The UK insurance sector must prepare for a regulatory revolution. It’s been brewing for some time, but the regulator is now signalling nothing less than a step change in its regulatory approach. Its use of supervisory technologies ( SupTech) is about to move from being a useful tool, to becoming its core way of monitoring for misconduct. The implications for insurers will be profound.
The FCA’s much delayed 2021/22 business plan is a curious document. It read less like a business plan and more like a statement of intent. And it was written very clearly with the Treasury, its political master, as the key audience. The narrative was all about setting expectations for regulated firms and making sure they delivered on them. Substantive mentions of customers seemed few and far between.
And from the first paragraph, the FCA’s new CEO, Nikhil Rathi, took centre stage. This is his statement of intent, delivered with a thump or two of the regulatory table. Now, of course, he has to deliver it. And his key tool for doing so is suffused throughout the report – supervisory technologies, the use of data and analytics for supervisory purposes (more here).
The Three Challenges
Let’s look at how he does this. He starts with three ‘distinct challenges’, the first of which is “taking advantage of data and technology to increase our ability to act decisively in the interests of consumers”. And from this follow the two other distinct challenges. They will be ‘more assertive’, largely because supervisory technologies gives them a better evidence base for doing so. They will be ‘more adaptive’ as products and services evolve because supervisory technologies moves them nearer to real-time supervision.
So the regulator is moving from being behind the market, picking up the pieces after the event, to being alongside the market, addressing issues as they develop. This is the huge step change that the regulator has been preparing for. And the new CEO is nailing his reputational colours firmly to that SupTech mast. That’s clear from the business plan’s narrative structure and tone.
Back in 2015, I wrote a paper for the Chartered Insurance Institute, in which I outlined the emergence of panoptic regulation (now referred to as SupTech). It showed how the future of insurance regulation would be built on two things.
Firstly, the streaming of real time data from regulated firms into a regulatory data loch. This would cover premiums, claims, communications, wordings – you name it. And secondly, this data would then be analysed in real time by various forms of analytics, looking for signs of consumer detriment. This would not just be about what is happening now. It would just as much involve the use of predictive analytics to identify where signs of detriment were likely to emerge.
In the intervening years since the first signs of panoptic regulation, the regulator has been working to build its capabilities and to train and test its models. It hasn’t been doing this alone – other countries’ regulators have been conducting similar exercises and sharing the results.
The Key Partner
An absolutely key partner for the FCA has been the UK’s leading pool of data science expertise, the Alan Turing Institute (ATI). The FCA and the ATI have been working together since 2016, to help refine both the data lochs and the analytics.
A sign of the confidence with which the FCA felt this was preceding came in a statement by the then interim CEO of the FCA, giving evidence to the Treasury Committee in February 2019. Speaking about his organisation’s ability to track discrimination in insurance pricing, Chris Woolard said that the FCA had “the resources and expertise to pick inside those insurance models.”
OK, some of you will be saying, show us some evidence. Well, even back in July 2018, the FCA was talking openly about using:
- graph learning to identify networks of potentially collusive behaviour;
- supervised learning and random forest techniques to predict the probability and location of an adviser mis-selling financial products;
- visual analytics to identify potentially misleading advertisements.
Note that second point. This is the use of SupTech to predict the probability and location of mis-selling. It will be in cases like these that Rikhil Rathi’s talk about ‘assertive and adaptive’ will be measured.
An Undimming Spotlight
So what is so profound about SupTech? It places insurers under an undimming regulatory spotlight. As their decisions generate data, so that data will be collected. That data generates outcomes, which will be analysed in respect of the firm and the person whose decisions triggered them. As those outcomes drift outside of FCA benchmarks, so a human supervisor will be alerted and a call put into the insurer for a meeting and explanation. And the time delay will be measured now in weeks, perhaps days, not years. No need to wait for complaints or speaking up reports to come in.
It will feel like the regulator is in the room where you’re making those decisions, watching who is making them and weighing up how fair, how free of conflicts of interest, how honest those decisions are. And in turn, that will influence the culture framing those decisions. Senior insurance people, knowing that outputs from their decisions were being monitored in near or real time, will orientate their decisions more and more towards the ethical. That is what makes panoptic regulation so powerful.
When I’ve discussed this with clients, a common reaction has been that this feels unethical. Yet those same firms are often doing something very similar with clients. In other words, collecting their data, analysing it and using it to make decisions. So what the regulator is doing feels pretty similar. And remember: this is not new regulation; this is new capabilities to apply existing regulations.
Until recently, I’d been a bit puzzled by how the FCA would be interpreting the data coming in from an insurer. How would they gauge whether that data was signalling something good, middling or worrying. I’m less puzzled now, for recently articles by ATI people have shown that their support has moved beyond advisory and encompassed particular data projects. What is coming out of those projects will provide the calibration tools that the FCA need.
Insurance is in the middle of some profound changes. Many insurers have become so focussed on this that they’ve not noticed how the regulator has been preparing its own profound changes too. Those insurers who have picked up on this have been reflecting on what this means for how they work. They’ve realised that engaging with InsurTech means engaging with SupTech too. And this has then influenced their strategies for digital and customers, and their work on purpose as well.
What SupTech does is put a mirror before the decision maker, not unlike a common technique for weighing up how ethical a decision might be. Called the front page test, it asks you to consider whether you would be prepared to see your decision feature on the front page of a leading business paper? If not, then why are you making it? In SupTech terms, would you be prepared to see your decision pop up on the regulator’s digital dashboard? If not, then why are you making it?
This changes the narrative from ‘can I’ to ‘should I?’ Might this hinder innovation though? Not really. It simply means the innovation will be viewed through a more reflective lens. And that can only be good for long lasting, beneficial change in a market whose success at social integration was always going to put it under careful scrutiny.