Nuclear Power: The Missing Cost-Benefit Analysis

Bright Future by Goldstein/Qvist - book cover

An influential1 new book A Bright Future says the world need lots more nuclear energy to slow the carbon emissions that cause climate change. Renewable energy sources are okay as far as they go, the authors argue, but so far can meet only a small fraction of our needs, especially in the developing world where electricity demand is increasing rapidly. Carbon-free nuclear energy can fill the gap, they tell us.

As an ordinary citizen, when I hear about nuclear energy nowadays, my first thoughts go to safety. I get a mental picture of the Fukushima meltdown in 2011. I was willing to give the authors the benefit of the doubt, however. I got their book, and the first thing I looked for was their cost-benefit analysis. I wanted to see how they factor in the chances of another disaster on the scale of Fukushima, Chernobyl, or Three Mile Island, and the potential costs to society of that.

But I didn’t find a cost-benefit analysis. There is none in the book.2 This raised alarms bells for me. Most students of economics and public policy learn cost-benefit analysis as a basic tool.3 Maybe it is missing because neither author is an economist. (One is an engineer, the other a political scientist, and the book’s forward is written by a psychologist.)

Instead of a cost-benefit approach, the authors offer a remarkably Panglossian view of those disasters. For example:

“… [T]here was no nuclear disaster at Fukushima. There was a natural disaster of biblical proportions, a small consequence of which was a very expensive and disruptive but nonlethal industrial accident at the Fukushima plant…”

Chapter 7 of A Bright Future

Never mind that the accident came close to spewing enough radiation downwind to make the city of Tokyo uninhabitable.4 At least they do acknowledge that the accident was very expensive. But then they fail to take that expense into consideration, in their analysis.

In the spirit of providing a more solid economic foundation to their case for nuclear energy, here I provide a first draft of the cost-benefit analysis that is missing from their book.

Regarding the disaster danger, a simple model would start by looking at the issue the way an insurance executive might. Suppose you are asked to insure a new nuclear plant against liability for an accidental meltdown. You compile some actuarial data based on industry experience. For example:

n300Number of nuclear plants in the world 5
d3Number of past disasters (Three Mile Island, Chernobyl, Fukushima) 6
c$442 billionAverage cost of a disaster (including wider economic or medical costs, not just direct cleanup.) 7

Using that data, there’s a d/n = 1% chance of a plant having a disaster over its lifetime. The expected (actuarial) cost of such a disaster is 1% X c = $4.4 billion 8

By this model, to issue an insurance policy that covers a plant for its lifetime, you would need to charge a premium of at least $4.4 billion. According to the book authors, the cheapest nuclear reactors cost $2 billion, so our insurance policy inflates the cost considerably. What does that do to the economic case for nuclear?

Of course, a thorough cost-benefit analysis will take into account much more information. But a first approximation such as this is useful to highlight the orders of magnitude we are dealing with.

The authors rightly point out that burning coal to produce electricity imposes enormous costs on society in terms of human lung disease as well as global warming. On a cost-benefit basis, maybe nuclear still will look better than coal, if they would do the math. On the other hand, our model above does not account for the risk of a terrorist attack on a nuclear plant, which we have never seen, and thus have no experience rating to plug in.

On the subject of terrorism, the authors basically tell us, not to worry:

Since the 9/11 attacks of 2001, the public imagination easily conjures up the potential dangers of terrorists crashing airplanes into things…. Analysis after 9/11 concluded that a fully loaded Boeing 767 jet would do little if any damage to a nuclear reactor even if it scored a direct hit.

Chapter 7

That’s more or less all they have to say about terrorist threats to nuclear plants.

This is a stunning failure of imagination (or maybe an intentional one?). How about a squad of suicide bombers driving a big truck packed with ammonium nitrate fertilizer and kerosene? This simple concoction, the home-made equivalent of dynamite, is far more explosive than would be jet fuel in the imagined airplane attack.9 How many nuclear plants around the world are vulnerable to such a group driving up to the front gate, killing the guards, then getting to the reactor building before any security force can stop them?

I also am annoyed at the way they use an apples vs. oranges line of argument. The authors describe new “fourth generation” nuclear technology (and beyond) that promises to be safer than what is in use today. This takes a certain leap of imagination, as none of this technology yet exists. On the other hand, they pound home the limitations of existing renewable energy technology. But renewable technology also is advancing quickly, such as that related to hydrogen fuel cells and underwater turbines, along with continual improvements in the efficiency of photovoltaics and power storage. To put nuclear and renewables into a head-to-head comparison for the future requires us to make projections about the advance of all these technologies. Such comparison is beyond the scope of this article, as it also was beyond the scope of their book.

The authors point out that adoption of carbon taxes could shift energy usage patterns in a positive direction. This is a policy approach I wholeheartedly endorse, as do most economists. They complain about how difficult it is to advocate raising taxes in a democratic society, and they are correct with respect to that observation as well. At the same time, they envision a seeming fantasy land where those same democratic societies, that balk at entirely rational carbon taxes, will see the light as to the true costs of coal vs. nuclear. If that is going to happen, we need the purported experts to realistically confront the cost-benefit trade-offs of all these choices. This book is not the place to look for any analysis of that kind.


  1. Summary versions as op-eds appeared in Wall Street Journal 11 Jan. 2019 and New York Times 6 April 2019.
  2. In fact, the term “cost-benefit”, searched in various permutations in the e-book, does not occur in the book.
  3. See for example Cost-Benefit Analysis Course, Wagner Graduate School of Public Service, New York University; also classic text Cost-Benefit Analysis by Mishan (I used his first edition in college in 1970s).
  4. See for example, speech by Allison Macfarlane, former head of the U.S. Nuclear Regulatory Commission, “Safety Expert: Fukushima Nuclear Disaster Could Have Been Worse — And Nearly WasIndependent News, 3 Mar 2016. Also see chapter 5 “Meltdown” of Japan: The Paradox of Harmony, by Keiko Hirata and Mark Warschauer, Yale University Press, 2014.
  5. Approximation. There are 442 reactors worldwide but some plants have multiple reactors. (USA has 99 reactors at 61 plants.) We use plants not reactors in the model because, if there a disaster at a plant, is likely to affect multiple reactors (as at Fukushima).
  6. This underestimates potential dangers because there have been other serious accidents that luckily did not turn into disasters.
  7. Using average of three disasters: Chernobyl $700B from The Financial Costs of the Chernobyl Nuclear Power Plant Disaster: A Review of the Literature, Jonathan M. Samet and Joann Seo, University of Southern California, 21 April 2016; Fukushima $626B from Real cost of Fukushima disaster will reach Yen 70 trillion or triple government’s estimate: think tank, Japan Times, 1 April 2017. To be conservative we have used $0 for Three Mile Island because comparable estimates of economic cost not found in our brief research.
  8. Probably an under-estimate here because many plants are in mid-life, so the possibility of further accidents during their lifetimes is not taken into account in the experience statistics.
  9. It is essentially the material that Kansas City terrorist bomber Timothy McVeigh used in 1995.

Leave a Reply

Your email address will not be published. Required fields are marked *