Thursday, February 18, 2010

Nuclear fallout

Among the biggest nuclear power hazards in the United States have been financial ones. Of 259 orders for nuclear power reactors from 1955 through 2005, 124 were cancelled and 3 terminated for other causes. Making it through to an operating license were 132 reactor orders [1], a success rate of 51 percent. Nuclear fallout continued with early shutdowns of the licensed reactors. Through 2005, 28 had been shut down, leaving 104 in operation. Operating lives of the reactors shut down ranged from less than 1 year to 35 years, all less than the expected 40 years, with an average operating life of 14.1 years [2]. Survival rate to mature operation was 40 percent.

Of the 124 reactor orders cancelled, 15 were under construction. All those abandoned reactors occurred between 1982 and 1988 for construction that started between 1972 and 1977 [3]. Financial losses totalled $22.3 billion at times of abandonment, or $60 billion adjusted to the start of 2010. Following is a list of the abandoned United States nuclear power reactors (reactor name, location, start year, stop year, adjusted loss):

WNP-1Hanford, WA19751982$2.0 billion
WNP-4Hanford, WA19751982$1.4 billion
WNP-5Satsop, WA19771982$1.6 billion
Zimmer 1Moscow, OH19721983$5.5 billion
Cherokee 1Gaffney, SC19761983$1.7 billion
WNP-3Satsop, WA19771983$1.6 billion
Midland 1Midland, MI19731984$4.4 billion
Midland 2Midland, MI19731984$7.6 billion
Marble Hill 1New Washington, IN19771984$4.7 billion
Marble Hill 2New Washington, IN19771984$2.4 billion
Perry 2North Perry, OH19741985$2.7 billion
Bellefonte 1Hollywood, AL19741988$9.3 billion
Bellefonte 2Hollywood, AL19741988$6.2 billion
Seabrook 2Portsmouth, NH19761988$4.9 billion
Watts Bar 2Spring City, TN19771988$4.1 billion

This discussion and the list of abandoned reactors do not count military or research programs, including the Clinch River Breeder Reactor at Oak Ridge, TN, abandoned in 1983 after spending about $1.6 billion in then-current dollars. Nor do they count losses from the 109 power reactors cancelled before construction, which have rarely been reported and never summarized. Those losses may have totalled $100 billion or more in 2010 dollars, based on a Missouri rate increase approved, but never implemented, for an unbuilt reactor [4]. The last reactor order cancellation occurred in 1994, ending a long run of heavy losses for electric utilities.

The period from the mid-1970s through the mid-1980s was a wild ride for the nuclear power industry. For a few years around the Arab Oil Embargo of 1973 orders surged. Then the industry found new demands for electricity well short of projections. At the same time, federal regulations rapidly tightened safety requirements, a trend that strengthened after the disaster at Three Mile Island 2 in 1979, greatly increasing construction costs. Hyperinflation of the late 1970s further increased costs, as interest rates on construction loans spiraled. When planned in 1971, the Vogtle plant in Georgia was estimated to cost $0.66 billion for four reactors, but when completed in 1989 it actually cost $8.87 billion for two reactors, including financing [5]. Cost escalation per reactor was about a factor of 9, adjusted for inflation.

The U.S. began a second half-century of nuclear power with promises of "third generation" reactors that would be safer, more reliable and more predictable to build and operate. The new reactor designs are supposed to be certified for operation before starting construction, unlike the earlier generations that were certified only on licensing to operate. There are more than twenty orders pending for those reactors. However, risks of failure remain high. As of early 2010 the construction cost for 1,200 MW of nuclear power capacity was estimated at around $7.9 billion [6], while the construction cost for the same coal-fired capacity was estimated at around $4.2 billion [7]. As of 2010, no proposed reactor could be built without a loan guarantee from the federal government. Protracted negotiations were needed to set loan guarantee fees. There also remain liability caps for major disasters (the Price-Anderson Act of 1957) underwritten by the federal government.

Dr. Paul Joskow, an MIT economist, estimated the cost of bulk power from "third generation" commercial nuclear reactors at 6.7 cents per kWh in 2006, based on $2.4 billion in construction costs for 1,200 MW of capacity [8]. In comparison, he estimated costs from 3.8 to 5.2 cents per kWh for power from coal-fired and gas-fired units. Dr. Joskow based his estimates on so-called "overnight" costs of construction, without interest or inflation, but of course reactors cannot be built overnight, so his models include some allowances for cost growth.

In view of the nuclear power industry's history, radically underestimating construction costs and cost growth, it would be prudent to substitute current, fully loaded cost estimates in place of the "overnight" costs, hoping that actual cost escalation will not break the budget of allowances in economic models. At $7.9 billion for 1,200 MW of capacity, the cost for nuclear power in Dr. Joskow's estimates would grow to 17.7 cents per kWh [9].

Recently Dr. John Parsons, an MIT Sloan School economist, showed that published costs for "third generation" nuclear plants use different approaches [10]. Some bundle in transmission upgrades; some include interest; some allow for cost escalation. The variations produce large discrepancies. Dr. Parsons' estimate of the "overnight" cost in 2007 for the power plant in reference [6] works out to $5.7 billion for 1,200 MW of capacity.

Probably more significant than specifics of estimates is a pattern of cost growth. The works of both Dr. Joskow and Dr. Parsons indicate that they trusted a 2002 "base case" of $2.00 a watt for the "overnight" cost of a "third generation" nuclear power plant. As of 2007, Dr. Parsons' estimate became $3.95 a watt, for an average of 5 plants with 10 reactors [10]. Cost estimates from these economists nearly doubled in only 5 years.

Only radical increases in costs of fossil-fuel plants and prices of fossil fuels could compensate for such huge differences between costs of power from nuclear and competing new sources. For the forseeable future, nuclear power looks like a high-risk, high-cost option for the United States, and for that reason it is unlikely to grow rapidly, even with strong support from the federal government.

[1] U.S. Energy Information Agency, History of energy in the United States: Nuclear Energy (2002), available at

[2] U.S. Energy Information Agency, Nuclear generating units 1955-2008 (2009), available at U.S. Nuclear Regulatory Commission, Commercial nuclear power reactors formerly licensed to operate (2005), available at

[3] Compiled from public records, Cancelled nuclear units ordered in the United States (2005), available at

[4] U.S. PIRG, A history of action in the public interest, 1980s (2002), available at

[5] U.S. Energy Information Administration, Vogtle nuclear power plant, Georgia (2009), available at

[6] Rob Pavey, Feds back two new reactors at Plant Vogtle, Augusta Chronicle, February 16, 2010, available at (two 1,100 MW units).

[7] Stacey Roberts, Utility's rate-rise request tapered, Arkansas Democrat-Gazette, October 15, 2009, available at (one 600 MW unit).

[8] Paul L. Joskow, Prospects for nuclear power, a U.S. perspective (2006), available at See Comparative Base Load Costs, page 23.

[9] See Craig A. Severance, Business risks and costs of new nuclear power, Electricity Journal 22(4):112-120 (2009), draft version available at Extended analysis in this study predicts bulk cost for nuclear power from "third generation" nuclear reactors of 25-30 cents per kWh.

[10] John E. Parsons, Financing new nuclear generation (2009, updating Future of nuclear power, 2003), available at See Table 4.