Nuclear Power Economics: Costs, Financing, and Market Competition

This article delves into the complex economics of nuclear power, examining its investment costs, financing mechanisms, and how it competes with other energy sources in various market scenarios. It also explores the impact of carbon pricing and the role of different energy technologies in decarbonizing the grid.

Understanding Investment Costs and Construction Timelines

The initial discussion revolves around the investment cost of building a power plant, a formula applicable to any technology. A key factor influencing this cost is the construction timeline. Long construction periods, typical for "stick-built" technologies like traditional nuclear plants, significantly drive up costs due to interest accrued during construction. This highlights the motivation behind factory fabrication, which aims to reduce construction times and, consequently, overall costs.

The difference between "overnight costs" (excluding interest during construction) and "investment costs" (including it) is crucial. The calculations presented assume a 4% real interest rate, which is considered somewhat low but reflects current financing trends. This lower interest rate biases the numbers, making investment costs appear smaller over long timescales.

Plant Lifetimes and Annual Capital Charges

Nuclear plants are traditionally licensed for 40 years, but most operate for around 60 years, making them exceptionally long-lasting. For societal benefit, a 60-year operational lifetime is used to calculate annual capital charges, rather than the official 40-year licensing period.

The annual capital charge is derived from the investment cost using a specific formula that accounts for the plant's lifetime (n periods) and an interest rate (r). The choice of 'r' is critical: while accountants might use a nominal interest rate for loan repayments, for a true understanding of costs in today's dollars, a real interest rate is preferred. This approach allows for ignoring the effects of inflation on future revenues, simplifying the analysis of real costs.

Data Sources and Adjustments

The data for plant lifetimes and costs are primarily sourced from the Energy Information Administration (EIA), based on historical plant surveys. However, the operational lifetimes for nuclear and coal plants are adjusted from 40 to 60 years to reflect current life extensions and demonstrated feasibility.

Capital Charges Across Technologies

Comparing capital charges reveals why certain technologies are rarely built:

• High Capital Costs: Solar-thermal and fuel cells have very high capital costs, explaining their limited deployment.
• Natural Gas: Natural gas plants, including those with and without carbon capture and storage (CCS), are significantly cheaper in terms of capital charge, often under $100 per kilowatt per year.
• Solar with Storage: Solar with four hours of storage is comparable to natural gas in capital cost, but a fully solar grid would require more storage (e.g., 12 hours).
• Wind: Wind power has a higher capital charge than natural gas but is still competitive.
• Geothermal: Geothermal is more expensive in terms of capital charge but is dispatchable and nearly zero-carbon, making it a competitor to nuclear in regions like the Southwest United States.

It's important to note that these comparisons initially focus only on capital charges, as many of these technologies (solar, wind, geothermal, nuclear) are fuel-free and thus have very low variable costs. Solar, in particular, stands out as the cheapest carbon-free energy source, driving massive installations in countries like China.

Financing Nuclear Plants: A Unique Landscape

Despite high costs, nuclear plants are built due to unique financing mechanisms:

• Government Support: In the United States and South Korea, nuclear projects often receive "risk-free rate loans" or "loan guarantees" from the federal government. These guarantee repayment to lenders, allowing projects to borrow at very low, risk-free interest rates. In some cases, the lending bank is government-owned, effectively providing free money (to be repaid).
• Ratepayer Billing: For projects like Vogtle in the US, utilities received authorization to bill customers for nuclear power construction costs years before the plants became operational, significantly reducing financing burdens.
• Legal Settlements: Lawsuits, such as the one against Westinghouse for the Vogtle project, can also provide substantial funds.
• Low Interest Rates: The Vogtle project benefited from historically low interest rates following the 2008 financial crash, allowing them to issue bonds at rates lower than inflation, effectively making money on the loans.
• National Policy: In countries like China and France, nuclear power is often financed directly from national budgets or through government-owned utilities, bypassing competitive market financing. India also uses a hybrid model of national budget allocation and foreign loans.

These mechanisms highlight that nuclear power often operates outside purely competitive market dynamics, driven by national policy and strategic considerations rather than solely economic viability at market rates.

The Concept of Base Load and Screening Curves

The discussion then moves to how different technologies are dispatched based on demand. A "screening curve" is introduced, which plots the cost of electricity generation against the number of hours a plant operates per year (a proxy for capacity factor).

• Optimal Mix: The screening curve helps determine the optimal mix of technologies. For plants operating most of the time (base load), technologies with higher initial capital costs but lower variable costs (like geothermal or, potentially, nuclear) are preferred. For plants operating only during peak demand, technologies with lower capital costs but higher variable costs (like natural gas simple cycle) are more economical.
• Base Load Misconception: The idea that "we need nuclear for base load" is challenged. While nuclear is well-suited for base load due to its high fixed costs and low variable costs, any technology that is cheapest at high capacity factors can serve as base load. The more accurate statement is that nuclear is only suitable for base load due to its high capital costs and inflexibility in ramping up and down.
• Demand Curves: Actual electricity demand data, like that from ERCOT (Texas electrical grid), is used to create a "load duration curve." This curve sorts demand from highest to lowest, showing how many hours a certain amount of capacity is needed. By overlaying screening curves onto the load duration curve, one can determine the optimal amount of each technology to build.
• Modeling Assumptions: The importance of comprehensive modeling is emphasized. Omitting certain technologies (e.g., geothermal, hydro, wind, solar) from a screening curve analysis can lead to inaccurate conclusions about the optimal energy mix due to substitution effects.

Bidding and Marginal Costs

For existing plants, the annual payment for capital is a sunk cost. Therefore, their operational decisions are based on their "long-run marginal cost" – the cost of producing an additional unit of electricity, including maintenance. In market-based systems, power plants bid to supply electricity, with the cheapest bids being accepted first. The price paid to all dispatched generators is typically the price of the most expensive bid needed to meet demand.

Nuclear plants, with their very low marginal costs (close to zero, as they prefer not to shut down), often bid low to ensure they are dispatched. This can lead to periods of low or even negative electricity prices, especially with the influx of zero-marginal-cost renewables like solar. While some view negative prices as a problem, they are market signals for generators to reduce output and for consumers (like battery storage) to increase demand, helping stabilize the grid.

Levelized Cost of Electricity (LCOE)

LCOE is presented as a single metric for comparing electricity costs across technologies. It represents the net present value of all investment, maintenance, and fuel costs divided by the total energy produced over the plant's lifetime. A critical assumption embedded in LCOE is the plant's operational capacity factor. While useful for comparing base load technologies, LCOE is less suitable for technologies with highly variable dispatch profiles (e.g., peaker plants).

The Impact of Carbon Pricing

A sensitivity study on the "social cost of carbon" demonstrates its profound impact on the optimal energy mix:

• Natural Gas: Without a carbon tax, natural gas is dominant. As a carbon tax is introduced, the cost of natural gas (without CCS) increases, making it less competitive.
• Geothermal: Geothermal, with its low emissions, becomes a strong base load contender.
• Nuclear's Role: Nuclear power's competitiveness is highly dependent on its cost and the social cost of carbon. If nuclear can be built cheaply (e.g., at Westinghouse's projected prices and construction times), it appears earlier in the decarbonization pathway. However, if nuclear costs remain high, it only becomes competitive at very high levels of decarbonization (e.g., 90% or more) and with a significant carbon price.
• Regional Differences: The optimal mix also varies geographically. In regions with poor renewable resources (e.g., Northern United States), nuclear plays a larger role in decarbonization. In regions with abundant renewables (e.g., Southern United States), wind and solar dominate, and nuclear's role shrinks significantly.

Limitations of Current Models

The models discussed, such as GenX, are powerful optimization tools but have limitations:

• Dispatchable Technologies: The screening curve methodology is not suitable for non-dispatchable technologies like wind and solar, as it assumes plants can be run when demand occurs.
• Assumptions: Models rely on specific assumptions about technology costs, construction times, and operational parameters. For example, the MIT study on nuclear energy assumed very fast construction times for nuclear plants, which may not reflect real-world experience.
• Omitted Technologies: Excluding technologies like hydro or geothermal can lead to inaccurate results, as they could offer competitive alternatives.
• Forecast Errors and Ancillary Services: Current models often do not incorporate forecast errors, essential reliability services (like grid stability), or the dynamic nature of demand over time.

Ultimately, the role of nuclear power in a decarbonized future depends on a combination of factors: its ability to reduce costs and construction times, the societal value placed on carbon emissions, and the availability and cost of other low-carbon, dispatchable energy sources.