Building pre-trade analysis into trading protocol selection

Published on 13 October 2021

As traders become more confident in their use of different electronic trading protocols, they need clarity on when and where to use them. Pre-trade tools are providing real support in this area as more formal ‘decision trees’ and less formal qualitative assessments are both used to assess the path to best execution.

Gareth Coltman, global head of trading automation at MarketAxess, tells us how buy-side firms are establishing best practices to get more efficient at selecting the right protocol, including better pre-trade analytics and increased automation.

Dan Barnes: Welcome to Trader TV – your insight into trading for professional investors. I’m Dan Barnes.

The range of trading protocols available to fixed income traders on the buy-side is increasing exponentially. This creates complexity on the trading desk, but also creates better execution outcomes. Joining me today is Gareth Coltman, Global Head of Trading Automation at MarketAxess. We’re going to be discussing how traders can manage these different trading protocols efficiently.

Gareth, welcome to the show.

Gareth Coltman: Thanks, Dan. Thanks for having me.

Dan Barnes: Can you tell us how do you see execution choices expanding in the fixed income market today?

Gareth Coltman: Yeah, I think we’re seeing a sea change really in trader behavior. We’ve got this rapidly evolving market structure. I think most notably, the rise of ETFs has created growing, systematic liquidity provision, but also, I think an increasing confidence from traders.

That technology can now solve for parts of the trading workflow that perhaps in the past was seen as the domain of voice trading only. So we’ve seen the emergence of order book protocols. MarketAxess has launched Live Markets, which is our order book protocol in the US. We’ve seen sessions-based protocols emerging, both traditional, kind of dealer to dealer flow, but also professional clients. Algo is using those types of protocol and we’ve obviously seen portfolio trading recently growing as well.

Dan Barnes: There are different criteria that they might be using for validating best execution. Can you give us some examples of what they may be?

Gareth Coltman: Historically, traders are looking to use competitive prices i.e cover price or other quotes coming back on their RFQ to validate best execution. But achieving the best execution is always a trade off. So obviously, if you want to achieve best price, you may have to adjust the timing of your order. You might need to think about perhaps sweeping up natural liquidity, which is going to be closer, potentially to the mid or the upper side of the spread, instead of just going straight to market and paying the bid ask spread to get something done straight away.

Obviously, again, perhaps you’ve got large size to do, but there’s a cost in doing that. How do you make a decision as a trader with? The best thing to do is to put all of that size into the market immediately or perhaps to wait. But that is changing, right? So we’ve seen increasing transparency, definitely seeing better predictive models. MarkeAxess has this composite plus pricing model as an example. So I think there are now more reliable ways to be able to predict outcomes, but also measure the cost of those choices.

And that’s going to give traders increasing confidence that, just as an example, in this case, I’m better waiting and I’m going to execute through the day and perhaps get my residual down to close.

What we’ve seen, undoubtedly over the last few years is this kind of abundance of data, but an inability to be able to actually consume that data in a practical, constructive way that would help traders make choices. And that’s been a huge challenge. And in terms of making a choice about how I approach the market, whether that is about protocol choice, or timing choice, or even choice of counterparty or the provider, it’s all about the data and all about my ability to kind of process and evaluate that data in real time. So ideally, what I’d be able to do is create a decision tree based on that data based on the likelihood of certain outcomes. This is going to be my approach to market,

Dan Barnes: How might a trader prioritize ttheir orders and what difference might that make to the trading protocols they choose to use?

Gareth Coltman: If the objective for the portfolio manager is just to get cash to work as quickly as possible, that’s going to affect what the trader does in terms of their choice in terms of their pathway to market. The more discretion that portfolio manager can give the trader, the more the trader is able to add alpha or add price improvement to that trading process. Assuming the orders come without specific instructions, that trader is going to take a look at what’s available in the market, perhaps looking at what’s happened historically and make a decision at that point about, ‘where do I think I’m most likely to achieve my objectives?’

We’ve got this increasing fragmentation, I guess you could say in terms of choice, availability of protocol and different protocols will suit different sizes of order or types of securities in terms of liquidity. And then on top of that, I’ve got a whole range of different counterparts and ways that I could trade.

And I think what we will see emerge is an increasing reliance on technology to help make those choices. So we’re already seeing that with these kind of predictive models like CP Plus, where we’re able to say, this is the likelihood of this outcome, but even evaluating all of those different outcomes is a complex thing to do. And so I think we’ll see an increase in the use of automation or at least kind of trade assist technologies that help clients make those choices.

Dan Barnes: What’s an efficient way of being able to access all of those different trading protocols?

Gareth Coltman: We have a vision for a future of automation that allows clients to go beyond simply automating the submission and execution of RFQ, where a client is able to set their objectives or intention in terms of how they want to trade, creating some parameters around things we’ve talked about, like urgency, appetite for price improvement, and then having the system suggests back to them a workflow which has the best opportunity to achieve those objectives.

So that may be that if a client says, ‘this particular order is a large block, but I’m building a position, so I’m happy to do this over time. I want to remain discrete.’ Then they may have a suggested workflow, which would, fx try and sweep natural liquidity out of the market.

If it’s a workflow where timing is more paramount, perhaps they’d use things like RFQ or order book and so on. And I think what automation allows the client to do is instead of having to worry about, well, which is the best protocol for me right now with this order and look at all of that data and try and make all those decisions. The system is able to help make some of those decisions for the trader.

Dan Barnes: Gareth, that’s been great. Thank you very much.

Gareth Coltman: Thanks, Dan. It’s been a pleasure.

Dan Barnes: I’d like to thank Gareth for his insights into trading protocols today, and I’d like to thank you, of course, for watching. To catch up on our other shows or to subscribe to our newsletter, go to TRADERTV.NET.