The FX Live Scribe

We recently ran our FX in-person meeting for our London based buy side community, looking into workflow & execution technology and desk structure & management efficiency. Some of the key discussion points, takeaways and photos from the sessions are captured below in the form of a scribe…

<ul> <li style='font-weight: 400;' aria-level='1'>Using algos for order placement needs questioning – how well do you understand how your algo interacts with liquidity?</li> <li style='font-weight: 400;' aria-level='1'>Peer to peer had many benefits when it comes to minimising market impact and information leakage.</li> <li style='font-weight: 400;' aria-level='1'>The number of options available in the algo space adds to a traders already complex role – how can you pick the correct algo providers? </li> </ul>

<ul> <li style='font-weight: 400;' aria-level='1'>Algos are a value add, a great evolution on traditional voice trading, but how they interact with liquidity needs to be better understood. Not all trades should be executed via an algo.</li> <li style='font-weight: 400;' aria-level='1'>Understanding the true cost of a trade is important – sometimes the cost is not just monetary.</li> <li style='font-weight: 400;' aria-level='1'>We are all doing peer-to-peer already</li> <li style='font-weight: 400;' aria-level='1'>We want to know when to use algo over RFQ</li> <li style='font-weight: 400;' aria-level='1'>P2P requires critical mass and the buy side must commit</li> <li style='font-weight: 400;' aria-level='1'>Peer to peer liquidity needs a critical mass</li> </ul>

<ul> <li style='font-weight: 400;' aria-level='1'>Beware of concentration of the FX market with 5 or 6 banks</li> <li style='font-weight: 400;' aria-level='1'>Algo wage in the street is only 15-20%</li> <li style='font-weight: 400;' aria-level='1'>It’s the liquidity that determines good algo execution</li> <li style='font-weight: 400;' aria-level='1'>Fragmented market and rivalry responsible for slow algo uptake</li> </ul>

<ul> <li style='font-weight: 400;' aria-level='1'>Understanding liquidity, requires algo is integrated </li> <li style='font-weight: 400;' aria-level='1'>Increased internalisation will lead to a more fragmented market and worse performance</li> <li style='font-weight: 400;' aria-level='1'>When and how to use algos needs to be established</li> <li style='font-weight: 400;' aria-level='1'>How effective is using algos to execute FX?</li> </ul>

Why are our clients not yet using FX algos as much as equities traders are? Algos usage globally 20% FX compared to equities 70% FX is way behind in terms of data analytics on the FX side. They do work and they are the way we should go. In the equities space it’s all exchange traded so that is the big difference. For FX, it means that there’s a struggle to access every bit of liquidity in the market at that time. We use algos a lot, probably around the 20% mark. Volatility is so high, in general the risk transfer you don’t take on – risk goes along with the cost of execution. What we need to do is find that sweet spot. This is what we’re doing from our side.

Currently, FX data and analytics lags behind that of equities, however buy side heads agreed that FX algos on the back of this are still critical and the direction in which FX trading is headed. In the equities space it’s all exchange traded so that is the big difference. In contrast, for FX, there remains a struggle to access every bit of liquidity in the market at that time. Still, many members use FX algos frequently (some stated circa 20% of trading). Over the past year volatility has been so high, so in general the risk transfer you don’t take on, and risk goes along with the cost of execution. The buy side are on a continual journey to find this sweet spot between risk and best execution.

Analytics is still a major area of development for FX, and buy side heads are looking to keep investing heavily in this area. Post trade data and analytics are now an industry norm, but FX still lags some way behind their peers in Equities for pre trade and intra trade analytics. Buy side are looking to improve access to execution data, and standardise this to inform insights on when and where to execute. Once data is standardised internally, buy side heads hope to use this to feed into automation efforts and drive efficiencies and automation frameworks into their trade lifecycle.

<ul> <li style='font-weight: 400;' aria-level='1'>Cost benefit analysis</li> <li style='font-weight: 400;' aria-level='1'>Still a big need for qualitative data</li> <li style='font-weight: 400;' aria-level='1'>Stress scenario</li> <li style='font-weight: 400;' aria-level='1'>Glad the market is growing because it shows us options</li> <li style='font-weight: 400;' aria-level='1'>Balance sheet impact on pricing FFX</li> </ul>

<ul> <li style='font-weight: 400;' aria-level='1'>There is a need for advanced integrated OMS for trade generation</li> <li style='font-weight: 400;' aria-level='1'>FX trading is still difficult with restricted currencies in particular</li> <li style='font-weight: 400;' aria-level='1'>Restricted currencies are a challenge to automate, often this is out of buy side control</li> <li style='font-weight: 400;' aria-level='1'>Generating spot and hedge off of cash traders is a challenge</li> <li style='font-weight: 400;' aria-level='1'>No one seems a 100% satisfied with their OMS</li> </ul>

<ul> <li style='font-weight: 400;' aria-level='1'>Automation is a moving target</li> <li style='font-weight: 400;' aria-level='1'>How to avoid capturing custodian flow?</li> <li style='font-weight: 400;' aria-level='1'>Spot + Hedge together as an independent TCA</li> <li style='font-weight: 400;' aria-level='1'>Restricted currencies through a single provider rather than via each separate custodian would be beneficial</li> <li style='font-weight: 400;' aria-level='1'>Similar pain points across the board and similar needs so nice to know it isn’t us being alone in it</li> </ul>

<ul> <li style='font-weight: 400;' aria-level='1'>Still problems with automating workflow</li> <li style='font-weight: 400;' aria-level='1'>Pain points for FX workflow: email, similar hedging</li> <li style='font-weight: 400;' aria-level='1'>FX algo decision tree trade selection process</li> </ul>

<ul> <li style='font-weight: 400;' aria-level='1'>Automation needs to stack with data </li> <li style='font-weight: 400;' aria-level='1'>Still a lot of work to do to improve automation</li> <li style='font-weight: 400;' aria-level='1'>No Standardised process for how to execute trades</li> </ul>

Buy side agreed that whilst automation has truly kicked off in FX, there is still a long way to go. In particular, automation of any sort is still extremely difficult when trading restricted or emerging market currencies, and many buy side reported having issues with or being unsatisfied with the functionality of their O/EMS. Some of the key advice shared by our buy side members included; ensuring you have an understanding that automation is a moving target and there is no specific end goal, there will always be another step or another process that can be simplified and made low/no touch; It is important to to start small when automating trades. Begin with small ticket, vanilla trades and build out automation with rule-sets to build confidence in automation internally before you look to expand this to larger tickets in the future. It’s also important to remember that automation is not there to replace the trader, but to supplement and complement their role moving forward.

<ul> <li style='font-weight: 400;' aria-level='1'>Many constraints exist for workflow interoperability currently – its difficult for providers to cater for all people</li> <li style='font-weight: 400;' aria-level='1'>Resource constraints are a burden, moving towards more no touch is the dream</li> <li style='font-weight: 400;' aria-level='1'>Although there is a move towards low touch or low touch there are issues to be resolved still </li> <li style='font-weight: 400;' aria-level='1'>No touch cash in ticket and leverage in our transactions</li> <li style='font-weight: 400;' aria-level='1'>It’s not integrating, but designing a solution that is challenging and requires low cost and interaction</li> <li style='font-weight: 400;' aria-level='1'>Container solution v walled garden -> evolution of FX space (FX models)</li> <li style='font-weight: 400;' aria-level='1'>I know a bit about technology, but not enough</li> </ul>

<ul> <li style='font-weight: 400;' aria-level='1'>Peer to peer vs central limit order book leads to further development</li> <li style='font-weight: 400;' aria-level='1'>Combined platform for all asset classes is the way forward</li> <li style='font-weight: 400;' aria-level='1'>TT settlement with P2P solution</li> <li style='font-weight: 400;' aria-level='1'>Need for platform that combines buy side liquidity alongside bank liquidity</li> <li style='font-weight: 400;' aria-level='1'>Challenges in data integration between systems – cost of resources</li> <li style='font-weight: 400;' aria-level='1'>Charles river is the app store for FX</li> </ul>

There are still many constraints in workflow interoperability for buy side heads of desk, and it is difficult for providers to cater for everyone. As trading desks and traders themselves become more multi-asset focussed, there is also a need for your tech stack to be efficient and easy to use for not just seasoned FX veterans, but potentially for equity traders to come in and trade FX within the same system, seamlessly and efficiently. The ultimate goal for many FX desks is to move more flow towards low or even no touch, but resource constraints and securing internal buy-in are challenges that still need to be overcome.

Buy side members were in agreement that the trading desk of the future will be fully focussed on automation, with centralised platforms allowing access to multiple 3rd party algos and other solutions, and the buy side wanting the option to create decentralised technology stacks and solution wheels to best fit trading circumstances at the time of trade. Buy side members also feel that blockchain powered peer to peer FX markets will be developed before 2030, which will significantly impact the future role of the trader. While this was a question that cannot yet be answered, key considerations now include; what skill sets will traders need in the future to meet ever changing technology demands; How will the role of the trader develop internally; Will roles be more focussed on viewing markets, or will traders remain execution focussed?

NDFs have offered considerable benefits for emerging markets (EM) FX execution for several of our buy side members. Some EM trades are particularly problematic, and to improve NDF trading buy side heads have firstly established which custodians they work with who have similar mandates and were engaged in the market, then established through analysis when is the best time of day to execute to minimise market impact. As a result of this, the highly illiquid EM currency pairs have remained voice traded, and they have been able to stream the rest electronically.

<ul> <li style='font-weight: 400;' aria-level='1'>Algo execution in NDFs is rapidly advancing</li> <li style='font-weight: 400;' aria-level='1'>Still a bit of a wait and see, not all are concerned or interested in the electronification so drivers not quite there to progress it further and faster</li> <li style='font-weight: 400;' aria-level='1'>Electronification – helping to prove best execution</li> <li style='font-weight: 400;' aria-level='1'>NDF trading is often a small part of overall flow. Hence not always the priority for investment ROS challenge</li> <li style='font-weight: 400;' aria-level='1'>NDF electronic execution would be a great tool as long as the liquidity is there and the pricing points are reliable to show best execution</li> <li style='font-weight: 400;' aria-level='1'>Banks will only stream in a certain amount and then disappear? No fixing risk in data but for more so for broken dates</li> <li style='font-weight: 400;' aria-level='1'>Buyside demand for streaming NDFs will be the deciding factor in its development</li> </ul>

Where are you at on the NDF electronification journey so far? At a tier one asset manager, they launched some funds many years ago called ‘double-decker’ funds. They were essentially emerging market funds for Japanese clients. They found rolling Brazil trades particularly problematic and felt they had to find better ways to execute them. At this time, they were trading a yard a day. However, a lot of the custodians had similar mandates and were engaged in the market at the same time. They then looked at when the best time to execute was in order to avoid noise in the market. It was important for them to roll their positions before the custodians did. They decided to execute via voice and NDFs. They now trade all other currency pairs; from Asian NDFs to Latam. They primarily trade illiquid NDFs (Korea, Brazil) via voice, and for the rest they stream them electronically. When you are trying to trade with Egypt for example, it is difficult from a user perspective as the FX risk at the moment is fairly low. They predict that ‘FX will be back’ and Asia will be a big part of the focus for next year. There are some algos out there that help you trade NDFs. The beauty of trading NDFs is you can trade on whatever date you like. However, if you go down the broken dates route, you don’t necessarily get the best outcome; you might have a good price, but not the best price.

<ul> <li style='font-weight: 400;' aria-level='1'>Clearing could save a considerable amount of capital expenditure (up to 30!)</li> <li style='font-weight: 400;' aria-level='1'>Still don’t know enough about UMR eventual outcomes</li> <li style='font-weight: 400;' aria-level='1'>It seems that most look at UMR as an exercise of managing exposure rather than how clearing might help</li> <li style='font-weight: 400;' aria-level='1'>For execution a clearing price would be beneficial</li> <li style='font-weight: 400;' aria-level='1'>Complexity cost of adopting clearing can be a barrier to reaping benefits</li> <li style='font-weight: 400;' aria-level='1'>It could be pricing that drives the move to clearing NDF</li> </ul>

Why aren’t you clearing your FX? The short answer is there is no regulatory need to be clearing our FX. I don’t think there is a pricing or best execution need either. UMR regulation would be what brought us into scope, so we’ve set up a working group to look at this. Been running for a few years now, but it seems at the end of the day for us that nothing will be in scope. For us to be in scope, our exposure has to be above 8 billion, across all asset managers, not just one. So when we trade for our clients, we aren’t trading for ourselves under our own ISDA, but for our clients. We had to look at which of our clients may be above the 8-bil cap, any we thought were, we reached out to them, about 20 globally, and asked them to feedback on whether they would be in scope. Around 15 did, then we had to establish which of those clients trade NDFs – of which there were 5 or less. Initial margin needs to be above 50 million also, we are nowhere near that margin. So it doesn’t look like there is a benefit for us here as its not regulatory mandated for us. But would like to see if that is really the case, and if clearing offers us any tangible benefits for the execution and pricing of our trades.