This is the second FSI report to examine what’s happening around SupTech, the technologies used by supervisors to support their regulatory objectives (more here). The first FSI report highlighted some interesting case studies, such as the FCA’s experimentation with the use of supervised learning and ‘random forest’ techniques to predict the probability of an adviser mis-selling financial products.
There was a sense of excitement about that first FSI report. That can’t be said about this second FSI report, which leaves the impression that SupTech is hard work, raises challenges and will take time to deliver noticeable results. That makes it a more realistic appraisal of how things stand in SupTech.
Before looking at those challenges, it’s worth remembering that even a patchy application of SupTech tools could impact what an insurer is doing quite significantly. For example, initiatives by the French and Korean regulators to analyse what’s being said in telephone based marketing, sales and advice will make the detection of mis-selling much easier. That would then expose those insurers to distribution restrictions, as happened in Korea in 2012.
Another example is the work being done on the analysis of key information documents and online advertisements. The impact here will be the automated blocking of such material, as well as an evidence trail for customer redress and automated compensation. Why wait for the customer to complain, when the mis-selling has already been established?
How Things Stand
If you look at SupTech in the context of a typical hype cycle, I think it is now beyond the ‘peak of inflated expectations’ and battling with the ‘trough of disillusionment’. Soon it will be moving up the ‘slope of enlightenment’ towards the ‘plateau of productivity’.
What’s being battled with at the moment are things like…
- the lack of expert resource in the combined fields of data science and regulation;
- adapting the culture of supervision to make better use of SupTech;
- dealing with the evolving nature of insurance regulation;
- creating better links between SupTech findings and the regulatory follow-up it prompts.
These are not insurmountable, especially given how one national regulator can share what they’ve learnt from a certain type of SupTech approach with other nations’ regulators. That’s not something that happens with the application of technologies in private firms.
There are two concerns I have with how the SupTech movement appears to be moving forward.
Firstly, the focus seems to be on making supervisors’ lives easier. On one level, this is hardly surprising, but actually, the problems that really need addressing are the ones that consumers have. There will be overlap of course, but I believe the relative positioning should be 'consumer concerns' to the fore.
Consulting with supervisory staff is seen in the report as a bottom up approach, when it is in fact a top down one. A truly bottom up approach would involve engaging with consumers and learning what they would like to see such technologies address.
And there lies the second concern. The report is often orientated around the technologies that are being applied, when it should be orientated more around the risks that consumers are facing the most. No where in the report are fairness, conflicts of interest or discrimination mentioned, which are where most consumer concern lies.
Insurers have to watch out for two developments. The first is when the present regulatory drive to accumulate data starts to plateau. The second is when the connect between the SupTech and the regulatory expertise begins to work smoothly. Those two events will point to the application of supervisory technologies moving up a gear or two. And that will herald its use becoming much more productive for regulators, which could be much more worrying for some insurers.