The Last Look…
Posted by Colin Lambert. Last updated: February 20, 2023
There appears to be an eagerness around the industry to create a buzz around “peer-to-peer” spot FX trading, but I am not sure I quite understand why.
It may just be a question of language, but the fact is, if someone says “peer-to-peer” there is a suggestion (at the very least) that this is yet another initiative to disenfranchise the banks (DeFi is another area where this seems to be the ambition). Some of the P2P initiatives, such as the recently unveiled Loop FX, make a point of embracing the banks, others cite their P2P platform as being great for the bank algo businesses, which they may be, but others genuinely believe that there is potential in a “pure” P2P model.
The problem, as I see it, is represented in a couple of uncomfortable facts. The first is that the banks remain comfortably the largest segment in the FX market and, along with one or two non-bank firms, are genuine risk absorbers of the kind many buy side clients want. The second is, the FX industry already has a spot P2P model, and often it is an absolute nightmare!
Last week saw a Bank for International Settlements paper published which highlighted how dealer constraints had a direct, and negative, impact on liquidity in FX markets. I am not totally onboard with the paper – it cites capital impacts but is analysing spot markets where that is much less of an issue, if one at all – but the core of its argument is sound; the FX market functions better when the major dealers are able to participate fully.
How then, do people expect large buy side firms, who often have limited experience in FX markets, to step in and play that role? Many of these firms rarely want to take FX risk, so will they be willing to wait for a match? I very much doubt it.
There is an even more uncomfortable truth for those pushing pure P2P, or even the use of bank algos to match this flow, as noted that is that the current model we have for it doesn’t work. The core premise of the pure P2P concept is that there is significant off-setting flow amongst buy side firms, but if that exists, one has to ask the question, why does the 4pm Fix work better?
The last month end saw a significant move in USD/CHF and USD/CAD and on regular occasions there are equally significant moves in other pairs. This is not about the length of the fixing window, it is merely an observation that the vast majority of flow in these windows are from the buy side, and there are always significant imbalances that create market impact.
It is also remarkable to me that some people think there is nothing in excluding the banks from these sessions because there is significant risk among these firms. Why shut out the majority of the market just for the sake of saying you are “peer-to-peer”. The fact is that the reaction of many bankers to the phrase “peer-to-peer”, is often one of hostility, although not amongst the prime brokerage and algo teams!
I do not understand the apparent eagerness on the part of some people to exclude a (large) chunk of the FX market. Using the phrase “peer-to-peer” is only likely to make it harder to achieve success.
So here’s an idea. Why not actually call these models what they really are – dark pools? There is a body of opinion that dark pools have negative undertones to the regulators, but if they are well-managed, with clear rules, I don’t see that. A significant proportion of US equity trading still takes place in dark pools, why? Because the buy side firms don’t want to go to the public markets unless they have to. They actually prefer the model, but – and this is the important factor – they don’t care who they match with, as long as they do.
Spot FX already has a dark pool of course in BGC’s platform, but that is limited to the major dealers and as such is equally a closed shop. That said, it is, reputedly, a very successful model, with banks telling me they are pushing some impressive numbers through the platform on a daily basis.
As far as I can see, the dark matching model has real potential in FX, because there are a lot of off-setting interests, as there are in equity markets. What I do not understand is the apparent eagerness on the part of some people to exclude a (large) chunk of the market. Using the phrase “peer-to-peer” is only likely to make it harder to achieve success, whereas a dark pool, which operates in exactly the same fashion (and probably has minimum interest thresholds to keep information sniffers out), but that is open to everybody (subject to the aforementioned rules) could be the equivalent of leaning on an open door.