Trading Desk Automation: Q&A with Capital Group and Nordea Asset Management

Equities, Finance, Trading

Simon Steward, Head of European Equity Trading at Capital Group and Rolf Molkjar, Head of Equity & Futures Trading at Nordea Asset Management are both dedicating a lot of time to scaling up trading desk automation at their firms.

Although at different stages on their journey, a number of the challenges and opportunities they are facing are similar.  In this joint Q&A, both Simon and Rolf talk about the importance of data, building better relationships with Portfolio Managers and the value their internal quant teams offer.

What are the main drivers of automation for your firms?

Simon: For us, the main driver of automation is about efficiency. It’s really not about replacing people or even cost-cutting, it’s simply about streamlining processes that are taking up trader bandwidth by implementing technological solutions. This can obviously take many different forms, so we rely on our trading teams to be able to help identify where there are opportunities. It’s also important that there are clear alignments with our business requirements and the needs of our clients.

Rolf: I’d agree that the purpose is mainly to empower our traders. Automation is a tool to facilitate that. By removing orders that we deem eligible for automation, our traders are able to handle more complexity and focus even more on our most difficult orders where the human element can add a lot of value. If we succeed in that, we are able to improve our execution performance as well as client and trader satisfaction. Automation in itself is not a goal for us.

Simon: Absolutely, you don’t want to automate just for the sake of it.

Rolf: Another thing we think about is that automation allows us to explore and test a lot of things in an easier and more consistent way than we would otherwise be able to, which we believe increases execution performance little by little over time.

What is the biggest challenge you’ve both encountered so far?

Rolf: I think there has been a lot of challenges already! The further down the road we get the more we run into as well – there are always things that you can never prepare for. If I name one, it was getting our traders onboard the project in the beginning. I think this became a challenge mainly because we did not spend enough time on making it crystal clear what our purpose with automation was. If that doesn’t resonate with the desk, you will not have everyone move in the same direction.

Simon: Key decisions around what kind of technology you need to incorporate and whether you ‘buy vs build’ are things we’ve found challenging. This requires careful consideration at the onset around relevant skillsets internally or ensuring you engage with the right third party provider who understands your needs. Establishing a solid rule base approach from the outset has been useful because automation can cause lot’s of downstream implications within businesses. The biggest challenge though is always time! Being able to narrow the time window from the point of concept or identification of an opportunity to the actual roll out is critical as the markets are evolving so quickly.

What are the challenges you face when it comes to accessing and using performance data to drive automation?

Simon: As with all data it’s about accuracy and having enough of it to make statistically relevant decisions.

Rolf: With more and more data, you realise that things are not as simple as we sometimes model it as well! As an example, if we think A, B and C are parameters that affect our execution performance, and we solve for these, we might find that actually there are ten other parameters that affect execution performance that are all interlinked. At the same time, the marketplace is not static either – Market structure is always changing and market participants change the way they execute. This is not an excuse for not doing the analysis, it is just an example of some of the challenges we have faced.

Simon: I’d say It comes down to understanding how you are measuring success. Let’s take the small subset of flow we have identified as being suitable for automation as an example. We have an initial framework around TCA that we utilise for our algorithmic business overall, but we have to be flexible and be aware that there are potentially different variables or outputs that we need to factor in.

Where are you now on the one click to zero click journey? And how much of your current flow is 100% automated?

Simon: The one click to zero click journey is very much front and centre for us currently. This year we have made large strides around this with a subset of our flow executed on a zero click basis (or fully automated). I think it’s important to raise at this point that even when you automate it still requires oversight and in the case of our zero click flow this sits with the traders on the desk. We’re currently at around 1-2% of our overall flow. Although a small percentage, this has provided us with an initial platform to be able to potentially upscale. Based on the structure of our business and our areas of focus this initial automation project will be critical for us as we move forward.

Rolf: We are still at one-click and on track to go to zero click early next year. Currently approximately 40% of our orders are now executed in our systematic process.

How have Portfolio Managers received the drive towards greater automation?

Rolf: The purpose resonates a lot with our portfolio managers as well. Simply put, we have two types of portfolio managers; fundamental and quantitative based. The fundamental PMs welcomes the opportunity to engage even more with the traders on how we optimally execute their orders, and our quantitative based PMs likes the data-driven execution model. Again, automation in itself is not the goal, it is more about the benefits that comes with it.

Simon: We are in an ever changing business landscape and technology is still one of the most fundamental differentiators so everyone is getting onboard with the fact that automation will play a part in success for the reasons I stated earlier. We are using data as part of the story that we tell to our investment group and we use examples from existing automated flow to illustrate the benefits. Like Rolf highlighted, I have seen over the years that the best way to communicate these kind of changes is to engage people early and explain the process, the reasoning and the benefits to achieve the optimal outcome for all concerned.

Do you have a quant team in place to help you? What is the greatest benefit they offer you, if you do?

Simon: We are hugely fortunate to have dedicated resources in the form of a quant team who cover us globally with a presence in each region (and I’m glad to say that presence is growing!)

Rolf: We have an in-house quant team of 4 people, each of them covering an asset class. The benefit of having them sit in one team together is that they are able to share best practice across asset classes and challenge our traders a lot more because of that. Also, the mind- and skillset for in-depth quantitative analyst is not necessarily the same as a traders, so having a dedicated quant team provides an opportunity to have the best of both worlds.

Simon: I agree, our team offer us so many areas of expertise; whether its data interpretation, market structure insights or project driven observations. They have definitely become core to our trading strategies and decision making around automation and system developments. Having the quant team embedded within our process has enabled us to ask better questions of our counterparts and venues alongside helping us evaluate and validate decisions around strategies and automation opportunities. Their insight is absolutely invaluable!

Previous Post
Rules-Based Execution In Rates: Fixed Income Report
Next Post
EMS Platforms Benchmarking: Fixed Income Pulse Report

Related content

Menu