Tick Size: To Change Or Not To Change?

By Arnaud Bossard, Senior Quant Analyst, BMLL.

First published by TabbFORUM, 30 April, 2024

Tick size is critical to the price formation and liquidity provision process and rarely changes. But when it does, granular Level 3 data can shed light on the impact of tick size changes and what it means for market participants.

A deeper dive into a tick size change, and how it reversed back.

A key feature of trading whilst using limit order books, is the tick size. The tick size is the smallest increment of price that orders can be changed, and is critical to the price formation process, giving a fixed number of potential prices for any buyer or seller. The choice of tick size is an important tool in liquidity provision, and crucial in determining liquidity, position sizes or ensuring tighter spreads. It is a hot topic of debate for regulators, exchanges and market participants across multiple asset classes. Indeed, there are ongoing debates, from Japan to the USA, on what an optimal tick size regime is.

Whilst the debate over optimal tick sizes is common, and there are many experimental models, an actual change in tick size is rare. What is more rare, is a change that is then reversed, however that can happen.

Here, we will explore an example. This case occurred on one of the most liquid Futures contracts on Eurex: the EuroStoxx 50 (FESX), which halved its tick size in June 2021, before reverting back nine months later.

Tick Size evolution on FESX contracts

Digging deeper

Using Level 3 historical data we are able to dig deeper to examine the impact of the Eurex Deutschland decision to reduce that tick in outright contracts of the EuroStoxx 50, and consider potential reasons for the adjustment back to the original value.

The two plots below show the aggregated Level 2 Limit Order Book (LOB) before and after the first change (on the 2021-06-21) for a similar price range:

Typical LOB snapshot for FESX Futures before the first tick size change (tick=1.0)

Typical LOB snapshot for FESX Futures after the first tick size change (tick=0.5)

We can see that, unsurprisingly, we get twice as many price levels after the tick change. We can also see the volume at each price level going down. If the typical volume was two times smaller, the total volume in the book depth would be retained, but, in this example, we can see it is four times smaller. Notice that despite a smaller tick, the order book is still tick size constrained (an asset is tick constrained when its quoted spread is most of the time equal to the minimum tick).

Change over time

In the plot below, we show the historical evolution of the average volume at a price level for the front month future between May 2021 and September 2022. It confirms the previous observations: a decrease in volume bigger than a factor of two. It is, as we had seen, four times smaller on the September 2021 maturity and even goes down to more than ten times smaller at the end of the March 2022 maturity.

It is also worth noting that, after the tick changes back in 2022, it does not go back to the previous level.

Volume per Level evolution on FESX contracts

Positive impact

A natural question is what does that mean for market quality. We can observe some obvious positive effects: a smaller bid-ask spread at first level, less time and a bigger probability to execute for passive orders in the book.

Bid-Ask Spread evolution on FESX contracts

Time to be Executed on FESX contracts

Probability to be Executed on FESX contracts

Not everything has improved

But, the picture is more balanced than it might appear at first sight. We had observed that the volume per level was significantly smaller than the expected factor of two, leading to greater execution costs for large orders.

Max Size for Aggressive Trades on FESX Contracts

Let’s look, for example, at the top 2% largest aggressive orders. First of all, they are not insignificant as they typically represent 30% of the total volume traded. Moreover, the necessity to consume multiple levels and thus go deeper in the book seems to have discouraged some agents from trading aggressively on large sizes.

STF Spread for 1000 contracts on FESX

We can also see that the sweep to fill (STF) spread (defined as the spread between effective prices on each side of the book for a given aggressive quantity) for orders of size 1000 (a typical order of magnitude for largest orders in the top 2% as seen on the previous plot) has clearly increased during the period of time when the tick was 0.5.

This increase of the STF spread, especially on the last maturities during the 0.5 tick phase, corresponds to a significant decrease of the volume available at each price level as we have seen on the fourth plot. A further analysis shows that the STF spread is proportional to the inverse of that average volume per level.

Conclusion

In this particular example, we see that using the granular Level 3 data to build derived metrics (such as the STF spread) reveals a trading behaviour that may explain why Eurex has reverted the tick change after less than a year. For markets wishing to implement a similar tick change, it is worth noting that trading did not go back to ‘normal’ on some metrics after the tick size came back to its initial value.

More generally, using high quality historical data and metrics can help market participants on a variety of aspects including best execution and market impact, as they consider implementing a significant change in production.

Read the article on TabbFORUM here.