Read time: 6 min

And Finally…

It’s almost two years to the day since I first started discussing latency floors, so with EBS rolling out its latency floor mechanism last week, I suppose the time is right to have a look at some of the themes that are dividing the industry – this being one of them.

A latency floor, along with the  “Green Room” concept seems to be the new fault line among platform providers, some eschew the concept claiming it is unnecessary, others see it as the panacea to the ills that have befallen FX at the hands of high frequency traders. Clearly much of the opinion formed is done so along business lines, so perhaps we need to take some of the emotion out of the issue?

As an interested but neutral bystander I think the latency floor has some value in the industry, but only as part of a package of measures. I am a little concerned that the concept is a reflection of a desire to return to simpler times – and that, I am afraid, ignores the reality of modern markets and is impossible.

Clearly there are behaviours in the FX market that are unfair and unacceptable, and most platforms accept this but too many hide behind their rules – preferring to promote the rules in place rather than actually enforce them.

Increasingly when I look at the FX market through the eyes and mechanisms of a trader I see a messy landscape where everyone is trying to second guess every bid and offer. In other words, we are at risk of over-analysing everything (and yes I do get the irony of me analysing this!) in a search for the source of the problem. I don’t think the problems have changed that much over the past couple of years – there is a lot of false liquidity out there, people are finding it increasingly difficult to execute large amounts quietly and speed is still something that provides an advantage – but I do think technology has changed how we look at the issues.

I have often thought that an MQL (minimum quote lifespan) that is set at a reasonable level reduces or eradicates the opportunities for price flashing. The latency floor is designed to reduce the opportunities for flashing as well as for “sniping”, but the latter raises a huge issue. A participant could be accused of sniping every time they hit a price and the market moves in their direction, and the liquidity providers’ main protection against sniping is “last look”. The problem is, last look often has the same characteristics of flashing – I don’t like the price so I’m off. That is illusory liquidity as much as an HFT firm stuffing hundreds of quotes into a system with no intention to deal.

In other words, the liquidity mirage is driven as much by some banks’ desire not to sniped or picked off as it is HFT firms flashing prices in. Of course, it is important to note that some liquidity providers are a little quick to throw accusations at counterparties where it is not deserved – the market maker has no inalienable right to make money on every trade immediately. Sometimes, you would hope (especially if you are a “client-centric” organisation) your customers make money from their trading or at least hedge well?

What we are witnessing at the moment is the result, in the case of EBS, of some serious analysis of behaviour on its trading platform, and its subsequent attempts to make things fairer. Unfortunately for EBS, as well as other platforms that to seek to balance things up by means of a latency floor, they may be acting in response to issues that are “old market” rather than “new”.

Why do I say this? Well because it’s hard to get away from the feeling that the industry is responding to human concerns – this in an era when 80% of spot trading (minimum) is electronic, and as such, takes advantage of speed. Like it or not, banks, HFTs, asset managers and even corporates, are using algorithms to execute – and these algos promote fragmentation through aggregation.

The real problem, as I see it, is that too many banks are in a position where they have the MIS available to them that can do nothing but show them where they get things wrong (or their customers right). The reality of aggregation is that you often deal when you are best (or worst for you) price – that is the real value of the concept to the aggregator, you get to see across the market and hit what is for you the best price, or more likely you hit the best average price.

Hitting best average price inevitably means someone gets hit when they don’t like it and it is here that I think we are too “old market”. More than ever, liquidity providers prices are in competition – not like it used to be when it was seven or eight banks pricing – now we have something like 20 or 30 people pricing. This means any slackness or technology slippage is punished – and the MIS will no doubt tell you that in a hurry.

Some LPs hide behind last look as a defence and in some circumstances a latency floor acts as a defence for these people. I am not sure they need defending. I can see why platforms want to protect their LPs, without them they are nothing, but are we risking going “old market” by making things too good for them?

The EBS move is interesting because it has a few non-bank market makers that do a lot of business on the platform and it has probably involved these firms in the consultation process. Just as probably, they appear to have accepted the case for slowing things down, which highlights how this is definitely not a bank versus HFT split.

I fully expect to hear from platform providers criticising the move – it won’t be the first or last time a floor has been the subject of derision – and their argument is equally valid, make sure that everyone has the same latency. The problem is, many don’t, equally they don’t have the budget to spend to making up ground, so what do they do? I think the answer to that lies in the increased market share of a few players at the top of the industry and the lower volumes on many public platforms – those without the budget are going somewhere else and it’s aggregation.

The modern platform has to have balance, and that means no last look and, probably, an MQL and latency floor. My worry is that the world has moved on and now it is all about aggregation. If one looks at EBS (and for that matter ventures like ParFX) they increasingly look the domain of the major players (Thomson Reuters avoids this due to the different make up of its core currencies and the importance of local players in those markets). For lower tier players the cost of competing doesn’t add up so they are aggregating single bank platforms.

This is why I suspect, as I have argued before, that EBS Direct is the future of that business rather than the ECN – latency floor or no latency floor. The latter is another attempt to make things fairer, but I remain unconvinced that it will have a huge impact in a changing world.

I started by noting that it is two years since we started discussing latency floors, but it is about a decade since I started arguing that the single dealer platforms will be the biggest competition for ECNs. At face value, with regulation seeking to curb some of the activity on the SDPs I was probably wrong, but on the other hand, it could be argued that more flow than ever is going to the top group of dealers – it’s just doing it through aggregation.

Our world has changed to the degree an LP is as much likely to get hurt by a bank-sponsored algo execution strategy used by a corporate as it is by an HFT firm trying to arbitrage a market. So much execution now uses speed and the ability to look across markets that trying to slow things down could be counter-productive. Better, it would seem, is an approach to establish tight parameters that everyone can adhere to.

Too slow? It won’t be a huge undertaking to catch up. Too fast? You’re going to have to rein it in. It’s hardly innovation and progress, but it may create the flat market landscape that we all desire.

Of course, I actually don’t believe such a landscape is possible – there will always be one group or another in the ascendancy – it’s the nature of the beast – but that is a discussion for another day.

Colin_lambert@profit-loss.com

Profit & Loss

Share This

Share on facebook
Facebook
Share on google
Google+
Share on twitter
Twitter
Share on linkedin
LinkedIn
Share on reddit
Reddit

Related Posts in