As always when I write something vaguely controversial I feel the need to follow up on last Thursday’s column – because some people either got the wrong end of the stick on what I was trying to say or I feel the even greater need to beat them down on their views of last look.
There is something about this topic that stirs the industry up more than anything else since the days of counter accusations of clients being read by their LPs and those same clients, according to the LPs, spraying the market; and in many ways last look is the modern day version of that argument. It exists to protect the LPs from predatory traders and, as I outlined last week in this column, some LPs could stand accused of abusing it.
So to the several discussions I have had about last week’s missive. I want to start off by summarily dismissing those that merely reiterated the counter argument I noted in last week’s column – no, I am sorry, you are not “waiting for the order to come back in the client’s favour” – you are waiting for the trade to come back into your favour. And again – if the flow is that toxic they why bother quoting for it, or can your ego not allow you to turn it down? If the flow isn’t toxic why are you artificially holding it?
I happen to believe this is not a question of ego, rather it is a manifestation of something I have written about a lot over the past few years – entitlement. There are LPs out there who do not understand that a genuine LP will take a hit on some flow – it won’t be deliberate but the law of averages dictates it will be so. There are firms out there who privately brag about how they have one, maybe two, losing days in a year from their FX business. That is not the profile of a genuine LP, it is one of a participant claiming to be an LP but in reality using last look, or perhaps cherry picking flow to an extreme level.
More thoughtfully, there were people who pointed out – correctly – that there are many clients out there who don’t mind high reject rates and several described the use of last look as a negotiation process, it was really a question of getting the hold time right to suit both parties. I accept this (you know what? Actually I don’t, but never mind) but it misses the point I was trying to make. As things stand with the FX Global Code, last look is to be used for a market and credit risk check and some of you argued that a longer hold time for certain flow was a market risk check. Fine, but as I – and I am glad to say, many others – see it, that makes the decision to implement a longer hold time a commercial one. An LP can do a credit check in microseconds and a market risk check according to the update pulse of whatever venue they are using – there is no flexibility around this, it doesn’t matter who they are quoting, their ability to conduct that risk check doesn’t change in temporal terms.
There was something of a misunderstanding on some correspondents’ part over exactly what my broader message was then, so let’s make it clear. Either the FX Global Code needs to be updated to say this practice is OK and what disclosures are necessary around it, or it needs to add a line stating that last look is not to be used for commercial purposes and that variable and asymmetric hold and response times are not good practice.
If the decision is to go the first route then transparent and good practices surely dictate that every client is told their average synthetic hold time over a month for each channel through which they trade (i.e. the hold time before the market and credit check takes place) and how that differs to the LPs average for all other clients – not just customers who trade in a similar fashion. This does not need to be, or should not be, a public disclosure, but if the LP is not forthcoming with the data, it is the type of information the client should demand if they are to hold their LPs accountable.
There were some who believed I was calling for the banning of last look – as much as I would like to be able to do that I understand this would, in the context of a modern market, leave LPs vulnerable to predatory trading practices, so no, there is no intent here to seek a ban on its use. Equally, one or two people thought that my suggestion that the LPs widen their spreads, was exactly the commercial decision that I was railing against and again, sadly, I obviously wasn’t clear enough. I have no problem with LPs taking a commercial decision when deciding how to treat certain flow, my problem is with how they do it and how transparent they are about it. Widening a spread is fully transparent – it says, “I don’t like your flow but I think at this spread it can’t hurt me”. Sneaking in an artificial delay for some clients but not others, latency buffering as it was once called, is anything but transparent as things sit.
One correspondent shared some analysis with me they had done as part of a consultancy gig, which indicated that for flow with a relatively high level of toxicity the spread would only need to be widened by between 0.00003 and 0.00005 to make it economic. Obviously all flow is different in its own way so generalising is dangerous, but the fact that someone volunteered this information means that it can be worked out. Is there an LP out there that doesn’t benchmark the value of the flow over the course of milliseconds to seconds? Simply work out how far away the market is on average at 100ms and build that into the spread you quote – no artificial hold time, a quick price that if dealt upon will satisfy the customer and, probably, the LP.
There should not be a stigma attached to wider pricing, but sadly a sense of entitlement has also emerged amongst buy side firms whereby they expect such a valuable (and hugely under-valued) commodity to be more or less given away – and it has to be said that the LPs themselves in their hunger to grab as much business as they can have fuelled this.
My final thought on the feedback is a worrying one. There were plenty of people in agreement with me, which is heartening, but amongst those that weren’t there was a hint of defensiveness around the whole use of last look. All feedback is gratefully received and remains on background, but it bothers me how quick some LPs are to jump to the defence of various uses of last look beyond what it is expressly (not expressly enough clearly) meant in the Code.
I may be wrong and even a little paranoid, but this defensiveness immediately makes me think these people have something to hide in how they use last look – and if that is the case then I fear that this camp will once again strive to ensure the Global FX Committee skirts what remains the one serious issue facing the FX industry when it comes to behaviour. It is to be hoped that the reformers win the day – if they do not, then last look will not go away as an issue, it will continue to fester until one day it becomes the next grenade to land in our laps with the pin pulled.