Skip to main content

TCA is dead….long live TCA…

in viewpoint

TCA is dead….long live TCA…

"It may be hard for an egg to turn into a bird: it would be a jolly sight harder for it to learn to fly while remaining an egg. We are like eggs at present. And you cannot go on indefinitely being just an ordinary, decent egg. We must be hatched or go bad."C. S. Lewis

Transaction cost analysis (TCA) as we know it has already begun to go bad. Its origins in FX were really to allow asset managers to show their clients that their FX executions were in the right ballpark, but the new breed of buy-side FX traders (many of whom have moved from the sell-side) are looking for much deeper insights, such as:

“How can I identify and measure any information leakage during the trade execution window?

My LP says they are internalizing the vast majority of my flow, minimizing market impact. How can I verify that?

How quickly can I slice this large order into the market with minimal market impact?

What is the precise cost of rejects from each of my LPs?

Exactly how much is it costing this Fund to have trades directed mainly to its custodian?

None of the above can be answered accurately with conventional TCA applied to our OTC market. There are two fundamental problems: The first is the external nature of most TCA offerings. An external system or provider can really only operate by ingesting snapshots of the trading and market activity (discrete timestamps) and measure distances between them. The second issue is the age-old problem of the reference rates used. Current wisdom calls for a constructed external tape of the FX market, but this is not the same market that the asset manager is trading in. It may be close, but it is not the precise, dealable pricing that the user experiences. Even more challenging is that each underlying fund may have different LPs, fragmenting the real market data even further.

The current best answer to both of the above problems is to build your trading system inside the best “big data” database you can find and then capture every single action and data-tick in real time. TCA (or any other form of reporting) then simply becomes a dashboard view into the database in real-time. Yes, I am talking my own book, but the prescience we displayed when building EBS Institutional inside a kdb+ database was no accident. Data gurus know this is the way to go, evidenced by the fact that the F1 Aston Martin Red Bull racing team adopted the same kdb+ data technology only this year. There is almost no greater use case for accurate and rapid data analysis than Formula 1 today, where hundreds of sensors on the cars are streaming telemetry data back to the team for real-time analysis. Only when you know exactly when and why things occurred and what else was going on at precisely that time can you make informed decisions and manage the outcome effectively.

Today’s more stringent best execution requirements need exactly the same high performance embedded data capture and analysis tools as F1, not only to justify decisions made in the context of everything going on at that moment, but to drive superior performance while proving compliance with the rules. On anothernote entirely, regulators are also using this technology (the Australian Securities and Investments Commission began five years ago), and now expect to conduct investigations at an almost atomic level. Why take a knife to a gun-fight?

This advanced form of embedded TCA is hugely flexible, as reporting of all types becomes just another view into the data. Need a new report? How about knowing what your intraday credit utilization is in real time? Or alerting you when a Fund is close to exceeding concentration limits with any LP? What if your compliance team needs to view outlier trades in real time? You already have all the data in the system, so it’s a simple matter to create and implement a new dashboard view in a few days.

Conventional TCA is much less flexible and the process really works the other way round. When you have a need for a report, you then go looking for how to acquire the data. This has hard constraints, adds cost, takes time, and each report becomes its own project, rather than being another “view” as part of a homogeneous environment.

The other great advantage of this embedded analytics is the business model, which clearly impacts cost. The current breed of independent TCA firms needs to be paid for. Even when those firms are acquired, there seems to be an ongoing expectation that those firms are revenue engines. As OTC spreads continue to decline, this layering on of additional cost pressures is highly problematic. However, if, you use new technology to enable the production of deep insights directly from inside the trading platform there is no conflict of business models demanding to be paid.

But what of independence?” I hear you cry. There is really no need for the TCA system to be independent of the trading system. Yes, there is probably a need for your TCA to be independent of your liquidity providers, but, with one or two exceptions, your trading platform is normally independent enough. The key concern is again about reference data. The model I am proposing compiles data sourced from your own liquidity providers. This is precisely your available “market” and enables accurate and informed decision making. There is a valid concern that the endogenous nature of the data set may limit the quality or “independence” of the analysis. This is very easily solved with an additional set of externally sourced data for those analyses that require it. We have a natural advantage because of our proximity to primary market data from EBS and now the CME, but we also include other external data where necessary. This is the only true independence needed, and again avoids the need to pay for an additional business model layer.

This brings us to a new era, where data validates instinct and experience, satisfies regulatory compliance and client demands, improves performance, and facilitates actionable conversations between asset managers and their liquidity providers. This is game-changing. To enter this era you need to adopt new technology, which collects all the data natively. You just need to hatch that egg and enjoy the transformation!


This content was sponsored by CME Group.