Looking into the Dark: Are you missing liquidity that’s addressable to you?

Optimising intraday trading - uncovering Dark and Addressable liquidity trends

By Roger Ahanonu, Senior Quantitative Analyst, BMLL.

First published by TabbFORUM, 13 August, 2024

As the fragmentation of market liquidity in Europe continues, the ability to examine historical market behaviour is increasingly important for trading practitioners. Understanding the constantly evolving liquidity landscape is crucial, particularly for optimising algorithmic trading, setting benchmarks for best execution and maximising execution alpha. Accurate interpretation of intraday trading behaviour is essential, as it helps market participants make informed decisions on where to route orders based on historical liquidity patterns. This analysis ensures that traders can navigate the complex market dynamics effectively and achieve optimal execution results. In this article we explore how intraday trading can be optimised through comprehensive trend analysis of both Dark and Addressable liquidity.

Intraday liquidity based on trade type - are they all Addressable?

By leveraging high quality historical data, traders can gain a realistic understanding of the types of liquidity likely to be available during specific timeframes, providing valuable insights for developing an optimal trading strategy. Trade-by-trade datasets, normalised across all venues and enriched with flags such as MMT (Market Model Typology), provide deeper analytical insights. *see note (i).

Figure 1 highlights the intraday volume profile, normalised by Trade Type across all liquidity types for instruments primarily listed on Euronext Paris. This analysis encompasses both on-exchange and off-exchange trades, on- book and off-book transactions as well as Addressable and Non-Addressable liquidity. The data covers trading activity from 08:30 am to 19:30 pm local time.

Particularly notable is the volume during the Closing Auction uncross, which dominates the intraday profile. Removing this segment of liquidity allows us to see a clearer view of the liquidity patterns throughout the day.

So without the uncrossing trades, how does the intraday liquidity pattern change? Figure 2 displays the volume profile excluding liquidity from uncrossing periods, such as the Opening Auction, Closing Auction, and Periodic Auction. With this type of liquidity excluded, we can clearly see the significant size of the Non-Addressable Special Price component within the volume profile.

Special Price component, a BMLL classification for non-addressable trades which includes negotiated trades, Non Price Contribution to Discovery Indicators trades (marked TNCP under MIFID II), dividends, and other technical trades make up the largest percentage of volume throughout the trading day, approximately one third. This clearly demonstrates that the inclusion of this type of Non-Addressable liquidity can significantly impact the accuracy of a volume profile.

MMT Flag - shedding light on Dark trading

Next I focus on Dark intraday liquidity. When analysing the Dark volume profile for stocks listed on Euronext Paris across all trading venues, a typical smile shape emerges, as Figure 3 illustrates. This is characterised by higher trading volumes at the start and end of the trading day, specifically around the Opening and Closing Auction periods, with a dip in volumes during the middle of the day. Additionally there is a notable pickup in volume at 14:30 pm local time, coinciding with the opening of the US market.

When analysing the intraday volume profile for Dark trading in XPAR Primary instruments, broken down by venue operators, and the specific trading locations, a similar pattern emerges, as Figure 4 indicates.

Although this demonstrates the general pattern of Dark trading, MMT flags can be utilised to provide deeper insights into the data.

Figure 5 illustrates the intraday volume profile for Dark trading in Europe, broken down by MMT flag level 3.5. Here, a value of ‘S’ corresponds to a flag of RFPT, indicating a Reference Price Trade. These mid price reference trades are accessible to market participants, and account for 54% of the total dark liquidity during the day on average.

Are these high-level classifications sufficient for market participants to optimise Dark trading strategies? We examined additional MMT flags to identify whether there are other meaningful classifications that could aid in developing more informed trading strategies, especially for algorithmic trading. Figures 6(a) and 6(b) below illustrate the volume profile for dark trading, categorised by MMT Flag level 1. It highlights the differences in the market mechanisms where dark trades occur. As expected, the majority of the volume (85%) is associated with MMT Flag Market Mechanism 3, which indicates trading on Dark Order Book. In contrast, a smaller portion of the volume, 15%, falls under MMT Flag Market Mechanism 4, representing Off-Book trading, including Voice or Messaging Trading. This category encompasses trades executed on venues such as LISX.

The remaining component, which accounts for <1% of the observed Dark volume, consists of 'Dark' trades marked by Market Mechanism 1. These trades usually involve hidden liquidity, such as iceberg orders, executed on a Central Limit Order Book and are categorised under MMT flag Transaction Category D. Notably, there are no trades associated with Market Mechanism 4, which involves Quote-Driven trading.

In summary, dark trading is predominantly conducted through anonymous order books, with a smaller portion handled through off-book methods. Understanding this distribution using MMT flags helps market participants optimise their Dark trading strategies by choosing the most appropriate venue and mechanism based on their trading objectives and needs

Optimise intraday volume profiling for individual instruments by ensuring accurate trade classification

Market participants can enhance intraday volume profiling for their individual orders through precise trade classification. We used LVMH (EPA: MC) as an example to illustrate the importance of accurate trade classifications.

Figure 7 displays the intraday volume profile of Addressable Liquidity for LVMH traded across European venues (XPAR: MC), while Figure 8 illustrates intraday profile of Non-Addressable liquidity for the same stock. Notable differences in the shapes of the volume profiles are particularly evident at the beginning and end of the trading day.

As expected, the highest proportion of the day's volume for both Addressable and Non-Addressable liquidity is concentrated around the market close, with approximately 17% of the total daily volume traded within this 5-minute window. Of this, 14% is Addressable liquidity, and 3% is Non-Addressable. With detailed historical data, users can further analyse and segment this information to gain a deeper understanding of both types of liquidity.

Next we examine the Dark intraday liquidity pattern for LVMH. Figure 7 illustrates the Dark liquidity for LVMH traded across all European venues throughout the day. It reveals a similar ‘smile’ pattern to the broader Dark liquidity volume profile shown in Figure 3, although the volume change between each 5 minute time frame is less smooth. As with the broader pattern, there is a noticeable increase in Dark liquidity during the opening and closing phases of the market. Users can further enhance this analysis by examining Market Mechanisms categories by leveraging MMT flags.

Efficiently identifying and isolating key liquidity sources

For trading practitioners, it is essential to access both normalised Market States, and Trade Type fields, including categories such as Lit, Dark, OTC. This capability allows users to easily identify addressable versus non-addressable liquidity and isolate relevant liquidity types in their analysis. This in turn allows practitioners to more accurately plan and execute strategies targeting specific types of liquidity (i.e. the Dark), where understanding the nuances of intraday volume trends is imperative for achieving targeted benchmark performance for their clients.

The example datasets provided in this article are available in the BMLL Data Lab which offers rapid access to Level 2 and Trades Market Data within a Python Jupyter notebook environment. This setup facilitates efficient analysis and allows users to quickly generate volume profiles for individual instruments, indices, or across the entire market.

With these tools, market participants can effectively pinpoint where dark liquidity and addressable volume are concentrated, enabling precise trading decisions.

_____________

Note (i): The industry-wide adoption of common data standards is key to overcome the challenge of producing consistent data across a highly fragmented market structure. The FIX MMT is an initiative that helped create a uniform trade flagging standard supporting conformance with MiFID/MiFIR post-trade transparency requirements. FIX MMT trade flags have been instrumental in the BMLL normalisation of trade data. They support the unambiguous categorisation of transactions, essential for all market participants running advanced analytics, with use cases for buy-side firms, sell-side brokers and market operators.