The EIOPA consultation is in response to a growing number of regulators looking at how non-life retail business is being priced. Ireland, the Netherlands, Sweden, Italy and of course the UK have each looked into differential pricing and found problems. The oldest by far is the NAIC in the US, which in 2015 identified a series of pricing techniques which it suggested state regulators treat as unfairly discriminatory (more here).
Insurers may think that ‘they have seen this before’ and ask ‘what is likely to change as a result?’ Here in the UK in particular, insurers are busy implementing the price walking ban and so will be tempted to think that ‘this is already something we are doing’.
One for UK insurers to Watch
Well, I would advise UK insurers to not be too casual and dismiss this EIOPA consultation. Yes, there’s likely to be overlap between what will emerge out of EIOPA’s current consultation and what UK insurers are already having to address. Yet, as I described back in 2020 (more here), the UK price walking ban failed to address the full range of price optimisation practices. It addressed a particular issue – walking - rather than pricing practices per se. Why? Because that was what the super-complaint had been based around.
If EIOPA is prepared to really get into the detail of present day digital pricing practices, then it should find the whole range of pricing practices that drive issues around fairness and discrimination. And if that were to happen, then UK insurers should watch for how the different practices end up being treated, for it is possible that the FCA might then replicate them in some form or another.
Surely the FCA has enough on its plate, some of you will say. That may be the case, but the FCA did say in their final report on price walking that they would be monitoring the use of price optimisation techniques. So the EIOPA consultation could provide the FCA with indicators for the scope and depth of what might be called the next round of regulatory scrutiny of pricing practices.
What Pricing Practices?
What ‘wider pricing practices’ are we talking about here then? The NAIC back in 2015 suggested that pricing practices such as price elasticity of demand, propensity to shop around and propensity to ask questions or file complaints, be deemed ‘unfairly discriminatory’. That’s a key term in US insurance regulation.
To what extent are such practices common in UK personal lines? Well, back in 2014, I wrote that “around half of UK motor insurance is said to be priced according to some form of price optimisation”. Eight years on, that percentage is likely to be bigger.
And more recently, I looked at how UK insurers were tracking claims complaints. A Bank of England / FCA report had looked at how insurers were using predictive analytics to target claims that had a high likelihood of customer dissatisfaction or complaint. The reason given was so that they could be flagged in order for a human to then monitor the claim and intervene if required. However, back in 2014, the FCA had found complaints tracking was being used to offer lower settlements. My point therefore is that the linking of price and complaints within the UK market is just as easy as the linking of settlements and complaints.
Let’s look now at the detail of the EIOPA consultation. What should UK insurers be noting about it? I’m going to look at eight features that stand out.
What is Differential Pricing?
While every other regulator to-date has been happy to describe these pricing practices as price optimisation, EIOPA has gone for ‘differential pricing’ instead. They felt the former sounded too optimistic and the latter more neutral. I think they’re wrong on this. Just because one form of a word denotes a positive outlook, doesn’t mean that another form of the word that denotes an act or process can’t then be used.
This is not nit-picking. Price optimisation has been widely used over the years, by a variety of organisations. It is the established term to describe these practices and EIOPA’s lack of action on those practices over the years (remember the NAIC was back in 2015) isn’t a credible base upon which to adopt new terminology. What’s more, their preferred term, ‘differential pricing’, is so neutral that it will be interpreted along a sliding scale of existing practices from acceptable to unacceptable. A good many underwriters will see ‘differential pricing’ as ‘business as usual’, on the basis that that’s what they’re already doing. What EIOPA saw as an improvement is more like a step back.
Cost of Service
The NAIC in the US and the FCA in the UK (see above) have both found evidence that insurers will price optimise on the basis of cost of service. Back in 2015, the NAIC gave two examples of this: the propensity to ask questions and the propensity to file a complaint. In other words, ask for clarification of something in your policy, or complain about the service you’ve received, and your next renewal premium will be higher as a result.
EIOPA on the other hand seem to be lumping cost of service in with risk. In other words, the benchmark norm for establishing if differential pricing was taking place comprised both risk and cost of service. This may not sound like much of an issue, but to me, it points to EIOPA taking a less focussed approach to price optimisation. I’m concerned about that being replicated in how EIOPA handle responses to the consultation.
Surely, some of you will ask, costs of service are a normal component of an overall risk price – underwriting doesn’t come free! And some business is more expensive than others to service – it’s only fair that this is reflected in the price. And I get both of these things, but at the same time note that the issue about optimising costs of service is that people who want to understand what is covered, and people who are not happy with a settlement, shouldn’t be charged more.
Across a reasonable sized portfolio, there will be a range of enquiries and complaints, and to be fair, these have to be absorbed into the overall cost of service. That’s what EIOPA are thinking, but that is not how the market is moving.
Risk and Character
A key question that needs to be addressed when weighing up price optimisation is what is meant by a risk factor. Insurers see aspects of character as now central to assessing the risk being presented. Meanwhile, consumers see insurers’ data gathering practices in relation to character as intrusive, partial and likely to result in bias.
EIOPA haven’t done enough to nail down this question and bring clarity to what exactly they’re looking to address. If they signal to the market that risk related pricing practices are fine, then they will in affect be handing regulation of pricing over to the market (blunt but true). This would be like putting consumers right back where they started, for example, in 2015 in the US and 2018 in the UK.
There is a debate emerging around what exactly is a risk factor and what is just opportunistic data gathering by insurers. Insurers’ approach is to see any data that might (via an AI model) be found to be related to risk in some way as a risk factor. This then justifies collecting all data in order to check on that. However, this obviously raises issues around privacy, autonomy, fairness and bias. EIOPA needs to address this question of ‘what is a risk factor’ in order to make their consultation meaningful.
Predictions
A lot of the talk about price optimisation is framed in terms of the present. And the EIOPA report follows much the same line. Yet the real value that insurers are seeking from price optimising techniques lies in their capacity to predict at least the near future, perhaps even beyond that.
Focussing on the present creates a real danger of those predictive techniques being at best down played, at worst ignored. EIOPA should have explicitly brought predictive analytics into their consultation, for it is where I believe most insurers are focussing their price optimisation work at the moment.
Vulnerability
All of us will at some point experience life events the pattern of which will be picked up in insurers’ data gathering, to then trigger some optimisation of their next premium. Some of those life events will be uplifting (a new job, run performance or marriage, for example), while some will be the exact opposite (a loss of job, health problem or divorce, for example). Price optimisation looks for both types of events and weighs them up as pricing opportunities.
This was behind Andrew Tyrie’s comments in his 2019 speech to the Social Market Foundation about digital technologies exploiting vulnerabilities that many, if not all, of us will experience at some point in time. His description of us as ‘all vulnerable now’ was not meant to downplay the life experience of people more traditionally classified as vulnerable in some way. Instead, he was emphasising that vulnerability is something that doesn’t happen, for many of us, to someone else. We will all experience it in some form and to some scale, and so we should all take an interest in it. It is not someone else’s problem.
EIOPA doesn’t seem to have grasped this change in how price optimisation can exploit both macro vulnerabilities that some experience, and micro-vulnerabilities that a much wider pool of people will experience. It’s another example of how their consultation scope will undermine their findings.
Life or Non-Life
Another scoping issue with EIOPA’s consultation is its focus on non-life business. Sure, it’s much clearer to understand (and so address) price optimisation in the non-life market. Non-life business is short term with pricing that can vary. Life business is long term, making price optimisation less of an issue. Isn’t it?
No - that’s no longer the case. Some parts of the life market have been experimenting with premium reviews for a while, using wellness programmes centred around devices and rewards to leverage premium change. Price optimisation is central to the value that insurers want to gain from this. Combine this with the sensitive nature of the data upon which that optimisation is based and EIOPA’s focus on non-life business becomes quite an omission.
Governance and Control
The EIOPA consultation takes quite a long look at ‘supervisory expectations’. They raise questions around how to use various forms of product governance and around supervisory approaches. It looks like fairly standard stuff, and therein lies the problem. Existing product governance within European insurers has achieved little to-date in terms of questioning and controlling the use of certain price optimisation techniques.
Might this be down to old regulations needing to be updated for modern digital practices? I don’t think so. Fairness, bias and autonomy are not new issues. And there has been enough meat on existing regulations for a questioning regulator (like the NAIC on behalf of state regulators) to have taken action before now. As that hasn’t happened, I believe that EIOPA’s obvious preference for controlling price optimisation through product governance to be wishful thinking. Of course responsibilities will be confirmed, policy promulgated and reports produced, but unless the obvious problems with the three lines of defence are fixed (more here), governance alone will fail to keep the use of price optimisation within acceptable bounds.
In my opinion, EIOPA is making the same mistake as regulators like the FCA, by looking to insurers for confirmation of compliance and monitoring on a macro basis. Instead, they should be looking to consumers for confirmation of outcomes, and monitoring on both a micro and macro basis.
Never Forget Claims
It’s true that optimisation techniques have arisen predominantly around pricing, but recently years have seen a growing use of such techniques in claims. As I mentioned earlier, the FCA have found ‘propensity to complain’ influencing settlement offers. What I’ve heard from within the market is the view that any push back on the optimisation of price will just cause more of its use in claims. This is claim’s ‘willingness to accept’ version of pricing’s ‘willingness to pay’.
I can almost hear EIOPA people saying ‘hey, we’re got to put some limits on our consultation’ and I understand that. At the same time, more attention to this earlier (such as soon after the NAIC in 2015) would have allowed European regulators to address developments like claims optimisation much earlier. Optimisation is taking place across the life cycle of insurance, and EIOPA needs to acknowledge this, even if their consultation then just focusses on pricing.
What to Watch For
Insurers in the UK and EU may be tempted to respond to this consultation within the scope that EIOPA has given it. It would be convenient but relatively short-sighted. Regulators have invariably been trailing public sentiment about price techniques used by insurers and this consultation is no different.
To secure the long term trust of consumers, insurers need to review their optimisation practices on wider terms. In particular, I recommend doing so in terms of the eight aspects I’ve outlined above, namely…
- keep thinking in terms of price optimisation, in order to recognise the step change in opportunity such practices present;
- make sure cost of service as recognised as a price optimisation issue;
- make sure the risk from character is seen as a price optimisation issue;
- see optimisation as encompassing both present and future prices;
- adopt a digital rather than analogue approach to vulnerability;
- see price optimisation as an issue for their entire retail portfolio, not just the non-life part;
- look beyond product governance for delivering accountability for the pricing practices being used.
- include claims, for the greater risk to digital strategies is there.
Sounds like a lot of effort, some of you will say. And indeed, it may well be so. To understand how much effort, look to US insurers, dealing with state regulators long versed in the issues associated with optimisation. Some of those states are very against optimisation, while some are more relaxed about it. They’ve experienced the impact of reining in their optimisation practices in places like California, a market bigger than the UK or any EU country.
My point is that the cost of not making the effort to understand the impact of optimisation practices could be the greater of the two. Just as ‘price walking and the super-complaint’ was just a matter of time, so I believe is the push back on certain uses of optimisation. And to be honest, it will be more impactful.
