Public participation in domestic regulatory decision-making

What role the public can and should play in regulatory policy decision-making is a source of debate in many countries. The overall trend in many Western democracies is for greater public participation, which may take various forms, ranging from simple communication by government about how decisions were arrived at, to the active participation of public representatives or citizens in the decision-making process itself. The traditional view has been that decisions involving technical or scientific content should be left in the hands of experts and scientists, with public participation limited to the lesser end of the continuum (Busch, Reference Busch2003). However, this view is increasingly being challenged, and governmental, scientific, and industrial bodies in many countries are beginning to pay greater heed to the public, seeking to involve it in regulatory decision-making involving science (Rowe and Freyer, Reference Rowe and Lynne2000; Martin and Richards, Reference Martin, Evelline and Jasanoff1995).

This departure from the traditional view is prompted by a number of factors, including a practical recognition that implementing unpopular policies may result in reduced support for government. It also has a more normative basis, with public participation arguably being an ingredient in enhancing the legitimacy of government regulation. The concept of legitimacy is a central one in the field of administrative law, which is largely concerned with ensuring that the exercise of authority is justified (Danner, Reference Danner2003). Lindseth (Reference Lindseth1999) sees it as ‘a broad, empirically determined societal acceptance of the system’. In a democracy, legislative branches of government obtain their legitimacy from accountability to the electorate. Administrative agencies, however, have no such direct accountability, and the question of legitimacy therefore becomes more contested. Thus, issues of procedure loom large. In his well-known work, Why Do People Obey the Law, Tyler (1990) reported his research findings that, while people are interested in getting what they want, they are more concerned with being able to exercise control over the process in ways they deem fair. A key means of exercising such control is through public participation in the relevant processes. In the area of regulatory decision-making, governments throughout the world have sought to increase opportunities for public participation, particularly in areas of social importance such as environment and health policy (Markell, Reference Markell2005). Ensuring public participation in regulatory decision-making is seen as a means to enhance government legitimacy (Markell and Tyler, Reference Markell and Tom2007; Markell, Reference Markell2005).Footnote ²

Arguments for greater public participation in regulatory decision-making around SPS measures are in many cases predicated on the notion that, while scientists may be able to tell the public what the risks are, they have no particular expertise in decisions as to whether risks are worth taking. Thus, public participation is necessary to preserve the legitimacy of science (Busch, Reference Busch2003; Martin and Richards, Reference Martin, Evelline and Jasanoff1995; Kriebel and Tickner, Reference Kriebel and Joel2001).

Public participation raises a concern that resulting regulation may be less than optimal, from both a domestic and global welfare point of view. Green (Reference Green1997) argues, for example, that in the context of environmental policy, the US system of public participation has demonstrated potential to shift the incentives of policymakers in inefficient and irrational directions, leading to the ‘pollutant of the month’ syndrome where stringent regulations are put in place due to public demand whether or not such controls are necessary or desirable. Even where regulations are considered optimal domestically, the fact that consumers are differentiated across countries by their level of risk perception and the risks they are willing to take may cause trade tensions as noted above; for example, where the exporting country contends that the importing country's regulation is excessive due to unfounded public fears.

Risk in Western society

Over the past three decades, a number of commentators have identified a growing culture of risk aversion in Western societies as people have reacted negatively to new man-made products and industrial practices (Taverne, Reference Taverne2005). Beck (Reference Beck1999) writes that we are living in a ‘risk society’ in which the risks of scientific and technological innovations to individuals, society, and the environment are perceived to have escaped the control of society, its institutions, and national boundaries. He argues that this risk society questions the continuing appropriateness of traditional concepts of authority based on hierarchy and neutral expertise. The science that created the new dangers, he suggests, cannot be relied upon to protect us from those dangers.

A slightly different perspective on the topic is offered by Neal (Reference Neal and Jones2000) who suggests that we live in a ‘non-risk society’. He finds that people are obsessed with, and intolerant of, any risks, great or small and that the culture of risk aversion in Western societies is so strong and pervasive that heavy handed and/or irrational responses to risk are not seen as a cause for concern. Perhaps this is because, as Lowrance (Reference Lowrance1986) argues, there are many hazards that, while we know enough, scientifically, to ‘worry’, we do not have enough knowledge about to know how much to worry. For example, scientists can detect tiny traces of pesticides in food, but in many cases cannot determine whether the chemicals will have any adverse health effect. Lowrance (Reference Lowrance1986) finds that ‘in our knowing so much more, though imperfectly, and aspiring to so much more, we have passed beyond the sheltering blissfulness of ignorance and risk-enduring resignation. This has given rise to considerable social apprehensiveness, which is affecting both the outlook of individuals and the functioning of institutions.’

Public vs. expert risk perceptions

On the one hand, the public is prone to fear too much and to overestimate the probability of some risks. On the other hand, they are also prone to underestimate other risks (Fraiberg and Trebilcock, Reference Fraiberg and Trebilcock1998). For example, on the evidence currently available, cigarette smoking poses a greater scientific risk to health than eating hormone-treated meat, yet in many European countries, the first is tolerated and the second is banned (Sampson, Reference Sampson, Robertson and Kellow2001).Footnote ³ Such lack of correlation with science-based risk assessments has led to arguments that consumers are ignorant, sometimes hysterical, and misinterpret statistical data (Wilkinson, Rowe, and Lambert, Reference Wilkinson, Gene and Nigel2004). Weinberg (1987) comes to this conclusion after looking at public concerns about low-level insult radiation. He argues that ‘the connection between low-level insult radiation and bodily harm is probably as difficult to prove as is the connection between witches and failed crops. That our society nevertheless has allowed this issue to emerge as a serious social concern … is hardly less fatuous than were the witch-hunts of the Middle Ages.’

Yet the picture is hardly so black and white as these comments might imply; many factors play a role in public risk perceptions. Researchers across a number of disciplines have sought to understand what influences peoples' risk perceptions (Slovic, Reference Slovic1987). Understanding these influences is important in the consideration of how much credence countries ought to be allowed to give to public sentiment in justifying trade restrictive measures under the SPS Agreement. Three key drivers of risk perception can be identified from the literature: psychological, social, and cultural.

Public risk perception and regulatory decision-making

From a trade liberalization perspective, if one views public risk perceptions as flawed, any influential role played by those perceptions in the domestic regulatory decision-making process will be viewed as problematic and will raise important questions about the legitimacy of the resulting regulation. For example, Fraiberg and Trebilcock (2002) relate public risk perception to what they describe as the dismal record of regulatory performance in the area of health and safety. They cite a 1996 study by Hahn which found ‘reason to believe that most regulations implemented in the US since 1990 would not pass a cost–benefit test’ (Hahn, Reference Hahn and Hahn1996). Similar studies are cited by Slovic (Reference Slovic and Paustenbach2002), who finds that great disparities in monetary expenditures designed to prolong life may be traced to public perceptions of risk (see also Breyer, Reference Breyer1993).

Yet Slovic (Reference Slovic2000) argues that discrepancies between expert and public perceptions of risk should not simply be explained away by suggesting that the public is irrational or foolish. Rather, he finds room for respecting public risk perceptions. He argues that the fact that lay people focus on more factors (largely qualitative) than just the number of lives at stake, justifies paying respect to their judgments. He argues that because people pay attention to such crucial factors as whether a risk is dreaded and/or known, they are in a way capable of thinking better, and more rationally, than experts. That is, people's ‘basic conceptualization of risk is much richer than that of experts and reflects legitimate concerns that are typically omitted from expert risk assessments’ (Slovic, Reference Slovic2000). He refers to this as the public's ‘rival rationality’ to that of experts, and suggests that each side must ‘respect the insights and intelligence of the other’.Footnote ¹³ This argument supports a ‘populist’ approach to risk regulation, which suggests that in a democracy, government should follow the will of the people rather than that of a technocratic elite. It thus contrasts with a technocratic approach based on the notion that because ordinary people are often ill-informed, regulators ought to follow science and evidence, not public opinion (Sunstein, Reference Sunstein2001). Wilkinson, Rowe, and Lambert (Reference Wilkinson, Gene and Nigel2004) concur with Slovic, arguing that, while consumers do not understand ‘risk’ in the same way as scientists, their non-statistical, more qualitative, and heuristic-based risk perceptions are often arguably quite sensible. Pointing to the case of genetically modified (GM) foods, they suggest that when there are familiar and relatively risk-free alternatives to GM foods available, who is to say that consumer avoidance of GM foods is illogical?

Sunstein is one of the strongest critics of Slovic's view that the public displays a ‘rival rationality’, arguing (2001) that ordinary people deal poorly with the topic of risk.Footnote ¹⁴ One of his criticisms of the notion relates to the fact that people's concerns differ so notably among countries. Sunstein (Reference Sunstein2001) questions why, if lay people really have a richer rationality, these cultural differences are to be found? For Sunstein, the conclusion that people deal poorly with the topic of risk provides justification for a more ‘technocratic’ approach to regulatory decision-making where decision-makers should follow science and evidence rather than the public. Thus, while he concedes that experts do indeed make mistakes, he considers that in their area of expertise they are more likely to be right more of the time than are lay people.

Kahan et al. (Reference Kahan and Paul2006) take issue with Sunstein's view. Their model, which sees risk judgments as being inextricably bound up with cultural values, counsels a regulatory model that seeks to include public input through procedures that are ‘genuinely deliberative, open, and democratic’. Thus, like Slovic's ‘rival rationality’ approach, they would give weight to public perceptions of risk. In the final analysis, there would seem to be merit in a regulatory system that contains elements of both the populist and technocratic approaches. In other words, as discussed in Section 4, both scientific expertise and public sentiment have a valid and important role to play.

The rationale for the requirements

The SPS Agreement was negotiated on the basis that requiring scientific justification for standards that deviate from international norms would make it more difficult for countries to shelter domestic industries behind restrictive health and safety regulations (Josling, Roberts, and Orden, Reference Josling, Donna and David2004). This approach is based on the characterization of science as an objective discipline that is free of values and leads to definitive knowledge (Derry, Reference Derry1999). This objectivity, it is argued, will allow panels and the Appellate Body to identify those SPS measures that are genuinely intended to protect health.

The characterization of science as objective has traditionally been accepted outside, as well as inside, the scientific establishment. It has played a key role in the status of science, particularly in Western society, where it has been able to set the standards and boundaries of what is and what is not acceptable knowledge and practice (Woolgar, Reference Woolgar1988). Yet the perception of science as objective is subject to challenge both in academia and in the wider world. The academic challenge focuses on whether objective knowledge is possible and the extent to which science is socially constructed. Opposing camps have questioned, for example, whether the achievement of scientific consensus is a rational process, or a political process; and whether scientific theories are accurate reflections of a pre-existing reality, or merely social constructions based on negotiations between interest groups (Derry, Reference Derry1999).

Derry (Reference Derry1999) argues that extreme views do not do justice to the subtle complexities involved in questions of the proper relationship between science and values. He argues that science and values are related to each other, but the relationship is not hierarchical. Baber (Reference Baber2004) offers a useful insight in this regard, suggesting that science is both constructed and natural. He describes it as being like a map; it represents a reality but it is also a political and social construct. Thus, he argues that there is no need to choose one over the other. This author takes the approach that there are indeed objective facts to be discovered, and that the scientific disciplines have much to offer in providing a baseline of legitimacy for health regulations. Yet, it also recognizes that pure objectivity is often neither possible nor plausible, particularly in the health and ecological sciences.

In the wider world, the challenge to science focuses more on its actual use in different contexts, including in regulatory decision-making (Derry, Reference Derry1999). A crucial aspect in this regard is its vulnerability to manipulation and capture: either to support regulations that have no rational underlying scientific merit, or to question the validity of scientific evidence on which regulations are based. Michaels and Monforton (Reference Michaels and Celeste2005) use the term ‘junk science’ to describe the manipulation of science to support unfounded regulations, referring to it as ‘the mirror image of real science, with much of the same form but none of the substance … it is a hodgepodge of biased data, spurious inference, and logical legerdemain’.

It is important to note in this context that science has been criticized for being elitist and non-responsive to public concerns. Recent surveys show a range of attitudes towards science. A recent Eurobarometer survey, for example, found that on average only 17% percentage of respondents in the EU had faith in science to resolve any problem. A 2005 survey in the UK found that, while a clear majority of people (around two-thirds) trust scientists, there was significant variation of trust in scientists working for different organizations, with those working for industry or government much less trusted than those working in universities or for charities (Ipsos Mori, 2005).

Objections to the SPS Agreement's scientific evidence requirements

As noted in the Introduction, there is considerable disagreement in the literature about the role of science in the SPS Agreement, with arguments both for and against its use. One of the most common concerns is that the Agreement's scientific evidence requirements exclude consideration of public sentiment. Dayna Nadine Scott (Reference Scott2005) argues that by relying on science, the SPS Agreement precludes any consideration of broad social, cultural, and ethical concerns: ‘establishing itself as the only accepted discourse through which actors may broker claims about particular technological practice, risk discourse produces a zone of silence within public debate, muting competing critiques of technology framed in social, economic, or cultural terms’. She argues further that the Agreement ‘ignores the role of culture in the perception and construction of risks’.

Joanne Scott (Reference Scott and Weiler2000) argues that only the language of science ‘resonates in the application of the SPS Agreement’. This is problematic for her, as she goes on to elaborate:

Joanne Scott (Reference Scott and Weiler2000) criticizes what she describes as the Appellate Body's unwillingness to countenance the possibility that an SPS measure might be rational below the threshold of ‘more than theoretical risk’. The threshold, she argues, is the same regardless of cultural sensibility or filters which mediate a society's relationship to risk; and regardless of citizen preferences. She posits that ‘in the event that science neither proves the existence of risk, nor proves that there is no risk, there is scope for ‘rational’ debate as to whether this theoretical risk should be tolerated. But it is a debate which will inevitably transcend scientific rationality, thus shattering the fragile illusion of objectivity and universal commensurability, and reducing the potential for ‘scientific universalism … [to be] used to overcome the particularism of legal systems.’Footnote ¹⁵

The critics of a science-based requirement for justification of trade-restrictive health regulations do not suggest a credible alternative means of sifting out protectionism from genuine health regulations. It is suggested here that this is in part because there is no obvious or helpful alternative. Howse's (Reference Howse2000) view is instructive in this regard. He argues that the scientific requirements in the SPS Agreement should be understood not as usurping legitimate democratic choices for stricter regulations, but as enhancing the quality of rational democratic deliberation about risk and its control. As he states:

Howse's argument is consistent with the claim made in the following section that science is not inherently destructive of democracy.Footnote ¹⁶ The SPS Agreement's scientific evidence requirements should not be viewed as precluding countries from employing a democratic approach to regulatory decision-making, which considers public sentiment. How this can be achieved will be discussed following an examination of the use of science in domestic regulatory decision-making.

The use of science in domestic regulatory decision-making

To what extent are countries' risk regulatory decision-making processes aligned with the SPS Agreement's requirements? The discipline of risk assessment, as required by Article 5.1, gained popularity in the US in the context of health regulation in the 1980s in order to help regulators make decisions in the face of large quantities of scientific and medical data concerning potential health hazards posed by physical and chemical agents in the environment (Paustenbach, Reference Paustenbach and Paustenbach2002). Risk assessment methodology has evolved significantly since then and risk assessments are now widely accepted practice not only in the US but in most other developed countries (Paustenbach, Reference Paustenbach and Paustenbach2002). In many cases, this move towards use of risk assessment goes hand-in-hand with a legal mandate to use science as a basis for regulatory decision-making in public health and environmental matters.Footnote ¹⁷

Risk assessment is generally seen as just one aspect of the broader process of risk analysis, which also includes risk management and risk communication. Traditionally, risk assessment has been seen as a separate aspect of the process where scientific questions are isolated and addressed in an objective manner at the beginning of the regulatory process, while pure policy choices are confined to the second stage of risk management (Turnley, Reference Turnley and Paustenbach2002).Footnote ¹⁸ However, the distinction between risk assessment and risk management, while intuitively appealing, is increasingly challenged. The US National Research Council (2000), for example, takes the approach that risk analysis is an iterative process, where risk managers move back and forth between the parts of the risk analysis using different sciences, cultural frameworks of the respective countries, and domestic and international political and economic considerations in making their decisions.

It should be noted that there is no explicit division in the SPS Agreement between risk assessment and risk management. However, the Agreement's rules do cover the normal activities that one would expect to see under both of these categories and broad distinctions may be drawn in regards to which rules are applicable to what activities. As noted, Article 5.1 requires countries to perform a risk assessment, as defined in Annex A(4). Article 5.2 specifies various factors that must be taken into account in conducting that risk assessment, including available scientific evidence. The SPS Agreement further provides that once a risk assessment has been conducted and there is evidence of a risk, countries must ensure, inter alia, that negative effects on trade are minimized (Article 5.4), that the least-trade restrictive option is taken (Article 5.6), and that there is no discrimination or disguised restriction on trade (Articles 2.3 and 5.5). While the Agreement does not state so specifically, the rules in Articles 5.4, 5.6, 2.3, and 5.5 all apply to what would be considered the risk management phase of risk analysis.

Risk assessments aim to estimate the likelihood of an adverse effect on humans, animals, or plants (as well as ecological systems) posed by a specific level of exposure to a chemical or physical hazard (Paustenbach, Reference Paustenbach and Paustenbach2002). They provide relevant information to policy and decision makers, but will not be the only information used in the final policy decision; social, economic, and political factors will also be considered during the risk management phase. In particular, the information obtained from a risk assessment does not answer whether or not the risk is acceptable to the population in question (Wooldridge, 2001). For example, a scientific risk assessment may help in setting a standard designed to limit the probability that an individual will develop cancer after a lifetime of exposure to a particular chemical substance to no more than one chance in a million. But the choice of the one-in-a-million goal is a policy choice that cannot be determined by science.

A defining characteristic of risk assessments is uncertainty, which exists when there is insufficient information to attach objective, or even subjective, probabilities to the risks in question (Kerr, Reference Kerr2003).Footnote ¹⁹ Uncertainties introduce room for subjective judgments and assumptions as risk assessors decide how the information they do have ought to be interpreted. When these uncertainties are considered, it becomes clear that the process of risk assessment is, as Powell (Reference Powell1999) describes it, a ‘messy business’. It is not so much an exercise in finding scientific truths as it is a complex set of judgments about whether assessors believe the evidence presented.Footnote ²⁰ The US National Research Council (1983) finds that many risk assessments can be neither proven nor disproved scientifically and estimates that a typical risk assessment consists of about 50 separate assumptions and extrapolations. These assumptions will be influenced by culture, politics, and subjective opinion. Different assumptions can alter the outcome significantly, depending on whether they are risk-averse, risk-tolerant, or risk-neutral. Powell (Reference Powell1999) refers by way of example to different rules developed in US agencies to interpret the pathological evidence on dioxins that have resulted in the Food and Drug Administration's cancer potency estimate for dioxin being almost an order of magnitude smaller than the Environmental Protection Agency's (EPA) estimate, and the Centre for Disease Control (CDC) falling somewhere in between. The inconsistent estimates result from the agencies applying the same linear cancer model but making a variety of different scientific assumptions and data treatments. In some cases, agencies will establish so-called ‘science policies’, whereby risk assessors are told what assumptions to apply in cases of uncertainty. For example, the EPA requires risk assessors to assume that positive effects in animal cancer studies indicate carcinogenic potential in humans (Walker, 1998).

A question arising from the uncertainties and complexities inherent in risk assessment, and the use of science policies, is how – in the context of a risk assessment – one can determine whether the decision made is a rational one? There are no codified rules in science for determining scientific rationality and, as a result, two scientists may arrive at different conclusions as to the rational support for a particular line of reasoning (Crawford-Brown, Pauwelyn, and Kelly, Reference Crawford-Brown, Pauwelyn and Smith2004). Despite the lack of rules, Crawford-Brown, Pauwelyn, and Kelly (Reference Crawford-Brown, Pauwelyn and Smith2004) argue that it is still possible to use the concept of an epistemic threshold whereby it can be asserted that no risk estimate should be considered as a candidate for scientific belief if it has not passed at least some minimum requirement of quality. In other words, it must have minimal epistemic status. They note that this threshold is one that is aspired to by the scientific community through the peer review process for research studies and scholarly publications.

The irony of risk assessment, with its attendant uncertainties and inevitable value judgments, is that it can be manipulated either to support or to oppose regulation. It can produce risk estimates that err on the side of risk acceptance, while, at the other end of the scale, it can produce risk estimates that are extremely conservative and unwarranted by the scientific evidence available. If this aspect of science is not recognized by WTO adjudicators, there is a risk that almost any trade restrictive SPS measure could be justified in the name of science if a domestic government plays their cards strategically. Section 4 will make suggestions as to how the risk of such an outcome can be minimized while still enabling science to be employed in a manner consistent with public participation.

Precaution in risk analysis

Many commentators have argued for the use of the so-called precautionary principle to bridge the gap between scientific uncertainty and risk regulation (United Nations, 2005). Interpretation and implementation of the precautionary principle takes a number of forms and it is impossible to identify any one uniform description. One of the best-known definitions is that set out in Article 15 of the 1992 Rio Declaration, namely, that ‘[w]here there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation’. Other definitions take a stronger stance, for example requiring action to be taken in the absence of scientific certainty (Goldstein and Carruth, Reference Goldstein and Russellyn2004).

Despite the lack of precise definition, it is possible to identify certain characteristics that can be labeled ‘precautionary’. Cosbey (Reference Cosbey2002) identifies the following: (i) preventative anticipation (the notion of taking action in advance of full scientific proof of its necessity); (ii) room for error (because we are dealing with complex systems, a deliberate margin for error should be left open); (iii) proportionality of response (the cost of proposed measures must not be out of proportion with the expected benefits, including costs avoided); (iv) onus of proof (should be on the proponent of any new product to provide an adequate level of evidence of its safety); (v) a search for greater certainty (any precautionary measures taken must be accompanied by a search for greater scientific certainty, and periodic re-evaluation of the measures in light of new evidence); (vi) openness of process (there must be transparency in the process of decision-making, timely distribution of information, and mechanisms for input from all those affected); (vii) emphasis on finding alternatives (if alternative products or technology exist that have the same valuable qualities without the same risks of negative effects, the alternatives are to be preferred).

The application of these kinds of precautionary actions in risk assessment would recognize the uncertainty inherent in the process and would dictate that decisions regarding what evidence to believe reflect conservative or cautious approaches to risk. This would mean, for example, that in cases of doubt, greater weight would be given to evidence suggesting risk, and that the vulnerability of sensitive members of the population be used to establish safety factors (Prévost, Reference Prévost2005). In risk management, precaution would entail giving effect to conservative risk assessments by allowing for the possibility that benign products may be banned or strictly regulated. From an international trade perspective, recognition of the precautionary principle, where there is any doubt about safety, would entail giving the benefit of doubt to importing countries in the form of precautionary protection, rather than to the producing country, even if the risks have not been completely assessed or cannot be proved with certainty to exist (WTO, 1997).

It would be absurd to suggest that risk regulators should never act with caution. As Sunstein (Reference Sunstein2005) observes, both individuals and sensible governments will embrace some form of precaution in the face of many risks. Hrudey and Leiss (Reference Hrudey and William2003) argue that ‘done well, risk management is inherently precautionary in the sense that it should make use of effective risk assessment to predict, anticipate, and prevent harm, rather than merely reacting when harm arises’. They argue that the key question is not whether to be precautionary but how precautionary we ought to be in specific cases.

Arguably, a generalization might be made that the threshold for precautionary action ought to be higher when the potential risks involve serious or irreversible harm, and lower when the potential harm is only minor (United Nations, 2005). In cases involving potentially serious and irreversible damage to human health, such as a risk of cancer, exercise of greater precaution is arguably defensible. It may also be defensible in cases where economic damage is at issue. For example, in an instructive discussion of plant health, Campbell (Reference Campbell2001) notes that biological pest invasions are usually irreversible and that it is far easier and less expensive to prevent an invasion through precautionary action than to combat one once it has occurred. Conflict over the use of precaution is inevitable, however, where it is suspected that it is being used to disguise what is in essence a protectionist measure. The place of the precautionary principle in the SPS Agreement is discussed in relation to Article 5.7 of the SPS Agreement in the following section.

In the domestic sphere

For many governments, if they wish to maintain public support for international trade liberalization, it is critical that they find ways to involve the public in trade-related decisions that impact upon their lives. Public health policy, as well as plant and animal health concerns, are areas of great interest and concern to many people. Governments therefore have much to gain domestically from ensuring that the public has a say in regulatory decision-making processes regarding SPS measures. A process that acknowledges and accepts a place for public participation throughout the process will be a more democratic process and is likely to be more widely accepted among critics of the trade system. The implications of this are noted by Skogstad (Reference Skogstad2001) who argues that if international science-based regulation is to be widely accepted as legitimate, it will likely have to rest on a more indeterminate understanding of scientific knowledge than has previously been the case. She suggests that such a view of science inevitably gives countries more scope to devise public policies consistent with the preferences of their citizens.

As noted, risk management has traditionally been understood as the phase of risk analysis where public values enter the fray, while risk assessment has been understood as a factually grounded, objective (that is, based on science), and value-free, analytic exercise requiring precise identification of possible harms to health and use of formal, expert-based assessments of the likelihood of such harms (Busch et al., Reference Busch2004). As the examination in Section 3 highlighted, however, this account is misleading and misstates the true nature of risk analysis which is increasingly being understood as a process where risk assessment and risk management overlap in many respects.

A number of critics argue that risk assessment in the domestic context is ‘undemocratic, morally bankrupt, and ethically impoverished’ and that it is a means to rationalize government decision-making and its alleged dominance by industry (Charnley and Elliot, Reference Charnley, Donald Elliot and Paustenbach2002). It is argued here that risk assessment should not be written off so quickly. First, risk assessment characterizes risk; but people make political and moral decisions about how much risk is acceptable. Risk assessment is used to inform decision-making; but risk decisions are, ultimately, public policy choices. Second, the presence of uncertainties in risk assessment speaks to the possibility for domestic governments to involve the public in shaping the approach taken by scientists in making assumptions and interpreting data. This is something that is increasingly being recognized by governments.

In a report prepared for Health Canada, Jardine et al. (Reference Jardine2003) emphasize that stakeholder involvement and communication are key elements that should be included in a framework for human health and ecological risk assessment as well as management. This must involve clearly communicating data sources and assumptions to the public (and other interested stakeholders) (Post, Reference Post2006). Jardine et al. (Reference Jardine2003) note that the importance of stakeholder involvement has been accepted in international fora, citing the outcome of a 2000 international workshop on risk-based decision-making, which concluded that:

Jardine et al. refer to a large number of instances where recognition has been given to the importance of public participation, including in the risk assessment process.Footnote ²² For example, the US Presidential/Congressional Commission on Risk Assessment and Risk Management, 1997 found that stakeholder collaboration provides opportunities to bridge gaps in understanding, language, and perceptions. It also found that valuable information or perspectives may emerge during any stage of the process and that it is therefore important that the process be both iterative and flexible enough to accommodate changing or new information. ‘Information gathering, research, analysis of risk or options, and public involvement may clarify or redefine the problem, change the focus to a different problem or identify other risks in a broader context.’Footnote ²³ The US National Research Council (1996) has made similar arguments, arguing that risk management requires an ‘analytic-deliberative’ process whereby analysis and deliberation are complementary and integrated throughout risk characterization. Each step of the process must, it argues, have an appropriately diverse participation or representation of the spectrum of interested and affected parties, of decision makers, and of specialists in risk analysis.

The issue of public participation raises the question of whether a regulatory decision based on scientific judgement will lack legitimacy if that judgement is at odds with public opinion. On the one hand, the notion that popular opinion will can be thwarted because it does not accord with scientific opinion cuts against the concept of legitimacy in a democracy. On the other hand, if the regulatory process is seen to be fair and inclusive, there is a much higher chance that legitimacy will be maintained, even where scientific and public opinion diverges, if there is opportunity for communication between stakeholders throughout the risk analysis process (Jardine, 2007). Jardine (2007) emphasizes that public attitudes cannot change scientific facts. What they can do, she argues, is ‘change how information is interpreted and used in decision-making processes, thereby validating decisions about these risks’. She notes that some researchers have suggested that stakeholder participation functions as a peer review process that complements the scientific peer review process.

In practice, countries tend to do a poor job of involving the public in risk analysis and there are no easy answers to finding the ideal means and level of involvement, with different situations requiring different responses. However, as countries experiment, more will become known about successfully managing public participation.

Under the SPS Agreement

Despite the limitations of science as a decision-making tool, it arguably remains the best tool we have to sift out unwarranted protectionism from genuine health protection measures. Without science, we are left at the mercy of competing unproven claims and perspectives. Even if we accept that public risk perceptions are often rational, they cannot be the only guide. As discussed in Section 2, lay people's perceptions of risk are influenced by different factors, some of which may incorporate protectionist agendas. As Cross (Reference Cross1992) observes: ‘powerful economic interests cannot change objective truth, but they can change public perception. Money and media are influential.’ It may be argued that, from the perspective of trade liberalization, some public risk perceptions are more ‘legitimate’ than others. Those which are the result of psychological factors or cultural worldviews ‘internal’ to the individual may be seen as more ‘legitimate’ than those social factors that may encompass protectionist interests. However, even if public sentiment could be determined as arising only from the more ‘legitimate’ factors, accepting public sentiment as a sole justification for a trade restrictive regulation would give countries potential to justify almost any chosen regulatory measure. The SPS Agreement's commitment to trade liberalization would be severely constrained if any measure could be justified in the name of a population's risk perceptions, regardless of how arrived at. Therefore, it is argued that we need a measure that is capable of being rationalized and assessed, even if that process is not perfect. Science, it is argued, has this capacity.

However, the limitations of science as a decision-making tool must be noted. Science is vulnerable to manipulation and capture that may defeat the original purpose of WTO rules endorsing a science-based system. Various stakeholders may seek to manipulate science, including scientists themselves, as well as governments, employers, and industry sponsors. As well, in many cases, lack of scientific data may mean that some decisions are inevitably based in part on non-scientific considerations. Domestic governments wishing to comply with the SPS Agreement should pay attention to this problem and seek to minimize the risk it poses by ensuring transparency and independence of risk analysis work carried out by government agencies. WTO panels and the Appellate Body should also consider the problem when considering what the SPS Agreement demands of countries. Trebilcock and Soloway (Reference Trebilcock, Julie, Kennedy and Southwick2002) have addressed this point, arguing that the regulating Member ought to bear the burden of proving that its risk regulation reflects at least a credible minority of scientific opinion in terms of the scientific credentials and research methodology of those holding the relevant opinion. This test hints at the need for a threshold of scientific rationality in its mention of research methodology. It will be suggested here that this is an appropriate way of proceeding; the discussion below suggests in more detail what role disputes panels and the Appellate Body should play when reviewing scientific evidence.

Panels should not put themselves in a position of deciding which of two or more differing scientific opinions put forward by the parties to a dispute is preferable (in effect acting as a supra-national ‘Science Court’). Rather, their role should be to ensure that the scientific evidence relied upon by an importing party passes a minimum threshold of rationality as determined by the risk assessment procedure followed. The discipline of risk assessment is still evolving, with a variety of frameworks used, within and across countries. It would be both unrealistic and unwise to require countries to follow a specific method of evaluation in carrying out a risk analysis. This would interfere unduly with regulatory autonomy, as well as the progress of science itself. As such, WTO panels should not take on the task of dictating how regulatory agencies perform risk assessments. However, panels should require some level of procedural conformity from countries relying on risk assessments to justify trade restrictive regulations. Dispute panels should ask whether countries have conducted a rigorous risk assessment that is consistent with protocols agreed by the international scientific community. That is, there must be a sense that the process followed is one that is reasonable in the circumstances, given the current state of scientific knowledge, and that domestic agencies have not arrived at an arbitrary or manifestly absurd conclusion. Some common ingredients can be distilled from the myriad of risk assessment frameworks in use. At the highest level of generality these include the need for scientific rigour, consistency, and transparency.

Dispute panels should consider using peer review processes as a guide for their role in reviewing countries' compliance with the SPS Agreement's scientific evidence and risk assessment requirements. Peer review is used as a tool by leading scientific journals in ensuring the rationality of the results submitted, and is also used by regulatory agencies who use science as a base for decision-making. Peer review may be described as a critical review of a work product (e.g., risk assessment) by experts who are independent of those who developed the product. It is a process designed to ensure that activities are, inter alia, technically accurate, competently performed, and properly documented (US Environmental Protection Agency, 2000). In the context of the SPS Agreement, panels could draw from the process of peer review in order to ensure that countries' risk assessments have a sound and credible basis. They would ideally do so by employing experts to conduct an in-depth review of the risk assessment, including the assumptions, calculations, extrapolations, alternative interpretations, methodology, acceptance criteria, and conclusions reached. The review would thus be procedurally based, focusing on the underlying scientific merits of the risk assessment, rather than on the actual decision reached. The purpose of the examination would not be to second-guess the scientific judgments made, but to filter out ‘junk science’ used to justify protectionist interests.Footnote ²⁴

In the language used by Crawford-Brown, Pauwelyn, and Smith (Reference Crawford-Brown, Pauwelyn and Smith2004), such an approach would enable panels to focus their attention on whether the result of a risk assessment achieves a threshold of rationality. The kinds of matters to be examined by experts will vary according to the specific circumstances (depending on factors such as the amount of information available, and the methodological difficulties encountered) but in each case panels should pose specific questions and concerns to be addressed by the experts as well as inviting general comments concerning the credibility of the risk assessment. These will include the following kinds of issues:Footnote ²⁵

• • whether the results of the risk assessment support the conclusions reached;

• • whether the risk assessment was based on the best available information that is in accord with current scientific thinking;

• • the degree of transparency of the process;

• • whether the risk assessment documented uncertainties in the data;

• • whether the risk assessment used appropriate models, datasets, and assumptions on which to base a scientifically credible decision;

• • whether the models used, with their associated datasets and assumptions, are able to answer the questions as stated in the risk assessment;

• • whether the modeling approaches used are suitable for assessing the suspected risk to human, animal, or plant health;

• • whether the assessment was based on the most appropriate studies, and whether there are other data or studies that are relevant (that is, useful for the hazard identification or dose response assessment) for the risk assessment; and

• • whether the choice, use, and interpretation of the data employed in the risk assessment is appropriate and scientifically sound.

To conclude, the suggestion here is that WTO dispute panels focus on the procedure followed by countries in their regulatory decision-making processes. Panels should seek to ensure that risk assessments relied upon are conducted in a rigorous manner, and that risk estimates meet a minimum threshold of scientific rationality. Where necessary (which it will be in most cases), they should seek assistance from expert panels familiar with scientific methodology. It is not realistic to suggest that any formula will enable panels to wave a magic wand and answer the question of whether or not a trade restrictive measure should be justified to protect health. The nature of the world we live in does not allow for such certainties. Rather, by ensuring that a minimum procedural threshold has been reached in terms of risk assessment, panels can allow countries flexibility to make decisions, while seeking to prevent purely speculative or ill-founded regulations that, in essence, are designed to protect local producers.

Where does this conclusion leave public sentiment? As the discussion of risk perception noted, public sentiment may be genuinely concerned with health and should be allowed to play some part in the risk analysis process; indeed, it is arguably important in a democracy that both scientists and regulatory decision-makers take such concerns into account. Such concerns should be accepted as having value (their own ‘rationality’), even if they are based on alternative perceptions than those held by experts. It is worth noting here Howse's (Reference Howse2000) argument that ‘if citizens believe they need a certain regulation, however “deluded” such a belief is, their utility will be reduced if they do not get it in the sense that they will believe themselves exposed to a risk they believe to be significant.’Footnote ²⁶ In other words, the utility from regulation comes not only from reduced likelihood of an event that one disvalues, but also from the psychological security that results from one's belief about the protection one is receiving (Chang, Reference Chang2004).

This interpretation of the SPS Agreement's science-based framework is reconcilable with a role for public sentiment. Such reconciliation rests on the acknowledgement that science incorporates assumptions, judgments, and values, and the argument that the scientific risk assessment process need not (and should not) silence public sentiment as some scholars contend it does. Rather, it is suggested that the scientific risk assessment process is capable of including public debate and alternative views of risk.

In this regard, science's limitations (vulnerability to manipulation and capture) also give it an advantage; they lend it a certain malleability as a decision-making tool. This interpretation counters the critique of some authors that science is a technocratic and elitist institution that ‘by necessity’ excludes the public voice. The public voice need not be excluded; where it is, the fault lies not with science per se (science is, after all, simply a means of obtaining knowledge about the world around us) but with the manner in which its tools are employed. Undoubtedly, it can be employed in an elitist and exclusionary manner. But it is the role of national governments to make sure that it is employed in a democratic manner by ensuring transparency and independence of risk analysis work carried out by government agencies. It is important to remember that we could abolish the use of a scientific benchmark in the SPS Agreement, and governments could continue to ignore public values as they wished.

However, incorporation of public sentiment in decision-making should not be allowed to compromise the scientific integrity of the process of risk analysis. As Cross (Reference Cross1992) suggests, ‘the people need not be foreclosed from risk determination, but reality (as ascertained through the scientific method) must remain as a check on the powers of government to act on public perceptions. Government systems should be constructed so as not to defer automatically or even presumptively to public perceptions of risk, unchecked by scientific data.’ The problem with relying fully on public sentiment is the lack of ability to then protect against the spectre of disguised protectionism. The social factors that influence public risk perceptions in particular leave open the potential for public opinion to be manipulated by protectionist interests.

It might be questioned whether it is possible to devise some procedural mechanisms whereby public sentiment could be evaluated and articulated in a systematic way. Charnovitz (2005) suggests as possibilities holding a referendum or poll of public opinion.Footnote ²⁷ However, given that public preferences may be driven by numerous factors, including protectionist influences, the problem arises that it would be extremely difficult and likely impossible to get to the bottom of what drives public opinion. Charnovitz (2005) raises another problem with finding a systematic means of ascertaining public opinion, namely that it would be problematic to imply that the WTO should show greater respect for public opinion than for policies chosen through political representation, technocratic expertise, or judicial authority. Noting that the WTO paradigm is that a government speaks for its citizens, he recalls that the WTO does not normally look inside a country to judge the democratic nature of its decisions. In other words, in the WTO, an assertion of a position by a government is conclusive even if the position fails to reflect popular will within that country.