Broaden traders’ pre-trade execution choices through post-trade price data

Published on 6 August 2019

Better prepared buy-side traders make better execution decisions. Chris Bruner, head of US credit at Tradeweb outlines how buy-side traders can quantify execution quality through the use of pricing tools in the pre-trade, at-trade and post-trade workflow. Quantifying execution quality, delivers feedback to buy-side desks so they can be ready for any orders, even in volatile markets.

Dan Barnes Welcome to Trader TV Fixed Income – your insight into the trading climate for professional bond investors. I’m Dan Barnes. Buy-side traders need data to build a picture of what is trading where in the corporate bond market today, but it’s hard to get. Chris Bruner, head of U.S Credit at Tradeweb, is going to explain how execution policy can be quantified, and how that quantification can be used. Chris, welcome back to Trader TV.

Chris Bruner Thanks, Dan. Nice to be here.

Dan Barnes So tell us, how easy is it for bond traders to quantify execution quality today?

Chris Bruner It’s traditionally been fairly difficult in the corporate bond market, and I think the necessary input is a great price at the time of the transaction. And because of the disparate nature of the securities, that’s been pretty difficult to achieve intraday. I think the overall pre-trade picture gets better and better, but actually, knowing for every trade you do that there is a high quality indication of what the level is, contemporaneous to where you trade, has been something that you couldn’t really achieve until fairly recently. One thing that we’ve been working on for the last few years to help with that, is what we call Ai-Price. It’s a service that prices over 17.000 corporate bonds in real time. So every 30 seconds, it’s spitting out our estimate for where we think that bond would trade, if it traded right now. So, the idea is really to be truly real-time, working from a trader’s point of view of where they think that price would be. And we have a lot of ways to validate the quality of that now. So fx, on our automated RFQ technology we use that price as the check, and it won’t trade unless it’s close enough to where Ai-Price says the price should be. Or at least it will get kicked out to be manually looked at by a trader before the execution occurs. So what we see is we know that the price is getting better and better coming out of the algo engine, because over 80 or 85% of all of our auto-RFQ trades are actually executing where we predicted it was before. It has been traditionally quite hard to build transaction cost analysis tools, or know exactly where a bond is trading before you trade. Now that we have something that has broad coverage and we know it’s accurate to within a fairly good araband, you can actually now deploy things like transaction cost analysis, that were available and other more mature asset classes into credit, and give clients the tools that they’re used to having, to look at actual execution cost.

Dan Barnes What are the elements of execution that can be captured in that sort of snapshot?

Chris Bruner It’s quite disparate and growing, I’d say. So traditionally it’s been messages and indicative pricing runs, and we try to provide as many of those things on our screen as we possibly can. At Tradeweb we’re very focused on providing services and data for high touch trading, as well as for low touch trading or for electronic RFQ protocols. So, we’ve got traditional sources that people are used to parsing out of messages. We’ve got a growing set of axe and inventory content, which is much more API driven, it’s connected up via programing interface. Fx we’ve got about 30 billion of live inventory and axes on the platform, that we’re constantly trying to sharpen up for clients so that they know if it’s there, it’s real. And if you wanted to engage on it, hopefully someone will show you a strong price, because they advertised it out to you. And if you combine those types of datasets along with Ai-Price, now you’ve got a much better picture before you’re trading for every security, really. Not that you can trade necessarily all those securities in huge size on an amazingly tight price, but you get a much better picture about which ones you can, which ones you can’t, and everything in between, basically.

Dan Barnes So what sort of demands are your users making on you at the moment?

Chris Bruner So, I think, Dan, what users are asking for us is how do we take these data services and transaction cost analysis tools to help them make more informed decisions about execution? We’ve worked for quite a while to provide a range of trading protocols for our clients. We focused on high touch flows to make larger sized trades more efficient, to make the processing easier, all the way through to all-to-all RFQs, all the way through to fully automated RFQs on the AiEX. So, you don’t even have to see them come out of your OMS, you just see executions flowing back in. So, we’ve had this wide range of protocols, but one of the key things for clients is, ‘how do I use which protocol, and which situation, and for which types of trades?’ And one of the things we hear the most is, we need to move beyond just best desk rules, and I need three quotes fx, to something which is really related to the average transaction cost, and breaking down that average transaction cost by liquidity score, by size, or by side, or by market condition.

Dan Barnes And then in terms of at-trade and post-trade, how would you describe the way you’ve responded to those demands?

Chris Bruner With at-trade, what we’re trying to do is help inform their execution decision tree. And what we see clients doing now is moving beyond just a binary decision tree, which was, for really large trees, I call people up and I get on chat and I do voice trades. And for smaller trades, I do electronic RFQ. It’s now becoming more at-trade, they might be looking at the outputs from our transaction cost analysis tools, to find where that transaction cost below, say, for certain sizes and liquidity scores, it’s below a certain threshold. They’re very comfortable to try to fully automate that. And one of the questions now is, ‘how much can we automate?’ The other way at-trade is really by providing new protocols, such as portfolio trading. Now you can trade whole baskets of bonds with single providers and soon multiple providers, instead of just doing one RFQ at a time. And then post-trade, I think most of the focus is going on to research into the quality of the data that drives transaction cost analysis. By no means is that over yet. We keep trying to refine both the scope of the bonds we can price as well as the quality of the pricing. And the more we use the the pricing for different activities in the trading context at-trade, the better the post-trade analysis gets that we can build off the back of it.

Dan Barnes You’ve talked about at-trade and post-trade, how does that feed back into that pre-trade process and how do you see that supporting your clients?

Chris Bruner It varies a lot by client, and and that’s interesting for us, because some clients, they just want the data and we can do that. We try to be very careful about how we do that. So we get the 30 billion of axes, as I mentioned on the platform, we have the ability to put that in pre-trade to a client’s own systems if they want to do most of the work themselves. But we try to protect the data, make sure there is legal usage for both the people that are really providing the data, which in this case is the sell-side. In some cases it’s us as well, because it’s our all-to-all inquiry, and then the clients can use it the way they need to in their own process. And then down the scale, there’s people that need us to do everything.

Dan Barnes Chris, thank you so much.

Chris Bruner Thanks, Dan. It’s good to be here today.

Dan Barnes I’d like to thank Chris Bruner of Tradeweb, and of course you for watching. To catch our reports in other markets, or to subscribe to our newsletter, go to TraderTV.NET.