So it’s “WMR Day” and I’m worried.
The new methodology for the WMR benchmark fix comes into being today following the aborted start in December and while the development is to be welcomed – I think most people saw the FX Benchmark Group recommendations as a sensible, proportional response to the existing problems – what I have seen over the past two weeks has me a little concerned, not for the industry (for once!) but for the end-user.
The jury remains out on exactly how many banks will charge for providing access to the Fix, but my understanding is that the majority will charge in some shape or form; of course it only takes one not to, to trigger a race to the bottom again! And, while I am on the subject, readers may wish to check out this story that appeared on Bloomberg recently – if nothing else it shows how little goodwill journalists there have for the FX market, as the what I consider outrageous headline “Banks Will Charge Extra for Not Manipulating FX” demonstrates.
Either way, back to my concerns.
I have just spent two weeks in London and arrived back with not only a chest infection that may mean this column makes slightly less sense than normal, but also a month’s worth of material for this column – starting with the feeling that price action around the 4pm Fix is going to change dramatically and not necessarily for the better.
The problem started nagging me when the third bank showed me its WMR Algo that it is going to use to execute client business (at the end there were six and probably more exist). Sharp eyed and regular readers with a good memory will no doubt be surprised at this turn of events given how, when this problem first reared its head my solution was for these customers to use algos, but I want to stress I thought they should use algos at more appropriate times of the day – not the 4pm Fix.
Obviously it is the right thing to do to hand over execution of these trades to the algo or agency team, and most banks are prepared for this either through taking a trader off the desk to execute manually in a sealed environment, the creation of dedicated execution teams to handle all orders (including stops), or handing the orders to the algo execution teams which are segregated from the mainstream FX trading business.
So why am I worried? Quite simply I always worry when I see several similar strategies executing in the market at the same time. As one unafraid to give a prediction (the motto is “sometimes right, sometimes wrong, always certain”), I could see things working out as follows.
The banks receive the Fix orders and at the open of the session they stick their bids and offers into a mechanism (EBS Fix seemed to be the mechanism of choice among bankers I spoke to). These orders will be placed by the algo with discretion, meaning there is a real chance that most of the matching will take place in the first 30 seconds to one minute.
All well and good, but we still haven’t solved the problem of the residue. Most of the algo teams told me their algo would passively bid and offer the balance and fill the order that way, but what if there are seven algos all chasing the same ultimate bid or offer? If the banks are able to attract other flow to their liquidity channels (and, importantly, the algo can access those liquidity sources), this problem may not exist; but when it comes to professional traders one senses that they will be watching the market carefully around the Fix to see which way the wind is blowing.
In other words they will choose not to stand in front of the runaway truck; rather they will wait for the orders to force the market in their direction before executing post-Fix.
So we could be in a position whereby the algos are all bidding increasingly aggressively, and then switching into full aggressor mode to fill the order within the window, leading to exaggerated price action towards the end of the window. This is likely to mean a reversal of the FXBG’s analysis, wherein the one-minute window saw a surge in volume but not in volatility. I suspect that we will see a smoothing of volume but micro volatility spikes within the five-minute window.
Interestingly, traders I spoke to about this agreed with me, but algo providers typically did not.
But of course, the banks don’t care what happens in the market. In fact no one really cares what happens in the market, as long as they can hit a benchmark. And this is my problem (but no one else’s). I suspect that simple analysis of execution around the Fix will increasingly show a wider gap between arrival price (I prefer to use commencement price, given how orders have to be submitted early) and Fix price, caused by the need of the banks to execute within the Fix window.
One can only imagine the carnage in the market if this happened over one minute, but thankfully we don’t have to worry about that. In fact we don’t have to worry about a thing because the benchmark will be set. It just bugs me that the quality of execution will be inferior.
Either way, welcome to the brave new world. I want to stress this does not mean me diluting my support for the FXBG’s recommendations; rather it represents a reinforcing of my original opinion (and a point raised in the FXBG report) that customers using the Fix may want to do some serious assessment of their execution practices.
By all means use algos, I have always believed this is the most sensible and fair solution, but also can I humbly suggest that users demonstrate a little market “nous” and execute away from the still flawed mechanism?
Colin_lambert@profit-loss.com Twitter @lamboPnL