35

My coworker told me women pay less car insurance than men. So I did some research and found some sites proving that.

For example: According to How Age And Gender Affect Car Insurance Rates, women tend to pay less for car insurance than men. How can a person pay less based solely on their gender? How can this be legal? Is this only in the US?

2
  • Comments are not for extended discussion; this conversation has been moved to chat. Comments are for improving the question.
    – Oddthinking
    Commented Apr 21, 2022 at 0:26
  • 2
    What country are you enquiring about?
    – CSM
    Commented Apr 21, 2022 at 17:03

4 Answers 4

28

Yes, in general women do get cheaper car insurance.

Some jurisdictions have banned this - several states in the US prohibit charging women more. Here is the EU guidance from 2012 where gender discrimination was banned in car insurance:

This ‘unisex' or gender-neutral pricing means men and women with the same characteristics (e.g. age, state of health depending on the product) should pay the same price for the same product. Pricing will have to be based on other risk factors than sex, such as driving behaviour in the case of car insurance. This means people will no longer have to pay more, or less, simply because of their gender.

Insurance companies used gender to set premiums because, in aggregate, women drive more safely than men. Insurers have more data and more complex models now than they did in the past, and many have chosen to use more specific predictors of driving safety for each individual to set premiums.

Does this mean that there's no longer a difference in how much men and women pay for insurance? No, the use of models with more predictors has actually made that gap wider.

What appears to be at work is that car insurance companies set a price very much according to all the other data they can find on you – without actually asking your gender. So the quote you get back reflects the risks attached to your occupation, how much you drive, the sort of car you drive and whether you have made any modifications to the car.

I asked Confused.com to explain the worsening premiums for men, despite gender equality. It said: “The continuing disparity between men and women could be linked to the fact that certain male-dominated occupations may have a poorer claims experience.

“Also, on average, men may tend to drive larger and more costly vehicles. The more expensive/high-spec the vehicle, the more likely it is that the cost of repairs will be higher, and therefore this is reflected in the premium charged.”

This is an expected outcome. It was easy to use gender as a model predictor because it was covariate with the specific driving habits that made insurance riskier. Insurers now measure those driving habits directly, but that doesn't change the fact that they're still covariate with gender, and that there will be differences in the average between car insurance premiums for men and women. Even in jurisdictions that haven't prohibited gender discrimination, it's mostly explicitly gone, as the insurers who haven't upgraded to more modern actuarial models have a competitive disadvantage compared to the insurers who have.

2
  • 10
    Excellent answer. I like that it shows that women still “get cheaper car insurance” even in legal jurisdictions where insurers are not allowed to use gender as a factor in insurance pricing. And this happens not because insurers are breaking the law, but because they have found other ways to accurately assess a driver’s risk without even asking about gender. In the end, the statistical fact that women are less risky to insure finds a way to get translated into statistically accurate pricing decisions, regardless of how our moral intuition feels about it.
    – Dan Romik
    Commented Apr 22, 2022 at 17:49
  • 4
    This suggests that if we look at intra-gender differences in car insurance pricing, we should see a difference to before: relative to average insurance costs, women who drive less safely than other women do on average will now likely pay more than they did before, while men who drive more safely than other men do on average will now likely pay less than before.
    – Jasmijn
    Commented Apr 22, 2022 at 21:30
47

It appears to be rational behavior for auto insurers to take a driver's gender into account when assessing the risk the driver presents to them and consequently how much they should charge that driver for insurance. In fact, it is rational for insurers to take into account any information about you that might affect the statistical risk to them from selling you insurance.

However, you are correct that this practice is seen as problematic by many, and goes against society's general desire for people to be treated equally and not discriminated against based on their sex or other attributes. This has led 7 US states so far to pass laws either fully or partially banning insurance companies from taking gender into account in setting auto insurance rates. Those US states are: California, Hawaii, Massachussetts, Michigan, Montana, North Carolina and Pennsylvania. The law in California went into effect on January 1, 2019, and appears to be the latest state law of this type. See also this press release from the Consumer Federation of America (where the seven states I named above are listed), and this Washington Post article.

The implication from the sources I've cited above is that charging women less than men for auto insurance is still legal in most US states. Presumably it is also legal in many other countries, and, as it is rational for insurers to engage in this practice (in an amoral sense that ignores any ethical implications and just looks at the pure economics of the question), presumably is practiced in at least some of those legal jurisdictions in which it is not prohibited.

1
  • Comments are not for extended discussion; this conversation has been moved to chat.
    – Jamiec
    Commented Apr 29, 2022 at 8:04
11

In Hungary in the past it was perfectly valid to charge differently based on various characteristics. This included the age and gender of the owner, or even auxiliary details, like the legal colour of the car. Some of these practices have been banned (as already mentioned in other comments the gender one for example in the whole of EU), but in Hungary to make sure companies don't cheat on these requirements it is now mandatory for every insurnace company to publish the algorithms they use to calculate insurance premiums. These algorithms are uploaded to the Hungarian Central Bank's website (note, the link is in Hungarian) and everyone can download and check them if they'd like, and you can verify for yourself whether their algorithms includes any difference between the various genders (now banned), or the colour of the car (now also banned).

In the past it was also a legal requirement to publish these algorithms every year in large national newspapers, so they were fairly simple, usually not longer than 4 pages long for every company, and it allowed people to actually calculate their premiums offline. However since this requirement was dropped, most companies started to simply publish 100+ page long tables with every make, model, engine power and postcode combination having their own entry.

Some of them are still moderately simple, for example KOBE is a fairly small insurance provider, and they still have a mostly easy to understand algorithm, that fits within 4 pages. Not sure how good Google translate (or any other translation system) would be on this PDF but you could check the various rules they apply, and see that gender for example is not included (but other things like age still is).

Do note this only applies for the 3rd party mandatory insurance, comprehensive insurance cover is usually bought separately and doesn't have this strict regulation on it, and the algorithm can be kept secret. However they are still not allowed to base the caldualtion based off on the gender of the owner (or driver)

5
  • 3
    A long time ago I heard about the car insurance being more expensive for red cars in Switzerland. Supposedly people who chose a red care drive more agressively.
    – Florian F
    Commented Apr 22, 2022 at 10:19
  • 1
    @FlorianF I believe in the UK a green car is statistically the safest.
    – Neil
    Commented Apr 22, 2022 at 16:14
  • 1
    @Neil, I guess as long as it's not British racing green...
    – Zeus
    Commented Apr 26, 2022 at 1:54
  • I like open algorithms, but if the aim was just to ensure that gender or colour were not taken into account, wouldn't it be sufficient to ensure that these characteristics are simply not provided for the quote?
    – Zeus
    Commented Apr 26, 2022 at 1:57
  • @Zeus these details are still needed to be provided for legal reason to the provider at some point. Like for online quotes it's easy to only ask these after the quote has been provided, but the law predates online systems by a couple of years
    – SztupY
    Commented Apr 26, 2022 at 8:38
6

Drivers are amongst the most highly 'statistised' group in the world. The stats gathered by insurers are in many ways intrusive but that allows them to gather huge amounts of data, and then tweak their policies to be both competitive and highly specific at the same time.

I worked at an insurance quoting provider (over 1M quotes per day), and they did so much A/B testing, on huge samples, to work out if (for example) men were more likely to buy a policy for a few $$ more on a Tuesday, then they were on a Friday. These kinds of tests were then fed back into the quoting engine to refine the policy offers.

Statistically, women ARE safer drivers, and should get lower premiums, but the only reason (in the EU at least) that women drivers aren't allowed to have lower premiums is due to sex discrimination legislation.

If allowed, the insurers would probably offer the lowest premiums, if they were allowed to find out what colour your hair and eyes are, as this (according to their research) could also be a determining factor on how safe a driver you are.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .