When Eli Lederman stood up in front of a packed room at January’s Finexpo financial technology conference, the audience was hoping for a new revelation about an alternative exchange that has generated more press coverage than deal flow. What they got was a plea for legitimacy, as the CEO of Turquoise reiterated the exchange’s raison d’etre.
Turquoise was formed by a consortium of investment banks but is now, as Mr Lederman took pains to point out, a private, “entrepreneurial” company with those banks as shareholders. It was initially billed as the “London Stock Exchange killer” that would be best placed to take the opportunity offered by the Markets in Financial Instruments Directive (MiFID) to compete with the European exchange monopolies. But, months after the MiFID deadline, Turquoise still has not matched a single trade – and will not until September 2008.
Technology concerns
Many observers believe the exchange’s competitors have stolen a march on it by launching on time, but chiefly, concerns focus on the speed and capacity of its technology. While many expected it to choose a mainstream provider, it picked the relatively obscure Cinnober of Sweden.
In Cinnober’s defence, the 120-strong software company has grown enormously in the past few years. It has also won business from some of the many multilateral trading facilities and alternative exchanges that have appeared as markets liberalise, as well as from trade data platform Boat.
Paul Pickup, director at consultancy Trading Technology, says the light-touch nature of Cinnober’s Java platform makes it a good fit for start-ups. “Their technology comes from the internet stable and is light and inexpensive from a hardware perspective. They have a thin client screen, which is also good for penetration of members in a start-up environment,” he says.
“I have repeatedly stated that their technology is not proven in high-volume markets, which is why the selection for Turquoise in particular is not a good fit, although they would dispute that.”
Cinnober CEO Jan Arpi counters: “We have been upgrading our platform continually since the company started. We have improved the latency features, we have added functionality on, for example, dark liquidity pools and we have improved the smart order routing functionality.”
Cinnober’s technology, while relatively low profile, is not wholly untested, having been used in the American Stock Exchange since 2002 and previously in the Nordic power exchange Nordpool, as well as FishEx, the electronic salmon trading market. However, none of these, as Mr Pickup notes, come close to the level of volume that Turquoise is expecting.
“[The American Stock Exchange] had a volume record of 539m quotes in the system on one trading day. I think the average latency was somewhere in the range of four milliseconds,” Mr Arpi says. “Four milliseconds is not bad, but it’s not today’s or tomorrow’s industry standard. It needs to be half of that or less. That’s what we are building now in the new platform.”
Functionality matters
Besides, Mr Arpi says, although latency figures are what clients typically ask for in the first instance, what really matters is functionality. “Is it a smart way of routing orders? Is it a smart way of making the deal where it’s most effective? Do I get the best price? Those questions are much more relevant. Every provider of a trading system is able to meet reasonable latency features,” he says.
Cinnober’s charm for its clients – which include Project Alpha, the “Canadian Turquoise” that is challenging the Toronto Stock Exchange’s monopoly – seems to be its rapid rollout. Cinnober practically offers a white-labelled, off-the-shelf exchange product. Based on standard Java, it can be tailored rapidly to fit a business case and, to some extent, devolves responsibility for technology and infrastructure from the sponsors, enabling them to avoid taking on the millstone of legacy infrastructure that currently hangs around the necks of many incumbents.
With Mr Arpi making the right noises in terms of functionality, and when it is examined through the lens of the new exchange landscape – one of commoditised technology and rapid change – the decision to pick Cinnober appears to make more sense.
There is another major consideration that has often been overlooked – and that is that Cinnober is an independent provider. If Turquoise and its parent banks want to decouple themselves from the existing European infrastructure, then using technology from the two leading providers – OMX, owned by the Nordic exchange group of the same name, or Atos Euronext Market Solutions, owned by NYSE Euronext – would have meant paying the old guard.
