Why aren't premiums better correlated to risk? It seems like that would be an obvious competitive advantage for an insurance company, if other insurera are mispricing risk.

It’s a mystery to me too.

Similarly, the cost savings from increasing my deductible just isn’t there. Going from $1000 to $2000 should be more savings than a handful of dollars per year, since that will save insureco on just about every single claim.

There’s no way they’re charging me $100+/month if my risk of a claim is 2-3% per year.

Increasing your deductible from $1000 to $2000 also eliminates the insurer's claim administrative expenses for claims in the range of, say, $1500 to $2500, since you'll report claims in that range if your deductible is $1000 but not if it's $2000.

But insurers' models are complicated. One factor to consider here is that the population with $1000 deductibles isn't the same as the population with $2000 deductibles. For example, some of the latter group have a higher deductible because they're financially stable enough to afford a $2000 emergency expense, but others have it because they're financially unstable enough to need the cheapest coverage they can get. There's a strong negative correlation between an individual's financial stability and their risk to an insurer, so in net, it wouldn't surprise me if the latter group is riskier on average.

Also, even if raising your deductible reduces your expected claim costs by more than a handful of dollars per year, insurers in most of the US aren't required to pass that savings on to you - they can price coverage based on your willingness to pay (see [0] and [1]). On the bright side, auto insurance is arguably about the closest thing there is to a perfectly competitive market in the US - your insurer has lots of competitors, and if they're overcharging you relative to your risk, it's easy to find a better deal elsewhere.

[0] https://www.naic.org/cipr_topics/topic_price_optimization.ht...

[1] https://www.npr.org/2015/05/08/403598235/being-a-loyal-auto-...

I’m not in US, which is probably why I dream of your rates.

Our government policy (Ontario Canada) is to ask insurers how to reduce premiums, then the government lets them reduce/cap benefits and premiums go up anyway.

It’s one of the many markets we would open to more competition. A lot of US insurers would salivate at our rates.

Yes, I don't think people outside of Ontario understand how hosed we are on car insurance. I admit we have fraud, but do we have more fraud than norther Illinois? No. Do we have different weather than northern Illinois? Again, no.

Because of Pareto principle. 80% of the insurance risk could be estimated with 20% of variables.

Which is more then enough because most of the insurance pricing is determined by sales and administration costs and not by price of risk itself.