The current and future projected cost of new electricity generation capacity is a critical input into the development of energy projections and analyses. The cost of new generating plants plays an important role in determining the mix of capacity additions that will serve growing loads in the future. New plant costs also help to determine how new capacity competes against existing capacity, and the response of the electricity generators to the imposition of environmental controls on conventional pollutants or any limitations on greenhouse gas emissions.

The current and projected future costs of energy-related capital projects, including but not limited to new electric generating plants, have been subject to considerable change in recent years. EIA updates its cost and performance assumptions annually, as part of the development cycle for the Annual Energy Outlook (AEO). For the AEO2011 cycle, EIA commissioned an external consultant to develop current cost estimates for utility-scale electric generating plants.[1] This paper briefly summarizes the design of the project and provides a summary of its main findings, including a comparison of the new estimates to those used in AEO2010. The final section discusses how EIA uses information on cost and other factors in modeling technology choice in the electric power sector.

Developing updated estimates: key design considerations
In order to maximize its value to EIA and external energy analysts, the project focused on gathering current information regarding the "overnight"[2] cost for a wide range of generation technologies, while taking care to use a common boundary in the costing exercise across those technologies. The cost estimates for each technology were developed for a generic facility of a specific size and configuration, and assuming a location without unusual constraints or infrastructure needs.

Current information is particularly important during a period when actual and estimated costs have been evolving rapidly, since the use of up-to-date cost estimates for some technologies in conjunction with estimates that are two, three, or even five years old for others can significantly skew the results of modeling and analysis. Where possible, costs estimates were based on information regarding actual or planned projects available to the consultant. When this information was not available, project costs were estimated by using costing models that account for current labor and material rates that would be necessary to complete the construction of a generic facility.

The use of a common boundary for costing is also very important. From experience in reviewing many costing studies for individual technologies, EIA is well aware that differences in practices regarding the inclusion or exclusion of various components of costs can have a large impact on overall cost estimates. This includes the categories of civil and structural costs (e.g., allowance for site preparation, drainage, underground utilities, and buildings), project indirect costs (e.g., a construction contingency), and owners costs (e.g., development costs, preliminary feasibility and engineering studies, environmental studies and permitting, legal fees, insurance costs, property taxes during construction, and the electrical interconnection costs, including a plant switchyard and tie-in to nearby transmission).

Summary of updated overnight capital costs estimates and comparison to information used in AEO2010
Table 1 summarizes the updated cost estimates for the generic utility-scale generation plants represented in EIA's model, including 7 powered by coal, 6 by natural gas, 3 by solar energy, 2 each by wind, hydro, biomass, and geothermal power, and 1 each by uranium and municipal solid waste. For some plant types there are several options shown to better represent the range of plants that might be built and their costs. For example, both single unit and dual unit advanced pulverized coal plants are shown, because many plants include multiple units and the costs associated with the dual unit configuration might better reflect the costs of most plants built.  Similarly, solar photovoltaic technologies include a relatively small 7 MW system and a much larger 150 MW system, because there is such variance in the sizes of the facilities being considered. The nominal capacity of the generic plants ranges from a 7 megawatt (MW) solar plant to a 2,236 MW advanced dual-unit nuclear plant, reflecting the significant variation in the scale of utility applications. Each technology is characterized by its overnight capital costs, heat rate (where applicable), non-fuel operations and maintenance costs, and, though not shown in Table 1, its environmental characteristics.

• Coal & Nuclear: The updated overnight capital cost estimates for coal and nuclear power plants are 25 to 37 percent above those in AEO2010. The higher cost estimates reflect many factors including the overall trend of rising costs of capital intensive technology in the power sector, higher global commodity prices, and the fact that there are relatively few construction firms with the ability to complete complex engineering projects such as a new nuclear or advanced coal power plant. The study assumes cost-sharing agreements between the project owner and the project construction contractors are reflective of those recently observed in the marketplace. As shown in Table 1, dual unit coal and nuclear plants generally have lower overnight costs per kilowatt than single-unit plants, reflecting their ability to take advantage of redundancies and scale economies in onsite infrastructure such as wastewater management and environmental controls to reduce the estimated total per-kilowatt cost of the project.
• Natural Gas: The updated cost estimates for natural gas combined cycle and combustion turbines generally remained similar to those of AEO2010
• Solar: The overnight capital costs for solar thermal and photovoltaic technologies dropped by 10 percent and 25 percent, respectively. The decrease in the cost of photovoltaics was due to the assumption of larger plant capacity and falling component costs.
• Onshore Wind: Overnight costs for onshore wind increased by about 21 percent relative to AEO 2010 assumptions. This is based on a specification for a new, stand-alone wind plant including all owners' costs and may differ from other reported costs in the literature, which are not fully characterized and may include sites that are built along side existing plants (and are thus able to avoid some amount of infrastructure costs).
• Offshore Wind: While offshore wind plants have been built in Europe, there have only been proposals in the United States, with final permitting only recently issued on the first of these proposals. The updated costs, some 50 percent higher than AEO 2010 estimates, are consistent with substantial first-of-a-kind costs that would likely be encountered when building projects in the United States, which largely lacks the unique infrastructure, needed to support this type of construction.
• Geothermal: Geothermal costs are highly site-specific, and are represented as such in the AEO estimates. The updated cost estimate is over 50 percent higher than the same site in AEO 2010.
• Biomass: Biomass capital costs are largely unchanged from AEO2010. However, the technology represented by the costs has changed significantly. Prior estimates were for a highly efficient plant employing gasification and a combined cycle generator; the new estimate is for a significantly less efficient direct combustion boiler. The lower operating efficiency (and therefore higher operating cost) for the biomass plant considered in the updated cost estimate implies a reduced attractiveness of investment in new biomass generation at an overnight cost similar to that for the more efficient biomass plant characterized in AEO2010.

While estimates of the current cost of generic electric generation capacity of various types are one key input to EIA's analysis of electricity markets, the evolution of the electricity mix in each of the 22 regions to be modeled in AEO2011[3] is also sensitive to many other factors, including the projected evolution of capital costs over the modeling horizon, projected fuel costs, whether wholesale power markets are  regulated or competitive, the existing generation mix, additional costs associated with environmental control requirements, load growth, and the load shape. Almost all of these factors can vary by region, as do capacity factors for renewable generation, operations and maintenance costs associated with individual plants, and cost multipliers applied to the generic estimates of overnight capital costs outlined in Tables 1 and 2. The next section provides a brief overview of some of the relevant issues, which are described in more detail in the description of the Electric Market Module included in the 2010 edition of the documentation for EIA's National Energy Modeling System.

EIA's analysis of technology choice in the electric power sector
Estimates of the overnight capital cost of generic generating technologies are only the starting point for consideration of the cost of new generating capacity in EIA modeling analyses. EIA also considers regional variation in construction costs, the structure of wholesale power markets that affect financing costs, the length of time required to bring each type of plant into service, and the capacity availability factors for solar and wind generation plants. EIA also accounts for three distinct dynamic forces that drive changes in plant cost over time. One is the projected relationship between rate of inflation for key drivers of plant costs, such as materials and construction costs, and the overall economy-wide rate of inflation. A projected economy–wide inflation rate that exceeds projected inflation for key plant cost drivers results in a projected decline in real (inflation-adjusted) capital costs. Projected capital costs also reflect projected technology progress over time. Learning-by-doing, which allows for additional reductions in projected capital costs as a function of cumulative additions new technologies, has a further effect on technology costs.  See the AEO2010 assumptions and model documentation for more details.[4]

Levelized cost is often cited as a convenient summary measure of the overall competiveness of different generating technologies. Levelized cost represents the present value of the total cost of building and operating a generating plant over an assumed economic life, converted to equal annual payments and expressed in terms of real dollars to remove the impact of inflation. Levelized costs, which reflect overnight capital cost, fuel cost, fixed and variable O&M cost, are a useful indicator of the competitiveness of different generation technologies. For technologies such as solar and wind generation that have no fuel costs and relatively small O&M costs, levelized cost changes in rough proportion to the estimated overnight capital cost of generation capacity. For technologies with significant fuel cost, both fuel cost and overnight cost estimates significantly affect levelized cost. Thus, while Table 2 shows little change between the updated capital cost estimates for natural gas combined cycle plants and those used in AEO2010, improved supply prospects for natural gas that will be incorporated in AEO2011 result in lower projected prices that in turn lower the levelized cost of gas-fired generation and improve the attractiveness of operating and adding gas-fired generation technologies.

It is important to note, however, that actual investment decisions are affected by numerous factors other than levelized costs. The projected utilization rate, which depends on the load shape and the existing resource mix in an area where additional capacity is needed, is one such factor. The existing resource mix in a region can directly affect the economic viability of a new investment through its effect on the economics surrounding the displacement of existing resources. For example, a wind resource that would primarily back out existing natural gas generation will generally have a higher value than one that would back out existing coal generation under fuel price conditions where the variable cost of operating existing gas-fired plants exceeds that of operating existing coal-fired plants.

A related factor is the capacity value, which depends on both the existing capacity mix and load characteristics in a region. Since load must be balanced on a continuous basis, units whose output can be varied to follow demand generally have more value to a system than less flexible units or those whose operation is tied to the availability of an intermittent resource. Policy-related factors, such as investment or production tax credits for specified generation sources, can also impact investment decisions. Finally, although levelized cost calculations are generally made using an assumed set of capital and operating costs, the inherent uncertainty about future fuel prices and future policies, may cause plant owners or investors who finance plants to place a value on portfolio diversification. EIA considers all of these factors in its analyses of technology choice in the electricity sector.

In sum, while overnight cost estimates are important inputs for EIA modelers and other analysts, they are not the sole driver of the choice among electric generation technologies. Users interested in additional details regarding these updated cost estimates should review the consultant study prepared by R.W. Beck for EIA in the appendix of the compete report linked to above.

Footnotes
1 EIA's electricity modeling includes both combined heat and power (CHP) technologies as well as a variety of distributed generation technologies, but those technologies were not addressed in the study, which focused on technologies within the electric power sector.

2 "Overnight cost" is an estimate of the cost at which a plant could be constructed assuming that the entire process from planning through completion could be accomplished in a single day. This concept is useful to avoid any impact of financing issues and assumptions on estimated costs. Starting from overnight cost estimates, EIA's electricity modeling explicitly takes account of the time required to bring each generation technology online and the costs of financing construction in the period before a plant becomes operational.

3 In AEO2010 and prior editions, the continental U.S., excluding Alaska, was divided in 13 regions for purposes of electricity modeling. The 22 region model that will be used starting with AEO2011 will allow for better representation of policy boundaries and market structure at the State level.

4 Assumptions and model documentation for the Annual Energy Outlook 2010 are available at http://www.eia.gov/oiaf/aeo/index.html.