How has the use of TCA evolved over the years?
There is an incredible, unquenchable thirst for data and trading performance numbers that is growing and growing. Some of this has been regulatory driven of course but there remains the quest – and rightly so – to utilise TCA for actionable insight and additional alpha and the users are getting smarter and smarter in how they approach this.
Investment firms have become a lot more joined up around TCA certainly in the last 10 years and I mean this in terms of internal interactions around the TCA numbers. The trading desk and the compliance teams form a feedback loop and dovetail a lot better than they used to around TCA. They operate together and joined up as opposed to separate business functions looking at and for entirely different data points/benchmarks within TCA. Again, the same with management and the TCA stakeholders, there seems to be more synchronisation around the analysis internally than there ever was.
The way firms trade now differs whether that be by algo usage, venue interactions, strategy attached to orders and the TCA has had to evolve with these changes, allowing the user to maximise the benefit from the solution and meeting business needs. Firms are a lot better at tagging orders, providing strategy and algo fields whereas eight years ago, it was sometimes a challenge to get Fix Tag 30 (Last Market) which provides the venue the fill was carried out on.
How does the use of TCA differ between various asset classes?
We have seen the expansion of the multi-asset class TCA driven by Mifid II. If you would have suggested fixed income TCA 15 years ago, people would have looked at you in disbelief. Now it’s a mainstream product being used globally by investment firms. The same applies for OTC derivatives. Again, firms are using solutions to meet their requirements in this asset class.
There is a large demand for multi-asset TCA, and a lot of this is driven by Mifid II and the changes in best execution requirements. FX and equity TCA are now mature products within the TCA suite and people are familiar with what can be done on these asset classes. It’s the newer asset classes that are seeing the most demand, for example bonds, CDS and OTC derivatives. Investment firms still need to provide a proof of best execution across hard to value assets using either an in-house or vendor solution. The approach to these asset classes is very different – it has to be. Each asset class has to be taken on its own merits, as they have unique market microstructures and nuances which have to be taken into account.
The limitations of the data in asset classes like CDS for example, leaves the user across the board with a fairly simplistic analysis. That’s not to say that won’t change going forwards with improvements to data. Someone running granular analysis on equity TCA in their approach will differ drastically with what can be done for example on an OTC derivative. We see much more of a compliance use case in these asset classes as opposed to the preferred actionable insight approach.
How can the use of TCA be improved/what are the key areas for growth linked to TCA?
Data quality is probably the largest most important piece within TCA, whether that be the client data or the benchmark data that the client data is measured against. Data quality has improved significantly over recent years. If you go back 15 years, orders would be placed over the phone, with no accurate timestamps, and very little transparency. Today, multi-asset orders are feeding through electronic platforms, in some cases with millisecond timestamp precision. This has led to an improved set of execution data for TCA purposes. The buy-side have also been good at pushing the sell-side to provide the necessary data points to enable them to run the analytics they want to see. Transparency has improved greatly, and this will continue to improve.
Analytics are becoming more sophisticated and the thirst to measure execution quality with new benchmarks and metrics is growing all the time. This is particularly prevalent in the newer asset classes people are starting to analyse. I think players in the TCA space who don’t acknowledge the part AI will play in TCA could potentially end up losing in the long run. I do think AI in the electronic trading space will create clear winners and losers. There will be less human interaction in the execution process as intelligence evolves. It’s inevitable.
AI does exist today for TCA; there are offerings that can look at an order from a pre-trade perspective and ascertain the optimal approach to execute, looking at historical data, patterns in momentum, liquidity, volatility, news stories, etc. The machine can make these decisions in seconds, whereas a human would have to spend a fair amount of time collating all this information. As times, data, technology and regulations change, so will TCA. AI will play a part in an intelligent, efficient execution process across all assets, and I see this as a growth area for TCA.