Dodd-Frank mandates the central clearing of most swaps, ostensibly to reduce counterparty risk, but market participants are already asking the same kinds of questions about this arrangement that they asked about CDSs in the first place. If protection buyers had diversified their positions among several counterparties, are they actually better off taking on the risk of a single derivatives clearing organization (DCO) instead?
That all depends on the capitalization and risk management practices of the DCO, and the competition between DCOs to increase their clearing business might serve to weaken both. Add to that the fact that most end users will interact with the DCOs through a futures commission merchant (FCM) and the counterparty risk picture becomes even more convoluted.
Rule 2 – Value at risk (VAR) is almost never the notional amount of the position. In fact, VAR is almost always less than the notional amount.
Let’s take a look at the typical fixed-floating interest rate swap (IRS), where Party 1 agrees to earn LIBOR, currently say 1 percent, and pay a fixed rate, say 2 percent, for five years, and Party 2 agrees the opposite. If Party 1 should immediately default on its obligation, the maximum total loss to Party 2 would be 10 percent of the notional (2 percent X 5 years), and that is only true if LIBOR goes to 0 percent. If LIBOR stays at 1 percent, Party 2’s VAR would be 5 percent of the notional (since Party 1 would immediately stop paying LIBOR). Party 1’s VAR is much more variable, and potentially larger, although nowhere near the notional amount of the position. So when the press mentions the trillions of dollars of IRSs outstanding, we should remember that the actual risk involved is a small fraction of that amount.
For certain derivatives, of course, VAR can be a much larger fraction of notional. In CDSs, for example, the protection seller is theoretically on the hook for the full notional amount, at least in physical delivery CDSs. However, should the seller default at the time of a credit event, the protection buyer would be out the difference between par and the securities’ value at that point, perhaps 70 percent of the notional. Should the protection buyer default, however, the seller’s VAR is much less, which brings us to...
Rule 3 – VAR is almost never symmetrical for both sides of a derivative. This is obvious in the case of CDSs, but is also the case in many other derivatives. In an IRS, for example, the floating receiver has significantly more VAR than the fixed receiver, even though his current income may be lower. For contracts for differences (CFDs) and total return swaps (TRSs) the VAR is more symmetrical, generally not completely. For commodity swaps, as for commodity futures, the VAR is not symmetrical, simply because a commodity’s price can only fall to zero, but it could theoretically rise an unlimited amount.
The commodity case brings to light another symmetry issue – the absolute value of the swap parameters. If we look at the previous IRS case, for example, the absolute value of LIBOR has a practical impact on the floating receiver’s VAR. If LIBOR is at 2 percent, the potential for a sustained rise of 6 percent is significantly higher than if LIBOR is at 10 percent, at least in the minds of most market participants. However, the initial margin (IM) requirements for many listed contracts are symmetrical.
Whether regulators and DCOs recognize the asymmetrical nature of VAR in setting IM requirements remains to be seen.
Rule 4 – Calculating counterparty risk is a multi-part, multi-step process. Take, for example, the calculations an FCM needs to make before allowing a clearing customer to enter into a new trade. The FCM must consider the following:
Internal customer financials, which are updated monthly or quarterly
Current derivatives and collateral positions, which are updated daily or in real time
Product and market data, which is generally updated in real time
All this data must be compiled, evaluated and a conclusion drawn almost instantaneously.
The data flow for trade approval is set out below:
Needless to say, large FCMs with thousands of customers will have to make hundreds of these calculations every day and their financial health will depend on their accuracy. This is one of those cases where an investment in some good technology will pay off handsomely, both in safety and in market share.