Richard Hills
Episode 14: Spread Cost Precision Matters
Spread Curves: It's Complicated
Average bid-offer spreads are the industry's preferred liquidity benchmark. Investors use them with volume and volatility data to formulate asset and security selection decisions. Traders use them for pre-trade cost estimates and scheduling, and post-trade analysis that sits at the heart of best execution. In the ETF world they underpin an Issuer's obligations to monitor the market quality of their products. But the measures are a complicated beast, with multiple dimensions and nuances. The average can deceive.
With European ETF assets at record levels and post-MiFID II scrutiny on best execution intensifying, precision in spread measurement matters more than ever. On April 7, 2026 a single spike in spreads over a five-minute window at 11.10 am UK time pushed the most-traded group of European-listed US Equity trackers from its 20-day average of 2.7 bps to nearly 12 bps, over four times its normal level. That 300-second window was large enough to leave a footprint on the 20-day average for the entire month. This episode continues the spreads work begun in Liquidity Matters Episode 3 and Episode 8, with finer resolution and a closer look at what the daily average conceals in a specific ETP context.
For the analysis we built an interactive chart of 12 of the most traded European-listed ETFs and one ETP, grouped by underlying benchmark into US Equity, Global Developed, European Equity, Physical Commodity and Commodity-Sector Equity. Spreads are measured as consolidated EBBO, time-weighted within each instrument and volume-weighted within each group, in five-minute bins for 20 trading days in April 2026. The upper panel shows the 20-day average; the lower panel shows individual daily curves. Use the date selector and the legend filters to explore the data yourself and hover over the chart to view the data by bin. You can easily compare a normal trading day against April 7 in the same asset class, or switch asset classes to see how the regime structure changes across the market.
Producing this picture is complicated; nanosecond-aligned data from every venue is required to ensure the correct sequence of bid and offer changes across up to five lit books, for thousands of traded ETP and equity instruments every day. Let's look at the key features.
The intraday curve has structure
The spread curve is not random. It resolves into four distinct but predictable regimes for equity ETFs, with a fifth regime exclusive to commodity-linked products around the London gold fix.
At the European Open (08:00–08:20 UK) spreads widen sharply as ETFs and their underlyings begin trading simultaneously: European Equity averages 9 bps in the first five minutes (against its daily average of 6.2 bps) and Global Developed 13.1 bps (against 5.3 bps) with the latter taking 40 minutes longer to find the morning level. The continuous trading floor that follows is the calmest part of the daily session, with each group trading below its own daily average: US Equity at 2.3 bps (against its daily average of 2.7 bps). The Global Developed group also shows lower spreads in the morning at 4.5 bps (against 5.3 bps) reflecting the substantial US weighting. European Equity hugs its daily average for most of the day at 5.8 bps in the morning (against 6.2 bps), more so than any other class.
At 13:25 (UK time) spreads widen for every group in anticipation of the US macro announcements release window; even the European Equity group spreads move out 10.6 bps, reflecting the read-across in global sectors. A short pause back at average spread levels evolves into the US Open at 14:30 - the most volatile part of the day for around 90 minutes, where US-underlying product spreads widen sharply to take into account new information.
Our chosen Gold physical ETP and Commodity-Sector Equity ETFs are among the most traded ETPs at the moment and therefore make the cut for inclusion in this episode but they only represent the gold and silver markets. The ETFs reach 5.1x their morning level reflecting the US weighting of NYSE and TMX listed miners and the ETP’s response to the increase in COMEX futures liquidity.
Within each regime, spreads are mean-reverting, usually pulling back within minutes of any out-of-regime spike. Market makers price expected information flow into their quotes to reflect risk premiums and uncertainty, and that flow is not constant through the day.
Why variation matters as much as the level
The regimes describe a shape. The level on any given day varies sharply with the effect of shocks and outliers. For example on April 14, the European Equity ETF group reached 45.5 bps in a single five-minute bin at 13.30 - over four times the 20-day average of 10.6 bps for that window. Strip that day out of the calculation and the 20-day average at 13:30 falls from 10.6 bps to 8.8 bps.
In the next episode of Liquidity Matters we will investigate how spread volatility and level evolves on ‘messy days’ compared to the average, and how random events often have an overweighted impact on average spreads. Precision is required to adjust and normalise these events.
Asset classes add a further dimension
The regime structure also differs substantially between asset classes, and not only in level but in shape with a wide range even within the same benchmarks and the country venues they trade on. Commodity-Sector Equity ETFs mostly track the US equity session and have a fifth discernible regime around the London gold fix at midday when spreads step up from 8.4bps to 17bps and remain there for the rest of the day. We will also investigate this in the next episode.
Conclusion
A daily aggregate is not precise enough to measure the true cost of liquidity, nor to identify trends in that cost over time. Variation is heavily skewed, since spreads blow out but never blow in, and a few outlier moments can pull a monthly aggregated average well above that typically experienced on a normal trading day. The scale of data involved in creating precision spreads runs to over two billion data points a day at nanosecond timestamp granularity.
For portfolio managers who use ETPs for asset exposure at country and sector level, for diversification and inflow management, selection of the right product from a stable of similar products relies on metrics such as spreads to assess liquidity characteristics. Issuers know this and rely on quality data to compare their products with competitors for both passive and rapidly growing active styles. Getting access to high quality, precision data remains a major operational challenge for investors, trading firms and ETP issuers alike.
Note on Methodology: the data behind this episode is the bid/offer spread at touch, consolidated across relevant venues. The xyt product set includes effective spreads with mark-outs (covered in Liquidity Matters Episode 8) and spread at depth for a range of notional value equivalent trades. These improve comparison with other products and trends over time and help to identify the liquidity premium, reflecting the spread as it applies to orders of different sizes. We will return to these topics in future episodes including a deep dive into spreads associated with stress days.
