D. Graham, Ph.D., Administrator
Office of Information and Regulatory Affairs
Office of Management and Budget
Executive Office of the President of the United States
Prepared for The International Society of Regulatory Toxicology
and Pharmacology Precautionary Principle Workshop, Crystal City,
VA, June 20, 2002.
seems that each new year brings new concerns about risks to human
welfare, quality of life, and environmental protection. Most public
concern in the United States relates now to the war against terrorism
and the economic hardship that many citizens continue to face.
There are also continuing concerns about the health and safety
risks that people face from the consumer products, production
methods, and wastes that support the high standard of living we
now enjoy. And there are some broader concerns about what economic
development is doing to the global environment and the welfare
of future generations. Given the explosion of information about
possible risks in life, it is difficult for the public to know
which of the many possible risks justify attention by policy makers.
Bush Administration believes that science should have a strong
role in setting risk-management priorities. Even in the relatively
simple case of health risks, sound science is critical. First,
there is the basic hazard question: What is the degree of certainty
that any hazard exists? For example, how confident are we that
the magnetic fields emitted from electric powerlines cause an
elevation in cancer rates among people living near those powerlines?
Second, if the hazard exists, a probability assessment is required
to distinguish a significant risk from a negligible one. For example,
if magnetic fields do cause cancer, how much is the probability
of cancer increased by a lifetime of exposure to a nearby powerline?
Third, the number of people exposed to a hazard needs to be considered
because population exposures contribute to the public health significance
of the hazard. Fourth, the severity of the health effect is relevant.
Even in the case of a dreaded disease such as leukemia, we need
to be sensitive to the distinctions between those forms of leukemia
that are untreatable and rapidly fatal versus those forms that
can be treated effectively by the tools of modern medicine. In
other words, the number and quality of life years lost from disease
need to be considered. Although these basic scientific questions
have been framed for health risks, it is feasible to frame a related
set of questions for other types of risks, such as threats to
natural resources and global ecology.
of how much scientific knowledge there is, public officials need
to consider the degree of public concern about possible risks
and how that compares with expert assessments. The emerging science
of risk perception and communication is exploring these issues.
How much does the public understand and care about a hazard? How
volatile is public opinion likely to be in the face of new evidence
or dramatic portrayals of danger? Does the hazard have characteristics
that elevate public concern, such as an unfair distribution of
risks and benefits? These social science questions remind us that
the challenge of risk management will not be resolved by hard
science alone. We need to consider the type and degree of public
participation that is appropriate for decisions about risks.
these sweeping questions about risk, what can we possibly hope
to accomplish in this brief conference? I would frame our task
as follows: There is a subset of possible risks in daily life
that are subject to substantial scientific uncertainty - often
on all four of the scientific questions that I mentioned - but
that, for one reason or another, trigger significant public concern.
Under these circumstances, what is the appropriate role for precaution
in the responses of risk managers? I have in mind risk managers
in both the public and private sectors.
do I mean by precaution? I can assure you that I do not intend
to define any universal precautionary principle. As you know,
the US government supports precautionary approaches to risk management,
but we do not recognize any universal precautionary principle.
We consider it to be a mythical concept, perhaps like a unicorn.
do believe that a dictionary definition of "precaution" is a useful
starting point for the conference. WEBSTER'S Second College Edition
of the NEW WORLD DICTIONARY defines precaution as "care taken
beforehand", or more precisely, as "a measure taken beforehand
against possible danger". I presume that the word "beforehand"
means before science has resolved all the key technical questions
about the hazard of interest or at least before the actual occurrence
of the event. Precaution is a well-respected concept: people practice
it regularly in the stock market and in hospitals to cope with
the uncertain risks of life.
Americans have experienced the pain and suffering that can result
from insufficient precaution in risk management. The health risks
of smoking, the neurotoxic effects of low doses of lead, once
used as an additive to gasoline, and the respiratory diseases
from exposure to asbestos in the workplace: each became major
public health problems in the United States. Public health historians
teach us that these problems could have been reduced or even prevented
altogether if early signals of danger had stimulated precautionary
measures by risk managers.
should not belittle the scientific challenges in each of these
examples. Consider tobacco. Although the causal link between smoking
and lung cancer now seems obvious, in the middle of the previous
century it did not seem obvious to many well-trained and thoughtful
physicians. They argued that they had treated many smokers for
a lifetime who never developed a significant lung ailment. Likewise,
they had treated patients with lung cancer who were not smokers.
The science of toxicology did not resolve this dilemma because
it proved difficult to produce tumors in laboratory animals with
tobacco exposures. The field of science that proved to be most
decisive is the one that some people now trust the least: epidemiology.
There was in fact a large statistical study of the health of British
physicians that played an important role in building the medical
consensus against smoking. Interestingly, epidemiology also played
a pivotal role in uncovering the neurotoxic effects of lead at
low doses and the diseases associated with exposure to asbestos
on the job.
we knew that scientific progress would always verify early signals
of danger or show that hazards are worse than predicted, then
the challenge of precaution would be much easier. A recent report
from the European commission makes the point that uncertain environmental
hazards often prove to be worse than anticipated. Yet the dynamics
of science are not so easily predicted. There are in fact many
cases of postulated or claimed hazards that have not been confirmed.
indications that drinking coffee might cause bladder cancer were
not confirmed. In the 1970's, the US Food and Drug Administration
declared a virtual war against the artificial sweetener saccharin,
after animal tests revealed bladder cancers following administration
of huge doses to rodents. The American people resisted the FDA's
conclusion and possibly for good reason. After thirty years of
biological experiments and large-scale statistical studies of
the consumers of saccharin, it now appears that the prediction
of a human cancer risk may have been incorrect. Indeed, the US
government recently took steps to remove saccharin from its official
list of carcinogens. More recently, scientific findings were publicized
claiming that low doses of chemicals now in widespread use may
be doing harm to the endocrine systems of the human body. Yet
further science has revealed that some of these early findings
about so-called "endocrine disruptors" cannot be replicated by
predictions of risk are also fallible. When I was a college student
in the 1970's, there were scholarly predictions of a Malthusian
global catastrophe, in part stimulated by reports of the Club
of Rome. There were also economists predicting that the world
price of oil would rise so high, due to limited petroleum reserves,
that the price of gasoline at the pump in the United States would
increase dramatically. Looking back, some of these predictions
may have been erroneous. As an academic, I contributed to the
erroneous predictions when I forecasted that front-seat airbags
would save 9,000 lives per year in the United States. It now appears
that the correct number will be somewhere around 3,000 lives saved
we contemplate the role of precaution in risk management, we must
remember that sometimes possible risks prove far worse than expected;
other times predictions of doom simply do not materialize.
is therefore useful to draw a distinction between the role of
precaution in the scientific assessment of risk and the role of
precaution in risk management. When analysts assess risks, they
may introduce conservative assumptions or safety factors into
the analysis to account for unknowns. These protective practices
may be intended to establish an upper bound on the true yet unknown
risk. In the United States, the technical community is now trying
to bring more valid data into the risk assessment process. When
considering the role of precaution in risk management, it is appropriate
for policy-makers and the public to inquire about the degree of
precaution embedded in the risk assessment.
use of precaution in risk management is sensible but susceptible
to misuse. If precaution is taken to an extreme, it can be very
harmful to technological innovation. Consider the following thought
experiment: Imagine it is 1850 and a decision is made that any
technological innovation cannot be adopted unless and until it
is proven to be completely safe by the proponents of the innovation.
Under this scenario, what would have happened to electricity,
the internal combustion engine, plastics, pharmaceuticals, the
computer, the Internet, the cellular phone and so forth?
the United States we have also learned the hard way that the urge
for precaution can lead to unfortunate outcomes. In energy policy,
for example, some of us regret our historical decisions regarding
nuclear energy. The possible risks of nuclear power generation,
coupled with the desire for precaution and rising costs of construction,
caused a virtual halt in the construction of new nuclear plants
in the United States. Thirty years later, we now find ourselves
even more deeply dependent on fossil fuels, which are a major
source of environmental concerns and calls for precaution. Part
of the answer lies in cleaner coal technologies, renewables and
energy conservation, but it also may be very unwise to foreclose
the advanced nuclear option.
people can disagree about what is precautionary and what is dangerous.
Consider whether the diesel engine should be used in passenger
cars and light trucks. Regulators in the State of California have
set the tailpipe emission standards for particles and nitrogen
dioxide so stringently that it may not be feasible to offer diesel-powered
cars for sale in the future. California's regulators see this
rule as a measure to protect public health from the known or possible
health risks of smog and soot in the air. Meanwhile, European
regulators and finance authorities have facilitated the growth
of the diesel-engine market in Europe to the point that a substantial
share of new cars in Europe are equipped with diesel engines.
From a global climate perspective, the pro-diesel policy in Europe
looks precautionary since the diesel offers significant fuel efficiency
advantages over gasoline-powered vehicles. Yet we should also
not forget that the price of gasoline at the pump is three to
four times larger in Europe than it is in the United States, reflecting
European tax policies.
diesel example reminds us that a zero-risk policy is rarely feasible.
More often, policy makers are engaged in an exercise of risk selection
and we should not permit any rhetoric about complete safety to
obscure this truth.
preparing for this conference, I re-read the European Commission's
February 2000 Communication on precaution and related comments
from committees in the European Parliament. I was encouraged by
these documents, even though we have many differences of opinion
about specific risk-management issues and even though the documents
do not provide a definition of the precautionary "principle" while
asserting its existence. Based on these documents, I detected
the following points of possible conceptual agreement.
precaution is a necessary and useful concept but it is also subjective
and susceptible to abuse by policy-makers for trade purposes and
other reasons. I was encouraged by a recent decision from the
European Court of Justice. The Court sided with the EC over France
on a BSE matter, suggesting that France may have overinterpreted
the precautionary principle. Second, scientific and procedural
safeguards need to be built into risk management decisions that
are based in part on precaution. Third, adoption of precautionary
measures should be preceded by a scientific evaluation of the
hazard and, where feasible, a formal analysis of the benefits,
risks, and costs of alternative precautionary measures. Fourth,
concerns for fairness, equity and public participation need to
be reflected in risk management. Finally, the set of possible
precautionary measures is large, ranging from bans or product
restrictions to education or warnings to market-based reforms.
Even the initiation of a targeted research program to better understand
a possible risk is a precautionary measure. As we consider the
case studies in this conference, we might consider what mix of
measures are appropriate in various circumstances.
conclude on notes of both optimism and caution. Precaution is
a perfectly sensible concept that is built into many health, safety,
and environmental laws in the United States. At the same time,
it may be wise to apply a precautionary approach to any attempt
to enact a universal precautionary principle into American law.
you very much for the opportunity to deliver these opening remarks.
I look forward to questions and comments.