Underwriter Certification of Nuclear Power, 2.0:

Underwriter Certification is slowly evolving. We now have a 100 page manual. But I don't expect many to read it. This post is a distillation. It supercedes earlier posts on this regulatory system.

Figure 1. Sultana boiler explosion, 1600+ killed

Balancing Benefit versus Harm

The core problem in the regulation of hazardous, beneficial activities is finding the right balance between safety and economy. If the regulation is too lax, the result will be disproportionate harm. In the worst case, the technology will be rejected and the benefits foregone. If on the other hand regulation makes an activity overly costly, resources will be expended which could be better allocated elsewhere; and in the worst case, the technology will be priced out of existence.

International Trade and the Classification Societies

Figure 2. The Lutine Bell at Lloyds of London. The bell is rung every time an overdue ship is reported: once if the ship was lost; twice if the ship was found. In 1799, the Lutine was lost with a large amount of gold bullion. The bell commemorates Lloyds ability to pay off even very large claims.

This issue probably emerged first in ocean transportation. The activity was both highly beneficial, and at the time highly hazardous. The solution was the Classification Society system. Shipowners needed insurance. The insurers needed to understand the risk they were insuring. So they set up ship inspection services called Classification Societies. The Classification Societies, usually shortened to Class, not only inspect ships, they set the rules by which ships are built, and certify that a ship complies with those rules. To do this, they recompute all critical design calculations. A Class employee signs off on every drawing. During construction, a Class inspector witnesses and signs off on a long list of tests and trials. For this service, the Classification Society charges about 2% of the price of the ship.

There are roughly a half-dozen major Classification Societies. They are paid by the shipowners. The Classification Societies must compete for shipowners. This sets up a balancing mechanism. If a Classification Society is too strict, it loses business. If a Classification Society is too lax, its certification becomes meaningless and won't be accepted by the insurers.

Overall the Classification Society system has done an excellent job of delivering the benefits of international trade. The real cost of transporting goods across oceans has decreased by multiple orders of magnitude. And seaman safety has improved markedly, albeit from abysmal levels.\cite{tromedy}

High Pressure Steam and the Boiler Inspection Services

Figure 3. Sultana leaving Memphis before explosion.

With the advent of high pressure steam in the 19th century, all sorts of formerly unimaginable wonders became possible. Trains, ocean liners, electricity. At the same time, a whole new danger was unleashed on the public. In the mid-19th century, fatal boiler explosions were running at better than 100 per year.\cite{thurston-1903} Between 1837 and 1878, there were at least 10 steamboat explosions that killed 20 or more people. The worst of these occurred on the Mississippi near Memphis on April 27, 1865. The steamboat Sultana was badly overloaded with Union soldiers on their way home at the end of the Civil War. At least 1600 were killed when 3 of her 4 boilers exploded.\cite{spignesi-2002}.

This prompted several civic leaders in Hartford, Connecticut to form the Hartford Steam Boiler Inspection and Insurance Company in 1866. The deal was pass our inspection, and we will provide reasonably priced insurance for your boiler. Boiler explosions became increasingly rare, despite order of magnitude increases in steam pressure and temperature.

In Germany, at almost exactly the same time, a group in Mannheim formed the Gesellschat zur Uberwachung und Versischerung vor Dampfkesseln (Steam Boiler Inspection and Insurance Association) in response to a boiler explosion at a brewery.¹ The insurance and inspection functions soon developed into separate entities. In Germany, the inspection services were called TUV's (Technischer Uberwachungsverein).

Like the Classification Societies, the TUV's must compete with each other for the inspection business. Like the Classifications Societies, they must find the sweet spot between being overly strict and overly lax.

The system has worked well, delivering the benefits of high pressure steam, allowing technological improvement, --- the Sultana's boiler pressure was 10 bar, modern boilers are close to 300 bar improving efficiency by more than a factor of 3 --- with eventually an excellent safety record.

Both the Classification Society and TUV systems have built in balancing mechanisms. They don't always get the balance just right, but they cannot stray too far from the sweet spot. Underwriter Certification is a variant of the successful Classification Society and TUV regulatory systems.

Figure 4. UK boiler explosions, 1866 to 1908.

Underwriter Certification

Underwriter Certification(UCert) is a market based regulatory system. It harnesses human nature, rather than pretending bureaucrats are self-sacrificing saints. It depends on competition to balance economics and harm, competition among vendors, competition among underwriters, and competition among Certification Societies. But it is not a free market system. Federal and local governments play critical roles in making the system work. UCert assigns the following roles to the four major players: the Feds, the local government, the underwriters, and the Certification Societies.

National Government

At the Federal level, UCert requires the government to:²

1. Promulgate and enforce a firm, fixed table of compensation payments.

The unique hazard associated with nuclear power is the potential for a large release of radioactive material. Anyone who suffers true harm from such a release must be compensated. This is not just a matter of equity. Such compensation internalizes this harm, ensuring that the potential harm is part of the developers' design and operating calculus.

The compensation must be based on a generous estimate the Lost Life Expectancy from the dose profile each individual incurred, according to a reasonably realistic dose-response model, and nothing else. The exposure payment would be both automatic, no fault, not subject to dispute, and exclusive. There would be no requirement to demonstrate any harm. The UCert Manual describes the compensation system in considerable detail.

To make this work, each plant will be required to have a network of radiation sensors throughout the surrounding area. This grid will be maintained by a federal Nuclear Monitoring Agency (NMA). In the event of a release, the NMA will be responsible for estimating each individual's maximal dose rate profile, distributing the resulting compensation payments, and collecting the money from the insurers. Once again the details are in the The UCert Manual.

The NMA would also measure plant boundary dose rates, and enforce the plant boundary dose rate triggers, including collecting any fines. See below. This must be on a strict pass/fail basis. There can be no judgement calls.

2. Specify liability cap methodology.

The Federal government shall specify the methodology by which the liability cap for any particular site will be determined. The goal is a reasonable, worst case, third party casualty cost. This must be done transparently by elected officials, not bureaucrats, using the table of compensation payments. The actual number for any site will depend on the local population distribution, geography, and weather conditions, as well as the plant size. The cap can have a well-defined inflator; but, once the rules are set, they cannot be changed for existing plants.

3. Require iron clad liability insurance for the cap amount.

If a plant's insurer cancels the insurance and the plant is unable to come up with a replacement, government shuts the plant down. If necessary, it can commandeer the plant and take over the operation.

If the compensation cap is exceeded in a casualty, then the compensation payments above the cap shall be paid out of public funds, spreading the harm across all taxpayers.

4. Prosecute alleged negligence.

The tort system must be eschewed. Only government entities would be allowed to prosecute individuals or sue the plant. In the event of a conviction, the proceeds from any fines or penalties go to the local and state government, not individuals and their ambulance chasers. Insurance against such fines would be illegal. Simple negligence can result in bans from nuclear industry employment.

5. Set food and land contamination limits.

Set the legal contamination limits above which produce must be taken off the market. Set the contamination limits above which the land cannot be farmed or grazed until the contamination falls below the limit. These limits should be set by elected officials, not unelected bureaucrats. Once set they cannot be changed for existing plants, unless the plant is compensated for any additional cost.

6. Set Business Compensation Shutdown Level.

As long as the dose rate at a business' location is above this level, a business can shut down and be compensated for the loss of profits. Similarly, employees of the business can refuse to show up for work and be compensated for the loss of wages.³ This Shut Down Level cannot be changed for existing plants, unless the plant is compensated for any additional cost.

7. Require any property insurance to have large deductibles.

Goal is to make sure shareholders suffer if plant screws up. Utility goes bankrupt and the shareholders are wiped out, before rate payers are asked to bear any of the costs of a casualty.

8. Require portion of top management compensation to be deferred.

These will the first funds used in the event of a release. Goal is to make sure top management suffers if plant has a problem.

9.Facilitate prototype testing

For untested technologies, the insurers will almost certainly require intensive, full scale, stress testing of prototypes. The federal government should encourage this by providing a remote, testing facility on a user pays basis. Each tenant would pay rent and other usage fees based on how much of the facility's land and services it required. Each would build its prototype entirely with its own money. Each would also be required to leave its site in an approved condition at the end of its tests.

The feds also set the bond that must be posted to be allowed to test. The bond will be based on a reasonable, worst case, third party casualty cost. It can be based on a 100% release of the volatile fission products. But it must use a reasonably realistic dose-harm model in estimating the Lost Life Expectancy associated with this release. For sufficiently remote test locations, the amount of this bond should be affordable. The bond should also cover the costs of decommissioning and removal.

The government does not decide who gets to test. If the bond is set properly and a developer is able to post that bond, he should test. If he can't, he should not.

10. Do not stifle competition; enforce it.

Foster multiple underwriters. Foster multiple Certification Societies. Foster multiple vendors. For the system to work, there must be competition at every level. Aggressive application of anti-trust and price fixing legislation will be required.

11. Set plant boundary dose rate limits.

These limits are triggers whose violation indicates something is not working right, and corrective action, penalties,and/or other intervention is required. These trigger levels need not, and usually are not, harmful dose rates. In most cases, they are and should be set far below the dose rates at which any radiation health effects have been reliably observed. The government must make this distinction clear.

If a government decides to change any of the trigger levels, any existing or under construction plants must either be grandfathered or compensated for any additional costs imposed on the plant by the change.

12. Provide protection from military attack.

Providing protection against national or sub-national group attacks is the function of the nation's military. This should be true of all high value facilities, many of which such as the transmission grid are far more vulnerable than nuclear power plants. Deploying a small army at each plant is an extremely inefficient way to attempt to protect the plants, and does nothing to prevent the real threat: missile and aircraft strikes. This can only be done by the country's military.

Nuclear plants should be required to provide only normal industrial security, similar to large chemical facilities such as Bhopal or large dams such as Banqiao; both of which are more vulnerable and have at least as much potential for widespread harm as a nuclear plant.

If a release is caused by war or a terrorist attack, the compensation scheme is the same, but the insurance and cap do not apply. Instead the compensation is paid out of the federal treasury. The cap if any will be set by federal lawmakers on an ad hoc basis. This is consistent with the national government's responsibility to protect the nation from foreign and domestic enemies.

One function the federal government does not do is approve or disapprove an individual plant. That is the role of the local government where the plant is located. There should be no federal incursions into state and local sovereignty, such as Environmental Impact Statements.

Local Government

No power plant, nuclear or otherwise, or any large industrial facility should be located in a community that does not want it. Each such community should balance the benefits and costs of such a development and make its own decision. The local community will see only a tiny proportion of the benefit to society as a whole of the cheap, reliable, zero pollution, and near zero CO2 electricity that the plant will provide. But it will bear the brunt of any problem at the plant. It should require compensation for taking on this risk.

This compensation can take many forms including:

1. Hundreds of good, steady jobs.

2. Property and other tax revenues.

3. Parks and other public facilities in the plant's buffer zone, which the local government should set.

4. District heating in cold locales.

5. Here's my favorite. Shares in the plant's ownership and therefore a portion of the plant's profits. This would give the community access to everything a shareholder sees and some say in the choice of directors, and indirectly in the choice of management.

What we need is auctions. Local communities interested in hosting a plant would prepare a package, delineating what they would require in order to agree to host a plant. They could ask for anything they want. The only requirement is the agreement is irrevocable for the life of the plant. The community that offered the most attractive combination of site and compensation package would be the winning bidder.

Underwriters

The insurers play a central role in this system. They want to write business, but only profitable business. They require that a Certification Society acceptable to the insurer confirm that a plant meets all the Certification Society requirements. Based on the Certification Society reports the insurer decides whether to insure the plant and at what premium. If the Certification Society reports that a plant is no longer in compliance with its requirements, the insurer cancels the insurance. The insurer can also adjust premiums according to performance.

If prototype testing is required, the insurers provide the testing bond. They will do this only if they are convinced that selling this bond will be profitable.

The policy would require the insured to report any significant defects or upsets to his insurer and his Certification Society on pain of losing his insurance. These reports could effect his premium; but the main purpose is to alert the other plants and vendors of a potential problem, if indicated. Insurer and/or Certification Society associations would develop to facilitate this sort of information transfer, as they have in other industries.

Certification Societies

Certification Societies set the rules with which a plant must comply including the component tests during construction and the trials before acceptance, witness those tests and trials, and, if all goes well, certify that the plant is in compliance with the Certification Society's rules. This is an ongoing process. The Certification Society will monitor operating plants, periodically inspect the plant and, if it is not satisfied, report to the insurer that the plant is no longer in compliance with its requirements.

Certification Societies are hired by a plant in order to obtain the certification required to buy insurance. This sets up a balancing mechanism. If a Certification Society is too strict, it loses business; too lax and its approval will not be accepted by the insurers.

For a new design, the Certification Society will write the test protocols, and oversee the prototype testing. This will be a step by step process in which the next step is only undertaken if the Certification Society is comfortable with the results at that point.

Overall, Underwriter Certification is very similar to the system by which we regulate ocean transportation and high pressure steam. The main difference is in the radiation exposure compensation system. The Feds dictate how the compensation shall be calculated and require insurance for a worst case exposure under that compensation system. Even here we have somewhat similar precedents such as the Montreal Convention which governs compensation for commercial airline casualties.

Is Underwriter Certification Feasible?

UCert at Fukushima

To address this question, the Gordian Knot Group (GKG) applied UCert to the Fukushima release, using the rules delineated in the Underwriter Certification Manual. The UCert Manual uses the Sigmoid No Threshhold (SNT) radiation harm model which is also described in the manual.

Even though the sensors failed when the grid went down, we have quite a bit of dose rate data from Fukushima. We have a combination of ground based and aircraft base measurements which can be extrapolated back to the initial numbers, since we know the isotopic compensation of the release.

From this data, GKG estimated the initial max and min ambient outdoors dose rates for each of 12 towns near the plant. Each town's population was divided into 11 equally sized groups, the lowest dose group had a maximal dose rate of the minimum ambient dose rate, the next to bottom had a maximal dose rate of the minimum plus 10% of the difference between the max and min ambient dose rate, the next group added another 10% and so on until we get to the top group with a maximal dose rate equal to the highest ambient dose rate for the town. I did the same thing for the businesses in the town.

We have a good idea of the Fukushima decay rates. Once we had an estimate of the initial maximal dose rate for each group in each town, it was straightforward to estimate the corresponding dose rate profile for each group, and the corresponding Lost Life Expectancy (LLE) and compensation Table 1 summarizes the results.

On this table GKG also estimated the actual SNT LLE's under the assumption that the Location Factor for the first 40 days was 0.10, after which the Location Factor rose to 0.20, as people resumed their normal habits.⁴ The actual LLE's are far below the maximal LLE's, on which the compensation is based, because the factor of more than 5 reduction in dose rate is combined with the better than quadratic relationship between dose rate and cancer incidence.

Under these assumptions, the total actual SNT LLE is about a year and a half which is concentrated almost entirely in the three towns closest to the plant. The total maximal LLE is about 63 years leading to a total Radiation Exposure Compensation of about 8 million dollars.

The GKG estimated the Earnings Compensation by assuming the average wage and worker fractions shown in the 3rd and 4th leftmost columns. This came to 24 million dollars. The Lost Profits will be a fraction of this, since for most businesses the payroll is larger than the profits.

These numbers involve a great deal of hand-waving and a string of heroic assumptions. But there is good reason to believe they are in the ball park, and, given the conservative nature of most of the assumptions, probably on the high side. They indicate that in most releases the Lost Earnings Compensation will be much larger than the Radiation Exposure Compensation, which make sense since releases such as Fukushima will result in no detectable radiation harm to the public. Even if the release is well managed, the disruption will be more harmful than the radiation.

The fact that almost all the radiation harm was in the towns nearest the plant indicates the importance of buffer zones. If people had been living right next to the breached reactors, the maximal dose rates and the compensation would have skyrocketed.

To first order the release at Chernobyl was 10 times larger than Fukushima. According to SNT, harm goes at the 2.18 power of the release. A guesstimate of the compensation at Chernobyl would be 150 times that at Fukushima. This would lead to 1.2 billion in Radiation Exposure Claims, and 3.6 billion in Lost Earnings Claims under Underwriter Certification.

Insurability

The non-life insurance market is extremely large, collecting over 2 trillion USD in premia in 2013. Nuclear liability insurance has been a very lucrative business. In the US there is a single provider of this insurance, American Nuclear Insurers (ANI), which lays off the risk to a consortium of underwriters at Lloyds and elsewhere. The annual per plant premium for $500 million in coverage is roughly $1,000,000. Since there are about 70 such plants, a little more than 7 years of revenues covers the entire underwriters' risk. It's an extremely profitable arrangement. Over the last 50 plus years, the ANI consortium has paid out $150 million in claims, $71 million at TMI, while collecting over 4 billion dollars in premiums, and further profiting from the investment returns on that money.

In a normal market, sellers would be scrambling to do more of this business, pushing the price down. And in fact underwriters estimate that ten to fifteen billion dollars of nuclear liability insurance is available, at a cost of 0.1 to 0.2 cents/kWh.\cite{wna-liab} This would be plenty to cover a Chernobyl under Underwriter Certification with SNT based compensation. Given adequate buffer zones, UCERT is feasible.

Practicality

Most of infrastructure for Underwriter Certification is ready to go. The insurance markets exist, as do the Classification Societies and TUV's, both of which have already done nuclear work. In Germany, the state has delegated nuclear plant certification and inspection to the TUV's. The American classification society is called the American Bureau of Shipping (ABS). ABS certified and then inspected the Savannah, the USA's only nuclear power commercial ship. ABS and Bureau Veritas, the French classification society, are currently working with floating nuclear plant developers. The system is already in place. We just need to decide to use it.

Take Away

We could have Underwriter Certification, with a few strokes of a pen.⁵ We could have pollution free, very low CO2, reliable, on-demand, 3 cents per kilowatt-hour electricity. What are we waiting for?

1. The UK moved slightly sooner. In 1859, the Steam Boiler Assurance Company was formed in Manchester.

2. Ideally, UCert would be combined with an incrementally increasing CO2 tax, together with dumping all subsidies and mandates to all sources of power. UCert does not depends on a carbon tax, but a properly designed CO2 tax is essential to a smooth, efficient transition to a low carbon society. Technology specific subsidies and mandates are the opposite of a pollution tax. They introduce market imperfections and inefficiencies. In their most pernicious form, the tax credit, they are a blatant, in-your-face transfer of wealth from the poor to the rich. They are the worst of all possible approaches to the problem.

3. If a business owner decides to stay open despite the dose rate being above the compensation level, she will still get the compensation as if she had shut down. Similarly, if an employee shows up for work under these conditions, he will still be compensated as if had stayed home. During that period, his income will double.

4. The maximal dose rate on which the compensation is based, assumes the individual stands unmasked outside his residence 24/7. The Location Factor is an estimate of the fraction of that dose which he actually receives, mainly because of time spent indoors. During plume passage, masking up with N95 masks can cut the actual dose by up to a factor of 100.

The compensation program assumes no evacuation. People would be free to evacuate if they wish; but their compensation would not change.

5. Of course, those few strokes of a pen will include a complete rewrite of the Atomic Energy Act. The NRC would disappear. It would be replaced by a far smaller Nuclear Monitoring Agency, whose only job would be measuring the dose rates, and enforcing compensation in the event of a release.