Across asset classes the electronification of trading is leading to increased levels of automation, but are some trades impossible to automate?
If you start from the right place, automation can be added in part at least, to mast trades argues. Charlie Campbell-Johnston, head of integration and workflow solutions at Tradeweb. He says that the successful automation of any part of the trading workflow must begin with a clear view of investment and trading goals.
Traders can then potentially add automated elements, across the spectrum of asset classes, if they can pinpoint the data needed to determine the outcome of a trading decision, and that decision is well-defined. The upside can be considerable, leading to increasing numbers of enquiries from trading desks with potential partners.
Dan Barnes: Welcome to Trader TV – your insight into trading for professional investors. I’m Dan Barnes. We’re seeing automation increase across both the buy- and sell-side across different asset classes. Joining me today is Charlie Campbell-Johnston, Head of Integration of Workflow Solutions at Tradeweb to talk about which trades he’s helping to automate and which trades can’t be automated. Charlie, welcome to the show.
Charlie Campbell-Johnston: Thank you, Dan.
Dan Barnes: Let’s start with the parameters which guide best execution, what sort of objectives do most trading desks have and how might they be affected by the types of funds and portfolios they’re trading for?
Charlie Campbell-Johnston: No two trading desks really have the same set of parameters, but broadly, it will be a combination of certainty of execution, speed, price, i.e. the transaction costs. But yet there will be different clients or fund managers with different trading objectives. So fx, one case would be an index house looking to trade at month-end, a high volume of tickets over a short period of time. Certainty of execution and speed are pretty important. If you’ve got a more active fund manager fx, transaction costs is likely to be far more of a consideration. So the ability to limit your transaction costs, either in cash terms, basis point terms and so forth, potentially at the expense of a likely higher hit rate may be more appropriate. And we’ll tend to work a lot with clients to be able to define those rules on an order by order basis to basically be in tune with those execution objectives.
The other thing I would say is that on the back end, it’s really important for our trading desk to be able to prove to their own clients that they’ve acted in accordance with these rules. So people will often use our transaction cost analysis tool to say, ‘this automated ruleset has had a positive impact on our transaction costs and therefore we could potentially look to extend it into other asset classes or other use cases.’ If it’s a question of comparing against the close, or maintaining a client hit rate to the certain level. Again, those same tools can be used for that same objective.
Dan Barnes: Do you typically see different execution parameters that are specific to certain asset classes?
Charlie Campbell-Johnston: One of our largest asset class adoption is in the ETF space. Quite often ETF trade is come from equity backgrounds, so they have a different approach to automation. Automation seems very much as something that should be embraced. In that case fx, speed is important. There’s less concern about market impact because these things are quite often and traditionally traded on exchange. But if you go further down the liquidity spectrum, that may be more considerations. More considerations such as; ‘should I be showing my RFQs to the whole market? Should I be showing my side before I trade?’ For a fully cross-asset, class desk or even one that trades a single asset classes, but with a wide range of liquidity within that asset class, an automation tool needs to combine really a commonality of approach.
What I mean by that is from a technical or a workflow perspective, you’re able to start off in one asset class and potentially move into other asset classes without doing a completely new build. So if you have a full understanding of the technical things you need and the workflows that are available to them, this then allows you to dip your toe in the water in something that you’re very, very comfortable with and then rapidly expand as you get used to it, whilst also understanding that there are different nuances in different asset classes.
Dan Barnes: In terms of the barriers to the success of a trade, how does automation specifically overcome those?
Charlie Campbell-Johnston: I think this depends on the definition of success and what the trading desk is looking to achieve. If it’s purely a scale play, the benefits are clear. Increased capacity, decreased operational and transaction costs. The traders role is still paramount, but the nature of what a trader is doing is changing over time. A lot more integration with market data sources and analytics, but also from the original case of a capacity issue. We’re seeing that automation is actually bringing new types of trading activity to the fore. One example would be the use of systematic funds in the leveraged space, being able to use automation to trade in a different way that they had before. And to that point, if you are looking to set up a fund, what we’ve built is the ability to be able to integrate into our automated tool, direct from a trader spreadsheet.
So the set up cost, which previously had been the implementation of the order management system or an execution management system, can now be removed if someone wants to get up and running very, very quickly. So fx, we’ve seen a marked increase in the index space of people using a functionality that allows orders or what we call time release orders, to be sent out to a certain point to the day. That allows a trader to trade in a very specific time with specific execution objectives, often in a very busy time that allows them to free up time for other kinds of trading activity. The focus on that may be to get a large amount of trades done at a certain amount of time, and at the closest point to the close.
Automation works well when markets are benign. OK, so when counterparties have their auto quotas on, you can be pretty confident of getting responses from a certain amount of counterparties according to certain conditions. But what we found, particularly about this time last year when the COVID-19 crisis started, is that people want to be able to adapt, that is to say still use automation, but also be able to adapt those rule sets to varying market conditions.
In volatile markets, a hybrid model, if you like, a sort of lite version of automation is something that really comes into its own so that traders can still benefit from the full array of parameters, settings and so forth, but they still have that last minute control to say, ‘OK, I’m prepared to pay a little bit more, a bid of the spread. I’m prepared to wait a little bit longer.’ And we’ve seen that trend continue quite significantly, but particularly when things are a little bit more challenging.
Dan Barnes: How can traders ensure that all the right factors are in place to allow them to start building an automated trading capacity?
Charlie Campbell-Johnston: Working with a trusted partner with experience in this space does give you a little bit more comfort. Second, even though these trading desks do have various different trading objectives and so forth, they still have the same things that they need to address. They want to be able to know that the partner they’re working with has the experience to understand that, but also has worked with people like themselves to be able to get a little bit of an understanding of what options are available. And that’s important for us, because it means that the nature of our interaction with clients is very much a collaborative innovation, so that we can make sure that we are designing protocols and so forth that fit in with not only one client’s objective, but also understanding how that can be sent out to other people.
Dan Barnes: Which trades can’t you automate, then?
Charlie Campbell-Johnston: The time may not be always appropriate for certain people, particularly in times of high volatility. There may be a full retrenchment from automation, but what we have seen is that those who’ve adopted it have been flexible enough to understand how they can adopt it and not overly rely on it, to make sure that they can use it as part of their overall strategy.
Dan Barnes: That’s been great, Charlie. Thank you.
Charlie Campbell-Johnston: Thank you very much, Dan.
Dan Barnes: I’d like to thank Charlie for his insights today and of course, you for watching. To catch up on our other shows, or to subscribe to our newsletter, go to TRADERTV.NET.
