Published 30 March, 2023. Last updated 30 March, 2023.

Nuclear power plants can provide reliable electricity generation with low air pollution and low carbon emissions. Some countries have chosen to embrace nuclear power, some countries reject nuclear power, and some have regulatory cludge that makes new nuclear power plants prohibitively expensive.

Some countries accept net costs of billions or tens of billions of dollars per year and thousands or tens of thousands of lives lost per year to resist the technological temptation of nuclear power.

This case study is part of the resisted technological temptations project. The goal of this project is to understand situations where some actor might have expected to capture substantial value from pursuing a technology, but did not as a result of concerns about downsides that would not directly affect that actor.

Nuclear power plants use controlled nuclear reactions to create heat and generate electricity. A heavy element, usually uranium, undergoes a fission reaction and splits into lighter elements, releasing energy. This energy is used to boil water, which drives a steam turbine, which generates electricity.

Different countries have dramatically different relationships with nuclear power. Some countries have embraced it. France produces 70% of electricity using nuclear power. Some countries, like the United States, use nuclear power but have made it prohibitively expensive to construct new nuclear power plants. Some countries previously used nuclear power, but have chosen to abandon it. Germany produced 30% of its electricity using nuclear power in 2010, but will complete shutting down the last of its nuclear power plants in April 2023. Some countries, like Italy, have completely rejected nuclear power. Nuclear power is a resisted technological temptation for some countries, but not for others.

Since the main benefit of nuclear power is electricity generation, calculating the net benefit requires comparisons to other ways of generating electricity.

The financial costs of nuclear power are highly disputed. Comparisons with other forms of electricity generation are difficult because different sources have different cost profiles. For nuclear, most of the cost occurs during construction. We can compare different power sources by their overnight construction cost,¹⁾ shown in Table 1. This does not include the fuel cost, which is important for coal and natural gas,²⁾ or the capacity factor,³⁾ which is important for wind and solar. The levelized cost of electricity (LCOE) combines these factors to make a more fair comparison, as shown in Table 2. There are significant discrepancies between estimates for both of these numbers. Nuclear construction costs have also varied widely over time and between countries, as shown in Figure 1.

Table 1: Overnight construction costs in the US, estimated by the EIA and Lazard. Units are 2010 US dollars per kilowatt of electricity production capacity.

Table 2: Levelized cost of electricity in the US, estimated by Lazard and Devanney. Units are 2010 US dollars per megawatt hour of electricity produced. Note that Devanney’s numbers are from 2014, before the dramatic decrease in the cost of natural gas in the United States. Lazard also estimates the LCOE for maintaining existing nuclear power plants to be \$29/MWh. Devanney also estimates the cost of nuclear power under a better regulatory regime to be \$30-40/MWh.

Figure 1: The overnight construction costs of nuclear power plants in individual countries vs time. Costs have increased in most countries since 1970, except South Korea, although the magnitude of the increase varies widely.⁸⁾

Nuclear power has several other benefits, compared to other electricity sources. Like fossil fuel plants, it can be operated all the time, and so supplies baseload power. Nuclear power plants can also be designed to be dispatchable,⁹⁾ although this reduces their capacity factor.¹⁰⁾ Like renewables, they do not emit carbon dioxide when they produce electricity. Life cycle greenhouse gas emissions¹¹⁾ are comparable to renewables.¹²⁾ Nuclear power also does not release air pollution, which means that it kills far fewer people than fossil fuels.¹³⁾

Nuclear power also has some unique challenges.

Even though coal, for example, kills far more people than nuclear power, many people think that nuclear power is more dangerous. The risk profile for nuclear power is different from coal. Air pollution from coal is continual, and causes increased risk from lung cancer. Nuclear accidents are big, dramatic events, which dominate the news. This does not seem to be the entire explanation. Nuclear accidents are much more infamous than lethal dam collapses¹⁴⁾ or gas explosions.¹⁵⁾ The association of nuclear power with nuclear weapons and a disproportionate fear of radioactivity makes nuclear power seem more dangerous than it is.

The technology associated with nuclear power can also be used to produce nuclear weapons. When more countries use nuclear power, there is more of a risk of the proliferation of nuclear weapons. This does not, however, explain why countries which already are capable of producing nuclear weapons resist nuclear power.

Nuclear power plants require a significant amount of water for cooling. This water does not become radioactive, but it can increase the temperature of local water supplies, unless large cooling towers are built.

Nuclear power plants generate high-level radioactive waste, in the form of spent fuel rods. The volume of this waste is small, but its long lifetime makes it difficult to deal with. Some countries have long-term storage sites,¹⁶⁾ some countries reprocess it to recover new fuel,¹⁷⁾ and some countries store it on site at nuclear power plants.¹⁸⁾

As noted above, different countries approach nuclear power in very different ways. This is a resisted technological temptation for some countries, but not for others.

Countries’ relationship with nuclear power can be placed into four categories: (1) countries that embrace nuclear power, exemplified by France, (2) countries that use and plan to continue using nuclear power, but regulations make it prohibitively expensive to build new nuclear power plants, exemplified by the United States, (3) countries that have chosen to phase out nuclear power plants after Fukushima, exemplified by Germany, and (4) countries that have rejected nuclear power plants for decades, exemplified by Italy.

Nuclear power is definitely not a resisted technological temptation in France.

France embraced nuclear power in the wake of the 1973 oil crisis: “In France, we do not have oil, but we have ideas.”¹⁹⁾ The Messner Plan transitioned France’s electricity generation to being predominantly nuclear. France currently produces 70% of its electricity using nuclear power.²⁰⁾

The Messner Plan was announced without public or parliamentary debate. There have been significant anti-nuclear protests in France, but they have been much less successful than in other countries.

Several other countries also use nuclear power extensively. Ontario, Canada produces 60% of its electricity using nuclear power.²¹⁾ It accounts for over 40% of the electricity generated in Sweden and Ukraine and over 30% of the electricity generated in Finland.²²⁾ South Korea was working towards producing 60% of its electricity using nuclear power, but this has stalled at about 30% after Fukushima and a scandal involving counterfeit parts in 2013. China and India are both building double digits of new reactors, but nuclear power’s share of electricity generation in these countries is still small.

The United States uses nuclear power and does not have plans to stop. It generates about 20% of the electricity in the US. However, almost no new nuclear power capacity is being built. Other than Vogtle Units 3 & 4, which began construction in 2013 and are currently being turned on, no new nuclear reactors have been built which begun construction since 1978 in the US.²³⁾ There is almost a de facto moratorium on new nuclear power in the United States.

The anti-nuclear-power movement in the United States is about as old as commercial nuclear power. In the early 1960s, a nuclear power plant at Bodega Bay, California, near San Francisco, was canceled due to local protests.²⁴⁾ As these local protests grew and extended their influence elsewhere in the country, they created organizations like the Friends of the Earth and the Union of Concerned Scientists. These organizations have not been able to directly defeat nuclear power politically on a national scale, even after a partial core meltdown at Three Mile Island.

Instead, the Nuclear Regulatory Commission added increasingly strict safety regulations, which caused the cost of building new nuclear power plants to increase dramatically. Reactors which began construction in the late 1970s took 2.4 times as long and cost 9 times as much as reactors which began in the 1960s. Especially damaging were changes to the regulatory code made during construction. New plants stopped being built under this stricter, and still frequently changing, regulatory regime. For more details of how this occurred, see discussions by Jason Crawford²⁵⁾ and Brian Potter.²⁶⁾

We are uncertain why this occurred. It might be the result of a failure to align the incentives of the NRC with the goal of building new nuclear power plants. It also might have been an intentional strategy by anti-nuclear activists to slow the development of nuclear power. Distinguishing between these possibilities would require figuring out the motivations of various individuals on the NRC during the 1960s and 1970s, which is beyond the scope of this page.

Estimating the direct cost of nuclear power being too expensive to build requires a bit of a calculation, which can be found in the Appendix. Under a different regulatory regime, the levelized cost of electricity for nuclear power could be $20/MWh less than coal and similar to natural gas. If 20% of the United States’ electricity generation switched from coal to nuclear, this would reduce the direct costs of electricity in the US by \$16 billion/yr. For comparison, the total sales of electricity in the US are about \$400 billion/yr.²⁷⁾

The two main indirect costs seem to be the contribution to climate change, measured by the social cost of carbon,²⁸⁾ and premature deaths due to air pollution. Our World In Data has estimated both of these:²⁹⁾ Nuclear power causes 0.03 deaths per TWh³⁰⁾ and 3 tons of CO2 equivalent per GWh, natural gas causes 3 d/TWh and 500 t/GWh, and coal causes 25 d/TWh and 800 t/GWh. This corresponds to an expected 20 d/yr and \$100 million/yr of climate change caused by nuclear, compared to 2,000 d/yr and \$20 billion/yr of climate change avoided if natural gas is replaced or 20,000 d/yr and \$30 billion/yr of climate change avoided if coal is replaced.

The benefit resisted by the United States’ nuclear regulatory regime is tens of billions of dollars per year and thousands or tens of thousands of premature deaths per year.

The regulatory induced cost increases seem to have been the largest in the United States, although something similar occurred in many countries starting in the late 1960s (see Figure 1).

The German anti-nuclear movement has a similar origin to the American one.³¹⁾ The movement began with local opposition to a nuclear power plant in Wyhl, which subsequently developed into anti-nuclear-power organizations, including the Green Party. Regulatory cost increases were not as severe as in the United States, but they were sufficient to stop any new construction from starting after 1982. An SPD-Green coalition government from 1998-2005 decided to slowly phase out nuclear power, although this decision was reversed after Merkel came to power. In 2010, nuclear power provided about 20% of Germany’s electricity.

In 2011, an earthquake and tsunami along the coast of Japan caused nuclear meltdowns and hydrogen explosions at the Fukushima nuclear power plant. The Japanese government has confirmed one death from radiation,³²⁾ and no measurable increase in cancer rates as a result of released radiation is expected, with estimates ranging from zero³³⁾ to a few hundred.³⁴⁾ Massive anti-nuclear protests broke out across the world, including in Germany. In response, Merkel’s government decided to phase out nuclear power by 2022. This has been delayed slightly by the Russian invasion of Ukraine, but the last of Germany’s nuclear power plants will be shut off in April 2023.

The decision to abandon nuclear power in Germany was clearly political, made at the highest level of government. It occurred in response to a highly publicized crisis involving nuclear power. This crisis was vaguely like a warning shot in that its consequences were not that terrible, but it rallied public and political opinion against the risks posed by a technology.

Germany’s decision to reject nuclear power has been extensively debated in public, which makes it easier to estimate the amount of value foregone by resisting the technological temptation. Some of the nuclear power has been replaced by electricity generated by burning coal, especially since the Russian invasion of Ukraine cut off much of Germany’s natural gas supply.³⁵⁾ Cost estimates for this decision can be found in a paper by Jarvis, Deschenes, and Jha.³⁶⁾ Shutting down half of Germany’s nuclear power plants cost \$1.6 billion/yr in additional operating costs, \$1.8 billion/yr due to the social cost of carbon, and caused 1,100 additional premature deaths per year due to air pollution. This compares to \$0.2 billion/yr gained by avoiding nuclear waste and the risk of accidents.

Germany is willing to pay billions of dollars per year and accept about a thousand deaths per year to resist the technological temptation of nuclear power.

Other countries which are currently phasing out nuclear power include Spain, Belgium, Switzerland, and Taiwan. Japan shut off all of its nuclear reactors immediately after Fukushima, but has restarted some of them since Abe came to power in 2014.

Italy has resisted the technological temptation of nuclear power for decades.

In 1987, in response to the Chernobyl disaster, Italy held a referendum on nuclear power. The referendum involved three technical questions, and the anti-nuclear position won 70-80% of the vote for each of them.³⁷⁾ Italy shut down all four of its nuclear power plants by 1990 and has not reopened or built any since.

In 2008, Italy elected a pro-nuclear government which called this decision a “terrible mistable, the cost of which totalled over €50 billion.”³⁸⁾ If we take this as a serious estimate, then Italy has foregone €1.8 billion/yr to avoid using nuclear power. This government started planning the construction of new reactors, but a referendum in 2011 (after Fukushima) saw 94% of voters reject the plan.³⁹⁾

Nuclear power is also illegal in Austria, Denmark, Ireland, Georgia and Uruguay.

A map which summarizes the current status of nuclear power in various countries is shown in Figure 2.

Figure 2: The status of commercial nuclear power in countries across the world. Research reactors are not included. Dark Blue: Operating reactors, building new reactors. Light Blue: Operating reactors, planning new build. Dark Green: No reactors, building new reactors. Light Green: No reactors, new in planning. Orange: Operating reactors, stable. Red: Operating reactors, decided on phase-out. Black: Civil nuclear power is illegal. Gray: No reactors.⁴⁰⁾

Here are a few key takeaways from this technological temptation:

• Political decisions can determine whether or not a technology will be used. Technological development is not inevitable.
• Politicians and the public can react decisively to a warning shot, although this reaction is not always well considered.
• Small groups of advocates can successfully block individual projects, and then build this success into national movements.
• Regulatory cludge can dramatically slow or even stop a technology.

Primary author: Jeffrey Heninger.

Here is an estimate of the amount of direct value foregone by the prohibitively high costs of nuclear power in the United States.

The total size of the US electricity market is about 4 PWh/yr.⁴¹⁾

Shifts of about 20% of the market between different sources seem to be achievable without restructuring too much of the rest of the electricity infrastructure. From 2010-2022, the share of electricity generated by coal fell from 45% to 20%, the share of electricity generated by natural gas grew from 24% to 40%, and the share of electricity generated by wind grew from 2% to 10%.⁴²⁾ Increasing the share of nuclear power up to the levels seen in Ontario or France would likely involve rebuilding other parts of the electricity infrastructure to accommodate it or involve making nuclear power plants dispatchable, which increases their effective costs. For this calculation, I will assume that 20% of the United States’ electricity production shifts from either coal or natural gas to nuclear.

Estimating the direct financial impact involves a comparison between the levelized cost of electricity (LCOE) for nuclear power under a different regulatory regime with the LCOE for coal and natural gas.

Devanney does this directly.⁴³⁾ He claims that nuclear power could be generated for as little as \$30/MWh, compared to \$50MWh for coal. His estimate for natural gas is before the recent price drop, so it is no longer accurate.

Lazard’s estimates of LCOE are consistently higher.⁴⁴⁾ His range for nuclear power includes \$150/MWh. Dawson estimates that construction costs are 80% of the total cost of nuclear power.⁴⁵⁾ South Korea’s nuclear construction costs are a factor of 4 lower than nuclear construction costs in the US.⁴⁶⁾ If we were to achieve South Korean construction costs, the LCOE for nuclear power would be \$60/MWh. This compares to Lazard’s estimates of \$70-160/MWh for coal and \$40-70/MWh for natural gas.

For this calculation, I will say that nuclear power could cost \$20/MWh less than coal and similar to natural gas under a different regulatory regime.

Multiplying this by 20% of the United States electricity generation would result in \$16 billion/yr in direct financial benefit from switching from coal to nuclear. The direct financial benefit from switching from natural gas to nuclear would be marginal.