Foreign exchange market participants should be relieved that the continuous linked settlement (CLS) system did exactly what it was designed to do as the US broker Refco collapsed into Chapter 11 bankruptcy in October. There are no reports of money being paid away without the opposite side of the transaction being received, which can be largely attributed to the success of CLS’s design.
Refco’s spectacular collapse will have revived memories of Baring Brothers’ demise in 1995 and, for those with really long memories, the failure of Bankhaus Herstatt in 1974. On that occasion, when German regulators closed down Herstatt, the bank had received its Deutschmark side of FX transactions in Europe before it had made any of its dollar payments in the US. The result was that counterparties were left holding unsecured claims against the insolvent bank’s assets.
Settlement risk, or Herstatt risk as it is often termed, has been a cause for concern ever since. The unregulated FX market was of particular concern, especially as it grew in size through the 1990s. Systemic risk as the result of a massive settlement failure was regarded as a major worry; the response was the creation and implementation of CLS.
However, some participants in the FX market heavily criticized CLS before it went live in September 2002. It was seen as too expensive. Furthermore, many did not believe it was actually needed. By the time of its launch, FX volumes had not only stalled but, according to the Bank for International Settlements, were actually in decline. Furthermore, the FX market had taken its own steps to mitigate elements of Herstatt risk, including initiatives such as netting agreements between counterparties.
Nobody can say definitively that FX trades involving Refco would not have settled if CLS had not been in operation. But the fact that there were no problems is very much a plus, particularly as far as risk managers and credit controllers are concerned.
“There were no problems with CLS. It was its first proper test and it worked very well. All our outstanding trades in the system settled in good order,” says Mark Clatworthy, managing director, operations, at RBC Capital Markets.
Jonathan Butterfield, executive vice-president, sales and communications, at CLS Bank, adds: “There were no fails with Refco. If trades were left in the system, we got payment versus payment (PVP). All participants had full oversight of their trades’ status and daily net positions as the situation developed.”
However, according to several sources, severe settlement problems did arise, and many are blaming CLS for these. “It didn’t function well. Nobody lost any money in the system, so, in that sense, it worked well,” says the global head of trading at one top FX player.
He adds: “All the Refco trades were matched in CLS but at least one bank pulled its trades out prior to settlement. That meant nobody settled with Refco via CLS on the Monday and Tuesday (October 17 and 18). Refco then filed for Chapter 11 and it was game over. Positions were then cancelled and marked to market with either a loss or a profit.”
Pointing the finger at CLS in this case might be unfair. The criticism could stem from a lack of full understanding of how CLS works and what exactly it does by those at the front end of the deal chain. Settlement is something that typically only concerns traders when a deal goes wrong and results in a loss.
CLS is not a clearing house, as many wrongly assume, so it does not guarantee a trade in the event of a counterparty’s collapse. It does not, in fact, even guarantee settlement. What it does do is ensure that if trades are entered, one side of the transaction cannot be paid away unless there is PVP, which is the simultaneous exchange of the two currencies involved in the deal.
According to the head of e-commerce at another bank, the problem was not so much with CLS as with the fact that some banks pulled their trades with Refco from the system. “People were reneging on transactions with Refco. Everybody was out for themselves. People pulled out of trades selectively,” he claims.
Butterfield confirms that some banks did start to pull or “rescind” trades from CLS from Friday, October 14. Banks are allowed to do this unilaterally up until midnight Central European Time on the business day before the trade’s value or settlement date, although they should normally have a good reason for doing this.
“A rescind in CLS represents nothing more than a statement that you’re going to settle in a different manner,” says Butterfield. “Late rescinding should only be used for serious credit reasons.”
He adds: “Refco was a matter of confidence, not liquidity. This was not like Barings. If you do rescind, there are several ways to settle. You can settle outside CLS, require payment first, net settle the trade or enter into a close-out agreement under your Isda [agreement]. It is an individual bank’s decision to elect to rescind unilaterally prior to the midnight before value date deadline.”
These are interesting points. The fact that all trades with Refco within CLS settled suggests that liquidity was not the reason to pull deals out of the system. However, several banks did. The more spurious rumours are that this was done to allow certain trades to be netted or closed out under Isda agreements.
CLS naturally will not comment on whether those banks that rescinded trades pulled all of their deals with Refco out of the system. “CLS did work as advertised and that’s that,” says Butterfield.
Some are claiming that once trades were rescinded, CLS was unbalanced and that then the system did not function properly. Butterfield denies this. But word will have got out quickly that trades were being pulled, which might have prompted a frenzy of deals being rescinded. After all, few institutions will have wanted the honour of being the last bank to settle with Refco, just in case money did go missing.
What motivated banks to rescind trades initially remains open to conjecture. However, it is clear that taking trades out of CLS could have exposed those institutions that rescinded to settlement risk, unless they had already decided they were not going to settle. It seems inconceivable that banks would have done it so they could selectively cancel trades, but those are the allegations that are circulating. Not surprisingly, there is a fair amount of bad feeling in the FX market as a result.
