davis_prospects for nuclear power_jep.26.1

7/31/2019 Davis_prospects for Nuclear Power_jep.26.1

1/18

Journal of Economic PerspectivesVolume 26, Number 1Winter 2012Pages 4966

II n September 2007, the U.S. Nuclear Regulatory Commission (NRC) receivedn September 2007, the U.S. Nuclear Regulatory Commission (NRC) receiveda license application or a proposed nuclear power reactor to be built ina license application or a proposed nuclear power reactor to be built insouthern Texas. The application marked the frst new license applicationsouthern Texas. The application marked the frst new license applicationin almost three decades. During the ollowing year, the NRC received 16 licensein almost three decades. During the ollowing year, the NRC received 16 licenseapplications or a total o 24 proposed reactors. The time was right, so it seemed,applications or a total o 24 proposed reactors. The time was right, so it seemed,or a nuclear power renaissance in the United States. Natural gas prices were at theiror a nuclear power renaissance in the United States. Natural gas prices were at theirhighest level ever in real terms. The 2005 Energy Policy Act provided loan guar-highest level ever in real terms. The 2005 Energy Policy Act provided loan guar-antees, production tax credits, and other subsidies or new nuclear plants. Manyantees, production tax credits, and other subsidies or new nuclear plants. Manybelieved that the United States was close to enacting legislation that would limitbelieved that the United States was close to enacting legislation that would limitemissions o carbon dioxide.emissions o carbon dioxide.Then everything changed. Natural gas prices ell sharply in 2009. Legislation toThen everything changed. Natural gas prices ell sharply in 2009. Legislation tolimit carbon emissions stalled in Congress. The global recession slowed the growthlimit carbon emissions stalled in Congress. The global recession slowed the growtho electricity demand. And in March 2011, an earthquake and tsunami knocked outo electricity demand. And in March 2011, an earthquake and tsunami knocked outpower at the Fukushima Daiichi Nuclear Plant in northern Japan, causing partialpower at the Fukushima Daiichi Nuclear Plant in northern Japan, causing partialmeltdowns at the plants three active reactors and large-scale releases o radioactivemeltdowns at the plants three active reactors and large-scale releases o radioactivesteam. Since 2009, only a single additional license application has been fled withsteam. Since 2009, only a single additional license application has been fled withthe NRC. The project proposed or southern Texas has been canceled, and ew othe NRC. The project proposed or southern Texas has been canceled, and ew othe applications pending with the NRC are moving orward. Fukushima has hadthe applications pending with the NRC are moving orward. Fukushima has hadperhaps an even stronger impact worldwide, leading Germany, Switzerland, andperhaps an even stronger impact worldwide, leading Germany, Switzerland, andItaly to announce plans to phase out their nuclear power programs and causingItaly to announce plans to phase out their nuclear power programs and causingChina to suspend approvals or new reactors.China to suspend approvals or new reactors.Nuclear power has long been controversial because o concerns about nuclearNuclear power has long been controversial because o concerns about nuclearaccidents, storage o spent uel, and about how the spread o nuclear power mightaccidents, storage o spent uel, and about how the spread o nuclear power might

Prospects or Nuclear Power

Lucas W. Davis is Assistant Professor at the Haas School of Business, University ofLucas W. Davis is Assistant Professor at the Haas School of Business, University ofCalifornia, Berkeley, California. His e-mail address isCalifornia, Berkeley, California. His e-mail address [email protected]@haas.berkeley.edu..doi=10.1257/jep.26.1.49

Lucas W. Davis


2/18

50 Journal of Economic Perspectives

raise risks o the prolieration o nuclear weapons. These concerns are real andraise risks o the prolieration o nuclear weapons. These concerns are real andiimportant. However, emphasizing these concerns implicitly suggests that unlessmportant. However, emphasizing these concerns implicitly suggests that unlessthese issues are taken into account, nuclear power would otherwise be cost eectivethese issues are taken into account, nuclear power would otherwise be cost eectivecompared to other orms o electricity generation. This implication is unwarranted.compared to other orms o electricity generation. This implication is unwarranted.Throughout the history o nuclear power, a key challenge has been the high cost oThroughout the history o nuclear power, a key challenge has been the high cost oconstruction or nuclear plants. Construction costs are high enough that it becomesconstruction or nuclear plants. Construction costs are high enough that it becomesdifcult to make an economic argument or nuclear even beore incorporatingdifcult to make an economic argument or nuclear even beore incorporatingthese external actors. This is particularly true in countries like the United Statesthese external actors. This is particularly true in countries like the United Stateswhere recent technological advances have dramatically increased the availability owhere recent technological advances have dramatically increased the availability onatural gas.natural gas.The chairman o one o the largest U.S. nuclear companies recently said thatThe chairman o one o the largest U.S. nuclear companies recently said thathis company would not break ground on a new nuclear plant until the price ohis company would not break ground on a new nuclear plant until the price onatural gas was more than double todays level and carbon emissions cost $25 pernatural gas was more than double todays level and carbon emissions cost $25 perton (Wald 2010). This comment summarizes the current economics o nuclearton (Wald 2010). This comment summarizes the current economics o nuclearpower pretty well. Yes, there is a certain conuence o actors that could makepower pretty well. Yes, there is a certain conuence o actors that could makenuclear power a viable economic option. Otherwise, a nuclear power renaissancenuclear power a viable economic option. Otherwise, a nuclear power renaissanceseems unlikely.seems unlikely.

The First Boom and BustThe First Boom and BustThis recent ebb and ow in the nuclear power sector recalls a much largerThis recent ebb and ow in the nuclear power sector recalls a much larger

boom and bust that occurred starting in the 1960s and 1970s. Figure 1 plots U.S.boom and bust that occurred starting in the 1960s and 1970s. Figure 1 plots U.S.nuclear power reactor orders rom 1950 to 2000. By 1974, there were 54 operatingnuclear power reactor orders rom 1950 to 2000. By 1974, there were 54 operatingreactors in the United States with another 197 on order. This period was one oreactors in the United States with another 197 on order. This period was one ogreat enthusiasm or nuclear power. U.S. coal prices were at their highest level evergreat enthusiasm or nuclear power. U.S. coal prices were at their highest level everin real terms, and utilities were orecasting robust growth in electricity demand intoin real terms, and utilities were orecasting robust growth in electricity demand intothe distant uture.the distant uture.11The U.S. Atomic Energy Commission (1974) predicted that byThe U.S. Atomic Energy Commission (1974) predicted that bythe end o the twentieth century hal o all U.S. electricity generation would comethe end o the twentieth century hal o all U.S. electricity generation would comerom nuclear power.rom nuclear power.Instead, reactor orders ell precipitously ater 1974. Over the next severalInstead, reactor orders ell precipitously ater 1974. Over the next severalyears not only were new reactors not being ordered, but utilities began suspendingyears not only were new reactors not being ordered, but utilities began suspendingconstruction on existing orders. Less than hal o the reactors on order in 1974construction on existing orders. Less than hal o the reactors on order in 1974were ever completed. Much has been written about the problems that aced thewere ever completed. Much has been written about the problems that aced thenuclear industry during this period (or example, Joskow and Yellin 1980; Joskownuclear industry during this period (or example, Joskow and Yellin 1980; Joskow1982; McCallion 1995). Part o the explanation is that concerns about saety1982; McCallion 1995). Part o the explanation is that concerns about saetyand the environment began to take a more central role. In 1974, the Nuclearand the environment began to take a more central role. In 1974, the NuclearRegulatory Commission was created to replace the Atomic Energy Commission,Regulatory Commission was created to replace the Atomic Energy Commission,which had previously been charged with both regulating and promoting nuclearwhich had previously been charged with both regulating and promoting nuclear1 For historic coal prices, see U.S. Department o Energy (2011a, table 7.9 Coal Prices, 19492010).

Natural gas was much less important during the 1970s because modern combined cycle technology hadnot yet been widely introduced and because shortages associated with ederal price controls on naturalgas limited the availability o natural gas or electric generation.


3/18

Lucas W. Davis 51

powera combination o duties which many viewed to be in direct conict. Thispowera combination o duties which many viewed to be in direct conict. Thisnew organization was to oversee the saety and security o all aspects o nuclearnew organization was to oversee the saety and security o all aspects o nuclearpower, including the initial licensing o reactors, the handling o radioactive mate-power, including the initial licensing o reactors, the handling o radioactive mate-rials, and the storage and disposal o spent uels. Beginning in the 1970s, it alsorials, and the storage and disposal o spent uels. Beginning in the 1970s, it alsobecame more difcult to site nuclear power plants. Communities began challengingbecame more difcult to site nuclear power plants. Communities began challengingnuclear power projects in ederal and state courts, leading to extended constructionnuclear power projects in ederal and state courts, leading to extended constructiondelays and changing public attitudes about nuclear power.delays and changing public attitudes about nuclear power.Utility regulation also experienced structural change at this time. During theUtility regulation also experienced structural change at this time. During the1950s and 1960s, economies o scale, decreasing commodity costs, and relatively1950s and 1960s, economies o scale, decreasing commodity costs, and relativelylow ination led to steady decreases in the nominal cost o electricity. Public utilitylow ination led to steady decreases in the nominal cost o electricity. Public utilitycommissions and consumers were pacifed with prices that remained essentiallycommissions and consumers were pacifed with prices that remained essentiallythe same in nominal terms year ater year. Joskow (1974) explains that ination inthe same in nominal terms year ater year. Joskow (1974) explains that ination inthe early 1970s, wreaked havoc on this process that appeared to unction sothe early 1970s, wreaked havoc on this process that appeared to unction sosmoothly beore . . . and most major frms ound that they had to raise prices (somesmoothly beore . . . and most major frms ound that they had to raise prices (someor the frst time in 25 years) and trigger ormal regulatory reviews. These reviewsor the frst time in 25 years) and trigger ormal regulatory reviews. These reviewsled to increased scrutiny o utilities capital expenditures, and in particular, invest-led to increased scrutiny o utilities capital expenditures, and in particular, invest-ments in nuclear plants.ments in nuclear plants.Then in March 1979, one o the reactors at the Three Mile Island plant inThen in March 1979, one o the reactors at the Three Mile Island plant inPennsylvania suered a partial core meltdown. Although not a single personPennsylvania suered a partial core meltdown. Although not a single personwas injured, the accident intensifed public concerns about nuclear saety. Thewas injured, the accident intensifed public concerns about nuclear saety. Thecombination o severe public concern about the risk o nuclear accidents andcombination o severe public concern about the risk o nuclear accidents and

Figure 1

U.S. Nuclear Power Reactor Orders

Source:Author based on data rom U.S. Department o Energy (1997).

0

10

20

30

40

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000

Three Mile Island

Chernobyl


4/18


escalating construction costs put nuclear projects in an extremely vulnerable posi-escalating construction costs put nuclear projects in an extremely vulnerable posi-tion. By the time the Chernobyl disaster occurred in April 1986, expansion o U.S.tion. By the time the Chernobyl disaster occurred in April 1986, expansion o U.S.nuclear power had largely halted. Today in the United States, there are a total onuclear power had largely halted. Today in the United States, there are a total o104 nuclear power reactors at 65 sites, accounting or 20 percent o U.S. electricity104 nuclear power reactors at 65 sites, accounting or 20 percent o U.S. electricitygenerationgeneration(U.S. Department o Energy 2011a, table 8.2a).(U.S. Department o Energy 2011a, table 8.2a).22All o these reactorsAll o these reactorswere ordered prior to 1974.were ordered prior to 1974.Nuclear reactor construction outside the United States ollowed a similarNuclear reactor construction outside the United States ollowed a similarpattern with a substantial boom in the 1960s and 1970s, ollowed by a long periodpattern with a substantial boom in the 1960s and 1970s, ollowed by a long periodo decline, shown in Figure 2. In addition to the United States, the other large-o decline, shown in Figure 2. In addition to the United States, the other large-scale early adopters o nuclear power were the United Kingdom, France, Germany,scale early adopters o nuclear power were the United Kingdom, France, Germany,Canada, and Japan. By the 1990s, construction had moved to Eastern Europe and inCanada, and Japan. By the 1990s, construction had moved to Eastern Europe and inparticular Russia and the Ukraine. The increase in construction 20082010 comesparticular Russia and the Ukraine. The increase in construction 20082010 comesprimarily rom China, which today has more reactors under construction than anyprimarily rom China, which today has more reactors under construction than anyother country.other country.

2 U.S. Department o Energy (2011a, table 8.2a) also reports that U.S. net generation o electricity in

2010 includes coal (45 percent), natural gas (24 percent), nuclear (20 percent), hydroelectric power(7 percent), and wind and other renewables (4 percent). In Davis and Wolram (2011), my coauthor andI examine in detail operating perormance at U.S. nuclear plants.

Figure 2

Nuclear Reactors under Construction Worldwide

Source:Author based on data rom International Atomic Energy Agency (2011).

0

50

100

150

200 Other

Eastern Europe

Western Europe

United States and Canada

1950 1960 1970 1980 1990 20102000


5/18

Prospects for Nuclear Power 53

Historical Construction CostsHistorical Construction CostsNuclear power plants are characterized by high construction costs and rela-Nuclear power plants are characterized by high construction costs and rela-tively low operating costs. Later in the paper, I present estimates o levelized costs,tively low operating costs. Later in the paper, I present estimates o levelized costs,which acilitate comparisons with other generating technologies with dierent costwhich acilitate comparisons with other generating technologies with dierent costprofles over time. It is worth starting with construction costs, however, because theyprofles over time. It is worth starting with construction costs, however, because theyrepresent a large share o the total cost o nuclear power.represent a large share o the total cost o nuclear power.Nuclear power plants are enormous acilities with high construction costs. TheNuclear power plants are enormous acilities with high construction costs. Thesheer scale o commercial-sized nuclear reactors means that most components mustsheer scale o commercial-sized nuclear reactors means that most components mustbe specially designed and constructed, oten with ew potential suppliers worldwide.be specially designed and constructed, oten with ew potential suppliers worldwide.These components are then assembled on site, and structures are constructed toThese components are then assembled on site, and structures are constructed tohouse the assembled components. All stages o design, construction, assembly, andhouse the assembled components. All stages o design, construction, assembly, andtesting require highly-skilled, highly-specialized engineers. Dierences in reactortesting require highly-skilled, highly-specialized engineers. Dierences in reactordesign and site-specifc actors have historically meant that there was little scope ordesign and site-specifc actors have historically meant that there was little scope orspreading design and production costs across multiple projects.spreading design and production costs across multiple projects.Figure 3 plots overnight construction costs or selected U.S. nuclear reactorsFigure 3 plots overnight construction costs or selected U.S. nuclear reactors

rom the U.S. Department o Energy (1986). The overnight cost is the hypotheticalrom the U.S. Department o Energy (1986). The overnight cost is the hypotheticalcost o a plant i it could be built instantly and thus excludes fnancing and othercost o a plant i it could be built instantly and thus excludes fnancing and othercosts incurred during plant construction. Costs are reported in year 2010 dollarscosts incurred during plant construction. Costs are reported in year 2010 dollars

Figure 3

Construction Costs or U.S. Nuclear Reactors by Year o Completion

Source:U.S. DOE (1986), table 4.Notes:Figure 3 plots overnight construction costs or selected U.S. nuclear power plants rom the U.S.

Department o Energy (1986). The fgure includespredictedcosts rom the same source or a handul oreactors that were under construction but not yet in operation in 1986.

1970 1975 1980 1985 1990

0

Costperkilowattofcapacity(in2010dollars)

2,000

4,000

6,000

8,000

Browns Ferry 3

North Anna 1

LaSalle 2

Byron 2

Diablo Canyon 2

Braidwood 1

Waterord 3Perry 1San Onore 2

Millstone 3

Limerick 1

Clinton 1

Shoreham

Nine Mile Point 2


6/18


per kilowatt o capacity. The fgure reveals a pronounced increase in constructionper kilowatt o capacity. The fgure reveals a pronounced increase in constructioncosts, particularly or plants completed during the 1980s.costs, particularly or plants completed during the 1980s.33Plants also kept taking longer and longer to build. As shown in Table 1, reactorsPlants also kept taking longer and longer to build. As shown in Table 1, reactorsordered during the 1950s took on average about fve years to build, whereas reac-ordered during the 1950s took on average about fve years to build, whereas reac-tors ordered during the 1970s took on average 14 years. Most studies attribute thistors ordered during the 1970s took on average 14 years. Most studies attribute thisincrease in construction time to a rapidly evolving regulatory process. A joke in theincrease in construction time to a rapidly evolving regulatory process. A joke in theindustry was that a reactor vessel could not be shipped until the total weight o allindustry was that a reactor vessel could not be shipped until the total weight o allrequired paperwork had equaled the weight o the reactor vessel itsel. Regulationrequired paperwork had equaled the weight o the reactor vessel itsel. Regulationalso contributed directly to construction costs. The Nuclear Regulatory Commissionalso contributed directly to construction costs. The Nuclear Regulatory Commissionimplemented revised saety codes and inspection requirements leading in severalimplemented revised saety codes and inspection requirements leading in severalcases to extensive reactor redesigns (Cox and Gilbert 1991; McCallion 1995).cases to extensive reactor redesigns (Cox and Gilbert 1991; McCallion 1995).An interesting point o comparison is France. Ater the United States, FranceAn interesting point o comparison is France. Ater the United States, Francehas more nuclear reactors than any other country, and 75 percent o electricityhas more nuclear reactors than any other country, and 75 percent o electricitygeneration in France comes rom nuclear power. Grubler (2010) fnds that 58 reac-generation in France comes rom nuclear power. Grubler (2010) fnds that 58 reac-tors in Frances main nuclear program were constructed at an average cost thattors in Frances main nuclear program were constructed at an average cost thatincreased over time rom $1,000 per kilowatt o capacity in the 1970s to $2,300 inincreased over time rom $1,000 per kilowatt o capacity in the 1970s to $2,300 inthe 1990s. The cost escalation is less severe than is observed in the United States,the 1990s. The cost escalation is less severe than is observed in the United States,but still somewhat surprising. As I discuss later, in many ways the French nuclearbut still somewhat surprising. As I discuss later, in many ways the French nuclearprogram was the ideal setting or encouraging learning-by-doing, so one might haveprogram was the ideal setting or encouraging learning-by-doing, so one might haveexpected costs to decrease over time.expected costs to decrease over time.

Financing RisksFinancing RisksThe long period o time required or construction means that the cost o capitalThe long period o time required or construction means that the cost o capitalis a critical parameter or evaluating the viability o nuclear power. Even or a lowis a critical parameter or evaluating the viability o nuclear power. Even or a low

3 See Koomey and Hultman (2007) or a more recent study o U.S. nuclear construction costs, and Mooz(1978), Komano (1981), and Zimmerman (1982) or studies o the earlier period. Joskow and Rose(1985) examine increases in construction costs or coal plants during the 1960s and 1970s fnding signif-

cant cost increases associated with measurable environmental-related technologies such as scrubbers andcooling towers, as well as a large increase in residual real costs that they attribute to changes in environ-mental regulation and to an unexplained decline in construction productivity.

Table 1

U.S. Nuclear Reactor Orders and Construction Time

Number ofreactors

ordered

Percenteventually

completed

Construction time (in years) for completed reactors

Decade Average Minimum Maximum

1950s 6 100% 4.5 3 71960s 88 89% 8.6 3 221970s 155 25% 14.1 8 26

Source:Authors tabulations based on U.S. Department o Energy (1997). Construction time is calculatedas the dierence in years between when a reactor is ordered and when it begins commercial operation.


7/18

Lucas W. Davis 55

cost o capital, the extended construction period imposes fnancing costs that are acost o capital, the extended construction period imposes fnancing costs that are asubstantial part o total project costs. However, nuclear projects typically ace a costsubstantial part o total project costs. However, nuclear projects typically ace a costo capital well above the risk-ree rate. These large-scale projects have a historicallyo capital well above the risk-ree rate. These large-scale projects have a historicallyhigh risk o deault. The high cost o capital that they ace reects the number ohigh risk o deault. The high cost o capital that they ace reects the number orisks that threaten the proftability, and even viability, o a nuclear project.risks that threaten the proftability, and even viability, o a nuclear project.More so than in most other investments, nuclear power plants ace substantialMore so than in most other investments, nuclear power plants ace substantialregulatory risk. The Nuclear Regulatory Commission has recently adopted severalregulatory risk. The Nuclear Regulatory Commission has recently adopted severalnew procedures intended to streamline the regulatory process. These reormsnew procedures intended to streamline the regulatory process. These reormsinclude pre-approving standard reactor designs, an early site permitting process,include pre-approving standard reactor designs, an early site permitting process,and combining construction and operat ng licenses which previously were appliedand combining construction and operat ing licenses which previously were appliedor separately. It remains to be seen how these procedures will work in practice.or separately. It remains to be seen how these procedures will work in practice.Regulatory approval is also required at the state and local level, and it can be aRegulatory approval is also required at the state and local level, and it can be areal constraint on plants. For example, in 1989 New York Governor Mario Cuomoreal constraint on plants. For example, in 1989 New York Gov ernor Mario Cuomoand the Long Island Lighting Company closed the Shoreham Nuclear Power Plantand the Long Island Lighting Company closed the Shoreham Nuclear Power Plantover long-standing concerns about how nearby residents would be evacuated inover long-standing concerns about how nearby residents would be evacuated inthe event o an emergency. The plant was 100 percent completed and had beenthe event o an emergency. The plant was 100 percent completed and had beenconnected to the grid, yet was never used to produce a single kilowatt hour oconnected to the grid, yet was never used to produce a single kilowatt hour ocommercial electricity.commercial electricity.Nuclear power is also sensitive to ederal energy policy. The enthusiasm orNuclear power is also sensitive to ederal energy policy. The enthusiasm ornuclear power in 2007 and 2008 was driven in part by the prospect o a ederalnuclear power in 2007 and 2008 was driven in part by the prospect o a ederalcap on carbon emissions, and so when the key legislative vehicle (H.R. 2454, thecap on carbon emissions, and so when the key legislative vehicle (H.R. 2454, theWaxmanMarkey bill) stalled in the U.S. Senate in 2009, it was a signifcant blowWaxmanMarkey bill) stalled in the U.S. Senate in 2009, it was a signifcant blowto the economic viability o new nuclear plants. In the last ew years, the Obamato the economic viability o new nuclear plants. In the last ew years, the Obamaadministration and some members o Congress have voiced support or a ederaladministration and some members o Congress have voiced support or a ederalclean energy standard under which a proportion o total electricity generationclean energy standard under which a proportion o total electricity generationwould be required to come rom sources that do not generate carbon emissions.would be required to come rom sources that do not generate carbon emissions.Such a policy could be a considerable boost or nuclear power, but the exact ormSuch a policy could be a considerable boost or nuclear power, but the exact ormo such legislation, or how likely its adoption would be, is unclear.o such legislation, or how likely its adoption would be, is unclear.Investors in nuclear power also ace the risk that ossil uel prices could decrease.Investors in nuclear power also ace the risk that ossil uel prices could decrease.In the United States, natural gas prices typically determine the marginal cost oIn the United States, natural gas prices typically determine the marginal cost oelectricity, so a decrease in natural gas prices reduces profts or nuclear plantselectricity, so a decrease in natural gas prices reduces profts or nuclear plantsthat sell power in wholesale electricity markets. Global availability o natural gasthat sell power in wholesale electricity markets. Global availability o natural gashas increased dramatically in recent years with improvements in horizontal drillinghas increased dramatically in recent years with improvements in horizontal drillingand hydraulic racturing technology. Natural gas producers have long known thatand hydraulic racturing technology. Natural gas producers have long known thatshale and other rock deposits contain large amounts o natural gas. It was not untilshale and other rock deposits contain large amounts o natural gas. It was not untilrecently, however, that these resources could be accessed at reasonably low cost.recently, however, that these resources could be accessed at reasonably low cost.Figure 4 plots U.S. natural gas prices rom 1990 to 2011 and a price orecastFigure 4 plots U.S. natural gas prices rom 1990 to 2011 and a price orecastthrough 2030. During the long period o relatively low natural gas prices, therethrough 2030. During the long period o relatively low natural gas prices, therewas not a single new nuclear plant ordered in the United States, and the surgewas not a single new nuclear plant ordered in the United States, and the surgein orders in 2007 and 2008 came at the same time that U.S. natural gas pricesin orders in 2007 and 2008 came at the same time that U.S. natural gas pricesreached their highest level ever in real terms. The baseline orecast rom U.S.reached their highest level ever in real terms. The baseline orecast rom U.S.Department o Energy (2011b) predicts that U.S. natural gas prices will remainDepartment o Energy (2011b) predicts that U.S. natural gas prices will remainunder $5 per thousand cubic eet through 2022. I true, this is a signifcant chal-under $5 per thousand cubic eet through 2022. I true, this is a signifcant chal-lenge or nuclear power.lenge or nuclear power.


8/18


Finally, investments in nuclear power ace considerable technology risk. OverFinally, investments in nuclear power ace considerable technology risk. Overthe 40-plus year lietime o a nuclear plant, the available sources o electricity gener-the 40-plus year lietime o a nuclear plant, the available sources o electricity gener-ation could change considerably. An alternative, lower-cost technology could comeation could change considerably. An alternative, lower-cost technology could comealong, or perhaps a technology that is known today such as wind or solar couldalong, or perhaps a technology that is known today such as wind or solar couldquickly become more cost eective. An alternative technology or carbon abate-quickly become more cost eective. An alternative technology or carbon abate-ment could become practical, like some orm o carbon capture and storage, whichment could become practical, like some orm o carbon capture and storage, whichwould render moot one o the advantages o nuclear power. New energy efciencywould render moot one o the advantages o nuclear power. New energy efciencytechnologies might reduce electricity demand.technologies might reduce electricity demand.

Recent International ExperienceRecent International ExperienceMore recent evidence on construction costs comes rom nuclear reactorsMore recent evidence on construction costs comes rom nuclear reactorscurrently being built in Olkiluoto, Finland, and Flamanville, France. Much hascurrently being built in Olkiluoto, Finland, and Flamanville, France. Much hasbeen written about these reactors because they are the frst new reactors to be builtbeen written about these reactors because they are the frst new reactors to be built

in Europe in many years, and because they use a next generation design thatin Europe in many years, and because they use a next generation design thatincorporates several new saety eatures into a reactor design that is widely usedincorporates several new saety eatures into a reactor design that is widely usedaround the world. Construction in Finland began in 2005 and was expected toaround the world. Construction in Finland began in 2005 and was expected to

Figure 4

U.S. Natural Gas Prices and Applications to the U.S. Nuclear Regulatory

Commission (NRC)

Sources:Author based on data rom U.S. Department o Energy (2011a, b) and U.S. Nuclear RegulatoryCommission (2011).

Vertical lines indicate applications to NRC

Baseline price orecast, DOE (2011b)

1990 2000 2010 2020 2030

Priceperthousandcubic

feet(inU.S.2010dollars)

0

2

4

6

8

10

12


9/18


be completed in 2009 at cost o about $2,800 per kilowatt o capacity. A series obe completed in 2009 at cost o about $2,800 per kilowatt o capacity. A series oproblems and delays have now pushed operations back to 2013, and costs are nowproblems and delays have now pushed operations back to 2013, and costs are nowestimated to be about twice the original estimate. Similarly, construction in Franceestimated to be about twice the original estimate. Similarly, construction in Francebegan in 2007 and the reactor was expected to be completed by 2011 at a cost obegan in 2007 and the reactor was expected to be completed by 2011 at a cost o$2,900 per kilowatt. Completion has now been pushed back to 2014 and the project$2,900 per kilowatt. Completion has now been pushed back to 2014 and the projectis reported to be 50 percent over budget.is reported to be 50 percent over budget.These experiences provide a reminder about problems that can occur duringThese experiences provide a reminder about problems that can occur duringreactor construction, particularly given the lack o recent construction experi-reactor construction, particularly given the lack o recent construction experi-ence. Both projects were delayed substantially when government saety inspectorsence. Both projects were delayed substantially when government saety inspectorsound problems. In Finland, the concrete oundation o the reactor building wasound problems. In Finland, the concrete oundation o the reactor building wasound to be too porous. In France, inspectors ound cracks in the concrete oun-ound to be too porous. In France, inspectors ound cracks in the concrete oun-dation and steel reinorcements in the wrong places. Project managers have beendation and steel reinorcements in the wrong places. Project managers have beenblamed in both projects or hiring inexperienced contractors and or providingblamed in both projects or hiring inexperienced contractors and or providinginsufcient oversight.insufcient oversight.Construction costs have tended to be lower elsewhere. Du and Parsons (2009)Construction costs have tended to be lower elsewhere. Du and Parsons (2009)report a mean overnight cost o $3,100 per kilowatt rom fve reactors completed inreport a mean overnight cost o $3,100 per kilowatt rom fve reactors completed inKorea and Japan between 2004 and 2006. Construction costs rom plants recentlyKorea and Japan between 2004 and 2006. Construction costs rom plants recentlycompleted in China are reported to be even lower and an important area or uturecompleted in China are reported to be even lower and an important area or utureresearch is to examine these costs in detail.research is to examine these costs in detail.Several studies have attempted to synthesize this recent international construc-Several studies have attempted to synthesize this recent international construc-tion experience with historical U.S. data and engineering studies to estimate currenttion experience with historical U.S. data and engineering studies to estimate currentconstruction costs or the United States. Table 2 reports estimates o overnightconstruction costs or the United States. Table 2 reports estimates o overnightconstruction cost rom two such studies. MIT (2009) estimates $4,200 per kilowattconstruction cost rom two such studies. MIT (2009) estimates $4,200 per kilowatto capacity or nuclear, compared to $2,400 and $900 per kilowatt o capacity oro capacity or nuclear, compared to $2,400 and $900 per kilowatt o capacity orcoal and natural gas. U.S. Department o Energy (2010) predicts somewhat highercoal and natural gas. U.S. Department o Energy (2010) predicts somewhat highercosts particularly or nuclear plants, citing increased prices or plant componentscosts particularly or nuclear plants, citing increased prices or plant componentsand key commodities and arguing that costs will be driven up by the act that only aand key commodities and arguing that costs will be driven up by the act that only alimited set o construction frms have the ability to complete a project o this scale.limited set o construction frms have the ability to complete a project o this scale.Both studies were completed prior to Fukushima and thus do not incorporate anyBoth studies were completed prior to Fukushima and thus do not incorporate anycost increases due to recently elevated regulatory scrutiny.cost increases due to recently elevated regulatory scrutiny.

Adding fnancing costs to these estimates implies that a typical two-reactorAdding fnancing costs to these estimates implies that a typical two-reactor2,000 megawatt plant could cost more than $12 billion. The long period o time2,000 megawatt plant could cost more than $12 billion. The long period o timesince nuclear power plants were constructed in the United States means that thesince nuclear power plants were constructed in the United States means that the

Table 2

Power Plant Construction Costs Excluding Financing or Nuclear, Coal, and

Natural Gas

Cost per kilowatt of capacity (in year 2010 dollars)

Source Nuclear Coal Natural gas

MIT (2009) $4,200 $2,400 $ 900U.S. Department o Energy (2010) $5,300 $2,800 $1,000


10/18


relevant experience that had been accumulated by companies involved with nuclearrelevant experience that had been accumulated by companies involved with nuclearengineering and plant construction has atrophied substantially (Joskow and Parsonsengineering and plant construction has atrophied substantially (Joskow and Parsons2009). There is some scope or importing nuclear engineers and other proessionals2009). There is some scope or importing nuclear engineers and other proessionalswho have worked on more recent nuclear projects in other countries, but the overallwho have worked on more recent nuclear projects in other countries, but the overalllevel o nuclear construction activity worldwide over the last 20 years has been solevel o nuclear construction activity worldwide over the last 20 years has been solow that the available global talent is limited. Moreover, the supply o nuclear plantlow that the available global talent is limited. Moreover, the supply o nuclear plantcomponents is now more limited than it was during the frst wave o nuclear powercomponents is now more limited than it was during the frst wave o nuclear powerplant construction. For example, there is currently only one acility in the world thatplant construction. For example, there is currently only one acility in the world thatcan produce the nuclear-grade heavy-steel reactor vessel needed or a boiling watercan produce the nuclear-grade heavy-steel reactor vessel needed or a boiling waterreactor, and there is currently a long waiting period or these orgings and or otherreactor, and there is currently a long waiting period or these orgings and or otherkey nuclear components (Ives, McCabe, and Gilmartin 2010).key nuclear components (Ives, McCabe, and Gilmartin 2010).These construction cost estimates contain considerable uncertainty, which isThese construction cost estimates contain considerable uncertainty, which isitsel a barrier to investment. Pindyck (1993) uses a model o irreversible investmentitsel a barrier to investment. Pindyck (1993) uses a model o irreversible investmentto illustrate how uncertainty over the prices o construction inputs and over govern-to illustrate how uncertainty over the prices o construction inputs and over govern-ment regulation aecting construction costs can lead investors to delay investmentment regulation aecting construction costs can lead investors to delay investmenton nuclear projects. One o the economic arguments made in support o the subsi-on nuclear projects. One o the economic arguments made in support o the subsi-dies or new nuclear plants in the 2005 Energy Policy Act was that they would helpdies or new nuclear plants in the 2005 Energy Policy Act was that they would helpresolve this uncertainty about construction costs.resolve this uncertainty about construction costs.

Levelized Cost EstimatesLevelized Cost EstimatesThe total cost o producing electricity depends both on construction costs and onThe total cost o producing electricity depends both on construction costs and onoperations and maintenance expenditures, including uel. These variable costs tendoperations and maintenance expenditures, including uel. These variable costs tendto be low or nuclear, potentially osetting the higher cost o construction. Table 3to be low or nuclear, potentially osetting the higher cost o construction. Table 3reports levelized costs or electricity generated in the United States rom nuclear,reports levelized costs or electricity generated in the United States rom nuclear,coal, and natural gas, the three primary orms o baseload electricity generation. Thesecoal, and natural gas, the three primary orms o baseload electricity generation. Theseestimates are based on a cash ow model developed in an ongoing series o studiesestimates are based on a cash ow model developed in an ongoing series o studiesat the Massachusetts Institute o Technology (MIT 2003; MIT 2009; Joskow 2006; Duat the Massachusetts Institute o Technology (MIT 2003; MIT 2009; Joskow 2006; Duand Parsons 2009; Joskow and Parsons 2009). For these estimates, all costs includingand Parsons 2009; Joskow and Parsons 2009). For these estimates, all costs includingconstruction, operation, maintenance and uel are calculated and discounted back toconstruction, operation, maintenance and uel are calculated and discounted back tothe present using an assumed cost o capital. This total cost is then levelized over thethe present using an assumed cost o capital. This total cost is then levelized over the

lietime o a plant in constant dollars to yield the long-run average cost o producinglietime o a plant in constant dollars to yield the long-run average cost o producinga kilowatt hour o electricity. This is equivalent to the real price per kilowatt hour thata kilowatt hour o electricity. This is equivalent to the real price per kilowatt hour thatthe plant would need to receive over its lietime in order to break even.the plant would need to receive over its lietime in order to break even.Under the baseline assumptions, nuclear is not competitive with either coal orUnder the baseline assumptions, nuclear is not competitive with either coal ornatural gas. The frst row o Table 3 reports the base case estimates reported in MITnatural gas. The frst row o Table 3 reports the base case estimates reported in MIT(2009). The levelized cost o nuclear power is 8.7 cents per kilowatt hour, compared(2009). The levelized cost o nuclear power is 8.7 cents per kilowatt hour, comparedto 6.5 cents or coal and 6.7 cents or natural gas. This gap widens in the secondto 6.5 cents or coal and 6.7 cents or natural gas. This gap widens in the secondrow ater updating these estimates to reect higher construction cost estimatesrow ater updating these estimates to reect higher construction cost estimatesrom U.S. Department o Energy (2010). The third row updates the estimates torom U.S. Department o Energy (2010). The third row updates the estimates toreect changes in uel prices since 2009. Uranium prices have increased modestly,reect changes in uel prices since 2009. Uranium prices have increased modestly,but uel expenditures represent a relatively small proportion o the total cost obut uel expenditures represent a relatively small proportion o the total cost onuclear power, and at this higher price, even ater including costs or conversion,nuclear power, and at this higher price, even ater including costs or conversion,


11/18

Lucas W. Davis 59

enrichment, and uel abrication, nuclear uel costs are still less than one centenrichment, and uel abrication, nuclear uel costs are still less than one centper kilowatt hour. Moreover, the medium- to long-run supply o uranium is highlyper kilowatt hour. Moreover, the medium- to long-run supply o uranium is highlyelastic, with substantial known reserves worldwide with a cost o recovery belowelastic, with substantial known reserves worldwide with a cost o recovery belowcurrent uranium prices (MIT 2003, Appendix 5.E; OECD 2009). Fossil uel pricescurrent uranium prices (MIT 2003, Appendix 5.E; OECD 2009). Fossil uel pricesare extremely important or the prospects or nuclear power, and the cost estimatesare extremely important or the prospects or nuclear power, and the cost estimatesin the third row reect somewhat higher coal prices but also considerablyin the third row reect somewhat higher coal prices but also considerably lowernatural gas prices. With these updated prices, the levelized cost o electricity romnatural gas prices. With these updated prices, the levelized cost o electricity romnatural gas is just above 5 cents per kilowatt hour, compared to more than 10 centsnatural gas is just above 5 cents per kilowatt hour, compared to more than 10 centsper kilowatt hour or nuclear.per kilowatt hour or nuclear.These estimates ollow the MIT studies in applying a somewhat higher cost oThese estimates ollow the MIT studies in applying a somewhat higher cost ocapital to nuclear power. As discussed earlier, this reects the high risk o deaultcapital to nuclear power. As discussed earlier, this reects the high risk o deaultand numerous orms o risk aced with nuclear projects. It is worth noting, however,and numerous orms o risk aced with nuclear projects. It is worth noting, however,that even without this risk premium, nuclear still has higher levelized cost thanthat even without this risk premium, nuclear still has higher levelized cost thancoal or natural gas. The model also assumes a 40-year lietime or nuclear, coal, andcoal or natural gas. The model also assumes a 40-year lietime or nuclear, coal, andnatural gas plants. Over hal o U.S. nuclear plants have received license extensionsnatural gas plants. Over hal o U.S. nuclear plants have received license extensionsto 60 years. Incorporating a longer lietime into the model makes nuclear lookto 60 years. Incorporating a longer lietime into the model makes nuclear lookbetter, but not by very much. The increased net revenue is ar in the uture so withbetter, but not by very much. The increased net revenue is ar in the uture so withdiscounting there is only a modest decrease in levelized costs. Moreover, coal anddiscounting there is only a modest decrease in levelized costs. Moreover, coal andnatural gas plants are also tending to be used or more than 40 years, and one wouldnatural gas plants are also tending to be used or more than 40 years, and one wouldwant to incorporate those longer lietimes as well.want to incorporate those longer lietimes as well.It is important to emphasize that these levelized cost estimates depend onIt is important to emphasize that these levelized cost estimates depend ona series o empirical assumptions, many o which can be only partially verifed.a series o empirical assumptions, many o which can be only partially verifed.Perhaps most importantly, alternative assumptions about nuclear construction costsPerhaps most importantly, alternative assumptions about nuclear construction costsor natural gas prices can begin to change the outlook considerably. Moreover, theseor natural gas prices can begin to change the outlook considerably. Moreover, thesecost estimates are or the United States and may not easily generalize to other coun-cost estimates are or the United States and may not easily generalize to other coun-tries. Construction costs vary substantially across countries due to dierences in thetries. Construction costs vary substantially across countries due to dierences in thecost o labor and other inputs, as well as dierences in the regulatory environment.cost o labor and other inputs, as well as dierences in the regulatory environment.

Table 3

Levelized Cost Comparison or Electricity Generation

Levelized cost in cents per kWh

Source Nuclear Coal Natural gas

MIT (2009) baseline 8.7 6.5 6.7Updated construction costs 10.4 7.0 6.9Updated construction costs and uel prices 10.5 7.4 5.2

With carbon tax o $25 per ton CO2 10.5 9.6 6.2

Source:These calculations ollow MIT (2009) except where indicated in the row headings.Notes:All costs are reported in 2010 cents per kilowatt hour. Row 1 reports the base case estimatesreported in MIT (2009), table 1. The cost estimates reported in row 2 incorporate updated construction

cost estimates rom U.S. Department o Energy (2010). Row 3, in addition, updates uel prices to reectthe most recent available prices or uranium, coal, and natural gas reported in U.S. DOE (2011a).Finally, row 4 continues to incorporate updated construction costs and uel prices and, in addition, addsa carbon tax o $25 per ton o carbon dioxide.


12/18


Another key actor is the availability o natural gas. Global capacity to transportAnother key actor is the availability o natural gas. Global capacity to transportliquefed natural gas is increasing rapidly but is still insufcient to have eliminatedliquefed natural gas is increasing rapidly but is still insufcient to have eliminatedprice dierences across countries.price dierences across countries.

Incorporating ExternalitiesIncorporating ExternalitiesLevelized cost estimates are typically designed to reect theLevelized cost estimates are typically designed to reect the private costs ocosts oinvesting in dierent orms o electricity generation. Thus they provide a basisinvesting in dierent orms o electricity generation. Thus they provide a basisor determining whether dierent types o plantsor determining whether dierent types o plants willbe built, but not necessarilybe built, but not necessarilyor determining whether dierent types o plantsor determining whether dierent types o plants shouldbe built.be built.The ourth row in Table 3 incorporates a tax o $25 per ton o carbon dioxide.The ourth row in Table 3 incorporates a tax o $25 per ton o carbon dioxide.

As a point o comparison, the Federal Interagency Working Group (2010) adopts aAs a point o comparison, the Federal Interagency Working Group (2010) adopts acentral social cost o carbon dioxide o $25 or 2015. Under this scenario, nuclearcentral social cost o carbon dioxide o $25 or 2015. Under this scenario, nuclearcontinues to have the highest levelized cost. The levelized cost o coal increases bycontinues to have the highest levelized cost. The levelized cost o coal increases by2 cents per kilowatt hour, but the levelized cost o natural gas increases by only about2 cents per kilowatt hour, but the levelized cost o natural gas increases by only about1 cent, not nearly enough to close the gap between nuclear and natural gas. More-1 cent, not nearly enough to close the gap between nuclear and natural gas. More-over, this static comparison based on current uel prices ignores that coal and naturalover, this static comparison based on current uel prices ignores that coal and naturalgas prices would likely all in response to a carbon tax. For both coal and natural gasgas prices would likely all in response to a carbon tax. For both coal and natural gasthere is a range o dierent sources available, much o which has a marginal cost othere is a range o dierent sources available, much o which has a marginal cost oextraction below current prices.extraction below current prices.Fossil uel plants also emit large amounts o local and regional pollutants.Fossil uel plants also emit large amounts o local and regional pollutants.Muller, Mendelsohn, and Nordhaus (2011) calculate that the external costs romMuller, Mendelsohn, and Nordhaus (2011) calculate that the external costs romsulur dioxide, nitrogen oxides, and particulates average 3.5 cents per kilowatt hoursulur dioxide, nitrogen oxides, and particulates average 3.5 cents per kilowatt houror coal, but only 0.1 cents per kilowatt hour or natural gas. Thus, incorporating theor coal, but only 0.1 cents per kilowatt hour or natural gas. Thus, incorporating theexternal costs o these pollutants improves the prospects considerably or nuclearexternal costs o these pollutants improves the prospects considerably or nuclearpower versus coal, but does little to close the gap versus natural gas. A comprehen-power versus coal, but does little to close the gap versus natural gas. A comprehen-sive welare analysis would also incorporate the negative production externalitiessive welare analysis would also incorporate the negative production externalitiesrom coal and natural gas. Perhaps most importantly, recent increases in shale gasrom coal and natural gas. Perhaps most importantly, recent increases in shale gasproduction have raised environmental concerns about water consumption andproduction have raised environmental concerns about water consumption andcontamination o drinking water. These costs are not yet well understood. However,contamination o drinking water. These costs are not yet well understood. However,the levelized cost estimates give some sense o how large these externalities wouldthe levelized cost estimates give some sense o how large these externalities wouldneed to be in order to make nuclear power the low-cost option.need to be in order to make nuclear power the low-cost option.There are also external costs associated with nuclear power. Included in theseThere are also external costs associated with nuclear power. Included in theselevelized cost estimates is a spent uel waste ee o 0.1 cents per kilowatt hour. Sincelevelized cost estimates is a spent uel waste ee o 0.1 cents per kilowatt hour. Since1983, the Department o Energy has collected this ee rom U.S. nuclear reactors,1983, the Department o Energy has collected this ee rom U.S. nuclear reactors,intended eventually to fnance a centralized storage acility or spent nuclear uel.intended eventually to fnance a centralized storage acility or spent nuclear uel.Currently, most spent nuclear uel is stored on-site in spent uel pools and dryCurrently, most spent nuclear uel is stored on-site in spent uel pools and drycasks. A comprehensive welare analysis would need to include both the privatecasks. A comprehensive welare analysis would need to include both the privateand external costs o this on-site storage. MIT (2010) and U.S. Nuclear Regulatoryand external costs o this on-site storage. MIT (2010) and U.S. Nuclear RegulatoryCommission (2011) discuss details o the nuclear uel cycle.Commission (2011) discuss details o the nuclear uel cycle.

Considerably harder to quantiy are the risks rom nuclear accidents. SinceConsiderably harder to quantiy are the risks rom nuclear accidents. Since1957, the PriceAnderson Act has indemnifed U.S. nuclear plant operators rom1957, the PriceAnderson Act has indemnifed U.S. nuclear plant operators romaccident liability above a certain cap, currently $12 billion. A Fukushima-typeaccident liability above a certain cap, currently $12 billion. A Fukushima-type


13/18


accident in the United States could easily cause damages well above this cap. It isaccident in the United States could easily cause damages well above this cap. It istoo early to measure the long-term external costs o Fukushima, but an early studytoo early to measure the long-term external costs o Fukushima, but an early studyestimates that radioactive contamination could cause 1,000 total cancer deathsestimates that radioactive contamination could cause 1,000 total cancer deaths(von Hippel 2011). As a point o comparison, cancer deaths rom Chernobyl are(von Hippel 2011). As a point o comparison, cancer deaths rom Chernobyl areestimated to be approximately 14,000 (Cardis et al. 2006). In addition to cancerestimated to be approximately 14,000 (Cardis et al. 2006). In addition to cancerdeaths, one would want to incorporate the costs rom other health outcomes, asdeaths, one would want to incorporate the costs rom other health outcomes, aswell as the pecuniary and psychological costs associated with relocating peoplewell as the pecuniary and psychological costs associated with relocating peopleliving near the accident site.living near the accident site.Perhaps hardest o all to measure are the risks associated with the proliera-Perhaps hardest o all to measure are the risks associated with the proliera-tion o nuclear weapons. This could come through the misuse o nuclear acilitiestion o nuclear weapons. This could come through the misuse o nuclear acilitiesto produce weapons materials, or rom a dirty bomb in which stolen radioactiveto produce weapons materials, or rom a dirty bomb in which stolen radioactivematerials rom any source are dispersed using conventional explosives. These risksmaterials rom any source are dispersed using conventional explosives. These risksare particularly acute in countries like France, the United Kingdom, and Japan thatare particularly acute in countries like France, the United Kingdom, and Japan thathave acilities or reprocessing nuclear waste. MIT (2003) and MIT (2010) discusshave acilities or reprocessing nuclear waste. MIT (2003) and MIT (2010) discussthese issues.these issues.Incorporating the external costs o nuclear power would urther increase theIncorporating the external costs o nuclear power would urther increase thegap between the levelized costs o nuclear and natural gas. An important prioritygap between the levelized costs o nuclear and natural gas. An important priorityor uture work is to refne measures o these external costs and incorporate themor uture work is to refne measures o these external costs and incorporate themexplicitly into levelized cost analyses. However, given current market conditions inexplicitly into levelized cost analyses. However, given current market conditions inthe United States, it becomes difcult to make an economic argument or nuclearthe United States, it becomes difcult to make an economic argument or nuclearpower regardless o the magnitude o these external costs. The frst challengepower regardless o the magnitude o these external costs. The frst challengecontinues to be construction costs, which are high enough that nuclear power strug-continues to be construction costs, which are high enough that nuclear power strug-gles to compete with natural gas even i one ignores these external costs completely.gles to compete with natural gas even i one ignores these external costs completely.

Learning-By-DoingLearning-By-DoingWhat would it take to reduce nuclear construction costs? One possibilityWhat would it take to reduce nuclear construction costs? One possibilityis learning-by-doing. In 2004, the Senior Vice President o the Nuclear Energyis learning-by-doing. In 2004, the Senior Vice President o the Nuclear EnergyInstitute testifed in ront o the U.S. Senate that nuclear construction costs wouldInstitute testifed in ront o the U.S. Senate that nuclear construction costs woulddecrease by 2030 percent ater the frst ew plants (Fertel 2004). In part on thedecrease by 2030 percent ater the frst ew plants (Fertel 2004). In part on thebasis o this testimony, the 2005 Energy Policy Act was drated to include loanbasis o this testimony, the 2005 Energy Policy Act was drated to include loan

guarantees, production tax credits, and other subsidies or new nuclear plants. Iguarantees, production tax credits, and other subsidies or new nuclear plants. Ilearning-by-doing could push construction costs down, this could change the equa-learning-by-doing could push construction costs down, this could change the equa-tion considerably or nuclear power. A substantial literature in economics indicatestion considerably or nuclear power. A substantial literature in economics indicatesthat learning-by-doing matters in a variety o markets (Alchian 1963; Joskow andthat learning-by-doing matters in a variety o markets (Alchian 1963; Joskow andRose 1985; Irwin and Klenow 1994; Benkard 2000; Thornton and Thompson 2001;Rose 1985; Irwin and Klenow 1994; Benkard 2000; Thornton and Thompson 2001;Kellogg 2011), and several studies have examined learning-by-doing in the construc-Kellogg 2011), and several studies have examined learning-by-doing in the construc-tion o nuclear power plants.tion o nuclear power plants.Recall that the time pattern o construction costs in Figure 3 did not provideRecall that the time pattern o construction costs in Figure 3 did not provideany immediate evidence o learning-by-doing. Instead, construction costs tended toany immediate evidence o learning-by-doing. Instead, construction costs tended toincreaseconsiderably over time. Several studies have nonetheless attempted to disen-considerably over time. Several studies have nonetheless attempted to disen-tangle learning-by-doing rom industrywide actors that were changing over time.tangle learning-by-doing rom industrywide actors that were changing over time.Using data rom the early nuclear builds, both Mooz (1978) and Komano (1981)Using data rom the early nuclear builds, both Mooz (1978) and Komano (1981)


14/18


fnd evidence o modest amounts o learning-by-doing in nuclear plant constructionfnd evidence o modest amounts o learning-by-doing in nuclear plant constructionthat accrue to the construction company in charge o the project, but no evidencethat accrue to the construction company in charge o the project, but no evidenceo industrywide learning-by-doing. Zimmerman (1982) also fnds learning-by-doingo industrywide learning-by-doing. Zimmerman (1982) also fnds learning-by-doingor the construction company and some evidence o spillovers across companies.or the construction company and some evidence o spillovers across companies.Using a longer panel, McCabe (1996) fnds evidence o learning-by-doing or bothUsing a longer panel, McCabe (1996) fnds evidence o learning-by-doing or boththe construction company and the utility managing the project, but does not testthe construction company and the utility managing the project, but does not testor industrywide learning.or industrywide learning.Learning-by-doing is important or the prospects o nuclear power becauseLearning-by-doing is important or the prospects o nuclear power becauseit provides a plausible mechanism by which nuclear construction costs couldit provides a plausible mechanism by which nuclear construction costs coulddecrease below the levels reported in Table 2. This is true regardless o whetherdecrease below the levels reported in Table 2. This is true regardless o whetheror not this learning-by-doing is privately captured. Who captures the learning-by-or not this learning-by-doing is privately captured. Who captures the learning-by-doing is important, however, or government policy. The economic argument ordoing is important, however, or government policy. The economic argument oran industry-specifc subsidy hinges on there being learning-by-doing that is notan industry-specifc subsidy hinges on there being learning-by-doing that is notcaptured by individual companies. I learning is ully appropriable, then frms acecaptured by individual companies. I learning is ully appropriable, then frms aceefcient incentives or investment and no government intervention is necessary.efcient incentives or investment and no government intervention is necessary.In addition, while there is almost certainly some industrywide learning-by-doing inIn addition, while there is almost certainly some industrywide learning-by-doing innuclear, there is also likely to be learning-by-doing in emerging energy technologiesnuclear, there is also likely to be learning-by-doing in emerging energy technologiessuch as wind, solar, and biomass. When there are a number o competing alterna-such as wind, solar, and biomass. When there are a number o competing alterna-tives, as in electricity generation, many economists avor broad-based subsidies thattives, as in electricity generation, many economists avor broad-based subsidies thatdo not single out individual technologies (Schmalensee 1980).do not single out individual technologies (Schmalensee 1980).Tied up in this discussion is a key tradeo between innovation and standardiza-Tied up in this discussion is a key tradeo between innovation and standardiza-tion. On the one hand, it is important to continue allowing or new and better reactortion. On the one hand, it is important to continue allowing or new and better reactordesigns with enhanced eatures or reliability and saety. On the other hand, requentdesigns with enhanced eatures or reliability and saety. On the other hand, requentredesigns make it harder to spread engineering costs across projects. The frst waveredesigns make it harder to spread engineering costs across projects. The frst waveo U.S. reactors were manuactured by our dierent companiesWestinghouse,o U.S. reactors were manuactured by our dierent companiesWestinghouse,General Electric, Combustion Engineering, and Babcock & Wilcoxeach withGeneral Electric, Combustion Engineering, and Babcock & Wilcoxeach withseveral dierent designs. At the time, such dierences were inevitable. The Unitedseveral dierent designs. At the time, such dierences were inevitable. The UnitedStates led the way in the development o commercial nuclear reactors and theStates led the way in the development o commercial nuclear reactors and thetechnology was evolving rapidly. Still, this diversity o designs provides a possibletechnology was evolving rapidly. Still, this diversity o designs provides a possibleexplanation or the lack o immediate evidence o learning-by-doing (Lester andexplanation or the lack o immediate evidence o learning-by-doing (Lester andMcCabe 1993).McCabe 1993).France oers a useul comparison on this point. Development o nuclearFrance oers a useul comparison on this point. Development o nuclearpower in France began later and with much less design variation. When Electricitpower in France began later and with much less design variation. When Electricitde France began seriously building reactors in the 1970s, it adopted a single designde France began seriously building reactors in the 1970s, it adopted a single designor all o its reactors. With one exception, all nuclear power reactors currently inor all o its reactors. With one exception, all nuclear power reactors currently inoperation in France are o exactly this same design (International Atomic Energyoperation in France are o exactly this same design (International Atomic EnergyAgency 2011). In addition, Electricit de France has long enjoyed a high degreeAgency 2011). In addition, Electricit de France has long enjoyed a high degreeo regulatory stability due to its close relationship with the French National Saetyo regulatory stability due to its close relationship with the French National SaetyAuthority and broad public support or nuclear power. Given this high degree oAuthority and broad public support or nuclear power. Given this high degree ostandardization, the apparent coststandardization, the apparent costescalationin French construction costs is particu-in French construction costs is particu-larly striking.larly striking.

Some within the U.S. nuclear industry claim that the United States is headedSome within the U.S. nuclear industry claim that the United States is headedmore toward the French model. For example, Michael Wallace, chairman o a majormore toward the French model. For example, Michael Wallace, chairman o a majornuclear power company predicted a couple o years ago (as quoted in Kanter 2009)nuclear power company predicted a couple o years ago (as quoted in Kanter 2009)


15/18

Lucas W. Davis 63

that new reactors would be standardized down to the carpeting and wallpaper.that new reactors would be standardized down to the carpeting and wallpaper.Perhaps the industry will quickly coalesce around a very small number o reactorPerhaps the industry will quickly coalesce around a very small number o reactordesigns, but this is not obvious based on applications received to date by the Nucleardesigns, but this is not obvious based on applications received to date by the NuclearRegulatory Commission. Among the 17 applications that have been received, thereRegulatory Commission. Among the 17 applications that have been received, thereis a mix o both pressurized water reactors and boiling water reactors, manuac-is a mix o both pressurized water reactors and boiling water reactors, manuac-tured by fve dierent reactor manuacturers (Areva, Westinghouse, Mitsubishi,tured by fve dierent reactor manuacturers (Areva, Westinghouse, Mitsubishi,GE-Hitachi, and GE). At a minimum, it seems clear that the French approach oGE-Hitachi, and GE). At a minimum, it seems clear that the French approach osupporting a single reactor design is not going to be adopted in the United States.supporting a single reactor design is not going to be adopted in the United States.

ConclusionConclusionNuclear power continues to generate enthusiasm based on its potential toNuclear power continues to generate enthusiasm based on its potential toreduce greenhouse gas emissions. A single pound o reactor-grade uranium oxidereduce greenhouse gas emissions. A single pound o reactor-grade uranium oxideproduces as much electricity as over 16,000 pounds o coalenough to meetproduces as much electricity as over 16,000 pounds o coalenough to meetthe needs o the average U.S. household or more than a year.the needs o the average U.S. household or more than a year.44 While burningWhile burning16,000 pounds o coal generates thousands o pounds o carbon dioxide, sulur16,000 pounds o coal generates thousands o pounds o carbon dioxide, sulurdioxide, and nitrogen oxidesnuclear power is virtually emissions-ree.dioxide, and nitrogen oxidesnuclear power is virtually emissions-ree.Nuclear power, however, is not without challenges. Fukushima has broughtNuclear power, however, is not without challenges. Fukushima has broughtto the oreront ongoing concerns about nuclear accidents and the handling andto the oreront ongoing concerns about nuclear accidents and the handling andstorage o spent uel. These external costs arestorage o spent uel. These external costs are in additionto substantial private costs.to substantial private costs.In 1942, with a shoestring budget in an abandoned squash court at the UniversityIn 1942, with a shoestring budget in an abandoned squash court at the University

o Chicago, Enrico Fermi demonstrated that electricity could be generated usingo Chicago, Enrico Fermi demonstrated that electricity could be generated usinga sel-sustaining nuclear reaction. Seventy years later the industry is still trying toa sel-sustaining nuclear reaction. Seventy years later the industry is still trying todemonstrate how this can be scaled up cheaply enough to compete with coal anddemonstrate how this can be scaled up cheaply enough to compete with coal andnatural gas.natural gas.

I am thankful to David Autor, Severin Borenstein, Chad Jones, John List, John Parsons,Timothy Taylor, Catherine Wolfram, and seminar participants at the University of

Tennessee, Brookings Institution, University of California Energy Institute, Berkeley Energy

and Resources Collaborative Symposium, California Public Utilities Commission, and the

University of Chicago for helpful comments.

4This is my own back-o-the-envelope calculation based on U.S. Department o Energy (2011a), table 8.2a

Electricity Net Generation, table 8.5a Consumption o Combustible Fuels, table 8.9 Electricity EndUse, table 9.3 Uranium Overview, and table 12.7b Emissions rom Energy Consumption or Elec-tricity Generation.


16/18


Reerences

Alchian, Armen. 1963. Reliability o Progress

Curves in Airrame Production. Econometrica31(4): 67994.

Benkard, C. Lanier. 2000. Learning andForgetting: The Dynamics o Aircrat Production.American Economic Review90(4): 103454.

Cardis, Elisabeth, et al. 2006. Estimates othe Cancer Burden in Europe rom RadioactiveFallout rom the Chernobyl Accident. Interna-tional Journal of Cancer119(6): 122435.

Cox, Alan J., and Richard J. Gilbert. 1991. AnEconomic Evaluation o the Costs and Benefts oDiablo Canyon. In Regulatory Choices: A Perspectiveon Developments in Energy Policy, edited by Richard

J. Gilvert, 26089. Berkeley: University o CaliorniaPress.

Davis, Lucas W., and Catherine Wolram. 2011.Deregulation, Consolidation, and Efciency:Evidence rom U.S. Nuclear Power. EI@Haas WP217, Energy Institute at Haas.

Du, Yangbo, and John E. Parsons. 2009.Update on the Cost o Nuclear Power. MITCenter or Energy and Environmental PolicyResearch Working Paper 09-004.

Federal Interagency Working Group. 2010Technical Support Document: Social Cost o

Carbon or Regulatory Impact Analysis: UnderExecutive Order 12866. http://www.epa.gov/otaq/climate/regulations/scc-tsd.pd.

Fertel, Marvin S. (Senior Vice President and

Chie Nuclear Ofcer o the Nuclear Energy Insti-

tute). 2004.Testimony beore the Senate Energy andNatural Resources Committee, U.S. Senate. March4. http://www.nei.org/publicpolicy/congressionaltestimony/energysubcmteertelextended/ .

Grubler, Arnul. 2010. The Costs o the FrenchNuclear Scale-Up: A Case o Negative Learning byDoing.Energy Policy38(9): 517488.

International Atomic Energy Agency. 2011.

Nuclear Power Reactors in the World. 2011 Edition.Vienna: IAEA.

Irwin, Douglas A., and Peter J. Klenow. 1994.Learning-by-Doing Spillovers in the Semicon-ductor Industry. Journal of Political Economy102(6): 120027.

Ives, Nathan, Steve McCabe, and Gary

Gilmartin. 2010 Nuclear Renaissance and theGlobal Supply Chain: Avoiding Pitalls, RealizingBenefts. Public Utilities Fortnightly, December 1.

Joskow, Paul L. 1974. Ination and Environ-mental Concern: Structural Change in the Process

o Public Utility Price Regulation.Journal and Lawand Economics17(2): 291327.Joskow, Paul L. 1982. Problems and Prospects

or Nuclear Energy in the United States. InEnergy,

Economics and the Environment, edited by GregoryA. Daneke, 23154. Lexington: D.C. Heath andCompany.

Joskow, Paul L. 2006. The Future o NuclearPower in the United States: Economic and Regula-tory Challenges. MIT CEEPR Working Paper06-019, Massachusetts Institute o TechnologyCenter or Energy and Environmental PolicyResearch.

Joskow, Paul L., and John E. Parsons. 2009.The Economic Future o Nuclear Power.

Daedalus138(4): 4559.Joskow, Paul L., and Nancy L. Rose. 1985. The

Eects o Technological Change, Experience, andEnvironmental Regulation on the ConstructionCost o Coal-Burning Generating Units. RAND

Journal of Economics16(1): 117.Joskow, Paul L., and Joel Yellin. 1980. Siting

Nuclear Power Plants. Virginia Journal of NaturalResource Law1(1): 167.

Kanter, James. 2009. In Finland, NuclearRenaissance Runs into Trouble. New York Times,May 28. http://www.nytimes.com/2009/05/29/business/energy-environment/29nuke.html?pagewanted=all.

Kellogg, Ryan. 2011. Learning by Drilling:Inter-Firm Learning and Relationship Persis-tence in the Texas Oilpatch. Quarterly Journalof Economics126(4): 19612004.

Komano, Charles. 1981. Power Plant Escala-tion: Nuclear and Coal Capital Costs, Regulation,

and Economics. New York: Van Nostrand ReinholdCompany.

Koomey, Jonathan, and Nathan E. Hultman.

2007. A Reactor-Level Analysis o Busbar Costsor U.S. Nuclear Plants, 19702005.Energy Policy35(11): 563042.

Lester, Richard K., and Mark J. McCabe. 1993.

The Eect o Industrial Structure on Learning byDoing in Nuclear Power Plant Operation. RAND

Journal of Economics24(3): 41838.McCabe, Mark J. 1996. Principals, Agents, and

the Learning Curve: The Case o Steam-ElectricPower Plant Design and Construction. Journal ofIndustrial Economics44(4): 35775.

McCallion, Kenneth. 1995. Shoreham and theRise and Fall of the Nuclear Power Industry. Praeger,

Westport.MIT. 2003. The Future of Nuclear Power. An

Interdisciplinary MIT Study. MIT Energy Initiative,

Massachusetts Institute o Technology.MIT. 2009. Update of the MIT 2003 Future ofNuclear Power. An Interdisciplinary MIT Study.
http://www.epa.gov/otaq/climate/regulations/scc-tsd.pdfhttp://www.epa.gov/otaq/climate/regulations/scc-tsd.pdfhttp://www.nei.org/publicpolicy/congressionaltestimony/energysubcmtefertelextended/http://www.nei.org/publicpolicy/congressionaltestimony/energysubcmtefertelextended/http://www.nytimes.com/2009/05/29/business/energy-environment/29nuke.html?pagewanted=allhttp://www.nytimes.com/2009/05/29/business/energy-environment/29nuke.html?pagewanted=allhttp://www.nytimes.com/2009/05/29/business/energy-environment/29nuke.html?pagewanted=allhttp://www.nytimes.com/2009/05/29/business/energy-environment/29nuke.html?pagewanted=allhttp://www.nytimes.com/2009/05/29/business/energy-environment/29nuke.html?pagewanted=allhttp://www.nytimes.com/2009/05/29/business/energy-environment/29nuke.html?pagewanted=allhttp://www.nei.org/publicpolicy/congressionaltestimony/energysubcmtefertelextended/http://www.nei.org/publicpolicy/congressionaltestimony/energysubcmtefertelextended/http://www.epa.gov/otaq/climate/regulations/scc-tsd.pdfhttp://www.epa.gov/otaq/climate/regulations/scc-tsd.pdf


17/18


MIT Energy Initiative, Massachusetts Institute oTechnology.

MIT. 2010. The Future of the Nuclear Fuel Cycle.An Interdisciplinary MIT Study. MIT Energy Initia-tive, Massachusetts Institute o Technology.

Mooz, William E. 1978. Cost Analysis of LightWater Reactor Power Plants.RAND Report, Preparedor the Department o Energy. R2304-DOE, June.RAND Corporation.

Muller, Nicholas Z., Robert Mendelsohn,

and William Nordhaus. 2011. EnvironmentalAccounting or Pollution in the United StatesEconomy. American Economic Review 101(5),164975.

OECD Nuclear Energy Agency and Interna-

tional Atomic Energy Agency. 2009. Uranium 2009:

Resources, Production, and Demand. (The RedBook.) OECD, Nuclear Energy Agency.

Pindyck, Robert S. 1993. Investment o Uncer-tain Cost. Journal of Financial Economics 34(1):5376.

Schmalensee, Richard. 1980. AppropriateGovernment Policy toward Commercialization oNew Energy Supply Technologies.Energy Journal1(2): 140.

Thornton, Rebecca A., and Peter Thompson.

2001. Learning rom Experience and Learningrom Others: An Exploration o Learning andSpillovers in Wartime Shipbuilding. American

Economic Review91(5): 135068.U.S. Atomic Energy Commission, Ofce o

Planning and Analysis. 1974. Nuclear Power Growth,19742000. WASH-1139(74). U.S. Department oEnergy.

U.S. Department o Energy (DOE), Energy

Inormation Administration. 1986. An Analysisof Nuclear Power Plant Construction Costs. DOE/EIA-0485. U.S. Department o Energy.


Inormation Administration. 1997. Nuclear PowerGeneration and Fuel Cycle Report 1997. DOE/EIA-0436(97). U.S. Department o Energy.


Inormation Administration. 2010. Updated CapitalCost Estimates for Electricity Generation Plants.November. U.S. Department o Energy. http://

www.eia.gov/oia/beck_plantcosts/.U.S. Department o Energy (DOE), Energy

Inormation Administration. 2011a. AnnualEnergy Review 2010. DOE/EIA-0384, October.U.S. Department o Energy.

U.S. Department o Energy (DOE), EnergyInormation Administration. 2011b. Annual EnergyOutlook 2011. DOE/EIA-0383, April. U.S. Depart-ment o Energy.

U.S. Nuclear Regulatory Commission (NRC).

2011. Information Digest 20112012. NUREG-1350,Volume 23, August.

von Hippel, Frank N. 2011. The Radiologicaland Psychological Consequences o the FukushimaDaiichi Accident. Bulletin of the Atomic Scien-tists67(5): 2736.

Wald, Matthew L. 2010. Giant Holes in theGround. MIT Technology Review November

/December, http://www.technologyreview.com/energy/26542/.

Zimmerman, Martin B. 1982. Learning Eectsand the Commercialization o New Energy Tech-nologies: The Case o Nuclear Power. Bell Journalof Economics13(2): 297310.
http://www.eia.gov/oiaf/beck_plantcosts/http://www.eia.gov/oiaf/beck_plantcosts/http://www.technologyreview.com/energy/26542/http://www.technologyreview.com/energy/26542/http://www.technologyreview.com/energy/26542/http://www.technologyreview.com/energy/26542/http://www.eia.gov/oiaf/beck_plantcosts/http://www.eia.gov/oiaf/beck_plantcosts/


18/18


davis_prospects for nuclear power_jep.26.1

Documents