As an example, Hurricane Katrina hit the Gulf Coast back in 2005, and it was a total disaster. Flooding, of course, was not covered by private insurance (it's been like this for a very long time), only wind, so there was a lot of arguing over whether insurance companies had to cover losses because the storm surge was actually caused by wind, and a lot of damage was solely caused by the hurricane-force Category 5 winds. I'm probably leaving some important details out...

Anyway, the aftermath was that private insurance companies in that area refused to insure wind damage to buildings. It just wasn't worth the risk. So the state (Mississippi) came up with something called the "wind pool" which was government-provided insurance for wind damage. The problem was, it was really expensive. So a lot of people never did rebuild, a lot of people left (or moved farther inland, where the insurance was cheaper), etc.

The bottom line is that enabling stupidity as far as building location only works long-term if it's being subsidized. If insurance reflects the true risk of building in a certain place, then it really limits that kind of activity.

Reasonable insurance is expected cost * probability + enough margin for a small profit + cover bad bets.

The gap between a reasonable rate and the rate paid is a gift paid to homeowners with above average wealth and builders with below average sense.

Since builders can earn more building a home that will be flooded out in 7 years with a higher selling price by virtue of beings water front property but only if the government subsidizes insurance they are directly incentivized to do so.

Most of the initial premiums are based on 100 year flood plain maps based on recent years of rain (which may not have been typical), assume that rivers are static objects that do not change course, assume that coastal sand beaches are static and not dynamic.

By letting people stay in places that flood repeatedly and removing the risk for home owners/purchasers, the NFIP incentivizes building and rebuilding in areas that flood. According to this article, there were 37000 homes at the time of writing that had flooded four times and gotten at least 5000 in coverage payout (I could be wrong but last time I checked that is the most I could get in earthquake insurance when I read the annual offer around 2000 from my homeowners insurance). The NFIP is underwriting pretty well off people's lifestyle of living on a beach/river front/bayou. https://www.seattletimes.com/business/flood-insurance-rebuil...