A lot of people have been talking to me about an old theme recently, only it has been given a new name – information leakage. Typically I have always referred to this as signalling risk, it means the same after all and has the same outcome, but either way there are more and more people worried about it.
In some ways this is just another element in changing market behaviour, whereas a few years ago a single ticket trade for $10 million was normal, nowadays it is a rarity, especially on the public platforms – everything is about breaking the ticket down into small packets.
The problem with breaking the larger ticket down has been discovered by some players and they have moved to more “full amount” trading with their LPs, however the tricky matter of information leakage is now being discovered by players further down the food chain in terms of average ticket size – and their complaints are not finding such a receptive audience among their LPs.
Not that long ago hedge funds, for example, didn’t mind a bit of market impact. If it cost them two or three points slippage on the order they didn’t care because it was a speculative position and the market impact could easily have been 10 points or (much) more. The trouble now is that LPs generally are much wiser to their tactics and they quote these accounts wider if they see them in a bilateral environment. This, in turn, has accelerated the move by those hedge funds to the more transparent markets where, paradoxically for such a transparent environment, they are anonymous!
And it is here, where these more aggressive traders have been forced, that they are having their problem with information leakage – and it is now that they are trying to solve the issue and finding it a complex issue without an obvious solution.
Back in the day, a trader could line up 15 LPs, hit them all for 10 million, have market impact and not care. Now they don’t have access to the 10 million streams (or if they do it is very wide thanks to their history) and so they have to get into the position (and more importantly out of it) in ones and twos – hence the information leakage. As I stated in this column recently, the CIA’s adage of “a secret shared is a secret squared”, works in markets as well – the more you trade the higher chance there is you being spotted.
There is actually a more fundamental reason for the information leakage of course, and it’s one that is not often mentioned because it doesn’t suit the agenda of most people involved at regulatory or participant level – and that is electronic trading.
Put simply, everything we do online, in any walk of life, leaves a digital footprint and as such it can, and will, be spotted. It’s just a question of how long it takes and with some firms seemingly spending as much time on technology aimed at sniffing out these orders as actually predicting short term trading patterns, that time horizon is typically getting shorter by the week.
So it is increasingly difficult in the current environment to hide the parent order, what is to be done? Sadly the only obvious solution runs the risk of incurring regulatory disapproval and possibly sanction – and that is to occasionally trade the opposite way, to throw the hounds off the trail. This treads uncomfortably close to spoofing (although I would argue there is a definite intention to deal) and just doesn’t sit well in an environment in which everyone must know what everyone is doing. The regulators want transparency, even if it undermines an end user’s performance.
Even halting the order when the market starts getting stretched rarely helps, I am told, because the pattern seekers are used to those sorts of gaps and very quickly see the order resume. Using multiple platforms in different products? Yep, they can spot that as well apparently, so simply there is no way around it – welcome to the modern FX market.
It is because they have researched this problem that some of the larger buy side firms have reverted to full amount trading or alternatively taken all of their execution in house to take out one very small piece of the pipeline that leads to leakage. For those firms without the track record of a longstanding relationship, or are viewed as too clever by half by the majority of their LPs, no such solution exists unless they can convince their LPs that they are full only with them.
The chances of that happening are remote because quite frankly a lot of LPs are over-sensitive already and are questioning those clients with whom they have had a great relationship – how will they react to someone they have thought of as “toxic” for the past two decades?
Another problem is what constitutes full amount? Is it the parent order in its entirety, or just the child order being offered to one LP only, to be followed by another? In the latter case there is also the question of how long the LP should be allowed to hold the risk? It could be argued that $10 million in USD/JPY only takes a few seconds to clear out and as such the executing trader should be clear to go with another clip and some LPs understand and accept that.
What is being forgotten, however, is that the LP clearing the risk is having impact and causing information leakage – it’s not always about who is doing the trading (it is sometimes of course), the basic act of trading will provide that digital signature no matter whether it’s the originating trader or the LP.
Someone suggested to me today that hedge funds and the like that are having these issues need to put something in their disclosures to investors as it could, if it persists, impact returns. Certainly investors may need to be aware of these issues but given how in the hedge fund world the big appear to keep getting bigger (and this has to be an issue across markets) I am not sure anyone seems to care!
Going back to FX, there are a couple of things that could help. The first is obvious – deal with a big internalising LP. The problem here has already been discussed, though, does that LP want to see your business at the type of spread that you have become accustomed to?
The other is a little old-fashioned and actually doesn’t really solve the problem, but it does delay it (an information speed bump perhaps?) By all means initially execute electronically, but if and when the market starts looking overloaded – which would prompt a pause by the executing algo – take that pause, and then trade a few clips by voice.
There are all sorts of complexities that would need to be taken into account, and I am sure that the first challenge will be from those seeking to measure best execution because they may not have a benchmark or it may be 0.00003 worse than top of book in EUR/USD. The fact is though, the very act of trading by voice (especially in a era in which people are fully aware of their obligations when it comes to order confidentiality) would slow the process down and, possibly, throw off a few of the “sniffers”.
Some firms already recognise best execution can include leaving top of book alone in the public markets of course, perhaps this message needs to spread a little further, because I feel confident in saying that an order execution that involves less information leakage will ultimately prove to be better than one in which a group of influential players can spot it after a few seconds.
Information leakage is a very real problem in the FX market now and it is starting to impact the execution experience of the end users (who I think, and the regulators will correct me if I’m wrong, are still the ones we are meant to be looking after?) The problem is the demand for transparency and for everything to be measured against an irrelevant, and frankly misleading, benchmark that is top-of-book.
Like most things FX there is a pendulum effect here and currently I would argue that it has swung against those trying to execute larger tickets. Some will argue, with no little credibility, that for some firms it is a question of chickens coming home to roost for they have exploited the LPs chase for market share for a decade or more and now have to deal with a new reality.
I accept that, but this issue is also starting to impact genuine end users and for that reason there should probably be a much deeper debate over how the industry goes forward – including a discussion on topics such as counter orders within the parent order, that some people will feel uncomfortable discussing.