Framing the Analysis: Base-Case Assumptions and the Effects of Policy

To assess the competitiveness of advanced nuclear technology in comparison with other base-load options, the Congressional Budget Office estimated the levelized cost of alternatives under a reference scenario reflecting the agency’s best judgment about future market conditions and the policy environment before the enactment of the Energy Policy Act of 2005, and under alternatives that consider the effects of both carbon dioxide charges and EPAct incentives. To calculate those costs, CBO adopted base-case assumptions about an array of technical and economic choices confronting investors in new electricity-generating capacity.¹

Some of the assumptions underlying this analysis are very uncertain. For instance, large power plants using carbon capture-and-storage technologies exist only as blueprints, which makes predicting their costs difficult. Also subject to uncertainty and controversy are the costs of building a new nuclear plant in the United States, where no reactor has been ordered since the 1970s, when substantial cost overruns were the rule. Even the risks of investing in conventional coal power plants are heightened because utilities cannot anticipate with certainty whether charges on carbon dioxide emissions will be imposed in the future and, if so, at what level. For those and other reasons, financial markets may require a higher rate of return for any investment in new base-load capacity. To account for those uncertainties, CBO assessed the levelized cost of alternatives over a range of values for critical assumptions about plant costs, fuel costs, and the rates of return required by investors to finance new capacity. The results of such calculations indicate the potentially significant impact of those uncertainties on CBO’s estimates of the levelized costs of the various technologies.

Levelized cost is the minimum price at which a technology option produces electricity and generates enough revenue to pay all of a utility’s costs and still provide a sufficient return to investors. In its analysis, CBO projected expected cash flows in order to find the minimum real (inflation-adjusted) price of electricity at which revenue exceeded costs by enough to encourage investment in the construction of new capacity based on each technology. (A detailed description of the approach used to estimate levelized costs in this study can be found in CBO’s Web supplement, The Methodology Behind the Levelized Cost Analysis.)

Levelized costs affect investment decisions made by both merchant generators and regulated utilities. If the levelized cost of a technology exceeded anticipated prices for electricity, merchant generators would be unlikely to invest in new capacity based on that technology because the expected return would not justify the amount of risk they would have to incur. State utility commissions commonly direct regulated utilities to meet anticipated demand for new capacity using the technology with the lowest levelized cost.

Both types of utilities typically fund costs that are incurred before a plant begins operating through a combination of debt and equity financing. The revenue that results from the sale of electrical power is first used to pay the plant’s operating costs, including the purchase of fuel. After deductions are made for corporate income taxes and debt payments, the remaining revenue is paid to equityholders. CBO estimated the lowest constant real price of electricity at which the return to equity was adequate to attract the investment for up-front costs.

Whereas utilities’ decisions are made on a site-by-site basis, the levelized costs estimated by CBO are intended to give a representative cost for each technology. One technology would probably not have the lowest cost in all parts of the country and other factors could be considered, but a technology with the lowest representative levelized cost would most likely be a common choice for new capacity.

Another consideration is that levelized costs do not indicate which technology produces electricity most efficiently because they capture the utility’s cost of generating electricity rather than the actual cost. Government incentives that directly subsidize electricity generation or transfer financial risk from investors to the public decrease levelized costs in comparison to actual costs. Because corporate income taxes increase utilities’ costs relative to actual costs, they could make efficient, capital-intense technologies relatively costly. Last, levelized costs include only those costs that markets and current laws require utilities to pay. For example, a carbon-emitting technology might have the lowest levelized cost in the absence of a carbon dioxide charge but not be the most efficient technology because the levelized cost would not account for the damage caused by carbon dioxide emissions.

The base-case assumptions necessary to estimate the levelized cost of plants that employ alternative technologies include the cost of building a plant as if it was built and paid for immediately (so-called overnight costs); the return that investors require to finance that construction and subsequent operations; and the cost of operating the plant (largely composed of fuel costs). As a first approximation, CBO relied on the Energy Information Administration’s most recent projections for its base-case assumptions and compared those with assumptions adopted in two prominent studies of generation alternatives, one conducted by researchers at the Massachusetts Institute of Technology (MIT) and the other by analysts at the International Energy Agency.²

CBO’s base-case assumptions include overnight costs of about $2.4 million for each megawatt of capacity for new nuclear plants and innovative coal plants but lower costs for conventional coal, conventional natural gas, and innovative natural gas technologies. For nuclear and innovative coal and natural gas technologies, the assumptions are intended to represent plants built over the next decade but do not incorporate the first-of-a-kind costs that are assumed to be covered by federal research and development programs. The estimate for nuclear plants, taken from the EIA’s most recent analysis, is roughly 10 percent above the estimate of overnight costs used in MIT’s study, which was published in 2003, before construction costs for most types of power plants surged. CBO also calculated construction costs for each technology using alternative assumptions designed to capture plausible variations in those costs. For nuclear and innovative coal technologies, CBO considered construction costs ranging from about $1.2 million per megawatt of capacity to roughly $4.8 million per megawatt of capacity. The breadth of that range reflects the uncertainty associated with the cost of building new nuclear plants in the United States and is wide enough to capture plausible further increases in construction costs, which could affect conventional fossil-fuel plants as well.

CBO’s assumption about the cost of building new nuclear power plants in the United States is particularly uncertain because of the industry’s history of construction cost overruns. For the 75 nuclear power plants built in the United States between 1966 and 1986, the average actual cost of construction exceeded the initial estimates by over 200 percent (see Table 2-1). Although no new nuclear power plants were proposed after the partial core meltdown at Three Mile Island in 1979, utilities attempted to complete more than 40 nuclear power projects already under way. For those plants, construction cost overruns exceeded 250 percent.3 (An average of 12 years elapsed between the start of construction and the point at which the plants began commercial operation. The overruns in overnight costs did not include additional financing costs that were attributable to post-accident construction delays.)⁴

Construction Starts			Average Overnight Costs^a
Year Initiated		Number of Plants^b	Utilities' Projections (Thousands of dollars per MW)	Actual (Thousands of dollars per MW)	Overrun (Percent)


1966 to 1967		11	612	1,279	109
1968 to 1969		26	741	2,180	194
1970 to 1971		12	829	2,889	248
1972 to 1973		7	1,220	3,882	218
1974 to 1975		14	1,263	4,817	281
1976 to 1977		5	1,630	4,377	169

	Overall Average	13	938	2,959	207

Source: Congressional Budget Office (CBO) based on data from Energy Information Administration, An Analysis of Nuclear Power Plant Construction Costs, Technical Report DOE/EIA-0485 (January 1, 1986).

Notes: Electricity-generating capacity is measured in megawatts (MW); the electrical power generated by that capacity is measured inmegawatt hours (MWh). During a full hour of operation, 1 MW of capacity produces 1 MWh of electricity, which can power roughly 800 average households.

The data underlying CBO’s analysis include only plants on which construction was begun after 1965 and completed by 1986.

Data are expressed in 1982 dollars and adjusted to 2006 dollars using the Bureau of Economic Analysis’s price index for private fixed investment in electricity-generating structures. Averages are weighted by the number of plants.

The base-case assumption adopted in this analysis for nuclear power plants’ overnight costs recognizes that history but also allows for countervailing factors, such as changes in the U.S. regulatory process and other countries’ recent experience with new reactor designs. In 1989, the Nuclear Regulatory Commission developed an alternative process for obtaining the licensing necessary to operate a nuclear power plant. That revised process is intended to reduce cost uncertainties by allowing utilities to fulfill more regulatory requirements before beginning construction, thereby reducing midconstruction design changes that contributed to overruns in the past.

The experience of a Japanese utility, the Tokyo Electric Power Company (TEPCO), in the mid-1990s also appears to support CBO’s base-case assumption about construction costs. According to the 2003 MIT study, verifiable data indicate that TEPCO constructed two advanced boiling-water reactors at costs and schedules close to manufacturers’ estimates.⁵ However, a Finnish utility that is building a reactor based on a different design, an advanced pressurized-water reactor, continues to have difficulty adhering to original cost estimates. By 2007, that project, initially estimated to cost €3 billion, had fallen at least 18 months behind schedule, causing costs to increase by €700 million.⁶

Even if construction proceeds on schedule, utilities still incur substantial financing costs because power plants take years to build, and financing costs for construction extend over the decades that a plant generates electricity. CBO’s assumptions about financing costs are a synthesis of the financial analyses presented in the studies by EIA and MIT. Those assumptions are encapsulated by the real rate of return that investors require to assume the risk of paying up-front construction costs.⁷ CBO used a real rate of return of 10 percent, which falls within the range of rates of return given in the other studies (see Table 2-2).⁸ The 10 percent rate of return was used for each technology, reflecting that the level of financial risk is similar across commercially viable projects. The MIT study assumed that a higher rate of return would be required for nuclear technology than for conventional fossil-fuel technologies; however, that 2003 study was published before much of the volatility in natural gas prices and when future federal carbon dioxide constraints may have appeared less likely.⁹ But nuclear plants could still be a riskier investment than competing alternatives, or the rate of return could vary for all technologies. In addition to the base-case assumption of a 10 percent rate of return, CBO considered the competitiveness of nuclear power under lower and higher rates (as shown in Table 2-2).

Notes: EIA=Energy Information Administration; IEA=International Energy Agency; MIT=Massachusetts Institute of Technology.

Real rates of return are rounded to the nearest quarter of a percentage point. CBO calculated those rates on the basis of the inflation rates and nominal rates of return used in each study. The underlying nominal rates of return for the studies conducted by EIA and IEA represent the weighted average cost of capital assumed in those studies. The nominal rates of return for the MIT study were constructed by taking the ratio of financing charges to balances in CBO’s replication of the MIT model.

The cost of fuel is one of the most significant operating costs included in CBO’s estimates of the levelized cost of options for generating electricity. The base-case assumption for nuclear power is that $8 (in 2006 dollars) in fuel costs are incurred for each megawatt hour of electricity generated (see Table 2-3). That contrasts with $16 for conventional coal-fired plants and $40 for conventional natural gas plants.¹⁰ Those assumptions are based on long-term projections by EIA. In the past, fuel costs have proved difficult to predict, particularly the price of natural gas. (See Figure 2-1 for fluctuations in fuel prices between 1995 and 2006.) In addition to those base-case assumptions, CBO also estimated levelized costs using alternative assumptions intended to capture most plausible variations in fuel costs.

		Advanced Nuclear	Conventional Coal	Conventional Natural Gas	Innovative Coal	Innovative Natural Gas


Construction
	Time (Years)	6	4	3	4	3
	Overnight costs (Thousands of dollars per MW)^a	2,358	1,499	685	2,471	1,388

Operating Costs
	Fuel (Dollars per MWh)	8	16	40	17	52
	Fixed operation and maintenance(Dollars per MWh)	8	4	1	6	3

Source: Congressional Budget Office based on data from the Energy Information Administration.

Notes: Electricity-generating capacity is measured in megawatts (MW); the electrical power generated by that capacity is measured in megawatt hours (MWh). During a full hour of operation, 1 MW of capacity produces 1 MWh of electricity, which can power roughly 800 average households.

The reference scenario excludes both the effects of prospective carbon dioxide charges and the impact of incentives provided under the Energy Policy Act of 2005.

Advanced nuclear technology refers to third-generation reactors. Conventional coal power plants are assumed to use pulverized coal technology, which produces energy by burning a crushed form of solid coal. Conventional natural gas power plants are assumed to convert gas into electricity using combined-cycle turbines.

Source: Congressional Budget Office based on data from the Energy Information Administration (EIA).

Note: The percentage changes are based on prices in 2006 dollars, with adjustments for inflation made using the gross domestic product price index. Prices for all fuels equal the average cost at which those fuels are delivered to power plants, as measured by EIA.

The cost of disposing of the used (spent) fuel generated by nuclear fission is currently unique to that fuel source. (However, if carbon dioxide charges were imposed in the future, they could be seen as similar because utilities could have to assume the full cost of their fuel choice—that is, they would have to pay for the damage inflicted by emitting carbon dioxide.) CBO’s levelized cost estimate for nuclear power includes a $1 per megawatt hour charge to cover the cost of such disposal.¹¹

Other important considerations in CBO’s analysis include the fixed cost of operating and maintaining a power plant (that is, the costs that do not vary with the amount of electricity produced) and the percentage of maximum electricity production achieved by the plant (the "capacity factor"). For nuclear plants, fixed operating and maintenance expenses include the cost of providing security and monitoring safety systems. On the basis of EIA’s analysis, the combination of such costs is assumed to be five times those incurred by a conventional natural gas coal plant and twice those incurred by a conventional coal plant. The fixed operating costs of innovative coal and natural gas plants are assumed to be higher than those of conventional fossil-fuel alternatives but lower than those of nuclear power plants.

Capacity factors range from about 80 percent to 90 percent, depending on the technology. Those factors represent the maximum rate at which each technology could physically operate, as determined by EIA. Historically, utilities have utilized natural gas capacity at much lower rates because, in comparison with coal-fired and nuclear technologies, natural gas has typically been an expensive source of base-load power. But because this study evaluates the competitiveness of natural gas as a base-load alternative, CBO assumed that options using that fuel would operate at their maximum capacity factor.

Expanding the reference scenario to include the effects of carbon dioxide charges and EPAct incentives changes CBO’s estimates of the levelized cost—and potential competitiveness—of each technological option. Carbon dioxide charges would raise the levelized cost of fossil-fuel alternatives but not the cost of nuclear power. Conversely, EPAct incentives reduce the levelized cost of nuclear power and innovative fossil-fuel options in comparison with that of conventional fossil-fuel technologies.

In addition to the uncertainty inherent in the base-case assumptions, some of the assumptions linking policy to levelized cost are subject to uncertainty. In particular, when estimating levelized cost, CBO assumed that the full benefits of EPAct incentives would be available to both nuclear and innovative fossil-fuel options. In some instances, CBO also used simplifying assumptions to incorporate policy into levelized cost. For example, carbon dioxide charges were assumed to be levied only on smoke-stack emissions (those resulting directly from the operation of a power plant). In reality, all technologies have the potential to produce additional emissions from the construction and decommissioning of a facility, as well as from the processing of fuel. Such additional life-cycle emissions are not included in this analysis because they are difficult to measure precisely and are probably an order of magnitude smaller than the stack emissions from coal or natural gas power plants.¹²

Carbon dioxide constraints could encourage the use of nuclear technology by increasing the cost of generating electricity with fossil fuels. The effect is most pronounced for coal, which emits nearly a metric ton of carbon dioxide for every megawatt hour of electricity produced. The effect on conventional generators fueled by natural gas would be less because they emit carbon dioxide at roughly half the rate of the average coal plant.

Because competing base-load alternatives emit carbon dioxide, the attractiveness of financing a new nuclear power plant depends on investors’ expectations about the costs of emitting carbon dioxide over the operating life of that plant. To the extent that carbon dioxide charges are expected, investment in new nuclear capacity would be more attractive relative to both the construction of new fossil-fuel capacity and the continued use of existing fossil-fuel capacity. Many investors appear to anticipate some form of carbon dioxide charge in the near future; a survey conducted by Cambridge Energy Research Associates in 2006 found that about 80 percent of utility executives expected a carbon dioxide charge to be implemented within the next 10 years.¹³

Although the imposition of carbon dioxide constraints would not directly decrease the cost of operating nuclear power plants, such a policy would increase the cost of operating fossil-fuel power plants, which in all their variants emit at least some carbon dioxide, and consequently make new nuclear capacity a more attractive source of base-load generation. Newly built power plants based on conventional fossil-fuel technology are designed to burn fuel more efficiently than plants built in the past, but their emissions would still be substantial enough for the cost of producing electricity to be sensitive to carbon dioxide charges. Innovative fossil-fuel power plants that capture and store carbon dioxide are assumed to emit only about 10 percent of the carbon dioxide discharged into the atmosphere by the lowest emitting conventional plants that burn fossil fuel—but they still emit carbon dioxide. As of 2007, such carbon dioxide capture-and-storage technologies had not been used at commercial power plants, but those technologies could be an option for new base-load capacity by the time new nuclear plants were deployed and might be the most competitive alternative to nuclear technology under carbon dioxide charges.¹⁴

CBO estimated the cost of emitting carbon dioxide using hypothetical charges based on the levels of carbon intensity for coal and natural gas reported in the MIT study. Those hypothetical charges were assumed to be proportional to the amount of carbon dioxide emitted by each technology (see Figure 2-2).

Notes: Electricity-generating capacity is measured in megawatts (MW); the electrical power generated by that capacity is measured in megawatt hours (MWh). During a full hour of operation, 1 MW of capacity produces 1 MWh of electricity, which can power roughly 800 average households.

Conventional coal power plants are assumed to use pulverized coal technology, which produces energy by burning a crushed form of solid coal. Conventional natural gas power plants are assumed to convert gas into electricity using combined-cycle turbines. Both innovative coal and innovative natural gas technologies are assumed to capture and store most carbon dioxide (CO2) emissions.

CBO’s analysis assumes that coal contains approximately 0.095 metric tons per million British thermal units (Btu) of CO2 and that natural gas contains 0.054 metric tons per million Btu. It also assumes that existing conventional coal technology burns 10.463 million Btu of coal per MWh of electricity and that existing conventional natural gas technology burns 8.401 million Btu of natural gas per MWh of electricity. See CBO’s Web supplement for assumptions underlying the analysis.

EPAct provides or extends numerous incentives for generating electricity using nuclear, renewable, and innovative fossil-fuel technologies. In general, the incentives lower the cost of nuclear and other innovative technologies in comparison to conventional fossil-fuel alternatives. Among the programs reauthorized under EPAct were the Nuclear Power 2010 program and FutureGen (originally authorized in 2002 and 2003, respectively).¹⁵ Under those programs, the federal government shares with industry the cost of researching, developing, and deploying advanced nuclear power plants and innovative coal-fired plants that incorporate CCS technology. Other incentives—in particular, loan guarantees, investment tax credits, and insurance against regulatory delays—are intended to encourage investment in nuclear power and innovative fossil-fuel technologies. Another set of EPAct policies provides production incentives for operating advanced nuclear power plants once construction is complete. (Table 1-1 lists and describes in more detail the incentives provided by EPAct.)

The lessons learned in developing and implementing a particular design for a new power plant could reduce the cost of building additional power plants with similar designs. First-of-a-kind costs could be especially large in the nuclear industry because the technology is complex, and utilities have little experience in navigating the revised regulatory process for obtaining a construction and operating license. To the extent that the original utility does not have exclusive rights to build any additional plants, part of the gains in knowledge and experience arising from the initial investment could be captured by other utilities that build plants later. As a result, utilities might underinvest in new technologies because they would not retain enough of the benefit such investment produced.¹⁶

The Nuclear Power 2010 pays a share of FOAK costs for advanced nuclear technology—to increase the amount of investment in its development.¹⁷ To estimate the reduction in costs attained by plants benefiting from that program, CBO included FOAK costs for licensing and design in supplementary analysis.

A second set of incentives encourages investment in the construction of power plants that use advanced nuclear and innovative fossil-fuel technologies. They include a loan guarantee program that insures the debt for such technologies, another insurance program that provides advanced nuclear technologies protection against the cost of certain delays in the start of operation, and tax incentives for investment in innovative coal technologies.

Such incentives could be viewed as countering negative effects on investment caused by taxes on capital income. For instance, corporate income taxes, as well as taxes on capital gains, dividends, and interest income, act as proportional surcharges on investing in the construction of power plants. Such taxes could cause utilities to prefer technologies that were less capital-intensive. In particular, capital costs make up a relatively large portion of the cost of producing electricity using nuclear or innovative coal power plants because those plants are relatively expensive to build. As a result, taxes on capital income might encourage utilities to build conventional fossil-fuel power plants, which have lower capital costs. Investment incentives could counter potential bias against capital-intensive technologies caused by taxes on capital income.

Loan guarantees and insurance against delays reduce the financial risk of investing in advanced nuclear power plants by transferring risk to the public. The reduced risk means investors would incur lower costs for financing construction and other activities before a plant began operating. However, economic theory suggests that such incentives cause recipients to invest in excessively risky projects because they do not bear all the cost of a project’s failure. The federal government also provides investment subsidies through investment tax credits, which reduce tax liability in proportion to construction expenditures.

A third set of incentives encourages not only the construction of nuclear plants but also their continued operation. Those incentives indirectly encourage investment by making operation more profitable. Incentives supporting operation include a production tax credit, a limit on liability for nuclear accidents, and a tax incentive to reduce the cost of disposing of radioactive waste, which is a byproduct of operating nuclear plants.¹⁸

Such subsidies could be viewed as compensating utilities that choose zero-emissions technologies, such as nuclear, for the potential public benefits of mitigating carbon dioxide emissions; however, such subsidies are inefficient and counterproductive in comparison to charges for emitting carbon dioxide.¹⁹ Because production tax credits reduce the price of electricity, consumers might use electrical power less efficiently and expand the gap between the price of electricity and its cost to society at the expense of the general taxpayer. Alternatively, a charge on carbon dioxide emissions that was representative of the damage those emissions cause would equate the price of electricity to its social cost. Such a price would lead to more-efficient use of electricity, because the utility and consumer, rather than the general taxpayer, would pay for the cost of carbon dioxide emissions.

See John Deutch and others, The Future of Nuclear Power: An Interdisciplinary MIT Study (July 2003); and International Energy Agency, World Energy Outlook (2006).

The calculation is based on data from Energy Information Administration, An Analysis of Nuclear Power Plant Construction Costs, DOE/EIA-0485 (1986). Those data include only plants on which construction was begun after 1965 and completed by 1986.

The 10 percent rate of return is based on 45 percent debt financing and 55 percent equity financing. Debt is assumed to be repaid at a rate of return of 8 percent over 20 years, and equity is assumed to be repaid at an average rate of return of 14 percent over the 40 years the plant is assumed to operate.

See "Coal Utilities Say They Do Not Fear Risk to Credit, Despite Moody’s Warning on Carbon Burdens," Platts Electric Utility Week (March 3, 2008), p. 1. According to that report, Moody’s Investors Services has warned that the prospect of future carbon dioxide charges may adversely affect the credit rating of utilities and thus raise the cost of capital for investment in conventional coal-fired generation.

See Department of Energy, Nuclear Waste Fund Fee Adequacy: An Assessment, DOE/RW-0534 (May 2001), p. 1.

After canceling the original FutureGen program—which would have funded the construction of a single coal plant with CCS technology that also would have generated hydrogen for commercial purposes—the Department of Energy now plans to fund research and development for multiple innovative coal plants that use CCS technology.

Although Nuclear Power 2010 and FutureGen have been two of the largest EPAct research and development incentives for the electricity industry, other cost-sharing programs exist for innovative coal technologies, renewable energy technologies, and fourth-generation nuclear reactors.

The EPAct expansion of the preferential tax treatment of decommissioning funds reduces the private cost of cleaning and securing a nuclear facility once it is retired, which primarily involves the disposal of low-level radioactive waste. The federal government also plays a role in the long-term disposal of spent fuel, but that program is not addressed by EPAct.

Study	Real Rate of Return (Percent)	Capital Recovery Period^a (Years)


CBO	8-3/4 – 12-1/2	40
EIA	9-1/4	20
IEA	8 – 11	40 – 25
MIT	8 – 11-1/4	25 – 40