January 18, 2008

Energy costs with externalities

Anti-nuclear people like to talk about the lack of complete business insurance coverage of nuclear power plant accidents. They ignore the uncovered external costs of other energy sources. They also ignore the disproportionate subsidies for wind and solar power. I will review research related to the subsidies, research, external costs and fatalities. I adapted this from a discussion I had on the Oil Drum.

I have a new article with a closer look at Feed in Tariff support of renewable energy

Geoffrey S. Rothwell of the Stanford Institute for Economic Policy Research examined the nuclear insurance issue

In economics, a subsidy 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" (The MIT Dictionary of Modern Economics, 4th Edition, 1992). Here, the "consumers" (of insurance/indemnification) are firms in the nuclear power industry and the "producer" (of insurance/indemnification) is the federal government. However, there is no subsidy payment unless there is an accident and damages are above the PAA liability limit. Because there is no payment, there is no "direct subsidy," although there is a
potential (or expected) subsidy.

Opponents of the PAA have used these estimates to argue for the ending of the "PAA subsidy" to the nuclear power industry. Without questioning the probability distribution assumption, they have followed the advice in Heyes and Heyes (2000, p. 99): "The implications for how anti-nuclear lobbyists should go about persuading regulators and governments that the extent of the subsidy which current law confers is unacceptably high are that it is likely to be more fruitful to ‘argue up’ consequences rather than probabilities." This has been done by claiming that the costs of a Chernobyl-like accident in the US would be more than $300 billion, without any discussion of the probability of such an accident in the US. See, for example, By focusing on one assumption (consequences) without considering other assumptions (probabilities), the anti-nuclear argument is incomplete.

In regards to government support and subsidies for different energy sources
A 2002 Cato Institute report showed that in the previous 20 years renewable technologies received $24.2 billion in US federal R&D expenditure, compared with $20.1 billion for nuclear and $15.5 for coal (adjusted 1996 dollars). The result of this was minimal electricity contribution from non hydro renewables, and 20% and 50% respectively contribution from nuclear and coal.

A 2006 study from Management Information Services on The US Energy Subsidy Scorecard showed that total federal incentives (of which R&D expenditure is only a part) from 1950 to 2003 totalled $63 billion for nuclear power, $111 billion for renewables, $81 billion for coal and $87 billion for natural gas (2003 dollars), lining this up against the resultant contribution to US energy.

Government support versus actual delivered electricity

R&D versus electricity generation

Focusing on R&D alone over 1994-2003, the study showed coal got $3.9 billion and nuclear $1.6 billion - both commensurate with their contribution to US electricity, while renewables other than hydro received $3.7 billion - vastly more than their foreseeable contribution.

Germany applies a mixture of incentives for renewables, such as a feed-in tariffs. The average feed-in tariff apart from solar PV is 8.5 c/kWh, or 16.4 cents including solar PV in 2006 (solar PV being 49 cents). The combined subsidy from consumers and government totals some EUR 5 billion per year - for 6% of its electricity.

Germany also provides producer subsidies to its coal industry amounting to EUR 68 per tonne for 34 Mt coal in 2000 - total EUR 2.3 billion.

EU energy subsidy analysis from 2004

External energy costs totals for energy. In the notes a discussion of the hypothetical severe nuclear accident. Chernobyl cost $370 billion. Equal to 10-20 years of excess coal or oil costs for the EU15 only. 2-6 years for world (US, China, India etc...) excess coal or oil costs.

Paul Scherrer Institut (swiss) for the study of energy costs with impacts and externalities included

333 final report on energy external costs

External energy costs added to costs of energy. High estimate on top and low estimate below. Nuclear price looks good.

Top ten energy related events for evacuees and costs

Chernobyl is put at US (2000) $370 billion. $6 billion for three mile island. when compared to the annual higher external costs for coal and oil. Then 10-20 years of EU only external costs balances out one Chernobyl. The coal and oil damage for the US and china and other non-EU countries would balance out the one time Chernobyl Chernobyl 3-5 times faster. Chernobyl happened once in 50 years with a particularly dangerous reactor.

The risk assessment for the modern reactors that we would be building should be considered This is important - no one is suggesting that we make more Chernobyl style reactors. Even all of the old Chernobyl style reactors now have containment domes. This would limit almost all of the worst case scenarios to Three Mile Island level accidents.

The AP1000 has a maximum core damage frequency of 5.09 x 10-7 per plant per year. The Evolutionary Power Reactor (EPR) has a maximum core damage frequency of 4 x 10-7 per plant per year. General Electric has recalculated maximum core damage frequencies per year per plant for its nuclear power plant designs:

BWR/4 -- 1 x 10-5
BWR/6 -- 1 x 10-6
ABWR -- 2 x 10-7
ESBWR -- 3 x 10-8

This means you would multiply the 1 in 100,000 to 3 in 100 million chances against the potential costs. All insurance risks and costs are calculated in this way. (Frequency times cost)

In the footnotes of one of the references there is a high range estimate of 5.5 trillion euro for a worst case damage event. How could a 5.5 trillion damage event happen ? I do not believe it is possible. Even blowing up like a nuclear bomb (which is impossible since the uranium is not pure enough) reactors are not close enough to population centers with the blast radius.

The world only has $140 trillion in financial assets.

Even estimates for the nuclear bombing of New York do not have direct economic damage at that 5.5 trillion level.

A worst case analysis for a terrorist attack on Indian point reactor. Has damage of 1.1 trillion to 2.1 trillion, where everything goes exactly to maximum damages. (700 to 1.5 trillion euro). All reactors are not near important financial centers, so the value of the surrounding areas would be less for all other reactors. However, the analysis has been that nuclear reactors would not release radiation if hit by a plane. So even that worst case scenario would not happen or it would involve several dozen people planting massive explosives or firing missiles to breach the containment dome while at the same time causing the reactor to meltdown before they could be stopped.

I believe the core damage event not breaching containment would cost $6 billion max (loss of reactor,that level of damage is covered under the insurance) and for the current reactors and processes 1 in 100,000. So once every 200 years for a slightly larger than current reactor fleet.

5 billion euro / 200 years = 25 million euro/year (covered under paid insurance)

1 in a million for some kind of containment breaching event
500 billion euro / 2000 years = 250 million euro/year
500,000 euro/year/reactor

A discussion by Heyes in 2003 about his 1998 estimate of nuclear insurance.

Heyes and Liston-Heyes noted an error in the way in which Dubin and Rothwell interpreted current insurance arrangements, and reapplied their methodology corrected for the reinterpretation. Heyes and Liston-Heyes’ correction reduced the estimates
of the subsidy substantially to $2.3 million/reactor/year.

I [ANTHONY HEYES] will let you in on a little secret: The two estimates and the methods used to generate them are, at best, unreliable and, at worst, deeply flawed. I can say that because I am one of the authors. I know squat about nuclear power. Do not get me wrong, the two papers are competent pieces of academic research and they deserved to be published in the reputable peer-reviewed academic journals in which they appeared. But the approach that they utilized is very much an experimental one, and one whose results can be highly sensitive to changes in underlying assumptions.

New reactors are 20 times safer. So one event in 4000 years for the same number of reactors.

Immediate fatalities counts by energy source. Latents for Chernobyl not counted [Latent estimate for Chernobyl 200-4000 total] and latents for fossil fuel air pollution not counted (4.5-6 million/year, World Health Organization statistics). Latents for oilwars not counted. Some of China's immediate coal deaths not counted.

Some anti-nuclear people talk about the importance of the speed of release of radiation. I say that speed of release of pollutants by itself is meaningless. It is what happens damage and death wise with the release. If speed effects the damage and death then it matters, but speed by itself is meaningless. Plus all pollutants need to be considered not just radiation.

The Ivy Mike nuclear bomb test of 1951 released 100 times the radiation of Chernobyl and it was released faster, but no one died from Ivy Mike.

I consider the 1.2 million global deaths from cars an outrage. More should be done to reduce those fatalities. Systematical adjustments like getting off of coal is a factor here. 6 billion tons of coal is moved (rail and trucks). Getting off of coal would reduce traffic accidents by about 3-10% and freight rail by (20-40%).

Social risk: A lot of wind power could have environmental effects. Drying out of peat bogs. Enough wind could effect weather. [Note: I am not against wind power, but people should not pretend that solar and wind are pristine and without some issues. They are very good, just like nuclear is very good. We need to not get lost in debating details of solar, wind and nuclear and then forget of the orders of magnitude difference for coal and oil.]There is the clear deaths from coal and oil as I have described, for some reason this is socially acceptable.

Uranium from seawater is often ridiculed, but it should not be. They dip the polymer adsorpant netting into the ocean and let the ocean currents flow by. They then pull out the netting and extract the uranium. It is like fishing with nets. You would not pump the water because the water is already moving.

If 2g-U/kg-adsorbent is submerged for 60 days at a time and used 6 times, the uranium cost is calculated to be 88,000 yen/kg-U, including the cost of adsorbent production, uranium collection, and uranium purification. When 6g-U/kg-adsorbent and 20 repetitions or more becomes possible, the uranium cost reduces to 15,000 yen. This price level is equivalent to that of the highest cost of the minable uranium. The lowest cost attainable now (2006) is 25,000 yen with 4g-U/kg-adsorbent used in the sea area of Okinawa, with 18 repetition uses. This is about $220 per kg (114 yen to 1 US Dollar in 2007) The price of Uranium is currently in the $80-120/kg range.

Note: that is the one thing is that we will only have to go to uranium from seawater in any big way in 50-500 years. The timing depends upon how quickly we make nuclear reactors that 50-100 times more fuel efficient and how well we develop more economic sources of Uranium.

It does take 1000 three MW wind turbines to equal one single 1 GW nuclear reactor. It takes ten times the steel and concrete to make those wind turbines. Plus the wind turbines and blades need to be built in massive factories. The wind turbines are 30-40 stories tall and the blades are larger than the wings of a jumbo jet.

More papers examining energy subsidies


DV8 2XL said...

I read this blog religiously and rarely comment as your articles are some of the most well written and informative on the web.

This entry is an especially good example, and will go into the list I keep of places to point to when my opponents play the insurance card in the nuclear debate.

Keep up the good work, just because we don't comment doesn't mean you're not being read.

bw said...

Thanks I appreciate your comment and citation.

I hope that by getting the right information and context out that better decisions, debates and discussions can be had.

Anonymous said...

Great stuff as always, but I am afraid much of the opposition is religious in nature rather than rationally based - but we can only battle on, and your technical know-how means that you are eminently qualified to lead the charge.

bw said...

Thanks Dave. I also appreciate your insiteful comments.

btw: one way that those who read and appreciate my articles can help would be to submit them to stumbleupon, reddit, digg and other sources. (Links are at the bottom of each article.)

Stumbleupon and reddit seem to be the most receptive to bringing traffic. Digg or slashdot would be but it seems tougher for my type of article to get traction on those populist sites.

thanks again

Eadwacer said...

I'm afraid I can't agree with the implied criticism in the first two charts. One purpose of R&D funding is to bring marginal technologies to the point where they can make major contributions. Complaining about overfunding solar or wind power today is like someone complaining about wasting money on coal research in the 17th Century, when what was needed was more efficient wood stoves.

bw said...

I am not complaining about overfunding of solars and wind, but I am saying that the complaint about society bearing the uninsured cost of a nuclear accident is invalid. Society pays for hydro, oil and coal externalities and those are higher than nuclear.

Coal companies and utilities do not pay for acid rain or air pollution. They do not pay insurance for it.

Solar and wind get subsidized just as all energy is subsidized.

What I do complain about are those who say that the choice should be to only fund solar and wind and geothermal when those technologies and industries are not ready to scale. It is not just a matter of money, it is a matter of new ideas and improvements taking time to mature. No matter how much money we had spent on developing nuclear in the 1900-1910 nothing would have happened, it was just not ready.

Solar and wind can continue at current levels and even a bit more. I particularly like the kitegen wind system. But we need to push ahead at a unimpeded pace with nuclear because it is ready to displace a lot of coal and oil, which are killing 4.5 -6 million people per year.

Al Fin said...

Nice job, Brian. As a more intelligent and sane Al Gore would put it, "Nuclear energy is our power bridge to the 22nd century."

Brad F said...

Is it just me, or is there a typo in the Issues in Science and Technology article? The graph titled "Distribution of Federal Energy Incentives
among Energy Sources, 1950-2003" clearly shows oil with the lion's share of incentives, nuclear at about 10% and renewables at about 5%. This contradicts the text on the following page, that renewables have received the second largest subsidy since 1950, although the graph, titled "Federal R&D Expenses for Selected Technologies, 1976-2003" appears to support the claim, until one notices the difference in the years.

I suspect it is correct that renewables have received the second largest subsidy since 1976, but probably not since 1950.

bw said...

The first chart is in regards to energy incentive subsidies and the next one is discussing R&D spending.

So renewables had the second most R&D spending (1994-2003) and were 4th in terms of overall subsidies.

See the original paper which I had linked in the article and provide again for more info

bw said...

An html version of that energy scorecard article

energy scorecard

bw said...

Here is an excellent analysis on Price Anderson that Bill Hannahan put out on

Regarding Price Anderson insurance.

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 giant 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?

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.

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?

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.

Authors of A Solar Grand Plan,

published in Scientific American, propose a solar plan that could be used by terrorists to kill millions of Americans.

Do you think this plan can get insurance? Nuclear does not have to be perfect, just better than any other practical option, which it is.

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