How much of a subsidy is the Price-Anderson Nuclear Industry Indemnity Act?

The answer is perhaps as high as a hundred billion dollars.

First some background. I testified in front of the Senate Environment and Public Works committee in July. In my testimony, “The High Cost of Nuclear Power,” I pointed out the obvious — that nuclear is a mature source of power that has benefited disproportionately from government support to date:

From 1948 to today, nuclear energy research and development exceeded $70 billion, whereas research and development for renewables was about $10 billion. From 2002 to 2007, fossil fuels received almost $14 billion in electricity-related tax subsides, whereas renewables received under $3 billion.

The Price-Anderson Nuclear Industries Indemnity Act caps the liability for claims arising from nuclear incidents. It reduces the insurance nuclear power plants need to buy and requires taxpayers to cover all claims in excess of the cap. The benefit of this indirect subsidy has been estimated at between $237 million and $3.5 billion a year, which suggests that it has been worth many billions of dollars to the industry. It could be argued that the value is considerably larger than that, since the industry might not have existed at all without it: “At the time of the Act’s passing, it was considered necessary as an incentive for the private production of nuclear power … because investors were unwilling to accept the then-unquantified risks of nuclear energy without some limitation on their liability.

One can make a case that such insurance was reasonable for a new, almost completely unknown technology in 1957. Extending it through 2025 is harder to justify. If investors aren’t willing to accept the risks of nuclear energy now, without taxpayers liable for any major catastrophe, perhaps the technology no longer deserves government support.

A certain senior member of the minority known for climate denial just submitted two (silly) written questions to me for the record on this:

[Questions in italics, answers in regular text.]

1. You state in your testimony that nuclear power,is “the beneficiary of some $100 billion in direct and indirect subsidies since 1948.” According to a recent Congressional Research Service report (, the Department of Energy states that the cumulative funding total from 1948 to 2007 is $85.01 billion. How do you explain the discrepancy between the Department of Energy’s total and yours?

A1: There is no “discrepancy.” The study you refer to is “Renewable Energy R&D Funding History: A Comparison with Funding for Nuclear Energy, Fossil Energy, and Energy Efficiency R&D.” That study clearly states that “This report provides a cumulative history of Department of Energy (DOE) funding for renewable energy compared with funding for the other energy technologies.” It is a report on federal R&D. There have been many other direct and indirect subsidies direct toward nuclear power from 1948 through today, including tax credits, loan guarantees, and the Price Anderson Act, which, by itself, it a subsidy worth many tens of billions of dollars (see below).

2. Citing the questionable “Wikipedia” website as your source, you state:
“The Price-Anderson Nuclear Industry Indemnity Act caps the liability for claims arising from nuclear incidents. It reduces the insurance nuclear power plants need to buy and requires taxpayers to cover all claims in excess ofthe cap. The benefit of this indirect subsidy has been estimated at between $231 million and,$3.5 billion a year, which suggests that it has been worth many billions of dollars to the industry.”

However, a Congressional Research Service report states: “Under Price-Anderson, the owners of commercial reactors must assume all liability for nuclear damages awarded to the public by the court system, and they must waive most of their legal defenses following a severe radioactive release ‘extraordinary nuclear occurrence.’ ( Please explain this discrepancy by citing the language fom the Price-Anderson Nuclear Industry Indemnity Act that requires taxpayers to cover all claims in excess of the liability cap.

A2. Wikipedia is not considered a questionable source. Indeed, it has been found to be as accurate as the Encyclopedia Britannica. However, since you (partially) cite the CRS report, “Nuclear Energy Policy,” let me cite the rest of it:

To pay any such damages, each licensed reactor must carry financial protection in the amount of the maximum liability insurance available, currently $300 million. Any damages exceeding that amount are to be assessed equally against all covered commercial reactors, up to $95.8 million per reactor. Those assessments – called “retrospective premiums” – would be paid at an annual rate of no more than $15 million per reactor, to limit the potential financial burden on reactor owners following a major accident. According to NRC, 104 commercial reactors are currently covered by the Price-Anderson retrospective premium requirement.

For each nuclear incident, the Price-Anderson liability system currently would provide up to $10.8 billion in public compensation. That total includes the $300 million in insurance coverage carried by the reactor that suffered the incident, plus the $95.8 million in retrospective premiums from each of the 104 currently covered reactors, totaling $10.3 billion. On top of those payments, a 5% surcharge may also be imposed, raising the total per-reactor retrospective premium to $100.6 million and the total available compensation to about $10.8 billion. Under Price-Anderson, the nuclear industry’s liability for an incident is capped at that amount, which varies depending on the number of covered reactors, the amount of available insurance, and an inflation adjustment that is made every five years. Payment of any damages above that liability limit would require congressional approval under special procedures in the act.

The Energy Policy Act of 2005 raised the limit on per-reactor annual payments to $15 million from the previous $10 million, and required the annual limit to be adjusted for inflation every five years. As under previous law, the total retrospective premium limit of $95.8 million is to be adjusted every five years as well. For the purposes of those payment limits, a nuclear plant consisting of multiple small reactors (100-300 megawatts, up to a total of 1,300 megawatts) would be considered a single reactor. Therefore, a power plant with six 120-megawatt pebble-bed modular reactors would be liable for retrospective premiums of up to $95.8 million, rather than $574.8 million (excluding the 5% surcharge).

Thus, each nuclear reactor has limited liability, and all of the reactors together have limited liability. That is why it is called the Price-Anderson Nuclear Industry Indemnity Act.

A 1992 U.S. Energy Information Administration (EIA) analysis, Federal Energy Subsidies: Direct and Indirect Interventions in Energy Markets, calls Price-Anderson, “A Federal regulation that continues to have a cost-reducing effect on the nuclear power industry.” According to the EIA analysis

These [liability] limits provide a subsidy to the nuclear industry to the degree private insurance premiums paid by operators of individual plants are reduced. In a 1983 study, the NRC concluded that the liability limits were sufficiently significant to constitute a subsidy. However, a quantification of the amount of the subsidy was not attempted. At issue are the probability distributions for various kinds of accidents and valuations of the consequences of accidents, all done on a plant-by-plant basis. The amount of the subsidy would then be found by calculating the differential effect on the insurance premium of imposing the liability limits.

EIA determined that the value of the subsidy to the nuclear industry as a whole was roughly $30 million per reactor per year, or $3 billion annually ($1991). The full subsidy value of the Price Anderson Act from its inception through today thus likely exceeds a hundred billion dollars.

Related posts:

13 Responses to How much of a subsidy is the Price-Anderson Nuclear Industry Indemnity Act?

  1. David B. Benson says:

    How many days in Iraq is that princely sum?

  2. David B. Benson says:

    Off-topic, but this looks to be a potential wedge or three:

  3. Boy you just like picking fights with the pro-nuclear folks, don’t you? The claim that PAA is a subsidy is highly debatable and of course won’t be determined in this thread.

    The nuclear industry, like you noted, would be liable for at least $10B in damages if an accident at a nuclear plant were to occur. The reason why some (mainly antis) consider this a subsidy is because the damage from a nuclear accident is “supposedly” many magnitudes higher than $10B. Therefore, the industry is “supposedly” paying less than what it should.

    To date, the US has seen only one nuclear plant pay damages for an accident – TMI. How much did TMI pay? Well according to the Wikipedia link you use, only $70M to date. The damages, just to be clear, basically went to businesses and individuals who lost revenues and salaries due to the evacuation.

    Now if the worst nuclear accident in the US cost only $70M, I would constitute that the nuclear industry is paying way more than what it should be paying. You can cite all the studies you want that try to guesstimate how much an accident would cost, but based on actual experience, the nuclear industry can easily cover an accident.

    Just to be clear, Congress decides who makes up the shortfall if the damages exceed $10B. They could decide to make the nuclear industry pay more than the $10B liability, not the taxpayers – in your Wikipedia link.

    The conclusions from the 1992 EIA document you cite are based on a 1983 NRC study that the NRC has now said is outdated (disclaimer found here). The NRC is currently updating that study with better data, greater experience and sophisticated computer models.

    The reason the nuclear industry continues to renew the PAA is because we’re about to build a lot of new nuclear plants. Like you said above, the PAA was created to entice companies to build plants.

    Last thing. Doesn’t a subsidy mean that the taxpayers actually pay something? So how is it a subsidy when the taxpayer hasn’t paid one penny due to a nuclear plant accident?

  4. I see that the disclaimer link didn’t come through. Try this.

  5. I guess the link gods are against me today. Here’s the actual address:

  6. john says:


    This is classic risk analysis and what you’re weighing with nukes is the potential (or perceived potential) for a truly cataclysmic accident against even an extremely remote possibility that it would occur. If the event is bad enough, the odds of it happening become less important — and the whole financial structure required to do these things would crash, without some very deep pockets underwriting those risks.

    In short, the cataclysmic accident doesn’t have to happen to require underwriting — it just has to be perceived as possible. And it is with nukes

    So it is a subsidy, in the sense none of these things would ever be built without it — and probably none will with it for sher economic reasons having nothing to do with accidents.

  7. Johan Simu says:

    In what manner is the hoover damn and other hydropower in america insured? What about large chemical industries?

    The insurance situation for large chemical industries and hydropower must be very similar to nuclear. Bhopal and Banqiao clearly demonstrated the huge damage those industries can cause.

  8. Ronald says:

    It’s just as if the government would pay for me to get a new house if there is flooding. That means I don’t have to buy flood insurance.

    For nuclear, it means they don’t have to raise electrical rates to pay for the insurance.

  9. BILL HANNAHAN says:

    Imagine that the terrorist attack on 9-11 never took place. Instead, suppose that on a busy weekday morning at about 11 AM, a design defect in the floor attach fittings of a World Trade Center building caused a mid level floor to collapse on to the floor below it.

    That started a chain reaction collapse that brought the building down. The upper floors tipped into the other WTC tower, triggering the same defect and bringing it down.

    There is no evacuation because there is no warning, and 40,000 people die in 30 seconds.

    A Boeing 747 takes off with a full load of fuel on a long international flight. One minute after takeoff it flies through the wake of another jumbo jet. The turbulence causes an undetected crack in the vertical fin to propagate, and the fin snaps off. The 747 yaws sideways, rolls onto its back and dives down through the roof of a large sports arena holding the national championship basketball game.

    200,000 pounds of fuel atomizes on impact with the floor and erupt in an enormous fireball inside the building, consuming all the oxygen and incinerating 40,000 people on live HD worldwide television.

    In 1997 the EPA determined that a human life was worth $5.8 million.!OpenDocument

    Corrected for inflation, that is $7.6 million now.

    The loss in each case would be $304 billion for human life, plus the property loss.

    The WTC did not carry this level of insurance. Should they have been prevented from constructing those buildings without adequate insurance?

    The airlines do not carry this level of insurance, should the airlines be grounded for lack of adequate insurance coverage?

    Suppose that a biogenetics scientist in a major pharmaceutical industry accidentally develops a virus that is more contagious than the common cold and more deadly than the HIV virus. He contaminates himself and his family, the virus spreads around the world and kills half the population. That would be a twenty five thousand trillion dollar loss. All the money in the world would not cover that loss.

    Should we shut down the entire drug industry and go back to life without medicine because it is not insured for all possible accidents?

    Dam failures have killed 8000 people in the U.S.

    In 1975 a single dam failure in China killed about 30,000.

    Dams in the U.S. are not insured for the maximum imaginable loss. Should we tear down all dams and give up our hydroelectric power, or charge much more per kWh to pay for insurance?

    Coal plants are killing over 20,000 Americans each year. That is a $175 billion loss each year that the coal plants are not paying for, a virtual subsidy.

    You are holding a wedding reception for 150 people in your home. An F5 tornado sucks your home and its contents up to 1,000 feet, grinds it into small pieces, and deposits the mess in a field 2 miles away, killing everybody.

    The tornado loss is $1.14 billion plus the property loss. Are you carrying that much liability insurance on your house? If not, should you be denied the privilege of owning a home?

    If we required every corporation and individual to obtain insurance coverage for the worst possible event no matter how unlikely, we would have no civilization at all.

    The Price Anderson Act requires that the utilities provide $10 billion in insurance coverage without cost to the public or government and without fault needing to be proven.

    It covers power reactors, research reactors, and all other nuclear facilities.

    It was renewed for 20 years in mid 2005, with strong bipartisan support, and requires individual operators to be responsible for two layers of insurance cover. The first layer is where each nuclear site is required to purchase US $300 million liability cover which is provided by two private insurance pools.

    The second layer is jointly provided by all US reactor operators. It is funded through retrospective payments if required of up to $96 million per reactor per accident collected in annual installments of $15 million (and adjusted with inflation). Combined, the total provision comes to over $10 billion paid for by the utilities. (The Department of Energy also provides $10 billion for its nuclear activities.) Beyond this cover and irrespective of fault, Congress, as insurer of last resort, must decide how compensation is provided in the event of a major accident.

    More than $200 million has been paid by US insurance pools in claims and costs of litigation since the Price- Anderson Act came into effect, all of it by the insurance pools. Of this amount, some $71 million related to litigation following the 1979 accident at Three Mile Island.

    American Nuclear Insurers is a pool comprised of investor-owned stock insurance companies. About half the pool’s total liability capacity comes from foreign sources such as Lloyd’s of London. The average annual premium for a single-unit reactor site is $400,000.

    Two teenage brothers are home alone. They break into the liquor closet and find a half gallon of tequila. The older boy challenges the younger brother, “Bet you can’t drink the whole bottle”. “Yes I can” says the younger boy, and proceeds to start chugging. He passes out without finishing it, losing the bet, and within the hour looses his life.

    This establishes that 64 oz. of tequila is a lethal dose. The Linear No Threshold (LNT) model says that if 64 people each drink one ounce of tequila one of them will be dead within the hour.

    This is how we calculate the risk of low level radiation.

    60 years of studying the effects of radiation has still not proven low level radiation to be harmful or beneficial. We can say with absolute certainty that the health effects of low level radiation are very small compared to other risks we accept without much thought.

    Google “radiation hormesis” for an interesting debate, or try this.

    The Chernobyl accident exposed millions of people to a small dose of radiation. The estimates of the number of deaths from Chernobyl over the next 40 years range from 4,000 (IAEA), to 100,000 (Greenpeace), based on the LNT theory.

    If radiation hormesis turns out to be valid the Chernobyl accident may prevent thousands of cancer deaths.

    The Chernobyl reactor had design defects that, combined with gross operator error, allowed it to go rapidly to 100 times the design power level, creating a powerful steam explosion that tore the roof off the building and dispersed fuel. It could never have been licensed in the US.

    If it had an appropriately designed containment building for that reactor design, the release would have been minor.

    Modern reactors have improved instrumentation and control systems, passive safety systems and strong containments designed to contain a full meltdown.

    Nobody is going to build another Titanic, or a De Havilland Comet, or a Chernobyl reactor.

    I cannot think of any industry that handles insurance coverage as well as nuclear power.

    I support the repeal of Price-Anderson and treating nuclear power like other industries.

  10. Cyril R. says:

    I stopped reading Bill H when he was trying to determine the value of human life.

    You’re a despicable propagandist Bill.

  11. Mark Walker says:

    I made it a little further, but Bill lost me at “radiation hormesis.”

    Still, I think Bill’s thought experiments are interesting and worth thinking about while keeping in mind that human ideas about risk and confidence are rarely as rational as seem assumed in Bill’s examples. Risk (insurance) and confidence (banking) are perceptions, and until human perceptions can be shifted we will have to live with them as they are.

    On the point of, “Nobody is going to build another Titanic, or a De Havilland Comet, or a Chernobyl reactor.”, one should look into Henry Petroski’s works addressing engineering failures.

    Success through Failure: The Paradox of Design
    Design Paradigms: Case Histories of Error and Judgment in Engineering
    To Engineer Is Human: The Role of Failure in Successful Design

    (BTW, the Titanic doesn’t really apply. It wasn’t designed to be an iceberg breaker.)

    I would agree that no sane person is going to build another monumental failure on purpose, although the continued use of light-water reactors makes one wonder…

    Unfortunately, not a single one of us, nor any collective group, has complete enough information or future knowledge sufficient to avoid such an outcome in the future for all activities. And, since the complex activities are more at risk from these deficiencies, we need to worry more about how complex things fail. There needs to be much more Systems Theory involved in the NPP debate and planning as we proceed.

    How did you like the recent construction stop on the Flamanville EPR caused by a contractor that took it upon himself to space the reactor foundation reinforcements differently than shown on the plan? And that was just one of the issues:

    Can’t get the concrete right. Trouble with the metallurgy. Sure sounds like the “good old days” of nuke plant construction we all came to know and love. Hey ASN, are we having fun yet?

    Guess what? They decided it was OK to proceed. What an unexpected outcome on a project that was too expensive to stop.

    Petroski shows, among other things, that it is typical for humans to scale up solid working designs without redoing the analysis but rather introducing additional safety factors. The process tends to lead to stupendous failures and loss of life and/or property as the scaling proves untenable.

    Nuclear plants “on steroids” (with apologies to Bill Wright on anyone? For myself, I’ll pass until we’ve stopped lying to ourselves about costs and benefit analysis.

    Want a nuke example of my “Petroski” argument? Look no further than the arbitrary lifetime expectancy change for currently operating nuclear power plants. Sure there’s a program to support the extension beyond the designed 30 year lifespan, but those actions are taken under duress of not having a replacement for the electrical power. How do *you* spell conflict of interest? Why should anyone believe safety hasn’t been eroded or compromised by the change?

    Heck, I wonder when we’ll get around tojust doing it the way the Russians are doing it:

    Environmental laws? We don’t need no stinking environmental laws!

  12. msn nickleri says:

    Environmental laws? We don’t need no stinking environmental laws!

  13. Jim Newberry says:

    “The benefit of this indirect subsidy has been estimated at between $231 million and . . . a year”

    Low ball statements like this from what is primarily a weapons agency (DOE) that was originally set up to promote “Atoms for Peace” should be questioned. Free market risk analysis from an insurance company might show this to be orders of magnitude too low (even for the upper end estimate).

    Price Anderson represents a destruction of free market capitalism by indemnifying a group of people to do what would otherwise be prevented by free market risk assessment. The 1950’s military agenda for this (electric companies all refused to invest originally) may have been to use the “Atoms for Peace” program to change postwar attitudes of the horror of Hiroshima and Nagasaki so the US Treasury could be tapped for what became tens of thousands of nuclear weapons. Even with direct and indirect taxpayer subsidies atomic fission reactor construction has been dead in the US for decades due to exposure to market forces and democracy. Price Anderson is an assalt on American ideals due to its neofascistic nature and should be repealed.

    A further point, the $10 billion discussed above is approximately the valuation of current proposals for a two reactor plant, so they cover their own investment. A meltdown (loss of coolant accident) can contaminate tens of thousands of square miles representing trillions of dollars of realestate. We should not indemnify a human activity so dangerous and uneconomical.