Pegah Esmaeili, Virtu Financial & Simon Bonde, Nordea Asset Management
The Nordic region has long been a key market for Virtu, as evidenced by its dedicated data centre in Stockholm underscoring its commitment. For many years, Virtu has closely collaborated with Nordic stakeholders including local banks, institutional investors, hedge funds, and smaller financial entities. Pegah Esmaeili, Virtu’s Head of the Nordic Region, explains that over the past year, top priorities emerging for regional clients have been gathering and normalising data, developing new algorithms, and creating and using next-generation models.
Simon Bonde, Team Lead Trading Quants at Nordea Asset Management, an investment firm with €264bn assets under management, explains the importance of data gathering and normalisation from a buy side perspective. “Having deployed algorithmic trading wheels and systematic execution frameworks for an extended period, we recognise the indispensable role of high-quality data in refining and scaling these solutions.”
Clients are increasingly interested in analysing their trading data to enhance trading results through optimal broker selection. Our experience shows that effective analysis requires clean data, proper classification, categorising broker algorithms, and normalising broker performance using a cost model. This aligns with the overall trend in the region of refining trade cost analysis programs with the goal of drawing insights and applying lessons learned to incrementally improve trading outcomes over time.
“We have leveraged the expertise and comprehensive data sets of our Transaction Cost Analysis (TCA) providers to normalise trading costs through cost models, facilitating an objective, apples-to-apples comparison of our brokers,” Bonde notes.
When behavioural change is based on post-trade results, it is imperative that pre- and post-trade cost models are aligned to provide consistency throughout the trading lifecycle and overall feedback process.
Recent analysis of Virtu’s Global Algo Peer universe data (comprising 9 million orders and $2.4 trillion traded between 2018 and 2023) shows Implementation Shortfall (IS) remains the most common wheel algorithm, with Close wheel algos gaining momentum. Additionally, the number of brokers on these algo wheels has remained stable or decreased. However, a key challenge for algorithmic brokers is that a one-size-fits-all approach does not work. For example, one client’s IS algo wheel flow may exhibit different characteristics than another’s, and the flow profile changes over time. As a result, rigidly using a single algorithm strategy consistently leads to sub-optimal outcomes, regardless of customisation. This is why human traders have multiple algo strategies available – they are tools to be deployed based on the trader’s view of the order and market conditions when trading.
Esmaeili explains that clients have found their latest “algorithm of algorithms” development called Switcher very useful for addressing this challenge. Switcher is a trading strategy that selects an appropriate algorithm for each order based on the order characteristics and current market conditions, with the ability to change algos throughout the order’s lifetime. It serves as a co-pilot for traders, managing low-touch or no-touch flows (such as an algo wheel), as well as discretionary flow when the trader lacks a view on the stock. Over the past year, results using this new strategy have been positive, and Switcher is now increasingly adopted by diverse global clients. A key driver for clients is transparency – while utilising machine learning techniques where suitable, they believe trading decisions should be deterministic, explainable, and auditable in real-time. This builds trust in the algorithm.
Clients in the region are also evaluating how to optimally integrate next-generation models, machine learning or otherwise, into pre- and post-trade processes. Trading desks have begun using these models to better explain trading costs’ sources, such as the proportion caused by market impact versus effects of market momentum or spreads. Advanced users take this further by using next-gen models to simulate historical trading, understanding if changes in participation, venue usage, or timing could consistently improve outcomes.
A common starting point is enhancing trade automation and algo wheel workflows, where models can provide a cluster or difficulty score, or even recommend an algowheel trading strategy based on order characteristics and environment. Another trend is fundamental equity and fixed income portfolio managers incorporating pre-trade cost estimates and liquidity scores into portfolio construction, inherently resulting in lower-cost orders for traders. While quantitative managers have used this approach for some time, these model outputs can also greatly benefit fundamental managers when easily integrated into workflows via API or OMS.
Bonde explains, “We have been strong proponents of our TCA providers developing cost models across all asset classes. While the process of collecting and normalising data across various systems and providers is undeniably time-consuming, it is an essential undertaking for us to foster a data-driven culture and achieve our goal of democratising data across asset classes in Global Trading at Nordea Asset Management.”
Virtu’s technology-driven approach to trading and market making has been particularly successful in the Nordic region, where it has larger market share in dark trading compared to the rest of Europe. From January to July 2024, buy-side and sell-side clients contributed to nearly 17% market share in POSIT dark pool trading in the Nordics, executing large block trades with each other. Being collocated at the Digiplex data centre in Stockholm also aids its Smart Order Router logic in overcoming the challenges of trading across primary exchanges and MTFs that are geographically distant from one another. This geographic proximity benefits clients trading in the Nordic region.
“We are particularly proud of the robust presence and impact we’ve established in the Nordic region, thanks to our advanced technology, execution capabilities and strong partnerships”, Esmaeili says. “What sets us apart is our complete range of product offerings, addressing customer needs at every stage of a trade, from pre-trade to post-trade.”
