> The risks of nuclear power are still largely unknown, as there have been only two major disasters over a few decades.
No, the risks of nuclear power are very well known, as there have been only two major disasters over quite a few decades, and the root causes of both of them are well known and easily avoidable.
> We don't know how much worse the reasonable worst case could be.
Yes, we do. We know that Chernobyl was worse than any reasonable worst case for any other reactor, because nobody is going to build another reactor with a positive void coefficient of reactivity and with no secondary containment, and then run uncontrolled experiments on it. And we know that Fukushiman was a reasonable worst case, because it subjected a just shut down reactor to zero decay heat removal for an extended period, and that is going to be the consequence of the worst possible accident that could happen to a running reactor, since any such accident will shut the reactor down.
In other words, in a sane world, the risks of nuclear power would be more insurable than the risks of, say, coal power, precisely because the nuclear risks are contained. No insurance company is going to sign up to liability for deaths due to respiratory failure from breathing coal dust over a period of many years. But in our insane world, we don't hold anyone accountable for such risks, so we end up treating nuclear, which is far safer per unit of energy generated than any source except solar, as if it were the riskiest of all sources.
You are talking about risks in a narrow technical sense. Insurance companies are more interested in the consequences of the disaster.
A reasonable worst case is something that could plausibly happen once in a century in the entire world with 10x more reactors than today. Maybe the reactor is located close to a major city. Maybe the company operating it has become incompetent and corrupt. Maybe the weather conditions are particularly unfavorable and most of the fallout goes towards the city. Maybe the meltdown was caused by a natural disaster that also makes evacuating the city harder. Maybe there are also widespread protests happening for an unrelated reason.
Or maybe the worst case is the incompetent and corrupt power company upgrading its reactors in a way that leads to a large number of meltdowns.
Remember that we are talking about what happens when the power company cannot pay and we don't want to socialize the risks. This is not something where the insurance company is allowed to set the terms. If courts and/or politicians ultimately determine that the power company is liable for $1 trillion, but its assets were only worth $100 billion, the insurance company must cover the rest.
That is a mistaken view; he's actually talking about liability. The risk of nuclear is smaller than the damage of the status quo, but it is so unexpected that we hold people liable when it materialises.
The damage from coal (not risk because the risk is technically 0 - the bad outcome is guaranteed by design) is substantially greater than the risk of a nuclear disaster. We just don't hold anyone liable for fossil fuel damage because then society would collapse. It is pretty stupid that we don't move to nuclear and highlights the inability of the anti-nuclear crowd to do cost-benefit analysis.
> No, the risks of nuclear power are very well known, as there have been only two major disasters over quite a few decades, and the root causes of both of them are well known and easily avoidable.
I think the statement that you were responding to would be better phrased as: "The risks of nuclear power are still largely unknown, since there have been only two major disasters over a few decades." Insurance companies are typically okay with insuring against high cost, high frequency occurrences as long as they can quantify it. If they can quantify it, they can set a price on it.
It is also worth noting that insurance companies were among the first adopters of digital computers because the better they can quantify frequencies and costs, the better they can set rates.
> I think the statement that you were responding to would be better phrased as: "The risks of nuclear power are still largely unknown, since there have been only two major disasters over a few decades."
That was how I interpreted the statement when I responded to it.
> Insurance companies are typically okay with insuring against high cost, high frequency occurrences as long as they can quantify it.
Insurance companies are also good at specifying insurable risks and providing incentives to the insured to manage them properly, even for rare but high cost events. That includes refusing to insure people who the insurance companies judge to not have good judgment. For example, no sane insurance company would ever have insured Chernobyl given its insane reactor design, nor would any insurance company have agreed to pay a claim when they found that the reactor operator ran an uncontrolled experiment on the reactor and that was what caused the accident.
As for Fukushima, as I commented elsewhere upthread, an insurer covering a reactor in an earthquake/tsunami zone would probably either not include tsunami inundation coverage in the policy, or would require a separate rider that would be priced separately. And such a rider would also have a list of requirements for coverage, which would include things like "don't site your backup power and switchgear where a tsunami might inundate it".
In other words, as I pointed out, the risks of nuclear power are well known, because we understand how reactors work and what the failure modes are and insurers therefore know how to write policies that set the right incentives for managing risk. That is true even though we have had only two major disasters over quite a few decades; in fact, as I pointed out, the fact that we have had only two such disasters, and both of them were due to root causes that are easily avoidable in the future, is a reason to be more confident that the risks are well understood and insurable.
Three Mile island was very close to being a major disaster, and upon further review had more radiation escape than was initially disclosed.
The list of all of these nuclear incidents is interesting, many of them seem to be various components of the system exploding for various reasons, which could have led to horrible outcomes but luckily did not: https://en.m.wikipedia.org/wiki/Nuclear_and_radiation_accide...
> Three Mile island was very close to being a major disaster
No, it wasn't. In fact, the plant operators made just about every possible mistake they could have made--and still no harm to the general public. TMI was actually an illustration of how good design of safety systems can protect the general public even when reactor operators make multiple bonehead mistakes. Unfortunately the media narrative, as in so many other cases, drastically misrepresented what actually happened.
> and upon further review had more radiation escape than was initially disclosed.
Yes, but still zero harm to the general public.
> many of them seem to be various components of the system exploding for various reasons, which could have led to horrible outcomes but luckily did not
I do not see any incidents with commercial power reactors that fit this description.
Yep, it's a perception issue that isn't helped by the fact that to normies, "nuclear" means "bomb", and the oil and gas industries have massively more lobbying dollars to get politicians to push their solutions above all others.
No, the risks of nuclear power are very well known, as there have been only two major disasters over quite a few decades, and the root causes of both of them are well known and easily avoidable.
> We don't know how much worse the reasonable worst case could be.
Yes, we do. We know that Chernobyl was worse than any reasonable worst case for any other reactor, because nobody is going to build another reactor with a positive void coefficient of reactivity and with no secondary containment, and then run uncontrolled experiments on it. And we know that Fukushiman was a reasonable worst case, because it subjected a just shut down reactor to zero decay heat removal for an extended period, and that is going to be the consequence of the worst possible accident that could happen to a running reactor, since any such accident will shut the reactor down.
In other words, in a sane world, the risks of nuclear power would be more insurable than the risks of, say, coal power, precisely because the nuclear risks are contained. No insurance company is going to sign up to liability for deaths due to respiratory failure from breathing coal dust over a period of many years. But in our insane world, we don't hold anyone accountable for such risks, so we end up treating nuclear, which is far safer per unit of energy generated than any source except solar, as if it were the riskiest of all sources.