Boiling The Frog: Nuclear Optimism Hides True Costs Till It's Too Late
July 24, 2009
by Craig Severance
There is a well-known story about how to boil a frog. If you try to throw a frog into a pot already boiling, he'll jump out. However, put a frog into a pot and slowly raise the temperature -- and you get frog legs for dinner.
The nuclear power industry seems to be pursuing this strategy, slowly releasing ever higher cost estimates for new nuclear power plants. If the public does not realize the true costs of a new nuclear plant, the industry can obtain political support for the Federal loan guarantees it needs. After the taxpayers are on the hook and a nuclear project is already underway, the full costs will become clear.
At that point, however, it may be too late for taxpayers and utility ratepayers to jump out.
The Frog Jumps: The Ontario Story. Last week the Ontario government put plans to build 2 new next-generation reactors on hold, after it received bids "more than three times higher than what the Province expected to pay", according to a story in the Toronto Star. The only "compliant" bid -- one where the supplier would be sufficiently at risk if costs exceeded the amount quoted -- was reportedly a $26 billion quote from Atomic Energy of Canada, Ltd, equal to roughly $10,800 per kW. (If this sounds familiar, recall my January 2009 study estimated a new nuclear project would most likely cost approximately $10,500/kW).
"It's shockingly high," the Star quoted Wesley Stevens, an energy analyst at Navigant Consulting in Toronto regarding the nuclear bid.
The Province had originally thought the two reactors would cost a total of only about $7 billion, which works out to about $2,900 per kW, according to 2007 estimates. The Star reports "During Ontario Energy Board hearings last summer, the power authority indicated that anything higher than $3,600 per kilowatt would be uneconomical compared to alternatives, primarily natural gas."
Rather than blindly accepting these estimates, however, Ontario authorities were wise enough to require nuclear vendors to submit bids accepting the risk of cost overruns. Once the suppliers were on the hook -- rather than ratepayers and taxpayers -- the true costs of new nuclear power became apparent.
Another bid was reportedly received from French nuclear vendor Areva LP, which also "blew past expectations", at $23.6 billion. However, that bid was ruled non-compliant as Areva was unwilling to sufficiently shoulder the risks of cost overruns.
It is not clear if these numbers include financing costs, as the process was confidential and the numbers reported by the Star were leaked from sources directly involved in the process. However, it would be highly unusual for a nuclear vendor to bear the cost of financing for a project, so the total cost of these plants is likely to climb even higher once financing costs are included.
The result of the risk-based bidding process was that Ontario, seeing the true costs of new nuclear power, has now "jumped away".
The Frog Slowly Boils: Nuclear Cost "Projections". The Ontario bidding forced a realistic look at actual costs to complete a new nuclear project. However, other proposed North American nuclear plants are still working with nothing but "projections", which invariably are optimistic.
When these "projections" are later found to have been ridiculously low, it will be utility ratepayers and taxpayers who bear the cost overruns.
The Pot Heats Rapidly. In 2003 MIT released its study "The Future of Nuclear Power", projecting that a new generation nuclear plant could be built for only $2,000/kW in "overnight" costs, in 2002 dollars. ("Overnight" costs are what it would cost to build a plant if it could literally be built "overnight" -- at prices in existence at the time, with no further inflation in prices, and no financing costs. Since this is not possible, a realistic estimate has to include future inflation in prices plus the cost to finance the plant.)
Many new nuclear power plant proposals in the U.S. and elsewhere were initiated based on these early optimistic projections.
Over just a 5 year period from 2002 to 2007, however, MIT's estimates of the "overnight" cost to build a new generation reactor doubled from the $2,000/kW projected as 2002 costs. MIT's 2009 Update now projects "overnight" costs to be $4,000/kW (in $2007), noting "the estimated cost of constructing a nuclear power plant has increased at a rate of 15% per year heading into the current economic downturn." The rapid inflation occurred because of worldwide economic growth causing strong demand for scarce resources, particularly from the rising economies of China, India, and Brazil.
Thus, after six years of hearing about a "Nuclear Renaissance", bad news is finally released that costs are now twice as high as previously expected.
Is Burner Now Shut OFF? If anyone looks at the way nuclear costs have been rising, they might be very afraid to "jump in" to a nuclear project. However, after admitting that nuclear costs have been on a rapidly increasing curve, the industry now says in effect it is "ok to jump into the pot" -- the burner of nuclear price inflation has been shut OFF.
MIT's 2009 Update now projects further cost increases will be no more than an assumed general economy inflation rate of just 3%/year. In other words, in "real" (inflation-adjusted) dollars nuclear construction costs will no longer go any higher.
Similarly, Florida Power & Light in its proposal for two new reactors at its Turkey Point station, starts with midrange $2007 dollars cost estimates of $4,070/kW. Although this is over twice cost estimates from just five years ago, FP&L projects that further nuclear construction costs will increase by only 2.5%/year. This is even less than the Consumer Price Index has risen in recent years.
Small Doses of Bad News -- The Pot Slowly Warms. It has now become clear how utility customers and taxpayers will find out that nuclear cost projections were overly optimistic -- a little bit at a time, as construction proceeds anyway.
Last week, the South Carolina Office of Regulatory Staff released its first Quarterly update of the projected budget for two new reactors at the South Carolina Electric & Gas Company's V.C. Summer nuclear generating station. The update showed SC E&G's share of the proposed project's budget increased by more than $500 million, from $6.3 billion to almost $6.9 billion (approximately 9%). Each new quarter, the proposed project budget will be updated, based upon the past five years' average cost increases.
However, in this backward-looking process, will it ever finally be realized that the utility has underestimated future expected cost increases? Will these ever be adjusted to develop a more realistic estimate of total costs to complete the project?
(Note: I do expect some updates to show a brief downtick due to the effect of the recession -- however, the recession is also cutting kWh usage and thus the need to build new power plants. The very factors that increase kWh usage also likely will restart the fires that cause power plant construction price increases.)
Oops -- You Mean We Can't Start Building Tomorrow? Another unseen feature of many optimistic nuclear industry cost estimates is the assumption that cost increases will be spread over an amazingly short construction period.
For instance, the MIT 2009 Update assumes construction will occur over a 5 year period from 2009 through 2013, with plant start-up in the year 2014! However, even the most optimistic schedule in new reactor cases (the SC E&G case noted above) has pre-construction costs beginning in 2007 with reactors coming on line in 2016 and 2018. This discrepancy is similar to estimated vs. actual construction periods in the first wave of nuclear power plants built in the U.S., as shown below:
Nuclear Reactor Construction Periods -- Estimated vs. Actual
Graph: Cooper, Mark "The Economics of Nuclear Reactors: Renaissance or Relapse?"
Assuming a very short construction period means inflationary increases in costs don't happen for very long, so they are not a big factor. However, even at low inflation rates, when you plot costs over a realistic time frame running 10-12 years and where most of the costs are spent in the back half of that time period, you realize that costs will rise very significantly.
The above discussion is based on the South Carolina project schedule and assumes the utility's own original schedule, without delays. However, delays do happen, especially to nuclear projects, such as a nearly 2-year delay recently announced for Progress Energy's proposed Florida nuclear plants. (See story here -- and including a bonus song at the end.)
Shocked and Amazed. As each new cost update is released , the nuclear industry will be "shocked and amazed" that costs continue to rise more rapidly than projected.
Only after billions have already been spent on a project will the true costs to complete the project become apparent. At that point, since an unfinished reactor produces exactly zero kWh, pressure will be intense to complete the project no matter what it costs, so that the billions already spent "won't be wasted".
This is exactly how projects costing an average of two to four times original cost estimates continued through to completion, in the first wave of U.S. nuclear projects. Only 9 nuclear projects were canceled after construction had already begun.
Frog Legs Anyone? And so -- the frog is slowly boiled. Those who wish to avoid this fate should take a lesson from Ontario and demand firm commitments from nuclear vendors. Ratepayers and taxpayers should not have to bear the costs of "unexpected" overruns. If the nuclear industry is required to be accountable, perhaps some real accounting will show us new nuclear's true costs.