Among the acres of retrospective pieces on “10 years since the Bank Crash” exploring regulatory change in financial markets, one simple fact is usually overlooked: For the average person, the crash brought not just perplexity but anger. There has been plenty of coverage of how baffled we all were over the complex derivatives contracts underlying the crash. But there has been far less focus on the emotional story: how best to understand and resolve the anger?
The new ‘behavioural regulators’ (notably the UK’s Financial Conduct Authority) are, on the evidence, well aware that anger is a significant human-factor risk. I’d argue that unresolved public anger has been a major factor shaping the new regulation – if not publicly acknowledged as such. The many regulatees that I work with remain seemingly unaware, or perhaps wilfully blind, to its significance.
That’s about to change. New approaches to research are now bringing that anger – and its behavioural drivers – more sharply into focus.
First, let’s remind ourselves of the broad impulse behind the new conduct regulation (or “behavioural regulatory enterprise”, as we political science geeks like to call it). For starters, think about this: Where does all regulation start from? Here’s where: It is always born from the event of failure. It’s a reflex. A response. Often framed in the heat of the moment, well-intended but ill-considered.
Why did we end up with a “Conduct” model of financial regulation?
Cast your mind back to 2008, to the catastrophe that governments then faced. At that point, crisis-hit administrations had to answer three pressing needs. First: to recover everyone’s trust in regulation – a tough call, when your regulators have just presided over a global crash. Second, as one minister from the fallen UK Government famously put it in his handover note: “There is no money left”. Having bailed out banks and introduced austerity, national treasuries could barely afford a new regulatory agency. Third, to manage public anger; ideally by finding someone to blame.
Introducing conduct regulation solved problems one and two. Even better, epic misconduct fines helped reduce public debt. Conduct regulators had voter appeal: they were cheaper to equip, and used “nudges” and other pop-psychology kit. Yet they initially failed to tackle that third problem: the need to blame, name and shame the banking leaders that the public saw as responsible. Without a lightning-rod to ground public anger there could not be the “emotional closure” necessary to allow the public and the financial industry to move on.
The need to fix this third problem is what now drives the UK regulator’s Senior Managers Regime (SMR). Banks must name the senior people responsible for risk – a popular (even populist?) approach – except that, for the SMR to work properly, it must prove that each SM personally intended or overlooked wrongdoing. That’s a tough ask. Behaviour is hard to measure. Its subtler yet most revealing nuances – like intentionality or propensity to rule-gaming – are really hard to capture.
Under the old regime, enforcers could only go after badly-designed or mis-sold products. Quite obviously, a financial product contract isn’t capable of actively expressing intent; it’s a passive, inanimate object. By contrast, Conduct regulators suddenly needed a whole new set of risk indicators for personal behaviour, including the intent of senior managers (and staff) to abuse customers and markets. Hence SMR, with its Culture Audits.
At conduct conferences around the world, I hear this: “The MI we have isn’t the MI we now need”. What everyone actually needs is a new analytic to identify intent. But how?
For four years from 2013-17, the new conduct regulators seeded the ground, reposting content from the latest psychological research journals. Now, technology breakthroughs are offering a better answer to the intent question. Advanced scientific analyses of data are making it possible to identify the hidden drivers of behaviour.
Can machines really understand intent?
We should of course cross-examine any new technique: How validly does the new “regtech” measure soft behavioural factors? How can data find anything as abstract as human intent, i.e. who was the “guiding mind” behind market abuse or a mis-selling enterprise? Impressively, the latest regtech techniques do exactly that – and, as a language analyst, I’m fascinated to see that the human communication is at the heart of these new insights.
The new analytics state-of-the-art is artificial intelligence that’s trained to identify layers of meaning (literal and inferred) within any form of conversation that your staff are having. Rather than just reading the words that people say, the AI learns language patterns that help it to discern underlying meanings.
This is big. Until recently, machines weren’t very good at understanding anything more than the surface layer of meaning: they used to blow up when asked to parse the difference between “time flies like an arrow” and “fruit flies like a banana”. Now, in a leap ahead, AI highlights clusters of meanings that are tells for various misbehaviours such as market manipulation, undue secrecy, or coercion. Banks lead the way in applying this technology, with the larger firms increasingly using AI-enabled language analytics as standard practice.
I’m fascinated by the possibilities this technology brings. So much so, that I’m researching it and will share what I’ve found here in further blog posts. For now, peg this thought: In a world where we have AI that’s clever enough to drive a car, how great will it be to put that level of problem-solving power to work to help stamp out abuse of markets or mistreatment of customers?
More to come…
Next blog, I’ll be taking the back off this magic box. Then, seeing how firms should use it to get ahead of the secret, “insider” games that a few of their staff play at clients’ (and employers’) expense. Do stay tuned.
Explore more expert articles about the changing world of conduct risk:
From Compliance Culture to RegTech Strategy: The Scenic Route
E-Comms Surveillance and the Importance of Intent: “It’s the thought that counts!”