When it comes to TCA for equity futures trading, how does this differ from other asset classes?
Equity futures TCA has many similarities with other asset classes but there are also some key differences. Within equity futures, there is a centralised exchange, deep order books and excellent liquidity in most contracts. Unlike equities, there is no fragmentation across multiple venues and price transparency is less of an issue compared to some other asset classes or instruments. The futures market also engages in rolls which is uniquely measured from a TCA perspective. Pre-trade, market impact and peer models are not as common and can be less detailed compared to equities and this is an area which might benefit from more investment by vendors and brokers alike.
In terms of similarities, the same benchmarks are used to measure performance versus other asset classes, such as IS, IVWAP, TWAP and open/close snaps, with the latter focusing on the equity cash times. Algo wheels within futures are now in full flow with similar structures to their equity counterparts, albeit with potential contract size nuances.
How has the use of TCA in the asset class evolved over the years?
Equity futures TCA has developed meaningfully over the last 20 years, as has been seen with other asset classes. Historically, the focus was on explicit costs, such as exchange fees and commissions, rather than implicit cost measurement. As with other TCA improvements in recent years, that has evolved to include analysis which can provide actionable conclusions.
Market data has improved, and internal order data has become more accurate with the recording of several different order lifecycle timestamps. This led to increased use of TCA benchmarking.
The rise in electronic trading supported more granular order data availability. The current age of increased automation in all areas of the industry has led to more efficient trade handling, more accurate market data and subsequent performance analytics. Market data vendors have been able to remove un-addressable blocks, as well as delayed prints and rolls, to ensure performance is calculated on liquidity which is addressable. Given the infancy of pre-trade and market impact models for equity futures, it’s still lagging equities in terms of available analytics, however that has improved significantly.
What should be front of mind for firms when it comes to building effective, workable equity futures TCA capabilities?
With any analytics framework construction, it’s vital that the market and order data being used is accurate, leading to actionable insights and conclusions. TCA results depend on the quality of order instruction tagging, which can vary significantly. For example, classifying roll orders will improve the efficiency of separating them out and suitably measuring this flow. The ability to measure performance versus each point in the order lifecycle allows information to be provided to portfolio managers and traders on any glaring issues such as delay costs and order profiling. As a result, being able to access this data via an O/EMS allows for a more detailed analysis.
Engaging with the trading desk on the most appropriate methodology for equity futures TCA measurement is essential as they will have specific knowledge of the market structure, which can enhance the analytical framework and resulting conclusions. Splitting results into various categories can highlight underperforming areas to be investigated. When flow is significant, the use of TCA vendors is advantageous. This includes algo wheel orders where contextual market data is provided such as spreads, volatilities, liquidity, and market momentum. These can then be used to normalise slippages versus market conditions, to treat brokers/algos fairly.
How can these tools be improved to best measure trading performance?
Order tagging could benefit from being more consistent in accuracy and population, rather than applying blanket instructions to orders. Even if portfolio managers interact with the trading desk in the office or via messaging services, the resulting TCA performances can only be driven by the quality of tagging available in the datasets.
Pre-trade and market impact models for equity futures are available but can lag their equity counterparts. Investment into these areas by TCA vendors and brokers could therefore distinguish them amongst their peers.
Algo wheel customisations are already available in the market, however they often take a while to complete. This is an area where improvements would again push vendors significantly above competitors.