Overview of Smart Order Routing for Efficient and Compliant Trading

Smart Order Routing (SOR) consists in processing orders on trading venues the most optimal way. On the surface, it’s pretty simple: as a broker-dealer you get from a customer a request to buy a certain amount of a company’s stock, you take a quick look at the different exchanges, maybe you will split this order if its volume is high, and that’s it. Well, not quite…

State of affairs

The two main factors complexifying SOR are

1. Fragmentation of trading venues: this fragmentation is more or less pronounced depending on the asset class (equities, derivatives, forex), with some mainly traded on exchanges and some almost exclusively OTC. Even for a pretty vanilla asset class like equities, there are several types of venues for executing an order: exchanges (like the NYSE), dark pools (pools of orders keeping them quite obfuscated for preventing large orders from skewing the market, often managed by major banks) etc. For each of them, there are a number of parameters to consider, like: price, liquidity (how abundant the order book for this asset on this venue in order to absorb absorb smoothly the order at hand), execution time, order rejection rate etc. Getting live data about the state of the available venues, navigating this information in order to design the best executions can be a daunting tasks

2. High Frequency trading: since its inception in the mid-2000 by pioneers like Dan Matthison, HFT captured a large part of the trading volume. According to the Reasearch Service of the US Congress, the algorithms behind, often highly sophisticated and efficient, capture now over 40% of stock trades in Europe. Lack of SOR awareness will make your orders get “eaten alive” by those bots, leading to compounding substantial losses.

Market drivers

The two main market drivers for the adoption of SOR solutions are regulation and ROI

Regulation

The regulations regarding SOR are mainly focused on ensuring best execution for the customers of execution venues. The main are:

• Reg NMS (US): published in 2005 by the SEC for the US market, is focusing on guaranteeing best prices for the orders, detrimental to other important factors like execution time etc., thus criticized, and could be revised in the near future

• MIFI II (EU): published by the ESMA, aims at guaranteeing best effort when executing orders.

With regulations come reporting duties, like mandatory BestEx (best execution) reports, Top 5 Venues Execution report etc.

ROI

SOR can generate a significant return on investment for trading venues and their customers, in particular through:

• lower execution costs
• lower operational expenses (in IT infrastructures, operations, maintenance…)
• competitive advantage for brokers, getting visible market differenciation

Technical challenges

• IT infrastructure: Seriously doing SOR implies to ingest and react to a constant tsunami of live data, requiring advanced, performant and reliable IT infrastructure deploying SmartNIC, implementing kernel by-passes, with parallelized GPU computation etc. Reducing the latency is key.
• Monitoring the execution quality through TCA (Total Cost analysis) is of paramount importance too, for business as well as for regulatory reasons. A TCA monitoring should comprise factors like:
• execution price, related to slippage prevention
  • execution latency (time for an order to get registered on the trading venue)
  • execution speed (by the trading venue)
  • fill ratio
  • spread (to the mid-point between bid and ask at execution)

Some of these metrics are not in the control of the brokers, who should those monitor them in order to dynamically adapt their order execution strategy when needed.

• Execution algorithm that should be:
  • explainable, so that its decisions can get reviewed by the broker and audited by the regulator
  • fast, for reducing latency and vulnerability of the orders. Will probably require a low-level implementation in C/C++/Rust, or CUDA if aimed for GPUs.
  • adaptable, self-learning, for adapting to changing conditions (price, liquidity, latency…) on the trading venues at hand

How to choose a SOR solution

Considering the asset classes you are trading, the most important points to consider when reviewing a SOR solution are:

• number and diversity of trading venues provided
• the quality of performance analysis for those venues
• user-friendliness of its UI
• metrics used to evaluation the transactions
• performance of the core technology / algorithms
• number and quality of the execution strategies provided
• if cloud-based: quality of the infrastructure, measured through latency, throughput, uptime guaranteed through a solid SLA

Steps to Implement

It could be a good idea to divide and conquer the SOR challenge through smaller, manageable and measurable steps, like:

1. TCA evaluation: which metrics are mandatory? Which ones would make sense from an operational viewpoint? From a strategic viewpoint?
2. Remaining regulatory tasks, typically report generation
3. Implementing / improving collection, ingestion and processing of live data related to the venues for the asset classes traded
4. Implementing / improving execution algorithm
5. Enhance IT-Infrastructure

Conclusion

For banks, funds, broker-dealers and every trading venue in general, SOR is of operational and strategic relevance, deeply impacting business and compliance. State of the art is evolving fast following always more fragmenting markets and efficient actors.

---

---