No fault incident which was not even claimed increasing my premiums

InsideInsurance

mikey72 wrote: »

1) Which way do you what the money to go?

2) What do you want to justify. I only need about 30 samples from garage costings where they have put colour on to show statistically either way.

3) I don't need to ask, I'm just proving blue cars should cost more.

1) Depends if I'm on commission or day rate for the insurer

2) Fortunately/ unfortunately insurers in the mass market space need a little more evidence than 30 cases. Many insurers have had millions of customers for decades.

Evidently there is at least a certain concern of lies, damned lies and statistics but given the premium is then adjusted for commercial considerations AFTER the statistical analysis there is little point in intentionally looking to skew it from "the truth".

That said, there is "expert judgement" to be involved in how much weighting you put on newer evidence -v- older. Evidently the bigger the sample the more statistically relevant the conclusion but if behaviours have changed over time you want to either weight the results or only take more recent years.

At the end of the day that's why the actuaries and other hardcore statisticians are paid £50k+

Once you've got "the truth" you can then adjust for commercial considerations as long as you dont threaten the stability of the overall book

mikey72

InsideInsurance wrote: »

..............At the end of the day that's why the actuaries and other hardcore statisticians are paid £50k+................

No doubt they have figures to show they're worth that much.

And, if they are, it's to ensure the premiums aren't set too low, I'm sure if they were set too high, no one would complain unless the customers left. And that's further checked by comparison to competitors.

InsideInsurance wrote: »

............Once you've got "the truth" you can then adjust for commercial considerations as long as you dont threaten the stability of the overall book

That's what I've been saying.
I have no doubt the maths can give the truth.
The problem starts when the truth is so big no one can see all of it, and if eveyone "knows" their little bit is right, it's very easy to add a few more bits here that no one can see. But it must be right, yes?
And if you can convince the punter the reason sounds good behind the truth, then it's pure profit for no work.

InsideInsurance

it's to ensure the premiums aren't set too low, I'm sure if they were set too high, no one would complain unless the customers left. And that's further checked by comparison to competitors

It is to show the technical premium is correct and that the commercial adjustments will not negatively impact the book.

Whilst you can push the prices up when you do so you reduce the amount of sales. Inevitably there is at least in theory a sweet spot where you balance profit per sale against volume of sales. Insurers use a number of methods to try and measure elasticity to predict conversion impact of any pricing changes

Competitor analysis is actually fairly difficult to do, especially as people differentiate pricing model by distribution channel etc. Insurers have invested a lot of money to try and identify "rate raiders" and block them (or send spurious data).

There are organisations that offer to give anonomised results in exchange for your own prices but they only test a dozen or so theoretical customers and so no way near a big enough sample to give any feed into individual element pricing decisions.

The problem starts when the truth is so big no one can see all of it, and if eveyone "knows" their little bit is right, it's very easy to add a few more bits here that no one can see. But it must be right, yes?

Whilst pricing is getting ever more sophisticated we aren't yet in a position that it is so big that "no one" has the oversight of the full picture.

What in theory could be a problem is being able to understand why X months/ years ago it was decided to apply a 20% deduction to the technical premium for people with X cars and living in the XXN postcode district and thus if that decision is still valid or if the discount should be removed.

I've seen this, and similar types of issues, happen many times before both in insurance and other industries.Its a pain in the !!!! to deal with and inevitably someone in IT discovers it and comes to your desk to ask what to do with it and stands there as if you'd be able to answer immediately.

This is evidently a governance issue and having a proper method of logging decisions and reviewing them is important.Workflow to promote them to live and a regular review of all adjustments. Unfortunately this also creates overheads and makes reacting more sluggish.

At the last place I was dealing with pricing every adjustment from the technical premium had to be reviewed every 6 months by the pricing committee and the systems were programmed to do this along with the requirement for changes to be documented and signed off by 4 heads of/ directors.

dalesrider

pmpm wrote: »

Can somebody please advise me?

But I made a big mistake by reporting this incident to my insurance company(first direct) (I was trying to be bit honest there :mad:!!!)

I called my old insurance company (First direct) but they were closed today.

1st point.... Since when has First Direct been a Insurance co?

2nd point... First direct is open 24/7/365/6. :j

Spiderham

mikey72 wrote: »

What conclusion?
That someone having a third party no fault accident is more likely to have a fault accident, or no more likely?

Sorry, to clarify the conclusion was that a claim costing money

mikey72 wrote: »

You can give me a comparison from statistical analysis of quality data, mainly six sigma, but you can include mtbf for semiconductors, on accelerated stress test, or autoclave, as they seems to work best.
If you know Taguchi, that's a fairly simple statistical technique.

Or just tell me the way you do it, and I'll try to keep up.

I was on about the "justification" for increasing the prices following a fault claim rather than the process of analysis. You are constantly saying you want to see it with your eyes and the analogy was intended to be an area where the background has to be taken on the trust of the expert.

If I came out and stated definitively my analysis/research (using a sample of all policies of an insurer, let's say 1,000,000 for a big one) showed on average 50% more claims, on average costing 30% more each per claim so cost per policy was 65% higher per policy when normalised for other risk factors would you believe me? If not what else would you need to be convinced?

Oh, by the way those figures are out of thin air.

mikey72 wrote: »

(Or I bet you I could show you why blue cars could have more accidents from your raw data, if you want to charge them more this month, I may need to be selective in the data I use, but I could justify that too)

Mikey, if an insurer feels that blue cars cost more in claims they can charge like that. Some may even do so.

If an insurer just wanted to raise the rates for some factor (has been known for over exposure in particular area, say to many of a vehicle make when the manufacturer has gone bust) they could just do it. It can also go the other way if a risk factor is shown to be higher risk but you decide not to increase prices because of the PR factor surrounding the issue. Insurers don't want to fudge the data to get a particular answer they want the real view of the situation separating the underlying pattern from the randomness so that a commercial decision can be made from full information. If something comes up contrary to "perceived wisdom" it will certainly be double-checked but if it is a true pattern people making the rate decisions do want to know about (sometimes they'll ignore it).

This has been about loading after a non-fault accident. We're all aware there is certainly a load from most insurers following this. Is this acceptable to people without in depth analysis as it's "obvious" and "common sense" so can be taken as gospel?

mikey72

Spiderham wrote: »

..........This has been about loading after a non-fault accident. We're all aware there is certainly a load from most insurers following this. Is this acceptable to people without in depth analysis as it's "obvious" and "common sense" so can be taken as gospel?

I think if you read the posts on here, most people don't think that increasing premiums after a non-fault accident is "common sense".
The majority indeed do agree it's "obvious" it's a "load" though.

dunstonh

mikey72 wrote: »

I think if you read the posts on here, most people don't think that increasing premiums after a non-fault accident is "common sense".
The majority indeed do agree it's "obvious" it's a "load" though.

"most people" dont understand and dont choose to understand risk & pricing. If people wanted to understand then the Daily Mail would go out of business.

InsideInsurance

dunstonh wrote: »

If people wanted to understand then the Daily Mail would go out of business.

Tits and !!!!!! will always sell

Spiderham

mikey72 wrote: »

I think if you read the posts on here, most people don't think that increasing premiums after a non-fault accident is "common sense".
The majority indeed do agree it's "obvious" it's a "load" though.

Sorry, this is why I shouldn't type at midnight. I missed out the sentence but what about fault claims?

This has been about loading after a non-fault accident. But what about fault claims? We're all aware there is certainly a load from most insurers following this. Is this acceptable to people without in depth analysis as it's "obvious" and "common sense" so can be taken as gospel?

That makes more sense now. I don't even know if there is a load from most insurers following a non-fault claim and wouldn't like to say either way, whereas I'm safe to say that on fault claims.