How the roll period affects market behaviour

Rolling a futures contract involves closing an existing contract and simultaneously opening a new one with a later expiration date. But when is the optimal time to roll? Typically, futures contracts have a fixed expiration date, and rolling occurs a few days or weeks before that date. However, the specific timing of the roll can vary depending on the market conditions and the preferences of the individual trader. Understanding the roll period is essential for traders who wish to maintain their exposure to an underlying asset or commodity beyond the expiration date of their current contract whilst minimising their trading costs (the so-called "rolling cost"). In this article, we will look at how Level 3 order book data can help participants understand when a roll occurs and how traders can use this to optimise their trading strategies and minimise market impact during a rolling period.

The method

To explore the impact of the roll, we use the BMLL Data Lab to analyse 45 pairs of WTI outright futures contracts across four years. We start with Level 1 metrics: liquidity at touch and spreads across the front-month (nearest future to expiry) and the near-month (next nearest contract to expiry). We then analyse the number of insert messages, a Level 3 metric, across the same period. By comparing the metric types, we can demonstrate how Level 3 metrics provide earlier warnings of when rolling will occur in the order book.

The analysis

By examining the distribution of time that liquidity at touch was greater for the front-month than for the near-month, we demonstrate that 2-3 days before expiry, there is a distinct change in market behaviour (Exhibit 1). As the front-month comes within a few days of expiration, there is a significant shift in liquidity at touch. A drop in liquidity at touch in the front-month contract makes it more expensive for traders to unwind their positions as the contract moves closer to expiration, adding to the rolling cost. Clearly, participants should look to begin their rolls before this change in available liquidity.

Exhibit 1: Distribution of time at which the front-month had greater liquidity than the near-month

Liquidity 9

Next, we examine the rolling period's impact on spreads by analysing the distribution of time in which the front-month had narrower spreads than the near-month. When examining Exhibit 2, we observe that spreads on the front-month widen significantly 2-3 days before the expiry date. To minimise rolling cost, market participants must be cognisant that as a future contract reaches expiration, spreads widen on the front-month and narrow across the near-month.

Exhibit 2: Distribution of time at which the front-month had narrower spreads than the near-month

Spreads 6

As seen above, a decline in liquidity and widening spreads before expiration can increase the rolling cost across the front-month. However, whilst Level 1 metrics such as liquidity and spreads are useful indicators of changing market conditions, they are not the most predictive. By examining a Level 3 metric, the average message insert count for the same sample set, we demonstrate that a change in the market dynamic occurs 6-7 days before the expiration of the front-month contract (Exhibit 3). The change in message insert count occurs earlier than in liquidity or spreads (Exhibit 1/2/3), exemplifying how Level 3 metrics are more responsive to evolving market dynamics.

Combining more responsive Level 3 metrics, such as insert message count, with Level 1 cost metrics, such as spreads, enables participants to better anticipate when there is likely to be a change in market dynamics ahead of the roll. Our analysis identifies a three-day delay between a change in the insert message count and widening spreads on the front-month contract. Therefore, if the increase in message count occurs between 7-8 days (1 day earlier than typically expected), participants should be aware that the increase in market activity can lead to a shift in liquidity and spreads earlier than expected. Participants with access to this data can better position themselves ahead of any change in the roll, minimising execution slippage during the roll and ensuring best execution.

Exhibit 3: Average message insert count across the front- and near-month

Insertcount 23

Exploring this analysis cross-asset

We carry over the methodology to two other futures complexes and examine the optimal rolling period, including Fixed Income and Equity Indices (Exhibit 4). The table below highlights the distinct optimal rolling periods of each futures complex. Based on the analysis conducted, we have found that the optimal day to start your roll is typically 1 day ahead of when the volumes typically move.

Exhibit 4: Optimal rolling period for crude oil, fixed income and equity indices futures



To minimise rolling costs, it is critical that participants truly understand how market dynamics evolve during the rolling period. To do this, traders must explore a broad time horizon and depth of data, which provides them with meaningful insights into the markets and securities they trade. The BMLL Data Lab enables participants to easily analyse the futures universe across ICE, CME and Eurex to understand how market conditions evolve over several years. By accessing BMLL Level 3 Futures Data, participants can go beyond the top of the order book and extract information unavailable to other practitioners enabling them to improve outcomes when trading the rolls.

If you have any queries about discovering the optimal rolling time for a futures complex, contact BMLL to find out more

Insights gained from BMLL Level 3 Data

Understanding market dynamics and the ability to navigate a vast and rapidly changing data set has never been more critical than now. BMLL is the leading, independent provider of harmonised, Level 3 T+1 data to the capital markets. We offer all market participants, including buy and sell-side institutions, flexible access to the most granular order book data and analytics to enable accelerated research, optimised trading strategies and the ability to generate alpha at unparalleled speed and scale. Level 3 Data from CME, Eurex and ICE is available to all market participants, covering Equity Indices, Fixed Income, Short-Term Interest Rates, Commodities, Digital Assets and FX.

For more information about BMLL Data & Analytics:

To receive more insight like this, register for BMLL Market Insight:

More information about the article

  1. The BMLL Data Lab is a highly scalable data science environment. For more information on the BMLL Data Lab:
  2. The WTI (West Texas Intermediate) futures contract is a widely traded financial instrument in the energy market. It is an oil futures contract that is based on the price of West Texas Intermediate crude oil. The contract is traded on the New York Mercantile Exchange (NYMEX). Market participants use it to hedge against price fluctuations or speculate on the direction of the oil market. Each WTI futures contract represents 1,000 barrels of crude oil and has a specified delivery month, usually at Cushing, Oklahoma.
  3. BMLL Level 3 Data provides full transparency of the order book, derived from every single insert, modify, execute or delete order message across every venue, available at daily and intra-day resolution.