Government Bonds: Autoexecution vs Autotrading

Published on 7 February 2019

Autoexecution of government bond trades can reduce risk and cost for investors, but is only available in a few markets. Traders need to take account of liquidity and investment goals to assess when auto-ex, auto-pricing and non-competitive trades best serve investment goals.

In January’s show, David Scilly, head of fixed income and Currency Dealing at First State Investments, and Frank Cerveny, head of markets and sales at MTS, discuss how to leverage automated trading in government bond trading, in order to deliver best execution.

Dan Barnes Welcome to Trader TV fixed income – your insight into the trading climate for government bond investors. I’m Dan Barnes. Government bond markets are more electronic and corporate bond markets, making them ripe for automation. In January’s show, David Scilly, head of fixed income currency trading at First State Investments, Frank Cerveny, head of markets and sales at MTS, will be discussing how traders can use automation to reduce cost and deliver better execution for investors in the rates space. David, welcome to Trader TV.

David Scilly Thanks very much for having me.

Dan Barnes How varied is automation in the markets you trade?

David Scilly So the response summaries can be quite varied. Obviously, with the US treasuries when you go there, you can go out to just about anybody and you’ll get a price almost instantaneously and you can just execute on it straight away. Across the G10, that’s typically the case. Everyone has a nominal limit, if you like, which all counterparties can trade to. And when we talk about automation, we might want to also just talk a little about the automation of the pricing engines, and the fact that the order staging that we’re doing can also be automated, such that we don’t even look at the trades and they’re going through and they’re actually just going executed automatically. In the G10 you can pretty much do that, but then when you start getting into some of the smaller countries you might notice, that the response summaries sometimes are not as many as you thought coming back. Or you might also notice that the responses are taking more time to come through.

Dan Barnes And what effect does having more automation for a particular instrument have on the approach traders take to execution?

David Scilly Particularly in the government bond markets, there’s automatic pricing and there’s automatic execution. For me, you want to try and align those two things as much as you possibly can where it makes sense to do so. With us, if you’re just doing simple hedgin- type exercises, or you’re doing small portfolio, rebalancing-type trades, which are high frequency, but low volume, then definitely the sooner and the faster you can get them out and executed, the better. It’s more about the post-trade and the reporting aspects of it as opposed to the pre-trade, ‘I need to spend a lot of time thinking about this.’ And some portfolio managers, if they’re putting through a portfolio replication basket of trades, there could be 100 orders in there. And as a trader, not being able to aggregate them and get them off your dashboard as quickly as possible is something which is a bit of a concern to you. So, we find at the moment that doing low touch is sufficient in the sense that we can aggregate easily through our dashboard. We can get the trade over and done in an electronic platform, any of them. And that’s happening in a sufficient time frame such that we don’t feel like we need to order execute the whole thing now, which would be filtering them out to the point where we’re almost not seeing them, and they’re just getting executed in the background. The concern that we have a little bit about automation itself is, that you need to build some good sets of rules around what is okay for auto execution and what’s not. You certainly need to still have that oversight as a dealer to make sure that you’re getting sufficient response summaries, sufficient frequency of participants in those markets, so they’re not just going straight through without any oversight whatsoever.

Dan Barnes What does the desk need in terms of data and technology to engage in the automation process?

David Scilly You definitely need access to all the different trading systems that are out there. There’s probably, as I said, about three or four of them. I think you also need to have the pipes into all the people that are actually there doing the pricing. So, if you were going to take a no touch type philosophy, you need to have all the participants in that market. You also need to have a degree of instructions and parameters within your OMS, sometimes order instructions, these kind of things which would actually help you direct those trades to the to the smart order router. And then you need the smart order router itself, as I understand it, which actually would then direct to the different trading systems neededt. Probably one more thing which would be helpful, is to have real-time or not real time, but as quickly as possible trade reporting and transaction reporting, such that you can see, at the point of execution where you should be getting set, obviously, It could be that we have 200 million government bonds to buy on a particular day, and we’ve got all day to do it. So, you could just spend your time working with some of the algorithms, some of the execution venues just to get bits and pieces done over time. And because those markets are so liquid, they can clear that risk off relatively easily, and you do have that opportunity. But the real trick, I think, with these markets is that you don’t really know until after the fact what’s going to be left behind and what your actual market depth is at that point. And  this is where I think quite often you might find in a size like that you might be better served going to one particular bank who’s going to take it on in principle, take it into their balance sheet at a level which you think is reasonable, at that point in time.

Dan Barnes And that’s the challenge of best execution sometimes, isn’t it? Non-competitive trades or non-comp trades can be just as effective as trades where you put three banks in competition?

David Scilly Yeah, you’ve always got to look at the market characteristics at the time. You might be going into an event. You could be doing anything which requires you to get done in a shorter time frame. Or you might have all day, because the market’s been relatively stable and there’s not really as much risk, and you can just afford to dribble it out. The way you explain what the market’s doing, and the way you’re explaining the characteristics of the order, is the key part of the transparency that people want from best execution.

Dan Barnes What can a smart trading desk do in order to optimize its use of automation, to support best execution and therefore investment objectives?

David Scilly A smart trading desk definitely needs to see these things as opportunities, they need to be able to execute on them, if they think it’s appropriate to do so. But I think that they should also use the information that these venues are providing, to effectively change their methodologies when they need to change them. So, there needs to be enough smart trading desks and they will have enough reporting to be able to look at their strategy, look at how they approach that particular order with those characteristics, and say, ‘did we do it the right way or is there another way that we could have done this which would have actually improved the outcome for the client?’

Dan Barnes David, thank you very much.

David Scilly Thank you. It’s good to finally be here.

Dan Barnes Now we’ll get the latest updates on automation in government bond market, speaking with Frank Cerveny, head of markets and sales at MTS. Frank, welcome to Trader TV.

Frank Cerveny Thank you, Dan. It’s great to be here.

Dan Barnes So where do you see the most demand for automation in the rates lifecycle at the moment then?

Frank Cerveny Dan, it’s fascinating. Automation permeates every aspect of modern day life, but for some reason in fixed income, it’s lagged. Now, this is probably not fair on the sell-side, because they’ve been engaging automated pricing algorithms for many years. But on the buy-side, it really could be developed further. It’s at the point of trade, which is probably the most important or critical part of the actual execution, where it has lagged. You could say that we have electronified the RFQ fx, but there is so much more that we can do. And for that reason, for the last 18 months, MTS has been working with a number of buy-side clients, to create a rules-based engine to develop autoX-protocols, to allow the buy-side to trade more efficiently, safer, and in a more cost effective manner.

Dan Barnes How would you describe what buy-side firms are trying to gain from automation?

Frank Cerveny Well, first of all, they want to be more efficient. And one of the things that we have to do this, we’ve gone in and we’ve just studied the workflow of our buy-side clients. And literally we counted the clicks, and we observed them in a kind of natural environment. And this helped us to develop improved protocols for things such as list trading, but with auto execution, we’ve kind of taking it a step further. Now today, the demands on the buy-side, I think, are more difficult than ever. You’re being asked to trade an environment where politics and central banks unnaturally move the market. The sell-side has less inventory, yet at the same time, they’re being asked to trade more with less resources. And during this entire period, they’re also being scrutinized to prove that their error free, and that they’re justifying best execution. And so, what we’re doing is we’re trying to help them in this process. And I find it very interesting, because the mirror embracing of auto execution helps, I think, reinforce the investment process.

Dan Barnes What are the components necessary to achieve these objectives?

Frank Cerveny First of all, you need smart and disciplined people who really understand your investment and execution process, because it’s important to note, at MTS, we’re trying to leverage, not replace the individuals by introducing auto execution. Second of all, you need data. Everyone’s heard the adage: garbage in, garbage out. Well, this is super important to trading, so it’s hard for me to imagine creating any auto execution parameters that involve data, that is not data that you can rely on. One of the unique selling points for MTS is that we can combine the real, executable, live data from our MTS cash order books to the buy-side, so they can use it as a parameter to set limits about when and where they want to execute. Finally, flexibility is very important. When trading, you know, one size or one rule does not fit all. So, as markets change we allow the trader to change the rules and the parameters in which he’s basing his auto execution around.

Dan Barnes As MiFID II has brought in best execution rules, do you think that’s encouraged auto execution?

Frank Cerveny Not necessarily, but it should. What auto execution allows you to do is to set the rules and parameters around the way and approach you trade, to eliminate any of these unconscious biases or errors from your trading.

Dan Barnes When we talk about rates markets, the US market is very different to the other G10 markets, it’s very different to EM markets. How do you see automation playing in those different environments?

Frank Cerveny It’s no surprise that the markets that are more standardized and more commoditized, were first to automate; the equity markets, the FX markets, the rates markets, the Treasury market, fx. But now we’re seeing data sources improve, we’re seeing automation improve, we’re seeing client and sell-side algorithms to prove the point, where they’re comfortable embracing more complex markets. And we saw that first in the investment grade markets, but now we’re seeing it move to the high yield and even emerging markets, what we see increasingly as clients that first start trading in the most liquid instruments and the smallest of instruments, but as their comfort level grows and the ease of it grows, you know, so does their appetite.

Dan Barnes That’s great. Frank, thank you.

Frank Cerveny Thank you. Dan, it’s been a pleasure.

Dan Barnes I’d like to thank David Scilly of First State Investments, and Frank Cerveny of MTS, for their insights into automating government bond markets. And of course, you for watching. To catch our monthly reports on other markets or to subscribe to our newsletter go to TraderTV.NET.