Systematic Internalisers – Data holds key to challenges and opportunities in MiFID II era

MiFID II broadens the scope of the systematic internaliser (SI) regime, not only to cover more asset classes, but also to apply more stringent rules, with the aim of improving pre-trade transparency and improving price efficiency and client executions. To examine the far-reaching consequences and practical implications for liquidity seekers and providers, against a backdrop of tougher Best Execution requirements, Chris Hall and Mike O’Hara of The Realization Group talk to Ollie Cadman of Vela, Citi’s James Baugh, Sun Trading’s Jamal Tarazi (Sun Trading is now part of Hudson River Trading), Rosenblatt Securities’ Anish Puaar and Alasdair Haynes of Aquis Exchange.

Where and how will European market participants find liquidity? This fundamental question is under intense scrutiny in the early stages of the MiFID II regime, which began its roll out in January 2018. It has risen to the top of the agenda for investment firms because the new regulatory framework demands they now take all ‘sufficient’ steps to achieve best execution, rather than all ‘reasonable’ steps as previously required.

Unlike its equities-focused predecessor directive, MiFID II brings its core principles of transparency, accountability and investor protection to all major asset classes. This requires significant process change in OTC markets, with credit and swaps trading being steered onto trading venues, having been highly voice-brokered previously. But the new mandate to take all sufficient steps to deliver best execution – thus theoretically providing better outcomes to end-investors – is also extremely challenging in equities as MiFID II eliminates or reduces access to certain established and popular liquidity and crossing pools.

The emergence of the systematic internaliser (SI), defined in MiFID II as “an investment firm which, on an organised, frequent systematic and substantial basis, deals on own account when executing client orders outside a regulated market, an MTF or an OTF without operating a multilateral system”, is one of several MiFID-2 compliant solutions that are now used to achieve best execution for clients. These include conditional order & block trading venues using the large-in-scale (LIS) waiver, trading in auctions and new execution venues falling under the SI regime. To meet toughened best execution obligations, market participants must be able to efficiently tap liquidity in SIs, despite the many current uncertainties about their role in an increasingly fragmented trading landscape.

And unlike other jurisdictions, there is currently no consolidated price feed in Europe, defining best price across all liquidity sources. “Real-time price transparency of any individual instrument across all trading venues, including the SIs, is going to be crucial to demonstrating best execution. This can only really be achieved by normalised, low-latency aggregation of market data. In the absence of a consolidated tape, firms are going to want to create their own European best bid and offer record,” says Ollie Cadman, head of business operations, Vela.

Out of the dark

MiFID II bans broker crossing networks (BCNs) – effectively unregulated platforms operated by banks – and limits trading in dark pools. BCNs were liked by the sell- and buy-side as flexible, low-cost, low-impact liquidity sources. Often, they operated in the dark, i.e. pricing was not displayed in advance, which helped to minimise market impact. They also appealed to electronic liquidity providers (ELPs), which provided as much as 50% of flow in some BCNs.

“Real-time price transparency of any individual instrument across all trading venues, including the SIs, is going to be crucial to demonstrating best execution”
Ollie Cadman, head of business operations, Vela


From 2014-17, dark trading regularly accounted for 8-10% of European equity trading volumes, according to Rosenblatt Securities, with around half of dark trading conducted on BCNs, but far from all trades were block size. While BCNs have been banned, dark trading has been only limited, in recognition of its value in protecting block trades from adverse market impact. Orders above a specified ‘large in scale’ (LIS) threshold can be executed without pre-trade transparency, as can a strictly-monitored percentage of orders below the threshold, under a system of ‘double volume caps’ that has been operational since March.

SIs have been around in the equities markets since MiFID I, but have only become an attractive channel for liquidity provision with the demise of BCNs. MiFID II also extends the scope of SIs beyond equity-like instruments to bonds, derivatives, emissions allowances and structured finance products.

In effect, the SI is the MiFID II-compliant channel for large-scale, automated market-making services by banks and ELPs. An asset manager might route an order to a broker’s SI, for example, in the expectation of getting a better execution in terms of price, hit rate and in particular market impact, all of which are considered primary factors to determine the quality of interaction.

The bilateral design of SIs effectively divides ELP liquidity from that offered by banks and brokers. Anish Puaar, European market structure analyst at Rosenblatt Securities, suggests there are pros and cons in this separation of liquidity. “Instead of bumping up against ELPs in a BCN, the buy-side will now be doing this through an ELP’s dedicated SI. On one side that’s fragmentation, but on the other side it’s more transparent, because you know exactly who you’re dealing with. If trading with an SI is not giving you good execution performance, then you’ll cut it off, just like any trading venue,” he says.

“If trading with an SI is not giving you good execution performance, then you’ll cut it off, just like any trading venue”
Anish Puaar, European market structure analyst, Rosenblatt Securities


Alongside traditional market-making, banks’ and brokers’ SIs offer the buy side the prospect of trading against flow from central risk books, which could yield block trading opportunities in multiples of average daily volume. Aquis Exchange CEO Alasdair Haynes suggests that providing only their own capital via SIs will make banks’ liquidity provision efforts more efficient than when relying on ELPs. “Even with recent changes to capital regulations, we may see a trend towards banks doing more principal trading and capital provision for their clients, via SIs,” he says.

James Baugh, head of European market structure at Citi, agrees banks can be more targeted with the type of principal flow they present through SIs.

“Clients will face a different experience when interacting with bank versus ELP liquidity, stemming from their different approaches to unwinding risk,” he adds. “ELPs’ spread capturetype strategies inevitably involve some level of impact when they hedge the position on the primary exchange, which is very different to how banks manage that unwind.”

Multiple models

Each sell-side firm will have their own nuanced approach to running an SI, but there is likely to be even more diversity among ELPs.

According to Jamal Tarazi, European business development at Hudson River Trading, becoming an equity-only SI is a “natural extension” of the firm’s ten-year record of providing bespoke liquidity in equities and related markets. He says success for ELP SIs will depend on the quality of interaction with trading counterparts, i.e. the brokers looking for liquidity on behalf of buy-side clients. ELP SIs will offer a bespoke feed to each broking client based on the liquidity needs of their different client bases, rather than supplying a one-size-fits-all service.

“The distinguishing feature will be the understanding of how that interaction happens, i.e. how one fine-tunes liquidity provision models to the requirements of each client, through constant feedback,” says Tarazi. “Is the client looking for size, price improvement, or an urgent fill? All these factors get priced into the quotes.” Tarazi then adds “The most important feature in the SI regime is counterparty transparency. For the first time, the buy-side will be able to measure the performance of various market makers in a systematic fashion. With the data at hand, buy-side trading firms will hold the ELPs accountable for the liquidity they provide. That’s an exciting prospect for both parties!”

While SI operators will offer multiple options to liquidity seekers, Citi’s Baugh says their appeal will vary not only between institutions, but across the order life cycle. “A buy-side firm might not, at least in the early stages of a trade, want to place child orders with an ELP SI as this might offer valuable insight into the parent order,” he says. “Although you might not want to send signals to an ELP SI with 70% of the parent order left to complete, your routing strategy might broaden if you’re looking to get the last 10% done.”

Tarazi contends that this situation is not limited to ELPs, however. “Information leakage needs to be managed against all SIs, and one should look at performance characteristics for the parent order and not just the child orders”, he says. “The buy-side should let the data arbitrate the routing decisions for brokers.”

New paradigm for routers

In a fragmented post-MiFID II world, the interaction of routers with diverse SI models may initially add a further layer of complexity. As mentioned, ELP SIs can and are offering different price streams to multiple brokers, potentially offering a more attractive price feed to a broker’s long-only asset management clients than, say, quant-based hedge funds. 

“Buy-side firms will have to work very closely with brokers to understand how they’re utilising this liquidity.”
Jamal Tarazi, European business development, Hudson River Trading


“If a buy-side client utilises multiple brokers, they might find different outcomes from the same SI depending on whether the order was routed via broker one versus broker two,” says Tarazi. “Performance will vary due to differences in brokers’ technology stacks, how their algorithms work, and the extent and quality of their interaction with individual ELP SIs. Buy-side firms will have to work very closely with brokers to understand how they’re utilising this liquidity.”

Buy-side outcomes may depend on whether broker’s routers communicate with SI via ‘blind ping’ or first consuming price feeds. With blind ping, the router sends an order to an SI without detecting whether it offers liquidity that can be executed against, thus potentially giving up order information without necessarily getting filled.

With SIs disseminating tailored feeds above standard market size (SMS), broker X’s SI won’t necessarily want to distribute its higher-value quote streams to rival broker Y’s router, for reasons of competition. “Consuming data first to see if there’s a match before sending an order gives a higher fill rate and less information leakage,” says Puaar. Interaction between price feeds and routers is of course at a nascent stage. While router functionality must be reappraised, SIs are also on a steep learning curve in providing price feeds.

“With a smart order router, you can sweep across multiple venues, including dark orders, but with SIs they are reacting to quotes rather than sending orders. It will take firms time to tweak their policies appropriately, i.e. figuring out how much liquidity they want to expose to SIs,” says Ollie Cadman of Vela. “Some SI operators aren’t necessarily used to making data feeds available. We’ve been working with some to help them understand what a good feed looks like, discussing challenges around symbology and managing roll-outs and updates to

Citi’s Baugh notes that the market will need the support of technology vendors and market data specialists in “normalising and standardising” access to SIs. Nevertheless, he acknowledges the challenges of a more fragmented liquidity environment. “It could be much harder to establish best execution if ELPs are sending private streams of quotes to banks and brokers that are fundamentally very different,” he says.

Puaar sees parallels between the level of routing information required by the buy-side under MiFID II with the due diligence undertaken by US asset managers in light of past dark pool scandals. “How does your SI work, how are you consuming data, how are you routing to other SIs? How does the broker preference its own SI before routing to other venues? Does it place a proportion of client orders into its own SI before routing externally? They’ve asked similar questions before, but need to ask how things have changed with the introduction of SIs.”

Too much of a good thing

How much former BCN flow will SIs capture? Only time will tell. A key aim of banning BCNs and restricting dark trading was to bring off-exchange equities trading between brokers and their clients under similar rules as trades conducted on trading venues – i.e. regulated markets, multilateral trading facilities (MTFs) and organised trading facilities (OTFs) – in the interests of transparency and fair pricing to end-investors. Critically, policy-makers were keen to prevent OTC trading from undermining price formation on trading venues, by requiring SIs to meet similar transparency rules. Exchanges and MTFs are also making their bid for ex-BCN business, via innovations such as periodic auctions and conditional order venues. If too high a proportion of trading is conducted without pre-trade transparency, the credibility of the price discovery function is diminished, thus potentially damaging investor confidence.

SIs must meet MiFID II rules on pre-and post-trade transparency, quote and trade matching, best execution reporting, and reference data reporting, with pre- and post-trade transparency waivers for illiquid and LIS trades applying to both for SIs and trading venues. SIs can update quotes at any time, and may limit the number of transactions per client as well as the number of clients to whom firm quotes and provided (for illiquid / LIS trades).

The European Commission has closed a loophole that could have allowed SI users to avoid pre- and post-trade transparency requirements, but there remain key differences with on-venue trading. In particular, SIs do not have to provide pre-trade transparency for their prices for orders above standard market size, which offers scope to undercut trading venues. The latter have voiced concern that SIs could capture significant market share – thus undermining price formation – as investment firms strive to prove best execution. The
European Securities and Markets Authority (ESMA) has responded by making SIs abide by the same tick price rules as trading venues, but the lack of above-SMS transparency remains contentious.

“If you’re going to take a portion of the lit book and start bilaterally trading it in a non-transparent way, that could potentially create all sorts of issues. If the SI is used as another tool to get better pricing, fantastic; but if it becomes dominant, it’s actually a very bad thing,” says Haynes at Aquis, who suggests regulators might step in if SIs grow too rapidly.

Still in the dark?

At this early stage in the new SI regime, all market participants and operators are feeling their way gradually, not helped perhaps by teething troubles in a new reporting regime struggling to distinguish price-forming trades, for example. Rosenblatt’s Puaar remains optimistic that data flows will support best execution. “Once SIs have become embedded, firms will have the data to be able to compare the performance they get from SIs, lit exchanges, dark MTFs and block venues. They’ll be able to make a judgement as to where bank and ELP SIs should sit for different types of orders”, he says.

“MiFID II is making the market transparent so that we can make fair decisions, and ensure the right result for the end-investor”
Alasdair Haynes, CEO, Aquis Exchange


Under MIFID II’s RTS 27/8, brokers must identify their top five most used trading venues and SIs, on a quarterly basis. Most observers agree this is of value, but call for more timely publication schedules. “In sport, people now win or lose fairly because technology can deliver real-time transparency,” says Haynes. “MiFID II is making the market transparent so that we can make fair decisions, and ensure the right result for the end-investor. RTS 27 and 28 don’t think go far enough. We shouldn’t be waiting three months, and the data should be available for everybody, so they can judge where orders are being sent, and how different venues are being used. This greater transparency will ultimately decide how best execution is achieved.”

The timetable and mechanisms for inclusion in the SI regime also make it hard to predict today the level of liquidity they will eventually amass. While firms can opt in voluntarily, they will soon be obliged to do so if the volume of their trading activity with clients reaches certain thresholds. As the threshold is calculated as a percentage of overall trading in a particular instrument, a firm may fall in and out of the regime every three months (in the case of equities), i.e. when ESMA publishes the denominator.

“Some firms are already engaging with their brokers, deciding to avoid certain types of flow until they have sufficient data to make an informed judgement”
James Baugh, head of European market structure, Citi


ESMA will first publish total number and volume data for the first half of 2018 on August 1, with investment firms obliged to undertake assessments on SI status by September 1. Subsequently, ESMA will provide updates on the first day of November, February, May and August, with firms meeting their SI obligations from the 15th day of that month.

With so many elements of the new environment incomplete there is considerable uncertainty about where liquidity will coalesce. Citi’s Baugh sees a diverse range of approaches across the buy-side. “Some firms are already engaging with their brokers, deciding to avoid certain types of flow until they have sufficient data to make an informed judgement,” he explains. “Others are looking to effectively outsource best execution to their brokers. They’re trusting brokers to access liquidity across multiple destinations, but will look to make adjustments based on TCA if necessary.”

In a fragmented trading world, it seems, timely and granular information will be critical to firms looking to support and achieve best execution. “At the end of the day, it’s all going to come out in the data,” notes Puaar.

For more information on the companies mentioned in this article visit: