Cost of Money for Nuclear Projects – Effect of $880 Million Financing Fee – 30% Increase in Electrical Power Price

I have been musing a bit this morning about the impact of the cost of money for nuclear projects in the United States compared to the cost of money for other energy investments. I have a spreadsheet model that I built based on the methodology described in Appendix 5.A — Calculation of the Levelized Cost of Electricity of the MIT Future of Nuclear Power study.

Though I do not know ALL of the required input numbers, I can make some educated guesses based on published information. I thought it might be a useful exercise to determine just how large an impact an $880 million credit subsidy cost would have on the economics of the Calvert Cliffs Unit 3 nuclear power plant project.

Here are the assumptions that I made, just in case someone wants to check my math or challenge the basis for the numbers I chose.

Model Variables

Overnight cost ($/kWe) 6500
construction time (years) 5
Total construction cost ($/kWe) 6126
Debt fraction of initial investment (%) 80
Equity fraction of initial investment (%) 20
Nominal cost of debt (%) 4
Nominal cost of equity (%) 10
Plant life (years) 40
Plant net capacity (Mwe) 1600
Capacity factor (%) 85
Marginal composite corporate income tax rate (%) 39
Heat rate (BTU/kWh) 10400
Unit cost of fuel ($/mmBTU) (not including disposal) 0.39
Nuclear waste fee (mills/Kwh) 1
real fuel escalation 0.5
Fixed O&M ($/kWe/yr) 47
Variable O&M (mills/kWh) 8
Incremental capital costs ($/kWe/yr) 20
Decommissioning cost ($million/GWe capacity) 2000
Carbon emissions tax ($/tonne-C) 0
assumed rate of inflation (%) 2
Debt term 15
Assumed rate of electricity price inflation (%) 0.500
Depreciation schedule 15 year MARCS

Using all of those assumptions, I computed that the sales price for electricity required would be about 10.7 cents per kilowatt hour at the time that the plant started operating, assuming that the credit subsidy cost was zero. With a credit subsidy cost of $880 million, required to be paid at the beginning of the loan period when construction begins, the electricity price needed for exactly the same return on invested capital would increase to 13.85 cents per kilowatt hour. That represents a 30% price increase completely due to the cost of paying for loan insurance.

(Aside: These numbers are not levelized cost of electricity. They are the prices required to produce free cash flow providing a 10% return on invested capital in year 2 of the project. Due to the effects of depreciation on reported profits, the ROE based on after tax profits is lower during the first 15 years than the ROE based on free cash flow.

Once depreciation is complete and the loans are repaid, ROE on profit and free cash flow are equal to each other and increase to about 21% in the case of no CSC and a lower 17% (with 30% higher electricity prices) in the case where the CSC was imposed using the above assumptions. End Aside.)

I am not advocating that the CSC should be zero; I am trying to illustrate the effect of the proposed fee on the price of electricity.

Since Constellation Energy is not a regulated monopoly utility with a guaranteed customer base, the sales price of electricity cannot be fixed to provide a given rate of return. The company would have to compete in the open market to sell its product at whatever the going rate would be.

Like other companies that sell their products in unregulated markets, Constellation would have the option of accepting lower margins in order to maintain sales volume. However, Constellation, like any other company, has a limited amount of capital to invest and a limited amount of managerial attention that can be applied to capital investments. It is not hard to see why model results like this would cause the company’s board of directors to determine that they have less risky investments available.

In a rank ordering of available investment opportunities, this particular project would have fallen quite far down on the list, especially after a long period of fruitless negotiation. The recent announcements by DOE and OMB officials that express surprise merely indicate a lack of ability to read the handwriting on the wall. It is hard to believe that anyone would fail to understand that the company did all they could – even to the extent of obtaining a hearing on the topic with pointed questioning by the Senate Energy and Natural Resources committee.

Please do not read this as uncritical defense of Constellation Energy. It is also not a political commentary. I will take the OMB evaluators at their word and assume that the structure of the project as a merchant power plant without a guaranteed power purchase agreement is responsible for the high level of project risk that led to the associated fee computation result.

The point is that the story helps to illustrate why many observers have come to the conclusion that electricity is too important to the functioning of modern societies to be left up to market forces. That is especially true when it takes so long to build the most promising types of new generating capacity. The short term nature of market trading does not contain the mechanisms required to support the long range planning and sustained effort required for a reliable electrical power grid.

For the past several years, the Maryland government has debated the issue of regulating electrical power production again. Perhaps it is time to recognize that the effort needs to move forward. Regulation would alleviate the concerns about the financial risk associated with market generators building very large projects that take a half a decade or more to construct and put into operation.

About Rod Adams

20 Responses to “Cost of Money for Nuclear Projects – Effect of $880 Million Financing Fee – 30% Increase in Electrical Power Price”

Read below or add a comment...

  1. uvdiv says:

    I don’t buy this for a second. The whole point of the loan guarantee is that the government covers the the debts if the nuclear builder defaults. One, you don’t account for this reduction in RISK to the builder. Two, you don’t account for the effects of this guarantee on the cost of financing — “nominal cost of debt (%)”, “nominal cost of equity (%)”. The securities are guaranteed by the federal government, so they are unnaturally “safe” and their cost will be far lower. Nuclear financing costs are currently higher than fossil fuels because of risk (check your source’s Table A-5.A.4); with loan guarantees, they’d be far LOWER — they would get below-market rates for loans. Your electricity price should certainly be lower.

  2. donb says:

    Rod, it would be interesting to run the numbers again for Calvert Cliffs #3, but using the same financing that Shepherds Flat wind farm in Oregon got as you wrote about previously .

  3. JimHopf says:

    Off topic:
    For what it’s worth Rod, it appears that the residents of Zion, IL agree with you about what should be done with the local plant.,0,2370880.story

  4. Jeff Schmidt says:

    Does anybody have a good source of data on average energy prices ($/kWh) for the different types of currently deployed generation system?
    I ask, because $0.107/kWh sounds to me to be kind of expensive to start with? 13 cents an hour sounds so expensive they wouldn’t have any customers. . .?
    I know on my electricity bill, the “Cost To Compare” is listed as like $0.096/kWh. So, .107 sounds a little bit on the expensive side, but quite possibly close to being competitive (I’m not sure of the particulars of the energy market that Calvert Cliffs plans to serve – it may be that 10.7 is right within the median price in that market), but am I right in thinking that at > 13 cents, the price would be too high to compete?
    Now the next question: are there other reactor/plant designs which might come in cheaper? Perhaps the Calvert Cliffs plan, as it currently exists, is just too expensive and *deserves* to be scrapped, but that Nuclear as a whole can be more cost competitive? Does anyone have any links to good, current analysis of what ways in which Nuclear can be made more cost competitive, along with realistic estimates of what the price can actually be brought down to?
    What about the costs of newer designs like PBMR? Could a PBMR (or some other type of small, modular reactor) plant be more competitive? It would seem like, to be competitive with Coal, you would need nuclear plants that can generate in the 5-7 cent/kWh range, wouldn’t you?

  5. Rod Adams says:

    @Jeff – the Energy Information Agency publishes sales price data by state and by market segment.
    Please understand – I am focusing on PRICE here, not cost. In nominally competitive markets, prices are determined in a bidding process between sellers with buyers indicating how much they will buy at various prices.
    Of the $.107 per kilowatt-hour price that would be required to provide the equity investors in Calvert Cliffs unit 3 a 10% return on invested equity, only $0.019 is actually related to the unavoidable operational costs of running the facility. What that means is that if the facility is built, the owners will make sure that their submitted bids keep that plant running at full bore. They will use whatever difference there is between the cost of operating and the sales price to pay back the loans.
    Not operating would make it harder to pay back the money. That is a different situation for a “cheap” natural gas plant where it is often better for cash flow to stop operating when the price of gas is high enough so that it costs more than the value of the electricity that can be produced.
    Also, please note that CC unit 3 would be a first of a kind plant in the United States. The cost per unit capacity is considerably higher than what it would be for units 5-10 of the exact same design. Even though there is some world wide experience being developed for EPR, every new work force team will have its growing pains and learning curves.

  6. Jeff Schmidt says:

    Thank you for the reply. Yes, I do realize that the operational costs of a plant are extremely low, but if the plant operator/investors can’t make back a reasonable return on investment over the life of the plant, then it’s not cost effective to build the plant – in essence, it sounds like you are saying that if the plant can’t sell the power at 10.7 cents, the plant is operating at a loss even if the actual operational costs are low because it’s not recovering the costs fast enough?
    You know, that’s an interesting point about the first plant being the most expensive. I wonder if the industry could work out something where later plant operators pay (or perhaps, I think the designer/builder of the EPR design is Areva? Maybe Areva could pay) some kind of ‘royalty’ to the operators who assumed the financial burden of being the first 2 or 3 plants built, who shoulder the increased ‘first-mover’ costs?
    I kind of doubt that would actually happen, but if the nuclear industry wants to be able to bootstrap construction of new plant types, such an arrangement would be expeditious and beneficial to the who industry.
    By the way – I remember reading a year or two ago about a proposal for a new nuclear generation facility to be built in Piketon, OH, at the old DoE enrichment facility location, by Duke Energy, Areva, and a group of other partners. Do you know – I *think* the proposed plant they are considering is that same EPR design you mentioned, but I’ve not been able to get a response from Duke Energy (After submitting a contact request from the Duke Energy Ohio website, I basically got a response from a customer service rep that provided absolutely no information, and no contact info for anyone at Duke Energy who *would* be able to help me – in other words, a very polite “Go @#!& yourself”)?
    If the Piketon plant proposal is based on that same EPR design, and the Calvert Cliffs plant doesn’t get built first, is that likely to derail the Piketon plant proposal?

    • Rod Adams says:

      @Jeff – One of the big challenges associated with making the corporate financial decision to build a new nuclear plant is that many of the models used for the process either ignore or greatly discount the returns that might be available in the period after the loans are repaid.
      For the model that I ran with the FOAK EPR that has a construction cost of $6,126 per kilowatt (way up at the top end of many assumptions) the early years are a bit lean, but still profitably if the company is able to sell its electricity for $107 per MW-hour. That model includes the effects of inflation, so the price computed is in “then year” dollars, not 2010 dollars. I suspect that if energy supply decisions in the US keep going the way they seem to be going now, with everyone jumping on the “cheap gas” bandwagon and with mandated “renewable energy standards”, $107 per megawatt-hour will seem quite cheap in 2017 or 2018 when the plant might enter revenue service.
      It is also possible that good project management and knowledge sharing from the 4 EPR projects that have already been well started will help the Calvert Cliffs unit avoid some of the FOAK costs, so the actually construction could be a bit cheaper than assumed.
      I also hope that CC unit 3 will eventually work out the initial (inevitable) operational kinks fairly quickly so that by year 3 or 4 it will be able to achieve a capacity factor closer to 90 or even 95% than the 85% assumed in the model. That kind of capacity factor really improves profitability, but assuming you can achieve that level is not a good, pre-decisional choice. The hard nosed investors who are making the several billion dollar investment decision need to base their decision on realistic assumptions that allow some margin for the unexpected and still indicate a good return.
      Once the plant is paid off – and I assumed a loan term of just 15 years – the plant becomes a huge cash flow generator. Even if electricity prices increase at a rate that is 1/4th of the rate of general inflation (0.5% vice 2.0%) and even if nuclear fuel prices increase at the rate of inflation plus 25% (2.5% vice 2.0%) the plant will be making a half a billion per year in after tax profit – assuming a corporate tax rate of 39%.
      That means that the reluctant co-signature from the government would be providing a return to taxpayers of about $300 million per year in taxes collected on the profits being made by operating the plant.
      Ratepayers in the area would also be paying lower prices for electricity than they otherwise would because the supply from CC unit 3 would be increasing the overall availability of electrical energy, keeping the supply/demand balance a bit more strongly in the favor of customers.
      One more comment about your request for information about future plans – never be too surprised that you cannot get good information about what companies plan to do in the distant future. They do not really know and keep close tabs on a lot of events that might drive different decisions. Even if they did know, they have a strong incentive for not making their plans too clear to anyone, there are competitive advantages associated with a bit of mystery. Heck, in most companies most of the employees are not even in the know about future plans.

  7. Jeff Schmidt says:

    @Kit: Yeah, I didn’t expect the CSR to actually be able to answer my question, but I *do* expect that when CSR’s encounter a question they can’t answer, that it get forwarded on, or at least that you contact the customer and give them an alternate contact point for answering their question. The thing is, I looked all over their website and there’s just no good way to try to contact anyone for such a question.
    I’m thinking I might try the “Media” contact. I’m not really part of the media, but perhaps the PR people would be more helpful in answering basic questions about the project, from interested Citizens and residents of the State of Ohio.
    All I wanted to know was if Duke was still considering the project, and what type of reactor/plant design they were looking to build at that site.

  8. bigblue says:

    FYI: French utility offers to buy Constellation Energy’s stake in UniStar Nuclear

    • Jeff Schmidt says:

      I’ve worked Tech Support helpdesk for a number of years of my life (though I no longer do), at a couple of different organizations. Tech Support isn’t the same as CSR, but it’s somewhat similar. I don’t expect the CSR’s to be able to answer my question. I specifically asked them, when I sent in my question, to forward this on to someone in the company who would be better equipped to handle such questions. I had no idea they hadn’t submitted any application to the NRC – that’s part of why I was asking them.
      Keep in mind, I am NOT asking the CSR a question – I am asking DUKE ENERGY a question. Unfortunately, the general public will find it hard to get contact information for people inside the company, and is forced to go though the CSR’s first. From an organizational standpoint, CSR’s are only expected to deal with routine questions, as you previously said, but they should also forward on any non-routine questions to an appropriate party. If corporate policy is not setup to do that workflow, then it’s not the CSR who is telling me they can’t answer the question, it is DUKE ENERGY, the entire company, which is refusing to answer my very reasonable inquiries.
      It is up to the corporation to make sure that workflows are setup within their customer service organization to escalate questions that are not routine, but Duke Energy has failed to do this. That is a fundamental failure of their customer service organization and policies.
      As for why I asked them in the first place – they announced this plan in June of 2009 – I would have thought they’d have spent the past year and a half ‘planning’ this. I’m actually surprised they haven’t filed some sort of App with the NRC to start the process. This is why I submitted the form on their website – they made an announcement a year and a 1/2 ago, and have been silent since then, so I was curious if they were moving forward or not. I guess if they *are* moving forward, they are doing it slowly.
      Part of me wonders, since they were partnered with Areva at the announcment, if they are just waiting to see how other EPR projects (particularly CC3) go before they commit further?

  9. Momchil says:

    Why did you use 40 years for plant life. All reactor vendors markets their reactors for 60 years life.

    • Rod Adams says:

      @Momchil – good question. For a corporate decision analysis, the life of the plant does not actually come into play. Even the returns after the loan is paid off are severely discounted because they are so far off into the future as to not matter much.
      Again please note – my post was not about the cost of electricity but about the PRICE required to provide the cash flow that would entice a corporate board to approve the project. With a required price of $107 per megawatt hour in 2018 dollars, many board members would consider that a reasonable bet. With a required price of $138 per megawatt-hour the risk goes up considerably.

      • Jeff Schmidt says:

        It’s really a shame that all the financial planning seems to be done with such short-sighted models. Surely there must be at least a few people or companies out there who have both the money necessary to build a few nuclear plants, as well as the long-term planning mindset to make such an investment? I don’t really know, but I suspect one of the biggest reasons we don’t see more private capital going into building nuclear plants is, in large part, because investors see that past nuclear projects have been derailed, not by technical or project management issues with the plants, but because of endless lawsuits from anti-nuclear political activists and problems with the government regulatory system changing requirements constantly in the middle of projects – at least, I’ve heard some commentators say as much. Any of you have any examples of such problems that I could go look up?
        Do you think a, shall we say, ‘reasonable and stable’ regulatory environment which also didn’t allow lawsuits on so many grounds to stop or delay plant projects until the court found some actual wrong-doing, provide an environment where private capital was more readily available to fund these projects, so that government loan guarantees were less necessary?

        • Joffan says:

          Jeff, the classic example of derailment is Shoreham, but there are plenty of less drastic interventions by government that slowed nuclear construction projects to the point of loss-making.
          This catgory of project risk for past nuclear projects is so high that the loan guarantee fees must be reflecting them, either directly or indirectly, which means that the government is charging would-be constructors for risks inflicted by the government – even if a different layer. In fact, a protection racket, in common language. I’m sure there’s some law about that…
          Government loan guarantees, with fees that only reflected external risks, would hopefully encourage a most robust attitude to attempted interference with nuclear project schedules. A price guarantee, that all nuclear electricity would be bought for some reasonable rate, might be even better for encouraging build. Worked for solar.

  10. Joel Gehman says:

    Thanks for your thoughtful analysis. One point I think you might not have mentioned (or perhaps I overlooked it in your post and the comments above) is that the DOE offered suggestions to Constellation as to how it might go about reducing the risk of the project, and hence the fee necessary to guarantee the loan:
    In Constellation

  11. Rod Adams says:

    @Joel – I discounted the comment from the NT Times journalist about the supposed “offer” from the DOE to reduce risk by signing a power purchase agreement with a regulated subsidiary. I am not a lawyer, but I know a bit about rate cases and about the potential mine field of litigation that a holding company would face in trying to enforce a PPA with a regulated company that they own. I have no idea what the DOE official was thinking about and believe he or she seriously underestimated the difficulty of making such a deal happen.
    With regard to my assertion that nuclear is not different from all other human industrial activities with regard to the existence of learning curves, I stand by that assertion. There are, of course, some caveats to the inevitable law that humans can learn from experience. First of all, the activity has to be repeated; building unique new plants each time is not going to allow any learning to occur. In the few cases in the US where the same company used the same work force to build what was essentially an identical unit, the second unit cost less than the first one. In the single instance where three essentially identical units were built on the same site with the same work force, the third one was cheaper still.
    (I am talking about construction cost here – the effect of interest costs has to be factored out, especially when you remember how high interest rates were in the late 1970s through the mid 1980s.)
    There also was not much incentive for companies to try to capture the economy of unit volume – the sellers were engineering companies that liked designing new items and liked selling products for whatever price the market would bear. The customers were generally regulated monopoly utilities that actually made larger profits by investing more money in capital equipment – as long as they could convince their PUC that the investment was “prudent”. If the extra cost could be blamed on a change in federal regulations, the PUC’s were pretty consistent in allowing that as a prudent expense. In other words, there was little incentive to resist regulatory creep.
    The nuclear industry has demonstrated that it can learn – just take a good look at the experience with complex tasks like refueling outages or steam generator replacements. The industry is full of very bright people who really can reduce the effort involved in completing tasks – if they are provided the correct signals and allowed to maintain consistent designs that do not require learning a brand new task each day on the job.

  12. Lars Jorgensen says:

    Decommissioning cost ($million/Mwe capacity) 2000
    I suspect the units are off here – should be $million/Gwe.