So much of bank reporting, accounting, and regulation are directed toward quantifying risks on an individual and systemic basis. This introspection can be very comforting from certain perspectives, particularly since so much of it is wrapped in careful consideration. In the figures on derivatives exposures in the banking system, the Office of the Comptroller of the Currency (OCC) actually does a terrific job explaining its various machinations in terms of defining “risk”.
On an individual level, these risk measures make perfect sense to accomplish their narrowly defined tasks. In other words, the OCC has performed well in quantifying known unknowns (to borrow an unfortunate phrase).
The basic building block of this effort is Net Current Credit Exposure (NCCE). When thinking about bank risk as it relates to being a derivatives dealer, you certainly want to understand the potential for losses; credit risk. Since derivatives are synthetic instruments, notional values are essentially meaningless outside of secondary considerations about volume running inside the system. On an interest swap of any particular corporate bond, for example, if the company goes bankrupt it does nothing to either counterparty in the swap. All that matters is the cash flows that are exchanged based on interest rates.
So the OCC breaks down derivatives in terms of expected cash flows. Contracts with a positive value are classified as “receivables”, while negative contracts are “payables”. In terms of credit risk for the bank or system, which is really counterparty risk, we only really care about the receivables. The losses that come from derivatives only arise when your counterparty defaults; they stop paying you cash flows that you are expecting to receive. Your payables are already accounted for elsewhere on the income statement and balance sheet since they are your obligations.
Using my examples from yesterday, the credit risk for JPM on its swaps with JHH relate to JHH stopping its leg of the payments. Since the swaps are “in the money” on JPM’s side, only JHH is paying out at this point (that will change if interest rates rise enough). To calculate JPM’s total credit risk for interest rate swaps, then, the OCC takes JPM’s Call Report and totals the present value of all expected cash flows only from contracts with a positive market value (those receivables). That leads to Gross Positive Fair Value.
In the first quarter of 2013, that was $3.4 trillion in total interest rate swaps system-wide and $4.2 trillion for all derivatives. Those numbers represent the present value of all expected cash flows by US banks from their counterparties, and it represents a credit risk should everyone that is supposed to pay suddenly default.
However, since that does not account for hedges and offsetting positions, the OCC takes another step to refine that credit risk measure. If you believe that bankruptcy is the biggest risk to those receivables, then you have to factor in recoveries even through the bankruptcy process. Derivatives dealers often require a master netting arrangement that allows the bank to side step (essentially) bankruptcy proceedings and become first in line for cash flows (this is simplifying master netting, but that is really a separate and independently arguable topic for another time). From the OCC’s standpoint, dealers with master netting arrangements reduce the credit risk to receivables even in the event of bankruptcy. So for systemic risk measures, the OCC subtracts contracts with master netting arrangements from the Gross Positive Fair Value to arrive at the infamous Net Current Credit Exposure (NCCE).
NCCE, however, is just a snapshot of what derivatives contracts look like today. So that is a limited view on credit risk to banks, and thus has limited value. The known unkowns is really about what risk exists to banks should conditions change over time, which they inevitably do.
Here is where the most gray area exists. How do we measure potential future changes in credit risk? The OCC has come up with a formulaic estimation for Potential Future Exposure (PFE). Before Q2 2007, the OCC used total notional values and applied some related estimates of relevant factors (interest rates, spreads, volatility) to gauge how much receivables might change, and thus how much credit risk, at the upper end, might be contained but hidden (again, the known unknown). Add that estimate for PFE to NCCE and you come up with Total Credit Exposure (TCE).
The OCC responded to bank concerns about TCE with regard to master netting. After Q2 2007, PFE was changed to a calculation not unlike risk weighting bank assets for capital sufficiency. The OCC multiplies the notional value of the Gross Positive Fair Value by a conversion factor. In the case of interest rate swaps, the conversion factor is 0.015 for contracts with a maturity beyond 5 years; 0.005 for contracts maturing between 1 and 5 years; and 0.000 (zero) for contracts with a maturity of less than 1 year.
That means if a particular bank has $100 billion in notional IR swaps outstanding (on contracts with a positive fair value) all maturing in three years, the OCC’s math turns that into $500 million PFE. It then contorts that raw PFE further in light of something called the “net to gross” ratio, which is largely tangential to this post so I won’t go into that beyond noting that it essentially allows a netting factor to be incorporated in this “adjusted” PFE calculation.
What you take away from this process, particularly the conversion factor, is that even in the regulatory estimation of potential credit risk, interest rate swaps merit little concern. A conversion factor of zero on IR’s with a maturity less than 1-year shows that regulators don’t believe that these kinds of contracts are in any way problematic. Therefore, they do not show up on the adjusted PFE, and thus reducing the estimate of Total Credit Exposure.
After the full panic in 2008, derivatives dealers suddenly became very used to short-term IR swaps. Since swaps with actual corporate counterparties (the biggest “end user”) tend to be much longer in terms, again, yesterday’s post on the Maryland hospitals provides a useful example, that would suggest, and highly so, that the increase in short-term contract use has been for dealer book hedging.
If we use the 1-5 year conversion factor on all contracts with maturities less than 5 years, the Total Credit Exposure jumps by about 20% ($200 billion) in for the most recent years. The OCC, however, is first to tell you that such PFE calculations, even adjusted, overstates credit risks because risks are not uniform in terms of cash flows – among other things, if a counterparty defaults along the way, the swap is terminated and the receivables will only count to that point, whereas the PFE calculation above simply assumes maximum fair value to future cash flows all the way to maturity.
But is credit risk really the definitive measure of systemic risk?
After all, it isn’t really losses that get banks and turn them into cold fusion candidates. Even at the peak of the crisis, in the fourth quarter of 2008, actual derivatives losses were only about $847 million, a relatively paltry figure given the scale of credit turmoil. The worst quarter for derivatives charge-offs was actually the first quarter of 2011, with $1.6 billion related to monoline defaults.
What kills firms is not losses, but illiquidity largely driven by margin/collateral calls. AIG, for example, was taken out on liquidity and collateral problems, not losses.
The cascade of risk that defines the unknown unknowns relates to the vulnerability of a firm’s liquidity structure. In the case of derivatives and dealers, credit risk is a nice academic measure, but it doesn’t really tell us much about the potential for cascading liquidity violence. In the case of all the major failures in 2008, they all came down to an overabundance and overreliance on the maturity mismatch. These firms were funding themselves with overnight and very short-term liquidity, in stark contrast to their asset base. When short-term funding dried up, so did the firms’ ability to survive because they were trapped by maturity carry.
In the derivatives markets, we now see the same kind of process. Dealers appear to be hedging their client exposures by rolling over the shortest-term hedges. What would happen, then, if their ability to roll hedges was interrupted or diminished? The firms would appear much riskier, and would likely be subject to additional margin and collateral issues.
From this perspective, there is a lot of faith in the ability of dealers to get right with whatever might happen down the road. In other words, despite all these risk calculations and estimates, regulators simply assume dealers will act, and will be able to act, proactively, and hedge accordingly and successfully regardless of how modeled parameters change.
That brings the discussion to vega, most importantly. A dealer book, including VaR, is tied closely to implied or estimated volatility. The more volatility potential flowing through models in a dynamic hedging environment, the greater the need for derivative books to get adjusted. That means both an increase in hedging demand and a decrease in the ability to absorb someone else’s risk. With only four real derivatives dealers, even the slight impairment in one is a systemic issue.
We know dealer risk rose significantly in the first quarter simply by comparing Gross Positive Fair Values against Gross Negative Fair Values. In other words, dealers were unprepared or unable to lay off enough risk in the rising swap rate environment at just the beginning of the year (we don’t know yet how that played out in the broad credit market selloff that followed, but we can guess). That led to this contract disparity, severing a relationship that had been very, very stable since the middle of 2009.
The stable relationship between aggregate positive and negative fair values exists by virtue of dealers’ collective ability to manage their own risks. That there was such a noticeable and sudden change, particularly if I am correct about dealers overcrowding the floating payor side, either means dealers just let the risk ride through their books (unlikely) or could not obtain enough hedging opportunities to fully offset new modeled calculations for risks, especially volatility. The latter would explain why swap rates moved so dramatically relative to bond rates (the 10-year swap spread went from negative to positive while bond rates were still moving lower).
That leads in to the rise in collateral demands we have witnessed on derivatives dealers themselves. In addition to the client collateral reversal I described yesterday, we can surmise that even the slightest disruption to the risk relationship described in the positive to negative value dynamic would raise risk awareness among counterparties. The OCC does not provide a whole lot of information on collateral, particularly before 2009, but they note in early 2008 (to reassure any researcher, no doubt) that banks held collateral totaling 30% to 40% of NCCE against other banks and dealers.
Of course, by the third quarter of 2009, the collateral demanded of banks/brokers more than doubled in relative and absolute terms to 90%; an unknown unknown. None of the credit risks measures captured that kind of dynamic, where collateral and volume combine to form a troubled environment, feeding back into dysfunction elsewhere. At the outset of 2008, NCCE was only $465 billion, and that included AIG, Lehman and Bear Stearns. It also included Wachovia (at the time the fourth largest derivative dealer) with its NCCE of just $19 billion, and a TCE of only $48.8 billion.
As one lawsuit against Wachovia alleged, the bank in 2007-08 was, “…facing its own balance-sheet problems at the time, of pushing to get collateral wherever it could.” That never showed up in the credit risk numbers.
The bottom line to all of this is that risks in derivatives books go beyond simple credit risk. In fact, as recent history has shown, credit risk is the least of the system’s problems. It is liquidity and collateral that defines systemic problems, and they are the hardest to see coming, let alone model. What the data shows in the first three months of this year is potentially a change in the previous steady state of the derivatives markets, likely as a shock to dealer books on unanticipated policy expectations. However, that is not going to show up in the usual metrics and formulas.
Click here to sign up for our free weekly e-newsletter.
“Wealth preservation and accumulation through thoughtful investing.”
For information on Alhambra Investment Partners’ money management services and global portfolio approach to capital preservation, contact us at: jhudak@4kb.d43.myftpupload.com
Stay In Touch