March 10th 2012March 10th 2012

S P E C I A L R E P O R T

N U C L E A R E N E R G Y

The dreamthat failed

SRnuclearEnergy.indd 1 27/02/2012 13:27

1

THE LIGHTS ARE not going o� all over Japan, but the nuclear powerplants are. Of the 54 reactors in those plants, with a combined capacity of47.5 gigawatts (GW, a thousand megawatts), only two are operating to-day. A good dozen are unlikely ever to reopen: six at Fukushima Dai-ichi,which su�ered a calamitous triple meltdown after an earthquake andtsunami on March 11th 2011 (pictured above), and others either too closeto those reactors or now considered to be at risk of similar disaster. Therest, bar two, have shut down for maintenance or �stress tests� since theFukushima accident and not yet been cleared to start up again. It is quitepossible that none of them will get that permission before the two stillrunning shut for scheduled maintenance by the end of April.

Japan has been using nuclear power since the 1960s. In 2010 it got30% of its electricity from nuclear plants. This spring it may well join theranks of the 150 nations currently muddling through with all their atomsunsplit. If the shutdown happens, it will not be permanent; a good num-ber of the reactors now closed are likely to be reopened. But it could stillhave symbolic importance. To do without something hitherto seen as anecessity opens the mind to new possibilities. Japan had previously ex-pected its use of nuclear energy to increase somewhat. Now the share ofnuclear power in Japan’s energy mix is more likely to shrink, and it couldjust vanish altogether.

In most places any foretaste of that newly plausible future will bare-ly be noticed. Bullet trains will �ash on; �at panels will continue to shine;toilet seats will still warm up; factories will hum as they hummed before.Almost everywhere, when people reach for the light switches in theirhomes, the lights will come on. But not quite everywhere. In Futaba, Na-mie and Naraha the lights will stay o�, and no factories will hum: not forwant of power but for want of people. The 100,000 or so people thatonce lived in those and other towns close to the Fukushima Dai-ichi nuc-lear power plant have been evacuated. Some 30,000 may never return.

The triple meltdown at Fukushima a year ago was the world’s

The dream that failed

Nuclear power will not go away, but its role may never be more

than marginal, says Oliver Morton

A C K N O W L E D G M E N T S

CONTENT S

As well as those mentioned in the

text, the author would like to thank

the following for their generous

help: Graham Allison, George

Apostolakis, Jun Arima, Jean-Paul

Bouttes, Tom Burke, Ron Cameron,

Armand Cohen, Timothy Collier, Luis

Echávarri, Mike Fowler, Malcolm

Grimshaw, Grenville Harrop, Tatsuo

Hatta, Frank von Hippel, Tetsunari

Iida, Philippe Knoche, Shigeaki

Koga, Steve Koonin, Per Lekander,

Richard Lester, Edwin Lyman, Hervé

Machenaud, Brendan Mrosak, Elon

Musk, George Perkovich, Scott

Peterson, Vic Reis, Mycle Schneider,

Tatsujiro Suzuki, Nate Taplin,

Stephen Tindale, Masakazu Toyoda,

Jack Wan, David Wark and Fuqiang

Yang.

A list of sources is at

Economist.com/specialreports

An audio interview with

the author is at

Economist.com/audiovideo/specialreports

4 A brief historyFrom squash court tosubmarine

7 SafetyBlow-ups happen

9 Nuclear wasteLeave well alone

10 CostsBandwagons and busts

12 ProspectsOver the rainbow

SPECIAL REPORT

NUCLEAR ENERG Y

The Economist March 10th 2012 1

2 The Economist March 10th 2012

NUCLEAR ENERG Y

SPECIAL REPORT

2

1

worst nuclear accident since the disaster at Chernobyl in the Uk-raine in 1986. The damage extends far beyond a lost power sta-tion, a stricken operator (the Tokyo Electric Power Company, orTepco) and an intense debate about the future of the nation’s nu-clear power plants. It goes beyond the trillions of yen that will beneeded for a decade-long e�ort to decommission the reactorsand remove their wrecked cores, if indeed that proves possible,and the even greater sums that may be required for decontami-nation (which one expert, Tatsuhiko Kodama of Tokyo Universi-ty, thinks could cost as much as ¥50 trillion, or $623 billion). Itreaches deep into the lives of the displaced, and of those furthera�eld who know they have been exposed to the fallout from thedisaster. If it leads to a breakdown of the near-monopolies en-joyed by the country’s power companies, it will strike at some ofthe strongest complicities within the business-and-bureaucracyestablishment.

For parallels that do justice to the disaster, the Japanese �ndthemselves reaching back to the second world war, otherwiseseldom discussed: to the battle of Iwo Jima to describe the hero-ism of everyday workers abandoned by the o�cer class of com-pany and government; to the imperial navy’s ill-judged infatua-tion with battleships, being likened to the establishment’seagerness for ever more reactors; to the war as a whole as a mea-sure of the sheer scale of the event. And, of course, to Hiroshima.Kiyoshi Kurokawa, an academic who is heading a commissioninvestigating the disaster on behalf of the Japanese parliament,thinks that Fukushima has opened the way to a new scepticismabout an ageing, dysfunctional status quo which could bring

about a �third opening� of Japan comparable to the Meiji restora-tion and the American occupation after 1945.

To the public at large, the history of nuclear power is mostlya history of accidents: Three Mile Island, the 1979 partial melt-down of a nuclear reactor in Pennsylvania caused by a faultyvalve, which led to a small release of radioactivity and the tem-porary evacuation of the area; Chernobyl, the 1986 disaster in theUkraine in which a chain reaction got out of control and a reactorblew up, spreading radioactive material far and wide; and nowFukushima. But the �eld has been shaped more by broad eco-nomic and strategic trends than sudden shocks.

The renaissance that wasn’t

America’s nuclear bubble burst not after the accident atThree Mile Island but �ve years before it. The French nuclear-power programme, the most ambitious by far of the 1980s, con-tinued largely undisturbed after Chernobyl, though other coun-tries did pull back. The West’s �nuclear renaissance� much bruit-ed over the past decade, in part as a response to climate change,�zzled out well before the roofs blew o� Fukushima’s �rst, thirdand fourth reactor buildings. Today’s most dramatic nuclear ex-pansion, in China, may be tempered by Fukushima, but it willnot be halted.

For all that, Fukushima is a heavier blow than the previoustwo. Three Mile Island, disturbing as it was, released relatively lit-tle radioactivity and killed nobody. By causing nuclear safety tobe tightened and buttressed with new institutions, it improvedthe industry’s reliability and pro�tability in America. Chernobyl

If you can read thisyou must be

Indian

THREE MILE ISLAND MARCH 28TH 1979

CHERNOBYLAPRIL 26TH

1986

FUKUSHIMADAI-ICHIMARCH 11TH2011

Over 500

150-500

100-149.9

50-99.9

25-49.9

10-24.9

Less than 10

Generation of nuclear powerTWh, 2010

Interactive: Explore our guide to nuclear power around the world at Economist.com/reactors2012 Sources: "Nuclear Power in a Post-Fukushima World", Mycle Schneider et al; IAEA

0

100

200

300

400

500

0

50

100

150

200

250

300

350

400

1954 1960 1970 1980 1990 2000 10 12

Number of reactors in operation

Operablecapacity,

GWe

NorthAmerica

+0.6

LatinAmerica

-0.1

Africa0.0

Far East+2.4

WesternEurope

-0.3

EasternEurope

+0.4

World+0.7

Middle East and South Asia

+4.7

Average annual growth in nuclearcapacity, 2000-10, %

The Economist March 10th 2012 3

SPECIAL REPORTNUCLEAR ENERG Y

2

1

was far worse, but it was caused by egregious operator error in atotalitarian regime incapable of the sort of transparency and ac-countability needed to ensure nuclear safety. It put paid to nuc-lear power in Italy and, for a while, Sweden, but in general itcould be treated as an aberration of little direct relevance to thefree world’s nuclear programmes. Poor regulation, an insu�-cient safety culture and human error (without which the Japa-nese tsunami’s e�ects might have been very di�erent) are muchmore worrying when they strike in a tech-nologically advanced democracy work-ing with long-established reactor designs.

And if the blow is harder than theprevious one, the recipient is less robustthan it once was. In liberalised energymarkets, building nuclear power plants isno longer a commercially feasible option:they are simply too expensive. Existing re-actors can be run very pro�tably; their ca-pacity can be upgraded and their lives ex-tended. But forecast reductions in thecapital costs of new reactors in Americaand Europe have failed to materialise andconstruction periods have lengthened.Nobody will now build one withoutsome form of subsidy to �nance it or a pro-mise of a favourable deal for selling theelectricity. And at the same time as the costof new nuclear plants has become prohib-itive in much of the world, worries aboutthe dark side of nuclear power are resur-gent, thanks to what is happening in Iran.

Nuclear proliferation has not gone as far or as fast as wasfeared in the 1960s. But it has proceeded, and it has done so handin hand with nuclear power. There is only one state with nuclearweapons, Israel, that does not also have nuclear reactors to gen-erate electricity. Only two non-European states with nuclearpower stations, Japan and Mexico, have not at some point takensteps towards developing nuclear weapons, though most havepulled back before getting there.

If proliferation is one reason for treating the spread of nuc-lear power with caution, renewable energy is another. In 2010the world’s installed renewable electricity capacity outstrippedits nuclear capacity for the �rst time. That does not mean that theworld got as much energy from renewables as from nuclear; re-actors run at up to 93% of their stated capacity whereas wind andsolar tend to be closer to 20%. Renewables are intermittent and

take up a lot of space: generating a gigawatt of electricity withwind takes hundreds of square kilometres, whereas a nuclear re-actor with the same capacity will �t into a large industrial build-ing. That may limit the contribution renewables can ultimatelymake to energy supply. Unsubsidised renewables can currentlydisplace fossil fuels only in special circumstances. But nuclearenergy, which has received large subsidies in the past, has notdisplaced much in the way of fossil fuels either. And nuclear isgetting more expensive whereas renewables are getting cheaper.

Ulterior motives

Nuclear power is not going to disappear. Germany, whichin 2011 produced 5% of the world’s nuclear electricity, is aban-doning it, as are some smaller countries. In Japan, and perhapsalso in France, it looks likely to lose ground. But there will alwaysbe countries that �nd the technology attractive enough to makethem willing to rearrange energy markets in its favour. If theyhave few indigenous energy resources, they may value, as Japanhas done, the security o�ered by plants running on fuel that ischeap and easily stockpiled. Countries with existing nuclear ca-pacity that do not share Germany’s deep nuclear unease or itsenthusiasm for renewables may choose to buy new reactors toreplace old ones, as Britain is seeking to do, to help with carbonemissions. Countries committed to proliferation, or at least inter-ested in keeping that option open, will invest in nuclear, as maycountries that �nd themselves with cash to spare and a wish tojoin what still looks like a technological premier league.

Besides, nuclear plants are long-lived things. Today’s reac-tors were mostly designed for a 40-year life, but many of themare being allowed to increase it to 60. New reactor designs aimfor a span of 60 years that might be extended to 80. Given that ittakes a decade or so to go from deciding to build a reactor to feed-ing the resulting electricity into a grid, reactors being plannednow may still be working in the early 22nd century.

Barring major technological developments, though, nuc-lear power will continue to be a creature of politics not econom-ics, with any growth a function of political will or a side-e�ect ofprotecting electrical utilities from open competition. This willlimit the overall size of the industry. In 2010 nuclear power pro-vided 13% of the world’s electricity, down from 18% in 1996. A pre-Fukushima scenario from the International Energy Agency thatallowed for a little more action on carbon dioxide than has yetbeen taken predicted a rise of about 70% in nuclear capacity be-tween 2010 and 2035; since other generating capacity will begrowing too, that would keep nuclear’s 13% share roughly con-stant. A more guarded IEA scenario has rich countries buildingno new reactors other than those already under construction,other countries achieving only half their currently stated targets(which in nuclear matters are hardly ever met) and regulators be-ing less generous in extending the life of existing plants. On thatbasis the installed capacity goes down a little, and the share ofthe electricity market drops to 7%.

Developing nuclear plants only at the behest of govern-ment will also make it harder for the industry to improve its safe-ty culture. Where a government is convinced of the need for nuc-lear power, it may well be less likely to regulate it in the stringent,independent way the technology demands. Governments fa-

1Decline and fall

Source: Institute for Sustainable Energy Policies, Tokyo

Scenarios for Japan’s nuclear capacity, GW

0

70

80

60

50

40

30

20

10

2010 15 20 25 30 35 40 45 50

2020phaseout

No new build and 40-year lifetime

Previous government plan

Post-Fukushima closuresand 40-year lifetime

Barring major technological developments, nuclearpower will continue to be a creature of politics noteconomics. This will limit the overall size of the industry

4 The Economist March 10th 2012

NUCLEAR ENERG Y

SPECIAL REPORT

2

1

THE NUCLEAR AGE began 70 years ago on a squash courtin Chicago, under the watchful eye of a man with an axe. A

team led by Enrico Fermi, Italy’s greatest physicist since Galileo,had been building a nuclear �pile� for weeks, slotting pellets ofuranium and bricks of graphite into a carefully planned geome-try through which ran various �control rods� of cadmium. Thesquash court was the only convenient large space available onthe university campus. On December 2nd 1942 the pile hadgrown large enough to allow a nuclear reaction to take o� whenthe control rods were drawn back from its heart.

The axeman was there in case the reaction went out of con-trol. If it did he would chop through a rope, sending the maincontrol rod crashing back into place, absorbing the neutronsdriving the reaction and restoring stability. Like many of Fermi’sideas, it had the charm of simplicity. Today every commercialpower reactor has control rods poised to shut it down at a mo-ment’s notice, a procedure called a scram�in honour, so it is said,of Chicago’s �safety control rod axeman�.

On that �rst occasion, nothing went wrong. As the pile’sother control rods were mechanically withdrawn, radiationcounters ticked up. Once Fermi was satis�ed that they wereshowing a true chain reaction, he had the rods reinserted. A cele-bratory bottle of Chianti was opened. A coded phone call in-formed the head of the National Defence Research Committeethat �the Italian navigator has landed in the new world.� Theaxeman put down his axe.

The energy output of that �rst reactor was tiny: just half awatt. Today’s most powerful reactors produce ten billion timesas much energy in the form of heat, about a third of which can beconverted into electricity. Five gigawatts is an amount beyondeasy comprehension, the daily equivalent of the energy given

A brief history

From squash court to

submarine

Nuclear reactors and their uses have not changed

much over seven decades

vour nuclear power by limiting the liability of its operators. Ifthey did not, the industry would surely founder. But a di�erentrisk arises from the fact that governments can change theirminds. Germany’s plants are being shut down in response to anaccident its industry had nothing to do with. Being hostage todistant events thus adds a hard-to-calculate systemic risk to nuc-lear development.

The ability to split atoms and extract energy from them wasone of the more remarkable scienti�c achievements of the 20thcentury, widely seen as world-changing. Intuitively one mightexpect such a scienti�c wonder either to sweep all before it or berenounced, rather than end up in a modest niche, at best stable,at worst dwindling. But if nuclear power teaches one lesson, it isto doubt all stories of technological determinism. It is not the es-sential nature of a technology that matters but its capacity to �tinto the social, political and economic conditions of the day. If atechnology �ts into the human world in a way that gives it evermore scope for growth it can succeed beyond the dreams of itspioneers. The diesel engines that power the world’s shipping arean example; so are the arti�cial fertilisers that have allowed evermore people to be supplied by ever more productive farms, andthe computers that make the world ever more hungry for yetmore computing power.

There has been no such expansive setting for nuclear tech-nologies. Their history has for the most part been one of concen-tration not expansion, of options being closed rather thanopened. The history of nuclear weapons has been de�ned byavoiding their use and constraining the number of their possess-ors. Within countries they have concentrated power. As theAmerican political commentator Gary Wills argues in his book,�Bomb Power�, the increased strategic role of the American pres-idency since 1945 stems in signi�cant part from the way that nuc-lear weapons have rede�ned the role and power of the �com-mander-in-chief� (a term previously applied only in the contextof the armed forces, not the nation as a whole) who has his �ngeron the button. In the energy world, nuclear has found its placenourishing technophile establishments like the �nuclear village�of vendors, bureaucrats, regulators and utilities in Japan whoselack of transparency and accountability did much to pave theway for Fukushima and the distrust that has followed in itswake. These political settings govern and limit what nuclearpower can achieve. 7

The way the future was

The Economist March 10th 2012 5

SPECIAL REPORTNUCLEAR ENERG Y

2

1

o� by six bombs like the one that destroyed Hiroshima. Imaginethat energy coursing through a few swimming pools-worth ofwater at three times the normal boiling point, trapped in a steelcylinder under pressures found a mile below the sea, and youhave a sense of the hellish miracle that is a modern reactor.

If �ights had lasted a billion times longer 70 years after theWright brothers’ �rst one took o�, they would have gone a thou-sand times round the world and taken centuries; a billion timesfaster, and they would have run up against the speed of light.Even at the heady rates of progress that Moore’s law ascribes tothe computer industry (stating that the number of transistors ona chip doubles roughly every two years), things take 60 years toget a billion times better.

But such comparisons �atter nuclear technology. TheWright brothers wanted their �rst aircraft to �y as far and as fastas it could; the �rst reactor was designed to do things in as smalland safe a way as possible. At the time Fermi demonstrated the�rst controlled chain reaction, the uncontrolled ones that woulddevastate Hiroshima and Nagasaki were already being planned.Reactors capable of generating hundreds of megawatts of heatwere on his colleagues’ drawing boards. Nuclear power did notgrow steadily over decades the way aircraft and computers have

done. It blossomed fast, then held strangely steady.Using reactors to generate electricity was not an early prior-

ity. The reactors of the Manhattan Project�the wartime nuclearprogramme begun in earnest shortly after Fermi’s success�weredesigned to further the project’s only goal: making bombs. A nu-clear chain reaction, whether in a reactor or an exploding bomb,comes about when the splitting of a ��ssile� nucleus by a neu-tron produces neutrons that will go on to split further �ssile nu-clei. The �ssile nuclei in Fermi’s reactor were of a particular typeor �isotope� of uranium, U-235. In natural uranium, only sevennuclei in every 1,000 are of this �ssile sort. That was not a pro-blem for Fermi’s reactor. The addition of graphite�a �modera-tor��slowed down the neutrons that were being given o�,which made them better at splitting other nuclei and enabled achain reaction to take place even when �ssile nuclei were scarce.Bombs are not the place for moderation: to make a uraniumbomb you need a core that is almost entirely U-235. As separatingout U-235 is extremely di�cult, the Manhattan Project’s phys-icists were not sure they could provide it on the scale that bomb-makers would require.

That created a need for an alternative source of �ssile mate-rial, and reactors provided it. When a neutron hits one of thenon-�ssile uranium nuclei� the vast majority�it can turn it intoa new element: plutonium. Plutonium nuclei are �ssile, and get-ting a bit of plutonium out of uranium that has been sitting in areactor is far easier than separating uranium isotopes. Reactorscould thus serve as plutonium factories, and the early ones wereused exclusively for that purpose. By the mid-1950s some reac-tors in Britain and France were generating electricity as well; theyneeded to be cooled anyway, and using the gas that cooled themto drive steam turbines was good public relations. But their mainpurpose was still to provide fuel for bombs.

Rickover’s killer app

Almost all of today’s nuclear power plants have a di�erentlineage. Hyman Rickover, a redoubtable American submariner,saw a niche for nuclear power plants in submarines. The diesel-powered sort needed to take on air through a snorkel; nuclear-powered ones would be able to stay submerged inde�nitely. Buta graphite-moderated reactor would never be compact enoughfor submarines. Rickover eventually settled on a reactor designthat economised on space by using water as both moderator andcoolant�a pressurised water reactor, or PWR. In many ways thisis a poor compromise. Water cannot be heated much above350°C, even under pressure, and still stay liquid. That limits the

Fermi and his nuclear family

6 The Economist March 10th 2012

NUCLEAR ENERG Y

SPECIAL REPORT

2 e�ciency of energy conversion (the hotterthe better in such matters). And normalwater is not a very good moderator. So-called �heavy water�, which contains adi�erent hydrogen isotope, is better, but alot harder to come by than the �light� type.

To make up for poor moderation, alight-water reactor needs fuel enriched inU-235. It does not have to be enriched asmuch as uranium for bomb-making does,but the enrichment systems used to makefuel for such reactors can almost as easilybe used to make the weapons-grade ura-nium that bomb-makers need. This is thetechnological basis of the stand-o� withIran, which has claimed unconvincinglythat its enrichment facilities are just for re-actor fuel. Other combinations of fuel andmoderator would have allowed the use ofnon-enriched uranium, and could indeedhave got by without producing plutoniumas a waste product, thus establishing amuch clearer dividing line between nuc-lear power and nuclear weapons. But theywould not have powered the submarinesof the 1960s.

In the early 1950s nuclear physicists were for the most partunexcited by the light-water reactor’s potential. By the time thenuclear age was just ten years old they already knew of manykinds of fuel, various moderators (as well as designs that neededno moderation) and many ways of getting heat out of reactors.They were excited by �breeder� reactors that both burned and

created plutonium. America’s �rst wasbuilt in 1951, and the assumption was thatany nuclear-energy economy worth thename would make use of the technol-ogy’s miraculous ability to produce its

own fuel. The fascination continued for decades. Breeder reac-tors have been built in Russia, Britain, Germany, India, China,France and Japan as well as America. But they have not provedremotely attractive enough for commercial development in aworld which seemed to have plenty of uranium.

By 1952 bomb-makers had also massively multiplied thepower of their wares by adding nuclear fusion, in which energyis liberated by adding together small nuclei rather than splittingapart large ones. Schemes to make reactors on the same principlesprang up immediately. Sixty years on, this line of research con-tinues, at great expense, without any prospect of commercialplausibility.

Make mine a PWR

The navy did not need exciting ideas; it needed submarinepower plants that used available technologies. In Rickover it hadan organisational genius capable of creating the industrial baseneeded to provide them, choosing and training the naval engi-neers needed to operate them, and instilling in that cadre the me-ticulous safety culture needed to stop the reactors from goinghaywire. Under Rickover’s tutelage American industry learned

to make PWRs, which it went on to o�er to electrical utilities. SoPWRs became the mainstay of America’s nuclear-power indus-try as it grew up in the 1960s, with the boiling-water reactor(BWR)�a similar light-water design, less e�cient and unseawor-thy but in some ways simpler and possibly safer�providing analternative. BWRs currently make up 21% of the world’s nuclearcapacity, but that �gure is set to diminish, not so much becausethe Fukushima reactors were of that kind (old PWRs in a similarsetting would not necessarily have fared better) but because thefew countries that ever went for the technology on any scalehave little appetite for new plants (America) or none at all (Ger-many and Japan). The 68% of the world’s nuclear electricity from

PWRs is thus set to increase, with othertechnologies trailing way behind.

Such homogeneity in a 70-year-oldhigh-technology enterprise is remark-able. Seven decades after the Wrightbrothers’ �rst �ight there were warplanesthat could travel at three times the speed

of sound, rockets that could send men to the moon, airliner �eetsthat carried hundreds of thousands of passengers a day, helicop-ters that could land on top of skyscrapers. Include unmannedspacecraft, and there really were �ights a billion times as long asthe Wright brothers’ �rst and lasting for years. But aircraft werecapable of diversity and evolution and could be developedcheaply by small teams of engineers. It is estimated that duringthe 1920s and 1930s some 100,000 types of aircraft were tried out.

Developing a nuclear reactor, on the other hand, has neverbeen a matter for barnstorming experimentation, partly becauseof the risks and partly because of the links to the technologies ofthe bomb. And whereas there are lots of things you can do withaircraft, more or less the only practical things you can do with areactor are to make plutonium for bombs, power submarines,produce isotopes used in medicine and generate heat and elec-tricity. Only the last is big business, and it can easily be done byother means. So options have been closed down and eggs havebeen piled into single baskets. The world went with pretty muchthe �rst sort of reactor it saw deployed at scale, contenting itselfwith increasing its size and trying, over the years, to render it everless in need of the attentions of the axeman. 7

Admirable admiral

Developing a nuclear reactor has neverbeen a matter for barnstormingexperimentation

The Economist March 10th 2012 7

SPECIAL REPORTNUCLEAR ENERG Y

1

FOR THE SURFERS o�shore the wall will be almost invisi-ble, hidden behind the existing sand dunes and pine trees.

From the land it will tower 10-12 metres above the Hamaoka nuc-lear power plant’s perimeter road. It will be 1.6 kilometres longand two metres thick; its foundations will be deeper than thewall itself is tall. It will weigh the best part of 1m tonnes. This iswhat Japan’s Chubu Electric Power thinks it will take to stop atsunami a touch bigger than the one that hit Fukushima, which isslightly farther from Tokyo to the north-east than Hamaoka is tothe south-west. Chubu expects to have the wall �nished by theend of this year. Until then Hamaoka’s three reactors�two ofthem similar to those at Fukushima�remain shut down.

At Fukushima the 14-metre tsunami easily topped the inad-equate defences. It �ooded all but one of the plant’s back-up die-sel generators and trashed the pumps meant to dump the reac-tors’ waste heat into the sea. The plant’s reactors had beenscrammed 40 minutes before the wave hit, but although shuttingdown a reactor’s chain reaction lowers its heat output by about97% almost instantaneously, the other 3% takes some time todrop to negligible levels, and that still amounts to a lot of heat.With no electricity either from the grid or from diesel generatorsto pump the heat away, all that was available were back-up sys-tems powered by steam from the reactors themselves.

In part because of human error, they failed. The fuel in thereactors’ cores got hot enough to melt. The cladding on the fuelrods reacted with steam to produce hydrogen. Systems thatshould have �ushed the potentially explosive hydrogen out ofthe containment vessels around the reactors also failed, so thegas started to accumulate in the buildings housing the reactors.One after the other, three of the buildings blew up, releasing ra-

dioactive material and contaminating an area that in some direc-tions went well beyond the 20-kilometre evacuation zone.

The reactors at Fukushima were of an old design. The risksthey faced had not been well analysed. The operating companywas poorly regulated and did not know what was going on. Theoperators made mistakes. The representatives of the safety in-spectorate �ed. Some of the equipment failed. The establish-ment repeatedly played down the risks and suppressed informa-tion about the movement of the radioactive plume, so somepeople were evacuated from more lightly to more heavily con-taminated places.

The outcome could quite easily have been even worse. Fu-kushima had a lot of used fuel in spent-fuel ponds, which keep itcool and absorb its radiation. The explosion in building 4, whichhad no fuel in its reactor but whose spent-fuel pool was full, ledsome to believe that the water in the pool had drained away andthe spent fuel was melting, though in fact the explosion seems tohave been caused by hydrogen from building 3. A worst-caseanalysis by Japan’s Atomic Energy Commission, not publishedat the time, suggested that if the hot fuel was indeed left high anddry, people as far away as parts of Tokyo would need to be evacu-ated and everyone in the capital would have to stay indoors. Andthere are a number of big cities�London, New York, Hong Kong,Los Angeles�that are closer to ageing nuclear plants than Tokyois to Fukushima.

Just in case

The governing principle of nuclear safety is �defence indepth�. Seek �rst to prevent failure, then to correct failures notprevented, then to control the consequences of failure, then todeal with emergencies beyond normal control. The mighty wallat Hamaoka provides a good example. It should, in itself, protectthe plant from the worst that the sea can throw at it. But if thewall should fail, new pumps installed in watertight buildingswill make sure the plant can still dump heat in the sea. The threereactor units are being reinforced to keep the sea from ever reach-ing the locomotive-sized diesel generators on their ground�oors. More diesel generators are mounted on the roofs, wellabove any conceivable tsunami; on the blu� behind the reactorbuildings there is a gas turbine to provide further back-up, not tomention a �eet of small trucks with their own little pumps, readyto go where they are needed.

This is not the �rst revamping of Hamaoka. The plant sitson top of the Nankai Trough, where two tectonic plates grind to-gether. This action can produce powerful earthquakes such asthat of March 11th last year or the one that caused the Boxing Daytsunami in 2004. In the past decade Chubu has revamped theHamaoka plant to make sure that it could ride out a big earth-quake, reinforcing various parts of the plant and closing the twooldest reactors. On paper, at least, there is no plant in the worldmore earthquake-proof than Hamaoka.

There are, though, plants with greater claims to a safe de-sign overall. Two of the reactors at Hamaoka, commissioned in1987 and 1993, are what is known as �generation II� designs, dat-ing from the period after the industry settled on certain stan-dards in the late 1960s but before Chernobyl. The third, switchedon in 1999, is one of Toshiba’s Advanced Boiling Water Reactors,the �rst of the post-Chernobyl �generation III� designs. Model-ling based on experience with previous plants suggests that therisk of a signi�cant radiation leak from generation I reactors wasbetween one in 1,000 and one in 10,000 per reactor year. For gen-eration II it is between one in 10,000 and one in 1m. For genera-tion III it should be between one in 1m and one in 100m. Thesecalculations do not re�ect the absolute risks as experienced inthe real world; there have been �ve major releases of radioactiv-

Safety

Blow-ups happen

Nuclear plants can be kept safe only by constantly

worrying about their dangers

E U R A S I A N

P L A T E

N O R T HA M E R I C A N

P L A T E

P A C I F I C

P L A T E

P H I L I P P I N ES E A P L A T E

Tokyo

SOUTHKOREA J A P A N

Osaka

NORTHKOREA

INSET

Sendai1 2

Shimane1 2

Genkai1 2 3 4

KashiwazakiKariwa

1 2 3 4 5 6

1 2 3 4 5 6

7

Takahama1 2 3 4

Ohi1 2 3 4

Mihama1 2 3

Tomari1 2 3

Ikata1 2 3

Hamaoka3 4 5

Fukushima Daini1 2 3

Onagawa1 2 3

Tokai2

4

Fukushima Dai-ichi

1Tohoku/Higashidori

Tsuruga1 2

Shika1 2

Sources: Japanese government; JAIF

Nuclear power plantsin Japan, Feb 2012

In operation (2)Being inspected (35)Shut down (17)

400 km

Tokyo

K A N T O

TOHOKU

Fukushima

Iitate FukushimaDai-ichi

FukushimaDaini

TokaiNuclear plant

Area in need ofdecontamination

quakes and �ooding. A report fromAmerica’s Nuclear Regulatory Commis-sion re�ects similar concerns. As well assuggesting new rules, it calls for a new co-herence in the regulatory �patchwork�that has grown up to deal with highly un-likely events.

Never be satis�ed

The need to keep questioningthings�from the details of maintenanceprocedures to one’s sense of the worst thatcould go wrong�is at the heart of a suc-cessful safety culture. Mr Jamet gives theexample of a worker noticing that a dieselgenerator has been switched o�. It is notenough to switch it back on. You also haveto ask how and why it got switched o�,and what other consequences that may

have had. When you have got to the root of it, you not only haveto change procedure but also to make sure that all other similarplants know about the problem and how to solve it.

It was to help with this kind of e�ort that, after Three MileIsland, the American nuclear industry set up its Institute of Nuc-lear Power Operations. INPO, headquartered in Atlanta, regular-ly inspects power plants, using its own sta� as well as engineersfrom other operators, and o�ers lessons learned from mistakesthroughout the industry so that all plants bene�t from what hap-pens at any one of them. INPO brings fresh eyes and high stan-dards, and its reports can be scathing. Phil Sharp, a former con-

gressman who now heads a think-tank,Resources for the Future, and sits on theboard of a power company, Duke Energy,says that INPO meetings at which thebosses of nuclear operating companiesare called to account for their plants’ fail-ings in front of their peers are unlike any-thing he knows of in the private sector.

But despite the attentions of the NRC

and INPO, things can still go wrong atAmerican plants. In the late 1990s Davis-Besse, a nuclear plant in Ohio, had lessdowntime than almost any other plant inAmerica. Various signs of incipient trou-ble�air �lters clogging up too frequently,borate salts of unusual consistency andrusty colour building up�were not fullyinvestigated. When the reactor wasopened for maintenance in 2002, it wasdiscovered that boric acid had eaten ahead-sized hole pretty much all the waythrough the top of the reactor vessel. Onlya thin layer of stainless steel was holdingthings together. As the plant’s operator,FirstEnergy, later reported, �there was a fo-

cus on production, established by management, combined withtaking minimum actions to meet regulatory requirements, thatresulted in the acceptance of degraded conditions.� Disaster wasonly narrowly avoided.

George Felgate, a veteran of Rickover’s nuclear navy whospent almost three decades at INPO before joining the World As-sociation of Nuclear Operators, points to a seeming paradox atthe heart of nuclear safety: if, having made every provision forsafety, you think for a minute that an accident is not possible, youput yourself at risk of being proved disastrously wrong. This

ity (Three Mile Island, Chernobyl and the three reactors at Fu-kushima) in only 14,000 reactor years of operation. But the trendtowards safety seems to be real.

Two generation III designs thought to be particularly ad-vanced are currently under construction. The AP1000 from Wes-tinghouse (now owned by Toshiba) is being built in China andAmerica, and the EPR from Areva, a French company resultingfrom a merger between French and German nuclear-plant-builders, in China, Finland and France. The EPR is the biggestplant ever designed and has safety systems galore, which meansmore pipes, more wiring, more concrete and higher capital costs.The AP1000 aims instead for simplicity, with fewer valves,pumps and wires and a greater reliance on �passive� safety sys-tems that use basic physics to provide emergency cooling andother safety functions. A French insider, unwilling to be named,considers the AP1000 to be a far more creative piece of engineer-ing. The Union of Concerned Scientists, an American gingergroup critical of most real-world nuclear programmes (thoughnot of the technology per se), prefers the safety concept of theEPR, with its multiple back-ups.

Proponents of generation III reactors, which is to say prettymuch the entire nuclear establishment, think�with some rea-son�that they would have fared much better at Fukushima. In-deed some feel that the circumstances at Fukushima�a freakish-ly large wave, an old set of reactors with insu�cient safetyequipment and a poor operator, poorly regulated�limit its rele-vance elsewhere. Those operating reactors with lower risks of�ooding or earthquake, better emergency cooling systems and

more robust power supplies might see themselves as having lit-tle to learn from it.

Philippe Jamet, of France’s nuclear regulator, the ASN, in-sists on a broader view; that Fukushima demonstrates a shortfallin imagination, not just in Japanese regulators but also in peoplelike himself. �If you had asked me a year ago about an accident inwhich multiple units were left without power and cooling,� hesays, �I would have said it was not credible.� The ASN has intro-duced new requirements for nuclear plants based on the Fu-kushima accident that go beyond safeguards against earth-

8 The Economist March 10th 2012

NUCLEAR ENERG Y

SPECIAL REPORT

2

1

When the reactor was opened for maintenance in 2002,it was discovered that boric acid had eaten a head-sizedhole pretty much all the way through its top

Hamaoka defended in depth

The Economist March 10th 2012 9

SPECIAL REPORTNUCLEAR ENERG Y

2

1

stress on constant vigilance means that nuclear safety can neverbe a technological given, only an operational achievement.

In many places, and particularly in Japan, the industry hasfelt a need to tell the public that nuclear power is safe in some ab-solute way. This belief is clearly no longer sustainable. The onlyplausible replacement is to move from saying �it is safe� to say-ing �trust us to make it as safe as it can be,� and accepting that insome situations and some communities that trust will not al-ways be given.

Japan’s government is trying to restore the trust its peopleare now unwilling to give. It is moving nuclear regulation fromthe industry ministry, where civil servants were devoted tobuilding up the industry, to the environment ministry. But the re-sponse to Fukushima has, so far, been inadequate. Many ques-

tion marks remain. One of the more worrying is how muchdamage the earthquake did to the reactors. It is claimed that theyweathered the quake, but some experts, such as Masashi Goto, aretired nuclear engineer, argue that there is evidence of signi�-cant damage that speeded up the subsequent meltdown. Analy-sis of the spread of fallout suggests that the �rst releases camevery soon after the tsunami hit, if not before. With quakes a moreconstant threat than monster tsunamis, these are the sort of les-sons that Japan’s �nuclear village� needs to learn.

If the Japanese nuclear establishment�industry and regu-lators alike�wants to earn trust, it must be seen to be learning ev-ery lesson it can. It must admit how little it previously deservedtrust and explain clearly how it will do better in future. Eventhen, such trust will not always be given.

OF ALL THE di�culties nuclear power is heirto, that of waste has most �red the publicimagination. Building power plants that lasta century is one thing; creating waste thatwill be dangerous for 100 times as long isanother. For decades America has failed tocreate a long-term repository for the wastefrom its civilian reactors at its chosen site,Yucca Mountain in Nevada. Most othercountries have similarly failed, so the wastefrom today’s reactors piles up.

As it happens, long-term waste dis-posal is among the more tractable nuclearproblems. Temporary storage is a good start.Once fuel has cooled down in spent-fuelpools for a while, it can be moved to �drycask� storage. Such storage appears robust(dry casks at Fukushima, hit by the tsunami,show no sign of having leaked) and can bemaintained inde�nitely. It takes space andneeds to be guarded, but it can provide anadequate solution for a century or more.

That is if you do not want to reprocessthe fuel to recover the plutonium inside it. Ifyou are a nuclear engineer you may �ndreprocessing rather appealing, partly toshow that your nuclear programme is assophisticated as any and partly because itgets around the o�ensive ine�ciency oflight-water reactors. If all the uranium inreactor fuel was either split or turned intoplutonium which itself was then split, youwould get 170 times more energy than youget from just using the fuel once, and wouldhave opened the way to technically intrigu-ing breeder systems.

You will not, though, be attracted toreprocessing if you are an accountant. Itcosts a great deal, and the plutonium pro-duced is for the most part more of a liabilitythan an asset. If you are a plant operator youwill also have your doubts. Burning fuel towhich plutonium has been added has various

drawbacks, one of which is that it is muchhotter when it comes out of the reactor,straining the capacity of your spent-fuelpools. Nor will you be that eager if you areconcerned about the local environment;reprocessing plants have a bad contamina-tion record. And if you are sceptical aboutthe merits of nuclear proliferation, you willwant to keep reprocessing to a minimum.

Having avoided reprocessing, in thelong term you will want to �nd a safe deepunderground repository for the waste inyour dry casks. This need not be too hard.Find a community that may be willing to takeon the challenge (one that already has tiesto the nuclear industry might be thus pre-disposed) and that has access to a suitable

Leave well alone

The best thing to do with nuclear waste is to stash it away, not reprocess it

geological setting. Then have an opendiscussion of the issues, look at people’sconcerns and o�er ways to lessen themwhile recompensing the community for itstrouble. Set up arrangements by whichlocal people can reassure themselvesabout any threats to their health, perhapswith free medical treatment and tests.Don’t scrimp on investment in the com-munity. Then let them choose. Thechances are that they will say yes. Thiskind of approach seems to be working inSweden and Finland, and Britain is tryingsomething similar.

This is more or less the opposite ofwhat was done at Yucca Mountain. Whatsome now refer to as the Screw Nevada actof 1987 imposed the choice of site andschedule. Nevada politicians objected.Geological surveys threw up some pro-blems. Nevada’s caucuses moved up theelectoral calendar, meaning that presi-dential candidates were greatly helped byan anti-Yucca stance. Moreover, thestate’s senior senator, Harry Reid, becameSenate majority leader. President Obamadrew a line under the episode by �nallyabandoning the project in 2010, 12 yearsafter the facility was meant to have en-tered service.

But as well as providing a textbookexample of how not to handle long-termstorage, America also boasts a success.The Waste Isolation Pilot Plant in the saltcaverns of Carlsbad, New Mexico, startedtaking shipments of waste from the coun-try’s military programme in 1999.Throughout the life of the project the localcommunity has been consulted and, onoccasion, recompensed. WIPP is not en-tirely trouble-free, but it has achievedenough social and political stability tomake the best of its geological gifts.

In a cavern, in a canyon

10 The Economist March 10th 2012

NUCLEAR ENERG Y

SPECIAL REPORT

2

1

IN HAIYANG, ON the northern Chinese coast, and at San-men, farther south, an international consortium led by

Westinghouse is well into building two AP1000s, with two morein the works; China plans eventually to have 12 split between thetwo sites. If the plans go ahead, each site will have as much ca-pacity connected to the grid as the whole of Nigeria has today.Yet the two plants represent only a small fraction of China’s nuc-lear ambitions. Its pre-Fukushima plans to increase its nuclearcapacity from 10GW to 80GW by 2020 may fall behind sched-ule, but China still looks certain to build more new nuclearplants than any other country over the decade to come�andpossibly more than all others combined.

By nuclear standards, this is a big deal; China will add morenuclear capacity in those ten years than France has in total. Butfor China itself it is less big; nuclear will go from generating lessthan 2% of the country’s electricity to less than 5%. Ming Sung,who works for the Clean Air Task Force, an American think-tankin Beijing, points out that China is not betting on nuclear; it is bet-ting on everything that o�ers an alternative to coal. China con-sumes half the world’s annual coal output, and has the supplyproblems, dirty air and huge death toll (hundreds of thousands ayear from respiratory diseases) that go with it. Junda Lin of theChina Greentech Initiative points out that the 2020 target for nu-clear has to be seen in the context of a 200GW target for windand an extra 100GW of hydropower. The idea is to try everythingand see what works best.

Most of the plants China is currently building are genera-tion IIs derived from a French design it bought in the 1980s andnow built by Chinese companies, but there are also RussianPWRs in Tianwan and Canadian Candus in Qinshan. In Taishantwo EPRs are being built by Areva and the China Guangdong Nu-clear Group, which has a long-standing relationship with theFrench industrial base from which its domestic designs ultimate-ly derive. And then there are the AP1000s. Westinghouse wonthat contract in large part by promising to transfer the technologyin full to local companies, but it hopes that its expertise will al-low it to keep a prominent role in the Chinese industry.

After Fukushima the state council stopped approving newpower stations and called for re-evaluations of the seismic and

�ooding risks faced by those already built and under construc-tion. A new law expected later this year will take nuclear regula-tion away from the National Development and Reform Commis-sion, the state’s industrial planners, and hand it over to theenvironment ministry, thus splitting the role of cheerleadingfrom that of invigilation. Part of what passes for the Chinese gov-ernment’s legitimacy comes from the perception that it can man-age large-scale technology well. The backlash against China’shigh-speed train programme after last year’s accident at Wen-zhou, which provoked criticism and anger of a sort that Chineseleaders fear, would be dwarfed by what could be expected froma nuclear accident.

A sincerely self-interested desire to avoid accidents,though, will not necessarily translate into a model regulatory in-frastructure. A safety culture of constant questioning will not beeasy to instil. And China’s nuclear regulatory workforce is al-ready more stretched than that of other big economies in termsof employees per gigawatt under regulation.

Beside the seaside

Another new law will outline future plans for the industry.Some expect China’s nuclear boom to slow down in the wake ofFukushima, with new capacity perhaps reaching only 40GW by2020. And China could get proportionally more AP1000s andfewer of its own home-made designs, the safety of which maybe less assured. All China’s current plants are by the sea, both be-cause it is convenient for cooling and because that is where thedemand is. There have been plans for nuclear plants inland,cooled by rivers, but concerns about the availability of water indrier years to come and the risk of contaminating it may causethese plans to be shelved.

China’s expansion into nuclear power is hardly a market-driven development, but it helps that the plants involved lookcomparatively cheap. There are two ways of measuring the costof a nuclear power plant: the �overnight� cost, which counts upthe material and labour that goes into a new plant as if it had allbeen purchased simultaneously, and the �levellised� cost, whichis a measure of the total amount of energy a plant provides overits life divided by the total expenditure�construction, operation,maintenance, fuel and, eventually, decommissioning. One is thecost of the capacity to produce electricity, the other the cost ofthe electricity produced. The 2010 edition of the IEA/NEA Costsof Generating Electricity study puts overnight costs for Chinesegeneration II plants at $1,700 for every kilowatt of capacity, giv-ing a gigawatt plant a price tag of less than $2 billion. For theAP1000s the estimated costs are higher ($2,300/kW), and by the

Costs

Bandwagons and busts

Nuclear plants are getting ever more expensive. But

Asian countries may build them more cheaply

Exploding

Source: “The Economics of NuclearReactors: Renaissance or Relapse?”,Mark Cooper (updated)

*Construction cost if all material and labourwere purchased simultaneously †By a range

of analysts, utilities and academics

Overnight costs* of US nuclear power plants

2

Year of operation or estimate

Cost

, $20

09/K

W

0

2,000

4,000

6,000

8,000

10,000

1970 75 80 85 90 95 2000 05 10

E S T I M A T E S †

That is why Hamaoka, for all the tsunami protection andearthquake-proo�ng it has undertaken, is unlikely to reopen. It istoo close to Tokyo, and too close to the expected epicentre of avery big earthquake that might happen one day, for people everto think it truly safe. Oddly, this may be the world’s only nuclearpower station that could bene�t from a quake. Only if the bigone comes and goes, and Hamaoka rides it out unscathed, mightit be able to build some trust.

Meanwhile, its engineers are trying to work out why, whenthe plant was being shut down last summer, a pipe in a heat ex-changer in the most advanced of the reactors burst in a peculiarway, damaging other plumbing so that a few tonnes of seawatergot into the reactor proper. The damage it may have done has yetto be assessed. It was something that nobody had expected. 7

The Economist March 10th 2012 11

SPECIAL REPORTNUCLEAR ENERG Y

2

1

time the projects are �nished they may be higher still; these arethe �rst AP1000s being built anywhere, so its wise to expect sur-prises. Schedules are being stretched, and the Chinese contrac-tors for key parts of the third and fourth AP1000s are falling be-hind a bit, according to Westinghouse.

Still, almost anywhere else in the world, these �gureswould today be a source of envy�or incredulity. When compa-nies were beginning to pitch generation III reactors ten years ago,they claimed that better, standardised designs and improvedconstruction techniques would make them both safer andcheaper. In Western countries that second claim has gone by theboard. British studies in 2004, 2006 and 2008 put the overnightcost of new PWRs at $2,233/kW, then$2,644, then $3,000. Estimates from theMassachusetts Institute of Technology(MIT) rose from $2,208/kW to $4,000 overroughly the same time. The NEA quotescosts for an EPR in Belgium (now can-celled) at $5,400 per kW. Capacity �red bygas turbines, for comparison, can cost lessthan a �fth of that.

Real construction costs, which in-clude the cost of borrowing the moneyneeded, are even higher than overnightcosts. Construction costs for the twoAP1000s that Progress Energy hasplanned for its Levy site in Florida have re-cently been reported at about $20 billion,which works out at about $9,000 per kW

and strongly suggests that the reactors willnot be built.

Cost escalation has been the rulethroughout the industry’s history. In thelate 1960s what is now called the �greatbandwagon market� took o� in America.Companies selling plants they had no realexperience of building o�ered �xed pricesto make them attractive. Utilities keen toreduce their reliance on coal in an age ofclean-air standards took the bait. As orders�ooded in, costs started to climb. Projectsmeant to be completed in years draggedon for more than a decade, in part because of new environmen-tal concerns, in part because designs were revised as lessonswere learned. At the Vogtle plant, in Georgia, a pair of reactorsoriginally priced at $660m in 1971came in at $8.87 billion 16 yearslater. Half the projects ended up cancelled.

The French experience is often quoted as a positive coun-ter-history to the American mess. France had long been keen onenergy security. When it made PWRs based on a Westinghousedesign a national priority in the early 1970s, it brought a thor-ough-minded discipline to the matter, building its capacity re-gion by region, improving the designs as it went along and in-creasing the size of its plants to reap economies of scale. Havingthe same contractor and customer for so many plants allowedthe system to learn from mistakes and to re�t older plants tonewer standards. Even so, according to calculations by ArnulfGrübler of IIASA, a think-tank near Vienna, each of the six de-signs France has �elded has cost more per kilowatt than the pre-vious one had. He estimates that the four reactors built in the1990s cost between $2,267 and $3,252/kW in 2010 dollars, morethan twice the real cost of capacity built in the 1970s and early1980s. The �rst two EPRs to be built in Europe, in France and Fin-land, have both gone extravagantly over schedule and budget.

A decade ago the nuclear industry hoped that the combina-

tion of safe, low-cost generation III reactors and governments ea-ger to encourage lower carbon-dioxide emissions would lead toa �nuclear renaissance�. In the West those low costs have failedto materialise, so the renaissance is largely stalled. Whereas afew years ago Britain was talking of building eight new reactorsto replace its ageing �eet, only two are likely to make it in the nearterm. Steve Thomas, an economist at the University of Green-wich, argues that even with a �xed carbon price of �36/tonneand a guaranteed price for the electricity (both features of a cur-rently planned re-regulation of Britain’s energy market; today’sEU carbon price is under �10), those plans remain vulnerable.

In a capital-intensive industry such as the nuclear one, the

cost of capital is always crucial, and higher overnight costs mag-nify the problem. Calculations of the levellised costs of energyby UBS, a bank, show clearly that the cost of capital dominatesthe picture. For a plant costing $5,500 per kW, capital makes up75% of total costs in Europe and America. UBS reckons the level-lised cost of such a plant in Europe is 11% higher than the cost of agas plant. It would take a quintupling of the carbon price to wipeout that di�erential. And those calculations assume that it is aseasy to borrow to �nance a nuclear plant, with all its uncertain-ties and regulatory risk, as it is to �nance a gas plant, which isprobably unrealistic.

Step on the gas

In eastern Europe, where Russian dominance of gas mar-kets is a political issue and electricity markets are still quite regu-lated, governments may consider such a di�erential acceptable.The Czech Republic is about to tender for new generation IIIPWRs, and Poland has plans along those lines too. But in Ameri-ca things look very di�erent. Asked if Fukushima put America’snuclear renaissance on ice, Ernest Moniz of MIT replies succinct-ly: �No. Shale gas did.� For all the production incentives, loanguarantees and indemnity for costs due to regulatory change of-fered by government, the sharp drop in gas prices caused by new

A clear argument for nuclear power

12 The Economist March 10th 2012

NUCLEAR ENERG Y

SPECIAL REPORT

2

1

�THERE IS ONLY one reason for America to subsidise nuc-lear,� says Ernest Moniz, of MIT, �and that is the climate.�

He has a point. By 2020, carbon emissions since the start of the21st century will have surpassed those of the entire 20th. There isa real risk that emissions on such a scale will bring disaster to hu-mans, or to the natural world, or both. Nuclear power, whichproduces no direct carbon-dioxide emissions, should be able tomake things better.

Robert Socolow, of Princeton University, and his colleaguescalculate that if the world were to replace 700GW of coal-�redplant with nuclear reactors over 50 years�which would more orless triple its current nuclear capacity�it could reduce its annualemissions of carbon dioxide by 3.7 billion tonnes. Allowing forthe need to replace most of the current �eet over the same per-iod, that would mean deploying nuclear reactors at three timesthe speed of China’s planned record-breaking deployment be-tween now and 2020, and doing it for �ve decades straight. Buteven that would make only a minor dent in the problem. In 2010the amount of carbon dioxide emitted by industry was about 30billion tonnes, and was growing at 3% a year. At that rate, the sav-ings from such a beefed-up nuclear-power programme wouldcompensate for just four years of emissions growth.

The prospects

Over the rainbow

If there are better ways to split atoms, they will be a

long time coming

3Spot the potentialWorld energy consumption*, million tonnes of oil equivalent

Source: BP Statistical Reviewof World Energy, 2011

*Based on gross generation and not accounting for cross-border electricity supply.Converted on the basis of thermal equivalence assuming 38%

conversion efficiency in a modern thermal power station.†Wind, geothermal, solar, biomass and waste

0

100

200

300

400

500

600

700

2000 01 02 03 04 05 06 07 08 09 10

Renewable† Nuclear

sources of supply ruled out new nuclear plants in any marketwhere the two energy sources compete freely. According to UBS,the advantage of gas over nuclear in America is roughly twicewhat it is in Europe.

John Rowe, CEO of Exelon, an energy company that has tennuclear power plants in its portfolio, says that companies like hisno longer have any reason to build nuclear plants. All plans tobuild nuclear plants in parts of America where the electricitymarket has been deregulated are coming to naught. Some Ameri-can plants will still be built, but only in the south-east, where reg-ulators allow the cost of increasing a utility’s asset base to bepassed on directly�indeed pre-emptively�to its captive custom-ers. Thus electricity consumers in Georgia are already paying fortwo new AP1000s which in February got clearance from the NRC

to complement the two reactors at Vogtle. In Sumner, South Car-olina, two more AP1000s are under contract. Those four willprobably be all the renaissance America sees for some time.

If the West could build new reactors as cheaply as Chinacan, things would look di�erent. That it cannot is in part due tolabour costs. But the Chinese must have other advantages too.The levellised costs of modern Chinese coal-�red power stationsare lower than the competition’s even when the power stationsare not built in China. The same is true for cement works; Chi-nese companies operating outside China cannot build them ascheaply as they do at home, but they still easily beat the interna-tional competition.

Further evidence that a di�erent industrial approach cancut costs comes from South Korea. Like Japan, the country has lit-tle by way of indigenous energy supplies, and it too decided onnuclear power to solve that problem and bring new technologi-cal skills to its industrial base. It gets some 30% of its electricityfrom nuclear plants, much the same as Japan did before Fuku-shima, and more than any large economy other than France. In2010 KEPCO, the South Korean power company, sold its reactorsoverseas for the �rst time, beating the French to a contract in theUnited Arab Emirates; at home its overnight costs for such gener-ation II reactors are calculated at just over $1,500/kW.

The true costs in South Korean business can be hard tomake out. It would not be at all surprising if, working abroad forthe �rst time and having been keenly competitive in its bidding,KEPCO failed to deliver the UAE reactors on budget. And giventhat nuclear prices have gone up everywhere else, it is fair to ex-pect that they will do so to some extent in Asia, too. But if Chinaand South Korea can build reactors abroad at prices not muchhigher than those at home, nuclear may see its fortunes tick upelsewhere, argues David Victor, of the University of California.Both Westinghouse and EDF have plans for new reactors in theexport market that would be designed and sold in collaborationwith Chinese partners. Russia is keen to export PWRs too, but itscosts are not clear.

Inviting the Chinese to come in and build a nuclear plant isan unlikely step for a Western government (though the South Ko-reans are bidding on a Finnish contract). Some developing coun-tries, though, may be interested. This is a matter of concern forbackers of the American nuclear industry with an eye to nation-al security issues, such as Pete Domenici, a former senator. IfAmerica is not engaged in the market, how can it use its in�uenceto deter proliferation?

And it will indeed have less scope for such in�uence. Buteven at Chinese prices, nuclear energy is expensive comparedwith coal, and if other countries gain easier access to gas, asAmerica did, that will reduce demand too. Vietnam is enthusias-tic about nuclear reactors; other Asian countries, especiallythose in tectonically active places�such as the Philippines andIndonesia�may be less keen than they were before the great tsu-

nami. South Africa is talking of buying nuclear reactors. Indiahas big plans on paper, but a law that makes the designers (ratherthan the operators) of power stations liable in case of accidentsgets in the way (and buying Chinese reactors might be anathemaanyway). There is interest in the Middle East, but as Charles Eb-inger of the Brookings Institution, a think-tank, points out, thecountries talking about buying nuclear power in response torunaway electricity demand might do better to curb their hand-some consumer subsidies. They might also do well to invest inalternative energy. The sun’s nuclear reactor has been going for4.5 billion years, and extracting power from it is getting cheaperevery year. 7

The Economist March 10th 2012 13

SPECIAL REPORTNUCLEAR ENERG Y

2

1

No technology can solve the climate problem on its own.Even in combination, today’s remedies�renewables, nuclearand energy e�ciency�hardly seem up to the job. To have a rea-sonable chance of keeping down the rise in temperature to lessthan 2°C, industrial economies need to reduce emissions by 80%by 2050. The true scale of this challenge is not widely under-stood. A thorough study of options for such cuts in California,long a leader in energy e�ciency, concluded that with today’stechnology and plausible extrapolations of it, 60% was the bestthat could be done. If California can’t do better than that, saysJane Long, of Lawrence Livermore National Laboratory, who ledthe study, �neither can anyone else�.

Even getting close to such goals, though, is easier with moretechnologies than fewer. Even if nuclear can make only a smallcontribution, it could be worth having. The IEA’s 2011 World En-ergy Outlook calculates that, between now and 2035, an emis-sions path that keeps the 2°C limit plausible would cost $1.5 tril-lion more if OECD countries were to stop building nuclear plantsand other countries halved their nuclear ambition, largely be-cause much more would have to be spent on renewables.

Germany, long keen on renewables and squeamish aboutnuclear, provides an example. Its decision after Fukushima tophase out nuclear power entirely will mean that most of the lostcapacity will need to be made up with even more renewables,though it will also build new fossil-fuel plants and import elec-tricity from nuclear plants in France. As the new fossil-fuel plantswill probably run on gas, emitting less carbon dioxide than docoal plants which are also due for retirement, this may keep thecarbon in check. But electricity prices for industrial customers,according to an analysis by UBS, will rise by more than 60% inreal terms by 2020. Ottmar Edenhofer of Berlin’s Technical Uni-versity says this is a pretty middle-of-the-road projection.

The market, too, will probably need some re-engineering.Systems with a lot of renewables make life hard for fossil-fuelgenerators, which have to shut down when it is sunny andwindy and take up the slack when it is not. To get the fossil-fuelinvestment it needs, Germany may well have to pay for the ca-pacity built even if it stands idle, or guarantee rates of return.

But though renewables are expensive, so is building newnuclear plants; the bills Britain will face as it tries to meet carbongoals with new nuclear should keep any Schadenfreude in check.The cheap new supply-side route to lower carbon-dioxide emis-sions is to replace old coal-�red stations with new gas-�red ones,which emit half as much carbon dioxide per megawatt hour.Plumping for renewables or nuclear will cost a lot more.

Still, renewables are getting cheaper, through technologicalchange and through the bene�ts of mass production and marketcompetition. In the long run, technologies that get cheaper canbe expected to edge out a technology that has only ever got moreexpensive. In a low-emissions world, the role for nuclear will belimited to whatever level of electricity demand remains whenrenewables are deployed as far as possible.

That is a large enough role for some Greens to have becomenuclear converts. For the most part, though, they are thinkingabout a nuclear programme more exciting than the slow, expen-sive and only marginally helpful deployment of PWRs. There

are many alternative reactor designs, and each has its champi-ons. An international body called the Generation IV Internation-al Forum (GIF), co-ordinated by the NEA, is drawing up plans forprototypes using such ideas, all claiming to o�er improvementsover the current crop.

Yet the problems these new reactors solve are for the mostpart those that the industry wishes it had, rather than those it ac-tually faces. The GIF designs, and others, are mostly �fast� reac-tors that use highly �ssile fuel and unmoderated neutrons; theycan both burn plutonium and create it in copious amounts. If �s-sile material were in short supply that might be an advantage.But uranium is not currently in short supply, and it makes uponly a small part of nuclear energy’s costs. The ability to makenew nuclear fuel solves a problem that reactors will run intoonly if their use becomes massively more widespread. Whatnew reactors need is an advantage that will make them popularin the �rst place.

Indeed, at present the ability tomake plutonium is a disadvantage. Dis-suading countries with nuclear pro-grammes, or that want nuclear pro-grammes, from reprocessing their fuel toproduce plutonium is one of the core pri-orities in anti-proliferation work (the oth-er is trying to keep newly nuclear coun-tries from developing their own

enrichment systems). If established nuclear powers were to stopreprocessing (as Britain is doing), it might help to persuade oth-ers, such as South Korea, that it is better not to start. A new gener-ation of plutonium breeders would completely undermine thate�ort.

Admittedly, other kinds of breeders are available. Molten-salt reactors, which keep their fuel in liquid form, could be usedto turn thorium, of which the world has an abundant supply,into a type of �ssile uranium not found in nature, U-233. Thiswould be rather unsuitable for bomb-making and gets round thecontinuing use of U-235 or plutonium, so thorium molten-salt re-actors o�er the possibility of breeding fuel in a way that does notfacilitate proliferation.

Like some of the other GIF designs, molten-salt reactorsalso have novel safety features; but although safety is a conditionof getting into the game, it is hardly a means of winning it. If gen-eration III reactors, well operated, prove safe, why upgrade? Ifthey are not safe, who would trust generation IV? The way to

Long waves

Sources: “From Energy Innovation to Energy Transformation”, Steven Koonin and Avi Gopstein; EIA

US energy supply, %

4

0

20

40

60

80

100

1850 60 70 80 90 1900 10 20 30 40 50 60 70 80 90 2000 08

Coal

Gas

Oil

Nuclear

Hydroelectric

Wood Otherrenewables

In a low-emissions world, the role for nuclear will belimited to whatever level of electricity demand remainswhen renewables are deployed as far as possible

On the other hand a prac-tical reactor can be distin-guished by the followingcharacteristics: (1) It is being builtnow. (2) It is behind schedule. (3)It requires an immense amountof development on apparentlytrivial items. (4) It is very expen-sive. (5) It takes a long time tobuild because of its engineeringdevelopment problems. (6) It islarge. (7) It is heavy. (8) It is com-plicated.

Innovators need to be ableto take risks, to try variations ontheir ideas and to be able tolearn. They �ourish in unregu-lated markets. They frequentlydepend on venture-capitalfunds which are dwarfed by thecost of even a single utility-scalepower plant. They also need re-wards. Yet makers of nuclear re-actors cannot take risks thatmight compromise safety, andthey cannot try lots of di�erentthings because it would be tooexpensive. And even if they suc-ceed, all they will be making iscommoditised electricity. Pow-er stations are not conducive toradical innovation.

Nuclear reactors, as Philippe Jamet notes, last for centuries;the technology is, by its own standards, still young. Longevityand inertia ensure that even a disaster like Fukushima cannotwipe it from the world. But they also ensure that it cannot growfast. In energy in general, technologies mature and succeed eachother over decades. Nuclear seems likely to lag behind even inthis slow �eld. That does not mean it will not, eventually, play alarger role, but that it will get there slowly. Inside a reactor, thingscan change in milliseconds. Outside, it takes lifetimes. 7

SPECIAL REPORT

14 The Economist March 10th 2012

2

NUCLEAR ENERG Y

Offer to readers Reprints of this special report are available. A minimum order of five copies is required.Please contact: Jill Kaletha at Foster Printing Tel +00(1) 219 879 9144e-mail: jillk@fosterprinting.com

Corporate offerCorporate orders of 100 copies or more are available. We also offer a customisation service. Please contact us to discuss your requirements.Tel +44 (0)20 7576 8148e-mail: rights@economist.com

For more information on how to order special reports, reprints or any copyright queries you may have, please contact:

The Rights and Syndication Department26 Red Lion SquareLondon WC1R 4HQTel +44 (0)20 7576 8148Fax +44 (0)20 7576 8492e-mail: rights@economist.comwww.economist.com/rights

Future special reports

Previous special reports and a list of forthcoming ones can be found online: economist.com/specialreports

win will be on price.At the moment, those who want to bring down the cost of

nuclear power are not, for the most part, looking at big genera-tion IV reactors that will not be built for 20 years, if ever. Instead,they are thinking small. Particularly in America, small modularreactors (SMRs) of up to 300MW are all the rage. Some, such asthe 100MW mPower reactor o�ered by Babcock and Wilcox, arescaled-down PWRs. Others are more exotic.

Think small

Such reactors can reach markets which today’s big reactorscannot. Many utilities�and smaller countries�have little inter-est in gigawatt-scale plants: they prefer to build around 100MW

of capacity at a time. Small reactors might also open up new ap-plications, perhaps in desalination, district heating or even tran-sport. He Zuoxiu, a Chinese physicist critical of his government’srush to build lots of big PWRs, has suggested that SMRs for ships,both military and merchant, would be a good way to train up acadre of engineers and designers.

SMRs can also be slotted into underground silos, whichcuts down on civil engineering costs. Perhaps most promising ofall, they would be built in factories, not on site. That should makethem less subject to delay than manufacture in the �eld. And afactory building ten such reactors a year for years on end mightbe able to make signi�cant cost reductions through incrementalimprovements�economies of number as opposed to econo-mies of scale.

But these advantages do not add up to a conclusive case fora small modular future. Babcock and Wilcox claim overnightcosts per kilowatt of capacity for the mPower roughly on a parwith those of big PWRs like the AP1000. But in an industry thathas long pursued economies of scale, many are unconvincedthat smaller reactors can deliver the same costs per kilowatt.Atam Rao, who led the design of an advanced generation III re-actor, GE’s ESBWR, describes such claims as �complete BS�.Things like control systems are needed for all reactors, big orsmall. Providing them for each small reactor is bound to push upcosts. Will any utility really think it makes sense to �eld ten SMRswith ten control systems and ten safety systems rather than onebig PWR? Only if it has seen it done economically elsewhere.

In the end, that is the biggest pro-blem for proponents of new approachesto nuclear energy. If a radically new tech-nology, as opposed to an incremental one,is to take o�, it needs not only to be re-searched and developed; it needs to be de-ployed, and industry will not do this untilit has seen the technology work. It was theAmerican navy’s deployment of nuclearreactors that convinced the world thatthey could be used as power generators.And it was the experience of deployingthem that allowed Admiral Rickover, inthe 1950s, to sum up the gap between ideasthat might work and those that are in factworking, in a way that still seems spot on60 years later:

An academic reactor or reactor plant almostalways has the following basic characteris-tics: (1) It is simple. (2) It is small. (3) It ischeap. (4) It is light. (5) It can be built veryquickly. (6) It is very �exible in purpose. (7)Very little development will be required. Itwill use o�-the-shelf components. (8) Thereactor is in the study phase. It is not beingbuilt now. After the dream, the humdrum reality

Cuba March 24thManufacturing April 21stInternational banking May 19thChina’s economy May 26th