Friday, February 6th, 2015
by GEORGE BOLLENBACHER
Given all the debate and discussion on the evolution of the swaps markets, including some negative comments by one of the new CFTC Commissioners, one group we have been waiting to hear from on this subject is the futures industry, where exchange trading and clearing have been working successfully for decades. Granted, swaps aren’t futures, but the Futures Industry Association’s (FIA’s) input is certainly worth listening to.
Recently the FIA’s Cleared Swaps Operations Committee issued a set of recommendations on how SEFs and DCOs should interact in order to make the swaps markets work better. “FIA believes that these proposed practices will help facilitate the growth of the industry in a manner consistent with existing regulatory guidance.”
SEF Operations
The first section of the recommendations relates to SEF operations and, in particular, the thorny subject of certainty of clearing (CofC). Since any SEF trade that can’t be cleared must be cancelled ab initio, it is essential that both parties to a SEF trade, who may not know their counterparty until after the trade is done, be certain that the trade will clear. Thus the FIA’s first substantive recommendation is that, “SEFs should either generate or receive a unique limit token for all screened orders and should ensure that this limit token is passed to the DCO (even when delegating submission to a third party) and to the FCM (directly or via a hub) in all status messages.” There is a footnote indicating that detailed specifications for the token are being worked on.
This recommendation explicitly encompasses both the ping and push methods of pre-check, so it means that a SEF would either have to receive an indication that the trade would be accepted by a DCO or generate a message to the DCO/FCM/Hub asking if the trade would be accepted. Interestingly, the recommendations also indicate that, “This should include voice trades processed via the SEF (i.e., swap block trades done pursuant to the rules of a SEF), as well as transactions executed electronically.”
There are some additional recommended requirements for SEFs, such as informing the FCM in real time about order status, verifying that a transaction submitted to the DCO is the same transaction executed on the SEF, and allowing FCMs to send kill switches to the SEF in cases where an already approved trade is no longer clearable. Finally, the recommendations would make the SEF responsible for alerting parties to the trade if DCO confirmations do not happen on a timely manner.
Much of the CofC discussion has centered around its impact on latency, or the delay in transmitting an order due to credit checking. There’s not much in the FIA document addressing this question, and we might expect some of the recommendations to increase latency. Whether that becomes an issue remains to be seen, of course.
DCO Operations
Here the FIA recommends that, “DCOs should perform a pre-trade credit check on each transaction. Until this functionality is built, at a minimum, DCOs should provide FCMs with their limit utilization and total credit line using real-time messaging.” Unfortunately, this recommendation is overly simplistic, and technically incorrect. First, the DCO would need to perform a credit check on each order, since by the time a transaction has been done it’s too late. Second, where the FCM has omnibus customer accounts the DCO does not know who the customer is, and can thus only credit-check the FCM. The FCM must credit-check its customer. And, as I have said before, a customer opening many omnibus accounts at different FCMs can shield its true risk from, not only the DCO, but, worse, each FCM.
Next the FIA says that, “DCOs should be guided by the CFTC in respect of removing last look to the FCM (i.e, request/consent messaging) for any SEF trades that have had a pre-execution limit check (i.e., valid limit token).” I’m not sure how to interpret this sentence, but I think it means that, once a DCO has issued a token, the SEF doesn’t have to recheck before allowing an execution. If so, as my daughter used to say, “Well, duh.” There is, of course, the question of how to handle orders where the DCO’s (or even the FCM’s) approval changes once the order has been entered. If real-time messaging can’t deal with that case, then we may need some liability waterfall.
There is discussion in both the SEF and DCO sections about allocations for bunched orders, but no mention of what to do with an unallocated portion of a bunched order if that particular allocation can’t be cleared. This subject has been handled, in a way, by the standby FCM arrangement, where an FCM guarantees any unclearable allocations, but that means that all the SEF/DCO arrangements we have just discussed must also be applied to the standby FCM.
Finally, the FIA recommends that, “DCOs should document policies and procedures relating to the steps to be taken by the DCO following registration of a trade which the parties subsequently notify the DCO to have been an erroneous trade.” Here there are actually two types of events: one in which the trade occurred but one of the parameters was wrongly reported, and one in which a trade was wrongly reported to have occurred, but didn’t. In the first case, does the SEF or one of the parties, send a correction, using the USI/UTI assigned to the original. In the second, is a cancellation message sent? And must both sides of an off-SEF trade send the cancel/correct, or can the DCO rely on one side’s version?
So what does this all tell us? As it turns out, perhaps the futures industry is not particularly well suited to opine on how the swaps markets should work. Many observers have pointed out the fundamental differences between swaps and futures, so we shouldn’t be surprised if futures solutions don’t provide any silver bullets for the issues we’re currently seeing under Dodd-Frank and its foreign cousins. It’s on to the next white paper, I guess.
