A modest proposal for reforming the NRC

Figure 1. NRC's cost/release optimum vs society's optimum. For explanation, see There are no solutions, only tradeoffs.

The choir is an obstinate lot. Despite my manifestly compelling argument for Underwriter Certification, they insist on singing from the We Must Keep the NRC hymnal. It's some form of weird nostalgia for a terrible past. Just reform it, they chorus. OK, I give up. Let's reform the NRC. That means aligning the NRC's incentives with society's. We must close the gap between NRC's optimum on the tradeoff curve, Figure 1, and the societal optimum.

The NRC has about 4000 employees. Suppose each NRC employee were paid $1000 a year for each operating nuclear GW the NRC had approved, provided the electricity is sold at under 3 cents per kWh. This compensation would increase each plant's cost by 1.6%, about the cost of UCert regulation, But the carrot comes with a stick. For every public Lost Life Year due to release radiation according to the NRC's radiation harm model assuming no evacuation, each employee would be sentenced to a day in prison, a real prison. Not a Club Fed.

Now the NRC employees face a decision that is rather like society's. They must trade off wealth versus potential harm.¹ If Congress has picked the bonus and the prison sentence correctly, the reformed NRC's decisions would look a lot like UCert's. Unfortunately, we lose the inherently balanced nature of UCert. Congress could pick the numbers very poorly.

Another issue will be the NRC's radiation harm model. Under the new incentives, the NRC would quickly ditch any harm model that vastly over-estimates radiation cancer for the dose rate profiles that will be experienced by the public in a release. LNT would be summarily debunked and trashed. What were we thinking! The problem is the NRC will now come up with a model that under-estimates harm. Who wants to go to jail for a harm that is so tiny that it can't be detected?

I imagine the bureaucrats will look at the Green Table and decide that, as long as the dose rate is less than 20 mSv/d, there is no compensatable harm, for we have not seen any. One ex-NRC employee has already proposed such a no harm, no foul model. Congress may decide to step in and require a more conservative model. They might even go so far as to impose UCert's very pessimistic SNT model, although this will be strongly opposed by the bureaucrats as unrealistically conservative.

They will have an argument but, if Congress stands its ground, the NRC will have to accept SNT. Even a partially decarbonized US, will require about 1000 GW's of nuclear at which point each NRC employee will be making a million dollars per year. A TMI would result in no prison time.² Even a Fukushima every 10 years in a 50 year career, would mean a total of less than a year in prison under UCert SNT. The NRC will be flooded with job applications including mine.

The reformed NRC will come down hard on dense-packing. Nobody wants to go to jail in defense of bureaucratic infallibility. It will insist on adequate buffer zones. No more nonsense about micro-reactors in the basement. But a release will no longer be intolerable. Any regulation that can't show that it materially reduces the regulator's expected time in jail will be jettisoned in the drive for 3 cents/kWh.

At least as importantly, the NRC would suddenly become a big believer in competition. Anybody who can build something a little cheaper would be welcomed with open arms. If the incumbent vendors tried to collude and push up cost, the NRC would have them in court the next day. The NRC would test product, not certify QA systems. Substance would replace process. Test the weld, not the welder would become the mantra. Guarantees would replace N-stamps. Full scale, long term, stress testing in a remote location would replace years of paper analyses. If Congress also replaced the American tort system with the UCert radiation compensation plan, the reformed NRC would look something like UCert.

Of course, my very poorly defined bonus is not a good proxy for market pressures. If it were, I can see no way of implementing it. No way of getting from here to there. But the point is that to really reform the NRC, you must make the regulators' incentives consistent with societal welfare. If you put the regulator in charge of making tradeoffs, the regulators' tradeoff must be the same or nearly the same as society's. Unless you can do that, the only good NRC is a dead NRC.

1. Pollution including CO2 will have to be handled by imposing taxes on the polluters. But Polluter Pays is the way it should be.

2. A TMI would still be a multi-billion dollar loss for the utility and its customers, something the utility would work very hard to prevent. This is a hint that the NRC is essentially redundant. The plants that were built without an independent regulator early in both the US and French programs have safety records that are every bit as good as the plants that had the benefit of NRC regulation.