Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 189
8
Conclusions
In valuing risk reduction and human life extension, the substantive
regulatory problems encountered by benefit-cost analysts are often signif-
icant but are far from standardized. They vary in terms of the legal and
administrative context in which they appear, the types of potential threats
they pose to life and health, and the characteristics and availability of infor-
mation about those threats. The family of approaches and techniques we
have labeled benefi~-cost anahysis has not been systematically evaluated for
its application to specific topical and conceptual problems in environmental
health and safety. Moreover, as practiced in regulatory contexts, benefit-
cost analysis is limited by the complex administrative, legal, and political
process that characterizes health and safety policy making. The steering
commmitee, on the basis of its conference papers and discussions, believes
that the current challenge for those who perform or use benefit-cost analysis
is twofold:
· to acknowledge and adequately identify the limited role of benefit-
cost analysis and analysts in the entire process of regulatory policy
formation and enforcement and
to distinguish systematically applications of current risk-control
approaches and techniques that are appropriate for specific types
of regulatory issues.
These themes run through the comments of many conference partic-
ipants, including those with sharply opposing views on specific topics de-
bated in the various sessions. Several scholars, for example, point out that
benefit-cost analysis is a small part of a very large, continuous decision-
making process for health and safety regulation that also encompasses
Congress, many administrative agencies and interest groups, scientists, the
courts, and public perceptions. Some question the ability of those charged
with valuation analyses to convey adequately uncertainties associated with
the underlying scientific information on the risks of pollutants and toxins.
189
OCR for page 190
190
CONCLUSIONS
Others focus on the need for decision makers and analysts to consider a
broader range of factors, including the concerns of future generations, in
current analyses and decisions. Still others express a disbelief in the ability
of any analysis that depends on trade-offs at the margin to reach what are
considered by them to be absolute goals, such as health and environmental
protection. Some practicing benefit-cost analysts wonder if they are asked
to do too much with a set of tools that does not yet enable them to meet
those expectations.
Taking into account the range of perspectives present at the conference,
the steering committee noted several problems related directly to the role of
analysis and analysts and the need to distinguish appropriate applications.
For example, there is serious disagreement among scholars, analysts, and
others familiar with the use of benefit-cost analysis as to whether current
approaches to characterizing and valuing risks can or should be asked to
accommodate the full range of factors that decision makers must implicitly
take into account, especially for those problems involving intertemporal
or purely qualitative comparisons. Even those who would support what
they believe to be appropriate use of benefit-cost analysis in environmental
policy making express the view that current methods can be employed more
effectively in evaluating some types of issues than others- for example,
employment effects compared with quality-of-life effects. In particular,
there may be an irreducible tension between a desire to reduce value
considerations to a single metric and a desire for symbolic protective action
in health and safety regulation.
A related problem is the one of expectations versus constraints in
doing benefit-cost analysis for regulatory policy making. For any area
of regulation, legal, administrative, and political factors can set severe
constraints on the conduct of analysis, as well as on the weight given to
analytic results. In contrast, there are incentives, some in response to these
constraints, for analysis and analysts to broaden their scope to show that
all relevant factors have been treated.
Consequently, benefit-cost analysis cannot now be considered to be a
formal decision-making mechanism accounting, for example, for the need
for symbolic action or the full range of qualitative costs and benefits asso-
ciated with policy alternatives. Benefit-cost analysis is more appropriately
used in conjunction with other factors as a set of information-gathering and
organizing tools that may be used to support both decision making and the
presentation of information to the public.
Therefore, we believe there is a need to develop commonsense criteria
for applying current methods to problems encountered in health and safety
regulation, by defining in terms helpful to both analysts and critics alike
the limits to analysis and by developing systematic application of analytic
approaches and techniques to appropriately matched policy issues. Of
OCR for page 191
CONCLUSIONS
191
great interest to analysts and decision makers is the practical impact of
these issues on their work. In particular, conference participants returned
repeatedly to the implications of considering benefit-cost analysis as a set of
approaches intended to clarify and simplify certain parts of difficult prob-
lems raised in the course of environmental, health, and safeW regulation.
Much discussion focused on the current and potential roles for agency staff
analysts in increasing understanding among decision makers and the public
of the costs and benefits of regulation, including dilemmas and trade-oRs
inherent in real decisions.
Expression of these concerns, as well as suggestions for dealing with
them, fall largely into five categories reflecting the most important aspects
of the analytical process for environmental and other regulatory decisions:
the administrative and legal context in which analysis is done,
including its purpose;
· the overall approach that is or should be taken in benefit-cost
analysis;
issues related to the specific analytic procedures and techniques;
· the adequacy of underlying scientific information on risk (e.g., dose
response or exposure); and
· the way in which decision makers incorporate risk assessments and
analyses intended to clarify or support policy making.
Conference participants were asked to identify opportunities for im-
proving risk-control analyses by practical suggestions for analysts and deci-
sion makers at EPA and other regulatory agencies charged with integrating
the full range of intangible factors bearing on health and safety regulation.
In considering this question, they were asked to define the limits of formal
risk-control approaches and techniques.
Participants were not expected to nor did they reach consensus on any
recommendation for a single set of analytical procedures or on a single
appropriate role for analysis and analysts in the regulatory policy process.
The cross-cutting concern for most participants remains the improvement
of benefit-cost analysis as a practical set of tools for policy making. We
believe that improving analysis involves careful consideration of underlying
assumptions, concepts, and methods. It also involves assignment of ap-
propriate types and levels of responsibilities to analysts (and to decision
makers) in light of the analytical capabilities available at present.
THE CONTEXT OF DECISION MAKING
No actions within organizations take place in a vacuum the context
limits the possibilities or even favors the opportunities available to the
OCR for page 192
lg2
CONCLU3~0NS
analyst or the decision maker. These forces originate both outside the or-
ganization (e.g., in statutory pre- or proscriptions) and inside (e.g., in policy
priorities). In terms of environmental decisions by the federal government,
the interplay between the departments or agencies and the Congress con-
stitutes one important set of external constraints. Statutes and legislative
history, which vary considerably between agencies (and sometimes between
programs within agencies) in terms of their approach to benefit-cost anal-
ysis and other valuation techniques, constitute another set of constraints.
The reaction of the courts to the uses of valuation approaches in support
of environmental regulation decisions is another, the activities of environ-
mental interest groups and other organizations represent still another, and
the particular policy goals of agency leaders are another. In his paper on
the contextual setting for benefit-cost analysis, Melnick details the sources
of several significant constraints and their major effects on the conduct
and use of benefit-cost analysis for environmental decisions. In particular,
the adversarial tone associated with policy making for much of health and
safety regulation has the effect of providing extremely strong, but rarely
singular, messages to benefit-cost analysts regarding how analysis should be
done and used.
Statutes, court decisions, congressional intentions, and executive
branch directives provide multiple, sometimes conflicting, guidance to
agency decision makers and analysts on whether or how to include benefit-
cost considerations in risk-management activities. Many key environmental
health statutes operate with admonishments to "protect health" or "use best
available control technology (BACT)," rather than offering a calculus that,
where appropriate, includes trade-offs and cost-based alternatives. Ken as
a whole, judicial rulings and opinions have likewise exhibited inconsistency
across agencies, statutes, and even specific policy arenas. Congressional
attitudes are often negative toward valuation approaches, while OMB has
pressured decision makers and analysts to more fully integrate such ap-
proaches into risk management.
Regulatory analysis reflects these characteristics, in that many of its
underlying assumptions are constrained by these legal, administrative, and
political features. For example, an analysis of fugitive arsenic emissions
from copper smelting would vary considerably depending on whether it
is deemed important to preserve the industry in question (see the ap-
pendix, which takes up this question). It is clear that few expect federal
health, safety, and environmental regulation to become more consistent and
consensual, either legally, bureaucratically, or politically, without major leg-
islative intervention. The most compelling questions concern prospects for
major intervention and how environmental analysis might be done in light
of these characteristics. For example:
OCR for page 193
CONCLUSIONS
193
· How can benefit-cost analysis be integrated into statute-based reg-
ulatory management practices that present a wide range of stances
toward analysis (e.g., BACT versus health-based standards)?
Can we expect benefit-cost analysis to become more politically
feasible or acceptable?
In response to these questions, there were two types of action-oriented
approaches expressed at the conference, one focusing on changing the
environment for benefit-cost analysis and the other focusing on improving
analysis in response to key criticisms of its assumptions and techniques.
In particular, participants discussed the need and prospects for (1) creat-
ing a more consistent and conducive context for valuation approaches by
amending the key environmental health statutes and (2) modifying analyti-
cal techniques and/or including additional variables in valuation approaches
in order to generate support for the results of benefit-cost analysis among
the public agencies, the courts, and Congress.
Achieving improved statutory and judicial consistency and encourage-
ment for the use of valuation approaches in health and safety decisions
would most likely rest on three things. First, general agreement among
policy makers would be needed that there is a compelling need to com-
pare gains in environmental health with the cost of achieving those gains.
Second, agreement would be needed that previous improvements in the
techniques of benefit-cost analysis now enable sophisticated assessments
of variables, such as human life extension, temporal effects, and other
important issues. Third, acknowledgment would be needed that benefit-
cost analysis is or can be decoupled from political motivations, such as
the perception that valuation approaches may support or justify policies
favoring industry or other special groups to the detriment of the wider
public. However, agreement on the need for cost reduction and/or greater
efficiency in protective regulation clashes with concerns that efficiency can
be used inappropriately as a rhetorical device justifying efforts to remove
or reduce needed regulations regardless of their allocative implications.
While benefit-cost analysis cannot now be considered to be a single
set of valuation approaches, techniques, and procedures, there is a core of
common understanding among many analysts based on shared assumptions
regarding expected utility, the implications of scarce resources, and other
concepts. Indeed, explicit techniques for valuing risks, costs, and benefits
have existed for several decades. However, one effect of multiple, conflict-
ing guidance from the regulatory context- even for a single agency is that
regulatory requirements and practices for the use of valuation approaches
are accompanied by political and policy forces favoring or opposing their
application. Consequently, the burden to develop criteria matching appro-
priate approaches and valuation techniques to specific types of problems
OCR for page 194
194
CONCLUSIONS
may be too large. For example, many believe that the use of benefit-cost
analysis is not consistent with the Clean Air Act (benefits can be examined
but not costs) and the Clean Water Act (analysis of costs but not bene-
fits is allowed), and yet benefit-risk analyses are required by the Federal
Insecticide, Fungicide and Rodenticide Act.
There are at least two different sets of perspectives on these questions.
One view is that the crux of the battle over acceptability of benefit-cost
analysis is a debate about how much power should be given to those who
favor and perform such analyses. According to this view, the battle rages
between Congress and the regulatory agencies, as well as between groups
residing within the regulatory agencies. Another view reflecting genuine
puzzlement in the case of environmental trade-offs sees polarization and
controversy emerging from institutions (e.g., Congress, the bureaucracy)
that are considered on most subjects to be moderating and consensus
building.
For both views, it may be that the use of benefit-cost analysis for
environmental health touches on basic unresolved differences regarding the
appropriateness of applying cost considerations to certain classes of public
goods, such as those described in the papers by Railton and MacLean. If
so, these issues must be addressed at the level of institutions (i.e., Congress,
the President, and the courts), and possibly through the conduct of further
research, before specific guidance can be supplied to regulatory decision
makers and analysts. Such a resolution might emerge over the long term
or as the result of a major environmental or economic crisis. But, in the
absence of a crisis, public opinion and institutions might continue to have
difficulty integrating fully the protection of human health and safety and the
consideration of costs and benefits in allocating scarce economic and other
resources. Signals from the legal and administrative context, furthermore,
suggest that agencies may continue to find both strong supporters and
strong critics of attempts to enunciate agency-specific or government-wide
approaches to risk-control analysis for regulation.
On the second question- whether future improvements in analytic
techniques, approaches, and theories could prove compelling to the policy-
making system and therefore generate additional support for the consistent
use of valuation approaches support for (or opposition to) benefit-cost
analysis may not always be based on assessments of technical features.
Therefore the contextual forces for regulatory benefit-cost analysis may
not respond to technical improvements with greater support for its use in
regulatory decision making.
Nevertheless, it was clear to many participants that any advances in the
application of evaluation approaches must be grounded solidly in the quality
of the analysis and its ability to speak to values considered important by
citizens and policy makers, regardless of whether they are easily discussed
OCR for page 195
CONCLUSIONS
195
analytically. Consequently, much of the conference focused on ways to
account for these additional variables and questions in valuation approaches
used in benefit-cost analysis.
APPROACH
The term approach refers here to the way issues are conceptualized in
terms of a general framework for analysis. Benefit-cost analysts, as Melnick
notes, typically see policy issues in terms of opportunity costs, incentives,
and the expense associated with eliminating risks. He characterizes lawyers,
in contrast, as more likely to give preference to concepts of "command and
control," of establishing standards and penalties to be imposed if those
standards are not met. At least some environmentalists and politicians, he
says, are more likely to see the issues in terms of an absolutist, health-only
stance for which no level of effort is too great. 1b the extent that these
characterizations are accurate, they influence the way analytic studies of
valuation are viewed.
In order to agree on an approach to valuing health effects, we would
want it to pass three tests: (1) accuracy, (2) verifiability, and (3) accept-
ability or feasibility. Much of the technical work to advance health-effect
valuations has focused on improving the accuracy and verifiability of the
scientific data and benefit-cost techniques. MacLean, however, describes
some of the reasons why this third test presents special difficulties. He
points out that there are substantial differences in the positions favored by
different groups, differences that are difficult to resolve through appeals to
either basic moral principles or empirical preferences. MacLean describes
what for many makes the issue absolutely unique: human life is sacred, a
fact with many ramifications. For example, although few disagree that it
makes sense to discount expenditures or the opportunity costs of health
effects, MacLean claims that there are moral and logical difficulties when
the value of life per se is discounted. He argues that the existence of eco-
nomic consequences of health and safety decisions does not justify treating
human life as if it were exchangeable.
There remain unavoidable comparisons, however, of how many re-
sources should be devoted to improved physical well-being compared with
energy conservation, greater economic growth, or other outcomes. One
view is that such comparisons can be made more efficiently by relying on
a common metric to characterize all important possibilities rather than
examining the relative value of risk reduction in each case. Yet even within
the community of professional analysts, there is disagreement about the
best way of doing so. For example:
· Where are improvements most needed in valuation approaches?
OCR for page 196
196
CONCLUSIONS
.
Are there viable alternatives to formal valuation approaches to
support environmental decisions?
As the above discussion of contextual issues indicates, a major chal-
lenge for analysts is to develop and employ appropriate approaches to
analysis when there is disagreement regarding basic policy problems and
goals. For example, some have proposed that air and water pollution could
be reduced more rapidly by making standards less stringent. The reasoning
is that, in some cases, current standards so far exceed current attainment
levels that there is little chance additional efforts will lead to complete
compliance. Therefore, the argument goes, fewer such efforts are under-
taken than would be if standards were set at a lower level. This reasoning
is rejected by those who believe that high standards should be met. Still
others would acknowledge the difficulty of meeting some current standards,
but hold that they act as a necessary spur to obtaining even the most modest
pollution reduction results. In this context, how does the analyst choose
between methods of analysis appropriate to obtaining relative reductions
or to meeting the more absolute goals?
In addition, some analysts and observers are concerned about the ac-
cessibility of analytic approaches. Many accept that an important attribute
of an ideal approach should be that it generate results people can un-
derstand and can analyze and argue about. In this sense, Railton refers
to benefit-cost analysis as an information-yielding device rather than a
decision-making tool. There remains, of course, the question of accessibil-
ity and training. Results easily understood by another benefit-cost analyst
may not be fathomable by an interested, educated, but untrained person
or even by environmental scientists trained in other disciplines.
Machina presents what might be considered by some to be an unusual
position with respect to the various parts of the analytical process that
need adjustment. He argues, for example, that recent theoretical and em-
pirical findings pertaining to how people actually evaluate risks and assign
probabilities ought to be reflected in the way preferences are analyzed,
but not in the analysis of probable outcomes. As he states it, observed
"departures from the strictures of probability theory should be corrected
but . . . (systematic) departures from the strictures of expected utility theory
should not". This is because the former involve the determination of the
risks associated with alternative actions or policies, which are in fact mat-
ters of accurate representation, while the latter involve the willingness of
individuals, organizations, and society to bear these risks, which is a matter
of preference. He concludes that analysis must be designed to account for
actual preferences, even those that depart from the strictures of expected
utility theory.
OCR for page 197
CONCLUSIONS
197
Considerable support for this view is emerging among economists
and other risk-control analysts. However, there appears to be little or
no consensus at present regarding specific valuation approaches that need
reforming in light of this new evidence on preference formation and little
consensus as to how such approaches, once identified, might be modified to
systematically account for observed departures from expected utility theory.
Furthermore, there is little agreement as to alternatives to formal
valuation approaches in support of environmental decisions. Although
many decry weaknesses in current analytic approaches, few alternatives are
proposed. MacLean discusses several reasons that a single analytic method
for making environmental decisions should not be applied universally.
This is principally because the necessity to involve symbolic elements in
comparisons and trade-offs makes such decisions very dependent on their
context. He concludes that it may be necessary to treat different values
differently and that situational specificity, not measurement difficulty, is
the fundamental problem in such comparisons. As Railton points out, an
approach that yields more information can increase accuracy yet at the
same time be less decisive.
Regardless of the valuation approach used, most participants seemed
to favor greater provision of information, even at the cost of decreased
decisiveness. Still, there is also a recognition of the need to systematically
differentiate analytical-approaches in light of the problems or values they
are meant to address.
PROCEDURE
Procedure, as used here, refers to the way the benefit-cost analyst
actually goes about systematically examining values. The choice of specific
methods, metrics, and measures will play an important role in provid-
ing understandable, believable information for decision makers and other
interested parties in protective regulation. For example, the concept of
willingness to pay (i.e., setting the value of a good or service according to
what people are willing to forgo in order to have that good or service) has
gained adherents and wider use in health and safety regulation in recent
years. When used in hypothetical markets (e.g., environmental decision
making), it is nevertheless a troublesome procedure to those who would
prefer to incorporate the notion of experience in some way, and who would
attempt to discover what people, with enough experience, would want. An-
other possibility would be to balance willingness to pay with a measure of
willingness to sell (how much people would have to receive in order to give
up the good or service), but such a concept also runs into measurement
problems in hypothetical markets.
OCR for page 198
198
CONCLUSIONS
Questions are raised also about the advisability of relying on tort law
to determine the value of intangibles, especially because jury awards appear
to be highly skewed in liability cases. Another possibility is to search for
valuation methods and/or procedures that account for distributional effects,
for example, assigning premiums to consequences based on income. Other
questions include:
· Can balanced procedures, possibly including a combination of val-
uation techniques and nonquantitative factors, be found that would
meet the approval of Congress, OMB, and the courts?
· Is the use of a simple or single metric appropriate when there is
variation among individuals, not just in terms of how much of a
value is wanted or can be tolerated, but in terms of whether they
want that particular value at all?
· Are there ways to capture certain difficult conceptual problems,
such as intergenerational equity, in analytic terms?
In his keynote address in this volume, Russell describes the necessity
of making what he calls "cruel choices." In terms of the accommodation
sometimes needed to reach agreement about such choices, it may seem as
though valuation techniques make trade-offs too explicit. Fuzziness some-
times appears to have a positive function in that it allows the participants
to believe in the myth of successful joint accommodation. Russell argues,
however, that better information needs to be presented to the public, and
that our system is predicated on the legitimacy flowing from the support of
an informed citizenry. Presenting these trade-offs is a critical part of that
process.
One procedure for simplifying, analyzing, and presenting trade-offs is
the transformation of a range of value observations into a single set of
common metrics. In analytic practice, this most often means the use of
monetary value as the common numerator or denominator, or both. Railton
discusses the difficulty of relying on an expert-determined metric when
issues of welfare are involved. Scientists, he says, have more knowledge
about natural phenomena and thus would be more reliable than the rest
of us in describing events. But the most reliable source for describing
people's welfare outcomes may be those individuals themselves. It may
be quite difficult to develop a metric that adequately captures individuals'
perceptions of their well-being. Nevertheless, in practice the regulatory
process does implicitly differentiate certain factors. Some have observed
that regulators are willing for society to expend more to protect individuals
who are involuntarily at risk than to reduce those risks borne by people
who are easily able to make choices regarding their own exposure. This
corresponds, by the way, to many people's preferences concerning a variepr
of risks.
OCR for page 199
CONCLUSIONS
199
Similarly, the problem of comparing well-being across a significant
span of time (e.g., intergenerational comparisons of health) has provoked
considerable controversy about the acceptability of analytic procedures.
There are regulatory actions involving human health and life for which
there is a need to consider costs and benefits that might accrue in the
future. The traditional method- for comparing costs and benefits over
time-discounting" is predicated on the belief that costs borne by and
benefits accruing to individuals and society in the future should be adjusted,
or discounted, to make comparison with current expenditures or costs valid.
In the area of human health and safety, the sacredness of human life, as
MacLean and others characterize it, can make it difficult for policy makers
to agree with methods that seem to say that a life saved today should be
worth more than a life saved 50 years from now. On the other side of
the coin, if resources are limited, then some allocation must occur. Under
those circumstances, the hard question is: What is the best way to deal
with policy problems for which actions taken or not taken today will have
human health effects in the distant future?
In answer to these difficulties and questions, there is at present no
broad consensus among those interested in health and safetr regulation.
Some oppose discounting altogether, others would apply discounting to
costs but not benefits, and many would support the use of a very small
discount rate far below that often used in analyzing costs and benefits of
other government and private programs.
It is important to note that although (as previously discussed) benefit-
cost analysis may be only a small component of the entire regulatory
process, analytic procedures must be constructed to account for the major
characteristics of decision making. Harris raises the point that, whereas
the usual way of thinking about analysis is as an input to decision making,
decision making often effectively provides input for subsequent analysis,
and that the best strategy may be to treat the whole process as dynamic.
The factors that ought to be considered may be too complex to model
adequately using current analytic methods, and the only way of developing
appropriate information about outcomes may be to take action that can be
reversed if necessary.
SCIENTIFIC INFORMATION
The potential conflict between risk assessment and risk-control analysis
is often realized. In the schema of key components of regulatory agency
policy making presented in the introduction to this volume (see Figure 1),
risk assessments depend on assessments of data from epidemiologic, bio-
logical, and engineering studies. In turn, benefit-cost analysis depends on
OCR for page 200
200
CONCLUSIONS
the application of valuation analyses to scientific assessments of risks associ-
ated with health and safety problems. Both risk assessment and risk-control
analyses inevitably involve simplification of processes and outcomes in or-
der to achieve clarity, meet regulatory deadlines, or because the underlying
data are insufficient. For example, scientists frequently complain that risk-
control analyses fail to capture adequately the factors they judge important
in scientific determinations, whereas analysts express disappointment at be-
ing prevented from achieving the ideal in risk assessment, accurately and
precisely portraying the risks of concern to policy makers and the public.
This, both claim, is because some factors can be quantified with confidence
while others remain matters of judgment. Scientists, if generalization may
be permitted, tend to distrust the policy analysts as oversimplifiers, and
the analysts are frustrated because the scientists cannot provide clear-cut
answers. Several practical questions are embedded in this caricature:
Is it possible to adequately reflect scientific judgment in analytic
valuation studies?
Can uncertainties or gaps in data be adequately incorporated?
Can conflicting models or interpretations be integrated into analytic
valuation studies?
How can values or policy judgments be integrated into analytic
valuation studies?
The ways different disciplines treat data, uncertainty, and judgment
may underlie these issues. For example; some natural scientists point out
that their colleagues are currently unable to come up with a method for
quantifying,- for example, the carcinogenicity of arsenic that would satisfy
analysts interested in determining the costs and benefits of policy options
for dealing with fugitive emissions associated with copper smelting. These
scientists were greatly disturbed that an enormous body of information,
mechanistic as well as epidemiologic, about the neurotoxicologic effects of
arsenic was ignored in the analysis of the substance as a human carcinogen
because it is judgmental and therefore difficult to express in quantitative
terms.
In his paper on information and regulation, Harris raises a related issue
having to do with the timing and sequence of new scientific information
as it becomes available for benefit-cost analysis. The question is knowing
when to act and when to wait for additional information that might lead
to a different action (including no action). The challenge shared by the
scientist, the analyst, and the decision maker is to construct procedures
for knowing when to recommend action, further information gathering, or
both.
OCR for page 201
CONCLUSIONS
201
Scientists, analysts, and decision makers, however, point out a number
of other issues that follow from an expansion of the concept of risk assess-
ment to include other factors. For example, some emphasize the distinction
between narrow technical uncertainty associated with any scientist's esti-
mate of risk and the broader structural uncertainty associated with differing
estimates of risk by various scientists. Most important, others argue that a
major reason why better scientific estimates of health and safety risks are
not available for example, estimates of the risks of noncarcinogens is
that decision makers and the public have not demanded them. Estimates
of the risks of carcinogens are relatively more available largely because
Congress, the courts, and the public have demanded that resources be
committed to producing those estimates.
Finally, there is a significant argument that risk information is often
inadequate because scientists are reluctant to make estimates in light of
limited data. This may be a major contributing factor in the conflict between
analysis and scientific risk assessment. Benefit-cost analysis is often driven
by the need to make a decision, while those who perform scientific risk
assessments are reluctant to guess when information is deemed inadequate.
It may be that scientists will always be distrustful of benefit-cost anal-
yses, which, in order to reduce, simplify, and meet tight schedules, rely
on summary characterizations rather than detailed information and accept
increased level of uncertainty and a reduced level of subtlety in under-
standing any single piece of the whole picture. There is a conflict, however,
between presenting complete information that is so detailed and complex
that only the experts can understand it, and simplifying presentations so
that those without the technical training can understand but lose the detail
and subtlety of the data. There does not appear to be a quick technical
or procedural fix that satisfies both scientific sophistication and simplicity.
But many would insist on making evidence, and especially the assumptions
underlying conclusions, regarding both risks and the valuation of risks as
open and explicit as possible. This would both encourage careful question-
ing of those assumptions and contribute to trust among scientists, analysts,
decision makers, and the public.
DECISION MAKING
A widely shared expectation is that an individual or small group, usually
at the apex of the regulatory agency, holds responsibility for the decisions
and actions of that agency. Another shared expectation is that a policy
decision should be based on the relevant evidence. As indicated previously
in this discussion, the role of the agency analyst, as well as that of the agency
decision makers at the apex, is often limited by a context in which myriad
other voices, both outside and inside the agency, speak strongly. Moreover,
OCR for page 202
202
CONCLUSIONS
these two expectations can conflict, particularly when it is difficult to gain
general agreement on the type and quality of evidence needed to support
a decision or when some important evidence is qualitative, implicit rather
than explicit, or missing altogether.
In this type of setting, agency political appointees, rather than analysts,
are responsible for factoring in a full range of considerations bearing on
a problem in effect, to act on behalf of the agency and the public by
applying an informal metric trading off the economic, political, ethical, and
etiological concerns raised in the foregoing papers and then to produce a
decision reflecting the results.
Within this framework, benefit-cost analysis can be used to provide
more explicit information about some (e.g., economic) priorities and trade-
offs as input to a decision, or it can be used to justify decisions already
reached. It is not intended, however, nor can it give equal emphasis to all
possible considerations bearing on a regulatory decision. Rather, analysis
traditionally focuses on a few factors the decision maker would like to make
explicit or those few for which quantitative values are available.
· What are the prerequisites for analysis that informs rather than
appears to supplant the prerogative of the decision maker?
There is some ambiguity regarding how much authority we wish to
grant to the process of providing information and the experts who provide
it. Both analysts and policy makers have experienced pressure to use
benefit-cost analysis to treat explicitly a broader range of relevant factors.
This includes direct statutory and administrative pressures for agencies
to broaden or to restrict their analytical coverage of values, including
economic, Biological, ethical, and administrative concerns. In addition,
strong criticisms and attacks on risk-control analysis by those mistrustful of
the assumptions, approaches, and results of valuation approaches may play
a role in stimulating theorists and practitioners to increasingly incorporate
and analyze as many issues and aspects bearing on a decision as possible.
The analyst, consequently, faces incentives to include a wider range
of considerations within the analytical framework, considerations that may
be difficult to treat formally (e.g., ethical and political variables). How-
ever, as the number of variables to be addressed increases and as their
character changes (e.g., inclusion of variables that are intangible or those
that raise questions of interpersonal comparisons), the analyst's challenge
is to prevent the problem under consideration, as well as the set of al-
ternative decisions, from becoming fuzzier and less well defined. Some
worry that, as the analyst incorporates an increasing number of factors, in-
cluding those factors not often treated quantitatively, a formal but limited
OCR for page 203
CONCLUSIONS
203
set of approaches and techniques is being substituted for the more infor-
mal approach provided by the politically and legally accountable decision
maker.
Three principal prerequisites are often cited as expectations for analytic
studies: knowledge, candor, and clarity. With regard to the first prereq-
uisite, most studies need three kinds of knowledge: technical knowledge,
economic knowledge, and legal knowledge. There is a lot of uncertainty
in all three areas, and whatever is presented to the decision maker must
reflect the degree of confidence in the knowledge in those three areas.
Second is candor. If the people doing the studies know that there
are some weaknesses or driving assumptions embedded within them, they
should frankly state that to the decision maker. At issue is the generation
of trust among those who experience the effects of regulation, among the
decision makers, and among valuation analysts. Much has been written in
recent years about the decline in trust in government and other institutions.
We need not repeat that discussion here, other than to point out that many
factors affect trust and that it is important for analysts to demonstrate
substantive competence, lack of bias, and openness in their work.
Finally, whatever is presented to the decision maker should be clear.
Studies may be clear to the leading workers in the field and their peers,
but not to decision makers.
The peer review system represents one possible way to promote ac-
curate knowledge, candor, and clarity in the use of benefit-cost analysis in
regulation. Review by h~owledgable experts can help reveal inaccuracies,
weaknesses, driving assumptions, and so on. But it may not be capable of
ferreting out the technical and other jargon that confuses nonexperts. Con-
ference participants pointed out that, although peer review is commonly
used to examine the methods, data, and results of risk assessment, formal
peer reviews of benefit-cost analysis are rarely undertaken within regulatory
agencies.
Few would support applying a mechanical, strict benefit-cost rule to en-
vironmental regulation that is not tempered by judgment and consideration
of other factors. The major role of benefit-cost analysis in environmental
regulation should be as an organizing concept and as a way of helping the
decision maker think about the factors that need to be looked at. A process
involving peer review of benefit-cost analysis could assist decision makers
by malting assumptions clear and by noting the strengths and weaknesses
of results.
Railton's suggestion that benefit-cost approaches be treated as informa-
tion-yielding rather than decision-making devices has several implications
for the way scientific data and judgment could be integrated into decision
processes. Most important, it would suggest disaggregated information and
error bands or ranges rather than point estimates. It also suggests the
OCR for page 204
204
CONCLUSIONS
importance of presenting information about several alternatives. Harris,
moreover, points out that the best strategy may be to undertake actions
that are reversible before complete information is available, because imple-
menting regulatory or other control actions may be the only way to obtain
relevant information. Information generation may be an excellent role for
benefit-cost analysis in such a dynamic process.
CONCLUSIONS
The conference was intended to explore a set of philosophical, political,
informational, and administrative issues in the use of benefit-cost analysis
for environmental decision making. This broad scope yielded an equally
broad mix of ideas, approaches, and conclusions in both the papers and
the conference discussions. The ideas discussed at the conference and
presented in these papers may contain a few seeds that could bear fruit if
they are nurtured and developed. Certainly they include several concepts
and comments that speak directly to the practice of benefit-cost analysis.
Emerging from conference discussions is a strong sense that, at present,
the challenge in improving the application of benefit-cost analysis is to
design practical procedures and techniques that accommodate (1) consid-
erable situational variation; (2) the fairly limited role played by formal
risk-control analysis in the full process of identifying, regulating, and en-
forcing solutions to health and safety problems; and (3) the tendency for
both critics and supporters of analysis to-overemphasize its influence in the
regulatory process.
The problems of health and safety regulation are far from standardized.
Similarly, benefit-cost analysis is really a family of related techniques and
approaches, only a few of which have been systematically evaluated for
their application to specific topical and conceptual issues in health and
safety regulation.
Consequently, the application of benefit-cost analysis must be sup-
ported by better systematic distinctions or situational conditions that reflect
accurately key variations among regulatory responsibilities and problems
facing health and safety decision makers. For example, can we systemat-
ically distinguish among situations involving potential loss of life, sickness
and injury only, or both?
Once a set of systematic distinctions has been drawn, it would then be
important to identify attributes of benefit-cost analysis techniques appro-
priate to those situational conditions. At the simplest level, for example,
are there some problems for which the use of formal analytical techniques
are not currently appropriate? At another level, is it appropriate to use
discounting of future costs and benefits when human life is not at stake?
OCR for page 205
CONCLUSIONS
205
Finally, it is important to formulate processes for encouraging the
development, within the various federal regulatory arenas, of agreement on
systematic scientific distinctions, techniques, and approaches to valuation
of risks, costs, and benefits.
Taking account of the fact that some conference participants generally
support the use of formal analysis whereas others generally oppose its use,
the committee was nevertheless able to reach the following conclusions.
1. Among both those who would general) support the use of benept-cost
analysis and those who would oppose its current use, there is a common
recognition of genuine moral and ethical dilemmas underwing evahtaiion of
the costs and benefits of programs to regulate health and safer nsks.
a. Current approaches to characterizing and valuing risks cannot accom-
modate with validity the ~11 range of factors that decision makers are asked
to take into account, particularly those drawing comparisons across time.
Statutes and administrative orders requiring major regulatory decisions to
be based on least cost/most benefit analyses have, according to one view,
stimulated a new emphasis on cost reductions in regulatory decision mak-
ing. According to another view, such requirements fail to acknowledge that
the general approach to benefit-cost analysis is not developed well enough
to fully account for important factors, such as public opinion, quality of life,
and other social, psychological, and institutional issues important in any
program of regulation. Similarly, considerable disagreement exists about
the ability of risk-control approaches to adequately consider future gener-
ations with current techniques for discounting. Supporters have proposed
careful analysis of long-term impacts of regulatory policies and the use of
nonmonetary measures, such as lives saved/deaths prevented, in calculating
impacts. Critics are concerned that equity cannot be discerned and com-
pared across time in ways that can be formally measured. This suggests a
need for further development in the areas of temporal and nonquantitative
factors as a contribution to improvement in benefit-cost analysis.
b. Serious concerns enlist regarding the appropriateness of formalizing
approaches to issues such as intertemporal equip. At the heart of this
recognition is the belief that there is an irreducible tension between a
desire to reduce value considerations in regulation to a single metric and a
desire for symbolic action. This conflict emerges most clearly with respect
to issues involving human life (preventing deaths/saving lives) in regulatory
decision making. There is a concern that, even if benefit-cost analysis
approaches and techniques could be developed to account systematically
for a broader range of issues on a broader range of regulatory topics, it
may be morally or ethically incorrect to place values such as human life
next to other economic factors.
OCR for page 206
206
CONCLUSIONS
2. Formal benefit-cost amass of health and safety risk in regulation Is
at present on) a limited and incomplete part of a large, complex anemic and
dec~sion-making process. This process largely determines the weight to be
given to benefit-cost analysis in policy making, and it sets limits on what
can and cannot be accomplished with formal evaluation of risks, costs,
and benefits. For example, the character of statutes, regulations, court
decisions, and enforcement can all constrain the assumptions, methods,
and data that can be used in a formal analysis. And, once completed, the
scientific quality of the analysis is only one factor influencing reactions to
the analytical results.
3. There are often misperceptions regarding the influence of benefi~-
cost analysis in decision making. Critics often express the need to curb
the power of benefit-cost analysis in decision making, while supporters
would increase its influence. There are at least two general reasons why
this polarization and possible overemphasis may occur. First, the current
adversarial character of regulation for health and safety issues encourages
some to seek ways to challenge (or defend) outcomes on whatever grounds
may prove successful. Under these circumstances, analysis can become a
target or a pawn of advocates. Second, there may be standards and/or goals
set for analysis that it cannot in all instances meet. In a policy process
in which analysis seems to offer ways of treating difficult issues, there is
a tendency to transform political disputes into scientific or analytical ones
and then to expect the analytical process to substitute for the political
process. The consequence is to expect benefit-cost analysis to be able to
accommodate a wide range of important factors bearing on a problem
and then to be critical when it is more difficult to treat formally some
factors than others. The result can be to politicize benefit-cost analysis and
polarize positions (e.g., those who would increase the use of analysis and
those who would eliminate it in favor of absolute standards, such as the
Delaney Amendment and the Clean Air Act).
4. We believe that, among agency decision makers, the courts, Congress,
and analysts, there is no consensus regarding the use of a specific set of
analytical techniques for a specific purpose. At present there are a range of
techniques available for use in benefit-cost analysis of morbidity, mortality,
and other health and safety ejects. There is agreement among many
practitioners and scholars regarding the importance of considering a wide
range of costs and benefits of alternative policies; there are also, in many
instances, a number of technical procedures and methods that analysts can
use to estimate those costs and benefits. But there is little agreement on
decision rules an agency might use in choosing one procedure or method
over another in specific cases. In the analysis of programs to prevent
deaths- the example is regulation to restrict fugitive arsenic emissions from
copper ore smelting- an agency analyst might choose techniques based on
OCR for page 207
CONCLUSIONS
207
willingness-to-pay. It is well known that this approach emphasizes different
types of costs than would other approaches. For example, techniques based
on willingness-to-pay will emphasize expressed preferences, while those
based on human capital models emphasize forgone earnings. Although
there is considerable debate within some agencies and among some analysts
as to the appropriateness of specific techniques for valuing lives, there is a
need to evaluate both the types of regulatory problems and the attributes
and implications of techniques for benefit-cost analysis, in order to develop
a more systematic basis for deciding when to apply which technique.
Based on these conclusions, the steering committee makes three rec-
ommendations for improving the use of benefit-cost analysis in health and
safety regulation:
1. Benefit-cost analysis should be thought of as a set of information-
gathering and organizing tools that can be used to support decision making
rather than as a decision-making mechanism itself. Beating the approach
in this way has several implications for the way it is applied. It empha-
sizes disaggregated information and range rather than point estimates. It
encourages development of information about several alternatives rather
than a single policy. It also encourages viewing both analysis and decision
making as dynamic rather than static processes.
2. There is a need to more systematically match analytic methods and
techniques to Apes of health and safety problems encountered in the regulatory
process. Researchers and analysts recognize that techniques associated
with benefit-cost analysis emphasize quantifiable factors, especially those
that can be characterized in monetary terms. In addition, benefit-cost
analysis is most applicable to those health and safety problems or elements
of problems in which externalities are relatively limited and opportunity
costs are relatively clear. For example, the economic impacts of closing
a plant may be easier to characterize succinctly than its effects on the
quality of life in the surrounding community. This implies (a) the need
to improve benefit-cost analysis so that a more complete range of impacts
and factors can be considered systematically and (b) the need to identify
systematically sets or types of problems according to their susceptibility to
benefit~ost analysis. Therefore, blanket determination of whether a single
analytic approach should be taken to all regulators decisions does not seem
appropriate at this time.
3. Regulatory agencies should consider expanding the use offormal peer
review mechanisms in the area of benefit-cost analysis for health and safer
decisions. The steering committee believes that benefit-cost analysis should
be subject to systematic, consistent, formal peer review, which can be
used to assess appropriateness of assumptions, techniques, and approaches;
limitations of data and methods; and the formal or informal treatment
OCR for page 208
208
CONCLUSIONS
of moral and ethical concerns. Traditionally, the Office of Management
and Budget has been asked to assess major regulatory benefit-cost analyses
as part of its mission to encourage greater use and improved quality
in analysis, but it is not appropriate to ask OMB to devote its limited
resources to accomplish a full range of review that is necessary for a
complicated valuation. OMB could be asked to respond to plans for
expanding peer review for valuation analyses.- Such an expansion would
complement existing agency peer reviews for scientific risk assessment prior
to agency decision making.
Representative terms from entire chapter:
decision maker
