I’ve been asked to write a piece about insurance, with particular emphasis on some aspect or other to do with the recent floods. I can’t say too much about it, except that it’s going to be nice and light and amiable. So I’d better not say what the recent floods really make me think.- which is that one way or another, this looks like another biggish nail in the whole coffin of the concept of insurance. In the unlikely event that concepts can have coffins.
Insurance, it seems to me, thrived during a long period in which we understood that everyone faced risks that would be cripplingly expensive, but we had no way of quantifying how expensive each person’s risk wold actually be. From this came the concept of pooling risk: if a hundred people faced a 100 to 1 chance of a mishap that would cost them £100 to remedy, then everyone would be smart to pay a premium of £1 to save a potential £99.
Up to a point, insurance can cope with the availability of more accurate risk profiling data. Say we can establish that of those 100 people, 50 face a risk that’s twice as great as the others. No great problem in charging the riskier half £1.33, and the less risky half 66p.
But take that line of thought further, and eventually you end up in a place where the whole concept of insurance comes into question. This happens when, at the extreme, we can accurately measure the risk faced by each of the 100 individuals. If we know that there’s one who faces a 99% probability of whatever it is, and two who face a 98% probability, and so on, how do we set our premiums then?
The answer is that we have only two alternatives: either we ignore the information, or we can’t offer insurance.
At the moment, we’re in a curious limbo between these two options. Usually under instruction from politicians, regulators and others, we ignore some of the most important information – for example, life insurers don’t use the results of genetic testing; following a recent European Court ruling no insurer takes account of gender; and – coming back to those floods – there is for the time being an agreement that homeowners living in areas where the flood risk is low will subsidise those in areas where it’s high.
But elsewhere, insurers do use much other very detailed information – motor insurers, for example, use literally dozens of pieces of data in their underwriting processes.
Going forward, left to their own devices, underwriters would only go in one direction with this. Ironically, though, it’s a direction that would eventually bring about the end of their industry – not exactly a helpful outcome for them, or indeed for their customers.
Hmm. I think I’ll need a different angle for that nice, light, amiable piece.