If the freedom of speech is taken away then dumb and silent we may be led, like sheep to the slaughter.

- George Washington

Sunday, 6 March 2011

Insurance and Gender

Following the ECJ's decision last week, that insurers cannot charge people different premiums based on their gender, there has been a general uproar to say that decision was 'absurd' and 'equality legislation gorn mad' (blogosphere passim). I beg to disagree.

Take car insurance. Insurers have, traditionally, quoted higher premiums for male drivers on the basis that men have more accidents than women, and that men's accidents tend to be more serious and therefore more costly to the company. Men (and especially young men) are expected to take this on the chin and cough up the extra, somehow shouldering the blame of all their similarly-testicled colleagues, no matter what their own driving skills or accident propensity.

Now it is correct that, generally speaking, men are more likely to be a poor risk, and generally speaking, women are more likely to be a better risk. But how does that reflect on an individual driver, just passed his or her test, who has no accident record, poor or blameless, to fall back on? I am male and have been paying proportionately more than my share for about 40 years now, and yet I haven't cost any insurance company a penny in payouts. (I suppose I should allow a small amount for admin costs in dealing with an accident which was not my fault, and for which I successfully claimed all my out-of-pocket expenses.) A girl I work with has written off two cars in as many years.

Predictions based on generalities and populations are not the same as predictions based on observation of actual behaviour. A young male driver may be accident-free for the whole of his life, and yet he will pay more over his lifetime than his sister, who may be as accident-prone as Mr Bean. What 'most men' do is irrelevant. He is being penalised for something over which he has no control, and which has no causal relationship with his own behaviour: the behaviour of other males.

We say that 'men can run faster than women' and, for the generality of the population, averaged out, this is true. However, this does not mean that some women cannot run faster than some men. Paula Radcliffe and my good self would be a prime example. Despite the ownership of X and Y chromosomes and two testicles, I could not beat Paula in any race, whether a hundred-yard dash or a marathon. And yet, looked at from an actuary's viewpoint, I should stroll it at any distance because I am packin' a pair.

Imagine if there were to be a race between a thousand people, 500 men and 500 women. And imagine that, because men were 'generally' faster runners than women, all the men were given the handicap of waiting ten seconds after the starting gun, to give the 'slower' female 500 a fair crack of the whip. Sure, some men would reach the tape at the same time as the fastest women, but many men would be lagging far behind, and many women would be up with the leaders. It would be ridiculous to handicap an entire gender in a mixed sporting event. So why is it 'logical' to penalise someone on the basis of their reproductive organs in something like car insurance?

I know insurance companies aren't a charity, and they must relate their premiums to the likelihood of paying out. But why not base that on the individual, rather than the entire sex of which that individual is a part? There exists already, and widely accepted, a system of premium loading and discounting according to claims record. We already discount premiums for drivers who don't make claims and keep a clean sheet, and we already load premiums for drivers who can't stop wrecking their cars, or who drive while drunk, or who break the law and get points on their licences.

I would suggest that the insurance industry takes this one on the chin, but use discounts and loading much more intelligently. Let all new drivers, male and female, start at the same level - a basic premium based on the likely repair costs of the car. A ten-year-old Corsa would cost less than a new Audi, but that would be the only relevant factor. The basic premium would need to be fairly high at first, as the driver would be an unknown risk, but could come down significantly with every year that the driver was on the road without an accident. By 'significantly', I mean by far more than the current maximum of 60% no-claims 'bonus'. After all, if you can drive for three years without a bump, you must be doing something right. Conversely, any claim on the insurance would be evidence that the driver was a poor risk, and premiums could be raised accordingly. This loading would need to be much more brutal than at present if it were to replace the gender-based risk calculations and be cost-neutral for the insurance companies.

So John passes his test and buys a car. For the first year, his insurance costs him (round figures) £1000. But a year later it is only £600, and two years after that £300. He is a safe driver - proved - and his insurance is costed accordingly. Jane starts with the same premium of £1000, but six months later has a minor accident. At renewal time, she finds her premium is now £2000. A year later, she writes her car off and soon afterwards is caught driving over the limit. Her next insurance premium is £6000, and she decides to give up and catch the bus. Actual driving behaviour, good and bad, would have real and significant consequences - nothing to do with gender, and everything to do with real-world outcomes. That would seem to me to be a much fairer way of going about things. And it could well have beneficial effects on the accident stats.

The ECJ's decision seems to me to be the right one from the point of view of simple fairness. A driver should pay premiums according to how he or she actually behaves, not how they may be expected to behave on the basis of their gender. That is in line with principles of individual responsibility and matching actions with consequences, which is good for the moral fabric of society.

These are some initial thoughts, and there's probably something major that I have missed. Your thoughts are, as always, welcome.

Of course, we know that all the inscos will do is to rack up female premiums to match male premiums, double their profits, and hope no-one will notice. But that's the way the world works.



  1. I hadn't seen insurance costs, in all honesty, as much of an issue except for youngsters. Certainly not one of what I might consider discrimination, given that the sex-based model was based on observed results rather than any particular bias. At worst, a convenient, if somewhat lazy, generalisation which tends to even out with age anyway.

    Having thought about since the ruling though, I do tend to agree with the counter-argument you present here. Insurance is a personal transaction and hence should only really be affected by traits you personally exhibit, not ones tagged on by broad-based associations.

    The difficulty is, of course, how to measure those traits with any degree of accuracy. If there is a known discrepancy in the pooled risk of any given combination of say, gender and age, it can only be compensated for by a) finding other ways of evaluating that discrepancy without mentioning gender and age or b) not evaluating it and jacking up everyone's premiums to the worst-case level.

    a) isn't so easy. Chances are it would be more intrusive and require a more expensive administration process. Or a different methodology: accidents per mile covered would be better than accidents per year as a risk indicator, for example, but I don't much fancy any of the Orwellian baggage that comes with keeping track of people's road miles (and average speed...). Or the implementation cost.

    b) is, I would have thought, the more likely. Leave gender out, level down the starting male premium a little and raise the starting female premium to meet it. Then adjust via a more comprehensive no-claims model, probably including a claims-penalty element as you suggest.

    I have to disagree on your base cost, though. Insurance is not mandatory in law because it benefits customers to have their personal risk covered: it's mandatory in order to address third party - ie public - liability. In that respect, the cost or value of the car is irrelevant: what matters is the amount of damage it can cause to others. In part, that's baselined, in part it's factored into the vehicle group (along with elements of cost of replacement). Hence you also see anomalous groupings, where particularly theftworthy vehicles rub shoulders with particularly powerful, expensive or hard to control vehicles at the far end of the group spectrum.

    I can imagine more of that happening in the future, with older models actually attracting a higher grouping to reflect the attrition rates of their typical owners. Boy racer hot hatches, like a 10-year-old Corsa, for example...!

  2. Your comment re base cost - agreed. I was trying to make a general point without getting snowed under with ifs and buts. The repair cost is a relatively trivial part of the risk calculation.

    It's not a simple matter, though. There you go at the end, talking about hot hatch 'typical owners'! And don't get me started on life insurance ... and pensions ...

  3. I can see your point, but the wrong thing about this ruling (actually, not the ruling itself but the law that makes it possible) is that it is another intrusion of the state in something that is not its business.

    Are you sure you are a libertarian? :)

  4. Actually, no! I'm pretty resistant to labels of any kind, even Libertarian.

    You could argue that the ECJ is simply correcting an injustice, which is that people are being charged premiums based on other people's behaviour. It's not a free market, after all - you don't have the choice of whether to take out car insurance or not, so you are at the mercy of the insurers' ideas of risk, which may not be fair or reasonable but merely profitable. If there were a free market in insurance, with the buyer able to choose between competing models, then you would be right, but there isn't. From a Libertarian poi t of view, I would have thought thr ECJ judgement is a step in the right direction, at least under the current arrangements.

  5. Just my native cynicism, there. I expect the insurance companies to try and find the easiest way of penalising potential boy racers without actually mentioning the "boy" part, so I'm guessing sportier cars at the cheaper end of the market may see a disproportionate rise in loading...

  6. A lot of the girls I work with ain't gonna like that.

  7. I'd quarrel with a lot of your practical details, but, (to my surprise), I now agree with your basic point. I didn't though, before I read it. For an opinionated guy like me, it's always a surprise when someone changes my mind. Ouch, it hurts a bit.

  8. Zaphod, thanks for that. I feel quite honoured.


Comment is free, according to C P Scott, so go for it. Word verification is turned off for the time being. Play nicely.

Related Posts Plugin for WordPress, Blogger...