TCA is Dead! Long Live ESA!
A recent headline in a prominent trade publication proclaimed, “Demand for single-broker TCA dropping despite sell-side investment”. The headline summarized the views of a conference panelist, the EMEA head of microstructure at Citi, James Baugh. Mr. Baugh explained that, while polls of buyside traders have shown increased interest in TCA, his experience has been that client interest in sell-side-provided TCA had actually declined, despite increased sell-side investment. He went on to explain that this likely stems from the need for the buyside to get a multi-broker view of TCA.
I think Mr. Baugh is 100% right that the multi-broker view is key – namely that a single broker’s TCA simply cannot give a complete picture because a broker only sees a subset of trades. But I think an even bigger issue is that most users of conventional broker- or third-party-provided TCA have found it to be of limited value. The main issue is that conventional TCA can tell you what happened with past trades, perhaps broken out by order characteristics, portfolio manager, strategy, etc. But conventional TCA alone doesn’t provide rich insights into what to do in response to that information.
I suspect what the buyside community is really saying is that they desire enhanced analytics to improve their trading. Buyside firms realize that we are in an unprecedented age of data, computing power, and algo customization, and they want to harness the power of all three. In essence, what they are really wanting is Execution Strategy Analysis (ESA), not Trading Cost Analysis. The former is meant to evaluate strategies themselves to gain an understanding of how to actually change the trader’s or algorithm’s behavior to improve performance, typically via an iterative process of enhancement-evaluation-improvement. The difference between TCA and ESA is akin to addressing performance in school. TCA can be thought of like a report card – descriptive of the past, but providing limited information on how to improve any underperformance. ESA, on the other hand, is similar to a parent- or student-teacher conference, where more complete, contextual data can be used to develop a plan to improve performance as well as clearly defining how to evaluate performance and make any further enhancements.
Some brokers have developed excellent capabilities to provide ESA to their clients in the form of Execution Consulting teams. These teams undertake a client-specific deep dive into how the individual workings of the algorithm are impacting that client’s order flow. They then provide analysis along with recommendations on how to improve performance, typically via algo customization. At the other end of the spectrum, many brokers continue to rely on traditional TCA reports and methods, which provide limited data and often dubious advice. For example, I suspect most traders at one point or another have heard the following in a TCA review: “Your costs are high for this set of orders, which suggests that they have alpha. Our advice: trade more aggressively to get ahead and capture that alpha”. Of course, based on conventional TCA, it is incredibly difficult, if not impossible, to discern alpha from excess impact. So the advice to speed up very likely just makes the problem worse (which sadly may lead a further recommendation to become even more aggressive!).
But getting back to Mr. Baugh’s original point. ESA also needs to be done on a multi-broker basis for trading desks to truly optimize. Buyside desks need to do deep dives with each broker and then analyze performance across brokers. To accomplish this, buyside firms must either build out their internal trading research capability and/or hire (unconflicted) outside consultants to help.
So, when the buyside says they want to increase their focus on TCA, I believe what they are really saying is that they to increase their focus on ESA, not TCA, and that they want to do this across brokers. And, to Mr. Baugh’s point, this is something that a sell-side broker simply cannot do alone. (Shameless plug: The Bacidore Group can help).
 I could have referred to this as Trading Strategy Analysis, but wanted to avoid the TSA acronym.