Ethical Issues Facing the Food Industry

by

Paul B. ThompsonCenter for Biotechnology

Policy and EthicsTexas A&M University

College Station, TX

The array of ethical issues facing the foodindustry is extensive. It includes fair and justtreatment to food industry employees, especiallyas disproportionate numbers of minorities takejobs in food processing and food service. Issuesof food distribution and hunger continue to beimportant. In the calendar year 1992, however,these issues pale in comparison to those raised byfood safety and labeling. The impetus for thisissue is multiple, The Nutritional Food LabelingPolicy Act has mandated new labels intended toprovide consumers with consistent information oningredients that will be useful in dietary planning.Questions over labeling of foods derived from thetransfer of genetic materials are being asked byregulators, the food industry and by consumeradvocates. The so-called “Delaney paradox,” hasraised questions about the ethics of limiting riskfrom additives, while risk associated with wholefoods is unregulated. At the same time, lingeringquestions about risk from pesticide residues andmicrobial contamination frame a continuing debateover food safety, one which frequently returns tolabeling as a strategy for addressing consumerconcerns.

The balance of this paper outlines a frame-work for policy analysis, and demonstrates howethics bears upon each element of the framework.Contested issues in food biotechnology policy areused to illustrate the applicability of the frame-work for interpreting policy conflict. Althoughthis approach addresses several of the key points

where ethical concerns bear upon food biotechnol-ogy, the paper makes no attempt to survey the fullrange of ethical concern. What is more, the paperdoes not present a normative argument favoringone policy option rather than another. The ideathat ethics requires a particular set of policies forfood biotechnology is not argued in this paper.Instead, the purpose is to examine how ethicalarguments establish a burden of proof for policyevaluation. The thesis is that effective policymaking requires an ability to understand howdifferent types of ethical criteria bear on policy,Insensitivity to contrasting ethical approaches willonly prolong policy conflict.

The framework is then brought to bear onthe question, “What ethical considerations shouldbe brought to bear on labeling policies for foodproducts?” This question does not equate labelswith wjirnings. There are many ways to configurea labeling policy that do not imply health claims.The short answer to the question is that there aretwo kinds of ethical considerations. The first hasto do with the use of labels as tools to produceethically desirable ends such as good health andconsumer satisfaction, The second has to do withthe role of labels in protecting the principles ofconsent. While these two ways of evaluatinglabeling policies may converge, they may alsoindicate contradictory directions. The longanswer to the question uses a general approach toethics and policy to show why this is the case.

February 93/page 12 Journal of Food Distribution Research

A Framework for Policy Analysis

Schmid (1987) presents a theoretical frame-work for policy analysis where the laws, proce-dures and administrative decisions that seem to beinstruments of policy are analyzed in terms of theincentives they create for key actors. Schmid’sframework develops a public choice/transactioncost approach to public policy that permits ananalysis of how informal norms and standardoperating procedures interact with the formalapparatus of law and administrative decision mak-ing. Conventional economic policy analysisassumes that a policy’s costs and benefits can becomputed simply by examining the impact of lawsand administrative decisions upon production costsand consumer demand. Here the key insight ofSchmid’s approach is that the formal policy appa-ratus is one component in an ensemble of laws,norms and standard operating procedures. Thetotality of this ensemble imposes a structureuponan existing reality, and the combination of struc-ture and the situation as determined by physicaland biological facts determines individual incen-tives and opportunities. According to Schmid,economists have naively assumed that individuals’behavior is shaped merely by preference rankingsof exchangeable goods and have failed to examinehow shared norms and public policy shape oppor-tunities for choice.

For purposes here, Schmid’s frameworkwill be telescoped into four key elements. Theyare defined here with explicit attention to theanalysis of food safety and nutrition policy.

1.

2.

3.

Situation: the things that cannot be changed.This should be understood to include the physi-cal, chemical or biological processors thatdetermine food production and consumption.

Structure: the ensemble of laws, sharednorms, procedure+ and rules that are eitherproposed or in place in the status quo. Inaddition to the obvious elements of policy,structure includes norms that govern whatpeople regard as food.

Chduct: the behavior that will be produced asa result of the opportunities created when agiven structure is imposed upon the situation.

Journal of Food Distribution Research

4.

Conduct includes the production, processingdistribution, and consumption of food.

Pe@ormance: a given pattern of conduct willproduce an end ‘state ‘which consists of thepolicy’s consequences for affected parties.Health, disease, injury, profit and loss allqualify as components of this end state.(Schmid, Shaffer and Van Ravenswaay, 1983)

The framework provides general categories thatSI1OWa competent analyst to bring implicit fea-tures of policy out more clearly, and to examinehow policies produce end states. It is admittedlyquite general and is undoubtedly commensuratewith many different methods of policy analysis.

It is worth noting a few additional pointsbefore examining how ethics bears upon theframework. First, the framework is interpretivein that it will require judgment to assign specificvariables to any of the four elements. For exam-ple, the technology that is used to detect the pres-ence of a substance in foods uses physical andchemical principles. Technology is, in that sense,a part of the reality or situation on which a struc-ture is imposed. However, this technology haschanged so dramatically in the past four decadesthat it is probably more useful to think of it as acomponent of structure. The interpretation ofadministrative guidelines for food safety decisionsincludes standard operating procedures for the useof specific technological tests. As such, whentechnology changes, there is a sense in whichpolicy changes, too.

Second, the general category of perfor-mance can be taken to include the full range ofcriteria that would be applied in evaluating apolicy. As will become clear shortly, some suchcriteria have little to do with the end state pro-duced by the policy. The dominant practice inpublic policy analysis is to predict policy out-comes, and to report them as an end state, oftenas costs and benefits. This practice leaves thedecision to the responsible party or parties, bethey an administrator, a court or the Congress.Decision makers can and do apply criteria thatmake little if any use of predicted end states, butthe typical practice among analysts is to equate thepredicted end state with the policy’s performance.

February 93/page 13

Analysts writing on the banning of Alar, forexample, typically evaluate the policy in terms ofa trade-off between the economic value of theapple crop for producers and some minimal, evententative, reduction in risk for consumers. Thisapproach leaves open the possibility of comparingthese outcomes using a variety of criteria, butpresumes that the decision is based upon projectedpolicy consequences to the extent that it is defen-sible at all (Roberts and Van Ravenswaay, 1989),

How Ethics Bears on Policy

The assumption that consequences (or endstates) provide the basis for evaluating publicpolicy has its philosophical basis in the ethicalwritings of Jeremy Bentham and John Stuart Mill.These utilitarian philosophers argued that actioncan be justified only in light of the consequences,and they proposed the twin norm of countingconsequences for all affected parties and of maxi-mizing aggregate utility. Traditionally utilitarianphilosophy has been criticized for its insensitivityto the distribution of costs and benefits. In morerecent times, John Rawls (1971) has argued thatpolicies should benefit the worst-off groups ratherthan maximize aggregate utility. As Nozick(1974) noted, both utilitarian and Rawlsian egali-tarian theories evaluate policy by applying a normor decision rule to the end state that the structureis expected to produce. Many economists who donot think of themselves as either utilitarian oregalitarian would also assume that end statesprovide the side basis for evaluating policy. Thesearch for Pareto better solutions or et%cientlevels of pollution begins by predicting policyoutcomes. The debate is over whether the ac-counting is accurate and complete or which princi-ples to apply in evaluating end states.

To use the language developed here, theseare all pe~ormance or end state focusedapproaches to ethics. Their philosophical pedigreeextends back to Bentham, who hoped to reform anEnglish legal system based upon status and privi-lege. By turning the debate toward consequences,Bentham established a burden of proof for whichsocial rank and divine right were irrelevant.While common law based policy on past practice,Bentham’s theory held that it should be basedentirely on expected outcomes. Bentham assumed

that rationality consisted in actions chosen asmeans to an end. While one might disagree aboutends, Bentham thought that a rational person mustaccept that acts which fail to achieve the desiredend are to be rejected. The point here is thatstructure becomes a means toward an end. Whilecitizens in a democracy can be expected to havedifferent preferences, they must evaluate the rulesand regulations adopted by policy makers only asmeans to some end. Hence end states are thedominant performance criteria for public policy.

Nozick’s Anarchy State and Utopi&(1974)is an extended philosophical attack upon thisnotion of political rationality. He offered a nowfamous analysis of why the basketball player WiltChamberlain is entitled to great wealth, despitewhat Nozick thought to be the lack of any propor-tionate social value produced by his play. Theargument stressed that, given any initial distribu-tion of wealth, policies that confiscate moneyvoluntarily exchanged between Chamberlain andpaying fans must necessarily violate individualliberties. While a performance focused analystmight argue that the market structure permittingsuch exchanges is efficient (in that it maximizesutility or produces a Pareto better outcome), suchconsiderations are irrelevant for Nozick’s argu-ment. The point was that policies are justifiedonly when they conform to an antecedently deter-mined set of moral or political rights. The conse-quences produced by structures conforming (orfailing to conform) to this set are irrelevant.

Just as it is possible to differ over the per-formance criteria used to evaluate policy, it ispossible to disagree about which rights belong inthe template used to evaluate policy, For Nozickand other libertarians, the template will be nar-rowly confined to those that protect individualsfrom interference by others. For liberals such asRonald Dworkin (1977) or Henry Shue (1980),the list of rights may be much more expansive.The point here is that these approaches to policyare structurefocused. They establish a burden ofproof met only by 1) demonstrating that policyconforms to the antecedently chosen structuraltemplate; or 2) refuting the claim that a givenright belongs in the template. Arguments thatstress trade-offs, efllciency or other features ofend states do not meet either test.

Febrwy 931page 14 Journal of Food Distribution Research

FRAMEWORK ETHICS

Situation

Structure 4 ➤ Rights Theory

Conduct ● ➤ Virtue Theory

Performance 4 ➤ Utilitarian Theory

Figurel: How Ethica Beara on Policy

The tension between end state and structurefocused policy evaluation has a ready analogy inthe deba~e ov~r food labels. whethe~ requir~ bylaw or custom, labels are clearly a component ofstructure for food policy. The debateis: howshould labels be evaluated? Given a performancefocus, labels will be evaluated in terms of the endstate they produce for the producers and consum-ers of food products. Labels will be seen aseducational tools. One will want to know whetherthe label allows the consumer to make food pur-chases that more fully satisfy preferences. Policywill be seen as a trade-off between producer costs,consumer preferences, and health; and labels canaffect each of these outcomes in a variety of ways.A structure focused evaluation will see the matteralmost entirely in terms of informed consent.Labels will be seen as transforming the conditionsof consumer choice from those of mild coercion toimplied consent. Policies that protect individualconsent are acceptable; those which forecloseindividual consent must bear a very heavy burdenof proof before being judged acceptable. Labelswill be preferred even if people choose to ignorethem, or even if consumers have false beliefs thatlead them to make less than optimal choices.

Utilitarian or consequentialist ethical argu-menta, then, express norms or policy criteria thatfocus on performance, while human rights argu-ments identify characteristics of structure thatmust be in place without regard to consequences.The matter does not end there, however, It isoften a person’s conduct that is judged ethical orunethical. If a policy structure induces individualsto behave in ways that are unethical, there is a

basis for rejecting the policy. For example, apesticide policy which encourages producers tomisrepresent their use of chemicals would bejudged unethical, even if no rights are violated, orif no harm is done.

Many authors have taken up conduct-focused ethics in the past two decades. BernardWilliams (Smart and Williams, 1972) criticizedutilitarian arguments because they fail to addressthe character of the moral agent. AlisdairMacIntyre (1981) has criticized both rights theoryand utilitarian arguments for the emphasis thatthey place on an individual’s self-regarding wants.He argues that a better approach would take upvirtues and vices that are the reference points formoral character. These authors do not discusspolicy, however. The public choice approach topolicy analysis makes it possible to see how thesephilosophical ideas bear on policy by making itclear how situation and structure produce conduct.There is little doubt that most citizens wouldregard the conduct-focus as the most obviously“ethical,” despite the relative lack of attention ithas received by policy analysts. Public outrageover Congressional check bouncing, for example,is almost certainly focused upon conduct, ratherthan structure or performance.

With respect to food safety policy, the mostlikely relevance of conduct-focused evaluation isnot to producers and consumers, but to the ccm-duct of policy makers themselves. Argumentsagainst the practice of pricing life, for example,are best analyzed as an objection to the practice ofquanti~ing the value of life. Amette Baier

Journal of Food Distribution Research February 93/psge 15

(1986), Allan Gibbard (1986), and DouglasMacLean (1990) have expressed the concern withpricing life in this way. The problem cannot beresolved by adjusting the amount of valueassigned to lives; nor is the objection based on thesuggestion that human lives should be assignedinfinite value. The point is that persons of goodcharacter do not make decisions by attempting todecide how much others are worth. A policyprocedure which requires public servants toengage in such conduct is, in this view, a corruptand indefensible procedure, and the policies thatresult from it are tainted.

To sum up, ethics bears on policy in threeways. Traditional human rights arguments focusupon structure, applying a template of anteced-ently determined constraints in their assessment ofpolicy. The language of virtue and integrityfocuses on conduct, evaluating a policy in termsof the patterns of behavior it promotes. Finally,the end state or performance evaluation of policythat has become a staple of the social sciencesdraws upon the utilitarian tradition in ethics.Each approach establishes a burden of proof thatcannot be met by arguments grounded in either ofthe other two approaches. At the same time, eachapproach is deeply grounded in the culture andhabits of contemporary Americans.

Food Biotechnology, Policy and Ethics

An analysis of food labels that stressesrights and consent will establish very differentburdens of proof than does an analysis that evalu-ates labels according to their consequences.Matters of character and conduct enter only indi-rectly into the disputed areas of policies for con-sumer information, but could be decisive to theextent that they break a deadlock between thosefocused on structure and those focused on perfor-mance or outcome criteria. Recombinant bovinesomatotropin (BST) is the case that has spurreddebate. The substantive ethical issum raised bythe development and proposed release of BSTconcern unintended consequences for the dairyindustry, dairy animals, and for environmentalimpact of dairying. However, it is public accept-ability of milk produced using the new technologythat has produced the greatest anxiety (Thompson,1992). Opponents of BST have called for a ban

on the technology, and, short of that, for labelingof milk produced using recombinant BST(Hanson, 1991). The ethical evaluation of label-ing policies for food biotechnology has thereforealready assumed practical importance (see alsoHopkins, Goldberg and Hirsch, 1991).

The case of BST should serve as motivationfor thinking about ethical issues, but it is a poorexample of the scientific issues that will arise inconnection with the use of biotechnology for foodproducts. At least some of these products willpose diftlcult questions for risk analysis, and somemay pose quantifiable risks. By contrast, there isvirtually no scientific support for questioning thesafety of BST milk (Kroger, 1992). Future policydecisions will almost certainly be characterized bythe kinds of uncertainty that have hindered theapplication of science to public policy in the regu-lation of artificial sweeteners (Merrill and Taylor,1986) or of chemicals (Graham, Green andRoberts, 1988). In such cases, questions aboutthe extrapolation of data from animal studies, orabout the applicability of epidemiological datacaused regulatory policy to become embroiled intechnical and methodological disputes. Criteriafor scientific judgment and cross disciplinaryconflict over patterns of scientific inference arecrucial to the policy debate. Because the proce-dures and norms for scientific enquiry are them-selves matters of philosophical dispute, it is accu-rate to describe risk policy debates as philosophi-cal controversies (Hollander, 1991), Debatesover acceptable evidence extend philosophicalcontroversy into the interpretation of situation, ofthe basic facts that must be accepted as constraintsupon available policy options, These debates arenot, h~ever, ethical debates that conform to thepattern described above.

Although the debate over acceptable evi-dence will almost certainly recur in future foodpolicy decisions, the lack of scientific or technicalcontroversy over BST makes it a good examplefor considering ethical issues precisely becausedisagreement about the probability of harm doesnot confuse tie ethical issues. The biologicalfacts that make up the situation for BST milk arenot themselves a source of controversy, at leastnot among scientifically informed participants.Nevertheless, labeling requirements for BST milk

February 931page 16 Journal of Food Distribution Research

have been proposed. A policy that certified orrequired labels should be understood as an altern-ative to policies which regulate by removing orapproving products tout court. Labels thus becomea component of policy structure, to use the termi-nology introduced here. It is worth noting, how-ever, that labels might become a component ofstructure in any of several ways. One mightrequire labels that proclaim the presence of BSTmilk, or one might permit the use of labels thatcertify its absence. In either case, the precisewording of labels will be extremely important, aswill the procedures for assuring the integrity oflabels. The diversity of approaches to labelingimplies that it is not one policy proposal that isbeing discussed here, but a general class of poten-tial policies that would be evaluated in similarways.

Whatever labeling strategy is employed, apolicy using product labels can be expected tostimulate certain patterns of conduct by consumersand by processors and manufacturers. Someconsumers will read labels and will use informa-tion as a basis for food purchases; others will not.One would presume that consumers expressingconcern over BST in milk would use the label,while others might not. These patterns of conductwill lead to consequences that determine the per-formance of a labeling policy. Relevant conse-quences certainly include health benefits or risksto food consumers that are incurred as a result oftheir conduct. They also include costs to con-sumers in the form of higher food prices, and inthe trouble and inconvenience required for readinglabels. Costs to processors and manufacturers arealso a component of the policy’s performance.Given this account of situation, structure, conductand performance, it is possible to examine theethics of a labeling policy for BST.

Pe~ormanceFocussedEvaluationofhbels:As noted, product labels can be expected to pro-duce certain costs and benefits for consumers andproducers. In the case of BST, the scientificconsensus is that the health benefit to a personwho would use such a label is zero, Consumerswho express concern over the use of BST wouldderive some benefit from reduced anxiety if theyare able to satisfy their preference for non-BSTmilk. These benefits must be weighed against the

direct costs of labeling, costs which maybe signif-icant when their impact upon processing isassessed, Even the approximate value of thesecosts and benefits is largely speculative, but thepoint here is to see how consequence assessmentprovides an ethical basis for the evaluation ofpolicy, The policy is justified in terms of theacceptability and desirability of its consequences.The historical standard has been the utilitarianmaxim proposed by Jeremy Bentham in 1789: actso as to produce the greatest good for the greaternumber of people. Although there are many casesin which pure optimization rules such as the utili-tarian maxim may need to be modified (Thompsonand Stout, 1991), policies which do not providebenefits that compensate for their direct cost ofimplementation to government and to affectedindustries will always be difllcult to justify.

Structure Focussed Evaluation of Labels:Structure focused evaluation centers upon protec-tion of rights as a precondition to ethically legiti-mate application of state power. Two key criteriaare consent and fairness. The principle of govern-ment by consent of the governed is, in manyrespects, the foundational norm of democraticgovernment, while fairness, understood as equalitybefore the law and protection of minority rights,constrains the excesses of democratic decisionrnahg. Labels are an attractive component ofstructure because they make it possible to arguethat individual food consumers have been placedin a position to grant or withhold consent to foodborne risks (real or alleged), A commmer whochooses to purchase a labeled food item can legiti-mately be understood to have consented to thetransaction, so long as meaningful alternatives areavailable. A policy structure which does not allowconsumers to discriminate on criteria they havejudged to be important violates consent criteria.However, labels can also raise questions aboutfairness to the food industry. If the institutionalpractice is to use labels only in cases where seri-ous risks to health have been scientifically demon-strated, as has been the case for tobacco, then theapplication of a label to BST milk may violate therights of industry by unfairly prejudicing consum-ers against the product.

btduct Focussed Evaluation of Labels:Character and virtue are less clearly related to

Journal of Food Distribution Research February 93/page 17

labeling policy than are rights and consequences.It is not obvious how the presence or absence oflabels for BST milk would induce consumers toengage in unethical conduct, Some Americansmay be inclined to make moral judgments ofcharacter based upon a person’s dietary choices.Some religions require a dietary regimen for thedeviant, for example, and it is already commonfor vegetarians and non-vegetarians to make moraljudgments about one another. Even so, there islittle public consensus for such judgment. Themore relevant conduct is that of industry. To theextent that labels represent a form of disclosurethat would be required by norms of honesty ortruth-telling, a practice of labeling might bethought to promote ethical conduct on the part ofindustry. Ironically, disclosure will win far morepraise if it is voluntary. Hence, a labeling policywill win more praise from those who focus onconduct if it facilitates, but does not require,disclosure of relevant information. Conduct eval-uation does not provide a clear mandate for oragainst labels, however, and it will not be empha-sized in the comparison which follows.

Comparing Ethical Approachesfor Policy Evaluation

Structure focussed and performancefocuss~ approaches to policy evaluation establishdifferent and sometimes contradictory burdens ofproof. Evidence and argument which is highlyrelevant to a performance evaluation is oftenirrelevant to an evaluation in terms of rights andconsent.

There are a number of difficult philosophi-cal and economic measurement problems that mustbe addressed within a performance focussed evalu-ation (Giere, 1991). Two general points of dis-pute concern the quality of models used to predicthealth consequences, and the choice of a decisionrule to compare consequences, once they havebeen assessed. The debate over linear vs. thresh-old extrapolation of data (Schaffner, 1991) is anexample of the first problem. The debate overDelaney vs. de minimus is an example of thesecond (Jasanoff, 1991). The point here is to seethat the crucial burdens of proof established withina consequence evaluation approach differ fromthose in a rights-based procedure. Within a struc-

ture focussed evaluation, consumer choice isimportant to the extent that it satisfies criteria ofconsent; questions of whether consumers are madebetter off by the choices they make are irrelevant.

Rights based approaches to social theoryhave never assumed that governments have anyresponsibility to make socially optimal policydecisions. Rather, the first ethical responsibilityof government is a negative one: not to interferein the personal liberties or freedoms of its citi-zens. Accordingly, structure focussed evaluationof public policy stipulates a list or template ofrights, liberties, and possibly opportunities, muchlike the U.S. Bill of Rights. A policy is evaluatedin terms of its ability to satisfy or fit this anteced-ently determined template of rights. The list ofrights is adopted prior to entertaining any particu-lar policy option, so performance evaluation ofspecific policies is not a component in justifyingthe inclusion of a given right. The key right withrespect to food consumption is a general right tonon-interference in personal choices, provided thatpersonal choices do not violate complementarynon-interference rights of others.

The example of labels for BST provides anillustration of why performance focussed andstructure focussed approaches to understandinghow ethics bears on policy introduce distinctburdens of proof. A more detailed analysis oflabels would need to take up additional issues.For example, the traditional norm of caveatemptor has historically served as an informalcomponent of structure for consumer food deci-sions. The previous discussion of labels hasassumed that consumers have a right to any infor-mation4hey deem relevant about food choices, butcaveat emptor might be taken to qualify this right.Further discussion of issues specific to labelingpolicies cannot be undertaken within the con-straints of this paper. However, readers should beadvised that what has been said to illustrate thecontrast between structure and performance is farfrom being a complete ethical analysis.

The contrasting burdens of proof for perfor-mance philosophies and structure philosophiespresent a general problem for biotechnology poli-cy and food safety. Interested parties, includingscientists and regulators, can be so closely wedded

February 931page 18 Journal of Food Distribution Research

to one of these philosophies that they fail tounderstand the force of their opponents’ argu-ments. Someone who insists upon interpretingstructure-focussed rights arguments in terms of theend state produced by policy will simply miss thepoint of that argument. Rights arguments willappear as irrational or naive, failing to grasp theimportance of trade-offs that are thought to be themain focus of policy evaluation. It is not difllcultto find authors who appear to exhibit virtuallytotal insensitivity to the burdens of proof entailedby a focus on structure. In a 1990 article on foodsafety policy for recombinant DNA derived ani-mal growth hormones, Fred Kuchler, JohnMcClelland and Susan Offutt characterize theissue entirely in terms of performance or end-statecriteria. The idea that rights or consent are rele-vant is absent from their analysis, Doyle andMarth (1991) also offer an analysis focussedexclusively on end states.

Milton Russell (1990) has written that it isirresponsible for a public of!lcial to make policydecisions without attempting to assess the conse-quences to the fullest extent possible. He qualifiesthis commitment to performance evaluation byalso stating that “. . . legitimacy flows from anacceptance of the decision, or at least the decisionprocess, by those affected” (p. 22). This qualifi-cation stresses the importance of consent. Con-sent can involve the prediction of consequences,so long as predictions are part of an informationsharing process designed to build consensus. Buttraditional consent criteria stress the protection ofrights, and may leave the burden of predictingconsequences to the affected parties themselves.Despite his qualifying comment, Russell mayfavor performance criteria in making the predic-tion of consequences a strict requirement of ethi-cal policy making, even while he endorses theprinciple of consent, One aspect of the tensionbetween end state evaluation and principles ofconsent is the question of who assesses conse-quences.

Biologists or economists who predict theconsequences of biotechnology do not, generally,need the advice of citizens in the process of col-lecting data and making projections. They canproduce a prediction of end states without seri-ously involving citizens in their activity. The key

predictions for regulation of foods involve riskassessment and economic impact. If citizens areinvited to participate in decision making after suchinformation has been wllected, analyzed andpresented, the opportunity for a structure or con-duct focussed evaluation has been reduced, if notforeclosed, by the preponderance of evidencerelevant only to performance criteria. Sincestructure focussed criteria often stress the impor-tance of participation and consent, the lack ofcitizen participation in the scientific assessment ofrisk and of economic impact is doubly troubling.Citizens have been denied participation in theearly stagea of the process, and are faced with afinal decision procedure in which evidence that ispotentially irrelevant to consent criteria dominatea.

Structure and conduct criteria should not bearbitrarily excluded from a decision, as they arewhen scientists make risk assessments withoutsubstantial citizen participation. But this is not tosay that citizen assessment of risks should simplybe substituted for scientific ones. Frank Cross(1992) accuses me of doing just this in a recentpaper @hompson, 1990). The point is that endstate assessments cannot meet burdens of proofestablished by criteria that stress participation andconsent. Deborah G. Mayo (1991) has anextended discussion of how scientific and citizenassessments of risks might be compared. Peoplewho interpret the criticism of scientific risk assess-ment as a call for citizen assessment are display-ing a myopic focus upon performance criteria.Structure and conduct criteria establish burdens ofproof in which anyone’s assessment of likely endstates is largely irrelevant. Put another way, ifthe goal is to implement a policy that is likely tominim@e food related illness, then it seems obvi-ous that the best scientific techniques should beused to predict the incidence of illness associatedwith a given product or practice. However, if thegoal is to ensure that consumers have confidencein the food supply, an altogether different policymay be indicated. Consumer conildence may bevery imperfectly correlated with the probability ofillness. Confidence may be more closely corre-lated with participation and consent, with thestructure under which dietary decisions are made,rather than the end state that is produced.

Journal of Food Distribution Research February 93/page 19

The implication is that, while performanceevaluation is “objective, “ in the sense that it doesnot systematically favor any specific interests orpolitical ideology, the practice of performingextensive scientific studies can introduce a biasagainst ethical criteria that emphasize structureand conduct. Yet it is easy to find examples ofstructure or conduct focused criteria in the historyof American government. The framers of theU.S. Constitution were themselves structure-focussed in adopting the Bill of Rlghta, and con-duct-focused in proposing the division of powers.It is therefore reasonable to presume that demo-cratic decision making ought not be systematicallybiased against structure and conduct criteria,Barring specific arguments to the contrary, policymaking procedures should avoid domination byany one of these three philosophical approaches toethics, and should weigh evidence and argumentsin terms of the burden of proof to which they aremost clearly relevant.

None of this is to suggest that structure-focussed or conduct-focussed ethical criteria oughtto determine policy choices unilaterally. The risksand economic consequences are obviously relevantto policy choice. The obvious ideal is when allthree approaches to policy evaluation converge onthe same choice, but there may be good reasonswhy impacts or end-states should be the overrid-ing considerations in making some policy determi-nations where they conflict. It will be very diffi-cult to understand or state those reasons, however,if one is so deeply committed to consequentialistor utilitarian thinlchg that it is impossible tounderstand how someone could see it anotherway. The main contribution of ethics to betterpolicy making is to show how the opportunitiesfor conflict and consensus rest upon alternativevisions of the right and the good. A policy analy-sis framework in which situation, structure, con-duct and performance have been clearly distin-guished can contribute to that end.

Bibliography

Baier, Annette. 1986. “Poisoning the Wells,” inValuesat Risk, Douglas E. MacLean, ed.,Totowa, NJ: Rowman and Allanheld, pp.49-74.

Cross, Frank, B. 1992. “The Risk of Relianceon Perceived Risk,” in Risk: Issues inHealth and Safety, 3(l): 59-70.

Doyle, Michael P. and Elmer H. Marth. 1991.“Food Safety Issues in Biotechnology,” inAgricultural Biotechnology: Issues andC%oices,Bill R. Baumgardt and MarshallA. Martin, cd., West Lafayette, IN: PurdueUniversity Agricultural Experiment Station,pp. 55-80,

Dworkin, Ronald. 1977. T&ng Rights Seri-ously, Cambridge, MA: Harvard UniversityPress,

Gibbard, AlIan. 1986. “Risk and Value,” inValues at Risk, Douglas E. MacLean, ed.,Totowa, NJ: Rowman and Allanheld, pp.94-112.

Giere, Ronald. 1991, “Knowledge, Values, andTechnological Decisions: A DecisionTheoretic Approach,” in Acceptable Evi-dence: Science and Values in Risk iUan-agement, Deborah G, Mayd and RachelleD. Hollander, eds., Oxford: Oxford Uni-versity Press, pp. 183-203.

Graham, John D., Laura C. Green and Marc J.Roberts. 1988. In Search of Safety:Ckemicals and Clmcer Risk, Cambridge,MA: Harvard University Press.

Hanson, Michael. 1991. “Consumer Concerns:Give Us The Data,” in Agricultural Bio-technology at the Crossroads:Biological,Social and Institutional Clmcerns, Ithaca,NY: National Agricultural BiotechnologyCouncil, pp. 169-178.

February 931page 20 Journal of Food Distribution Research

Hollander, Rachelle D. 1991. “ExpertClaimsand Social Decisions: Science, Politics, andResponsibility,” in Acceptable Evidence:Science and Values in Risk Management,Deborah G. Mayo and Rachelle D.Hollander, eds., Oxford: Oxford UniversityPress, pp. 160-173.

Hopkins, D. Douglas, Rebecca J. Goldberg andSteven A. Hirsch. 1991. A MutableFeast.’Assuring Food Safety in the Era of GeneticEngineering, New York: EnvironmentalDefence Food).

Jasanoff, Sheila. 1991. “Acceptable Evidence ina Pluralistic Society,” in Acceptable Evi-dence: Science and Values in Risk Man-agement, Deborah G. Mayo and RachelleD. Hollander, eds., Oxford: Oxford Uni-versity Press, pp. 29-30.

Kroger, Manfred. 1992. “Food Safety andProduct Quality, “ in Bovine Somatotropinand Emerging Issues: An Assessment,Milton, C. Hallberg ed., Boulder, CO:Westview Press, pp. 265-270.

Kuchler, Fred, John McClelland, and Susan E.Offutt. 1990. “Regulatory Experiencewith Food Safety: Social Choice Implica-tions for Recombinant DNA-DerivedAnimal Growth Hormones,” in Biotechnol-ogy: Assessing Social Impacts and PolicyImplications, David J. Webber ed.,Westport, CT, Greenwood Press, pp. 131-144.

MacIntyre, Alisdair. 1981. Afier Virtue:A Studyin Moral Theory, Notre Dame: Universityof Notre Dame Press.

MacLean, Douglas E. 1990. “Comparing Valuesin Environmental Policies: Moral Issues andMoral Arguments. ” ValuingHealth Risks,Cbsts, and Benejits for EnvironmentalDecision Making: Report of a Conference,Washington, DC: National Academy press,pp. 83-106.

Mayo, Deborah G. 1991. “Sociological VersusMetascientific Views of Risk Assessment,”in AcceptableEvidence: Science and Valuesin Risk Management, Deborah G. Mayoand Rachelle D. Hollander, eds., Oxford:Oxford University Press, pp. 249-279.

Merrill, Richard A. and Michael R. Taylor.1986. “Saccharin: A Case Study ofGovernment Regulation of EnvironmentalCarcinogens,” in Agriculture and HumanValues, 3(1 and 2): 33-73.

Nozick, Robert. 1974. Anarchy State andUtopia, New York: Basic Books.

Rawls, John. 1971. A Theory of Justice,Cambridge, MA: Belknap Press of HarvardUniversity Press.

Roberts, Tanya and Eileen van Ravenswaay.1989. ‘The Economics of Safeguardingthe U.S. Food Supply,” EconomicResearch Service, Agriculture InformationBulletin Number 566, U.S. Department ofAgriculture.

Russell, Milton. 1990. “The Making of CruelChoices, “ in Valuing Health Risks, Cbsts,and Benejits for Environmental DecisionMaking: Report of a Clmference,Washington, DC: National Academy Press,pp. 15-22.

Schaffner, Kenneth F. 1991. “Causing Harm:Epidemiological and Physiological Conceptsof Causation, ” in Acceptable Evidence:Science and Values in Risk Management,Deborah G. Mayo and Rachelle D.Hollander, eds., Oxford: Oxford UniversityPress, pp. 204-217.

Schmid, Allen. 1987. Property, Power, andPublic Choice: An Inquiry into J2ZWandEconomics, 2nd Edition, New York:Praeger Publishers.

Journal of Food Distribution Research February 93/page 21

Schmid, Allen, James D. Shaffer and Eileen O.van Ravenswaay. 1983. “CommunityEconomics: Predicting PolicyConsequences,” Department of AgriculturalEconomics, Michigan State University, EastLansing, MI.

Shue, Henry. 1980. Basic Rights: Subsistence,Afluence, and U.S. Foreign Policy,Princeton: Princeton University Press.

Smart, J. J. C. and Bernard Williams. 1973.UtilitarianismFor and Against, New York:Cambridge University Press.

February 931page 22

Thompson, Paul B. 1990. Tlisk Subjectivismand Risk Objectivism: When Are RisksRed’?” Risk: Issues in Health and Safety,Concord, NH: Franklin Pierce Law Center,1(1):3-19.

Thompson, Paul B. 1992. “Ethical Issues inBST,” in Bovine Somatotropin andEmerging Issues: An Assessment, MichaelC. Hallberg ed., Boulder, CO: WestviewPress, pp. 33-50.

Thompson, Paul B. and Bill A. Stout. 1991.Beyond the Large Farm: Ethics andResearch Goalsfor Agriculture, Boulder,CO: Westview Press.

Journal of Food Distribution Research