Critics of nuclear power like to claim that nuclear power is heavily subsidized and one of the biggest alleged subsidies (and one that makes activists positively apoplectic) has to do with the Price-Anderson Nuclear Indemnity Act (PAA) and accident liability. It is obvious from recent comments made on Kos that many people are under the mistaken impression that Uncle Sam cuts annual checks to Acme Insurance to pay the liability premiums for the country's 104 nuclear plants and have suggested, for example, that those tax dollars would be better spent installing solar panels on everyone's roof. PAA isn't that kind of subsidy, if you want to call it that.
In 1954 Congress passed the Atomic Energy Act which made the development of commercial nuclear power for peaceful purposes a national priority. Lewis Strauss, then chairman of the Atomic Energy Commission and enthusiastic booster of all things nuclear, pushed and prodded the nation's utility executives over the next several years to begin building plants and even threw in an offer of seven years worth of free fuel to sweeten the pot, but could find few takers. The thing holding back the works was the question of liability.
Because nothing in life has zero risk, even the most ardent nuclear power supporter was unwilling to operate a power plant without adequate liability coverage. Insurance companies were willing to offer $60 million policies, far more than had been offered to any other industry up to that time, but it was still deemed inadequate by a long stretch. As a result, General Electric indicated that it would stop work on the Dresden station and Consolidated Edison declared they would not load fuel and run its reactor at Indian Point until the matter was resolved.
Enter Senator Clinton Presba Anderson (D-N.M.), Chairman of the Joint Committee on Atomic Energy, and Congressman Charles Melvin Price (D-Ill.), another member of the Joint Committee. They introduced legislation that required that each plant obtain the $60 million in available private insurance coverage and included a promise that the government would pay up to $500 million in damages in the event of a catastrophic accident. [As an irrelevant aside, although Sen. Anderson and Lewis Strauss both supported nuclear power, they hated each other's guts. Senator Anderson, along with Senator Lyndon B. Johnson, later sabotaged Strauss' confirmation as Secretary of Commerce. Anyway, sorry for the distraction. I return you now to the diary, already in progress.] ...expected that the operating experience gained would give the insurance industry the actuarial data needed to underwrite all, or at least most, of the risk. Hence, the Act was set to lapse in 1967. The Price-Anderson Act, for all intents, kick started the nuclear industry.
The Price-Anderson Act Today
Despite the expectation that PAA would be a temporary measure, it has been renewed and modified several times since it was signed in 1957 and currently possesses the following features:
Nuclear power plants are still required to obtain the maximum amount of private insurance available (now $300 million per plant).
In the event of an accident where the claims exceed the primary coverage, all nuclear power licensees are obligated to contribute $15 million/year per reactor until the claims have been paid or their individual liability (capped at $95.8 million) has been reached, for a total of about $10 billion.
PAA is “no fault”. Claimants can't seek punitive damages and nuclear plant owners are not allowed to shirk responsible (by suing equipment suppliers, for example).
The Price-Anderson “subsidy” derives from the idea that an accident could result in claims that far exceed the $10 billion the utility industry is obligated to pay and that plant owners are therefore getting an economic benefit from this limit to their liability.
The Price of Price-Anderson
A subsidy, according to the MIT Dictionary of Modern Economics, is a “payment made by the government (or possibly by private individuals) which forms a wedge between the price consumers pay and the costs incurred by producers, such that price is less than marginal cost”. Some, including myself in the past, have argued that, since the existing private liability coverage provided by PAA has proven more than adequate to cover all of the claims made to date (including at Three Mile Island) and there has never been a “payment made by the government”, that Price-Anderson is technically not a subsidy. However, some economists maintain this is muddled thinking. After all, if the government were to offer to pay compensation to victims in the event I had inadequate automobile coverage, I would derive economic benefits (such as reduced car insurance premiums) regardless of whether I ever became involved in a fender bender. So maybe it is a subsidy. For now lets not get wrapped around the axle over semantics. Whether we call PAA a “subsidy” or downgrade it to Plutoid status as a “potential subsidy”, it has a perceived value to the industry. So what is PAA worth?
Although various numbers have been thrown out before them, a pair of economists, Rothwell and Dubin, where the first (that I know of) to attempt a serious calculation and came up with $22 million per reactor per year in 1990. In 1998, Heyes and Heyes suggested some additional terms be included which reduced that number by a factor of 10 to $2.3 million/year. Any statements you see thrown around regarding the aggregate value of the subsidy derived from PAA (assuming they weren't pulled it out of a hat) more than likely can be traced to one of these studies.
Now, there are certain things a person can stick his chest out and say with confidence, like giving the spelling of “Mississippi” or reciting the first 50 thousand digits of pi. The value of Price-Anderson as a subsidy is not one of them. In fact, the point of coming up with these numbers wasn't to end the argument but to begin a dialog. Heyes wrote of the various analyses performed, “The two estimates and the methods used to generate them are, at best unreliable and, at worst, deeply flawed.” By that he wasn't confessing that he and his colleagues had done shoddy work, only that, the methodologies were “experimental” and the results highly sensitive to the underlying assumptions. Rothwell further indicated that the calculations depend crucially on the probability of a worst-case accident and the damages associated with such an accident. Therein lies the rub. The analyses relied on data from the Nuclear Regulatory Commission, but the NRC has practically disowned the studies tied to the consequences of a worst case event (CRAC-II, NUREG-1150 and the rest) as relying on unrealistic assumptions and overstating the consequences. Therefore they would lead to an overestimate of the size of a Price Anderson subsidy. The NRC has an ongoing project (SOARCA, for “State of the Art Reactor Consequences Analysis”) that should provide reasonable answers, but it won't be completed for several years.
Recently I've been wondering about the nature of subsidies. For example, after 9/11 Congress stepped in and capped the liability for the airlines because the industry would have completely collapsed otherwise. Does that count as a subsidy even though it took effect retroactively? Does a policy of intervening when markets malfunction have to be explicitly stated beforehand to count as a subsidy? The Fed did some funky things to intervene and prevent a junk mortgage-driven meltdown of the banking industry. Subsidy? I'm just asking.
One thing is sure, without Price-Anderson the 20% of the United State's electricity that is currently generated from nuclear power would almost certainly be coming from mostly coal and gas-fired plants instead, since they were (and continue to be) the only realistic options for providing baseload generation.