The global FX code – Digging deeper on ‘last look’

01 Jun 2021, 10:25

Share this

‏‏‎ ‎

This article counts towards your CPD Points

What would you see as being the code’s greatest success so far?

The GFXC has already firmly established itself as the point of reference for best practice in the eFX market and that is something that deserves to be acknowledged. Awareness of the code, what it stands for and the growing number of participants who are signing up also illustrates just how much a framework was needed – and perhaps the fact that inaction simply wasn’t an option. Failure to deliver here would have almost inevitably meant some kind of attempt at coordinated international intervention and regulators have endorsed the progress made so far. But with the code now poised to see its second major review, it seems clear that further improvements can be made.

What do you think needs to change?

As it stands, the code is arguably still too liberal. Whilst it has done a lot to advocate transparency and disclosure, questions still need to be asked as to whether the current approach offers too much latitude. A more prescriptive approach from the code in terms of exactly what is acceptable when it comes to aspects such as last look and the treatment of rejected orders is probably warranted.

What’s your view on last look?

Last look remains a significant point of contention. It’s easy to dismiss this as being one-sided, but the optimal situation – of all offers being available immediately at an unlimited market depth – is simply unachievable. Something has to give and whilst those counterparties who are sitting on the very highest levels of internal flow may have an advantage here, their motivation in this debate needs to be clarified. Last look provides an invaluable risk control mechanism, ensuring no extreme price movement has taken place and also that the trade remains within the counterparty’s operational limit – although again this should always be conducted on a symmetrical basis.

There is a concept known as ‘zero hold time last look’, which is a great step forward - but there needs to be a degree of realism as to what can be delivered here. That’s both in practical terms and also in order to keep the market functioning by allowing counterparties to manage risk. Even with the fastest processing, there remains an absolute time required to validate and execute the order, and still check it falls within operational limits. That said, this should be measured in a couple of milliseconds. If necessary, counterparties should be advised as to why the order was rejected, and the speed of this execution should also remove the ability for the provider to undertake any unscrupulous activity such as pre-hedging.

In this vein, the handling of rejected orders also warrants further scrutiny within the code. As a business we are strongly against the use of rejected order information in any subsequent real time decision making. Using such valuable data to bolster a risk management strategy at the expense of the client is in our opinion a wholly unacceptable practice.

Do longer hold times always result to worse pricing?

CMC Markets conducted research with Tradefeedr back in 2019 to analyse the impact longer hold times had on costs in the event trades were rejected. Whilst the obvious correlation holds true – greater latency results in higher rejection costs – it was interesting to note that the adverse impact wasn’t always in evidence if latency could be limited to just a few milliseconds.

When it comes to our customers, we don’t by default use last look, although by working in conjunction with them both in the onboarding process and throughout the relationship, we will find the most appropriate solution. After all, they should also have a vested interest in avoiding toxic or sharp flow. However, our investment in technology is paving the way for a move to zero hold time last look – something which will see an order validated against the client risk metrics and the price checked in a fraction of a millisecond. There’s still some fractional latency but this is transparent and looks instantaneous - whilst enabling risk to be properly managed.

Can transaction cost analysis (TCA) be done consistently across multiple liquidity providers?

At CMC we’re strong advocates of TCA and work with established third-party providers to ensure they can offer existing and prospective clients alike the insight they need to make better decisions. Ensuring market wide standards here would have the potential to add more clarity to the situation, although experience has shown that not infrequently – and perhaps quite surprisingly - the buy side have no interest in running these metrics. That shows what could be perceived as an atypically human aspect in this market, but one that perhaps also acknowledges that a degree of room for manoeuvre here and the value of personal or legacy relationships is truly valued, especially by the more experienced participants.

There are some genuinely smart tools on the market to track TCA and it’s vital that the buy side take this seriously to help make sure assessments are truly impartial. We work with a third party to ensure integrity here, but any moves to standardise within the code how this analysis is undertaken would have to be seen as a positive step.

Does the buy side have to accept that if they want access to the tightest spreads they will likely have to accept some form of last look in order to protect the LP against predatory flow?

That is a reality, but the last look window can certainly be shortened. As we already mentioned, the perfect world of unlimited depth, at-market pricing and instantaneous execution cannot be delivered in the long term if there is to be a sustainable, efficient market. With that in mind, it’s critical that the motivations of those who want to see further select reform to the GFXC are fully disclosed and properly understood. The eFX market is of such scale that it would be difficult to imagine monopolistic behaviour lasting for any length of time, but the potential for short term distortions to be overly beneficial to individual parties cannot be overlooked, either. Ultimately, participants need to understand and be comfortable with the total cost of the transaction, including the adverse impact of orders not being filled and re-quoted price being seen.

To conclude, which is more effective - eliminating last look or monitoring transaction costs appropriately?

Counterparties simply need to be aware of what’s going on, so that’s why the focus here really needs to be around transparency and disclosures. Last look serves a role when it comes to ensuring the trade is within acceptable parameters, but any hold time needn’t be for more than a few milliseconds. Those who claim otherwise ought to come under a high degree of scrutiny so the cause of any delay can be fully understood, alongside what they are doing in that window. And a robust approach to transaction cost analysis should always be taken too, even if you are being offered zero hold time last look. Ultimately that’s what enables counterparties to be held to account if the service delivered in the longer term falls short of those promises made at the outset.


Contact us

Find a member of our team in your local region and get in touch today

Meet the team

Contact us to find out how we can work together to open up new potential for your business. If you would like to speak directly to a local member of the team you can find their details here.