From market-makers to matcha: 10 years on an FX desk

I’ve just achieved – or endured? – the milestone of 10 years on an institutional FX trading desk. Either way, it feels like the right moment for a little reflection.

Let’s start with an assumption: this morning I tapped in with my contactless card for the Elizabeth line, sipped canned matcha, scrolled TikTok and started writing this article on Zoom – all before 8am. Things have changed for me since 2015. (As well as becoming somewhat depressing, you may say, looking at that list…). Surely, the FX market must have changed too?

Well … has it?

When I sat down to list the big shifts over the past decade, I hit a wall. Then it struck me that the limited changes over the past decade are telling.

Yes, e-trading has exploded. But just doing more of something that’s been around for 10 years wasn’t the compelling argument for a fundamental structural change I was hoping for.

The FX market has evolved, but perhaps not transformed

Auto-execution is arguably the primary change on the buy side. Yes, bank desks might be a little leaner and more hybrid in responsibilities. But if you dropped a retired FX salesperson from 2015 back onto a desk today, they’d probably find the landscape surprisingly familiar – albeit with a few more dashboards and a lot more credit and onboarding discussions.

So why, if both automation and electronification has made execution faster and more efficient, do buy-side trading desks often look the same size as they did 10 years ago? From my perspective, it’s because automation doesn’t necessarily replace work – it just changes the work and broadens the remit.

On our desk, a single team can execute frontier FX trades, historic rate rollovers, sovereign, supranational and agency bonds and interest rate derivatives in a day. My gut feeling is that a decade ago, that breadth of tasks would have required more headcount.

(And to the juniors reading this: yes, you’ve got more on your plate, but at least you can do it from home in your pyjamas on a Friday when you want to!)

This is a mature market, but there’s room for more change ahead

This raises an interesting question: is a broader remit really more efficient? From one angle, yes – you’re doing more with the same resources. From another, does specialisation suffer?

Beyond automation, other changes have been introduced under the guise of improvements, but are they really?

Many clients now pay for the credit they consume, which is arguably a healthier long-term model – but is it a bad thing that credit now costs a bank magnitudes more than it did previously?

Banks are better capitalised. However, this has made their appetite for derivatives more volatile – especially at global systemically important banks (G-Sibs). Is the marketplace more efficient if portfolio FX hedging costs remain the same, yet under the surface banks are coming and going, decreasing and increasing appetite as they suddenly work out that they’ve worked up huge risk-weighted asset (RWA) costs? Depends on your perspective.

The spot market has also been reshaped by non-bank market-makers; private funds have expanded and increasingly blurred into institutional trading; and best execution requirements have grown – adding an emphasis on justifiable execution and a defensible audit trail alongside the pursuit of truly ‘best’ execution.

All of which is to say: the FX market has evolved, but perhaps not transformed. The leap from 2015 to 2025 appears less dramatic than the leap from, say, 1995 to 2005 (although I wasn’t around, but then again, neither was ice cold canned matcha in London).

The next decade may yet bring that elusive transformation. For now, desks continue to operate under a familiar structure – only broader in scope, more automated and electronic in execution, and, at times, stretched thinner in specialisation.

This is a mature market, but there’s room for more change ahead – maybe enough for another article, if I can squeeze it in between my second matcha and an RWA discussion with a G-Sib….

Editing by Lukas Becker