The problem is that Deutsche is too big to fail — more precisely, that the new Basel III bank resolution procedures now in place are unlikely to be adequate if it defaults.
Let's review recent developments. In June 2013 FDIC Vice Chairman Thomas M. Hoenig lambasted Deutsche in a Reuters interview. "Its horrible, I mean they're horribly undercapitalized," he said. They have no margin of error." A little over a year later, it was revealed that the New York Fed had issued a stiff letter to Deutsche's U.S. arm warning that the bank was suffering from a litany of problems that amounted to a "systemic breakdown" in its risk controls and reporting. Deutsche's operational problems led it to fail the next CCAR — the Comprehensive Capital Analysis and Review aka the Fed's stress tests – in March 2015.
Major senior management changes were made throughout 2015 and Deutsche was retrenching sharply with plans to cut its workforce by 35,000. This retrenchment failed to reassure the markets. Between January 1st and February 9th this year, the bank's share price fell 41 percent and the prices of Deutsche's CoCos (or Contingent Convertible bonds) were down to 70 cents on the euro. Co-Chair John Cryan responded with an open letter to reassure employees: "Deutsche Bank remains absolutely rock-solid, given our strong capital and risk position," he wrote. The situation was sufficiently serious that the German finance minister Wolfgang Schäuble felt obliged to explain that he "had no concerns" about Deutsche. Finance ministers never need to provide reassurances about strong banks.
On February 12th, Deutsche launched an audacious counter-attack: it would buy back $5.4 billion of its own bonds. The prices of its bonds — and especially of its CoCos — rallied and the immediate danger receded.
Fast forward to the day after the June 23rd Brexit vote and Deutsche's share price plunged 14 percent. Deutsche then took three further hits at the end of June. First, spreads on Credit Default Swaps (CDSs) spiked sharply to 230 basis points, up from 95 basis points at the start of the year. These spreads indicate the market's odds on a default. Second, Deutsche flunked the CCAR again. Then the latest IMF Country Report stated that "Deutsche Bank appears to be the most important net contributor to systemic risks" in the world financial system and warned the German authorities to urgently (re)examine their bank resolution procedures.
Less than a week later, Deutsche's CoCos had collapsed again, trading at 75 cents on the euro. The Italian Prime Minister, Matteo Renzi, then put his boot in, suggesting that the difficulties facing Italian banks over their well-publicised bad loans were minuscule compared to the problems that other European banks had with their derivatives. To quote:
If this nonperforming loan problem is worth one, the question of derivatives at other banks, big banks, is worth one hundred. This is the ratio: one to one hundred.
He was referring to the enormous size of Deutsche's derivatives book.
In this post I take a look at Deutsche's financial position using information drawn mainly from its last Annual Report. I wish to make two points. The first is that although there are problems with the lack of transparency of Deutsche's derivatives positions, their size alone is not the concern. Instead, the main concern is the bank's leverage ratio — the size of its 'risk cushion' relative to its exposure or amount at risk — which is low and falling fast.
Deutsche's derivatives positions
On p. 157 of its 2015 Annual Report: Passion to Perform, Deutsche reports that the total notional amount of its derivatives book as of 31 December 2015 was just over €41.9 trillion, equivalent to about $46 trillion, over twice U.S. GDP. This number is large, but is largely a scare number. What matters is not the size of Deutsche's notional derivatives book but the size of its derivatives exposure, i.e., how much does Deutsche stand to lose?
There can be little question that this exposure will be nowhere near the notional value and may only be a small fraction of it. One reason is that the notional value of some derivatives – such as some swaps – can bear no relationship to any sensible notion of exposure. A second reason is that many of these derivatives will have offsetting exposures, so that losses on one position will be offset by gains on others.
On the same page, Deutsche reports the net market value of its derivatives book: €18.3 billion, only 0.04 percent of its notional amount. However, this figure is almost certainly an under-estimate of Deutsche's derivatives exposure.
It is unreliable because many of its derivatives are valued using unreliable methods. Like many banks, Deutsche uses a three-level hierarchy to report the fair values of its assets. The most reliable, Level 1, applies to traded assets and fair-values them at their market prices. Level 2 assets (such as mortgage-backed securities) are not traded on open markets and are fair-valued using models calibrated to observable inputs such as other market prices. The murkiest, Level 3, applies to the most esoteric instruments (such as the more complex/illiquid Credit Default Swaps and Collateralized Debt Obligations) that are fair-valued using models not calibrated to market data – in practice, mark-to-myth. The scope for error and abuse is too obvious to need spelling out. P. 296 of the Annual Report values its Level 2 assets at €709.1 billion and its Level 3 assets at €31.5 billion, or 1,456 percent and 65 percent respectively of its preferred core capital measure, Tier 1 capital. There is no way for outsiders to check these valuations, leaving analysts with no choice but to work with these numbers while doubtful of their reliability.
The €18.3 billion net market value of its derivatives is likely to be an under-estimate because it is based on assumptions (e.g., about hedge effectiveness) and model-based valuations that are will be biased on the rosy side. Deutsche's management will want their reports to impress the analysts and investors on whose confidence they depend.
The International Financial Reporting Standards (IFRS) used by Deutsche also allow considerable scope for creative fiddling, and not just for derivatives: weaknesses include deficiencies in provisions for expected losses and IFRS's vulnerability to retained earnings manipulation.
Experience confirms that losses on some derivatives positions (e.g., CDSs) can be many multiples of accounting-based or model-based projections of their exposures.
For all these reasons, the true derivatives exposure is likely to be considerably greater – and I would guesstimate many multiples of – any net market value number.
In short, Deutsche's derivatives exposure is much greater than €18.3 billion but only a small fraction of the 'headline' €41.9 trillion scare number.
Bank accounting is the blackest of black holes.
Deutsche's reported 3.5 percent leverage ratio
Recall that the number to focus on when gauging a bank's risk exposure is its leverage ratio. Traditionally, the term 'leverage' (or sometimes 'leverage ratio') was used to describe the ratio of a bank's total assets to its core capital. However, under the Basel III capital rules, that same term 'leverage ratio' is now used to describe the ratio of a bank's core capital to a new measure known as its leverage exposure. Basel III uses leverage exposure instead of total assets because the former measure takes account of some of the off-balance-sheet risks that the latter fails to include. However, there is usually not much difference between the total asset and leverage exposure numbers in practice. We can therefore think of the Basel III leverage ratio as being (approximately) the inverse of the traditional leverage (ratio) measure.
Armed with these definitions, let's look at the numbers. On pp. 31, 130 and 137 of its 2015 Annual Report, Deutsche reports that at the end of 2015, its Basel III-defined leverage ratio, the ratio of its Tier 1 capital (€48.7 billion) to its leverage exposure (€1,395 billion) was 3.5 percent. This leverage ratio implies that a loss of only 3.5 percent on its leverage exposure (or approximately, on its total assets) would be enough to wipe out all its Tier 1 capital.
If you think that 3.5 percent is a low capital buffer, you would be right. Deutsche's 3.5 percent leverage ratio is also lower than that of any of its competitors and about half that of major U.S. banks.
One can also compare Deutsche's reported 3.5 percent leverage ratio to regulatory standards. Under the Basel III rules, the absolute minimum required (Tier 1) leverage ratio is 3 percent. Under the U.S. Prompt Corrective Action (PCA) framework, a bank is regarded as 'well-capitalized' if it has a leverage ratio of at least 5 percent, 'adequately capitalized' if that ratio is at least 4 percent, 'undercapitalized' if that ratio is less than 4 percent, 'significantly undercapitalized' if that ratio is less than 3 percent, and 'critically undercapitalized' if its tangible equity to total assets ratio is less than or equal to 2 percent.
The Federal Reserve is in the process of imposing a 5 percent minimum leverage ratio requirement on the 8 U.S. G-SIB (Globally Systemically Important Bank) holding companies and a 6 percent minimum leverage ratio on their federally insured subsidiaries, effective 1 January 2018.
So, Deutsche's leverage ratio is (a) not much bigger than Basel III's absolute minimum, (b) 'undercapitalized' under the PCA framework and (c) well below the minimum requirements coming through for the big U.S. banks.
The 3.5 percent leverage ratio is also a fraction of the minimum capital standards proposed by experts. On this issue, an important 2010 letter to the Financial Times by Anat Admati and 19 other renowned experts recommended a minimum leverage ratio of at least 15 percent. Independently, John Allison, Martin Hutchinson, Allan Meltzer, Thomas Mayer and I have also advocated minimum leverage ratios of at least 15 percent, which happens to be close to the average leverage ratio of U.S. banks when the Fed was founded.
There are also reasons to believe that the reported 3.5 percent figure overstates the bank's 'true' leverage ratio. Leaving aside the incentives on the bank's part to overstate the bank's financial strength, which I touched upon earlier, the first points to note here are that Deutsche uses Tier 1 capital as the numerator, and that €48.7 billion is a small capital cushion for a systemically important bank.
The numerator in the leverage ratio: core capital
So let's consider the numerator further, and then the denominator.
You might recall that I described the numerator in the leverage ratio as 'core' capital. Now the point of core capital is that it is the 'fire resistant' capital can be counted on to support the bank in the heat of a crisis. The acid test of a core capital instrument is simple: if the bank were to fail tomorrow, would the capital instrument be worth anything? If the answer is Yes, the capital instrument is core; if the answer is No, then it is not.
Examples of capital instruments that would fail this test but are still commonly but incorrectly included in core capital measures are goodwill and Deferred Tax Assets (DTAs), which allow a bank to claim back tax on incurred losses if/when the bank subsequently returns to profitability.
Deutsche also reports (p. 130) a more conservative capital measure, Core Equity Tier 1 (CET1), equal to €44.1 billion. This CET1 measure would have been more appropriate because it excludes softer non-core capital instruments – Additional Tier Capital – that are included in the Tier 1 measure.
If one now replaces Tier 1 capital with CET1, one gets a leverage ratio of 44.1/1,395 = 3.16 percent.
Even CET1 overstates the bank's core capital, however. One reason for this overstatement is that the regulatory definition of CET1 includes a 'sin bucket' of up to 15 percent of non-CET1 (i.e., softer) capital instruments, including DTAs, Mortgage Servicing Rights,and the capital instruments of other financial institutions. The consequence is that Basel III-approved CET1 can overstate the 'true' CET1 by up to 1/0.85 -1 = 17.5 percent.
Yet even stripped of its silly sin bucket — which was a concession to banks' lobbying to weaken capital requirements — a 'pure' CET1 measure still overstates core capital. Basel III defines CET1 as (approximately) Tangible Common Equity (TCE), plus realised earnings, accumulated other income and other disclosed returns. Of these items, only TCE really belongs in a measure of core capital, because the other items (especially retained earnings) are manipulable, i.e., these items are not core capital at all.
And what exactly is Tangible Common Equity? Well, the 'tangible' in TCE means that the measure excludes soft capital like goodwill or DTAs, and the 'common' in TCE means that it excludes more senior capital instruments like preference shares or hybrid capital instruments such as CoCos.
The importance of TCE as the ultimate core capital measure was highlighted in a 2011 speech by Federal Reserve Governor Daniel Tarullo. When reflecting on the experience of the Global Financial Crisis, Tarullo observed:
It is instructive that during the height of the crisis, counterparties and other market actors looked almost exclusively to the amount of tangible common equity held by financial institutions in evaluating the creditworthiness and overall stability of those institutions and essentially ignored any broader capital measures altogether.
As a consequence, CET1 is itself too broad a capital measure but we don't have data on the TCE core capital measure we would really want.
The denominator in the leverage ratio: leverage exposure vs. total assets, both too low
Turning to the denominator, first note that the leverage exposure (€1,395 billion) is less than the reported total assets (€1,629 billion, p. 184). You might recall that the leverage exposure is supposed to take account of the off-balance-sheet risks that the total assets measure ignores, but it doesn't. Instead, the measure that does take account of (some) off-balance-sheet exposure is less than the total assets measure that does not. If this has you scratching your head, then your brain is working. The leverage exposure is too low.
If one replaces the leverage exposure in the denominator with total assets, one gets a leverage ratio equal to 44.1/1,629 = 2.71 percent, comfortably below Basel III's 3 percent absolute minimum. But this total assets measure is itself too low, because it ignores the off-balance-sheet risks, which typically dwarf the on-balance-sheet exposures.
In short the 2.71 percent leverage ratio overstates the 'true' leverage ratio because it overstates the numerator and understates the denominator.
Market-value vs. book-value leverage ratios
There are yet still more problems. The 2.71 percent leverage ratio is a book value estimate. Corresponding to the book-value estimate is the market-value leverage ratio, which is the estimate reflected in Deutsche's stock price. In the present context, the latter is the better indicator, because it reflects the information available to the market, whereas the book value merely reflects information in the accounts. If there is new information, or if the market does not believe the the accounts, then the market value will reflect that market view, but the book value will not.
One can obtain the market value estimate by multiplying the book-value leverage ratio by the bank's price-to-book ratio, which was 44.4 percent at the end of 2015. Thus, the contemporary market-value estimate of Deutsche's leverage ratio was 2.71 percent times 44.4 percent = 1.20 percent.
Since then Deutsche's share price has fallen by almost 43 percent and Deutsche's latest market-value leverage ratio is now about 0.71 percent.
So what's next for the world's most systemically dangerous bank?
At the risk of having to eat my words, I can't see Deutsche continuing to operate for much longer without some intervention: chronic has become acute. Besides its balance sheet problems, there is a cost of funding that exceeds its return on assets, its poor risk management, its antiquated IT legacy infrastructure, its inability to manage its own complexity and its collapsing profits — and the peak pain is still to hit. Deutsche reminds me of nothing more than a boxer on the ropes: one more blow could knock him out.
If am I correct, there are only three policy possibilities. #1 Deutsche will be allowed to fail, #2 it will be bailed-in and #3 it will be bailed-out.
We can rule out #1: the German/ECB authorities allowing Deutsche to go into bankruptcy. They would be worried that that would trigger a collapse of the European financial system and they can't afford to take the risk. Deutsche is too-big-to-fail.
Their preferred option would be #2, a bail-in, the only resolution procedure allowed under EU rules, but this won't work. Authorities would be afraid to upset bail-in-able investors and there isn't enough bail-in-able capital anyway.
Which consideration leads to the policy option of last resort — a good-old bad-old taxpayer-financed bail-out. Never mind that EU rules don't allow it and never mind that we were promised never again. Never mind, whatever it takes.
 CoCo investors feared that their bonds' triggers would be breached and that their bonds would be converted into equity, which might soon become worthless.
 See T. Bush, "UK and Irish Banks Capital Losses – Post Mortem," Local Authority Pension Fund Forum, 2011) and G. Kerr, "Law of Opposites: Illusory Profits in the Financial Sector," Adam Smith Institute, 2011.
 Indeed, as partial confirmation of this conjecture, on p. 137, the Annual Report gives an estimated 'total derivatives exposure' of €215 billion, nearly 12 times the €18.3 billion net market value of its derivatives book and over 4 times its Tier 1 capital.
 In its Annual Report, Deutsche highlights as 'headline' capital ratio its Tier 1 capital ratio, the ratio of Tier 1 capital to Risk-Weighted Assets (RWA): this ratio is 12.3 percent if one uses the 'fully loaded' measure, which assumes that CCR/CRD 4, the Capital Requirements Directive, has been fully implemented. However, the RWA measure is discredited (see here) and these so-called capital ratios are fictitious, not least because they assume that most assets have no risk.To illustrate, the ratio of Deutsche's RWA to total assets is only 24.4 percent, which suggests that 75.6 percent of its assets are riskless!
 These numbers refer to the CRR/CRD 4 'fully loaded' measures.
 See Board of Governors of the Federal Reserve System," Agencies Adopt Enhanced Supplementary Leverage Ratio Final Rule and Issue Supplementary Leverage Ratio Notice of Proposed Rulemaking." Press release, April 8 2014.
 For more on the sin bucket, see T. F. Huertas, "Safe to Fail: How Resolution Will Revolutionise Banking," New York: Palgrave, 2014, p. 23; and "Basel III: A global regulatory framework for more resilient banks and banking systems," (Basel Committee, June 2011), pp. 21-6 and Annex 2.
 For a more complete definition of CET1 capital, see Basel Committee on Banking Supervision (BCBS) "Basel III: A global regulatory framework for more resilient banks and banking systems," (Basel Committee, June 2011), p. 13.