TCA Comes of Age

Transaction Cost Analysis (TCA) is an essential requirement for buy side firms today, to satisfy regulatory requirements around best execution, to optimise costs, manage risks and to improve trading performance and execution quality.

So what are some of the challenges associated with TCA across different asset classes, and how can firms address these challenges, particularly from a pre-trade perspective?

Best Execution

TCA has become a fundamental tool to help firms meet their regulatory requirements around best execution. Both MiFID II in Europe, and the SEC’s order handling disclosure rules in the US, require transparency into how a firm’s routing decisions impact order execution quality for their clients. In electronically traded markets, market structures are highly complex. Countless factors feed into the decision process around where and how to execute. Firms therefore need strong TCA tools to help navigate through this complexity.

In the past, TCA mainly centred around post-trade analysis of transaction costs. Today however, firms are looking at an increasing number of real-time metrics and applying them pre-trade, using TCA to help select the right venue, broker/counterparty and algorithm for each trade, and to determine how to fine tune parameters on the fly. Firms therefore need to know how each venue, broker and algo is likely to perform under certain conditions, so they can route orders accordingly. And depending upon the asset class, different factors need to be taken into consideration.

Equities, FX & Fixed Income

According to a recent report from Greenwich Associates, almost 90% of equity trading desks now use TCA, compared with 60% in FX and 38% in fixed income.

In equities, leading firms use TCA to both ensure best execution and to preserve alpha, by assessing the potential market impact of their orders in real time. In order to conduct intelligent pre-trade analysis of liquidity pools (e.g. dark versus lit), trading sessions, brokers, algos and participation rates, firms need to be able to ingest and analyse accurate, timely, comprehensive and actionable data, both from internal PMS, OMS & EMS systems, and from external real-time and historical market data sources.

FX markets are much more decentralised and fragmented than equities, with a wider range of execution options available, including disclosed relationship-based pools, anonymous ECN-type pools, and Liquidity Provider (LP) algos. Where traders have the ability to view and analyse liquidity from each of those destinations in one place, in a true, real-time ‘top-of-book’, TCA can help with both sourcing liquidity and reducing slippage, and can be used to assess the likelihood of execution at quoted prices. Based upon their own firm’s flow, traders can determine how and where to execute for optimal performance, and decide whether to use an aggressive or a passive algo, or something in between (e.g. combining an aggressive component that goes to a disclosed LP with a passive component going to an anonymous ECN).

Historically, fixed income TCA has been the most challenging because of varying levels of liquidity in certain bond sectors. The lack of trading in specific names and bond issues means it is often difficult to reference an accurate benchmark to carry out a meaningful analysis of the execution quality. Even for bond issues where trading has been more active, the orders are often executed by voice rather than electronically and so access to the required execution datasets has been hard to come by.

In more recent years the percentage of bonds trading electronically has risen and so, in certain circumstances, getting access to this execution data has been easier but this is not the case across the board. A robust bond TCA analysis needs to aggregate pricing data from multiple sources, including indicative and firm quotes directly from dealers, from multiple platforms and executed prices from services such as TRACE.

Even this is often insufficient and so, to fill the data gaps, bond TCA often needs to include model/ theoretical prices and evaluated prices from independent firms. Here prices for illiquid bonds are calculated based on more liquid bonds with similar characteristics and these prices are used as the benchmark for the TCA. Also, as the bond market is RFQ-based, TCA must also consider prices quoted by dealers that were not hit/accepted into evaluation. Managing these large data sets and making actionable metrics available at the point of trade and post-trade is the goal for leading TCA providers.

Real-Time, Pre-Trade TCA

Pre-trade TCA can help with liquidity analysis, reduction of market impact, and improved execution quality, by using models that factor in multiple data points in real time. This is an area where AI and machine learning are increasingly being used.

By looking at live market volatility and liquidity characteristics throughout the day, along with historical data, AI-based models can analyse expected liquidity, participation, spreads, volatility and volume consumption, and provide recommendations around which broker and which algorithm would be most appropriate for a particular trade or strategy, together with appropriate participation rates.

Machine learning models built on convolutional neural networks, for example, can be trained to suit a firm’s investment objectives and trading styles. Users can see how the boundaries change over time and adapt the models by inserting new data or removing it. (A good example of this might be where a firm removes March and April 2020 data, anomalous because of the unprecedented volatility at that time).

Conclusion

TCA has evolved to the point where it is of value beyond just the trading desk. Integration of TCA across a firm’s entire workflow can provide actionable information not only to traders, but also to portfolio managers, risk controllers, back office and compliance staff, to better manage risk, optimise costs, and ensure regulatory compliance.

Leading TCA solution vendors are now working closely with their clients to help them get the most out of their data, so that they can use it both quantitatively and qualitatively to make the right decisions across their trading workflow.

As financial markets become increasingly complex and data-centric, as regulators push for more transparency, and as firms need to stay ahead in an ever more competitive market,
Advanced TCA not only helps buy side firms meet their regulatory requirements around best execution, but also gives them the ammunition they need to make better use of data across the trading lifecycle, from portfolio construction to trade execution.