No, it doesn't. Because I don't get insurance coverage for a nominal fee like $1 per month. I pay a non-trivial amount of money every month. So in my mind, 9

No, it doesn't. Because I don't get insurance coverage for a nominal fee like $1 per month. I pay a non-trivial amount of money every month. So in my mind, 90% of that is going towards the possibility that I'll eventually make a rare claim (float), 5% is going towards administration, and 5% is profit. I know those numbers aren't right, but lets say that they are. In that case, paying $100/mo for insurance and not making a claim for 10 years there's over $10,000 worth of loose money (more once you factor in interest!) sitting around that's just for me because according to your analysis insurance isn't risk pooling but risk smoothing. So then I make a claim for the whole car, $20k. Fine, half of that comes out of the money I already paid to them for risk smoothing, the other half should come from the future payments I'll make to them for risk smoothing. My rates should double in this case. But most people don't total their cars every 10 years, many people go their whole lives without ever making a claim. So at the end, they might have paid $50k in self-insurance risk smoothing premiums that the company gets to just keep, that's not right is it? Clearly no. So what people actually think is supposed to happen is that everyone pays about the same amount and some people get unlucky and make claims and that doesn't get counted against them. They're all wrong, but that's how insurance is sold so it's understandable. In reality I suspect that the breakdown of float, administration and profit is more like 30%, 30% and 40% which is why premiums can go up so much after you make a claim; you're actually making almost no real contribution towards the float at "regular risk" insurance rates. Of course this also neglects the risk pooling aspect.