Reflections on FILS Europe 2026, liquidity, competition and the peculiarities of fixed-income trading.
I've spent a good part of my career around trading desks, watching technology change almost every aspect of how we trade.
We have more venues, more liquidity providers, more sophisticated pricing, better analytics and more electronic execution than ever before.
All tremendous progress.
But after three days at FILS Europe in Antibes, I found myself asking a fairly simple question.
With all this technology, why is it still so difficult to work out where the real liquidity is? And why do traders still have to work so hard to find it?
One of the recurring discussions at FILS concerned the reliability of axes and indicative prices.
Michał Pondel from Allianz Global Investors talked about looking beyond axes towards measures such as dealer hit ratios and the firmness of prices.
Benjamin Mahe from BNP Paribas Asset Management went further, describing how unreliable axes can result in dealers being excluded from RFQ selection altogether.
I can understand the frustration.
An axe is supposed to tell me someone has an interest in trading a particular bond. But if that interest evaporates the moment I try to execute, what have I really gained?
Indicative prices have their place. They provide market colour and help with price discovery. But they're not necessarily executable liquidity.
We've become very good at distributing pricing information. I'm less convinced we've become equally good at telling traders which of that information they can actually trade on.
There's a big difference between seeing a price and being able to deal on it.
I recently had a conversation with a senior European government bond e-sales executive about executable streaming prices alongside traditional RFQ.
He raised an interesting point.
Under the RFQ model, dealers have an opportunity to understand a request before committing to a price. They can assess the size, consider their positions, look at what's happening across their client franchise and decide how aggressively they want to respond.
That information has value. It helps manage risk, but it also provides a commercial advantage.
With executable streaming prices, the dynamic changes. Dealers are committing to prices without assessing each individual request beforehand. They need to update those prices continuously, respond faster and manage their risk accordingly.
It's a different proposition for the bank, and I could understand his concern.
But I couldn't resist challenging it.
Haven't we been doing something similar in the futures markets for decades?
During my time at Barclays Capital, our fixed-income business was one of the bank's biggest users of futures.
When clients traded with the cash bond desks, the resulting interest-rate exposure could be hedged automatically in the futures market. The technology was already there to execute and manage that part of the risk at speed.
And that was years ago.
Of course, automatically hedging a trade isn't quite the same as continuously making executable prices in cash bonds. But it demonstrated how much could already be automated.
So I wonder whether technology is really the biggest obstacle.
Perhaps the more interesting question is what happens to the economics of liquidity provision when dealers have less information and less time to assess individual trades.
One thing I learned over the years is that an opaque market isn't necessarily a disadvantage to everyone.
For liquidity providers, information has value. Knowing who's buying, who's selling and having time to consider the risk can make a meaningful difference.
And let's not forget, banks are commercial businesses. They provide liquidity because there's money to be made, not simply as a public service. They commit capital, manage inventory, hedge risk and expect to earn a return.
They also pay considerable sums to trading venues to distribute prices and facilitate transactions, depending on the venue's commercial model.
Who ultimately bears those costs, and how they influence the prices clients receive, is an interesting question in its own right.
Greater transparency doesn't just change how traders find liquidity. It can change how liquidity providers compete and make money.
Of course, there's a fairly important difference.
Futures are standardised contracts, with liquidity concentrated in relatively few instruments. Buyers and sellers meet through central order books, with visible prices and executable depth.
Cash bonds are different.
A company issues a bond, investors buy it, and much of that inventory may sit in portfolios for years. You can have all the pricing technology in the world, but if nobody wants to sell the particular bond you're looking for, liquidity isn't magically going to appear.
This is especially true in corporate credit.
Government bonds are somewhat different, with generally deeper liquidity in benchmark issues and a more active market-making community. But even here, a newly issued benchmark and an off-the-run bond can be very different propositions.
And transparent, executable interdealer government bond markets already exist.
So I'm certainly not suggesting the entire cash bond market should operate like a futures exchange.
Transparency doesn't create inventory, and an executable price doesn't guarantee unlimited liquidity.
But surely there's room to make more of the liquidity that does exist accessible to a wider range of participants?
And that doesn't mean streaming is the answer to everything.
One thing that became clear at FILS is that no single execution protocol suits every trade.
Aria Pattinson-Speirs from MarketAxess made the point that static execution decisions don't work particularly well in such a fragmented market.
Benjamin Mahe reinforced the importance of moving between venues and protocols depending on the instrument and market conditions.
I agree.
RFQ, RFM, RFS, streaming and voice all have their place. Different instruments, trade sizes, liquidity conditions and execution objectives call for different approaches.
Sometimes a streaming price is exactly what you need. Sometimes you want several dealers competing for an RFQ. And sometimes a conversation with someone who understands the market is worth considerably more than another electronic indication.
The challenge isn't deciding which protocol wins. It's being able to use the right one for the trade.
Which brings me to perhaps the most frequently discussed subject at FILS.
Every additional venue, protocol and liquidity source potentially introduces another connection, another application and another set of information someone has to make sense of.
Who ends up dealing with it?
Usually the trader.
Guillaume Carreno from Crédit Agricole CIB described the proliferation of applications across the trader's desktop and the challenge of bringing everything into one coherent view.
Anyone who's spent time on a trading floor will recognise the picture.
Bloomberg, OMS, EMS, Excel, chat, venue screens, pricing tools, analytics...
Each has its purpose. Unfortunately, they don't always behave as though they know the others exist.
And trading isn't just about obtaining a price and executing an order. There's finding liquidity, assessing execution choices, managing orders, allocations, reporting and everything that follows.
Established platforms have spent years developing these capabilities, and traders quite rightly expect them to work.
Another source of prices, however attractive, isn't enough if it means sacrificing the tools and workflows people already depend on.
Perhaps the opportunity isn't to replace everything, but to make the various components work better together.
And then, of course, there's AI.
It would be difficult to attend a financial technology conference these days without hearing how AI is going to transform everything.
I'm sure it will play an important role.
But if we can't reliably distinguish an indicative price from executable liquidity, how much confidence should we have in an automated system making decisions based on that information?
And if the workflow is fragmented across half a dozen applications, are we improving the process or just making the same inefficiencies happen faster?
Maile Robichaud from State Street Investment Management spoke about interoperability being the next evolution of fixed-income trading.
That resonated with me.
We've had the ability to automate parts of trading for decades. Perhaps the next challenge isn't simply making things faster, but making the entire process work better together.
Otherwise, we risk building increasingly clever technology around fundamentally disconnected workflows.
I don't think the future of fixed income is about one execution protocol replacing another.
Nor do I think greater transparency alone solves the liquidity problem. Banks and other providers need to make money from the risks they take, and competition must work for both sides of the market.
What struck me at FILS was that we've developed an impressive range of ways to access and execute liquidity, but perhaps haven't made them particularly easy to use together.
More data doesn't necessarily mean better information. More axes don't necessarily mean more executable liquidity. More venues don't automatically mean better execution. And more automation isn't particularly helpful if the underlying workflow remains disconnected.
For me, that's where the opportunity lies.
Give traders access to the venues, dealers and protocols that suit their business. Help them understand where genuine liquidity exists, compare prices and execution costs, and make informed decisions without navigating half a dozen applications.
And do it without asking them to abandon the systems, tools and relationships they already rely upon.
I'll admit we're not entirely neutral on the subject.
At RocketFin, these are some of the challenges influencing our work on Ignite, our workspace-native Rates OEMS, which we introduced at FILS.
We're not trying to tell the market how it should trade or suggest that one venue or protocol has all the answers.
Quite the opposite.
We believe traders should have the freedom to access and execute liquidity in whatever way makes sense for the trade, with the information and tools they need brought together in a more coherent workspace.
We're still early in that journey, and there's plenty we want to learn from the people actually trading these markets.
But I came away from FILS with one fairly strong conviction.
The fixed-income market doesn't necessarily need more ways to trade. It needs to make better use of the many ways it already has.
No one wants one venue. Everyone wants one view.
I'd be interested to hear what others think. Are we genuinely making fixed-income trading simpler, or just adding increasingly sophisticated layers to an already complicated market?
Thanks to the many contributors at FILS, and to Lucy Carter and Nick Dunbar at The DESK and the team at The TRADE for their coverage of the discussions.