The most ethically questionable practice to have emerged out of the innovation currently transforming insurance is claims optimisation. As key pieces of accountability and data legislation go live in 2018, what should claims directors weigh up in relation to this controversial practice?
Claims optimisation uses data and analytics to identify claimants that are financially hard up and offers them cash settlements below the true value of their claim. It’s a natural extension of price optimisation, in which quotes are set not by the risk, but by the amount you’d be willing to pay.
How far advanced is its introduction into UK insurance markets? That’s unclear, although I’ve come across signs that it might already be influencing some settlements. I believe it’s on the cusp of moving from trial to implementation. So what should a claims director take into consideration when weighing up such an implementation? Here are some question to be weighed up.
Do policy wordings support the use of claims optimisation? Interestingly enough, my household contents policy does, in respect of almost all types of loss.
Yet is this sufficient? After all, contracts have implied terms relating to performance. As one law professor wrote recently, “We have promises of indemnification; we have narratives of mutual good faith.” Are these being upheld?
If the fundamental narrative of ‘being put back in the position you were in before the loss occurred’ is to be changed, then is it not in the public interest for that narrative to be clearly shared with customers?
And in weighing up that question, think not just of personal lines, but of commercial lines too – will that struggling company be offered ‘a discount for cash’?
Claims Fraud Questions
Insurers want to share more of their claims data (more here) in order to better prevent fraud. In many countries, the market enjoys exemptions from competition law in order to do so. Those exemptions would be at risk if claim settlements become detached from loss.
And if it becomes widely known that cash settlements are on the rise, what sort of signal would that send to fraudsters?
Some consumers think that insurers will often try to cut the cost of their claim. And the sector has at times played into this narrative. I hear talk at conferences of ‘pushing back’ on claims as a normal part of business. Meanwhile, claimants experience complex and mystifying supply chains that put forward settlement offers that are difficult to add up.
For some, this tips over into an attitude of ‘if they’re going to cut my claim, I might as well not hold back in what I claim for.” Opportunistic fraud is wrong, but will the public find settlements derived from claims optimisation any less wrong? That’s why I have described claims optimisation as the Achilles Heel of insurance fraud.
Claims optimisation relies on the use of some clever algorithms, that explore the data an insurer holds about the claimant. Any signs that a claimant might accept less will be picked up and used by the algorithm. Given the gender bias now recognised to be in the training data upon which many algorithms are trained, how fair will an optimised settlement offer be?
For large injury claims, will that algorithm exploit not just the gender pay gap, but gender differences in financial vulnerability and emotional reaction to risk? Remember that you can’t rely on dealing with this through controls over input fields like gender, for algorithms learn patterns from across huge and diverse datasets.
Insurers have actuaries who model claims patterns. They have auditors who monitor how well business risks are controlled. And they have compliance teams who check that regulations are adhered to. How will these people respond to claims optimisation? If claims are settled on vulnerability rather than loss (which is what claims optimisation amounts to), how comfortable will they be to sign their names to the outcomes it generates?
Each of those roles will be in the spotlight of the Senior Managers and Certification Regime from December 2018 onwards. The quality of actuarial work is already on the radar of the Financial Reporting Council as a ‘hotspot’ risk in data driven firms. Will claims optimisation become a development around which questions are asked about professional reputations?
Claims optimisation has been described as having a ‘cake and eat it’ feel to it. Insurers want access to as much of your data as possible, with as many legislative exemptions as to what they do with it as possible. The intention on actuarial grounds is to set ‘accurate’ prices, yet the data is now to be used to detach claims costs from actual loss amounts. How will legislators feel about that?
Trust and Integrity
Claims optimisation feeds directly and powerfully into the issue of public trust, and into the integrity of the people who lead the market. The public want to trust the sector to deliver their part of the insurance deal. Claims optimisation puts a hole in that trust, well below it’s waterline. It raises questions about the integrity of the people who make the important decisions.
It’s often said to me that such practices are necessary because ‘if we didn’t do it, others would and we’d end up losing market share’. And interestingly enough, it’s a narrative that at times I’ve heard the UK regulator sympathise with. Yet I can’t think that sympathy will survive the widespread adoption of claims optimisation. Given that the regulator raised concerns last month about some of the outcomes being generated by big data, “particular in relation to general insurance”, then that sympathy could already be on the slide.
It’s regularly said that insurers, in their embracing of big data and analytics, need to also embrace failure, and to recognise and act quickly on anything that doesn’t look sustainable. Claims optimisation needs to be embraced quickly as a failure in trust and integrity.