Broken Windows and Electricity Generation:

The Cost of Prematurely Closing Existing Plants


Travis Fisher
Institute for Energy Research
Washington, DC
Mark Febrizio 
Policy Associate
Institute for Energy Research
Washington, DC

 “There is only one difference between a bad economist and a good one: the bad economist confines himself to the visible effect; the good economist takes into account both the effect that can be seen and those effects that must be foreseen.” – Frederic Bastiat


The configuration of the electricity generation sector has long been shaped by public policy. From the early days when Thomas Edison lavished New York politicians with expensive dinners in order to build the Pearl Street Station to the evolution of modern electricity markets, policymakers have heavily influenced the makeup of the electricity generation industry.[1] In the context of current environmental policy—driven by regulations from the Environmental Protection Agency (EPA)—electricity generation faces new and unprecedented intervention by federal policymakers.

As energy economists, a fair question to ask in today’s policy climate is: Do policymakers recognize the economic tradeoffs in electricity generation policy?

When weighing the merits of different sources of electricity — such as affordability, reliability, sustainability — it is vital for policymakers to have the right data to inform their decision-making. Nevertheless, accurate information about the cost of electricity generation is harder to find than one would think.

Due to differences in capital costs, capacity factors and fuel costs, among other things, it is difficult to make judgments between natural gas, coal, wind, solar, hydropower, nuclear, and so on. However, the U.S. Energy Information Administration (EIA) offers a useful estimate of the levelized cost of electricity (LCOE)—“the per-kilowatthour cost (in real dollars) of building and operating a generating plant over an assumed financial life and duty cycle”—for new generation resources.[2] Essentially, EIA’s LCOE is an effective way of comparing the “overall [competitiveness] of different generating technologies,” which are newly constructed.[3]

Even with EIA’s estimates of LCOE, though, policymakers lack vital information about the cost of electricity generation. EIA’s LCOE would be sufficient for determining the most affordable energy sources for electricity generation in a world where existing generation sources are permitted to operate until the end of their economic lives. But in the current policy context, some power plants are retiring early due to regulations that either make them uncompetitive or outright force them to shut down. As a result, a cost comparison between existing resources and new sources is required to construct a complete economic picture.

A recent study commissioned by the Institute for Energy Research (IER) estimates the levelized cost of electricity for existing power plants, which allows policymakers to compare the cost of existing generation resources with the cost of new resources.[4] The following is a summary of IER’s findings:




Federal Policy is Driving Electricity Generation Choices

At the beginning of the 21st century, coal-fired power plants provided more than half of the electricity for the United States (52%), with nuclear (20%), natural gas (16%), renewable energy (9%, including hydropower), and 3% of petroleum and other liquids rounding out the list.[5] Federal policies are already altering the energy landscape by subsidizing certain types of energy, heavily regulating others, and, overall, modifying the economic incentives of electricity generation. By 2013, the most recent year of data in the EIA reference case, coal’s generation share declined to 39%, and generation by natural gas and renewable energy rose to 27% and 13%, respectively.[6] In April of 2015, the amount of electricity generated using natural gas briefly surpassed coal for the first time ever.[7]

Although low prices have spurred the expansion of natural gas and made coal less economical on the margin, federal policies are the key drivers of this trend and will only become more influential in the coming years. These policies fall into two main categories: (1) subsidies for wind and solar power, which artificially lower the cost of those technologies, and (2) environmental regulations, which add substantial costs to existing units and threaten to shut them down before the end of their economic lives.[8]

In particular, regulations promulgated by the U.S. Environmental Protection Agency (EPA) are inducing the retirement of existing power plants. A 2014 report by the Government Accountability Office indicated that EPA regulations were influencing power companies to accelerate the retirement of coal-fueled generating capacity.[9] According to the EIA, 40 gigawatts (GW) of coal-fired capacity and 46 GW of natural gas or oil-fired capacity will close by 2040 under the reference case. When factoring in the EPA’s Clean Power Plan, the proposed regulation of carbon dioxide emissions from existing generating units, “90 GW of coal-fired capacity and 62 GW of natural gas/oil-fired capacity retire by 2040,” assuming no further extension (or expansion) of the policies.[10]

Notably, the Clean Power Plan includes elements of both types of federal policies above—it encourages new wind and solar resources and discourages existing coal-fired resources. Building Blocks 2 and 3 of the plan mandate the expanded use of “lower-emitting power sources” like natural gas and increased capacity of renewable energy sources, respectively.[11]

New electricity demand is not driving the increase of new renewable generating resources. Between 2004 and 2014, electricity demand in the U.S. only grew by 0.3% per year.[12] Neither is the end-of-life closure of existing power plants. In fact, data from the Federal Energy Regulatory Commission (FERC) and EIA “indicate that most existing power plants could remain economically viable for years or decades beyond their current age.”[13] In essence, federal policies are prematurely shutting down economical existing units and replacing them with subsidized sources of electricity.

What is the significance of these changes to electricity generation, which have been accomplished in large part through regulations and subsidies, rather than driven by market forces? The unseen results of these policies are the additional costs that are incurred by retiring generation sources that are still functional. If the Clean Power Plan does indeed close 50 GWs of existing coal-fired generation resources, what does that cost in terms of lost low-cost electricity?


What is Not Seen: Analysis of LCOE for Existing Sources Presents Fuller Context

Simply examining forward-looking estimates like those of the EIA does not provide the full picture. If one merely looks at the benefits of new power plants and doesn’t assess the costs of destroying existing things of value, the analysis will be skewed and incomplete for policymaking purposes.

The late French economist Frederic Bastiat examined the concept of unintended consequences in his essay, What Is Seen and What Is Not Seen, where he discussed the less apparent ramifications of destroying things of value in order to spur economic activity in another area. His best-known example is that of a child breaking a shop window, causing the people of the town to take comfort in the increased business for the glazier to repair the window. Ultimately, Bastiat disagrees with that common sentiment and argues that breaking windows to encourage industry in another sector is harmful, since “society loses the value of things which are uselessly destroyed.”[14]

Similarly, policymakers must understand the full context of electricity generation and analyze the cost of breaking windows in the electricity generation sector—i.e. retiring existing generation resources before the end of their economic lives. In order to perform that analysis, policymakers need an apples-to-apples metric—the LCOE for existing sources. The IER study provides this much-needed cost information and illustrates the competitiveness of current resources with the politically favored, emerging ones.[15] In addition, although EIA’s figures are useful for comparing the costs of new electricity generation, the estimates do not “reflect the real-world operating characteristics of different generation resources on the power grid,” since the data refers to each source’s highest level of operation rather than historical levels.[16] Therefore, the IER study also adjusts EIA’s LCOE data using real-world capacity factors for each technology.

Due to the different roles that generating sources play in electricity production, it can be challenging to compare power plants that serve different purposes. Not only do varying capacity factors among resources make analysis difficult, but equating dispatchable and non-dispatchable technologies also poses problems. By listing traditional fuel sources like coal, natural gas, and nuclear separately from non-dispatchable ones in its levelized cost analysis of new sources,[17] EIA delineates between the two categories but doesn’t fully weigh the entailing tradeoffs.

One way that the study approaches such a comparison is to incorporate the costs that non-dispatchable sources impose on the rest of the power grid. Often, dispatchable generators are required to produce less power to accommodate generation from intermittent sources. Thus, as more and more renewable technology is added to the power grid, dispatchable power plants are forced to back down more frequently. We refer to the increase in LCOE for dispatchable sources due to lower output as the “imposed costs” of intermittent generation. While some fuel savings might be possible, the imposed costs result as dispatchable generators “run fewer hours without substantially reducing their fixed costs.” The estimated "imposed costs" of wind power are between $15.87 and $29.94 per megawatt-hour.[18] 

After including imposed costs, the IER report estimates that new wind resources come in at $112.80 per megawatt hour—a heavy premium over the LCOE of existing coal resources at $38.40 per megawatt-hour.[19] Factoring these costs into a broader assessment of costs and benefits is important when considering the context of fusing variable sources onto the grid and requiring current technologies to accommodate them.

Although comparing the LCOE for existing sources to new wind sources reveals a substantial differential, replacing existing capacity with new natural gas plants is also expensive. As more coal power is phased out in response to environmental regulations, natural gas capacity is expanding.[20] Even when these reliable and flexible resources are integrated onto the power grid, consumers pay cost premiums. Combined-cycle natural gas additions that substitute for shuttered coal capacity effectively add a 91 percent energy cost premium for consumers.[21]

The lesson here is clear: prematurely closing productive generating sources in order to integrate new ones, regardless of their reliability, increases costs for consumers. Even a new plant of the same source type would add on costs, such as construction debt and equity obligations, that aren’t at their lowest for most plants until 30 years of operation.[22] In the context of the broader debate about electricity generation, it is important not only to take into account the relative costs of different generation technologies but also to weigh the relative costs of new and existing sources.



As policymakers continue to make decisions about the energy future of the United States, energy economists strive to provide them with the proper tools to predict long-term trends and results. And as government policies motivate the closure of coal-fired power plants and catalyze their replacement with new sources of natural gas and wind, tools that assess the lasting effects of these changes are valuable inclusions to the discussion.

While EIA’s estimates of new levelized costs for new generation resources are an important piece of the puzzle, calculating the LCOE of existing generating units allows for more effective comparison because it provides a broader contextual perspective. Consequently, IER’s study offers valuable information for policymakers.

Furthermore, examining the levelized costs of new and existing resources reveals deeper truths about value creation and tradeoffs. While an intentionally broken window might spur economic activity in one sector (the window repair business), it does so at the expense of other worthwhile endeavors and ultimately destroys a thing of value. In the same vein, closing coal-fired power plants and propping up new natural gas, wind, and solar facilities might appear to create positive economic activity. It is only after we inquire about the relative value of existing power plants that we understand the cost of these “broken windows.”



[1] Travis Fisher, "History of Electricity," The Story of Electricity, August 29, 2014, http://instituteforenergyresearch.org/history-electricity/

[2] EIA, “Levelized Cost and Levelized Avoided Cost of New Generation Resources in the Annual Energy Outlook 2015,” June 2015, p. 1, http://www.eia.gov/forecasts/aeo/pdf/electricity_generation.pdf

[3] EIA, “Levelized Cost,” p. 1. Inserted word to correct spelling.

[4] Thomas F. Stacy and George S. Taylor, Ph.D., The Levelized Cost of Electricity from Existing Generation Resources, Institute for Energy Research, June 2015, http://instituteforenergyresearch.org/wp-content/uploads/2015/06/ier_lcoe_2015.pdf.

[5] “Electricity Generation,” Annual Energy Outlook 2015, April 2015, p. 24, http://www.eia.gov/forecasts/aeo/pdf/0383(2015).pdf.

[6] EIA, “Electricity Generation,” AEO 2015, p. 24.

[7] EIA, “Electricity from natural gas surpasses coal for first time, but just for one month,” July 31, 2015, http://www.eia.gov/todayinenergy/detail.cfm?id=22312.

[8] Stacy and Taylor, The Levelized Cost of Electricity from Existing Generation Resources, p. 39.

[9] GAO, EPA Regulations and Electricity: Update on Agencies' Monitoring Efforts and Coal-Fueled Generating Unit Retirements, GAO-14-672, August 2014, http://www.gao.gov/assets/670/665325.pdf

[10] EIA, “Proposed Clean Power Plan would accelerate renewable additions and coal plant retirements,” Today In Energy, June 5, 2015, http://www.eia.gov/todayinenergy/detail.cfm?id=21532.

[11] EPA, "Overview of the Clean Power Plan," August 3, 2015, p.4, http://www.epa.gov/airquality/cpp/fs-cpp-overview.pdf.

[12] Stacy and Taylor, p. 5.

[13] Stacy and Taylor, p. 5.

[14] Frederic Bastiat, “The Broken Window,” What Is Seen and What Is Not Seen, Chapter I, 1850, http://www.econlib.org/library/Bastiat/basEss1.html.

[15] Stacy and Taylor, p. 1.

[16] Stacy and Taylor, p. 1.

[17] EIA, “Levelized Cost,” p. 6.

[18] Stacy and Taylor, p. 26.

[19] Comparing the LCOE-New (adjusted to real-world capacity factors) for wind and the LCOE-Existing for conventional coal. See Stacy and Taylor, p. 4.

[20] EIA, “Scheduled 2015 capacity additions mostly wind and natural gas; retirements mostly coal,” March 10, 2015, http://www.eia.gov/todayinenergy/detail.cfm?id=20292.

[21] Comparing the LCOE-New (adjusted to real-world capacity factors) for CC gas and the LCOE-Eexisting for conventional coal. See Stacy and Taylor, p. 26.

[22] Stacy and Taylor, pp. 21–23.