A counterpoint (not necessarily my view but a view nonetheless) would be that if everyone pays into a single required insurance pool then the risk doesn't need to be accurately determined a priori. Each individual's rates might be the same, but they could be lower overall because there are more paying in. A low risk driver may pay more than they use, but it could be less than the market rate regardless.

The risk still needs to be accurately determined, even if not for each individual. If 100 people in a population will be injured and get a pay out of $100k, you need to know that to set general rates.

What I’ve seen in places that have government insurance is the government is constantly trying to backstop the insurance.

Why? It becomes a political hot potato when rates go up, so there is pressure to keep them down with lower premiums. The insurance body doesn’t care because the government will bail them out.

You can just update rates after the fact if necessary. I don't really see the problem with that. People with a bad history may get higher rates, just not based on risk analysis but based on their actual history.

The only way to convert "has a DUI" into a specific rate increase is through risk analysis of everybody with DUIs.

Why is that the only way? Seems to me you could just raise rates after someone caused actual damage instead.

By how much? Logically, the fair amount to raise rates would be the amount that compensated for their increased risk of doing it again. There you are, doing a risk analysis.

(If you had them directly pay for the damage they caused, that wouldn't be insurance, that would be a middle man for restitution.)

By some amount proportional to the damage caused, of course. The more times you cause damage, and the higher the damages, the higher your rates go over time.

I think my point is pretty clear without me coming up with a precise formula: instead of predicting rates, just adjust them based on history instead. After all, someone likely to cause repeated damage in the future is also likely to have caused it in the past.

The only major difference between the two approaches is that mine doesn't require someone to make predictions - which might be biased - in exchange for mine maybe under-charging someone with no negative history who is about to cause a huge amount of damage and then stop paying into the system (death, etc). But in those (rare?) cases, the fact that everyone is part of the system means that the single rare loss is amortized nicely and without bias.

>The more times you cause damage, and the higher the damages, the higher your rates go over time.

But by how much? If it's enough to fully offset the cost of the damage, that's not insurance, but a payment plan. If it's less than the cost of the damage, then you have to decide what it's going to be.

Again, I'm not going to come up with an exact formula. The point stands without one.

You can calculate using any history you like. Weigh past damages based on time between, or time in the past. Be creative.

>The free market is required to accurately price the risk for each driver.

That same market means some will decide that not having insurance is a better deal. Tying insurance to something like drivers licences increases the cost of not having insurance.

>I do not see how a state-provided solution would come up with "accurate" pricing for individuals.

Can the state not hire actuaries?

>Can the state not hire actuaries?

yes, but a sibling commenter has weighed in how it turns out: https://news.ycombinator.com/item?id=29719762

A government entity could perform this, no free market required. See residential flood insurance for a government based risk rating example.

IIUC, that is actually a form of government subsidy, which exists only because there are places that are uninsurable due to flood risk. It is politically unpopular to tell people they have to self-insure because they picked a risky to live. If we decided that were something we were willing to tell people, there'd be no issue with letting the market handle it, same as most insurances.

You mean the same policy that resulted in a whole bunch of people living in flood zones? And which we can't get them to leave because it would be politically unpopular to raise their insurance premiums to reflect their actual risk? That seems more like an example of exactly why we don't want the government -- which inevitably means politicians -- setting insurance rates.

Government provided residential flood insurance is a horrible example of moral hazard and should be eliminated. Taxpayers shouldn't have to subsidize investors who choose to purchase property in flood zones.