The New Crowded Derivatives Trade
At the end of 2012, the four largest banks in the world by assets were Deutsche Bank ($2.73 trillion), HSBC ($2.69 trillion), Mitsubishi UFJ ($2.67 trillion) and Credit Agricole ($2.58 trillion). Topping the US list were JP Morgan Chase ($2.35 trillion), Bank of America ($2.21 trillion) and Citigroup ($1.86 trillion). However, in addition to currency dynamics, these rankings are based on balance sheet presentations as they exist under different accounting regimes. The US banks fall under GAAP rules maintained by the FASB, while the European banks conform to guidelines created and implemented by the IFRS.
The main difference between the two accounting systems, as it relates to balance sheet size and bank exposure, is the treatment of derivatives. The IFRS bases its derivative presentation rules on gross values. GAAP rules favor netting in various forms.
If we were to harmonize balance sheet size according to the IFRS, the three largest banks in the world are actually the three US banks cited above, and it's not even close. JP Morgan Chase "grosses up" from $2.35 trillion to $3.94 trillion; Bank of America from $2.21 trillion to $3.54 trillion; and Citigroup from $1.86 trillion to $2.87 trillion. In fact, the seven largest US banks would "gross up" their balance sheets under IFRS standards by more than 57%, from $10.1 trillion to $15.9 trillion.
When looking at a bank's balance sheet in the modern age, it is sometimes very easy to get lost in the haze of complexity and forget the main purpose. A balance sheet is not just a measure of risk, in its basic form it is supposed to disclose the financial resources available to a bank should it suffer financial pressure. In the context of monetary mechanics, then, a bank's resources are the system's ability to withstand any "shocks", financial or otherwise. Can a financial firm or firms pay out claims without resorting to a massive fire sale of assets? This is the very concept that has garnered much deserved blame for contributing to the Great Depression.
In that respect, netting makes a fair amount of sense since it compares contracts in a positive market value position against contracts under water. The net amount, or what GAAP rules take as the standard for disclosure, does closely align with both risks and potential resources.
According to the Office of Comptroller of the Currency, at the end of 2012 the largest US banks owned contracts with positive fair values of $4.76 trillion. Against those were netted $4.3 trillion, for a netted current credit exposure (NCCE) of just $386 billion. That is the Comptroller's estimate of both potential current risk (potential future risk, or estimates of how far these values can change in the future, add $675 billion to netted estimated potential losses, but that's a topic for another day) and positive resources to the US banking system. To achieve these market values, the Top 25 US banks have engaged in $222.7 trillion in notional amounts.
Of that $222.7 trillion, 96.5% is traded OTC outside of any organized exchange; just a private contract between two parties (or three). $208 trillion (93.4%) belong to the top four US banks alone - the three cited above plus Goldman Sachs. Of those big four banks, 98.4% (including 100% at Goldman) are "for trading" and marked-to-market daily.
By far the largest exposure among these banks is interest rate swaps. While credit default swaps held so much of the public's imagination in 2008-09, interest rate swaps are where the action is now. Out of the $208 trillion at the big four, $128 trillion is interest rate swaps; $80 trillion of that are contracts maturing in less than one year. Trading revenue from interest rate swaps was $3.3 billion in Q4 2012 for these four banks, by far the most of any derivative type (total derivative revenue was $2.99 billion, lower due to the $567 million in negative revenue from credit derivative positions).
For the top 25 banks as a group, total interest rate swap revenue for the full year 2012 was $17.1 billion. That was up 52% over 2011, and 178% more than 2010.
But that isn't the fullest extent of bank activities in derivatives. While regulated commercial bank revenues for 2012 were $17.1 billion in interest rate swaps, bank holding company revenue actually totaled $23.9 billion. That was an 82% increase over 2011, and 383% more than 2010 (these revenue figures for bank holding companies are consolidated, meaning they include revenue from both the regulated commercial bank and the holding company together).
Bank holding companies are not regulated in the same manner as their bank subsidiaries. The Bank Holding Company Act of 1956 established the legal constraints over the legal specifications for a bank holding company. They are regulated by the Federal Reserve under Regulation Y, and are subject to capital rules and reporting requirements. But the 1999 passage of Gramm-Leach-Bliley allowed bank holding companies to become "financial holding companies" and thus engage in "other financial activities". That included securities dealing and underwriting at the holding company level.
If we include holding company derivative exposures, notional amounts go from $222.7 trillion to $287.6 trillion. We have no idea what the netted exposure is for the holding companies because the OCC does not provide any figures; the only information offered is the size of notional amounts and revenues they generate for the holding companies. Bank of America, for example, shows $42.5 trillion in notional exposure at the regulated bank, but $61.9 trillion consolidated at the holding company. That means there are $19.4 trillion in derivative contracts at the holding company alone, netted in some unknown fashion.
JP Morgan, the largest derivative holder, only shows a measly $500 billion at the holding company level, whereas Morgan Stanley grosses $2.45 trillion in its regulated unit but a hefty $42.5 trillion at its holding company parent.
What really sticks out, then, is the disparity in revenues. Across all derivative contract types, regulated banks "earned" $17.9 billion in 2012. Across all derivative contracts, holding companies, including the $17.9 billion at the regulated subs, earned $48.8 billion. Despite a pittance in total notional exposures compared to their regulated subsidiaries, holding company traders seem to be far more efficient at winning trades. It is left unexplained why this disparity exists at why it is so large, but we can infer a great deal from just doing the math.
Circling back to interest rate swaps, we know that this has been the focus of bank trading for now two-plus years, as the revenue growth demonstrates. Credit derivatives have been a big loser, especially as some legacy crisis and pre-crisis structures and portfolios are wound down and run off, so focus had to turn to some other trade to make up the difference. Made possible by QE's suppression of interest rates across the yield curve and even credit markets (look at high yield rate and spread movements in the past few years), banks and especially holding companies have done very well not fighting the Fed. That would suggest, very strongly, that banks, and especially bank holding companies, have been marginally paying floating in these interest rate swaps (and receiving fixed).
In addition to netting, derivatives also are largely collateralized to varying degrees. The regulated banks reported that 58% of their net current credit exposure was with other banks and securities firms; 1% from hedge funds; 6% from sovereign governments; and 34% from corporations and "other" counterparties. Of those two large exposures, regulated banks hold a total of 88% collateral (at fair value) against banks and brokers, but only 41% against corporations and "other".
For the bank holding companies, we have no idea what collateral they hold or to what extent. What we can discern from all of this information, especially the disparity in trading revenues, is that holding companies are taking riskier positions and holding far less collateral. The flipside to that is they likely are posting far less collateral (posting collateral is a relatively expensive proposition that reduces revenue and returns), particularly since the revenue numbers suggest a far higher proportion of "winners" (where collateral is being posted in the other direction).
If we are also correct in interpreting the relative skew of trading positions at holding companies as far more favoring paying floating, that coupled with the potential lack of collateral may have placed holding companies (and some of the regulated banks) in a relatively precarious place given the recent "volatility" in credit markets. Ever since the word "taper" has been extended into the credit trading lexicon, bond yields have backed up significantly, particularly in the high yield bubble.
If holding companies are now on the wrong side of that trade, that would mean a rash of collateral calls. The 10-year treasury swap rate has seen, by far, the most significant upward movement since the taper talk began. It was such a dramatic move that it would seem that some traders are being forced out of their positions and into more netting at the margins; that is to obtain an opposite trade to extract oneself from the original position. Given a matched duration between swap contracts and collateral, this begins to explain some of the shortage of collateral seen in the 10-year US treasury space (http://www.alhambrapartners.com/2013/06/04/repo-warning-returns/) - where QE is not at all helping.
For regulated banks and their holding company parents, these positions make perfect sense. Knowing that the Fed was going to go heavy into QE, they responded by taking one side of that trade, particularly since the Fed has been careful to suggest ZIRP would be around through 2015 at least. It seems that taper talk has caught them sideways and exposed (the renewed asset inflation and bubble markets like real estate and, again, high yield debt looks to have gotten some of the FOMC's attention).
This raises all manner of questions about interest rates in general. How can the Fed allow taper or even an unwind of QE with such a crowded trade on the floating side? Given Regulation Y, it's not as if the Fed isn't paying attention to holding company activities, though it's not likely they know much more than modeled aggregates instead of the exact positioning in the derivatives trades themselves (not unlike JP Morgan's CIO office and IG9). That means that holding companies are at least doing enough netting and "risk management" to maintain their Fed ratings (the Federal Reserve simply rates holding companies on a 1 to 5 scale), but there is so much black box here that it may not mean much.
It's a trap that the Japanese know all too well. Rising interest rates are not necessarily going to ruin US banks' credit holdings of US treasuries, there is already an accounting convention for that. They can simply swap any treasuries in their trading books to their "bank books", claiming their irrevocable intention to hold them to maturity and thus removing mark-to-market. Since there is little perceived credit risk in holding to maturity a US treasury, rising interest rates won't impact banks with price losses.
But the derivative markets are far different. Rising interest rates can and likely already have reduced liquidity through the same process that destroyed bank liquidity in 2007 - collateral calls. With so much of this both OTC and inside the opaque world of bank holding companies, we will have absolutely no idea how much impact interest rates can have, and neither will regulators directly.
That raises another set of problems, namely conflicts of interest. Rising interest rates would actually be beneficial to markets and the economy if they were natural and organic. But might the Fed maintain ZIRP and QE to save banks from liquidity problems? There has been suggestive evidence of this in Japan as Japanese banks have been huge derivative market participants since ZIRP and QE in Japan were instituted over a decade ago (and why JGB volatility has been so damaging since the latest and "greatest" Bank of Japan QE was enacted this year).
This is exactly what we should expect by artificial intrusions into erstwhile nominally private markets by central banks. Given the allure of "easy" money, it is little wonder banks not only rush to the trade on that side but also do so as risky and hidden as possible. As long as the banks are aligned with the central bank policy everyone wins (theoretically); it's a victimless crime (theoretically). But once the markets get so distorted and crowded based on nothing more than intervention and perceived intervention, there is a drastic inability to remove even the slightest degree. It is a feedback loop or trap.
It is never losses that bring down banks; it is illiquidity. That was as true of Lehman Brothers and AIG as it was a large proportion of banks in the first stages of the Great Depression. With collateral so vital to modern liquidity, and with QE and repo demand actively at odds with healthy circulation, the potential for problems is heightened, as it always is with crowded trades. It's not the huge notional amounts that are really concerning, it is where they are located and to what extent they are covered by collateral and truly enforceable netting.
