Reinsurers are an influential bunch. No start up firm could enter the insurance market without their hefty support. And most existing insurers buys large amounts of reinsurance in a mix of layers and slices to stabilise their accounts. This reliance is a double edged sword though, especially during a time of change. They must both encourage their insurer clients, and be cautious at the same time. So how are reinsurers performing? Are they getting the balance right in how they engage with the digital transformation of insurance? Here are ten ways of looking at how they’ve been getting on.
I have a concern about the internal consistency with which some reinsurers engage with questions of trust in digitising markets. I’ve seen functional chiefs confirm they would absolutely never do ‘practice X’, when I’ve read articles by their analytics team about ‘practice X’ having been a feature of their work for several years. Reinsurers are often big firms, but I don’t think this is a size issue. It’s more an issue of different internal cultures working to varying agendas. Given the influence that reinsurers have, this is a problem.
Accelerator / Brake
Some reinsurers raise questions about data ethics issues, while giving support to quite radical digital practices. This seems at odds with the precautionary mindset with which the reinsurance sector often works. It feels more like pushing the foot down on the accelerator, while gingerly tapping the brake. At best, this comes across as cautionary, at worst as contradictory. I’ve seen this confuse key audiences, who wonder why reinsurers raise important questions, when the responses gathered gain little traction?
Reinsurers have always adopted a serious and scientific mindset to their work, keen to show that it is grounded on careful analysis and review. And that is fine, but then, why does this seem to be abandoned when engaging with some insurtech start up ideas? Why do reinsurers sometimes lend their support to new underwriting, claims or counter fraud practices, when the science underlying those practices is questionable at best, controversial at worst. Of course scientists are human, as are reinsurers, and the shiny new idea always attracts. Are reinsurers being critical enough though about what they choose to support? Have they set their standards too low? Are they undertaking any ethical due diligence?
In my past dealings with the reinsurance market, I was struck by how long a memory some firms had. Their engagement with new technologies has sometimes had near catastrophic results. And so when presenting new risk transfer ideas, there seemed to be a mountain to climb to even get a hearing. No longer, it seems. Now it feels like every new idea needs to be grabbed before someone else gets to it. Yet history has a habit of repeating itself. Some of these present day enthusiasms will turn out to be controversial misjudgements, the ethical impact of which could be profound. We need reinsurers to be better than this.
Technological change is complex. Turning it into a successful business is even more complex. And clearly reinsurers have been able to cope with that complicatedness, given the widespread enthusiasm with which they’ve been embracing insurtech start up firms. So why do they talk so cautiously about data ethics? Why do they talk in public about the complicatedness of data ethics making it a difficult theme to move forward on? I’m not saying data ethics is easy, but in broad terms, it’s no more complicated than some of the data / analytics ideas that reinsurers are already engaging with.
Reinsurers are obviously going to play a crucial role in the insurance response to the emergence of autonomous vehicles on our roads. And the success of their long term involvement will rest upon their underwriting and claims being very carefully considered. What we need to learn more about, is just how critical reinsurers are being in say, their underwriting? To what extent are they interrogating the construction of the underlying algorithms? Is their due diligence picking up on concerns like those covered in this paper on ‘AI, disability and bias’? Third party motor liability is compulsory in many countries. This makes its underwriting (AV or otherwise) very much a matter of interest to the wider public. The reassurance of reinsurers matters.
Reinsurers are, to be honest, just a particular type of insurer, responsible as manufacturers for their products just like any other regulated firm. Yet they exert an interesting influence, leaning the insurance market in this or that direction through the rates and covers they make available. They’re also vociferous consumers of data, and therein lies this accountability question. Given that influence, given that appetite for data, where does their accountability for data ethics lie? If that influence has evolved (as I believe it has) into an even stronger ‘data pull’ influence than back in analogue times, then should their accountability now have a scope that reflects that influence, rather than just their products?
Let’s develop that ‘data pull’ influence a little more. The most fundamental trend in insurance at the moment (personalisation) feeds on data to an enormous and ever growing extent. And the actuarial science that underpins reinsurance relies on insurer clients pulling all available data about policyholders into their models. Yet has their influence, their appetite for data, their actuarial science, made reinsurers somewhat blind to the problems that personalisation could present? There are signs that personalisation is no longer seen as the singular future for the overall market, and that’s great, but it feels rather too much like the proverbial stable door. Being late to this debate, how can reinsurers catch up, or are they too committed now to pull back?
Just loan it
Let’s keep with that commitment to personalisation that many reinsurers have built their future around and develop it a little more. In fewer years than many reinsurers might expect, personalisation will effective turn reinsurers into loan companies. The better prepared will function like credit firms, the less well prepared like a ‘payday loan’ firm. Their product will amount to little more than easing short term cash pressures of front line insurers. After all, what’s the point of reinsurance, when the insurer’s personalisation model is simply going to recoup losses arising from an upturn in claims, by bouncing up the premiums charged to those consumers. Does personalisation herald the end of reinsurance as we know it?
I’m going to end on a challenging note. Is reinsurance taking personalisation seriously? If, as I and others believe, personalisation represents a systemic risk to both the efficacy of insurance markets, and public trust in them, then should reinsurers not consider it as up there with other systemic risks like, yes, climate change? Is all that data, that actuarial science, making reinsurers blind to this? It seems strange, given their renown far sightedness, that only now are one or two reinsurers turning their thoughts to the implications of personalisation. Let’s hope it’s the beginning of something significant.
I am conscious that this is a somewhat broad and high level look at reinsurance, data and ethics. Yet I hope it has given reinsurance people some food for thought. I do see some promising shoots of attention to data ethics amongst reinsurers. I hope these thoughts help their propagation.