How to build low-touch, multi-asset trading desks

Published on 17 February 2020

Leading asset managers are building low-touch, multi-asset trading desks to deliver automation wherever possible. Building automation into the trading process requires the asset manager to define best practices around volume, asset class and timeliness of liquidity, then building appropriate rules into the execution process.

John Adam, global director, Portfolio Management & Trading Solutions at FactSet, says trying a one-size-fits-all approach results in one-size-fits-none outcome. Instead, specific expertise in asset class trading is needed to ensure that the right parameters around execution quality are built into the execution management system, to deliver efficiency with no loss of quality.

Dan Barnes Welcome to Trader TV, I’m Dan Barnes. Asset managers can deliver highly efficient access to markets by automating trading across asset classes. John Adam, global director of portfolio management and trading solutions at FactSet, is here to tell us how firms are realizing these benefits. John, welcome to Trader TV.

John Adam Thank you for having me.

Dan Barnes So, tell me, what does automation look like in trading today?

John Adam It’s a labor-saving device for the traders. Specifically, they’re able to take the orders that are perhaps low parts of the average daily trading volume, residuals or otherwise straightforward orders to execute, and have some level of automation rules routing within the execution management system. That enables those orders to get to the market automatically with the appropriate monitoring so the trader can intervene if something goes out of band.

Dan Barnes How do you see this evolving over the short to medium term?

John Adam I think it’s made some structural changes to the buy-side trading desk. The most notable one that I’ve seen is we’ve begun to see desks, that rather than segment around asset-class silos, segment around high touch, low touch and no touch trading. What that means is there’s enough commonality now thanks to the technology that’s in place, that traders can trade multiple asset classes through the same view, and have rules and automation that allow them to trade across multiple asset classes. This, in turn, frees up their time to look at ad-trade analytics and choose their moments of intervention so that they can be the most impactful in their roles and add the most value to the portfolio lifecycle.

Dan Barnes Those low touch automated desks that are trader-cross assets, that’s a new and very exciting development.

John Adam It really is. I think that it’s bringing some cross pollination, too, to the way that we look at trading different asset classes, particularly OTC, for example, you can see a strong uptick in the amount of algorithmic trading. In FX, we’re beginning to see the foundations of automation in fixed income trading. The asset class become more electronic and we’re beginning to see more methods of executing than have ever been available for fixed income. In order for there to be automation, there must first be electronic liquidity. We have yet to figure out a way to automate a voice trade. However, particularly in places like IG credit, sovereign debt, we’re starting to see more and more electronic liquidity, which means that for the foreseeable future we’re looking at a bifurcated market for fixed income. There’s going to be the asset classes that are liquid enough to trade electronically, and therefore there’s going to be some level of automation that’s allowed, or rather there should be some level of automation that’s enabled for those asset classes. But again, in fixed income, experience and relationships matter. So for the illiquid names, the knowledge, the expertise of the trader is more important than ever.

Dan Barnes That’s very true. And so if I’m a head of trading, and I want to optimize automation on my trading desk today, what do I need to be thinking about?

John Adam I think the things that the head trader should be thinking about, are the basic set of rules or best practices that they want to see their traders executing day in and day out. That’s foundational to automation. I think with all of the discussion around it, it’s very easy to get blown off course by discussions around data science, around machine learning, and artificial intelligence. But really, at the heart of it, trade automation is about defining a set of best practices for how an order should be traded based on the volume, based on the asset class, even based on the time of day or the news and events around when that trade hits the desk, and beginning to codify those in the execution management system, so that we can see consistent behavior in the low touch trading. Once we’re sure that that’s working right, that can move to low touch, and most importantly, having the correct set of alerts and protocols around low touch and no touch trading, so that the trader can intervene when something happens that is not accounted for in the set of rules that I’m using to automate my trading. It’s very important when looking at trade automation that there’s an open platform to which I can integrate the various market side analytics that are going to drive that. And then also that I can get the data out of the system through reporting, through analytics, and be able to look at my performance post trade, so that I can assure that the rules are doing what I coded them to do.

Dan Barnes Taking best execution policies for different asset classes would seem to be a complex thing? If you’re turning them into a single process, do you see that as being challenging?

John Adam Yeah, that gets right to the heart of what’s challenging about best execution. If I try to reduce everything down to a one-size-fits-all policy, I get one-size-fits-none. So the technology is very powerful and very capable of automating the training for various asset classes. But we can’t forget that individual expertise in each of those asset classes, and knowing how they trade and knowing the DNA of that particular asset class is equally important. So the ideal best execution policy is flexible enough to understand that there’s different benchmarks based on how an order comes in. The type font comes in from, as well as the market side conditions. For example, if I’m trading a highly illiquid name, my best option might be to call the minimum number of brokers that have a specialty and trading a certain illiquid fixed income instrument. That way I can minimize my information leakage. And that’s been the same challenge with best execution that we’ve faced for years. However, if I’m creating a highly liquid equity, a different best execution policy should be in place. So I should have a consistent approach across multiple asset classes. However, the policy and therefore the automation, should be intelligent enough to recognize the different liquidity mechanics of each of those asset classes.

Dan Barnes How do you see trading desks putting automation into best effect today?

John Adam Perhaps it’s a boring place to start, but we have to put pen to paper and start with, ‘where is the desk spending their time today?’ ‘What are the most time consuming trades we have?’ ‘What are the ones that are the most impactful to the overall performance of our fund?’ When we start there, we can then look at various techniques for automating them, as well as distributing the order flow.

Dan Barnes John, thank you very much.

John Adam Thank you for having me.

Dan Barnes I’d like to thank John Adam and, of course, you for watching. To catch our monthly reports on other markets or to subscribe to our newsletter, go to TRADERTV.NET and ETFTV.NET.