To Be Notified When New Content is Posted, Go To Blog Page and Use RSS +/or E-Mail Subscribe Features
Nuclear Costs -- Who Has "Better Numbers"?
Who Pays if Things Go Wrong?
July 31, 2009
by Craig Severance
On July 10th I debated the Nuclear Energy Institute (NEI) in a head-to-head discussion in Washington. (See here for article on the "lively discussion".)
This week NEI published on its blog site NEI Nuclear Notes a very spirited critique (here) of my debate presentation. (My presentation was based upon my detailed study "Business RIsks and Costs of New Nuclear Power", published in January 2009 here.)
Now We're Talking. The NEI fight-back response is welcome in that we are blowing open the "Black Box" of hidden assumptions about the costs of new nuclear power. It is NOT a cordial discussion when one side won't disclose its numbers. I wrote about this in January:
It has unfortunately been the case over the last couple of years that some utilities have begun to claim that even rudimentary basics of their nuclear cost estimates must be hidden from the public as 'trade secrets". For instance, in the South Carolina Electric & Gas proposal to build two reactors now under consideration by the South Carolina PSC, there is literally a large "box" obscuring the bulk of the calculations....In a different case, Duke Energy claimed that it does not even have to disclose its new cost estimates for a proposed nuclear facility in Cherokee County, S.C. In the Duke case, C.Dukes Scott, South Carolina's consumer advocate... noted, "If the cost wasn't confidential in February," Scott said, "how is it confidential in April?"
We've come a long way since then, as we are now "duking it out" in a much healthier fashion.
What Will a Nuclear Plant Cost? It Depends on Whom You Ask. One of the most striking things about estimates currently circulating for new nuclear power is their range. They are "all over the map", as shown in this illustration below (with the nuclear column now circled by me) from Mark Cooper's recent study "The Economics of Nuclear Reactors: Renaissance or Relapse?":
Busbar Costs of Alternatives to Meet Electricity Needs
Cooper's study observes that the range of cost estimates depends greatly on the estimator. He notes "early consultants, government, and academics" produced estimates (now several years old) that were quite low. The age of those studies (before the rapid run-up in costs became apparent) is probably the biggest reason. The mid-range comes from recent utilty estimates. He comments: "Utilities, especially in the early phase of the regulatory process, have an interest in understating costs, as long as the estimates are nonbinding. Low-balling the costs helps to get the power plant approved." Finally, the higher end of cost estimates (as shown on the graph) come from "Wall Street and independent analysts".
Whose numbers would you trust the most?
Am I the "Outlier"? On the above graph, you can see that "Severance" is the upper end of the range, at $210/MWH (21 cents/kWh), which comes from backing the $2018 first year of full operation nominal dollar 30 cents/kWh back to Cooper's $2008. This is my upper case scenario. The lower cost scenario -- which I presented in the NEI debate -- in nominal dollars in the first full year of operation would be about 25 cents/kWh, and this comes in at around 18 cents/kWh in $2008.. My study includes roughly 2 cents/kWh for externalities (nuclear waste & decommissioning costs) not often included by others, Cooper recently e-mailed me noting "with this adjustment [backing down to $2008-- cs] your estimates of 18 to 21 cents are on spot with the upper end of my range 16 to 20 cents." (Note the "upper end" of the range was from Wall Street and independent analysts, so I'm in good company.)
How Do Estimates End up So Different? Total cents/kWh numbers are conclusions, and we are still lost unless we know where they came from. We can't just dismiss a source as biased, we need to know how the numbers ended up so different. How is this happening? We'll look at the recently released May 2009 Update to MIT's "The Future of Nuclear Power" study, versus my Study.
Starting Point -- "Overnight" Cost -- Essentially the Same. "Overnight" costs are what it would cost to build if you could "build the plant overnight" -- at today's prices with no further cost inflation, and no financing costs because you built it all overnight. You can't actually do that, so Overnight Cost is just a "starting point".
MIT's 2009 Update accepts the shocking news that Overnight costs to build a new nuclear plant increased from $2000/kW in $2002 dollars, to $4000/kW just five years later, in $2007. MIT notes "the estimated cost of constructing a nuclear power plant has increased at a rate of 15% per year heading into the current economic downturn."
My Study used essentially the same number, for my upper case scenario, starting with an "Overnight" cost in $2007 of $4,070/kW.
Factor in Current Slight Drop in Costs. NEI's critique of my Study claims I have not factored in the slight drop in the power plant cost index which has occurred during the recent recession. However, in the NEI debate I presented my Low Cost scenario, which begins with an "Overnight" cost of only $3,596/kW. This is a drop of almost 12% from my previous "Most Likely" Scenario and 10% less than the MIT Overnight Cost estimate.
The CERA power plant cost index as of 1st Qtr 2009 had fallen to 217 from its high of 231 at 3rd Qtr 2007, a drop of only 6%. Hold off on the confetti -- a 6% drop is nice, but not much of a backdown after costs doubled the previous 5 years. If that's all the greatest worldwide economic meltdown in decades can reduce power plant construction costs, they are apparently not going to come down very much.
By moving to the Low Cost case with its 12% lower Overnight Cost, this more than accounts for the recessionary slight drop in power plant construction costs.
How Long Will it Take? We can start from the same point, but a major flaw occurs when the amount of time needed to actually complete a nuclear project is underestimated. The longer it takes, the more the project is exposed to annual increases in costs, driving up total construction costs. The longer it takes, the more financing costs are incurred during construction.
This is a huge error of many optimistic nuclear cost projections. For instance, the MIT Update assumes construction takes place over just 5 years from 2009-2013, with the plant operational in 2014! This is far shorter than nuclear project schedules laid out in utility dockets today (e.g. the South Carolina Electric & Gas facility projects pre-construction expenses beginning in 2007, with operation of the two reactors beginning in 2016 and 2018, a total length of 11 years).
Cooper plotted the expenditure schedule in my Study (for a 2-reactor project) compared to Moody's study (for a 1-reactor project):
Construction Expenditures Across Time
Source: Cooper, Mark "The Economics of Nuclear Reactors: Renaissance or Relapse?"
NEI Notes proudly displays a table showing all the many ways the nuclear industry has worked to shorten the licensing and construction process. All of these changes are already assumed in the utility-developed nuclear project schedules. It still takes a long time to develop a nuclear Mega-Project. As noted in my Study, I did not assume any delays, though delays are chronic for the nuclear industry. (See here -- with a bonus song at end -- for story on Progress Energy's recently announced nearly 2 year delay.)
Severely understating the length of time to complete a project is like two runners comparing times, with one correctly considering the course as 500 meters, while the other is bragging about his much better times -- for only a 100 meter race!
This construction schedule flaw in optimistic nuclear estimates was clearly discussed in my recent articles (e.g. my "Boiling the Frog" article here) -- yet NEI Notes chose to completely ignore this issue.
Have Power Plant Cost Increases Essentially Stopped? There is a well-known story about a man who started to build a tower, but because he did not "count the costs" to complete the tower, had to abandon the project to the ridicule of his neighbors.
With a Mega-Project like a nuclear plant, you might start with cost estimates of around $4,000/kW -- at today's prices. However, by the time you actually start building the plant you will need to pay the "going rate" in the future years when construction actually takes place. Note from the expenditure schedules above that the bulk of the expenditures usually take place in the "back half" of the schedule, several years after the project originally begins moving through licensing and approval processes.
If power plant construction costs again increase at anything like their aggressive inflation of recent years (15%/year for the 5 years entering into the recession, as MIT noted), total construction costs will be astronomically higher than originally thought. My Study assumed this gets better, and moderates to only about 8%/year, a major decrease from recent experience.
Optimistic nuclear cost estimates acknowledge that recent (pre-recession) cost inflation for power plant construction has been severe. However, they assume the burner heating up cost inflation has now been shut OFF and future escalations will be no more than general economy inflation rates. For instance, MIT's study assumes only a 3%/year inflation in costs over the starting point Overnight Cost estimate. Similarly, Florida Power & Light's Turkey Point project assumes future cost increases will be only 2.5%/year -- even lower than recent Consumer Price Index inflation!
Who is right? This is the future we are discussing, so unless Christopher Lloyd comes roaring in with his DeLorean to tell us the answer, we won't know till it happens. However, there are clues.
"We're Not in Kansas Anymore". The NEI Notes article cites older experience with electric utility cost indices, arguing from this older experience that we should once again expect power plant costs to rise at very gentle inflation rates. However, if this is true, what explains the rapid growth in costs over the last several years?
Most analysts see the world economy is now changing in fundamental ways, to make room for the rapid growth of emerging economies, especially China and India. They see fierce competitiion for scarce resources as the major reason why power plant costs (along with prices for major commodities) shot up so rapidly just prior to the recession.
Source: China and India's Ravenous Appetitie for Natural Resources -- Their Potential Impact on Colorado, Colorado Geological Survey, link here to access excellent PPT
For an excellent discussion of these trends, I recommend the Colorado Geological Survey's Power Point (PPT) presentation cited above.
Hazard Duty Requires More Pay. The new MIT study includes an "IF Only" scenario -- "if only" nuclear projects could be built with the same "Risk Premium" for their financing costs as natural gas and coal projects.
"IF ONLY" is truly wishful thinking in this case. Nuclear power projects carry the inherent risk of managing a Mega Project stretching over many years, with exposure to cost increases, delays, and massive financing costs during this whole period. On top of these risks, the projects are so large they strain the financial resources of even the largest utilities.
Moody's considers new nuclear power projects so risky they described the risk for utilities as a "Bet the Farm Risk", in their June 23, 2009 pronouncement "New Nuclear Generation: Ratings Pressure Increasing". The report (cited in this Reuters article here) was essentially a "shot across the bow" to utilities planning new nuclear projects. Moody's warned these utilities that unless they acted soon to strengthen their balance sheets to be able to handle such a large project, they could expect their credit ratings to be downgraded.
IF Only Carbon Prices Helped Out. With all the major flaws discussed above, it is more wishful thinking still to believe that carbon trading can rescue nuclear power. After making very major leaps of faith on construction schedules and cost escalations, MIT's study still fell short of finding nuclear power to be competitive. So, they propose carbon taxes may finally make nuclear power compettive.
More realistic assumptions about new nuclear power's real costs, however, dispel the notion that carbon trading will solve its problems. For instance, in Citi's August 13, 2008 "Sell Downgrade -- Nuclear Renaissance to Nuclear Vaporware?" report on nuclear vendor Shaw Group, Citi found that at more realistic cost projections (as in Moody's study) for new nuclear power, an implied oil price of $397/bbl would be needed to justify new nuclear power (using oil prices as a "proxy" for energy prices in general).
Nothing approaching the kind of premiums for fossil fuel use needed to make new nuclear power competitive are likely to ever come about as a result of carbon pricing. A fictional new nuclear power might be helped, but the real new nuclear power doesn't even come close.
Isn't Everybody on the Public Dole? So we come back once again to asking the taxpayers for money. I find it simply amusing NEI Notes argues strenously all forms of energy production are subsidized, so why not nuclear power?
The charts and numbers they link to show one thing very clearly -- energy industries with lots of money for lobbying and campaign contributions (oil, gas, coal, and nuclear) have enormous subsidies from the Federal government. Meanwhile, solar and wind power have tax credits, set to expire in a few years.
NEI also argues that Federal Loan Guarantees are definitely the way to go. I couldn't disagree more. Look at the current situation. Even in this recession, the most economical projects (energy efficiency, wind power) are still obtaining private financing and still moving forward. The riskier, more expensive projects are not having as much success attracting private capital.
What do you know -- capitalism works! A Federal Loan Guarantee, on the other hand, is a means of state-controlled Central Planning for the economy. Instead of letting private investors and lenders choose the most worthwhile projects based on economics and risk, a political decision is made that certain projects "must be done". Ask the Soviets (remember them?) how well this works.
My own belief, as I stated in the debate, is that if Congress sets strong limits on carbon dioxide emissions, there should no longer be any need for public subsidies. Utilities will be under an emissions limit, and can bid out the most cost-effective ways to meet those limits.
Whose Money is On the Line? There are big differences in cost estimates for new nuclear power, depending upon who is doing the estimating -- a nuclear promoter, or an independent analyst.
Since enormous amounts of money are at stake, the big question then becomes -- whose money is on the line? If nuclear vendors want to project very low costs, let them back up those projections with firm commitments.
Instead, we see the industry turning to ratepayers and taxpayers to take the hit if their numbers are wrong.
In other words, the industry wants to play but won't put its money up. With a huge stash at risk, is this a good bet?
This article was first published on July 31, 2009.