In this brave new banking world of impenetrable bureaucratic morass designed to keep us all from ourselves (slogan: don’t panic, there’s a lot of new math), there are new acronyms for just about anything. To regulators, some groups of letters mean a lot more than they might for banks, while investors, loosely defined, focus on still others. One such term found itself in a perfect storm of the all-too-familiar unintended consequences of regulations.
There is, of course, a fatal conceit at the heart of all of this – that regulations alone can design a foolproof system. This is not to say that mistakes won’t be made or losses tallied, only that with each new layer of complexity the hubris unduly aligns the belief that dictation alone can prevent individual circumstances from ever again impacting systemic function. As if that attempt wasn’t risible enough, the way in which regulators go about accomplishing this utopian design is by using the last crisis as the template.
AT1 bonds are a relatively easy concept to understand from both bank and investor perspectives, but that doesn’t mean everything is straightforward. Governments wanted banks to give them more assurance that if there ever was a “next one” that bank investors would be specifically prepared to act as a capital buffer before “forcing” (TBTF doctrine survives everything) sovereign action and taxpayer loss. The fuzzy but officially embraced concept of “ring fencing” starts with any bank’s capital structure. A contingent capital security seemed the perfect answer; functionally a bond that pays slightly better than peer bank debt because in times of stress the bank can suspend irrevocably the interest payment if not convert it into common stock.
These contingent convertible bonds, colloquially referred to as CoCo’s or just cocos, were initially met with skepticism before orthodox central banking thrust enormous “reach for yield” in their direction. For investors, it seemed all relatively easy; that banks would only suspend interest if their capital ratios fell below a certain amount. Nobody cared much about the “certain amount” in the world of QE3 (then 4) for dollars and Mario Draghi’s promise in euros. Issuance soared, with Dealogic reporting that well more than €100 billion (actual estimates vary widely) in cocos were sold since 2013, including ~€45 billion in both 2014 and 2015. Those totals do not count cocos sold in dollar-denomination.
Even as the world shifted under the “rising dollar”, there was little actual alarm. Issuance slowed, but as Moody’s wrote in May 2015 there was every (mainstream) reason to fully expect the comprehensive outlook contained within Yellen’s “transitory”:
Moody’s Investors Service says that the global issuance of contingent capital instruments (CoCos) has slowed, but will likely recover in the second half of 2015, as banks strive to meet their capital requirements.
“Banks issued $47.5billion of CoCos between January and mid-May 2015, down from $53 billion a year ago. Although annualized the amount would total about $127 billion for 2015, or well below $175 billion in 2014, we expect a pick-up in the second half as banks fulfil their regulatory capital requirements,” says Barbara Havlicek, a Moody’s Senior Vice President.
It’s an amazing repetition, as this Moody’s statement about coco issuance could have just as easily passed as a similar statement about manufacturing PMI’s. Weakness was temporary since central bankers declared it was; because of this logical fallacy (appeal to authority, especially when that authority has nothing to offer but credentials) these expectations were ripe for unmasking. It would take something significant for it, however, as this overriding theme was terrifically entrenched. As late as September 2015, echoing how “transitory” was still thought to apply even after the events of August, Markit Financial Information Services was staunchly defending both financial services credit and cocos specifically.
Financials debt has proven to be one of the areas of relative strength in the current market with the Markit iBoxx $ Banks index and broader iBoxx $ Financials index both delivering positive total returns in the year to the end of August. This relative performance also hold true on the other side of the Atlantic with the iBoxx € Financials and iBoxx € Bank indices outperforming the wider universe of Euro denominated investment grade bonds by 40bps and 50bps, respectively, over the same time period.
The outperformance was largely driven by falling risk perceptions as the index option adjusted spread of both the Euro and USD bank indices have performed better than the wider bond market. Falling risk perceptions have been most felt on the subordinated segments of bank balance sheets with the Markit iBoxx € Banks Subordinated and iBoxx $ Eurodollar Banks Subordinated indices outperforming their more senior peers since the start of the year.
Because of this apparent market faith, Markit declared, “investor demand for the asset class [cocos] also looks to be healthy.” It didn’t quite turn out that way, as cocos present us with yet another indication that “something” changed in the middle of last year. Initial estimates for coco issuance and capital “needs” in Europe for 2016 were running around another €40 billion, less than 2015 (which was slightly less than 2014) but at least a faithful volume that keeps the regulatory fiction in place. Well into March 2016, however, actual issuance started the year with precisely €0.00.
It wasn’t until mid-March that a bank finally got a coco to issue, and there is reasonable conjecture that UBS had the ECB’s “upgraded” QE program to thank for it. Whether or not the European central bankers had designs on bank financing with their broad corporate bond buying spree is unclear, but it is quite likely they would at the very least have been pleased about the spillover. As investors came flooding back into the credit market, UBS issued the year’s first coco for $1.5 billion at 6.875% coupon, which the Wall Street Journal reported was the same interest rate the bank paid the year before (perhaps more a commentary on UBS than cocos in general?).
In between those two coco flotations was a world of difference. The markets being overrun by constant liquidations in January and early February included financial, meaning securities of banks and financial firms themselves. It was one of the most striking aspects of the liquidations, bringing up all the wrong sorts of memories about panic and systemic bank irregularity that really isn’t that far in the past. The clear center of that storm was Deutsche Bank.
The name of the firm itself gives it the reputation of size and strength, literally translated as Germany’s Bank. Last year’s stumbles need very little review as they have been covered in widespread fashion, including a good deal of commentary by me. The bank bet on Yellen’s recovery and was forced to pay the price for being so easily led by illusion, or at least begin to do so as we still aren’t anywhere close to figuring out just how much following the orthodox “global growth” projection will take out of the bank. DB, to start with, has reported its first annual loss since 2008, suspended the dividend on the common stock, and greatly enlarged its transition plans.
Into this background of high uncertainty the ratings agencies pressed even further upon outside tension. In June 2015, S&P downgraded Deutsche (along with Barclays and RBS) and then Fitch dropped the bank to A- in December. Negative comments and assertions continued all through the beginning of 2016. In early February, analysts at CreditSights Inc. published a report that suggested Deutsche may struggle to pay all coupons on its riskiest obligations if conditions for the bank did not improve. Coco investors took note.
It was the last kick to investors who had already been engaging in self-reflection by “sell at any cost.” The descent in coco prices was not limited to Deutsche’s, nor was it specific to cocos; DB’s credit default swap spreads, for instance, blew out to about 500 bps (related, undoubtedly, to hedging concerns about cocos and other DB debt).
The European Banking Authority (EBA) did its usual part to make the bad situation that much worse, turning out what looked to be an almost mini-run in bank securities, at least in wholesale terms as far as this wholesale banking world goes. Last December, the EBA issued an opinion on the MDA that suddenly cast a bit of doubt on exactly when banks might start restricting payments to cocos or even begin converting them. To that point, it was clear investors paid little or no attention to the MDA, preferring to focus instead on rates like 6.875% in combination with multiple global QE’s that seemed to add up to no-risk premium interest.
The actual details of the Maximum Distributable Amount aren’t really all that interesting or, as it turns out, really that significant. All the EBA opinion did was clarify that European banks should take into account both Pillar 1 and Pillar 2 requirements in addition to any added by the combined CRD buffers, while further asking the European Commission to revisit the CRR text (about Level 1). I could spend the next thousand words explaining what all that actually meant, but in reality what it did was achingly simple – at a time when investors who didn’t care about acronyms were suddenly confronted with banks suddenly and “unexpectedly” losing large sums of money again and with no propensity to explain why or how (“market turmoil”) they (the banks) managed to do that, these nervous coco holders were then given another dose in the form of more muddied interpretations about just when or where all these cocos might cease to be “risk free” investments.
If there are any adherents left for the full or even strong form of the efficient market hypothesis they should be made to account for the coco market in early 2016. It was yet another blow for the easily dismissed notion of a robust or “resilient” financial system that has been claimed both as a result of monetary policy and these “macroprudential” efforts. Cocos have already demonstrated the effects of pro-cyclicality which quickly spread globally. It may be another reason to thank (temporarily, anyway) the People’s Bank of China for stepping into the dollar market on and around February 11 to end it.
Obviously, the disorder in cocos was not limited to those under the bureaucratically fumbling auspices of the European Banking Authority. As noted above, cocos have been just as popular if not more so in dollars as euros. Further, as the Markit commentary mentions above, those in dollars were actually (and not surprisingly) eurodollar placements, meaning “dollars” rather than dollars. And for yet another example of how everything in the wholesale banking world is connected to everything else, the biggest flotations in “dollar” cocos were done for Chinese banks. You can try to conceive of the global economy outside of the global dollar short, but doing so will always lead to plain confusion as it is always there and is always financial at its root. The world “needs” the “dollar short” for global trade and legitimate purposes, but it is the financial sector that actually takes on the short to get it done.
In October 2014, Bank of China issued its first AT1 (coco) piece at a whopping $6.5 billion, more than three times subscribed with $21.8 billion in orders. Though the deal flow skewed Asian, meaning Asian demand, that only reinforces what the eurodollar actually represents. A little over a month later, Commercial Bank of China funded a large triple currency AT1 starting with a RMB 12 billion (about $2 billion equivalent) piece, $2.95 billion in eurodollars, and a final, marginal €600 million piece. According to Moody’s, the biggest issuers of cocos since 2009 have been the Swiss, including, not surprisingly, Credit Suisse #1 at $19.5 billion (not all in dollars) and UBS ($18.5 billion), and the Chinese; Bank of China ($18.9 billion very likely mostly in “dollars”) and Agricultural Bank of China ($17.8 billion).
When the coco market fell apart in early 2016, with no issuance at all for nearly the whole of the first quarter, that disrupted intended funding not just for Deutsche Bank and European firms expecting to raise that €40 billion on the year, but also the Swiss and Chinese heavily in terms of their intended “dollar” flow. I haven’t found any published expectation for what Chinese banks, in particular, were planning on funding via this market in 2016 but it is quite reasonable to assume it was significantly more than zero. Where might they have gone instead to replace those expected “dollars?”
Chinese banks were likely moved into shakier short-term “dollar” funding arrangements or to be put upon the PBOC’s rolling crisis management. From this perspective you can appreciate the Chinese central bank’s increasingly impossible tradeoff between “selling dollars” and internal RMB liquidity, and why it swings so assuredly from one to the other at these regular intervals. Unfortunately, there is very little information about cocos in the months since the Deutsche Bank-led scare of the January/February liquidations. While UBS was the first to open up the coco market back in March, there isn’t much to indicate that banks are back to sourcing “dollars”, euros, or anything else at even close to the same pace as last year’s level (which was already, again, less than 2014). In other words, just another tightening of the systemic ratchet.
For its part, the EBA in late February further clarified its opinion of the prior clarification with a lovely publication titled, appropriately, SSM SREP Methodology Booklet. In it we learn that for calculating the MDA trigger point the ECB will use CET1 instead of total regulatory capital, plus all the Pillar 2 and CRD buffers. They also noted that as these CRD buffers start to apply they are phased in while Pillar 2 capital requirements decline in response. In all cases, the EBA was critical of how MDA’s are being automated and that someone, somewhere might want to think about what that means. Whether or not that unwound some of the confusion over its December contribution to the systemic coco run isn’t clear, nor is it really the point.
This whole story revolves around how the eurodollar system connects each of the various pieces so that what used to be distinct and discrete systems are now interconnected nodes in a complex, intertwined network. Thus, the European Banking Authority can confuse investors about a strictly European bank under increasing scrutiny such that the pro-cyclicality of one portion of the banking structure that was built up supposedly in response to forestall the next crisis because it was assumed to look like the last crisis actually becomes a systemic interruption in that whole arena far beyond the boundaries of Europe or euros. The interruption of cocos meant not just trouble for European banks going about their European business but spreading like contagious disease into Chinese banks trying to conduct their “dollar” business.
In attempting to make the system more robust to these kinds of systemic “runs”, regulators have missed the true weakness. All one really has to do is read the SSM SREP Methodology Booklet to get a good sense of it: the problem is not banks, it is banking. Regulations focus on the strengthening and fortification of individual banks when their whole problem is really the wholesale network that practically guarantees serious doubt will never stay local. These are not really banks in the traditionally meaningful sense of the word because there isn’t any money in them; nor actual currency. They are creations of math-as-money and the more regulations try to strengthen them as such the more it only reinforces this virtual, complex arrangement. It really doesn’t matter how good a wholesale bank can be made via government decree if the system in which that bank operates is and remains highly unstable. We know that is the case because the EBA “opinion” in December would never have been noticed if it weren’t for the underlying systemic illiquidity of that time. Again, the problem isn’t so much Deutsche Bank or its cocos but that DB and its cocos could become the trigger point in these kinds of disruptive network interruptions because the network itself is already corrupted.
The lesson should be applied to monetary policy, as well, including the “unexpected” impotence of central banks. As the Fed, what role was there for the FOMC while all this was transpiring? It was a European issue, primarily, as cocos are not used extensively among US banks, but the end result was significant contribution to the global “dollar” problem and, at the very least, some spillover into US banking doubt. Bank stocks in the US followed the “dollar” very closely if not the spread of coco irregularity (almost guilt by association). Again, individual doubts were turned at breathtaking speeds into general systemic concern by general, systemic illiquidity.
At the very least, we should find out relatively soon whether the EBA’s updated MDA has cleared up any lingering doubts about Deutsche or even cocos themselves. Moody’s decided yesterday to further downgrade the bank’s senior unsecured debt to just Baa 2, only two notches above junk. The rationale for the ratings action was the most significant part, however, especially as all that ugliness from earlier in the year is supposed to be once more all behind us. The agency basically said that the DB’s plan to pare back its businesses and cut costs will be effective and “credit friendly”, but there are still growing doubts that they will ever make it that far, “unless there is a material and sustained improvement in the operating environment.”
Cocos or not, it is all about the “dollar” and its self-reinforcing effects upon the economy. As the economy goes, so goes the “dollar”, and as the “dollar” goes so does the “operating environment” for the whole world. It is a glimpse into a small part of how there really is something bigger than all the world’s QE’s combined.