Finding Blocks – A Case Study – Industry Collaboration in Action

Share


With Robert Barnes, CEO, Turquoise, Jonathan Finney, Head of Systematic Trading EMEAA, Fidelity International, and James Hilton, Co-Head AES Sales EMEA, Credit Suisse
Robert: In an order book environment that naturally leads to small trade sizes, investors that wish to outperform benchmarks are calling for innovation in electronic block trading.
To answer this call and still trade in the presence of anticipated MiFID II double volume caps, one needs a respected and working trading mechanism that can match orders above 100% of the Large In Scale (LIS) threshold determined per stock by ESMA.
Turquoise has a working LIS innovation: Turquoise Block Discovery, which matches undisclosed Block Indications that execute in Turquoise Uncross. Turquoise design reflects active collaboration with the buy-side and sell-side. 1 , 2
Jonathan: With impending MiFID II legislation, Fidelity as a whole has commended much of the innovation that has happened in the markets and been actively engaged on several projects, many from inception. However, as a large institution we also very much err on the side of caution and, when Turquoise approached us about their proposed new venue, we met this with cautious optimism. Our concerns were varied but were mainly focussed on due diligence, risk and control analysis and general policing of participant behaviour. We could definitely see potential in their proposals but we felt there was also a lot of granular detail which needed to be worked through for us to be comfortable as a firm from a best execution perspective.
Therefore, when Turquoise Block Discovery first launched, we were not one of the first initiators but we definitely had a interest in its success and were actively engaged with Turquoise, buy-side peers and sell-side counterparts in the meantime in discussing what we needed to be comfortable with engagement. By understanding how their reputational scoring worked, having increasing comfort in the transparency of their policing mechanisms, and attaining greater understanding of the types, size and quality of trades occurring, as time went on, we felt we were in an increasingly strong position to justify adding Turquoise Block Discovery to our selection of venues via brokers, especially using the conditional order type. I am pleased to say that, though success has been relatively scarce as adoption is still in its early growth phase, the quality of liquidity has scored well across a variety of external metrics.
James: There have undoubtedly been challenges to implementation of this solution. Even before you look at the technology development, you have to spend some time to understand and get comfortable with the model – it’s relatively unique. And clearly, we had to garner feedback from clients to see whether there would actually be a demand for it.
Once we’d established that it would be a viable model, and clients wanted to use it, we invested in the technology development. Luckily for the buy-side, the heavy lifting definitely falls on the sell-side. Supporting the full conditional nature of the service has required some effort, and there are still some key brokers working on that. Because of the model, Credit Suisse offers access to the service on an opt-in basis. So, whilst you have some brokers not offering full functionality, and buy-side having to positively opt-in, invariably it takes time for the momentum to gather. The key to the service is that there’s virtually no opportunity cost. We have been meeting a lot of clients to talk about the service, and explain why this is something we feel is worth supporting and the majority are signing up.
Key metrics
Robert: The metrics that matter include low reversion, large average fill size, and high firm up rates.
Low reversion
The design of Turquoise Uncross innovates with a random timing mechanism which prevents either buyer or seller from determining the instant of execution. This has been the subject of multiple studies by LiquidMetrix, the independent analytics firm that specialises in venue performance metrics and execution quality analysis. They repeated their analysis in October 2015 and, for the third year in a row, spanning before and after the 2014 launch, and again concluded that trades occurring on Turquoise Uncross had a far lower correlation with sharp market movements on primary venues than trades occurring on other continuously matched MTF Dark Pools.3
Large average fill size
Turquoise prioritises orders by size and features innovations that deliver a broker neutral venue that is reversing the electronic trend of shrinking trade size. Turquoise Block Discovery now averages more than €250,000 per trade, and this average is more than twentyfive times larger than the average €10,000 for electronic trades matched by continuous dark order books.
High firm up rates
We have more than a year’s worth of empirical measurements evidencing consistently high firm up rates. These high firm up rates result from a strict timing mechanism and robust automated reputational scoring, which measures the difference between the original block indication and the subsequent firm order.


Since launch, more than 95% of Order Submission Requests (OSRs) resulted with firm orders within the specified time window – under half a second – and with both price and size parameters which pass the minimum requirements of the service defined with input from buy-side and sell-side users. Furthermore, in more than 90% of all OSRs, the firm order not only met the minimum requirements but also was at least the full size of the originating block indication.
Jonathan: A key advantage as a venue is the ability to utilise the conditional order type. The conditional order type effectively allows us to continue working an order independently, whilst still having potential exposure should liquidity opportunities arise. Whilst, conditional order types are not a new phenomenon in themselves, this was the first example of widespread engagement via brokers in Europe and helped us address numerous concerns, including the real fear of missed opportunities by residing in one venue when liquidity is actually in another. By using conditional orders, the opportunity cost is minimised. Now, whilst Turquoise may argue that this was not the original intent of the venue, this becomes incredibly important in a post-MiFID II world where 4/8% caps on dark will most likely drastically change the European liquidity landscape. One way round this, currently still available via MiFID II proposals, will be the Large In Scale waiver but there-in lies the rub. Current trading behaviour (by both us and our peers) seeks to distribute the order among a variety of venues, thus reducing the opportunity cost of missing liquidity, rather than rest in one sole venue. In a LIS world, there will be much fewer orders that will have the ability to be distributed amongst a variety of “dark” venues and still maintain the LIS integrity per venue. Conditional orders help alleviate this by allowing a virtual presence in Turquoise Block Discovery and thus not disrupting our workflow, a very powerful liquidity-seeking tool in the trader’s arsenal.
Increasing routing flow
James: From a technology perspective, we’re fully operational, but we did spend a lot of time optimising our work flow. Routing to a venue where you’re sometimes trying to send as large an order as possible (albeit conditional), requires some careful development.
From a client perspective, we’re doing all we can to present the benefits to our clients. An awful lot has been done in the UK already, and we’re building momentum in Europe. But we need to do more to engage with our US clients, who perhaps aren’t as involved with how the overall European market is evolving. The industry has an opportunity to help build solutions to the liquidity challenges we face, and we believe Turquoise Block Discovery is one of those solutions.
Turquoise publishes weekly stats, so we can see that we’re a significant % of the executions. We’re continuing to sign up clients, but it’s important that the wider market engages. It’s crucial that the sell-side supports the full conditional service; otherwise you simply won’t see enough large blocks crossing.
The buy-side perspective
Jonathan: Whilst Turquoise Block Discovery would not suit all needs or all order profiles (nor do they profess to do so), they do provide a compelling solution to a variety of questions regarding liquidity-sourcing and order-routing in a post MiFID II world. The problem is that currently there is no incentive to move directly to an LIS model or at any point until it is forced upon us by MiFID II. Average trades sizes are fractions of LIS and current behaviour still needs to account for the large proportion of volume occurring in these smaller trade sizes. The dilemma then is how to transition market behaviour from small dark fill sizes to large-in-scale dark fill sizes in a manner which smoothes out the issues and kinks along the way. There are clearly venues where similar behaviour takes place currently, but there are several key differences which set Turquoise Block Discovery apart from this and would thus require a change in trader behaviour. The good news is that Turquoise, the sell-side, and the buy-side are all enthusiastic in their drive to find a solution to this and I believe these discussions will drive how our engagement changes with Turquoise Block Discovery and other venues over the coming year and more.
How important has it been for Turquoise to have a buy-side user group?
Robert: Buy-side participation has been integral to defining reputational scoring and other key features that work through sell-side algorithms that access Turquoise Block Discovery.
While, MiFIR Article 4(1)(c) can allow dark trading for orders received by a venue that are LIS compared with normal market size and not include them in the MiFID II double volume cap calculations, the challenge is to increase size of marketable orders where today less than 1% of trades on European order books trade in sizes greater than LIS.
The total value matched since launch is above €1 billion. While this amount is niche relative to the addressable market for electronic block trading, Turquoise Block Discovery has a profile – 50% of Turquoise Block Discovery Block Indications and 50% value traded already are in sizes greater than LIS – that is ripe for more activity into quality workflow that features low reversion, large average fill size, and high firm up rates.
Both buy-side and sell-side tell us that the way the buy-side works orders will have to evolve, and therefore we look forward to continuing our constructive cooperation with buy-side and sell-side.
Jonathan: The question can be expanded out from this to discuss buy-side user groups in general. As previously mentioned, MiFID II could prove to be a major disruptor to current market practice and there have been various innovations proposed amongst several market constituents. It is thus incredibly important for buy-side counterparts to get together and discuss advantages and short-comings of potential solutions to MiFID II-shaped problems and, in fact, market structure issues as a whole. Through intelligent debate we will be able to move forward as an industry in a way which would best serve our end investors, rather than working separately, championing different individual projects, ultimately increasing fragmentation, raising costs and hurting our clients.
With respect to Turquoise Block Discovery, it was through discussions within the buy-side group that we were able to increase practical (rather than theoretical) understanding of the mechanisms behind the venue and take comfort that our views were not extraordinary and which eventually led to its implementation.
We’d love to hear your feedback on this article. Please click here