that we now have assets we will generously peg at $60 billion. So we now have a $190 to $290 billion shortfall. The midpoint is nearly double the lass loss estimate.

Of that total, bankruptcy overseer Alvarez & Marsal has tried to claim that $50 to $75 billion was due to the disorderly failure. Even if you accept the high end of their range, you get $115 to $215 billion in losses. But Alvarez & Marsal has reason to exaggerate the “disorderly BK” losses. The experts I am in contact with who are working on the BK say that $15 to at the very top $30 billion in losses is attributable to the rushed process. That gives a range of $160 to $260 billion in losses to unsecured creditors.

By contrast, the FDIC in the report tries to claim the losses were only the asset side overvaluations of $50 to $70 billion of troubled assets and that the losses would have been only $40 billion. This is a gap of roughly $120 to $220 billion. We are supposed to trust them based on their willingness to engage in phony math like this? Yes, the Lehman entities had franchise value that was lost in a liquidation but the scale of the losses being realized demonstrates that the black hole in Lehman’s balance sheet was far greater than the FDIC is pretending that it was.

Assumes unrealistic 90 day preparation time. In this case, it starts with March 2008. With an investment bank or dealer firm, the FDIC is not going to have anything remotely approaching that sort of runway, even if the receivership was cooperative, which is unlikely. A firm under stress is going to be a hotbed of rumors. The sort of proctological exercise implied in a 90 day resolution plan can’t possibly be kept under wraps. And once a rumor gets out, creditors and counterparties will not have much faith in an untested process, particularly since, as Josh Rosner has pointed out, the fact that they might be subject to either a bankruptcy or a Title II unwind greatly increases uncertainty. In every other walk of life, a business is subject to only one possible resolution regime, not two.

The FDIC document asserts that the market would not have been alarmed by a heightened FDIC presence at Lehman because they’d presumably be rooting around another firm or two aggressively. Who are they kidding? Any unusual FDIC activity means a firm is at risk and no counterparty would want to be exposed (the paper does concede there could be “signaling problems”). Bear failed in a mere ten days. The idea that the FDIC will have the opportunity to pre-plan sales is highly unrealistic.

Indeed, consider what would have played out with Lehman. Fuld was firmly in denial that his bank was in trouble, to the point that he undermined negotiations repeatedly by insisting on too high a price and with the Korean Development Bank, insisting that they buy the entire firm when they were on the verge of doing a deal to buy only good assets (the FDIC asserts that Fuld would have thrown in the towel earlier, given the lack of a bailout option, but given Fuld’s refusal to hear various authorities tell him Lehman would not be rescued, the assumption that he would have been more rational with a different fact set seems optimistic). Given the need to get three regulators to agree to a Title II resolution (the Fed, Treasury, and FDIC), how quickly would they have agreed with Fuld aggressively protesting to each that he needed no help? And with someone like Fuld fighting hard to go it alone (and thus not cooperating with a resolution), the regulators would have had to remove him forceably, again making a long runway sales process impossible.

And let us charitably assume that the FDIC manages to get its full 90 days. The rumors would be all over the firm. The best people would be looking to depart, and senior people often market themselves with their teams. The risk of a 90 day process is the best teams would exit, diminishing the value of the franchise. Why should competitors buy large operating businesses when they can cherry pick the best people? Ironically, the ferocity of Fuld’s delusions probably kept staff in place; with examiners on site, it would be hard for a CEO to keep reality distortion working. In addition, Lehman did use very high levels of restricted stock in its bonus scheme to reduce mobility; given what happened to the stock, it proved not to be much of a golden handcuff.

The paper also assumes a resolution effort would have started in March 2008. I don’t see any basis for that argument. The officialdom went promptly into “Mission Accomplished” mode post the Bear rescue. Even though the SEC and Fed did increase supervision of dealers, including having the SEC task one staffer to Lehman and the Fed assign two on site, the numbers were grossly inadequate for a firm of that complexity and confirms a lack of recognition of how fragile the firms and financial system remained post Bear.

And note that John Mack was behaving precisely the same way as Fuld, insisting he could raise capital when all his peers were certain he was a goner. And guess what? He did get Mitsubishi UFJ to come in at the 11th hour. So other CEOs could convince themselves that they were acting like Mack, not Fuld, in insisting that they did not need help.

How likely is consensus among three regulators in the face of vociferous opposition and stonewalling? At a bare minimum, it will slow the process, meaning any resolution will take place after the patient has deteriorated further. As Bob Teitelman of The Deal noted:

Who believes that if Goldman, Sachs & Co. was flaming out, the feds would not flinch? Answer: no one with a measurable IQ. Resolution authority resembles proactive bubble defense: The optimal time to use it is before the anticipated corpse turns blue. But if Paulson had shuttered Lehman right after Bear collapsed, would he be praised, pilloried or prosecuted like a dog? Lehman would have howled, Congress would have whined, so try door No. 3. Resolution authority demands, well, resolution in the face of a spitting mob. And yeah, money; no free lunch here. To make it fly requires a hero — Volcker played that role once on inflation — willing to lose everything. Alas, such lunatics are rare, making resolution authority just a dusty prop from an old movie.

Ignores political complexities and gaming under bridge bank. The paper describes how, like the bridge facility provided under the Resolution Trust Corporation, the FDIC would have funded Lehman while it was being sold.

This is not as neat and tidy as the anodyne description makes it sound. First, the funding under the RTC was very controversial even though there really was no choice with lots of banks falling over. Indeed, Congress pushed to have it wound down as soon as possible, even though banking experts contend that keeping the facility open longer would have resulted in better results for taxpayers.

Second and far more important, how is the FDIC going to supervise a dealer firm? Remember, top management and the board will have been fired. The experience with AIG is that the staff executed transactions where they had discretion at prices extraordinarily favorable to counterparties. How is the FDIC going to prevent gaming when on the government dime? Consider this report from a correlation desk trader in early 2009 and then commentary on it:

During Jan/Feb AIG would call up and just ask for complete unwind prices from the credit desk in the relevant jurisdiction. These were not single deal unwinds as are typically more price transparent – these were whole portfolio unwinds. The size of these unwinds were enormous, the quotes I have heard were “we have never done as big or as profitable trades – ever“.….

I can only guess/extrapolate what sort of PnL this put into the major global banks (both correlation and single names desks) during this period. Allowing for significant reserve release and trade PnL, I think for the big correlation players this could have easily been US$1-2bn per bank in this period.

For those to whom this is merely a lot of mumbo-jumbo, let me explain in layman’s terms:
AIG, knowing it would need to ask for much more capital from the Treasury imminently, decided to throw in the towel, and gifted major bank counter-parties with trades which were egregiously profitable to the banks, and even more egregiously money losing to the U.S. taxpayers, who had to dump more and more cash into AIG, without having the U.S. Treasury Secretary Tim Geithner disclose the real extent of this, for lack of a better word, fraudulent scam.

I’ve since heard both theories, that per above, these wildly overpriced trades were part of a backdoor recapitalization, or that the AIG traders were currying favor with potential employers. Either way, it illustrates the potential for large scale abuse by traders operating with a government credit line.

Assumption that buyers will step forward for major operations quickly. Ahem, remember Continental Illinois? In 1984, the then seventh largest bank in the US failed, which until the run on WaMu was the biggest bank failure ever. It was such a garbage barge that the government wound up operating it and was not able to fully divest it until 1991.

Assumes a derivative contract termination process out of a parallel universe. Das took issue with many of the assertions made in the paper:

a) p 8, 9 – re derivative contracts – only some inter-bank contracts were unwound quickly. Many contracts were terminated later and left to run. Also contracts allow termination but depending on whether AET [automatic early termination] or non-default termination is selected, there is no legal obligation to terminate. This directly conflicts with the assertion that under the resolution system, the banks cannot continue the contract.
b) p 15, 37, 49 – apparently, the FDIC will value all the positions. I wish them luck. In many contracts, there are two experts and six opinions as to the value of the contract.
c) p 48 – in my experience, much of the international derivative activity by Lehman took place out of Lehman Brothers Special Financing, guaranteed by LBI or LBH (or both). The report seems to take a different view. I am curious why this may be the case.
d) p 49 – the FDIC are proposing to haircut everybody including derivative counterparties but how does this relate to:

(i) where the position is collateralised.
(ii) the relevant legal entity is not in bankruptcy and able to meet its obligation
(iii) the contract is not terminated by the counterparty
(iv) where valuations cannot be agreed.

Das elaborates further in a forthcoming paper on Lehman:

Filing for Chapter 11 or its equivalent generally constitutes a termination event under the ISDA Master Agreements. In the Lehman case, not all legal entities simultaneously filed for bankruptcy. Due to operational and legal reasons, some group entities did not file at all and continued to operate or were ultimately purchased by other firms. This means that all derivative contracts were not capable of being terminated. Where entities filed for bankruptcy, there were differences in timing which created problems, especially in relation to determining termination values.

Even where contracts were eligible for termination, only where AET provisions had been agreed did the contracts automatically terminate. Where the non-defaulting party had the right to terminate, it is not clear whether all parties chose to terminate the contract….

In the Lehman case, termination values proved problematic. Following the filing, market quotations were not available, especially in the case of more complex and difficult to value products. Even where available, there were significant differences in valuations for some transactions. The turbulent market conditions, ironically caused by the Lehman filing, exacerbated the difficulties. Differences in and confusion about termination dates compounded problems and may have contributed to substantial differences in valuation of contracts.

At the time of its filing, Lehman had around 1,200,000 derivative contracts with notional face value of around $39 trillion open.

It isn’t hard to see that disputes about valuation could have a significant impact on the level of losses.

As you presumably have gathered, the process of creating the FDIC paper has a lot in common with developing financial models: tweak enough key assumptions, and you can make the outcome you need to present look plausible, even when those who are close to the action know it to be unreasonable.

While the FDIC is to be commended for pushing to break up the banks, or at least have the major operating units more clearly segregated, misleading the public about the viability of special resolution in its current form is simply dishonest. It’s clear from a political standpoint why the FDIC believed it needed to do so, since saying that a supposedly tidier firm like Lehman can be resolved while a behemoth like Citi can’t is to force the biggest players to reconfigure their operations. But international dealers with derivatives exposures are the messiest financial businesses to put down, and pretending otherwise is a great disservice to taxpayers, who ultimately pay the freight for future financial firm misadventures.