Taking the Precautionary Principle
Seriously
Alan Randall
Precautionary Principle: when making multi-dimensional public
decisions, acting to avoid and/or mitigate potential harmful
consequences should be accorded high priority.
We already have techniques for risk assessment and risk management
Benefit cost, risk-benefit, expected value, expected utility
Insurance, self-insurance, self-protection – private, public
… Ordinary Risk Management
… but the whole PP movement is founded on the claim that something more,
something stronger than ORM, is needed.
Examples from PP statements
Uncertainty about harmful consequences does not justify failure to take precautionary action
(Bergen Declaration 1990)
Plausible but uncertain harm justifies precautionary intervention (UNESCO 2005)
Uncertain harm requires intervention, and the burden of proof is shifted to the proponent of the
proposed risky action (Wingspread Statement)
Applications
Multinational orgs – EU, UNESCO
National laws – US endangered species laws.
Regulatory policy – fisheries
International agreements – Montreal protocol
Trade disputes – hormones in beef, GMO commodities
International food aid -- Zimbabwe
PP Controversies
In lay discourse reactions range from …
PP offers an idea that seems like ordinary commonsense: extraordinary risks call for
extraordinary precaution
… to
PP would undermine business-as-usual and stifle innovation and growth. See, e.g.,
the Proactionary Principle: People’s freedom to innovate technologically is valuable
to humanity. The burden of proof therefore belongs to those who propose restrictive
measures…
Scholarly criticisms include:
PP is meaningless (self-defeating) … would forbid any and all risky alternatives, yet
even the no-action alternative involves risk (Sunstein).
PP is nothing special – just good science, risk assessment, benefit cost analysis, i.e.,
ORM (Farrow)
The PP debate is just a quarrel about who owns the null hypothesis.
There are almost twenty different formulations of the PP in circulation, so how can it
be taken seriously in real-world policy?
The PP conflicts with other important values that good law and policy would surely
respect, and it fails to provide clear instructions for resolving these conflicts.
My project
To define and justify a PP that can be taken seriously.
Neither foolish (not self-defeating, not readily invoked by false alarms)
Nor redundant (nothing more than ORM)
Must survive all of the criticisms above
Getting started
1. Consider the Hippocratic Oath – above all, do no harm. What can this mean?
If taken literally, many quite ordinary treatments would be forbidden
Perhaps it means do no harm in the net (BC)
Both action and no-action are risky (expected value or expected utility, ORM).
Perhaps it calls for an asymmetric calculus
Perhaps it is nothing more than an admonition to, before intervening, pause to
consider the unintended consequences
Implications
Addresses case 1: a novel and potentially threatening action
Identifies benefit cost, expected value, and expected utility approaches (ORM)
Also, 2 approaches outside of the usual range of ORM – stronger: asymmetric
treatment of uncertain but potential asymmetric harm; and weaker: sensitivity to
possible unintended consequences
Getting started
2. Consider the rivet-popper (Erhlich and Ehlich 1981)
“There is risk on both sides of the equation, and PP would paralyze innovation
and stunt economic growth, all because we are scaredy-cats.”
But surely the risk to the plane is at least a candidate for classification as
disproportionate and asymmetric relative to the risk to the economy.
Implications
Addresses case 2: a business-as-usual pattern of exploitation (or pollution) that
might overstress natural and/or economic systems, potentially resulting in a
novel and threatening outcome (regime shift)
Frames the discussion in terms of
collective threats (if the wings fail, all aboard the plane face extreme danger)
asymmetric and disproportionate threat
Specifying the PP – an ETR framework
If there is evidence stronger than E that an activity raises a threat
more serious than T, we should invoke a remedy more potent
than R
T. Threats relevant to my PP:
[Language notes: 1. harm, threat, risk. 2. risk, uncertainty, gross ignorance]
disproportionate and asymmetric
collective threats
arise from complex and novel phenomena (things and events).
Complexity possibility of regime shifts, tipping points
Novelty on one side of the equation vs familiarity on the other
Exacerbated by large spatial and/or temporal scale.
PP offers guidance beyond the scope of ORM if the choice set includes some actions
that induce disproportionate and asymmetric threats, and some actions that do not.
[PP offers little beyond ORM in cases where all possible actions involve similar threats.]
The nature of collective threats. The threats addressed by my PP are
collective threats.
Exposure is in considerable degree involuntary.
The threat is to a larger system to which we are inextricably bound
We are not sufficiently informed to exercise meaningful choice.
Individual defaults may plausibly include possible outcomes of very large
negative value – if I do not undergo the treatment I may die. For collective
threats, the default is much more likely to be: if we do not take the proposed
action, mortality may increase fractionally.
Default threats to individuals may well be of greater relative magnitude than in the
collective case
Conversion of frequencies among populations to probabilities for individuals is less
informative for individual choice than we might hope.
An example may help. The PP seems not well-adapted to guide decisions re the
permissibility of clinical trials.
Peterson (2006, 2007) has argued that the case of clinical trials is informative to the PP
debate: “… directly applicable to recent discussions of PP in medicine, biotechnology,
environmental management, and related fields”. I think not – my PP would not apply to
the clinical trials case (the threat is not asymmetric), but collective threats are different
and the PP is much better adapted to them.
R. Remedies may include
Prohibit the intervention / halt the business-as-usual stress
Avoidance, mitigation, and/or adaptation – beyond those that may be
recommended by BC or ORM considerations
Insurance, self-insurance, and self-protection – it takes a PP to justify a package
that includes ORM and some additional restraints.
There are two additional components of a comprehensive precautionary policy
Monitor, study, and reassess the threat
reduce or mitigate disproportionate negative impacts
E. What evidence would invoke PP? … plausible and credible scientific evidence
Commonsense notions of scientific plausibility and credibility.
US court decisions: Daubert, Kimho Tire, Cellular Telephone.
Judgments of scientific credibility are the province of scientists, not
ordinary
people. [See also Kahan et al response to Sunstein]
Campaign to politicize science and reduce scientific discourse to an ordinary
cultural quarrel: “junk science” (junkscience.com), “muddying the waters”
(Michaels and Monforton), “manufactured uncertainty”, “false balance”, and
“teaching the controversy”.
But … [a post-postmodernist defense of science as an institution capable of
resisting and surviving these challenges].
Tentative working definition of the precautionary
principle
In response to a scientifically credible threat of disproportionate and
asymmetric harm, we should take avoidance and mitigation measures
beyond those justified by expected welfare considerations.
On the possibility of a utilitarian PP. Can we imagine a utilitarian PP (utilitarian ORM
might logically include PPs that were consistently utilitarian)? Early attempts at a utilitarian PP
were problematic because extreme risk aversion was imposed in one guise or another, and
often unconvincingly. See also Peterson’s (2006) claim that the PP is incoherent.
Randall (e.g., 2007)argued that all of this (over)reaching for a utilitarian PP is much too
strained – there is no convincing argument against the idea that the scope for business-asusual utilitarian ethics might coherently be bounded by non-utilitarian constraints.
Margolis and Naevdal (2008) argued that thresholds, kinks, etc. that might generate a
utilitarian PP might be found in the response function rather than the utility function.
In the end the arguments of Margolis and Naevdal (who have justified a utilitarian PP for certain
particular specifications of how the world works), and Randall (who argues that a
coherent precautionary constraint on the domain of utilitarian ethics does not itself
have to be utilitarian) can coexist and in fact complement each other. Peterson’s (2006) case
that the PP is incoherent falls when faced with systemic complexity in a utilitarian universe, and
when faced with ethical pluralism.
Understanding the PP as a principle
Discussion of PP
commonly proceeds in legalistic terms, seeks enforceable interpretations
exhibits discomfort with competing definitions, and inconsistencies (either internal or with
other widely-honored values). See also Peterson 2007: PP cannot guide action and decision
unless it is structured and used as a stand-alone rule.
However, this discussion reflects a rather odd view of what can reasonably be
expected of a principle
Principles are general statements of normative moral positions for a class of concerns.
Some principles are recognized by many competing moral systems.
To call a moral statement a principle does not grant it lexical priority over other principles –
clashes between principles are resolved by weighing the principles and the facts of the case.
Yet, a principle is much more than a preference – we feel a serious moral loss when we have
to compromise a principle in a particular case.
Agreed principles provide not policies but a frame for policy resolution that anticipates
and accounts for competing and conflicting principles.
If PP is viewed as a principle, then it succeeds if it serves as a principle that
guides law and policy, in the sense that it embodies a considerable moral
intuition to be consulted along with other pertinent principles and values when
the issues it addresses come into play.
The PP and governance
Pragmatic policy processes submerge value discourse – “fact-based” policy:
no value matters until it matters
a special challenge when dealing with obligations to the future
Acting on principles in practical governance
The fundamental framework proposed and defended here would make business-as-usual
decisions according to familiar welfare criteria, while providing precautionary exceptions for
particular threats that are scientifically credible, disproportionate, and asymmetric. The
precautionary exceptions are justified by principles that would be honored by many different
moral systems:
Do not sell-out something unique and valuable for modest gain
Do not take inordinate risks for modest gain
Do not impose big risks on the public for modest private gain
Do not impose big risks on the future for modest immediate gain.
This framework would respect both the modern experience of technical progress
and increasing welfare, and the reasonable instinct for caution as we continue to
push at the frontiers of what can be known about our planet’s capacity to support
future welfare.
What can we reasonably ask government to do, in an era
when government seems afraid of governing?
Do more to get Ordinary Risk Management right
Benefit cost, risk-benefit, expected value, expected utility
Insurance, self-insurance, self-protection – private, public
Articulate clear principles and take them seriously in practice
Commitment to precautionary exceptions in cases that exceed a threshold
of concern
Commitment to transparent and inclusive public decision processes
Commitment to equitable sharing of burdens – esp. burdens undertaken
to benefit future generations
Bind managers to principles, but encourage them to interpret principles
in practice via a serious policy dialogue
Neither planning to the 4th decimal place, nor abdication to stakeholder
conventions