For the insurance example, you're describing insurance as a forcing function for better made, safer buildings. That's what building codes are for though, we shouldn't need to have both and building codes are a more efficient and direct way of ensuring safe buildings.

For car seats, I'm not sure how we could know that people wouldn't buy them. I don't expect anyone would propose dropping the requirement to see how the market responds, and probably rightfully so. If car seats are much safer though (and I'm obviously not disputing that), people that can afford one would buy it anyway.

> ... building codes...

I agree that in an ideal world that would be sufficient. But in practice, governments rarely deploy trained actuarial to make decisions, rather relying on politics and shoddy studies. Government codes also change very slowly. Insurance companies (whether private or public), under the financial incentive, are constantly changing their policies and rates in response to new data and calculations. I would be open to looking at studies that resolve this question one way or another.

> ... car seats...

I grew up in a poor global south country. Rich people, who clearly can afford them, don't buy car seats. Many people who live in countries where they are forced to buy car seats, when they come back on vacation don't use car seats for their kids. People can be very irrational.