How did the Russian sanctions impact your desk?
The sanctions imposed on Russia have cut off many Russian banks from the SWIFT global payments system and the global payments architecture and have limited the room for manoeuvres for all Russia-related market participants. For us at the trading desk, the most important thing was to always act in accordance with all sanctions. First, we had to have an overview of all our sanctions-related touch points in the back office and settlement processes in order to work off all transactions even under high pressure in a very restrictive environment.
This included questions like: which account is set up at which custodian bank? which counterparty and clearing house fits? and which regulation applies to all participant in the chain? The very close internal cooperation between our central trading desk, portfolio management, operations and legal and compliance helped us to master this task successfully, alongside the good relationship of trust with our long-standing counterparties.
In setting up and establishing the multi asset trading desk, the focus is on optimisation, electronification and automation. During this challenging time, it has proven invaulable to have senior trading specialists in our team, who are able to switch in volatile market phases to high-touch very quickly and who have very sound knowledge of market structures and their participants – thus ensuring a maximum of transaction flexibility while meeting all sanctions.
What is the liquidity landscape like?
We have certainly seen an impact to liquidity conditions this year due to various geopolitical developments, impacts of regulation along with central bank developments.
Fragmentation of liquidity in the FX market is an important topic, that should not be underestimated – the drivers are regulation and electronification, accelerated by market disruptions in the wake of recent crises. Unlike in the past, when market access and information flow went through the market maker, we now have a very different set-up. The control, management and therefore quality of execution depends entirely on our decisions. Therefore, we need to ensure that we have optimal access to the sell-side and various liquidity pools. In addition, the analysis of execution performance is essential to manage our flows – in terms of liquidity and performance. We are steering considerable foreign exchange flows, so we have to have the ability to align our trading and analysis set-up to the benefit of our Group-internal as well as external mandates. Not having the right set-up would be a significant disadvantage, especially in the less liquid currency pairs.
How do you use algorithms?
We use algorithms to select liquidity pools and implement different trading strategies. While we do not develop our own algos, we do benchmark and analyse a changing set of sell-side algos that we use, and differentiate their use by market phase, benchmark/ intention, and currency pairs. The responsibility of buy-side traders has changed significantly – algos are starting to take the place that voice used to dominate. When managed appropriately, they work really well because there is a lot of liquidity hidden even in tight markets. They allow traders to take direct control of order execution, and analysis during and after the trade provides insight into prevailing market conditions. Buy-side traders can use this information to adjust their future execution and improve the overall execution costs for our mandates. The volume we trade via algos has increased during the last two years and will further do so.
What’s your technology set-up like?
We use one OMS for all asset classes and multiple trading platforms. For FX, we use two platforms with different strengths for flow/cash and algos/FX options, but also to have a back-up in place. Improving market access is also an important and ongoing issue on the buy-side, new internal and external specifications, as well as scarce IT resources, can be an obstacle here in the desired flexibility. For this reason, we have also initiated an adjustment of the system architecture with the establishment of the multi asset trading desk. This will allow us to be more flexible in connecting relevant trading venues and information systems in the future, despite the higher hurdles (IT, interfaces, compliance gates). In addition, I see a great advantage in our set-up – to have one system or cockpit – with which all multi-asset traders have the same information available and easy access to all asset markets. This makes automation and low-touch trading easier to scale and leaves sufficient space for the significant amount of high-touch trades we have to carry out deriving from our core business.
How has automation developed in FX, for example in NDFs?
For G10 currencies – FX spot, forwards and swaps – we trade almost exclusively electronically via the trading platforms – the use of algos is well established and has proven itself in various market phases. Rule-based trading is an option for small volumes, depending on the cross. On the NDF side, we still trade a much larger portion via voice. But we also see the benefits of using algo in NDFs – however, the ability to tap internal liquidity or flows in NDFs is lower compared to G10 spot trading. Algo development in this area is still ongoing but has improved significantly over the past year.
Being able to efficiently trade even broken data is a significant advantage. Other obstacles such as contractual requirements, confirmations and determination procedures are well established on our side. We are also very well prepared for Phase 6 in UMR. Looking at the efforts made here, I assume that participants with lower trading volumes probably do not place the highest priority on electronification of NDFs. Generally, when talking about automation – aside from the major currencies – the liquidity landscape needs to be very well understood and kept in mind. It can dry up all of a sudden, making a fixed ticket amount approach for fully automated trading far too large. We have started to look at how we can improve our trading rules using machine learning to make sure that the system takes different factors into account and does not trade a position automatically that is too big.