A landmark decision by US insurance regulators could translate into significant disruption for the UK personal lines market if the UK regulator adopts a similar position. It’s certain to be one of the most awaited outputs from the FCA’s market study into the use of big data in general insurance.
A few months ago, in this post, I outlined how the National Association of Insurance Commissioners was consulting on whether to recommend to its members (the individual US state insurance regulators) that they ban the use of price optimisation in personal lines pricing. And the outcome of that consultation was a White Paper that did recommend that price optimisation be banned in the personal lines market. And it went even further.
The NAIC’s White Paper identified four types of price optimisation that they felt failed a key regulatory test, that ‘rates should not be … unfairly discriminatory”. And the four types were:
So far, 18 state regulators have banned price optimisation, most noticeably California and Pennsylvania. The number rose steadily during 2015 and is expected to rise further during 2016.
The scope of these four types of price optimisation is telling, covering not only classic economic levers such as price elasticity, but more prosaic kinds such as asking questions or making a complaint. Either which way, a large number of US insurers will now have to unroll their price optimisation software from sizable chunks of their personal lines portfolios.
Some believe that the ban represents a hurdle that some insurers will look for ways to work around. And one of the problems for US state insurance regulators will be to find evidence of price optimisation within their traditional rate filing arrangements. Those arrangements are unable to handle the sheer volume of data that insurers’ pricing models now throw up. And it is around this very point that another of the NAIC’s recommendations stands out as highly significant.
In Appendix D of the NAIC’s price optimisation white paper (on page 22) is a set of questions that state regulators can direct at insurers to ascertain exactly how their price models work. Some of the particularly interesting questions focus to the use of input variables sourced from external sources, such as:
In essence, the NAIC wants insurers to open up their pricing models so that state regulators can analyse them for sign of any rates that might be unfairly discriminatory. It represents a key move towards what I’ve called ‘panoptic regulation’: the regulator who plugs ‘conduct algorithms’ into insurer pricing models, with the ability for real-time regulation. It’s been described to me as moving insurance regulation from being behind the market, to being alongside the market.
For UK insurers, the $64,000 question is of course: will the FCA follow suite? That they will address price optimisation in their big data market study seems certain: failing to do so would sink their study’s credibility below the waterline, especially with Select Committee Members of Parliament with constituents struggling to pay continually rising premium.
So how might the FCA respond? I would be surprised if some form of brake was not applied to the use of price optimisation in the UK personal lines market. Whether that brake is an emergency stop one, or a light foot on the pedal to slow down its use, is difficult to gauge. If I was forced to come off the fence, I would expect a ban along the lines of the four types set out above from the NAIC’s white paper. Clearly, measures relating to asking questions or complaining are a no-brainer – they’ll be banned and any insurer found to have used them will have their regulatory reputation severely tainted. And if you take what the FCA’s behavioural economists concluded on pay day loan pricing, then price elasticity will go too.
The FCA’s payday loan market study (and speeches by their then CEO) signalled the emergence of panoptic regulation. Their big data market study will signal its emerging structure.