Baron, J., Gowda, R., & Kunreuther, H. (1993). Attitudes toward
managing hazardous waste: What should be cleaned up and who
should pay for it? Risk Analysis, 13, 183-192.
Attitudes toward managing hazardous waste: What should
be cleaned up and who should pay for it?
Wharton Risk Management and Decision Processes Center
The Wharton School
University of Pennsylvania
Philadelphia, PA 19104-6366
Wharton Risk Management and Decision Processes Center
Working paper #92-11-02
Pre-publication version of an article with the same title and
authors published in Risk Analysis, 13, 183-192.
This work was supported by NSF Grant SES-88-09299, and
grants from the Hewlett Foundation, Johnson and Higgins, and the
Center for Law and Economics, University of California, Berkeley.
We thank Bill Wymer for conducting an initial focus group and for
editing the questionnaire, David Goldstein, Ramkumar Kashyap,
Jacqueline Meszaros, Florence Nelson, Maruska Pinches, Penelope
Pollister, Matthew Robinson, Dan Rubinfeld, and Joanne Schwartz,
for other assistance. Frederick Allen, David Aylwart, Dennis
Connolly, Jan Edelstein, Marianne Lamont, Howard Platzman, Kate
Probst, Jackie Warren and members of the Wharton Risk Management
and Decision Processes Center Advisory Committee provided helpful
comments on an earlier draft. Preliminary findings were
presented at the annual meetings of the Association for Public
Policy Analysis and Management and the Society for Risk Analysis
in 1990 and 1991, respectively. A longer version of this report
is available from the third author.
Authors are listed alphabetically. Jonathan Baron is
Professor of Psychology, University of Pennsylvania; Rajeev Gowda
is Research Fellow, Science and Public Policy Program, and
Assistant Professor of Political Science, University of Oklahoma;
Howard Kunreuther is Riklis Professor of Decision Sciences, Risk
Management and Insurance, and Public Policy and Management,
Wharton School, University of Pennsylvania.
Hazardous waste policy in the United States uses a
liability-based approach, including strict, retroactive and
joint-and-several liability. To assess attitudes toward these
basic principles of liability, and toward priorities for cleanup
of wastes, a questionnaire was mailed to legislators, judges,
executives of oil and chemical companies, environmentalists, and
economists. The questionnaire consisted of abstract, simplified
cases, which contrasted basic principles rather than dealing with
real-world scenarios. Subjects were asked how they would
allocate cleanup costs between companies and government as a
function of such factors as adherence to standards, adoption of
best available technology (BAT), and influence of penalties on
Most subjects felt that if the company followed government
standards or used the best available technology (BAT), it should
pay for only a portion of the cleanup cost, with the government
paying the rest. In general, responses did not support the
principles underlying current law - strict, retroactive, and
joint-and-several liability. Most subjects were more interested
in polluters paying for damages than in deterrence or future
benefit - even to the extent that they would have `harmless'
waste sites cleaned up. A bias was found toward complete cleanup
of some sites, or `zero risk.' Different groups of subjects gave
similar answers, although more committed environmentalists were
more willing to make companies pay and to clean up waste
regardless of the cost.
The issue of liability allocation is at the heart of
hazardous waste policy and is essential to an understanding of
its subsequent performance. The intent of the Comprehensive
Environmental Response, Compensation, and Liability Act of 1980
(CERCLA, or Superfund) was that highly dangerous hazardous waste
sites should be cleaned up quickly. To facilitate this process,
the government established a fund from a variety of tax sources
that may be used to undertake cleanup. Strict liability and
retroactive liability were included in the law to ensure that the
government is able to receive compensation for the cleanup
expenditures that it undertakes or to induce responsible parties
to perform the cleanup. Strict liability means that the
government has to prove only a party's involvement at a waste
site, not its negligence. Retroactive liability means that any
party can be held liable for past actions even if it was
following prevailing government regulations at the time.
At some CERCLA sites where there are two or more potentially
responsible parties (PRPs), the rule of joint-and-several
liability (JSL) has been applied. This rule implies that the
PRPs as a whole, or a subset - even including just a single
party - may be held liable for all the costs associated with
cleaning up a particular site. In other words, JSL could act to
divorce the liability faced by a firm from its actual conduct and
responsibility. Controversy has emerged over the effect of this
doctrine on the cleanup process (e.g., Katzman, 1988; Schmit,
Anderson and Oleszczuk, in press; Wilkerson & Church, 1989; for
other recent references on the effectiveness of CERLA in general,
and of the Resource Conservation and Recovery Act of 1976, see
Acton and Dixon, 1992; Cheek, 1989; Environmental Protection
Agency, 1990; Probst & Portney, 1992; Russell et al., 1992).
The controversy about Superfund raises a number of questions
about attitudes concerning fairness in the allocation of cleanup
costs and attitudes about the tradeoff between cost and benefits
of cleanup. A few studies relevant to the tradeoff question have
been done. The cost-allocation question is related to
psychological studies of responsibility for harm. These studies
have typically found that average responsibility or liability
judgments are greater for negligent harm than for unintentional
harm, although some responsibility is assigned even for
unforeseeable accidents (Fincham & Jaspers, 1980; Shultz &
Wright, 1985; Thomas & Parpal, 1987). But none of these studies
has assessed the number of subjects who would favor strict
liability vs. negligence, none has examined the principles
underlying JSL, and none has used an environmental context.
We therefore undertook a survey to sample opinions about
both liability allocation and about priorities for cleanup. The
survey dealt with attitudes about basic principles, so we used
simplified cases in an attempt to make these principles salient.
We assume that attitudes about such principles play a significant
role in determining attitudes toward specific policies and are
therefore relevant to understanding the acceptability of
hazardous waste policy.
One aim of the survey was to specify the determinants of
percieved responsibility for cleanup of past waste, as indexed by
allocation of cleanup costs between a company and the government.
We shall describe the determinants examined later. The survey
also examined attitudes toward the desirability of cleaning up
hazardous waste. It has been argued that the public (compared to
expert judgment) puts too much emphasis on waste cleanup and that
much of the money spent on cleanup could better be spent on
reducing other health and environmental hazards (Environmental
Protection Agency, 1990; Kunreuther & Patrick, 1991). Russell et
al. (1992) estimated the cost of remediation at $750 billion, but
they point out that this figure could change by hundreds of
billions in either direction with a shift in policy.
Excessive emphasis on cleanup could result from patterns of
judgment that are biased with respect to a consequentialist
standard, i.e., the principle of bringing about the best overall
consequences. For example, people might prefer zero risk at one
site to a greater reduction in total risk (but non-zero risk) at
two sites. Such biases have been found in other studies of
individuals who have no particular stake in the outcomes (Noll &
Krier, 1990; Baron, in press a,b). In the type of mail
questionnaire distributed in this study, it is impossible to
determine the bases for subjects' responses, so conclusions about
biases must remain tentative, unless they are supported by more
careful studies in the laboratory.
The groups sampled were the principal parties interested in
the formulation and implementation of Superfund. Efforts to
insure return of the questionnaire met with varying success, so
samples were not representative of their respective populations.
We examine the various groups primarily to determine the extent
to which opinions vary across groups, and our samples may be
adequate for that purpose. The groups were as follows, with
return rates in parentheses after each group name:
Chief Executive Officers (CEOs) of Oil and Chemical
Companies (60 returned out of 418 mailed, 14.4%). We focused,
in particular, on those companies that are already contributing
to cleanup efforts through feedstock taxes under CERCLA (even
though they may be only a small fraction of the PRPs involved in
the cleanup effort). The names of the specific companies were
obtained from the Dunn's and Infotrac databases using Standard
Industrial Classification codes of the products subject to the
chemical and petroleum feedstock taxes.
Economists (57/200, 28.5%). This sample consisted of
20 responses from a random sample of the American Economic
Association members and 37 responses from a random sample of a
list of scholars provided by the Center for Law and Economics at
the University of California, Berkeley.
Environmentalists (94/269, 34.9%). This sample
consisted of 71 subscribers to the environmental policy
newsletter issued by the Council for Economic Priorities and 23
grass-roots environmental activists who attended a 1990 meeting,
organized by the Philadelphia branch of Clean Water Action, with
the EPA Administrator, William Reilly.
Experts on Hazardous Waste Policy (29/100, 29%). This
group consisted of participants reflecting the different
interested parties concerned with hazardous waste policy. All of
them participated in a 1990 conference on `Minimizing
Environmental Damage: Strategies for Hazardous Waste Management'
in the Philadelphia area under the auspices of the Wharton Risk
and Decision Processes Center and the Environmental Associates of
the Academy of Natural Sciences.
Judges (89/200, 44.5%). These were members of
Judicate, an organization of retired and semi-retired judges who
serve as arbitrators.
Legislators (104/1170, 8.9%). Questionnaires were
mailed out to members of the following groups: Chairpersons of
Environmental Committees of State Legislatures; Chairpersons of
Natural Resources Committees of State Legislatures; All members
of the US Senate; Members of US House Ways and Means Committee;
Members of US House Public Works and Transportation Committee;
Members of US House Energy and Commerce Committee; ten state
legislators each from the 15 states with the lowest and thirteen
states with the highest number of hazardous waste sites; a random
selection of some of the remaining members of the US House.
Since only 67 (9.2%) of these questionnaires were returned, a
second mailing was sent to a 336 federal and state legislators
who were members of committees that dealt with hazardous waste
policy. The staffer in charge of hazardous waste policy could
fill out the questionnaire reflecting the legislator's viewpoint
on the subject. An additional 27 questionnaires were returned
from this sample.
WHO SHOULD PAY FOR CLEANUP: THE ONE COMPANY SCENARIO
A. Nature of cases
The first part of the questionnaire asked subjects how they
would allocate the costs of cleanup between a single company that
disposed of waste and the government. The following background
information was provided: `In 1990, it has been discovered that
there is groundwater contamination around a landfill owned and
used solely by a well-established manufacturing company. The
Environmental Protection Agency has determined that the site
needs to be cleaned up.' Subjects then considered the following
cases, quoted below essentially in their entirety. (Subjects
were not given the case descriptors, such as `Negligence' and
Case 1 (Negligence): `The party had started disposing
of its wastes in the landfill in 1983, well after hazardous waste
legislation was passed. It failed to meet government standards
for hazardous waste treatment and disposal.' This case reflects
gross negligence by the company.
Case 2 (Standards): `What if the company had
followed government standards for hazardous waste treatment and
disposal? Although it followed the standards, it did not use the
best available technology to avoid pollution.'
Case 3 (Profitability): `What if the company did
not make a profit from the products that resulted in the
hazardous waste by-products? ... ' The rest of this case was
identical to Case 2, so that the two cases could be compared.
Case 4 (Improve) and 5 (Stop): These two cases,
printed on the same page, were designed to determine whether
respondents were sensitive to the consequences of penalties in
determining what proportion of the cleanup cost the company
should pay. The conditions were as in Case 2 except for the
Case 4: `The company in this case makes products that are
highly beneficial to people. ... The more the company is
made to pay, the more likely it is that the company and other
companies will adopt the best available technology to avoid
pollution in the future.' (Other details were identical to Case
Case 5: This case is the same as 4 except for the last
sentence, which reads, `The more the company is made to pay,
the more likely it is that the company and other companies,
instead of changing their technology, will stop making the
products that led to the waste.'
Those respondents who believe that a penalty should be designed
to act as a deterrent will favor a higher penalty in Case 5 than
Case 6 Best Available Technology (BAT): `Suppose that
in the early 1980s the company followed government standards and
used the best available technology but there was still
groundwater contamination from the landfill. Today there is
still no way that this leakage could have been prevented.' This
case can be compared to Case 2.
Case 7 (Hindsight): This case was designed to test
whether respondents would judge past actions by present
standards, ignoring the fact that a company may have done the
best thing possible at the time. It was identical to Case 6
except that the last sentence read, `Research developments show
that the technology used earlier was flawed and that the
leakage from the landfill could have been prevented if today's
state-of-the-art technology had been used in the early 1980s.'
(In a second version, given to different subjects, the waste is
deposited in 1991 and the contamination is discovered in 1995.
The responses to the two versions did not differ, so this
distinction was ignored.)
Case 8 (Innocent): `What if the company disposed of
all of its waste in a landfill in the 1960s, before any
legislation on hazardous waste management was passed? It did not
treat its wastes before disposal. The company thought that the
waste was not harmful. The company was in compliance with
existing standards.' (In a second version, the company
`treated its wastes before disposal, voluntarily, using the
best technology available at that time?' The responses to the
two versions did not differ so they are combined in the
analyses.) Those who feel that the innocent polluter should pay
a proportion of the cost would be strong believers in strict
Table 1 shows the average proportion of the costs that
subjects assigned to the company for each case. The groups are
ordered by the average percentage that they feel the company
should pay across all 8 cases. The Experts assigned the company
the same or a higher proportion of the cleanup costs than any
other group for each case. As might be expected, the CEOs
treated the company more leniently on the average, but all
significant differences in cost assignment among groups
disappeared when the experts were excluded from the analysis.
(When the experts were included, significant differences among
groups are found by a Kruskal-Wallis analysis of variance in
Cases 2, 3, and 4.) In general, then, the CEOs did not differ
from the other groups. The Experts (and, in the analysis of
liability rules, the Environmentalists) accounted for all the
significant group differences.
Mean allocation of costs to the company in the one-company
cases, by group, in order of overall willingness to make the
1 2 3 4 5 6 7 8
Negl. Stds. Profit Improve Stop BAT Hindsight Innocent
Experts 100 80 81 82 71 67 67 52
Envir. 98 69 68 72 70 51 51 52
Econ. 98 60 60 65 59 46 49 39
Judges 97 59 62 63 58 46 45 40
Legis. 97 63 63 67 63 40 42 38
CEOs 99 54 54 55 56 46 39 38
All 98 63 63 66 62 47 47 42
We classified responses to the eight cases into three
different liability rules as follows. We recognize that these
definitions are not precisely consistent with existing legal
terminology but are useful for distinguishing among individuals'
Strict: The company pays 100 percent of cleanup costs
for all eight cases.
Negligence: We define this term to mean that, if the
company followed existing government regulations and standards,
it would not be liable for any of the cleanup costs. Under this
rule, the company would pay 100% of the costs in Case 1 only
and 0% of the costs in the other seven cases.
Partial: This rule falls between the other two. The
company's liability is partial, and is inversely related to its
intent to avoid pollution, as evidenced by obeying government
standards, utilizing BAT and/or taking other precautions not
legally required. This rule implies that companies would incur
some but not all the costs for Cases 2 through 8. (9% of the
subjects assigned less than 100% in Case 1. All of these
subjects assigned fell into the Partial category for the other
cases, so they were included in it.)
Table 2 shows the percentage of respondents in each of the six
groups, as well as the entire sample, who subscribed to each of
the three liability rules. Across the eight cases, only 18
percent of the sample imposed a strict liability rule. Even
among the environmental group, only 21 percent favored this rule.
At the same time, there was even less sympathy for a negligence
rule, with 10 percent of the respondents favoring it (24% in the
economists, who favored it the most). The expert and
environmentalist groups tended to favor the strict rule much more
than the negligence rule, but the other groups did not differ in
this preference. Groups differed significantly in their relative
preference for strict vs. negligence: p=.032). When the
experts and environmentalists were excluded, however, group
differences were no longer significant. By far the most common
pattern of responses across the one-company cases was partial
liability, with 71% of all the respondents favoring this type of
Percentage of subjects in each liability category.
Strict Partial Negligence N of responses
Experts 31 62 7 (29)
Envir. 21 75 4 (84)
Econ. 20 56 24 (54)
CEOs 18 60 22 (60)
Legis. 15 77 8 (95)
Judges 12 81 7 (86)
18 71 11 (408)
Specific comparisons of cases shed light on the determinants
of cost assignments. Groups of subjects did not differ
significantly in the extent to which they showed these effects
(as indexed by differences between cases), unless differences are
reported. The following effects were found (or not found):
A company should pay less in cleanup costs if it follows
government standards, and less still if it exceeds standards.
Two thirds of the subjects assigned less to the Standards case
(Case 2) than to the Negligence case (Case 1); 33% assigned the
same amount; none assigned more (p=.000, sign test). Subjects
also assigned less to the company that uses BAT (Case 6) than the
one that simply follows government standards (Case 2). 25% of
the subjects assigned no cost to the company in BAT (vs. 15% in
Standards) and 27% assigned it the full cost (vs. 35% in
Standards). Over 45% of the subjects penalized the company less
in Case 6 than Case 2, vs. 6% who penalized the company more for
adopting BAT (presumably because they had forgotten their
response to Case 2). Groups differed significantly (p=.000):
the majority of judges, environmentalists, and legislators
assigned less liability when the company used BAT; but only a
minority of the other groups did so. Subjects were still more
lenient in Case 8 (Innocent) than Case 6 (p=.000, sign test),
although 24% still assigned the full cost to the company.
Groups differed (p=.020), with
environmentalists and experts being more inclined to make the
The effect of penalties on future company actions has
little effect on cleanup cost allocations. Over 75% of the
subjects (317 out of 410) made no distinction in their cleanup
cost allocation between the situation where the company was more
likely to adopt BAT (Case 4) and the situation where it was more
likely to stop producing a beneficial product (Case 5).
Fifty-nine subjects assigned less of the burden to the company in
Case 5 than in Case 4, and 9 did the reverse. Although this
difference was statistically significant (p=.000, sign test),
most subjects were insensitive to the effect of penalties on
future company actions. This result is consistent with the
findings of Baron and Ritov (1992).
No effects of profitability or hindsight were found.
Cases 2 and 3 (Profit) did not differ significantly: failure to
profit from the waste did not reduce the company's
responsibility. Cases 6 and 7 (Hindsight) also did not differ:
discovery of a new technology did not increase responsibility.
IV. JOINT AND SEVERAL LIABILITY: THE TWO-COMPANY SCENARIO
In four cases, subjects had to allocate costs between two
companies and the government. The use of two parties, both of
whom controlled the disposal site is a convenience that allows us
to focus on some of the underlying principles of allocation.
Psychological studies of equity (see Mellers & Baron, in press)
have examined the factors linking cost allocations to different
`input' variables. Wilkerson and Church (1989) suggest that the
proportion of cleanup costs assigned to different PRPs will
depend on the dominant rationale guiding the Superfund program.
For example, a deep-pocket rationale would imply that large
companies ought to pay more than smaller companies, regardless of
damage caused. A proportionality rationale, on the other hand,
might allocate costs on the basis of the proportion of total
hazardous waste that the PRP had deposited, assuming that the
firm would still be solvent after settlement.
A. Nature of Cases
The background information which provided an introduction to
the cases went as follows: `In the next three cases, it has been
discovered that there is groundwater contamination around a
landfill owned jointly and used only by two companies, Large Co.
and Small Co. Both companies generated the same kind of
hazardous waste. The Environmental Protection Agency has
determined that the site needs to be cleaned up. Small Co. has
about a tenth of the assets of Large Co.' After each case,
subjects were asked to allocate costs among the two companies and
the government. They were then asked to allocate costs on the
assumption that the government would pay nothing.
Case 1 (Size): `Both companies disposed of their
wastes in a landfill in the 1960s, before any legislation on
hazardous waste management was passed. Neither company treated
its wastes before disposal. Both companies thought that the
waste was not harmful, and both were in compliance with existing
standards. The proportion of the waste dumped by each
company cannot be determined.' If individuals were distributing
cleanup costs to the two companies on the basis of size, then one
would expect that approximately 90% of the company expenditures
would be allocated to Large Co. In the absence of other
information, we might expect size to be a good indicator of both
the ability of the company to pay and the amount of waste
Case 2 (Amount of waste): This case was identical to
Case 1 except that the last sentence was replaced with:
`Small Co. dumped 75% of the waste, and Large Co. dumped 25%
of the waste.' If the amount of waste is a determining factor
in determining cost allocation then Small Co. would pay 3/4 of
the cleanup expenses assigned to industry.
Case 3 (Orphan firm): The case was identical to Case 2
except that the last sentence was replaced with: `Small Co.
which dumped 75% of the waste, has since gone out of business and
cannot pay anything.' This case was designed for subjects to
determine who should pay for the shares of companies that are out
of business. In the `government can't pay' question, subjects
were simply asked whether they found it acceptable for Large Co.
to pay the full damage.
Case 4 (Harm): This case was identical to Case 2
except that the waste dumped by Small Co. was harmless. The
waste dumped by Large Co. was still harmful.
B. Factors Determining Cost Allocations
Mean percent cost assignments to Large Co., Small Co., and
government in the two-company cases. (NA = question not asked.
1 2 3 4
Size Amount Orphan Harm
Govt. can pay
Large Co. 39 21 34 52
Small Co. 17 37 NA 4
Govt. 44 42 66 43
Govt. cannot pay
Large Co. 70 35 NA 90
Small Co. 30 65 NA 10
Table 3 shows the average percentages allocated to the large
and small company. The main results were as follows:
Company assets influence allocation of costs. In Case
1 (Size), where the amount of waste that each company deposited
was unknown, subjects took size into account, allocating an
average of 39% to Large Co. vs. 17% to Small Co. when the
government paid. When the government did not pay, the
allocations were 70% vs. 30%, respectively. 73% of the subjects
assigned more of the cost to Large Co. than to Small Co.; none
did the reverse (p=.000, sign test). Some subjects (39% when the
government could pay, 32% when it could not) allocated costs in
proportion to size (i.e., 10 times as much to Large Co. as to
Small Co.). However, 27% of the subjects assigned equal amounts
to the two companies whether or not the government could pay),
despite the size differences.
Amount of waste deposited influences cost allocation.
As shown in Table 3, subjects allocate a larger average cost to
Small Co. than to Large Co. for the situation where the smaller
company deposited 75% of the waste (Case 2). Only 2 subjects
(0.6%) allocated the costs in proportion to company size, whether
or not the government paid any share. The majority of subjects
allocated costs according to the proportion of waste deposited
(54% when the government could pay, 67% when it could not).
Relatively few subjects allocated the costs equally between the
companies (12% when the government could pay, 8% when it could
Interestingly, many subjects regarded equal shares as a
better heuristic for assigning cost when they were uncertain
about the proportion of waste dumped by each company (Case 1)
than when they had this information (Case 2). This behavior is
surprising, since the size of the company should be a proxy for
the amount of waste dumped (if it is not explicitly specified) as
well as for the firm's ability to pay for cleanup.
The government should assume all or most of the orphan
share of companies who have deposited hazardous waste. When
Small Co. has gone out of business (Case 3) then, according to
the doctrine of JSL, Large Co. is forced to pay the orphan share
unless the government takes all or part of it over. When
contrasted with Case 2 (which is identical in all respects except
for the insolvency of Small Co.), the average cost share paid by
Large Co. increases from 21% to 34%, but so does the government's
share (42% to 66%) as shown in Table 3. Only 26% of the subjects
subscribed to the JSL doctrine by indicating that Large Co.
should bear the full cost burden. At the other extreme 36% felt
that the government should pay all the costs that would have been
incurred by Small Co. had they still been in business.
Subjects were also asked whether they would find it acceptable
for Large Co. to pay Small Co.'s entire share if the government
would not pay, as dictated by JSL. 57% of those who answered
this question said that this was `unacceptable even if it is the
only option.' The rest found it `acceptable' (12%) or
`acceptable if it is the only option' (31%). Groups differed
here (p=.001): judges and environmentalists found it more
acceptable for Large Co. to bear the full share, and CEOs found
it less acceptable. But 43% of the environmentalists (the most
tolerant group) still found it `unacceptable even if it is the
Companies should pay less for cleanup when waste is
harmless. Case 4 (Harm) concerned the harm caused by the waste.
When Small Co.'s waste was harmless, its average cost allocation
dropped from 37% (Case 2) to 4%, as shown in Table 3. Since
Large Co.'s waste was harmful, it is not surprising that its
share rose from 21% to 52%. 84% of the sample felt that Small
Co. should pay nothing for cleanup if it had deposited harmless
V. RELATIVE IMPORTANCE OF CLEANUP
A. Nature of Cases
The final major part of the questionnaire focused on whether
certain waste sites should be cleaned up. Subjects were given
the following four cases:
Case 1 (Harmless waste, company pays): `A
well-established manufacturing company disposed of hazardous
waste in a landfill owned and used only by this company in the
1960s, before any legislation on hazardous waste management was
passed. It did not treat its wastes before disposal. The site
is far from any human settlements, and there is no chance of it
being developed. Further, geological conditions prevent the
waste from leaking into groundwater that anyone uses. Hence, the
waste is completely harmless to people.' Subjects were
asked, `Should the company be made to clean up the waste?' The
possible answers were: `Yes, regardless of the cost.'; `Yes, but
only if the cost is reasonable.'; and `No, it is not worth
spending money to clean up the waste.' The first answer suggests
an unwillingness to consider other possible uses of the funds
spent on cleanup.
Case 2 (Harmless waste, government pays): `Suppose
that the company in Case 1 was out of business and could not pay.
How much should the government pay?' Subjects were given the
same options as in Case 1.
Case 3 (Own waste): `Suppose that there are two sites.
The waste in the first site was deposited solely by a
well-established company beginning in 1983, well after hazardous
waste legislation was passed. The company failed to meet
government standards for hazardous waste treatment and disposal.
The site is far from any human settlements, and there is no
chance of it being developed. Further, geological conditions
prevent the waste from leaking into groundwater that anyone uses.
Hence, the waste is completely harmless to people.
'The second site is on the outskirts of a fast-growing town.
At this site, the identical waste was deposited at the second
site by a second company, which is no longer in business and
cannot pay for cleanup. The second site can be cleaned up for
the same amount of money as the first site. The second site
is dangerous to human health.`
Subjects were asked to rank the following options:
1. The first company should not be made to do anything.
2. The first company should be made to clean up the first site.
3. The first company should be made to clean up the second
A ranking of Option 2 higher than Option 3 suggests an 'own
waste` bias. This case was inspired by the Exxon Valdez oil
spill, in which it was argued that the money spent on attempting
to clean up the spill would have done more good for the
environment if it had been used in other ways. The own-waste
bias is thus an opinion that companies should clean up their own
waste even if the money could be more effectively spent on other
waste. Of course, the idea that people are responsible for the
effects of their own behavior is built into law and custom, and
the idea of using penalties for other purposes is a fairly novel
Case 4 (Natural): 'The groundwater and drinking water
in two communities are contaminated by a chemical called XYZ,
which causes cancer. XYZ is sometimes found in the ground
naturally and sometimes as a component of hazardous waste
deposited in the ground. The level of XYZ is the same in both
communities, but the cause of the XYZ is different:
`Community A: XYZ got into the groundwater because a
company deposited it in a landfill. This company disposed of
all of its waste in a landfill in the 1960s, before any
legislation on hazardous waste management was passed, and it did
not treat its wastes before disposal. The company has since gone
out of business. It cannot pay anything.
'Community B: XYZ got into the groundwater through a
natural process. It was not produced by any human action.`
Subjects were asked which community was more important to clean
up. They had the option of responding that they were equally
This case was inspired by the comparison between the
cost-effectiveness of hazardous waste cleanup and remediation of
indoor radon. It has been argued that the health benefits per
dollar are higher for radon reduction than for cleanup of
hazardous waste, even though considerably more funds have been
allocated for cleaning up waste (Environmental Protection Agency,
1990). Spranca (1991), Baron (in press a), and Baron and Ritov
(in press) found a tendency to give nature the benefit of the
doubt in a variety of judgment tasks.
Case 5: (Zero risk bias). 'Two cities have landfills
that affect the groundwater in the area. The larger city has
2,000,000 people, and the smaller city has 1,000,000. Leakage
from the landfill in the larger city will cause 8 cases of cancer
per year. Leakage from the landfill in the smaller city will
cause 4 cases of cancer per year. Funds for cleanup are limited.
The following options are available:
`1. Partially clean up both landfills. The number of
cancer cases would be reduced from 8 to 4 cases in the larger
city and from 4 to 2 cases in the smaller city.
'2. Totally clean up the landfill in the smaller city
and partially clean up the landfill in the larger city. The
number of cancer cases would be reduced from 8 to 6 cases [7 for
half the subjects] in the larger city and from 4 to 0 cases in
the smaller city.
`3. Concentrate on the landfill in the larger city, but
partially clean up the landfill in the smaller city. The number
of cancer cases would be reduced to from 8 to 3 cases in the
larger city and from 4 to 3 cases in the smaller city.' Subjects
were asked to rank these options. Subjects who favored zero risk
should favor Option 2. If this bias is strong enough, some
subjects will favor this option even when the total reduction in
risk is less than it is for the other options. Viscusi et al.
(1987) and Ritov et al. (in press) have found evidence for such a
`zero risk' bias.
The questions on importance of cleanup were designed to
look for biases, that is, decision making practices that favor
options that do not seem to produce the best overall
consequences. The main results were:
A few people say they would pay any price to clean up
harmless waste. In Case 1 (company), 17% of the subjects
said that the waste should be cleaned up regardless of the cost;
in Case 2, where government pays, only 14% felt the waste should
be cleaned up. Groups also differed significantly: (p=.000 for
both). Environmentalists were most likely to endorse this answer
[40% for Case 1, 28% for Case 2]; experts were least likely
[3% for Case 1, 0% for Case 2]. In sum, many subjects seem
unwilling to consider other uses of money spent on cleanup, or,
at least, they are unwilling to admit doing so. (The case
specified that the waste is `completely harmless to people,' so a
subject might feel that the cleanup is worthwhile because of the
consequences for animals. At least for this reason, we cannot
assume that all responses are biased.)
People exhibit an own waste bias. The critical
comparison in Case 3 (Own waste) was the ranking of two options:
cleaning of harmless waste deposited by the solvent company or
cleaning of harmful waste deposited by a defunct company. Those
who ranked the first option higher are said to display own-waste
bias. 51% of the subjects showed this bias. To our knowledge,
no previous study has demonstrated the use of this type of
decision rule. We note again, however, that this bias is built
into the law, possibly for good reason.
Few people exhibit a naturalness bias. Over 90%
treated natural and human-caused waste equally. These
individuals did not exhibit a naturalness bias.
Approximately 8% of the subjects favored cleaning up the
company's waste; only 1% favored the natural waste (p=.000 by
sign test for the difference). The small number of subjects
showing the effect suggests that it is not important in practical
contexts. However, we took pains to make the effect obvious; in
real cases, naturalness bias might play a more important role.
People exhibit a zero-risk bias. Zero-risk bias is
defined as a preference for cleaning up a waste site completely,
even if the total number of lives saved as a result of this
strategy is lower than the number that would be saved if the
identical resources had been expended toward cleaning up two
different waste sites partially. Two versions of Case 5 (given
to different subjects) tested this bias. In both versions,
partial cleanup of both sites saved 6 cases of cancer. Complete
cleanup of one site combined with partial cleanup of the other
saved 6 cancer cases in Version 1 (the same as the other options)
and 5 cases in Version 2 (one less than the other options). In
Version 2, the zero-risk option is worse than the other two
Zero-risk bias in Version 1 was defined as ranking Option 2
(complete-reduction) as the best of the three alternative
choices. 18% of the subjects favored this option over the other
two. In Version 2, the bias was defined as not ranking the
complete-reduction option as the worst of the three options. (It
should be ranked worst because it saves fewer cancer cases.) 42%
of the subjects exhibited this `zero-risk' bias. (11% ranked the
complete-reduction option as best. The bias was related to group
in both versions [p < .01]: legislators, judges,
and environmentalists exhibited the bias most frequently, while
experts and economists showed it least.)
One of the criticisms that can be levied against the
Superfund program is that its strategy of requiring complete
cleanups at sites results in an inefficient utilization of
cleanup resources. The zero risk bias suggests a psychological
basis for this inefficient strategy.
Given their different interests, we expected to find large
differences among the six groups of subjects. In fact, very few
differences were found. Specifically, there were no
statistically significant differences in the effect of the
following factors on cost allocations: whether the company
followed standards; incentive effects through penalties; the
invention of a new technology; the size of the company; the
amount of waste; and the harmlessness of waste (for purposes of
cost allocation). The percentages of subjects who exhibited the
own-waste bias and the naturalness bias were also similar across
all six groups.
Those group differences that were found (previously noted)
suggest that environmentalists and experts tended to be more
inclined to clean up waste and more inclined to make companies
pay for it. (The attitude of these groups may perhaps be
explained by their inclination to support the current law. The
questionnaire did give subjects the option to indicate why they
responded to different cases, but subjects gave few such
explanations.) In addition, economists and experts were less
inclined to display zero-risk bias, as one would expect on the
basis of their professional training. (Many of the experts had
economic or statistical training.)
We were interested in whether classifying the subjects as a
function of their attitude toward environmentalism would show
sharper differences. Two questions near the end of the survey
asked subjects: `Do you consider yourself an environmentalist?'
(312 yes, 81 no) and `Are you a member of any environmental
organization?' (152 yes, 242 no). Of the 152 who answered yes to
the second question, all but 6 answered yes to the first as well.
Our environmentalism measure was the number of yes answers given
to the two questions together. Those who answered yes to both
questions were given a score of 2 while those who answered no to
both questions were given a score of 0. The mean
environmentalism scores of the six groups were (in rank order):
environmentalists (1.72); experts (1.31); legislators (1.15);
economists (1.08); judges (0.94); and CEOs (0.83). Although the
group differences are highly significant statistically, it is
apparent that environmentalists can be found in all groups.
Accordingly, we looked for rank order correlations, as
measured by Kendall's t statistic, between this
environmentalism score and the use of all the guidelines and
decision rules we have described. Although our analyses reported
earlier did not show consistent differences between
environmentalists and other groups in the assignment of costs to
companies, the present analysis found more consistent
correlations between environmentalism and assignment to companies
across the one-company cases. (Environmentalism was
significantly correlated, at p < .01, with greater assignment of
costs to companies in one-company Cases 2, 3, 4, 5, and 8.)
Likewise environmentalism was correlated with greater acceptance
of Large Co. paying Small Co.'s share when the latter had gone
out of business (t = .12, p=.007). (But 47% of those with
environmentalism scores of 2 still found it totally
unacceptable.) And environmentalism was associated with greater
willingness to clean up harmless waste regardless of the cost
(t=.19 for two-company Case 1, .18 for Case 2, p=.000 for
Our results suggest that three provisions of the Superfund
law - strict, retroactive, and joint-and-several liability - run
counter to the intuitions of most subjects in all groups, even
those who are familiar with the current law. Only 18% endorse
strict liability consistently, and only 26% endorse JSL in the
the Orphan case. In place of JSL, most subjects tended to favor
proportional liability based on the amount of waste and its
harmfulness, allocating costs on the basis of company size only
when the amount of waste deposited by each company was unknown.
Most respondents felt that the government should absorb some
or all of the costs if companies have followed standards or
utilized the best available technology. This result suggests
that people would favor a larger role for the government in
bearing cleanup costs.
Many subjects, especially environmentalists, say that waste
should be cleaned up regardless of the cost. Even if this
extreme response is designed to maximize impact, it could lead to
an expenditure of funds on one kind of cleanup that could be put
to more cost-effective use in other activities that would achieve
the same general goals. In addition, many subjects favor the
idea that polluters should clean up their own waste, even if the
money they pay to do this could be used more cost-effectively
elsewhere. A few subjects think that human-caused waste is more
important to clean up than otherwise identical natural chemicals.
And many subjects favor complete cleanup of some sites rather
than greater partial cleanup of several sites at the same or
All these biases together - even in different individuals -
could work to create public support for ambitious cleanup
programs. Given the fact that cleanup is expensive (even putting
aside the costs of litigation), it is prudent to ask whether the
money could be reallocated to achieve the same general goals at a
lower cost. If the public resists such questioning, then
education may be helpful in clarifying the relation between decision making
practices and consequences (see Baron & Brown, 1991).
Acton, J. & Dixon, L. (1992). Superfund and transaction
costs. Santa Monica, CA: The Rand Corporation.
Baron, J. (in press a). Heuristics and biases in equity
judgments: a utilitarian approach. In B. A. Mellers and J. Baron
(Eds.), Psychological perspectives on justice: Theory and
applications. New York: Cambridge University Press.
Baron, J. (in press b). Nonconsequentialist decisions.
Behavioral and Brain Sciences.
Baron, J. & Brown, R. V. (Eds.) (1991). Teaching decision
making to adolescents. Hillsdale, NJ: Erlbaum.
Baron, J., & Ritov, I. (in press). Intuitions about penalties
and compensation in the context of tort law. Journal of Risk
Cheek, L. (1989). Insurability issues associated with cleaning
up inactive waste sites. In H. Kunreuther & M. V. R. Gowda
(Eds.), Integrating insurance and risk management in
hazardous wastes. Boston: Kluwer.
Environmental Protection Agency (1990). Reducing risk:
Setting priorities and strategies for environmental protection.
Washington, DC: U.S. Environmental Protection Agency.
Environmental Protection Agency. (1992). A preliminary
report on the indirect effects of the Superfund program.
Washington, DC: U.S. Environmental Protection Agency, Office of
Policy, Planning, and Evaluation, Office of Policy Analysis.
Fincham, F., & Jaspers, J. (1979). Attribution of responsibility
to the self and other in children and adults. Journal of
Personality and Social Psychology, 37, 1589-1602.
Katzman, M. T. (1988). Joinder of defendants and induced
innovation in environmental torts. Environmental Law, 19,
Kunreuther, H. & Patrick, R. (1991). Managing the risks of
hazardous waste. Environment, 33, 12-15, 31-36.
Mellers, B. A., & Baron, J. (Eds.) (in press). Psychological
perspectives on justice: Theory and applications. New York:
Cambridge University Press.
Noll, R. G. & Krier, J. E. (1990). Some implications of
cognitive psychology for risk regulation. Journal of Legal
Studies, 19, 747-779.
Probst, K., & Portney, P. (1992). Assigning liability for
Superfund cleanups: An analysis of policy options. Washington:
Center for Risk Management, Resources for the Future.
Ritov, I., Baron, J., & Hershey, J. C. (in press). Framing
effects in the evaluation of multiple risk reduction.
Journal of Risk and Uncertainty.
Russell, M., Colglazier, E. W., & Tonn, B. E. (1992). The U.S.
hazardous waste legacy. Environment, 34, 12-15, 34-39.
Schmit, J., Anderson, D., and Oleszczuk, T. (in press). An
analysis of litigation claiming joint and several liability.
Journal of Risk and Insurance.
Shultz, T. R., & Wright, K. (1985). Concepts of negligence and
intention in the assignment of moral responsibility.
Canadian Journal of Behavioral Science, 17, 97-108.
Spranca, M. (1991). The effect of naturalness on
desirability and preference in the domain of foods, or
naturalness bias: A natural bias. Manuscript, Department of
Psychology, University of California, Berkeley.
Thomas, E. A. C., & Parpal, M. (1987). Liability as a function
of plaintiff and defendant fault. Journal of Personality and
Social Psychology, 53, 843-857.
Viscusi, W. K., Magat, W. A., and Huber, J. (1987). An
investigation of the rationality of consumer valuation of
multiple health risks. Rand Journal of Economics, 18,
Wilkerson, W. R., & Church, T. W. (1989). The gorilla in the
closet: Joint and several liability and the cleanup of toxic
waste sites. Law and Policy, 11, 425-448.
File translated from
On 7 Dec 2003, 08:33.