Market valuations have been under pressure in 2022 but financial data volumes have only gone one way: up. How has the dynamic impacted surveillance approaches for financial services firms?
In April, Eventus participated in the RegTech Asia 2022 conference. Vince Turcotte, Director Digital Assets, Asia Pacific, joined a panel examining how recent challenges of elevated trading volumes, remote work, and an increasingly volatile market environment have all intensified the pressure on financial institutions to catch signs of market abuse, as well as the need for technology to do it.
Data, the panelists agreed, is the key to everything.
But collecting, storing, analyzing, and even assigning responsibility for data remains an institutional challenge.
Joining Vince were Selina Tindall, a Senior Solutions Consultant at Smarsh, Damian Bearzatto, Head of Product & Capability, Risk Surveillance at Macquarie and Deepanvita Upadhyay, Head of Front Office Supervision, Conduct and Control at Nomura International Wealth Management, with Regulation Asia’s editor, Manesh Samtani acting as moderator.
In the old world, there was a simple proposition that time is money; it still is, but from the panel’s point of view – data now also equals both time and money and managing data is critical for profitability at financial services firms.
It’s not just the sheer volume of data that is inundating companies and challenging capabilities; it’s also a surge in the variety of data types, formats, and delivery mechanisms. All of it must be managed and monitored, not only for trade surveillance and monitoring, but for myriad functions in the post-trade process, as well as more general corporate operations.
Among the challenges that volume and variety present is that all data, not just trade records, are potential artifacts for trade surveillance, each with its own idiosyncrasies around storage and security, as well as formatting and transformation issues.
It’s worth noting that Validus can ingest data from across an organization—from simple flat files, to exchange drops, ISV feeds, to all manner of APIs—and run extract, transform and load (ETL) processes to reform disconnected sources into organized structures. The result is a private data lake, from which clients can extract and repurpose information while still accommodating relatively unsophisticated or inexperienced users—whose expertise or data needs may be in other fields or functions within the company.
We understand that in financial services, data has become a critical asset, and can represent substantial costs; but it should never become an obstacle to fast and accurate surveillance.
Financial data isn’t all numbers or unified
A data dimension that the panelists highlighted repeatedly is that financial services companies typically have massive data warehouses that handle different data libraries in several formats. Critical data does not exist in a singular usable format.
A byproduct of Validus data transformation efforts is a harmonized data library composed of client trade lifecycle data, market-price data, and client static data, which can all provide a foundation for surveillance and other post-trade functionality.
Remote work during Covid-19 has further complicated this issue; a firm’s data flow may come via messaging platforms or on the telephone, lacking the formal structure of an in-house order management system. Yet every client touchpoint remains a data point that may contain risk indicators.
For surveillance purposes, voice or messaging orders complicate real-time data capture, requiring additional data mapping to link what was traded relative to what is allocated to clients. Tracking the start of the trade, to its actual end, may also insert layers of data systems, potentially with different versions and formats. People working at banks and financial firms recognize the challenge this entails, but with the predominance of legacy systems, incremental upgrades and new technologies, the challenge can be multiplied as well as common.
Also common are surges in order-message volume during times of market volatility, as is being seen today. These instances push compliance teams to their limit while the legacy systems they use may create risky dilemmas. Surveillance officers have to choose between drawing samples and hoping for the best in a flood of “false positive” alerts, or recalibrating alerts to reduce workflow. In either case, they are leaving much to chance.
Integrating supervised machine learning (ML) into surveillance processes, helps manage workflow in this era of surging data. ML casts a wider net, examining all potential suspicious activity and replacing sampling techniques to ultimately produce a more precise set of candidate alerts. Company analysts can then review a manageable set of high-probability alerts, while lower probability candidates can be graded and stored.
Each type of alert has 30-40 features, and Eventus maintains training databases with tens of thousands of sample alerts to refine ML processes.
A second level of refinement is handled through Robotic Process Automation (RPA). Eventus works with clients to find opportunities to automate alert analysis via small snippets of code, helping analysts focus resources on the highest priority candidates, and freeing time previously spent with false positives.
Periodic management reviews of stored alerts also grade the performance of the ML and RPA models, ensuring that alert calibration is correct, and allowing for further fine tuning and grouping of alerts.
Down the road, the Reg-Asia panelists emphasized, financial firms will need to apply even more advanced data mining conventions to their surveillance, and not just from a process or productivity perspective. Increasingly, this is also the view from a regulatory perspective. As above, Eventus helps consolidate trade lifecycle data in a unified format, which can be accessed at any time to take a look back through activity. This is quickly becoming a baseline requirement.
Financial firms need to be able to connect activity involving multitudes of accounts trading across a variety of assets and markets, or multiple regulatory jurisdictions. Simply facilitating that activity is already a big data challenge, ensuring it is also free of fraudulent or collusive behavior is the fundamental part of that challenge that the Reg-Asia panel was organized to address.
Finance needs to future-proof
As practitioners on the panel pointed out, handling data is about more than formats or storage; data also has a lifecycle and context to consider, which may be complex but it’s that interdependence of time and context that financial firms will need to address going forward.
Automation, data analytics and machine learning can help them bring control back to their surveillance program. And that’s where Eventus brings its expertise, taking trade data in any format from any source, and transforming it into uniform, accessible and recyclable formats.
Time is still money and the less of it that financial firms have to dedicate to formatting data, the more value and risk reduction they can derive from it.
Watch the replay of the panel below.
Eventus is a leading global provider of multi-asset class trade surveillance and market risk solutions. Its powerful, award-winning Validus platform is easy to deploy, customize and operate across equities, options, futures, foreign exchange (FX), fixed income and digital asset markets. Validus is proven in the most complex, high-volume and real-time environments of tier-1 banks, broker-dealers, futures commission merchants (FCMs), proprietary trading groups, market centers, buy-side institutions, energy and commodity trading firms, and regulators. The company’s rapidly growing client base relies on Validus and Eventus’ responsive support and product development teams to overcome its most pressing regulatory challenges. For more, visit www.eventus.com.