A group of European and UK banks is considering building its own fixed-income trading software, in a move that could allow members to cut ties with the market’s biggest vendor, Ion Group.

The project, dubbed Cohesion, is in its early days but pitches were invited from potential builders of the technology this summer. The banks are now understood to be weighing whether to commit further time and money to the work.

“All banks need to be more efficient, so when you have a player that’s as dominant as Ion it’s a worry for everyone. It’s in the industry’s interest to have more competition,” says an e-commerce specialist at one bank, who is familiar with the project.

A group of banks came together to discuss ways that electronic trading could be done more cheaply. One idea was to build everything ourselves and own the software—and a variant of that continues.

Bank e-commerce specialist

The number and identity of the banks in the group is not certain. At least one bank that was part of the original group is understood to have dropped out; other banks have been invited to join.

Three sources with knowledge of the project independently name four banks—BBVA, HSBC, ING and Lloyds Banking—but could not confirm that all of them are actively involved. Those sources say a total of five banks are currently part of Cohesion. A fourth source claims there are six banks involved.

Spokespeople for all four of the named banks declined to comment for the record. One cast the consortium as “a working group” which is “exploring solutions”, rather than seeking to replace Ion’s fixed-income software, MarketView.

That is a fair description of the project’s origins, according to one of those involved.

“A group of banks came together to discuss ways that electronic trading could be done more cheaply. One idea was to build everything ourselves and own the software—and a variant of that continues,” says an e-commerce specialist at a second bank.

He says the plan is “to create a complete e-trading capability—buying rather than building some smaller components—and then stitch them together to create something comprehensive that could fill the space of a major bank’s existing infrastructure.”  

Many of the banks involved are “Ion banks”, he adds—and multiple sources described the project as a way to replace MarketView, or reduce the banks’ reliance on a system that is said to be in use at almost every bank with a fixed income trading business.

• READ MORE: WatersTechnology spent three months examining Fidessa to see what has transpired inside the vendor since the Ion acquisition. During a period of great change, a lot of questions—and worry—remain. Click here to read more.

Consulting firm GreySpark Partners is managing the project and ran the request-for-proposal process, in which as many as 10 firms are said to have pitched.

GreySpark managing partner, Frederic Ponzo, declined to comment on the identities of the banks, but did not deny the existence of the project or the firm’s own involvement. In an email, Ponzo said GreySpark “runs multiple RFPs for banks or groups of banks every year. This year was no exception.”

The project is sensitive because Ion is one of the biggest vendors to bank trading rooms at a time when the industry is under pressure to cut costs—particularly in fixed income.

Sources with knowledge of the project estimate the largest dealers are estimated to spend $10 million to $12 million per year licensing Ion’s software. Estimates of the costs faced by smaller dealers, like those said to comprise the bulk of the Cohesion group, range from $2 million to $3 million per year.

Those costs have contributed to friction between Ion and its customers, which Risk.net reported on in January. As well as trader screens, pricing and risk analytics, Ion packages can also include the so-called messaging bus—used to ship data and orders between different systems—and connectivity to scores of trading venues around the world. Those connections make it difficult for any bank to cut ties with Ion, undermining their negotiating position when renewing contracts that typically run for three years or more.

It also poses a number of challenges to the Cohesion banks. Some observers doubt the project’s chances of success.

One sticking point is that the banks all have existing contracts with Ion that expire at different points, meaning some would probably not be able to avoid renewing if Cohesion is taken further.

“Cohesion may not go ahead. The banks all have different build deadlines to hit before their contracts renew and they will then be locked in for another three-to-five years. So, they’ll be wondering whether to contribute to the funding if the results won’t be available in time for their own renewal. They could just sit back, let everyone else pay for it and then buy it off the shelf when a later renewal comes around,” says a third source with knowledge of the project.

There is no point holding a gun to someone’s head if you don’t have any bullets.

But a smaller consortium would struggle to muster the financial wherewithal for such a major project, observes the first e-commerce source: “You need to balance the need for critical mass in terms of funding with the need to avoid it becoming unwieldy. You don’t want to try to corral a large group. More than 15 banks would be too large. But if it’s less than five, then the burden of building it would be too great.”

It’s hard to say how big that burden might be.

Two sources with knowledge of the project offer quite different estimates of the time required—the first suggests somewhere between three and five years, while the second says it could be completed in two to three years “if you really put your mind to it, and had enough financial capital.”

The size of the required financial backing is also hard to estimate. As a rule of thumb, the second e-commerce source says it might cost 10 times the annual license fee to build the software from scratch—“clearly, it would help to mutualize those costs,” he says—but he warns that, once built, the software would need further work on testing and maintenance.

Other bank consortia have succeeded in creating software in the fixed-income space—for example, Neptune, the bond trading utility. Last month, a group of nine major banks announced the creation of a platform, DirectBooks, to streamline securities issuance.

These ventures were not going head-to-head with an existing, established service, however.

What Cohesion has in its favour is a conviction that current spending on incumbent vendors—not just Ion—is unsustainable.

“All the banks share a belief that this is a common problem and something they all want to fix,” says a fourth source with knowledge of the project.

“Something needs to happen,” says the second e-commerce source. “Especially in fixed income, banks just can’t sustain the costs they used to. People really need to innovate, and come up with something cool, or force the incumbent vendors to reduce prices.”

This hints at another possibility for participating banks. Cohesion could end up being a device to strengthen their hands in renewal negotiations, but a fifth source notes the project would need to be credible before it could be used as a deterrent.

“There is no point holding a gun to someone’s head if you don’t have any bullets,” he says.