SAME PROBLEM, NEW SOLUTIONS?

Dina Medland looks at why the buyside are not that happy with the current TCA offering.

A recent report on changing attitudes towards transaction cost analysis (TCA) both in the US and Europe by Aite Group, the independent research and advisory firm, finds that the buyside’s satisfaction level of TCA, in general, is low. There is a lot of inherent mistrust of data that is driven by lack of an accepted, industry-wide TCA framework.” (TCA Evolution, Through the Lens of Buyside Customers by Simmy Grewal and Sang Lee, Aite Group 2011).

Simmy Grewal, one of the authors of the report, says, “We found that buyside firms rely on a number of TCA products, sometimes as many as five, which highlights the fact that asset managers are not happy with their TCA.”

However, using multiple TCA providers at a time when traders are facing increasingly competitive and complex trading environments and thinner margins against a backdrop of rapid advances in technology does not necessarily point to high dissatisfaction levels. Michael Sparkes, director ITG, says, “People are using TCA more extensively than five years ago partly because of technology changes and partly because of market structure and fragmentation. The Aite report presents a very powerful graph charting the rise of electronic trading.”

‘While you could argue that the satisfaction data shows that users feel TCA has limitations, I interpret it to mean that people are now looking much more at what it is they actually want from TCA. Satisfaction levels reflect the fact that people want to do more with it and that is a challenge for the industry, not least because systems were not originally designed to address the questions now being asked” he adds.

While historically TCA has been very much a post-trade exercise that compares an execution against a benchmark such as the volume weighted average price (VWAP), there remain problems with that approach. “VWAP will be affected by your activity – the larger you are, the more you will be the VWAP. Where it is used we would recommend a participation-based approach rather than a time interval VWAP” says Sparkes.

“By far the most widely used measure in TCA is ‘implementation shortfall’, which looks at implicit costs around or slippage between when a trade enters the market, when it is executed and the final execution price. Our approach has been to develop models based on our client universe database to generate an ‘expected cost’ i.e. for an order with certain characteristics, traded in a certain set of market conditions. We believe this provides the fairest way of determining whether a client’s trading costs are in line with reasonable expectations or not, and highlights areas of particular strength or weakness” he adds .

TCA providers agree that the level of detail demanded varies enormously, depending on the client. Poor performance in difficult markets, especially in equities, has meant that there has been a significant increase in interest in TCA beyond the trading desk, as institutions look to mitigate the costs of trading.

Dave Hagen, director strategic accounts & trading at Linedata in the US says, “A lot of our larger clients are now doing their TCA on their own. I’ve just had a fascinating conversation with a gentleman who actually went out and hired three quants to build out his own proprietary TCA.”

There has been talk, he says, of ‘in flight TCA’, which essentially means in ‘real time.’ “What I am hearing now is ‘am I doing the best for this order at the moment’ – but I have been hearing talk about moving to real-time TCA for five years and have yet to see it” says Hagen.“There are no hard facts on this but I have a real sense that the European-domiciled clients I speak with seem ahead of the curve on this front.”

At Société Générale, Stephane Loiseau, head execution services appears to be at just that cutting edge. “TCA is not a post-trade event, it’s a lifecycle event, from when the order is first made to when it is settled it has to be a metric that is used all the way through” he says. Fragmentation in the market, scarce liquidity for macroeconomic reasons and regulation all pose challenges for traders, but the availability of tools now is also second to none, he adds.

“At SocGen we define TCA ourselves, and reduce participation in a venue as we go along. Trading strategy is a complex scenario with a lot of ‘ifs’ in play. By defining the objectives of the order, you can look at the tools and channels you are going to use, including choices of venues and brokers. Monitoring the parameters of the order as you go along, you can change the channels to adjust your trading strategy” says Loiseau. SocGen’s QVM approach is now about a year old and is used primarily on equities, “but we want to adopt this on other asset classes. We have already noticed trends that are helpful to investors, such as quantifying the impact of promotions on venues” he adds.

The absence of a consolidated tape in Europe remains “an issue compared to the US market.” “But there is a lot of information coming back to you and the question is how to make the most of it” he says.

Shifting landscape

Loiseau sees change as having taken place in the last 18 months, with innovation now taking place at a granular level. “For the most sophisticated buyside the focus is on mapping what services and tools are available and building their own mix” he says.

TCA providers say the shift in the TCA landscape has been “huge.” “People want to understand how they can look at their costs to improve performance. A big topic now is toxicity of the venues – how do you know one venue is better than another?” says Peter Weiler, EVP sales at Abel/Noser in the US.

“Buyside firms are not doing away with using vendors, but you do see a segment of the market which has a lot of resource do in-house work, especially if they are very quantitatively focused. The trend is towards using data to build algorithms instead of having an ‘algo sweep’ off the rack” he adds.

Clients are also increasingly interested in using visualisation tools to make TCA more “actionable” he says, and this is a part of a wider trend across the buyside, sell-side and quantitative firms.

Panopticon, the provider of real-time visual data analysis, recently conducted a live webinar with OneMarketData on how you can improve the efficiency of your TCA using a combination of visual data analysis, high performance tick history databases and complex event processing (CEP) engines. It included a “deep dive” demo showing how you can more efficiently analyse massive tick, quote and trade history data sets.

Visual representations such as heat maps and trend lines are “critical and the next step will be to bring those visualisation tools into the trading environment on a real-time basis” says Weiler. At IFS LiquidMetrix, the research and market data analytics company, managing director Sabine Toulson says, “ People are telling us that they want to go in two directions: they need specific information on the venues used to trade, but also data at a macro level in order to make judgements on the execution performance. ”

Not everyone agrees though on the compelling need for aggregate data, or with one of Aite’s conclusions that “As is the case with most aspects of financial performance, individual data is not as important as peer evaluation.” Sparkes points out that the report quotes 65% of respondents as saying that TCA could be built in-house, which would remove the importance of peer evaluation in the equation.

Nor does everyone get excited about visualisation tools and TCA. At SocGen Loiseau says, “Traders are not big fans of visualisation tools. They are good for the marketing brochure. We have had such tools here but would rather rely on the real-time indicators with no added latency.”

However, it remains clear that TCA is “moving downstream, to the trading level where we can shape trading. It is Behavioural Finance 101, it really is about building trading strategies from TCA data” says Weiler at Abel/Noser.