This is something else where I am amazed that they get away with it.
Assuming you are a two-adult household with two cars, it is usually cheaper to insure both cars with the same insurance company, and if you are a many-car household, much the same applies, you get discounts for each additional car.
However, it strikes me that the discounts are nowhere near big enough, because the marginal extra risk associated with one extra car is negligible, it is only the number of drivers which really matters.
Taking a two-car family (like mine), a lot of the time we are both in her car or both in my car, while the other car is safely parked up at home. In other words, a lot of the time the total risk is much the same as if we only had one car.
And there are some journeys which one or the other of us has to make, like picking up the kids or going to the supermarket, so the fact that I am using my car at any moment means it is less likely that she will be using hers. The total number of miles we drive between us is barely higher for us having two cars than if we only had one.
It's the same with a single, unmarried car enthusiast who has half a dozen useable cars on his driveway, he can only be using one of them at a time.
-------------------------------------------------
Ultimately, I'm not sure that compulsory private third-party insurance makes sense anyway.
a) There are three basic principles to insurance:
i. Risk spreading, where one specific party cannot bear the full cost if his project goes wrong, so he invites lots of other people to take a small part of the risk, and
ii. Risk pooling, where large numbers of people are running a very small risk of incurring what is for them a very large loss (like your house burning down). But from the insurance companies' point of view, that is not really a risk at all, they know that every year x,000 houses will burn down and that they will have to pay out £y million, so they divide that £y million by z million people with a house and everybody chips in a couple of hundred quid a year to a common fund.
iii. Self insurance. If you only own one home, you have to have fire insurance. However poor value it is, you wouldn't be able to sleep at night if you didn't have it. But what if you are Prince Charles and own thousands of homes? You know perfectly well that every few years one of them will burn down. It's cheaper paying for one to be rebuilt every few years than is to pay for insuring each one.
(Of course, we have to have some sort of criminal sanctions for unduly reckless drivers who kill or hurt somebody, or even if they don't, separate topic.)
b) But as most households own a car or two and nearly everybody is at risk of being hit by somebody else's car, whether as driver, passenger, pedestrian or owner of physical property, and the amount of damage you suffer as a result of a collision bears little relation to anything apart from sheer blind bad luck. This brings us safely into the realm of self-insurance - not for each individual but for the UK population as a whole.
(It's not like fire insurance where each owner has some influence over the risks, i.e. if you own a big house the total potential damage is much greater than if you own a small one; if you install loads of fire alarms, your risk is lower than if you don't etc.)
c) And the total amount of damage caused by thirty million vehicle owners to the rest of the population is a fairly stable figure (and tiny as a % of GDP). For some reason, the cost is recovered only from car owners, even though non-car owners are also at risk (of them or their front wall being hit).
d) So if you ask me, it would be cheaper and more efficient to bypass the insurance companies and just fund the third-party element out of general taxation (or out of fuel duty, or whatever seems appropriate), that way absolutely every risk is pooled in the same place, administration costs are minimal because you don't need dozens of separate companies collecting hundreds of millions of payments a year, you don't need expensive advertising.
e) This logic does not extend to fire and theft insurance, those are specific risks over which the owner has some control, the same as with fire insurance for your home (see above). There are plenty of people (like me) who simply don't bother with fire and theft insurance for their cars because worst-case having to spend a couple of grand on a replacement is well within my means, meaning that people like me wouldn't need this faff with insurance every year. Her Indoors is a bit more car-proud than me, so she probably would insure hers for fire and theft (even though I wouldn't bother with hers either), fair enough.