There is a clear trend towards 24/7 Equity Trading. But how much liquidity is available for participants during overnight trading?
A deep dive into the recent tech sector drop, the impact of the DeepSeek news, and how this affects the US Open.
Authors: BMLL’s Rob Laible, Head of Americas and Shaurya Thakur, Quantitative Analyst Americas.
First published in TabbFORUM, 19 February, 2025
The demand from Retail (especially in Asia), participation by Market Makers, traders looking to mitigate risk, along with extended trading hours combined with the recent tech sector drop driven by the DeepSeek news, provides the backdrop to this case study in overnight trading volumes and buy/sell behaviours.
We observe a persistent order book imbalance between the bid and ask sides and buy and sell volumes, contributing to sustained selling pressure, and driving a steady decline in execution prices. We also see relatively tight spreads and liquidity levels when compared to the primary market for some stocks, emphasizing the role of overnight trading in shaping price discovery for the next day's Market on Open price.
BMLL’s Rob Laible, Head of Americas and Shaurya Thakur, Quantitative Analyst Americas, examine how much liquidity is available for participants during overnight trading.
Understanding price movements, trading activity and liquidity during extended hours is key as we continue to move towards extended trading, to stay ahead of market movements, make informed decisions, and optimize execution strategies.
The recent selloff in the technology and semiconductor sectors, driven by the release of Deepseek's new generative AI models, underscores the significant impact that weekend news can have on equity markets. While most U.S. equity trading occurs during regular market hours, there is a growing trend toward extended trading sessions, including overnight trading. Platforms like Blue Ocean ATS now allow overnight trading, whilst exchanges like CBOE (EDGX), and the New York Stock Exchange (ARCA) have both announced plans to enable their exchanges to trade extended hours [1,2]. Additionally, 24X National Exchange has received SEC approval to operate as the first national securities exchange in the US allowing 23-hour trading each workday, subject to Equity Data Plans implementing necessary changes and a final SEC rule filing [3]. MOON ATS is also looking to enter the extended hours trading space [4].
These developments reflect a broader movement toward a 24-hour trading cycle, enabling market participants to react promptly to news and adjust their strategies ahead of the regular market session. Understanding price movements, trading activity, and liquidity during these extended hours is crucial for traders aiming to stay ahead in an increasingly dynamic market environment.
Comparing overnight trading dynamics with regular market hours
Our analysis focuses on the 24-hour trading activity of several tech/semiconductor sector tickers. We examined Blue Ocean's overnight trading session, specifically from Sunday, 01/26, at 8 PM to Monday, 01/27, at 4 AM. To benchmark this activity against the broader market during the corresponding day, we utilized BMLL’s CBBO dataset (CBBO). The CBBO dataset consolidates the best bid and offer across all U.S. exchanges and includes odd lot sizes, offering a more detailed and accurate representation of the market compared to the NBBO.
To conduct this research, we leveraged the BMLL Data Lab, a scalable Python sandbox environment that enables advanced data exploration, in-depth market microstructure analysis, and customizable research workflows. Additionally, we incorporated BMLL Level 3 (market-by-order) datasets (Level 3), allowing us to study liquidity and market dynamics at an unprecedented level of granularity. By analyzing the evolution of liquidity and order flow across different venues, we gained deeper insights into how overnight trading conditions compare to regular market hours.
Price Movement and Buy/Sell Side Aggressor
We start by looking at SOXS, the ETF for the Direxion Daily Semiconductor Bear 3x Shares ETF. This ETF has a strong inverse relationship to the semiconductor sector, reflecting heightened interest in bearish semiconductor positions. Using BMLL’s granular data during the overnight, we see significantly higher volumes in aggressive buys in the plot for Buy vs Sell Trading Volume and Price Movement (Figure 1), as represented by the green bars, as well as a steady uptick in price (blue line), highlighting an increased buying pressure during the observed trading session. Notably, traded volumes are consistent during the night session, with some increase during early hours.
Price Movement and Order Book Imbalance
We now focus on TQQQ, the ProShares UltraPro QQQ, a leveraged ETF aiming to deliver 3X the daily returns of the Nasdaq-100 Index. Again, utilizing BMLL’s data throughout the overnight trading session, we observe a persistent order book imbalance between the bid and ask sides (Figure 2). The higher red bars representing ask sizes indicate an excess of sell-side liquidity, while the comparatively lower green bars on the bid side reflect weaker buy-side support. This imbalance contributes to sustained selling pressure, driving a steady decline in execution prices and reinforcing the downward trend throughout the session.
Comparison with regular trading hours
Looking at the overnight session, we see some clear trends - a strong imbalance between buy and sell volume, and a price that reflected the weekend sentiment (and carried into the main market session). A natural question from a liquidity perspective is how much liquidity is available for participants during overnight trading?
Examining the spreads
The chart (Figure 3) illustrates the BMLL metric “time-weighted average spread” for TSLA over a 24-hour trading cycle, comparing various trading periods, including overnight trading, pre-market, market opening, and regular trading hours.
Overnight trading spreads for TSLA are observed to be comparable to the spread for CBBO dataset, which represents the broader market for TSLA. Additionally, during some trading periods, the overnight spreads are slightly tighter than the general market spread, indicating efficient trading conditions during these times. This comparison highlights the relative consistency of spreads across different trading periods during overnight hours.
Using Level 3 (Market-by-order) Data to understand Depth of Liquidity
Our analysis leverages BMLL's Level 3 (market-by-order) (BMLL) datasets to evaluate the depth of liquidity across different price levels over various time periods and trading venues, including Blue Ocean, ARCA, NYSE, and Nasdaq. The plot for Level 1 notional liquidity (Figure 4*) for TSLA highlights a significant imbalance on the ask side during the overnight opening period, primarily driven by selling pressure linked to the Deepseek news.
Beyond this initial imbalance, the overnight session demonstrates strong and even time-weighted notional values on both the bid and ask sides, with liquidity levels closely mirroring those observed not only in the broader market during regular trading hours but also on individual venues such as ARCA, NYSE, and Nasdaq. By leveraging Level 3 (market-by-order) data for TSLA, we directly compare liquidity dynamics across these venues, revealing that the overnight session maintains a competitive liquidity profile.
At deeper levels, the overnight session continues to show resilience. The plot for TSLA Level 10 notional liquidity (Figure 5*) underscores the substantial depth of liquidity, demonstrating that overnight trading effectively competes with market conditions at least 10 price levels deep. Notably, the overnight session outperforms some individual venues during regular trading hours, further reinforcing its growing role as a viable liquidity source. The ability to analyze order book depth across multiple exchanges using Level 3 data provides a more granular understanding of how liquidity develops outside of traditional market hours, allowing for more informed trading and execution strategies.
Price Changes Across Market and Overnight Sessions (Execution Prices)
The chart (Figure 6) highlights distinct price movements based on execution prices for the analyzed tickers during market trading sessions and overnight trading periods. It emphasizes the role of overnight trading in shaping price discovery, with key price movements often occurring before the regular market session begins. This analysis leverages BMLL's derived analytics (Analytics), which provides access to over 400 metrics to analyze price movements, liquidity dynamics, and order book characteristics.
- For most tickers, execution prices declined from the market's Friday close to the overnight Sunday opening, continuing into the overnight session due to the Deepseek selloff. A small reversal occurs from the overnight Sunday close to the Monday market open, but the downward trend resumes during regular trading hours.
- For tickers that closed higher than they opened during the market session, such as MSFT, INTC, and MSTX, we observe a mid-session decline followed by a reversal toward the end of the trading day, leading to a higher closing execution price. In contrast, during overnight trading, the decline persisted continuously without a similar recovery.
- For AMD and TQQQ, the decline in execution prices from the opening to the closing during overnight trading was more pronounced than the price movement during the market session, indicating stronger downward pressure in the overnight period. In contrast, for AVGO, NVDA, and TSM, the decline was significantly larger during the regular market session.
Final Insights from Extended Trading and Order level data
The Deepseek-driven selloff in the technology and semiconductor sectors underscores the importance of understanding overnight trading and its influence on the following day’s market session. As the demand to trade 24x7 continues to rise, and market structure evolves, extended trading hours are becoming more prominent. Overnight trading sessions provide an early indication of price movements and sentiment shifts. This is particularly relevant in a global trading environment where geopolitical events happen unexpectedly and do not conform to traditional market hours.
The combination of BMLL Level 3 data and the BMLL Data Lab environment with Blue Ocean’s market data provides critical insights into overnight trading, showing how liquidity spreads and volume evolve and change overnight. We can see that trade imbalances provide clear directional indicators, whilst spreads and liquidity remains competitive. As overnight trading grows, understanding the impact will be increasingly important for traders seeking to stay ahead of market movements, make informed decisions, and optimize execution strategies.
References
* The time intervals defining the trading buckets for the overnight session, market session, and individual exchanges (ARCA, NYSE, Nasdaq) are consistent across Figures 3, 4, and 5. The time intervals for the market session and individual exchanges are identical.