The Last Look…

The subject of non-bank market makers and the primary venues has raised its head again, with several people keen to discuss the presentations by LSEG and EBS to the European Central Bank’s FX Contact Group. What struck me about the snapshots provided is how it reinforces what has been a long-held view of mine – this particular piece of the FX market structure hasn’t changed much in the last decade.

You can find the presentations here, and they offer evidence of what I have been talking about for some time now, the NBMMs’ influence has stabilised – indeed may even have diminished a little, but it remains a big part of the primary market’s environment. Of course, what is often overlooked is this sector’s continued small part in the overall FX market. Yes, it may have an outsized influence in this key area (especially for data), but in the big picture? That is less clear, but looking at the BIS data from last year, if you include hedge funds, which is a decent clip of business, the sector has a 12% share of spot. Equally, assuming EBS and LSEG Spot Matching are in that $25-35 billion per day mark – neither firm reports CLOB volumes remember – than the non-banks are doing around $15 billion per day on the venues, a drop in the ocean of nearly $3 trillion traded in spot every day.

As is the case in so many other areas, there are nuances – a firm like XTX Markets operates very differently to other non-banks, who each have their little nuances, but the majority of whom have one advantage – speed. Equally, through its business model and longevity, XTX has direct relationships, so is involved in more business than probably shows up in the volume data, but the reality is even if all the PB volume in the world was non-bank market makers, it would still only be just over a third of activity.

These firms are vital to the primary markets, however, and by association, for the price formation process. It is notable that another view was reinforced in the presentations, bank prices are more stable and longer-term on these venues. No one should be surprised by this, mainly because, in an era in which order management has become so tricky, the safest solution to avoid conflict with a client is to post interest on the primary CLOB, and leave it there until hit. Equally, NBMMs will inevitably post for shorter periods because they are actively trying to be top of book to exploit their technology advantage.

There is one historical insight that stood out to me; how the NBMM make/take ratio on EBS peaked around 70% after the launch of EBS Prime in 2004 and then declined steadily to under 40% by 2013. This highlights how much this sector was responsible for the heyday of activity on EBS, when it was regularly averaging in the high $100 billions daily (and often over $200 billion). It also can be used as evidence as to why the banks pulled back from the venue soon after, commencing a decline in activity that was only really arrested a year or two ago (internalisation played a bigger role of course).

The number of NBMMs has also declined on EBS (and CME, where many still operate), this I see as a reflection of many wannabee firms finding they can’t compete with the top tier of both banks and non-banks, but also the opportunities in other, more lucrative areas. Many NBMMs have prospered in crypto, they are doing so in prediction markets, and some have discovered equities is a much more rewarding environment than FX.

banks seem more comfortable with how the NBMMs are playing…this is in part due to their own tech improvements, but also testament to the nuancing of the rules of engagement that have been taking place on the primaries over the past decade

The bottom line, however, remains plus ça change. Over the last decade there are a few nuanced changes – more “large maker/takers” NBMMs on LSEG and a notably longer bank participation and median order life on EBS for example – but the reality is what we have always known, NBMMs provide more fleeting ToB support, while the banks provide some backbone. Together, however, they seem to be providing a healthy ecosystem on these venues, perhaps this is why matters have stabilised at these venues and, potentially, started moving in the other direction after a prolonged period of steady decline?

Aside from the tremendous revenue streams offered by the data services associated with both platforms, the real barometer of how they are doing can be found in the volume numbers, and here, things are looking good for the rebound narrative..perhaps. While both platforms (and their volume data is across multiple venues or products remember) saw a decline in FY ADV from 2015 to 2025, both are tracking for their best year in some time in 2026. Over the first six months of the year, EBS has reported ADV of $71.8 billion, which, if sustained, will be its highest since 2019. Equally, LSEG’s spot FX ADV is $116.7 billion – which would be the highest since 2014.

There needs to be some context, however, we don’t know how much of this growth is on the other venues, and while the platforms’ environments are different, the comparative picture is potentially even healthier elsewhere. While both are significantly below their historical highs, since March 2020, which can be used as a useful benchmark for high activity in the “modern” FX market, four of LSEG’s top-10 month have been in 2026, while EBS has had two this year.

Elsewhere, however, FXSpotStream and SGX FX futures have had their six busiest months to date from January to June, while the first half makes up six of the seven busiest months at 360T’s spot franchise. Euronext FX has four of this year’s months in its top-10, including its busiest yet, while Cboe FX has four of its top-five and six of its top-10, all occurring in 2026.

So, a rising tide lifts all ships (and CLS this year has seen six of its 10 busiest spot traded volume months), and primaries have felt the benefit of that, but what of the future, and does the presentation to the ECB group offer any signs?

There is no doubt that a volatility spike or an event attracts the NBMMs to the market, and traders of all colours to the primaries, thus they become more active in volume terms. We are likely to see this environment continue, so there is no reason to expect activity to decline dramatically from here.

clients are voting with their business and proving that FX is a healthy and fair environment

The good news, talking to people in the market, is that the banks seem more comfortable with how the NBMMs are playing. This is in part due to their own tech improvements, but also testament to the nuancing of the rules of engagement that have been taking place on the primaries over the past decade.

They are unlikely, however, to engage more than they do, if my conversations are a reflection of the wider view – mainly because a ‘lit’ environment is still seen as the last place you want to clear risk of any size. Placing interest ‘on behalf of’ is one thing, looking to exit a chunk of risk, is another – although one caveat of this is how one e-FX trader observed that the market tends to react less to encountering a larger, resting, order, than it does when one is introduced at or close to, top of book. I think this is more a factor of a moving market meeting interest than anything else, but could be another positive sign.

Overall though, this looks like a market structure which is firmly established, where the lit CLOBs (and other venues) benefit from uncertainty and vol spikes, but the majority of activity still takes place behind closed doors. There are those that would change this structure, but the clients are voting with their business and proving that it is a healthy and fair environment. Others also believe that the diminished status of the primaries (in volume terms) is a concern, but I would disagree.

To me the primaries remain the barometer of the spot FX market, they provide a ‘clean’ dealing space for market data purposes and their volumes rise and fall with everyone else’s. Most importantly for the market data argument however, is a factoid in the EBS part of the presentation – one that reinforces a point my friend and market veteran Steve Flanagan has stressed so many times. During the yen intervention at the end of April, EBS traded at 98.4% of available price points. That is not to say there was a price every point on the way down, more that there was a price all the way down – a small but important factor. These prices were delivered by bank and non-bank players.

This, to me, shows that the FX market structure is efficient and up to the job, and if that is the case, then the premise at the start of this column is correct – there is little new in the presentations. They do, however, reinforce that what we have works, and should not be interfered with.

colin@thefullfx.com