The Last Look…

Not for the first time over the past decade or so, it seems that, for some, expectations around the growth of algo execution tools are not being met, which in turn raises the question, where does this leave some of the banks’ business models in the FX world?

It is not so much, according to some I have spoken with, that algo execution tools are not being used – they are – more that the projected growth just isn’t there. If you’re a user of algos, you probably have been for some time and you know the strategies and providers you like and are sticking with them. More pertinently, at the top end of town, a lot more buy side firms are using their own strategies, as one hedge fund manager put it to me recently, “If you use a third-party algo, that third-party will know intimately how your trading is going to progress – it gives away too much information.”

That latter comment actually just highlights how there is a still a need to build trust, the pricing business of an algo provider does not, or should not, have access to activities on the algo platform – that said, a pricing engine will be able to spot patterns and regular use of one firm’s algo strategy will only reinforce that knowledge. Not only that, but without a decent internalisation model, external LPs will also easily spot the flow through the market – it is notable that while the sophistication level of many algo strategies has barely changed over the years, the ability of LPs to “read” the market has improved multiple times over. The flow is simply less invisible than it used to be.

This has led, I am told, to a switch by certain counterparties wanting to go back to risk transfer for their bigger tickets, especially in uncertain times where volatility could spike at any time.

All of this leaves the banks with a bit of a conundrum – yes, there is more automated trading taking place, which is good for efficiency and the bottom line, but one area pegged (no pun intended) to boost revenues was algo execution services, and that just isn’t happening at many shops.

The desire to push clients to algo execution products rather than take the risk on themselves, has reinforced the impression that the dominant business model in FX at the banks is that of a broker. I am not sure it should be – as I have argued for some years now – because while regulation has forced the hand of many in the fixed income space, for example, where holding risk is an expensive process, in FX, especially spot, there is no great cost. The fact is, though, that is how many senior managers (often from the ‘FI’ bit of the ‘FICC’ business) see it.

The problem with the broker model, as the inter-dealer firms can tell you, is that it is predicated upon ever-increasing volumes if the demand for constant revenue growth is to be met. If activity remains steady – which is what I am hearing it is at the moment, compared to 2023 – the chances of year-on-year revenue growth are impaired.

There is a battle of ideas taking place between those who see nothing but positive growth in volumes and are content to stick with the status quo; and those who believe FX is the market where the banks should be taking on more risk to differentiate themselves – and (re)introduce an extra revenue source

This, in turn, leads the board to do the only thing they know how, cut costs through headcount reduction, with the aim of improving net revenues – which is what we have been seeing over the past few months at some places. To a degree, those banks cutting staff are relying upon achieving a fine balance – something they have historically generally failed to manage – we all see the continued uncertainty in the world as something that will drive higher volatility, but the banks are hoping they have left enough resources in place to handle the increased demands upon the business (work-life balance goes out the window at this stage).

Volatility may save them, but there is a bigger issue bubbling away, one that is related, to some degree, with the lack of real growth in algos. Clients are getting smarter about their execution and even customers that have been with a bank, or small group of banks, are starting to look more closely at their execution quality. This takes a fair few “free” dollars off the table as far as the LPs are concerned.

These firms are not paying the brokerage fees for using an algo, and they are not blindly giving up their business trusting that the price they get is good. I am told that this is being led by the asset manager community, that has finally woken up to the potential savings that can be made through smarter FX hedging (including, I have to say, what some people tell me is the increased use of non-month-end fixes to dilute the impact on those days).

Personally, since the ‘standing instruction’ debacle of more than a decade ago, I think asset managers have been treated fairly by the LPs, but it seems that more and more people on the execution teams are able to show they are adding value by executing in a slightly different, smarter, fashion. As this spreads through the community, the big banks may find that the pip or two that was once available, no longer is. It may be a quarter or half a pip, which works for some, but while the revenue stream doesn’t dry up, it slows to a trickle.

These managers are not using off-the-shelf algos, they are either adapting a third-party’s to their needs or building their own liquidity pools into which they can execute their own bespoke strategies – neither of which helps a service provider seeking the constant comfort of a brokerage-type stream.

One other thing to throw in for non-custodian banks – post-May, when North America is at T+1 in its securities markets, a lot of these asset managers will not have the right processes in place to handle the change, therefore they will, for a period at least, turn to their custodians. This won’t be extra flow to the market, it will come from other, non-custodian LPs. It is fair enough to argue that the banks these firms are dealing with shouldn’t be a counterparty risk, but then neither were quite a few banks that have disappeared over the past couple of decades. I suspect the general thinking is, if the firm isn’t ready, then the best execution policy is put on hold until such time a solution is arrived at to allow the manager to process and execute their trades in time for CLS inclusion.

A couple of conversations I have had this year tells me that this is a subject being debated in some banking circles, where there is something of a battle of ideas taking place between those who see nothing but positive growth in volumes (and therefore brokerage-type revenues) and are content to stick with the status quo; and those who believe FX is the market where the banks should be taking on more risk to differentiate themselves – and (re)introduce an extra revenue source.

Banks study their client bases very carefully…If they are, they are probably starting to see some unsettling signs

There is no doubt where I sit, after all I have argued this for some time – banks should be taking on more risk – and by ‘risk’ I mean for hours or days, rather than just milliseconds or a minute or two. It is significant to me that hedge funds are adding traders ahead of what is expected to be a busy time in markets, so the question should be asked, why aren’t the banks? It’s probably because too many can’t see the value in having risk in a spot business that they have grown used to as a brokerage-type model, but if the existing revenue sources dry up? That changes the narrative.

There is also the question of customer service. More clients are looking for risk transfer – one way to differentiate an FX LP is having inventory that can be used to help the client with their execution. I understand that many LPs will argue that is exactly what they do now, only the inventory is held for a few seconds and in not very large amounts. My argument is that at least one part of the business should be looking to maximise revenue from existing flow, as I have argued before, while also looking to use that inventory to improve execution quality for other clients. I am sure there is a way that a bank can come up with a smart formula to reward better client outcomes as well as profits.

Banks don’t always get it right when it comes to the business cycle – there is a standing joke in some of my circles that you can tell the top/bottom of the cycle if the banks are getting in out (never the right way of course) – but if they do think of FX as being a little different to fixed income and equities (diversification anyone?) then it strikes me they will see that shrinking revenues from client volumes can be compensated for through higher proprietary trading revenues.

Banks are very keen to tell us that the customer is at the centre of everything they do, which suggests they study their client bases very carefully. If they do, they are probably starting to see some unsettling signs, but the answer is not to blindly cut staff, more they should take a leaf out of their clients’ books and think smarter about their FX business in particular.

After all, the inter-dealer brokers had to evolve and diversify to stay alive…

colin@thefullfx.com