CLS’s previous reticence only served to convince me that those in the market who described it as a bad tax had a point and I had got the impression that CLS’s charges were somewhat discretionary. Now, however, having looked at the figures, I’m less convinced about both points.

It did take a bit of a sledgehammer approach – which I’m not averse to using – to get CLS to open up. Previously, I drew a blank from numerous sources on what it was charging. Everyone I asked immediately clammed up, as if they were witnesses to a crime who had been nobbled before the trial of the accused.

According to Jonathan Butterfield, EVP, marketing and communication, at CLS Bank, CLS’s pricing structure is not something it arbitrarily sets. Instead, feedback is gathered by its markets advisory committee and the ensuing recommendations are presented to the board.

Butterfield says that CLS naturally has had to respond to changes in the market. One of the biggest issues it has grappled with is the emergence of high-frequency traders. This has resulted in a far higher number of smaller-value tickets going through the system. Astonishingly, trades for less than $1 million now account for about 55% of the tickets that CLS processes. Presumably these are done on single-bank portals.

The cost to settle these small lots now starts at 40 pence per ticket, dropping to 25 pence for incremental volume. For other trades, settlement costs range from 80 pence to £1.10. Obviously, these are not the only costs incurred. Consultancy company Z/Yen, which has worked with more than 20 banks in benchmarking the processing cost of a FX trade over the past six years, has tried to determine how much operational and IT costs further rack up the price of settlement. In 2005, it found that operational costs accounted for another $1.90, while IT costs added $1.30. It is likely that both these have fallen since then. However, for market-sized CLS trades it appears that the cost per ticket is around $4.

This calls into question the viability of some prime brokerage offerings, which apparently start from just $2 per million. It appears some PBs are willing to give their FX business away free in return for a payback in other areas. However, I hear that one PB, which has a tariff starting at $8, is almost fighting new clients off. So even tight-wad end users are finding that you do get what you pay for and sometimes it is worth stumping up a little more for a better service.

Another discovery I made this week is that settling outside CLS is unlikely to prove a cheaper alternative, even if nothing ever goes wrong. Jeremy Smith, director of financial services at Z/Yen told me that in 2005 the operational cost for non-CLS trades rose to $4.10 from $1.90. He says that there is a reduced overall in the CLS environment cost due to greater and more efficient STP. CLS users make a maximum of 63 payments a day. Obviously, netting would reduce this but the industry is unlikely now to choose to settle outside CLS. For the moment, CLS has made it clear that it will not accept pre-netted trades for settlement.

CLS’s decision to talk to me has also put paid to another urban myth that had taken hold. It is not paying IBM a fee per ticket. Butterfield says it has a 10-year fixed management contract with the IT provider but he points out that if IBM did charge per ticket, payments would have soared, standing out clearly in the accounts. When CLS was launched, it was envisaged that it would handle 60,000 tickets a day. Its average is now 250,000, with a peak of twice that.

However, I would still like to see a little bit more openness from CLS. I did ask it for its rule book and constitution, but was told that would cost me the $5 million that CLS charges to join it. Euromoney can be quite generous when it comes to expenses but there is a limit. 

                                                               CLS blocks FXMarketSpace’s plan to pre-net cash trades

FXMarketSpace and the settlement monopoly

The 50/50 joint venture between CME and Reuters is a great concept, bringing very real exchange efficiencies to the cash FX market. However, as is always the case, FXMarketSpace is now finding that it is not simply going to be able to turn its screens on and see a wave of liquidity hitting it from other venues.

The company’s management team have been around long enough to have known that it would not get an easy ride from the beginning. But CLS’s refusal to let it submit netted trades for settlement – which almost looks monopolistic behaviour but which it can no doubt justify because it is an industry-owned utility – has clearly harmed FXMarketSpace’s prospects. FXMarketSpace is swift to point out that netting was not the cornerstone of its business model that some, myself included, have believed and perhaps reported.

History has shown that inferior offerings can win out. Old folk will recall the battle between VHS and Betamax for the video market. More recently, various trading venues have been set up, seemingly with industry backing, simply to exert pressure on incumbents to reduce costs. I don’t actually think this is the case with FXMarketSpace and that it is, potentially, a real alternative to the existing trading venues. The company’s oft-reported claim that it is looking to attract fresh players to the market sounds entirely feasible. If it is successful, it will ultimately bring even greater liquidity to the market, which can only be a good thing and perhaps even force through the ability to net, which removes another risk element from the deal chain.

RBS shuts its FX prop desk