Anyone who has bought or sold a home knows that, prior to closing, an inspector usually verifies the property's condition. The buyer doesn't simply accept the seller's representation that the home is sturdy and unlikely to collapse. An inspection lays bare the houses faults, giving the buyer comfort in the home's condition and absolving the seller of ongoing responsibility.
While such a careful examination might not be typical of most markets, the European financial markets are not most markets, and regulators are imposing a type of inspection requirement under MiFID II Article 17.
Article 17 and Regulatory Technical Standard 6 require that buy-side firms annually assess and validate the integrity of trading systems and maintain detailed records of such checks. Any firm that accesses broker algos directly from their O/EMS is likely to be affected and will need to have in place systems, governance and risk controls that affirm system integrity and prevent exceeding trading thresholds and the sending of erroneous orders, potentially disrupting the market. This is a high bar.
Whereas today it might be acceptable to lean on brokers for the evaluation of their own algos, the new regulatory regime wants the buy-side to more equally share this responsibility. The buy-side now enjoys the ability to contract out for best execution and trading strategies with minimal regulatory scrutiny and accountability. As of 3 January, the regulatory pressure will spread more equally to the buy-side, and it will be in their best interest to be in full control of the tools used for trading and trade analysis.
This is important as MiFID II could, in the opinion of Allan Goldstein (pictured), CFO, COO and CCO of New York-based Trade Informatics, create an environment where the buy-side reduce their reliance on traditional broker execution services, instead taking trading intelligence and best execution obligations into their own hands.
"Under Article 17, buy-side firms will be required to perform a series of pre-game checks on the algorithmic trading systems they use and will ultimately need to validate the integrity of those systems," says Goldstein. "How can this effectively be achieved if they rely on outsourced broker-sponsored algorithms? How will managers properly validate the integrity of a trading system and the results of sell-side trade analysis if they don't have full access to, and knowledge of, the underlying technology? Recent ESMA Q&A clarified that RTS 6 requirements will not apply when a firm transmits an order to another, who in turn uses an algorithm. Will traders resort to this antiquated workflow simply to avoid RTS 6? I think not."
But that's exactly the position managers may find themselves in when engaging directly with broker algorithms under MiFID II. While they might have a vague idea of how orders are being routed and filled, they almost certainly have no way of looking under the hood and often don't have complete control over routing preferences. And, they can't send the equivalent of a home inspector to evaluate and validate the technology they are using.
"I believe that in the long run, MiFID II obligations will compel the buy-side to disintermediate aspects of traditional sell-side execution services, and accelerate the trend towards the buy-side taking more control over their trade implementation process."
In January 2014, the Hong Kong Securities and Futures Commission ("SFC") issued rules requiring the sell-side to provide "training materials" to the buy-side in an effort to educate users about their strategies including details of testing and development details. However, in the MiFID II regime, will this suffice and is it really worth the effort to source, review and rely on sell-side provided documentation? As unbundling of research and execution becomes a reality, the current practice of spreading orders out to multiple broker-sponsored algorithms will quickly reveal that commoditisation of execution strategies is prevalent, even among those offered as "customised". So will the buy-side choose to skirt broker algorithms completely? Is this even a viable option?
"There exist today broker-neutral trading platforms that make use of historical trading data to create a truly customised systematic execution solution. By using these platforms, the buy-side can regain control of and transparency into their own trading strategies. They'll be able to understand, document and participate in every aspect of how execution strategies are built, tested and customised. By matching each alpha source to an appropriate execution strategy, the implementation of a portfolio idea is never separated from the idea itself. This approach is truly customised algorithmic trading informed by the alpha profile of the manager and is the best way to bring realised returns in line with expected returns."
So where does the sell-side fit into this scenario? Goldstein suggests: "In this scenario, the buy-side would continue to use brokers for market access, liquidity sourcing and market microstructure expertise optimising their experience at the point of sale. The sell-side smart order routers (SORs) are quite sophisticated today as they allow for child order placement optimisation. Parent orders should reside in the buy-side's system, brokers should only see child orders thereby minimising information leakage. So the buy-side can lean on their brokers to help them understand and direct where child orders should be going supported by 3rd party post trade venue analysis via TCA (Transaction Cost Analysis)."
On the market microstructure point, sell-side firms and market makers have been investing a significant amount over the last 12 months in smart order routers. In a deal worth USD 1.5 billion, high speed trading firm Virtu recently agreed to acquire KCG, which has built out a significant market access technology infrastructure, including sophisticated SORs, microwave transmitters and co-located servers.
"Microstructure detail deals with child order placement. Once the child orders are parcelled out by the buy-side, the sell-side knows what to do with them. The point I'm making is that the buy-side is going to become the owner of how and when those child orders are parcelled out," states Goldstein.
The upshot to this is that the buy-side will likely require better performing systems and a more robust approach to trading when they truly own the trading expertise with respect to parent orders and the sell-side takes on a specialised role as market microstructure experts.
"I think there will be a push by the buy-side. There should be a push in this direction. If not, and managers take a business as usual approach, using their brokers' commoditised algorithms is not likely to lead to performance gains. It will limit the amount of customisation and actual transparency to achieve sufficient steps toward best execution leaving a regulatory vulnerability," concludes Goldstein.
