SPECIAL ARTICLE
The Other Side of Nuclear Liability
Suvrat Raju, M V Ramana
The draft nuclear liability bill indemnifies the supplier of
a nuclear plant and caps the liability of the operator in
the event of an accident. The indemnity for suppliers is
meant to please multinational plant vendors who wish
to be free of liability even for accidents that result from a
design flaw. The cap on operator liability is far lower than
the potential damage that a nuclear accident could
cause. This clause is designed to facilitate the entry of
domestic big business into the nuclear market by
ensuring that domestic operators will not be held
responsible in the eventuality of damage following an
accident. Hence the bill transfers risks for a nuclear
mishap onto the people at large. Furthermore, it offers
almost no financial disincentive for unsafe behaviour on
part of the operators and suppliers of nuclear plants.
Suvrat Raju (suvrat@hri.res.in) is at the HarishChandra Research
Institute, Allahabad and M V Ramana (ramana@princeton.edu) is with
the Program on Science and Global Security, Princeton University, US.
48
T
he withdrawal of the civil nuclear liability bill from active
consideration in Parliament provides us with a short pe
riod to relect on some of the larger questions that are
often neglected in the arguments over the minutiae of the bill.
Not that the latter are unimportant, but neglecting the larger
questions is also unwise. To the extent that we discuss the bill
itself, we focus on its two major features rather than the details.
The primary reason for the government to withdraw the bill
was the strong opposition in Parliament and outside by various
political parties and other formations. At the same time, there is
also ample pressure, both foreign and domestic, to have some
legislation passed so that various new players can enter the
Indian nuclear energy market. Since the Manmohan Singh
government is committed to a large expansion of nuclear power,
it is expected that the bill will soon be revised and submitted for
approval. What might the contours of such a revised bill be?
To understand this, we irst broadly describe the expected
landscape of new nuclear commerce in India. In this landscape, a
supplier or vendor – a company like Westinghouse or General
Electric (GE) or Areva – will “supply” a nuclear plant to an Indian
operator. This operator will operate the plant to generate electri
city and sell it to the grid. Under the current Atomic Energy Act,
only the government, through the Nuclear Power Corporation of
India Limited (NPCIL) can operate nuclear plants. Although
private participation is currently disallowed, this might change
in the future.
The nuclear liability bill focuses on the operating lifetime of
the reactor and has two key clauses. First, in the event of an acci
dent, it indemniies the supplier from any kind of liability, regard
less of what caused the accident, i e, even if the accident was the
result of a design law. Second, it caps the liability of the operator
at Rs 500 crore, with additional compensation, if required, com
ing from the government. These clauses are designed to please
separate lobbies but are equally egregious.
The irst clause is largely a result of pressure from United States
(US) nuclear vendors. Before going any further, a clariication is
necessary. The nuclear industry today is so multinational in
nature that it may not be accurate to describe companies as
American or French. An example is Westinghouse, which manu
factures pressurised water reactors. Westinghouse was pur
chased by British Nuclear Fuels in 1998, which then sold it to
Japan’s Toshiba in 2006. In 2007, Toshiba sold 10% of the com
pany to Kazatomprom, the national uranium company for the
Republic of Kazakhstan. Besides these, the Shaw group, based in
Louisiana, US, owns 20% of the company and the Japanese IHI
Corporation a 3% share. Nevertheless, inasmuch as they continue
to maintain strong links with the US government, we describe
april 17, 2010
vol xlv no 16
EPW
Economic & Political Weekly
SPECIAL ARTICLE
companies like Westinghouse and GE as American. The US wants
India to accede to a speciic liability regime – the Convention on
Supplementary Compensation (CSC) – and is the major source of
political pressure on the government. However, other manufac
turers like Areva (France) also expect that India will accede to
some international liability regime.
All international liability regimes share the two features men
tioned above: they channel liability to the operator and cap this
liability, transferring inal responsibility for compensation to
governments. Both these provisions are in sharp contrast to
standard tort law; former Attorney General Soli Sorabjee (2009)
even suggested that these clauses would be struck down by the
Supreme Court. Why does the nuclear industry operate within
this special legislative framework?
Motivations for Liability Conventions
The answer to this is closely tied to the history of nuclear power
development in the US and its efforts to export nuclear techno
logy. The cap on total operator liability emerged in the irst con
text whereas the requirement that the supplier should be indem
niied from all legal liability came about in the second. There are
deep parallels between this history and the forces that are at play
in India today.
After the second world war, governments on both sides of the
Atlantic were keen to promote nuclear power. In the US, however,
for over a decade, electric utilities were reluctant to invest in
nuclear power, primarily because it was economically unattrac
tive. Government agencies offered “a series of incentives” that
“included a raft of subsidies, the promise of immediate proits,
sheltered investments, and the guarantee that unanticipated
costs could be passed on to consumers through rate increases.
Another potential advantage to private industry was the oppor
tunity to develop a capitalintensive technology, thus decreasing
dependence on labor” (Clarke 1985). These by themselves were
insuficient to induce suficient interest in nuclear power amongst
utilities. It was ultimately the threat the government would enter
the electricity generation market that catalysed the growth of the
civilian nuclear industry in the US.1 This was followed by a period
when utilities adopted nuclear power with great enthusiasm,
which lasted till around 1974.
Among the measures that the government undertook, was to
set up a liability regime. This was needed because a nuclear acci
dent could, as described later, cause immense loss of life and
property. Of more concern to the industry was the fact that the
claims for compensation that might be launched in the event of a
major accident could easily bankrupt the operator and supplier of
the plant.
One option was to obtain private insurance. But attempts to
get insurance coverage ran into two problems. First, the hazard
posed by nuclear power was entirely new, and there was no ben
eit to be had from “accumulated experience”; nor was there “the
opportunity to employ the trial and error procedure which
underlies so many segments of the insurance business” (Thomas
1958). The second more serious problem was that the cost of
damage from a nuclear accident could be immense, well beyond
the resources of any insurance company. And, the insurance
Economic & Political Weekly
EPW
april 17, 2010
vol xlv no 16
companies believed in the axiom, “if you write catastrophe in
surance you must expect to pay catastrophe losses” (ibid). It is
to deal with this problem that the US Congress introduced the
PriceAnderson Act. The act channelled all economic liability to
the operator of the plant, and limited even that to a ixed
amount, thereby making the government the ultimate insurer
in the event of an accident.
The notion that the supplier should be indemniied from all
legal liability came about when the US started to export nuclear
plants to western Europe. Western European governments were
keen to promote nuclear energy but, for various reasons, found
themselves importing American light water reactors (LWRs).2
American corporations were unwilling to bear liability for an
accident in a plant designed by them. The US Atomic Industrial
Forum conducted comprehensive studies examining how to
protect American suppliers from liability claims originating in
Europe. In an inluential publication, called the “Harvard report”,
it recommended legislative intervention to make it impossible for
anyone to ile a liability claim against atomic suppliers! Soon
afterwards, the irst international liability regime – the “Conven
tion on Third Party Liability in the Field of Nuclear Energy” (also
called the Paris convention) – was adopted in accordance with
these recommendations (Vanden Borre 2007).
Curiously, the US itself did not accede to the Paris convention
(or even later treaties like the Vienna convention). Under Ameri
can law it remains possible to sue the nuclear supplier although
economic liability is channelled to the “omnibus insurance” pro
vided by the PriceAnderson system. In fact, the US engineered
the Convention on Supplementary Compensation for Nuclear
Damage (CSC) in the late 1990s because of this and other incom
patibilities between existing liability conventions and Price An
derson system. Although the CSC shares the two principles men
tioned above, a “grandfather clause” allows the US to retain its
own tort law without any changes.
These unusual legislative measures were justiied by the
Atomic Industrial Forum 50 years ago by appealing to the “undis
puted fact that there is a vital national interest in the develop
ment of atomic power”. A very similar argument is advanced by
proponents of the liability bill and an analogous set of factors
drive attempts to create a common international liability regime
today. However, in the intervening 50 years we have seen the nu
clear industry battling soaring costs, public antipathy that is in
part a result of a catastrophic accident, and a persisting nuclear
waste problem. It would seem to be in the national interest to
turn away from nuclear power. But that possibility has not
entered the debate over the liability bill in any serious fashion.
We now turn to a more detailed examination of these two key
clauses in liability conventions.
Supplier Liability
Let us irst turn to another arena to understand what is at stake in
the notion of supplier liability. Speaking in February 2010 at a
hearing in the US House of Representatives, Henry Waxman, the
chairman of the Committee on Energy and Commerce, stated
In preparation for this hearing, the Committee analysed over 1,00,000
pages of documents from Toyota and the National Highway Trafic
49
SPECIAL ARTICLE
Safety Administration (NHTSA). These documents show that both
Toyota and NHTSA have received thousands of complaints of runaway
Toyota vehicles...Our review indicates that Toyota received as many as
2,600 complaints of runaway vehicles through its telephone hotline
alone. Over 700 of these incidents resulted in accidents. Toyota had
three responses: First, blame the driver. Second, blame the loormat.
Third, blame a sticky gas pedal. And NHTSA – without doing any
meaningful independent review – accepted Toyota’s explanations.3
The irst response is likely be seen as unfair, especially now that
Toyota’s head himself has admitted that the problem lay with
the company’s quick pace of expansion. Yet, it is the equivalent
of the irst response that is being written into law through the
nuclear liability bill.
The implicit assumption behind relieving nuclear suppliers
from liability for accidents is that designs are perfectly safe. But
can nuclear reactors reasonably claim to be completely immune
from design errors leading to accidents? The historical evidence
suggests otherwise.
After the 1978 Three Mile Island (TMI) accident in Pennsylva
nia, US President Jimmy Carter appointed a commission to inves
tigate the accident. The commission was headed by a mathemati
cian John Kemeny, and produced its report in 1979. Among other
things, the report observed that
several earlier warnings that operators needed clear instructions for
dealing with events like those during the TMI accident had been disre
garded by Babcock & Wilcox (B&W) [the supplier of the TMI reactor]
and the Nuclear Regulatory Commission (NRC).
Important among these was the September 1977 incident at the
DavisBesse nuclear plant, again a B&W reactor, when a pilotop
erated relief valve (PORV) stuck open and pressure fell, while the
pressure level showed by the appropriate gauge kept increasing.
The same was to occur in TMI and operators again assumed that
the pressure within the reactor was rising and shut off water low
to the reactor core. The DavisBesse incident was investigated by
both B&W and the NRC, but no information calling attention to
the correct operator actions was provided to utilities prior to the
TMI accident.
The Kemeny Commission also concluded that the “control
room was not adequately designed with the management of an
accident in mind” noting in particular that
over 100 alarms went off in the early stages of the accident with no
way of suppressing the unimportant ones and identifying the impor
tant ones. The danger of having too many alarms was recognised by
Burns and Roe (the architectengineer) during the design stage, but
the problem was never resolved...The arrangement of controls and in
dicators was not well thought out. Some key indicators relevant to the
accident were on the back of the control panel (Kemeny et al 1979).
Of course, the same design errors are unlikely to be present in
the reactors that were designed after TMI but there could be other
design errors, some of which may not have been identiied so far.
Indemnifying the supplier from liability has a serious conse
quence. Without liability, there is less incentive for the supplier to
design safe plants. Nuclear reactor designers have multiple con
straints. While they would obviously not want a reactor they
designed to undergo an accident, they also have to make sure
that the design is economical. Nuclear power, as is well known, is
already very expensive (see for example, Deutch et al 2003;
Ramana et al 2005; Ramana 2007 a and b). Thus, for example,
50
one reason that recent reactor designs have adopted passive
measures or components,4 as compared to the more traditional
active systems, is that “active safety systems are expensive to
build and operate” (Forsberg and Weinberg 1990). To the extent
that liability claims provide a signal, allowing potential lawsuits
would make reactor designers err on the side of safety rather
than economics.
There is a further concern. The ideological outlook that nu
clear designers function under often makes them discount safety
problems. For example, it has been long known that fast breeder
reactors have special safety problems posed by core rearrange
ments (see Wirtz 1978; Waltar and Reynolds 1981). Nevertheless,
the DAE has long pursued a breeder programme. While carrying
out studies of safety of its reactor design (Singh and Harish 2002),
it has also on occasion argued that safety concerns were com
pletely misplaced in the irst place. Thus, a DAE oficial argued
that the fast reactor community “ought to assert themselves and
destroy the sodium void phobia ...the necessity of a dome on the
top of the reactor vessel and the core catchers needs to be chal
lenged …after all, if the reactor can be designed to be inherently
safe or if the probability of failure of the shutdown function can
be brought down to 1e8 [1 part in 100 million] per demand, why
invest more funds for safety features” (Paranjpe 1992). This con
viction that the reactor is inherently safe comes in the way of reli
able safety studies, and indeed it has been shown that there are
serious problems with these studies (Kumar and Ramana 2008).
An example from a different technological arena of ideological
predilections within an organisation leading to an accident is the
January 1986 explosion of the Challenger space shuttle (see for
example, Vaughan 1996).
If vendors are indemniied against potential claims of safety,
their primary aim will then be to get the plant design approved
by the regulatory authority that needs to certify the design as be
ing safe. In terms of legal liability, then, the supplier’s job is done
once it has persuaded the regulatory agency, through whatever
means, of the safety of the design. The indemniication also does
not offer any inducements to the supplier to share new safety
concerns that come to light at a later stage with the operator or
improve the plant.
Moreover, this places a grave responsibility on the regulator
which must, often in a short period because of political pressures,
evaluate and approve a very complex design. The Indian regula
tory system is illprepared to accept this responsibility.
The organisation responsible for the regulation of nuclear safety
in India is the Atomic Energy Regulatory Board (AERB). There are
two problems with the AERB: it has limited personnel and facili
ties, and it is not autonomous. The AERB reports to the Atomic En
ergy Commission (AEC) that is chaired by the head of the DAE;
moreover, the AERB’s budget comes through the DAE. The AERB’s
lack of technical staff and testing facilities has been commented
on by Gopalakrishnan, a former chairman of the AERB
(Gopalakrishnan 1999). We note also that, as described earlier, the
Kemeny Commission also castigated the US NRC, which has greater
resources than the AERB, for errors leading to the TMI accident.
Thus, the DAE is not only in charge of ensuring rapid nuclear
growth, it is also the inal arbiter in matters of safety. A few
april 17, 2010
vol xlv no 16
EPW
Economic & Political Weekly
SPECIAL ARTICLE
months ago, the UK regulatory agency found laws in the designs
presented by Areva and Westinghouse for new nuclear plants and
withheld safety clearance (The Guardian 2009). Can one imagine
the AERB doing something similar? Not only would it face pres
sure from the manufacturers, but it would also have to contend
with its parent organisation – the DAE – which would have an in
terest in expediting the regulatory process.
The lack of autonomy in nuclear regulation is indicative of a
larger problem – the lack of independent sources of nuclear ex
pertise. Between nuclear vendors who are anxious to sell their
product, and a DAE, the sole oficial repository of all nuclear
knowledge, that is desperate to import reactors in large numbers
so as to increase nuclear power’s contribution to electricity gen
eration from a mere 3% to something more respectable, there is
no oficial body that policymakers could turn to for advice on
whether a reactor design is safe or not.
Cap on Operator Liability
We now turn to the issue of the liability cap on operators. As we
mentioned above, the cap on operator liability came about be
cause operators were unable to get the requisite insurance. One
reason was that it was very hard to estimate the probability of a
nuclear accident. More importantly, it was recognised that the
consequences of a nuclear accident would require compensation
that exceeded the capacity of the insurance market.
These two factors remain a problem to this date. From an actu
arial perspective, the civilian nuclear industry has over 13,000
reactor years of experience, with one catastrophic accident that
led to widespread radioactive contamination, i e, Chernobyl, and
a number of less severe accidents, of which Three Mile Island is
the best known example. This experiential basis is still too limited
to be certain about the probability of a major nuclear accident.
Further, “many nuclear safety related events occur year after
year, all over the world, in all types of nuclear plants and in all
reactor designs and that there are very serious events that go ei
ther entirely unnoticed by the broader public or remain signii
cantly underevaluated when it comes to their potential risk”
(Kastchiev et al 2007). Many of these accidents do not escalate
into major ones purely because of chance or the intervention of
human operators rather than any technical safety feature. From
the viewpoint of an insurance company, or indeed of the public,
such factors cannot be taken as assured.
The atomic energy establishment in India sometimes tries to
suggest that the probability of an accident can be calculated
precisely and is very low. For example, in November 2009, the
oficer in charge of the BARC safety council secretariat claimed at
the India Disaster Management Congress that “the probability of
occurrence of any major event is 106 ” (1 part in a million)
(Mishra 2009). This is disingenuous.
Theoretically, there are good reasons to be unsure about the
probability of an accident. At a narrow level, the nuclear industry
primarily analyses the safety of reactors using techniques called
probabilistic safety analyses (PSA) or probabilistic risk assessment
(PRA). The idea is to analyse an accident as resulting from a series
of failures, for which probabilities are assigned. The irst promi
nent application of PRA methodology to nuclear safety was in the
Economic & Political Weekly
EPW
april 17, 2010
vol xlv no 16
1975 reactor safety study in the US led by Norman Rasmussen
(NRC 1975). Following widespread criticism of the study, and es
pecially its executive summary, the NRC appointed an outside
panel to examine the Rasmussen study, which eventually submit
ted its report in 1978 (Lewis et al 1978). The panel identiied a
major problem with PRA methodology: that it is “conceptually im
possible to be complete in a mathematical sense the construction
of eventtrees and faulttrees” and this “inherent limitation
means that any calculation using this methodology is always
subject to revision and to doubt as to its completeness”. This
problem is practically inescapable.
A different, structural way of thinking about the same problem
is to think about the nature of the technology. The different sys
tems of a nuclear reactor interact in complex ways, making it pos
sible that multiple failures could interact in unexpected ways. A
second factor is the presence of tightly prescribed steps and un
changeable sequences in operation that must be adhered to.
Therefore accidents can escalate quickly, with few alternate path
ways to diffuse them (Perrow 1984). Safety interventions,
whether by humans or automatic safety equipment must occur
quickly, and be adequately planned for. Both these factors pose
challenges to safety and its demonstration. They also make it
more dificult to infer safe operation from past record; a system
could have relatively minor accidents, but many such failures
could combine unexpectedly in the future, leading to a much
larger accident.
There is, thus, uncertainty about the probability of an acci
dent. What of the magnitude of the damage? The Indian bill caps
operatorliability at Rs 500 crore and this igure has drawn criti
cism. In defence, supporters of the liability have argued that
other countries have similar igures. But this does not answer the
criticism. Even some of the highest operator liabilities in other
countries – in Japan, for example, plant operators must provide a
“inancial security amount” of $1.2 billion or about Rs 6,000
crore (World Nuclear News 2009) – are triling compared to esti
mates of the potential damage from a nuclear accident. The CRAC
II simulations conducted by the Sandia National Laboratory in
1982 examined the possible consequences of a worstcase accident
under worstcase conditions (a socalled “class9 accident”).5 For
the Indian Point nuclear plant near New York, this study esti
mated a property damage of $274 to $314 billion (in 1982 dollars).
These igures are debated by, for example, the US Nuclear
Regulatory Commission. But the key point is that the magnitude
is so much in excess of what is offered as liability.
Of course, these igures are speciic to that reactor and to the
location, i e, New York. But there is no corresponding study for
the case of, say, Jaitapur or Kovvada. This represents a serious
failure of process. Before the government even drafted such a
bill, it should have launched a transparent and comprehensive
evaluation of what the likely inancial damage would be in
the event of a catastrophic nuclear accident under Indian
circumstances.
Such an evaluation should ideally be carried out by a panel
that is independent of the nuclear establishment and involves not
just engineers and scientists but also economists, public health
professionals, doctors, psychologists, agricultural scientists, and
51
SPECIAL ARTICLE
environmental restoration specialists. It should take into account
the speciicities of the country and proposed locations, including
the largely rural population of India, the dependence on land and
water resources for livelihoods, the state of public health, the ca
pacity to deal with emergencies, and analysis of the effects of
other industrial accidents in the country. And inally the inputs of
the populations likely to be impacted should be taken into ac
count, perhaps through offering a draft of the results for public
comment. Without knowing whether the likely damages will be
of the order of Rs 10 crore or Rs 1,00,000 crore, how can there be
any discussion of where the liability limit should be placed?
However, contrary to common supposition, the liability cap is
not designed to please the us. In fact, American companies do
not really care how much the Indian operator needs to pay
provided that they are not liable for anything. The low cap on
liability is, very likely, the result of pressure from Indian big
business which wants to enter the lucrative nuclear power
industry. For example, in his 23 September 2008 statement as
Chairman of Reliance Power, Anil Ambani explicitly stated that
they expect the government to “make at least 15 other policy
changes to permit the entry of private players into the nuclear
domain”, including “enacting national legislations on civil liabili
ties” (http://www.reliancepower. co. in/ html/pdf/R_Power_
Chairmans_Speech%5B1%5D.pdf).
The present government is certainly responsive to such con
cerns. In a recent interview, Prithviraj Chavan insisted that “if we
don’t keep the caps reasonable, nobody will invest in India”
(Phadnis and Chatterji 2010). He could not have been referring to
foreign plant manufacturers (for whom the cap is zero!) so he
must have been referring to private investment in the operation
of nuclear plants.
Setting the cap at Rs 500 crore for private operators, again,
sends the wrong signal. The cost of a nuclear plant itself could be
as much as 60 times higher than this cap. For example, the reac
tor that Areva is constructing in Olkiluoto, Finland is anticipated
to cost more than $6 billion or about Rs 30,000 crore. In a simpli
ied scenario, consider a private company that is in the position of
asking the supplier for a design modiication that could lead to
delays and a cost increase of 10% of the plant value. Would the
company rather spend Rs 3,000 crore on this or take the risk of
paying damages of Rs 500 crore sometime in the future?
The above example neglects the loss of revenue that the com
pany would suffer in the event of an accident but including this
does not change the essential conclusion. Assume that the plant
has a small and constant probability of an accident every year r,
and an anticipated lifetime L. Further, say that the cost of the
plant is C and this is recovered over the time L (including a proit
rate will not change the essential result here). The liability cap is
What Is Maoism
and
Other Essays
Edited and Introduced
by
This book is motivated by a desire to rekindle an imagination of
socialism that brings to the fore the emancipation and fulfilment of
the basic human needs of the most exploited, the most oppressed,
and the most dominated on this earth.
CONTENTS
Acknowledgements
From the Intellectual “Underworld”: An Introduction
SECTION I : From Marx to Mao
What is Maoism? —Bernard D’Mello
SECTION II : Marxism: Its Revolutionary Essence
What is Marxism? —Paul M Sweezy
Notes on the Centennial of Das Kapital —Paul M Sweezy,
Marxism and Revolution 100 Years after Marx —Paul M Sweezy
SECTION III :Leninism and Its Stalinist Decomposition
Lenin and Stalin —Harry Braverman
The State and Revolution —Ralph Miliband
52
Bernard D’Mello
SECTION IV : Where It All Began: Maoism in China
The Formation of Mao’s Economic Strategy, 1927-1949 —John G Gurley
China: Contrasts with the USSR —Harry Magdoff
Mao, Rural Development, and Two-Line Struggle —William Hinton
On the Role of Mao Zedong —William Hinton
SECTION V : Maoism in India and Nepal
Recent Developments in the Naxalite Movement —Tilak D Gupta
Maoism in India: Ideology, Programme and Armed Struggle —Tilak D Gupta
Interview with Baburam Bhattarai: Transition to
New Democratic Republic in Nepal—WPRM (Britain)
SECTION VI : The Promise of Radical Democracy
Capitalism and Democracy —Paul M Sweezy
Socialism in Poor Countries —Paul M Sweezy
Socialism: Legacy and Renewal —Paul M Sweezy
Notes on Contributors
ISBN: 978-81-88401-23-9; pp. x+318; Rs.150
CORNERSTONE PUBLICATIONS, P.O. HIJLI CO-OPERATIVE
KHARAGPUR-721306, INDIA
e-mail: cornerstone_pub@sancharnet.in
april 17, 2010
vol xlv no 16
EPW
Economic & Political Weekly
SPECIAL ARTICLE
F. If the accident happens in the irst year, the company entirely
loses C and in addition must pay F. If the accident happens in the
last year of operation, the company hardly loses any revenue
but must still pay F. Neglecting the discount rate for simplicity,
the expected loss of the company (for small r) is approximately
(F + C/2) r L. A more detailed and general calculation is pre
sented in the Appendix (p 54).
Let us understand this formula a little better. Say that a plant
costs Rs 30,000 crore as above and the company inds that there
is a problem in the plant that leads to an estimated 1% chance of
an accident over its lifetime. This is, in fact, a very large and seri
ous risk. With a liability cap of Rs 500 crore, the expected loss of
the company is just Rs 155 crore. So, if it costs more than Rs 155
crore to correct this problem there is no inancial incentive for
the company to spend the money. In fact, even if the corrections
do not cost anything in actual expenditure, but merely lead to a
six month delay, at the beginning of the plant’s life cycle, at a 3%
discount rate, this would be equivalent to a loss of about Rs 450
crore which is more than the expected loss above.
Of course, this is not the whole story since the company would
probably be concerned about several intangibles like the loss of
reputation in the event of an accident. Moreover, the engineers
and safety analysts involved might be careful about safety just out
of professional integrity. However, the point illustrated above is
robust: if the liability cap is very small compared to the typical
sums of money that are spent on the plant, then the inancial lia
bility itself does not provide any disincentive for unsafe behaviour.
A Revised Bill
What might a revised bill look like? The political analysis above
makes a few predictions. The US is very clear that its suppliers
must be immunised from liability. And because the Manmohan
Singh government is beholden to the US for having shepherded
the waiver from the Nuclear Suppliers Group, it is very tough for
it to stand up to American pressure. Although acquiescing to this
demand will have serious safety implications as we described
above, the government is likely to push the American case. So a
revised bill will continue to channel liability to operators.
What about the issue of the cap on operator liability? The po
litical pressures here are fewer. It is true that private companies
– like the Tatas and Reliance – will not enter the nuclear market
unless they are assured that they will not be required to pay com
pensation for all the damage that they may be responsible for.
While the Manmohan Singh government would probably like to
funnel proits from the nuclear industry into the private sector,
this may not be a pressing priority.
Moreover, because the nuclear establishment in India is inter
ested in maintaining some sort of monopoly over reactor opera
tions, private companies are likely to be allowed to enter the
nuclear sector only as partners of the NPCIL. This is the case, for
example, with the joint venture between the National Thermal
Power Corporation (NTPC) and NPCIL, where NTPC has only a
49% minority stake (The Hindu 2010). This tells us that to assuage
political opposition, the government might well raise the liability
cap on operators. However, this may be accompanied with some
behind the scenes assurance that in any individual contract, the
Economic & Political Weekly
EPW
april 17, 2010
vol xlv no 16
liability for a nuclear accident might be transferred to the NPCIL
(which is a public sector company) freeing up the private partner
from this responsibility.
It is worth noting though that a large increase in the liability
cap would also help to ameliorate the lack of supplier liability in
the bill. The draft bill explicitly allows the operator a “right of
recourse” if the accident was caused by an act of “wilful or gross
negligence on the part of the supplier”. While this is no substitute
for allowing the victims to sue the supplier it does mean that sup
pliers would be concerned that operators would try and transfer a
large part of the liability onto them. This will induce them into
taking greater care, especially if large sums of money are involved.
A note of caution is in order here. While the CSC explicitly
allows the operator a right of recourse “if the nuclear incident
results from an act ... done with intent to cause damage,” the
Indian bill extends this to cases of “gross negligence”. It is quite
likely that American companies will be unhappy with this clause.
Omer Brown, a prominent legal analyst for the US nuclear indus
try complained that these terms are illdeined and “the differ
ences between negligence, gross negligence and wilful act are as
clear as the differences between a fool, a damn fool and a goddam
fool” (Varadarajan 2010). While it might be politically dificult for
the government to delete this clause from a revised bill, individ
ual contracts that the NPCIL signs with international suppliers
might well explicitly renounce this right of recourse.
We should mention here that in recent times, senior govern
ment oficials seem to be suggesting that the nuclear liability bill
just sets the “strict liability” of the operator; i e, to claim the
Rs 500 crore in compensation one does not need to demonstrate
fault on part of the operator but that existing tort laws will con
tinue to apply and that the liability of the operator (and perhaps
even the supplier) here would be unlimited. As the Hindustan
Times put it, “according to the government, the bill’s intent is not
to limit the compensation, but to ensure at least some immediate
compensation” (Hindustan Times 2010).
This suggestion contradicts the explicit assurance in the bill
that “no civil court shall have jurisdiction to entertain any suit or
proceedings in respect of any matter which the Claims Commis
sioner...is empowered to adjudicate under this Act.” As we de
scribed above, the cap on operator liability serves a purpose. It
protects corporations from the bankruptcy that might result from
a nuclear accident. And in order that this function is maintained,
ordinary tort law cannot apply to nuclear accidents if this bill is
passed; the government’s claims to the contrary are misleading.
Concluding Thoughts
The nuclear liability bill is a result of two different sources of pres
sure: foreign countries, most prominently the US, and domestic
big businesses which would like to come into the nuclear business
because of the potentially large proits. But, as we have argued
here, the two crucial clauses of the liability bill run counter to the
principles of natural justice. The bill, in its current form, or in the
modiied forms that we think are likely, must be opposed.
At the heart of the debate over nuclear liability is a conlict of
interest. On the one hand is the risk faced by nuclear suppliers and
operators that they may, some day, be forced to pay large amounts
53
SPECIAL ARTICLE
of compensation for catastrophic accidents that they can, through
their actions, lower the chances of, but never eliminate. On the
other hand, there are the risks faced by the inhabitants of the
regions close to a nuclear reactor and the interests of the people at
large who might be stuck with the burden of cleaning up the
Notes
1
2
3
4
5
In 1956, Senator Albert Gore (father of the former
vice president of the United States, Al Gore) and
Representative Chatield introduced a bill that
would direct the US Atomic Energy Commission to
build nuclear power plants to provide electricity for
its uranium enrichment and plutonium production
activities, instead of buying electricity from nearby
utilities. This cleared the Senate but Republican
opposition did not allow it to pass in the House,
proposing instead, “to give private industry more
incentives to build the nuclear plants”.
On the emergence of the light water reactor as the
leading nuclear reactor type see Cowan (1990).
See
http://energycommerce.house.gov/Press_
111/20100223/waxman.statement.2.23.10.pdf.
The background to this case was the numerous re
ports of incidents where cars manufactured by
Toyota had experienced uncontrollable accelera
tion leading to accidents, some of which were fatal.
Toyota later accepted that this was due to a dei
ciency in design and recalled millions of its cars.
A passive component operates without any exter
nal input, for example operators or equipment, to
activate the function. See Anders Martensson
(1992).
CRAC stands for Calculation of Reactor Accident
Consequence.
References
Clarke, Lee (1985): “The Origins of Nuclear Power: A
Case of Institutional Conlict”, Social Problems,
32, No 5.
Cowan, Robin (1990): “Nuclear Power Reactors: A
Study in Technological LockIn”, The Journal of
Economic History, 50, No 3.
Deutch, John et al (2003): “The Future of Nuclear
Power: An Interdisciplinary MIT Study”, Massa
chusetts Institute of Technology, Cambridge,
MA.
Forsberg, Charles W and Alvin Weinberg (1990):
“Advanced Reactors, Passive Safety and Accept
ance of Nuclear Energy”, Annual Review of
Energy, 15.
Gopalakrishnan, A (1999): “Issues of Nuclear Safety”,
Frontline, 1326 March.
Hindustan Times (2010): “PM Open to Talks on
Nuclear Liabilities Bill”, 1 April.
Kastchiev, Georgui et al (2007): “Residual Risk: An
Account of Events in Nuclear Power Plants since
the Chernobyl Accident in 1986”, The Greens/Eu
ropean Free Alliance, Brussels
Kemeny, John G et al (1979): “Report of the President’s
Commission on the Accident at Three Mile
Island”, Kemeny Commission, Washington DC.
Kumar, Ashwin and M V Ramana (2008): “Compro
mising Safety: Design Choices and Severe Acci
dent Possibilities in India’s Prototype Fast Breed
er Reactor”, Science and Global Security, 16.
Lewis, H W et al (1978): “Risk Assessment Review
Group Report to the US Nuclear Regulatory
Commission”, Nuclear Regulatory Commission,
Washington DC.
Martensson, Anders (1992): “Inherently Safe Reac
tors”, Energy Policy, 20, No 7.
Mishra, S K (2009): “Role of Regulator in Control of
Nuclear and Radiological Emergencies” in the
Second India Disaster Management Conference,
National Institute of Disaster Management, New
Delhi, 46 November (http://nidm.gov.in/idmc2/
PDF/Abstracts/Nuclear.pdf).
NRC (1975): “Reactor Safety Study (RSS), Study
54
results of a catastrophic accident. Whose interests should the gov
ernment protect? The Manmohan Singh government would like to
intervene legislatively to immunise multinational corporations
and limit the liability of domestic operators. A clearer indicator of
the nature of the current Indian state cannot be found.
Director: N V Rasmussen”, US Nuclear Regulatory
Commission, Washington DC.
Paranjpe, S R (1992): “Report on the Specialists’ Meet
ing on Passive and Active Safety Features of Liq
uidMetal Fast Breeder Reactors Organised by the
International Atomic Energy Agency at Oarai En
gineering Centre of Power Reactor and Nuclear
Development Corporation, Japan, 57 November
1991”, Nuclear Safety, 33, No 4.
Perrow, Charles (1984): Normal Accidents: Living with
High Risk Technologies (New York: Basic Books).
Phadnis, A and S Chatterji (2010): “Government Open
to Raising Nuclear Liability Cap”, Business Standard, 14 March.
Ramana, M V, Antonette D’Sa and Amulya K N Reddy
(2005): “Economics of Nuclear Power from Heavy
Water Reactors”, Economic & Political Weekly, XL,
No 17.
– (2007a): “Economics of Nuclear Power: Subsidies
and Competitiveness”, Economic & Political Weekly, XLII, No 2.
– (2007b): “Heavy Subsidies in Heavy Water”,
Economic & Political Weekly, XLII, No 34.
Singh, Om Pal and R Harish (2002): “Energetics of
Core Disruptive Accident for Different Fuels for a
Medium Sized Fast Reactor”, Annals of Nuclear
Energy, 29.
Sorabjee, Soli (2009): “Capping Nuclear Liability is a
NonStarter”, The Hindu, 13 December.
The Guardian, (2009): “Nuclear Reactors Contain
Safety Flaws, Watchdog Reveals”, 27 November.
The Hindu (2010): “NTPC, NPCIL Venture to Kickstart
Soon”, 18 January.
Thomas, DeRoy C (1958)” “Can We Insure against Lia
bility from Nuclear Incidents?”, California Law
Review, 46, No 1.
Vanden Borre, Tom (2007): “Shifts in Governance in
Compensation for Nuclear Damage, 20 Years after
Chernobyl” in Michael Faure and Albert Verheij
(ed.), Shifts in Compensation for Environmental
Damage, SpringerWein, New York.
Varadarajan, Siddharth (2010): “Liability Cap Low but
Washington Still Fears Lawsuits”, The Hindu,
8 March.
Vaughan, Diane (1996): The Challenger Launch Decision: Risky Technology, Culture, and Deviance at
NASA (Chicago: University of Chicago Press).
Waltar, Alan E and Albert B Reynolds (1981): Fast
Breeder Reactors (New York: Pergamon Press).
Wirtz, Karl (1978): Lectures on Fast Reactors, Ameri
can Nuclear Society, La Grange Park, IL.
World Nuclear News (2009): “Japan Revises Liability
Laws”, 23 April, http:// www.worldnuclear
news.org/RSJapan_revises_nuclear_liability _
laws2304091.html
Appendix
Let us try and make the calculations for the
inancial disincentive provided by the liabil
ity cap a little more precise. We use the follow
ing notation: C – overnight cost of construc
tion; α – discount rate; F – the total cap on
liability; L – the total lifetime of the plant;
r – the probability of having an accident per unit
time; i – the net income (revenue − costs) of the
plant per unit time. Here we consider the case
of a catastrophic accident; the moment this ac
cident happens, the plant shuts down, the opera
tor pays the ine and loses all future revenue.
Let us start with the discrete time case and
then pass to the continuous case. We divide
time into units of δt. The probability that no
accident happens for n units of time, and then
happens is:
p(n)=(1–rδt)nrδt.
(1)
The total income earned in this case is
I(n)=–C–
F
(1+αδt)n
S
+
n
iδt
j=1
(1+αδt)j
(2)
.
The probability that no accident happens in
the lifetime of the plant and the income earned
in that case is
L
δt
iδt
j=1
(1+αδt)j
I0=–C+ S
L
p0=(1–rδt) δt,
(3)
.
So, the expected income in L years ( δt units
of time) is
L
L
δt
〈IL〉=p0I0+ S I(n)p(n)
L
δt
=p0I0+ S
n=1
(
(
)
n=1
–C–
(1+αδt)n
)
+S
n
F
(
iδt
j=1 (1+αδt)j
n
(1–rδt) rδt
)
→ –C+∫0 ie–xαdx e–rL+∫0 –C–Fe–tα+∫0 ie–xαdx e–rtrdt
L
L
t
e–L(α+r) ((–1+eL(α+r)) (i–Fr)–CeL(α+r) (α+r))
α+r
=
(4)
In our problem, r is small (since, typically
r ≈10–4/yr or even smaller). So, we can expand
to irst order in r, which leads to
〈IL〉≈–
Cα+(–1+e –αL)i
α
F
i –αL
+ —(e –αL–1)+–(e
+Lα–1) r.
α
α2
[
]
(5)
The irst term is the proit that would be
made without accounting for accidents, while
the second term tells us the disincentive from
the revenue loss and ine. We get the answer
used in the text when we set the discount rate
α=0 and also set the proit to zero, i e, i=C/L.
In this limit, we see that
lim〈IL〉|i= C ≈ –(F+ 2 )Lr+O(r2 L2)
α→0
C
(6)
L
which matches with our previous answer.
If the discount rate is about 3% and the plant
is operating over a lifetime of about 30 years,
then we cannot use this formula but must in
stead use Eqn 5. The effect is to change a few
factors but the important features of the result
are unchanged.
april 17, 2010
vol xlv no 16
EPW
Economic & Political Weekly