The Last Look…
Posted by Colin Lambert. Last updated: December 6, 2022
The recent move by some venues to prioritise FX Global Code-only liquidity has helped dilute the sense that too many platforms hide behind the excuse that they can’t police behaviour effectively because of the fragmented nature of the market, but it remains clear to me that there are those that continue to have their head in the sand, and it’s about time consumers called them out on it.
It is coincidental that I am writing about this the week after we published analysis of a Euronext paper that seeks to help bring understanding to behaviour on the platform, you can read it, and my thoughts on the conclusions, here, and I should stress, what I am about to go into does not relate to that venue in any way.
The bottom line is, not for the first time, I have seen data from a platform that suggests to me at least, that non-Code compliers perform poorly, and provide worse outcomes for consumers.
I remember writing about round-trip times in a late column for Profit & Loss, in which I highlighted one LP on a non-ECN platform having a response time of over 2 seconds on rejects – and yes, that’s seconds, not milliseconds. Since then, things have improved, but it is far from an even improvement, and the laggards seem to be, surprise, non-Code signatories.
The data represents activity on a non-ECN platform, and generally speaking, things look pretty good. The vast majority of the top 20 are Code compliers (17 actually, which I understand is a fair representation of how it is across the industry) and hold times are, with one or two exceptions, the same, or close to the same no matter whether it is a fill or a reject. One or two Code compliers have what look to be lengthy hold times, in excess of 20ms, but they could be quoting largely in lower tier currency pairs where market data is slower, or be located a long way from the matching engine.
It is interesting that only one LP is actually at zero in terms of holds and rejects and, I must confess, it still grates on me that there are Code-compliers with longer reject holds than fill (about a third of compliers). One assumes, as Code compliers, that the asymmetry is disclosed and it is probably fair to point out that these LPs still have pretty low reject ratios at sub 3%, although I also feel obliged to point out these are actually at the high end of the scale when looking at all Code-compliers.
Non-Code signatories have an average hold time of 85ms – just the eight-times longer
Where things turn murky, inevitably, is when we look at the non-Code signatories; “performance” – and I use that word very loosely – blows out. Compared to the compliers’ average around the 6.5ms mark for holds on fills, the three non-signatories manage to hold for in excess of 40ms. It’s even worse on rejects, where compliers on average manage to get it done in just over 10ms, but non-Code signatories have an average hold time of 85ms – just the eight-times longer.
You want to know about reject rates? No problem. As noted, Code-compliers are sub-3%, and the non-signatories? Just the small matter of almost 25% – so one in four trades are rejected, and you can bet the cost of those rejects will be high.
It would be great if someone could explain to me how these LPs are not playing games. More importantly, it would be great if a platform could explain to me why they think this asymmetry on their own venues is acceptable? They should be embarrassed at allowing this type of thing to take place.
Too often I have been told they can’t police behaviour because of fragmentation, but in this instance, the performance difference of the three non-signatories is so stark, surely they have to ask questions?
It is up to the consumers to use this data and bring the platforms to heel on it, and some will still refuse to do so, which is both a shame and shameful.
In my op-ed of the Euronext paper, I observed that while occasionally non-signatories may perform better, this rather ignores the whole – specifically that Code-compliers are affected by what goes on and have to change their behaviour. The data described above merely reinforces the point – amidst a sea of participants doing the right thing, there are some rogue operators likely to bring the whole crashing down.
It is up to the consumers to use this data and bring the platforms to heel on it, and some will still refuse to do so, which is both a shame and shameful. There is, however, change in the air. More consumers are using data than ever before and they are starting to see the true picture, thanks to access to all the above datapoints plus things like cost of rejects.
This will not necessarily mean they will move platforms, customers rarely do that, but it may encourage them to add a second venue in time, one that will see more of their business. The other aspect is that this data will also be available to prospects – and if they do their due diligence right, they will go with the venues that not only have the best and fairest trading environment, but the ones that actually keep deviations from the norm in check and provide all the necessary data to allow the consumer to monitor its LPs.
The clock is ticking on this ridiculous situation, and those platforms allowing non-Code signatories to behave as they wish have due warning – adapt or become irrelevant.
