Car Insurance Named Driver Algorithms

Jd1399

It has come to the time of year where my search for the cheapest car insurance takes over. In my attempt to discover just how low I can get my insurance, I had a play around with named drivers.
I am currently a named driver on my partner's policy just in case we ever need to share driving on longer journeys. This caused a £40 increase to her policy when she added me on. However, if I added her onto my policy, it caused my premium to drop by £70. What I do not understand is that I am statistically the more experience driver. I have held my drivers license one year longer, I have a full four years no claims bonus and she only has one years no claims bonus as unfortunately she had to make a claim last year. 
I am not complaining as its cheaper insurance for us, but I am more concerned that I have done something wrong.
Should adding a named driver, with less driving experience (years since passed test) and less no claims bonus cause a car insurance premium to drop?

ciderboy2009

Did she add you on part way through her policy?  If so then the £40 was probably an amendment fee.

Jd1399

No both those circumstances are from new policies found via comparison sites. 

lindabea

I think you're trying to find some logic in a situation where none exists!!  I had a similar experience a couple of years ago; my partner was renewing his car insurance following a claim the previous year.  His quote was higher than the previous year's premium, however, when he added me as a named driver, the quote was lower than he had paid before the claim.  I suppose the 'logic' is that if I'm driving then there is less risk!!     

Jd1399

Yeah I was afraid of try to do so  
But in your circumstance there is some 'logic' as you are adding a more experience named driver to someone who's made a claim so the price drops. In broad terms. 

Whereas for me I'm adding some one who's made a claim to my insurance and the price drops. So the circumstance is reversed if that makes sense. 

lindabea

Well... you say there is some logic, and from an insurance company's point of view, it can be regarded as such.  But from our point of view, there isn't, because whenever we go out.(not that we do much these days), he does most of the driving anyway.  So, where is the logic in that?  And how do the insurance companies pro rata the 'driving  times' between the named drivers over the year to come up with a reduction in premium based on the fact that he will not be driving all of the time.  It's totally impossible; and that's why I say there is no logic in how they work it out.          

Sandtree

Jd1399 said:

No both those circumstances are from new policies found via comparison sites. 

This therefore muddies the water potentially as you are comparing a large number of different algorithms and only looking at the cheapest in both all cases. Even if its the same brand at number 1 it may be a different insurer if that brand is an intermediary.  You arent comparing apples and apples if you say Direct Line offered you a premium of £600 as IOD but Admiral offered £530 as a + spouse 

lindabea said:

Well... you say there is some logic, and from an insurance company's point of view, it can be regarded as such.  But from our point of view, there isn't, because whenever we go out.(not that we do much these days), he does most of the driving anyway.  So, where is the logic in that?  And how do the insurance companies pro rata the 'driving  times' between the named drivers over the year to come up with a reduction in premium based on the fact that he will not be driving all of the time.  It's totally impossible; and that's why I say there is no logic in how they work it out.          

You are over thinking it... insurance pricing has multiple considerations:

(Risk Premium/Claims estimate + Operational Costs) x Commercial considerations x profit margin x risk margin x Insurance Premium Tax

Commercial considerations can cover a range of aspects from growth strategy discounting to propensity modelling and can go wider than simply the premium presented... if your analysis shows the customer is a bargain hunter you show them the most basic product with all the optional extras removed whereas if they are in the mass-exclusive segment you show them your bells and whistles policy with several optional extras included by default.

The underlying risk premium (technical premium or host of other names) is not derived from logic in mass market consumer insurance but by simple statistics... if an IOD policy has a 10% higher claims experience than an Insured + Spouse then a 10% load is added. You can try and work out why that may be the case if you fancy but generally speaking that isnt a step gone through assuming the statisticians agree the delta is statistically relevant. That will cover cases that are 50/50 driver split and where the spouse is only there for emergencies. Its not a question asked and so will be based on an average.

Insurers always have to think hard about what questions to ask customers... long forms put customers off, how reliable is the answers going to be and can you prove things are a lie? So driver split may well have a material impact on premiums but you'd have no way of proving a false declaration and so its significantly devalued as a question. As mentioned previously, on some analysis done for a previous client of those claimants who had a theft or vandalism of their vehicle whilst at home and had declared the car was kept in the garage a very significant proportion didnt even have a garage and a similar proportion had a basic car port type thing that they'd declared as a garage. Thats the level of honesty on something thats possible to check, imagine how it'd be on something impossible to.

Jd1399

Sandtree said:

Jd1399 said:

No both those circumstances are from new policies found via comparison sites. 

This therefore muddies the water potentially as you are comparing a large number of different algorithms and only looking at the cheapest in both all cases. Even if its the same brand at number 1 it may be a different insurer if that brand is an intermediary.  You arent comparing apples and apples if you say Direct Line offered you a premium of £600 as IOD but Admiral offered £530 as a + spouse 

Interesting point thank you for the input. I didn't see it like that and in the end I suppose it does not really matter. I was just more curious as to how it worked. Both quotes are actually with Hastings Direct but I understand at the end of the day it's a complicated process that only a few people will fully understand. 

Sandtree

Jd1399 said:

Interesting point thank you for the input. I didn't see it like that and in the end I suppose it does not really matter. I was just more curious as to how it worked. Both quotes are actually with Hastings Direct but I understand at the end of the day it's a complicated process that only a few people will fully understand. 

The Risk Premium isnt THAT complicated, up unit a few years ago one of the largest UK private car insurers had a very simple system... premiums starts at £X and then a multiplier is applied based on one or two dimensional tables so for example the "Where Kept" was a one dimensional table with like:
Garage: 0.94
Driveway: 1.00
On Road 1.03
Other: 1.05
So if the premium was £1,000 at the point of hitting this question it could drop to £940 or raise to £1,050 depending on your answer.

Two dimensional tables cover compounding factors so back before it was made illegal age and sex was a two dimensional table as young male drivers paid more than young female drivers but middle age they were about the same.

Where things get messy is when you start adding in the commercial factors, if your propensity modelling shows someone is the type of customer that sticks around for many years then offer them a good price because you'll be getting revenue for the next few years whereas if someone is a switcher you'll only get premium this year from them so ensure you're not selling it at a loss.

These two elements are assessed independently and some may move hand in hand and others may move in opposite directions. So adding your high risk spouse to the policy may say from a risk perspective we need to bump up the premiums but for propensity it may mean your more likely to stick around so it'll lower the premium.

In my day pricing changes were discussed weekly and mountains of data is analyzed in preparation to both predict the impact of proposals and validate changes have had the intended impacts. These days AI/ML is starting to get more involved and that adds more fun to the equations

forgotmyname

Insurance and sense...

Adding a driver who had several accidents = cheaper quote.

Adding a 25 year old driver added £625 to the quote when they could insure the car for under £600

alone and my premium with no named driver was half that.

biscan25

I seem to have to add/ take off my parents each year, which changes the quote +/- £20. There's not a lot of logic. They live the other side of the country and wouldn't drive my car.
In recent times, insurance modelling uses machine learning, taking the human sense check out of the picture, which comes with some perverse outcomes. Admiral got in hot water for underwriting by name/ email address resulting in racist quotes.