Tell us more about Deutsche Börse’s new dark trading offering coming next week.
We’re very much looking forward to our launch on Monday (9 December) and are definitely optimistic. We have a good mix of clients – both international and German – who are ready to start from day one.
Once we got approval from regulators on 28 November – who of course look into every detail which takes time when it comes to Q&As etc. – we allowed time for everyone to get prepared.
The banks and proprietary trading firms who are ready to go have more than 40% of the market share on Xetra today, and we also have some more in the pipeline that are interested and in preparation.
We will see how it goes on Monday and the first weeks after launch, but we have a good idea of how liquidity in the midpoint book will evolve. As you know we have two order types, the sweep orders and the midpoint orders. Midpoint orders stay in the midpoint book if they are not immediately executed while sweep orders go directly to the lit book in case they don’t get executed in the midpoint book at entry. We actually designed this in a way that makes it a pure upside to use the sweep option compared to a regular order that you would send directly to the lit book: it has no latency disadvantage and it costs you the same.
One step further, if you’re lucky, you can find a price improvement saving half the spread compared to the aggressive execution in the lit book. Therefore, we expect that sweep will be used heavily even if not at all times. It’s an interesting offer where you can use the sweep functionality with no disadvantage at all and then the midpoint orders that will be passively written in the book will not have to wait for long to get executed.
In terms of the initial launch, we’re also starting with an attractive incentive programme for the midpoint orders to start with, just to ensure maximum efficacy.
We’ve previously spoken about the idea of an integrated approach to dark trading, why is this important?
I think it’s very important because when we took the decision to enter that segment it’s not an entirely new thing. There are dark pools all over the place but being the reference market (source for reference prices) is unique and allows us to tailor in a way that others simply cannot.
First of all, we have the real time midpoint prices that other venues also have to use so whenever an external dark pool offers trading in German equities they have to use those prices in order to calculate that midpoint and technically we have those prices without any delay.
This also means there’s no stale pricing, and thus there’s no risk that you have some players in our dark pool that have information advantages over others. This is different from external venues to which a latency arbitrageur could leverage a faster private connection in order to have information on changes in the reference price first transported to that venue before the reference price is even updated there.
The midpoint book is hosted on the same machine as the lit and incoming order messages are being sequenced strictly across both those books and we can process the sweep order execution in one matcher transaction.
That means in the midpoint book where it might find some immediate execution there and then it goes to the central limit order book. Essentially, it’s only one transaction and only after all those things have happened the next order message is being processed by our matching engine. Therefore, there’s no overtaking. I think that’s compelling.
Can you tell us more about how this is a client-led development?
The idea was born back in Spring 2023 when some of my colleagues came back from a roundtable with some clients and debriefed that midpoint was something we should be looking into.
From there it was basically a joint brainstorming between some clients and us on how we could best leverage our position as reference market and importantly how we could do a complementary offering to others that already exist.
Following that comes finding a release time. We have many ideas and allocate our developer resources carefully at our company that it’s almost like an internal competition as well when it comes to what can be prioritised, but here with clients so involved of course we are focused here. At the end of the day, it’s a business case and when it comes to meeting client needs it’s a back and forth – inspired by an initial idea and then build upon.
Following the roundtable, we continued bilaterally with those clients we felt were most interested and most committed to providing their insight, really digging into what exactly they wanted from this.
What are the main technical challenges Deutsche Börse is prioritising when it comes to further enhancing dark trading?
There were a couple of functionalities and features that we developed, guided by our clients, that we tailored especially for the special needs in midpoint trading. One of the priorities for example is that we have a minimum acceptable quantity (MAQ), something that when we consulted clients was flagged as essential.
Next thing, our clients were very specific on wanting dedicated MIC codes for the regulatory reporting. Essentially all the transactions from our midpoint book are reportable under a dedicated MIC only for this purpose.
Third, it may seem like a small detail but also an interesting thing when it comes to user friendliness: We’ve learnt that clients, if they use the sweep order, want to do it without even looking at the specific situation or the specific instruments. The result is that on Xetra Midpoint you can use sweep orders even in instruments or at times where the midpoint book is currently not available, without getting an order reject – it will simply be routed to the lit book.
Another focus is on self-match prevention, as exists in our lit books. Instead of re-using our existing functionality, we learned that actually it’s much more efficient in midpoint trading to use a “cancel passive” logic that allows a trader to switch from bid to ask by simply sending a new order that will automatically delete the older orders marked with the same ID on the other side while the new order is entirely sustained.
And last but not least, the matching algorithm that we use in our midpoint book is different from our central limit order book – optimised to maximise executable volume, reflect individual execution constraints such as MAQs, and checking for potential executions not only when new orders come in, but also if the reference price changes.