This article was written by Mike Googe, product manager and global head of transaction cost analysis at Bloomberg.
Futures volumes are at an all-time high, driven in part by the growth of passive investment strategies such as ETFs. Traders use futures to offset cash risk while trading individual lines of a basket to manage that risk. Futures are attractive for this purpose due to their high liquidity and cost efficiency. This growth has led to greater scrutiny of execution costs as investors look for evidence that everything is being done to protect returns.
Get access to the latest news and insights from Bloomberg’s electronic trading specialists.
Trading futures is more immediate than trading stocks or bonds, and analysis has typically focused on execution strategy. Calculating costs based on an Execution Interval Arrival Price, or VWAP, was a fundamental requirement, but increasingly sophisticated analytical requirements are emerging. So how does transaction cost analysis (TCA) in the futures space adapt to meet this demand?
Standards-setting sophistication
Listed markets have the luxury of complete price and volume data sets for benchmark calculation. When attempting to compare and evaluate execution strategies, an emerging theme is to focus on the performance of each individual fill. This analysis compares each fill to benchmarks that answer two important questions for traders:
- Have we detected any spread? (i.e. did we execute a lower price than the offer when purchasing or vice versa)
- Has the level of aggression led to unfavorable price movements, resulting in increased market impact?
These metrics aggregated to the order can provide important indicators of execution quality and can be helpful in selecting future trading strategies.
Volatility is crucial in derivatives pricing and affects the bid-ask spread. While the typical arrival price method considers the midpoint of the bid/ask spread at the given timestamp, there is an increased need to consider a distant benchmark (which takes the supply side into account when buying and vice versa when selling). Similar in principle to spread recording, this proves to be the benchmark of choice when looking at the order or route level in order to capture the impact of the prevailing bid/ask spread and not hinder the analysis by half the spread.
A benchmark that is particularly popular with passive strategies is the Market On Cash Close (MOCC). This measure is not an official order type and is very popular because it determines the price of the future at the time the underlying securities cease to be traded. This allows traders to allocate volume and price when the underlying markets are operating and when they are not.
Relative comparison
These represent absolute benchmarks. While traders instinctively assess good and bad results, pressure from external stakeholders to prove this has led to the need for a relative benchmark that provides a frame of reference for comparison.
The first thing that pops up is peer benchmarks. These measures have been successful in equity analysis for years and are becoming significantly more important in the fixed income and foreign exchange markets, but are new to futures. Peer benchmarks are community-sourced and based on aggregated community results, but should not be viewed as a comparison of participants. This is a comparison of flow profiles based on order and market condition characteristics. The approach is to compare observations at a meaningful level of aggregation to add value; Not so abstract that they are of little value, but also not so detailed that they lead to data scarcity at best and potential data leaks at worst. Reasonable approaches to minimum contributors and observations can usually alleviate these problems. Another challenge for a futures peer calculation is its temporary nature. Each contract only has a limited term, usually three months. So how can one make a meaningful comparison with such a short time frame? The answer lies again in the aggregations used. Using a single contract would be precise but would have limited data and only temporary value. So focusing on the contract and not the specific series offers the advantage of lengthy data.
Roles and role periods expand the challenge. Their contingency creates an imbalance between supply and demand as the roll date approaches, which can lead to skewed results. Again, using appropriate aggregations helps isolate roles from individual branches, and aggregating peer counts into a quarter would usually smooth this effect.
Gain insight
The next step is to organize the data to achieve the most value. Typically, participants strive for two outcomes. Efficient outlier detection and decision support insights. Increasing volumes require automated outlier detection to identify trades that do not meet execution quality thresholds. These thresholds vary depending on various factors, order difficulty, market dynamics, etc. However, the most frequently asked question is: “Where are thresholds set?” Experience and judgment guide this, but an independent frame of reference for decision-making is crucial, and here too, peers can -Data can be helpful. For selected combinations of groupings/benchmarks, proposed thresholds based on positive and negative percentiles of a distribution curve formed by these combinations can identify any order that falls in the top or bottom “n” percent of a similar flow in the peer community.
This flow profiling approach also provides insights for decision support. Collecting and grouping results by order-related factors (e.g. size/ADV, order type, etc.) and market conditions (e.g. volatility or momentum, etc.) allows for a fair comparison of performance and illustrates conditions and characteristics that influence trading performance influence. By increasing automation using tools like RBLD, incorporating these insights and using them with peer benchmarks, traders can reduce complexity and have more time to focus on more impactful orders. Finally, there is growing interest in considering traders’ order creation times and routing times to gain insight into the impact of timing and delays and to consider all traded elements (including FX exposure trades) for a holistic analysis of trading costs.
Diploma
The growth in futures trading is driving the development of matching analysis. Competition is driving the introduction of new measurements and approaches. Far from simply extending existing TCA approaches, futures markets must overcome clear challenges in capturing the nuances of this asset class. However, the underlying data offers participants a real opportunity to develop high-quality analysis that can only lead to better results.