I am speculating here a bit as my information is not first hand – I’ve never owned an aircraft and I’ve never looked at accident statistics in details. But I’ve heard the following statements:
Insurances usually work a lot with statistics and probabilities, so why do they decrease the prices exactly at the time an accident is statistically most probable to occur?
Don’t think of it as “Pilots with 500 Hrs” versus “Pilots with 499 hours”
In reality it’s a statistical thing, “ALL the Pilots with 0 to 499 hrs” versus “ALL the Pilots with 500 hrs to 50,000 hrs”
Neil wrote:
In reality it’s a statistical thing, “ALL the Pilots with 0 to 499 hrs” versus “ALL the Pilots with 500 hrs to 50,000 hrs”
That is not the meaning of “peaks at 500h”. It should be more something like this:
Insurances assume and average aircraft/mission (e.g. private use, training use, utility, commercial…) and price based on pilot experience, this gives low premium for high hours: on same conditions/missions a high hour pilot is less risky
When you aggregate this on all pilots it needs only to “work on average” for them to be able to get a profitable business when “risk netting across their pilot pool works” as long as that graph stays as it is over time and they have a big diverse pool of pilots.
Now from a single pilot perspective this leads to some positive ups and negative downs, a pilot who takes more risks that the average mission will get good value for money as he gets subsidized by those who under use their mission, then you have the moral hazard to account for in that curve: experienced pilots tends to take more risk….
Insurances are made with “or more” clauses for pilot experience. Let’s assume the most dangerous pilots statistically are 500h pilots. A “100h or more” insurance will cover the 500h pilots, so needs to be expensive.
lionel wrote:
Insurances are made with “or more” clauses for pilot experience. Let’s assume the most dangerous pilots statistically are 500h pilots. A “100h or more” insurance will cover the 500h pilots, so needs to be expensive.
But “700h or more” gets me far away from the most risky ones.
@Vladimir, on the maths you may get weird results trying to fit that curve with two premiums “experienced premium” “junior premium” as Neil mentioned, especially if you have no clue what sort of pilots you got inside your pool 
However, you have to take that graph as input to stay competitive (it is somehow a legal requirement)
Neil wrote:
Don’t think of it as “Pilots with 500 Hrs” versus “Pilots with 499 hours”
You can always have separate premiums for a 499h pilot vs 500h pilot but then you don’t have many in each pool category to get any diversification benefit and you will not be able to explain to management who is causing profit/loss, so most insurers will just slap the two with the same number and blame it on the 499h guy 
There are also two separate questions: why there is a peak in the hazard rate? why insurance does not follow that peak?
The first one is just an artifact: few accident on low hours pilots as they are much controlled and few accidents on long hours pilots as they are they have seen it all but also they are few so subject to from “survivorship/selection bias”, so there must be a peak somewhere: regulators seems to advocate 150h/200h/300h, some thinks it is 1500h, lot of insures believe in 100h (twice PPL time), most aggressive insurers thinks it is 75h (PPL time + 25h), empirically it seems 500h…
The second one is related to insurance business and pricing, they want to charge more on a large proportion of my pool that they “qualify and can justify as risky” or alternatively have more excess (which relate to the payout amounts not accident probability), ask for instructor checks….remember it will be really hard for a sales guy to do marketing of a premium curve that follow that hazard rate curve but as long as they have it “balanced on average” on their specific pool while keeping competitive pricing they should be ok in terms of profitability
Also without having sufficient data you can go into all sort of paradoxes: premium for an exotics aircraft is usually off due to few data points, they tend to be flown by more experienced guys and that data get re-applied to less experienced guy for a quote using the above curve, so they will get a “cheap insurance” for something that they will surely crash in their 46 hour…but they may not care as long as “less than 100h pilot accidents” stays in the empirical range
@Neil wrote:
Don’t think of it as “Pilots with 500 Hrs” versus “Pilots with 499 hours”
I wanted to join a group which owned an aircraft and they told me that I have to cover the full increase of premium because the insurance had exactly such a rule: if all pilots have >= 500h, the premium is lower than if there is at least one pilot with <= 499h.
Where did lbra get his statistic of accidents peaking at 500 hours?
I’ve never seen such a statistic quoted before. The closest I’ve seen is the first 500 hours being called the “Killing Zone” which comes from a book with that name, with very flawed statistical analysis. Basically the book says that most accidents happen to a pilot within their first 500 hours. They came to this conclusion because the vast majority of accidents happen to pilots with less than 500 hours. But they never took account of the fact that very few pilots ever reach 500 hours, with most giving up flying well before that. If most pilots have less than 500 hours, then most accidents will happen to pilots with less than 500 hours.
