Getting the most value from multi-asset TCA

Published on 17 September 2019

Transaction cost analysis (TCA) can have very positive effects on everything from trading to the investment process, if it is implemented well. However, while equity TCA is relatively mature, there are serious challenges in other asset classes, which require traders to track down robust data sources and conduct rigorous appraisals of the metrics to be used in order to extract the greatest value from those services.

Kevin O’Connor, head of analytics and workflow solutions at Virtu Financial, outlines the benefits and best practices that buy-side firms can achieve when their traders use TCA effectively, and offers his insights into the key hurdles that need to be overcome.

Dan Barnes Welcome to Trader TV – your insight into the trading climate for professional investors. I’m Dan Barnes. Consistency in trading performance can only be assessed if traders have reliable benchmarks and data. Kevin O’Connor, head of analytics and workflow solutions at Vertue, gives us his advice on analyzing best execution across asset classes. Kevin, welcome to Trader TV.

Kevin O’Connor Thank you for having me.

Dan Barnes So how are end investors and portfolio managers’ goals impacted by execution quality?

Kevin O’Connor Execution quality affects returns, and that’s the most important factor. The whole concept of measuring execution quality really came out of the portfolio theory of implementation shortfall; being defined as the difference between the paper or theoretical return of a portfolio and the actual return of the portfolio. The difference being the cost of execution, either explicit, such as commissions, taxes and fees, and implicit, which are the transactions cost that we’ve now associated with TCA. And transactions costs can also affect the portfolio construction process, especially in quantitative portfolio management. Asset managers will often incorporate expected transactions cost in the optimization process, so you have the expected return and the risk profile and some other factors. And then transactions cost, as an input and an optimizer, will actually select some securities, because they’re easier to transact, because they have similar risk and return characteristics. Why not take the easier one to trade? And so it actually can directly affect what portfolios ultimately get constructed.

Dan Barnes That’s really interesting, and that also sounds like something that would be of interest to non-quand funds too.

Kevin O’Connor Yes, the concept has been used in non-quand funds and as it relates to liquidity especially.

Dan Barnes So how do asset managers effectively assess execution quality and what difference does the asset class make to that?

Kevin O’Connor  Asset managers should really be looking at as much of the execution process and its effect on on parent orders or the portfolio as possible. So it means looking at traders, looking at brokers and venues, but also figuring out how that may affect overall performance at the parent order level. Now, in equities, that’s pretty easy to do or easier to do, because there’s available data out there and consolidated tapes and being able to take data from exchanges and resellers and compare it to all the different points of the investment process. In FX, the quality of the transaction data from customers might be good, but then the available market data, you have to collect and aggregate that yourself or use a provider to do that. In fixed income, there’s even less available market data out there. So sometimes it’s less of a precise measurement process, and more of an evaluation of, ‘did I choose the right style of execution?’

Dan Barnes And do you find that buy-side firms are quite open to the idea of measuring trading as part of the investment process, or is it quite segregated?

Kevin O’Connor I think that’s come a long way. That the interaction between portfolio managers and traders is as strong as it’s been, the indication of what the urgency is on the order or what the expected execution style should be is fed to the traders. Traders are working on metrics to give back to the portfolio managers to show what they saw in the markets upon receipt of the order, what the typical patterns are, and confirming with the portfolio managers what their default execution strategies will be, given what they’ve learned in the past.

Dan Barnes Where there is more data needed in markets like fixed income, how challenging is it to actually pull that data in?

Kevin O’Connor It’s quite challenging and I think that efforts are underway right now with execution providers and venues and fixed income to generate data sets, perhaps generate some value of pricing information that can be resold and put to the market. There’s efforts in place to try to organize all of the broker communications, and forms of broker runs, and information that are passed around the industry. It will take some time, I think, before that becomes a viable source for the average asset manager to give access to. But, you know, there is progress there.

Dan Barnes How does that sit in terms of the overall hurdles that asset managers need to overcome?

Kevin O’Connor Those hurdles remain the same, and they’ll probably be the same five years from now, whether it’s about data, and first of all, it’s being able to take data out of their order management or execution management systems. It’s being able to get access to market data that really represents the market that they saw, or they could have seen if they had access to other venues, and then being able to do something with that data transparency, to clean up some of the issues they might find in the data and actually attach it to the whole investment process.

Dan Barnes What are the strategies that asset managers can employ to overcome some of those hurdles?

Kevin O’Connor I think they first have to work a bit with their order management or execution management systems, to make sure that what goes to those systems are focused on getting clean data. Data that captures what the intent was and what the ultimate result was, in a normalized way that can be used. In the case of measuring algorithms there’s been a big push towards algo wheels, which are exactly that, a chance to normalize data so that each of the providers or brokers on the wheel can really be measured the same way, where the intent was clearly aligned and they had discretion over the orders. Lastly, if they’re using a provider, getting some transparency into what the methodology is for aggregating the market data that they’re using, and perhaps even getting the ability to download the results into their own facility, so that they can do some analysis on their own, not to necessarily check on the provider, but that could be one reason, but also to combine it with data they may not be giving to the provider.

Dan Barnes Kevin, thank you. That’s been fantastic.

Kevin O’Connor Oh, it’s been great. Thank you for having me.

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