This article was first published by Nathan Lynch on the Thomson Reuters Regulatory Intelligence service for financial crime, risk and compliance professionals.
The booming job market for staff with "data science" expertise has presented several pitfalls for banks and regulators as they compete for talent to improve their compliance processes within a highly competitive sector. Technology experts have warned that banks and regulators could "burn capital" if they rush to secure talent without first doing due diligence on the precise "data science" roles that they need to fill.
Managers need to assess the nature of their "data science" hiring needs before hunting for staff in an overheated job market, technology providers said. This involved ensuring the business has effective curation of its data assets and is prepared for an expansion of the data science and analytics aspects of the risk and compliance function.
Banks and regulators have embraced the data science revolution in recent years but are battling against a long-term shortage of experienced staff. The latest "Top IT Skills Report" from tech hiring consultancy DevSkiller found that the trend had only continued; data scientists are the most sought-after segment across the worldwide IT job market. The report found there had been a three-fold increase in the number of data science-related roles that recruiters were attempting to fill throughout 2021.
The hiring squeeze has affected banks and other financial services firms, as they seek to embed "data science" expertise within their risk and compliance teams. The competitive job market has been particularly challenging for regulators, however, as they are restricted in the salaries they can offer to candidates.
Demand for "data science" roles was growing rapidly. Organisations were struggling to find suitable talent with experience in the financial services sector, said Jacob Smith, a compliance recruitment consultant with JS Careers in Sydney.
"The challenge is finding candidates with suitable skills in what is a relatively new field, overlaid with the very high demand for candidates in the regulatory compliance field in general," Smith said. "Recruiting into the data science space is most certainly one of the most challenging areas at present. The salary levels being offered are comparatively high."
Dr Janet Bastiman, Chief Data Scientist at RegTech firm Napier, said organisations could exhaust capital quickly if they built teams without first setting out a clear set of objectives and target skills. There were many different skills sets and career roles within the broad category of "data science", which increased the risk of making the wrong hiring decisions, she said.
"The term 'data scientist' is such a big, broad area. I think it's something that's often misused and misunderstood," she said. "If you're hiring for a data scientist, you need to understand what it is you really want from them, or you'll end up making a lot of very expensive, incorrect hires."
Hit the books, before you leap
Compliance teams and regulators were rushing hiring decisions, often without a clear idea of the depth and breadth of the "data science" profession, Bastiman said. Organisations need to ask several questions as they complete recruitment and selection procedures. These include: whether the firm needs someone to carry out analytics to investigate data; whether a more commercial approach will mean hiring someone who provides business advice; or whether a firm needs to develop its own models rather buying them off the shelf. All of these skill sets — and more — come under the broad umbrella of "data science".
"It's a very difficult role to hire for. If you just dedicate some budget and say, 'I need a team of 20 data scientists', you could end up basically burning that budget and having nothing to show for it. So, understanding the return-on-investment (ROI) of what your teams do is absolutely critical," Bastiman said.
Tooling up before toiling on
The latest Cost of Compliance report from Thomson Reuters Regulatory Intelligence found firms were investing heavily in data science and automation as they embraced digital transformation.
"A sign that regulatory technology may be coming into its own is the shift in the time spent tracking regulatory developments — a key area where regtech solutions can be deployed," the report said. "The open chequebook era for compliance, which characterized the years after the financial crisis, is over."
Another common pitfall was to build a team of data scientists without first ensuring that the organisation is "tooled up" and ready to make the most of this expertise, Bastiman said.
"Particularly in the anti-money laundering (AML) space, it depends on the tooling you've got and what you're trying to achieve. Because getting a data science department together is a very expensive thing to do. You have to have things in place first, in terms of making sure that your data is ready and it's secure and it can be analysed," she said.
Organisations have also been caught out hiring people with an academic pedigree when an "industrial data scientist" is what is required. The opposite can also be the case where high-level strategic thinking is needed.
"These decisions need to be firmly directed because, depending on where you're hiring from, if you get academic data scientists they can spend months or years investigating a problem in an academic style. It's a very different skill set, to getting someone to do quick turnarounds, and maybe being less certain, but getting you results faster," Bastiman said.
Conduct and AML regulators have been investing heavily in data science and analytics in recent years to make better use of "intelligence-driven" supervision. Joe Longo, chair at the Australian Securities and Investments Commission (ASIC), said in March that "data science" would be increasingly fundamental to the effectiveness of regulation and oversight.
"Without continued investment in technology and data capability, ASIC runs the risk of its effectiveness being diminished and, at worst, of ASIC becoming irrelevant over time," he told an industry forum.
The UK's Financial Conduct Authority (FCA), meanwhile, has decided to build expertise in-house as it seeks highly specialised talent in a tight market. It has set up a "Data Science Graduate Programme" to train its own data scientists. These recruits are housed within the FCA's Advanced Analytics function, which leads the development and application of data science. The team's objective is to "deliver business value through the use of pioneering techniques."
One of the team's goals is to advance the agency's ability to identify harm. This may involve using machine learning to build classifiers to support decision making, or deploying natural language processing and network analytics to identify suspicious transaction patterns. The FCA has also created "data science units" across the organisation, which work closely with sector experts to analyse risk, triage cases and automate processes.
"This will result in harm being detected more quickly to protect consumers," the FCA said in its Data Strategy Update. Industry officials in London said that remained to be seen.
Keep on top of the latest industry insights and developments in AML
Sign up to our latest industry insights to stay ahead of the crowd when it comes to industry matters, regulatory developments, the financial crime headlines, and our favourite topic – technology. Or follow us on LinkedIn and Twitter.