Date:
Location:
Speaker: Barbara Fried, William W. and Gertrude H. Saunders Professor of Law, Stanford University
In exploring the question of her talk's title, Barbara Fried focused specifically on whether there is a coherent alternative to cost-benefit analysis for risk regulation. Fried discussed the possibilities and shortcomings of non-welfarist (and in particular deontological) approaches to the problem of harm to others. She concluded that the principal approaches to date could not yield a coherent alternative to welfarist aggregation.
Fried began by noting that the central disagreement between deontological approaches and welfarist ones concerned the moral significance of the fact that a course of conduct will, or is very likely to, result in harm to actual persons. The fact that welfarist approaches like cost benefit analysis aggregate costs and benefits to all, thereby permitting conduct that harms one group of people in order to help another, is regarded by deontologists as one of its most troubling features.
Fried noted that most liberal deontologists acknowledge that some tradeoffs are inevitable, and that welfare considerations must and should be taken into account in making them. But they treat a prohibition on harm to others as a presumptive side constraint on the imposition of any risk that might otherwise be recommended by cost-benefit analysis. Fried suggested that this insistence on side constraints was not a meaningful alternative to welfarist approaches in deciding what conduct one can engage in and what one cannot. She attributed this misstep to two problems in the non-welfarist literature. First, she attributed it to confusion between the allocative question of what risky conduct we permit in society and the distributive question of who should bear the costs associated with permitting or prohibiting the risky conduct. While the two questions are interconnected, they are nonetheless distinct. Secondly, she suggested that the theory of compensation has been given prominence over a theory of action, because the institutional requirements of the law (specifically, tort law) have focused attention on the former.
Fried then distinguished between ex post and ex ante analyses of risk, and contrasted three approaches. The first is a simple ex post analysis, whereby harm is a necessary precondition for judging conduct wrongful, and sufficient to create at least a prima facie case of wrongfulness. This approach suggests that we cannot judge rightness or wrongness of conduct until we know what harm issues from it.
The second is an ex post analysis, accompanied by a "peek back" to ex ante. With this approach, harm is a necessary, but not sufficient, factor for judging wrongness. This approach is prominent in the legal literature. While actual harm is treated as necessary for liability, it is not sufficient. A finding of "negligence" (phrased variously as "wrongful conduct," or "an absence of due care," "fault," etc.) is required as well. Whether the conduct was negligent or not is established by a "peek back" to the position of the actor at the time of choosing the action. Fried suggested that the determination reached from the hypothetical ex ante perspective may be hard to distinguish from some version of cost-benefit calculus.
The third approach is an ex ante approach, with a "peek forward" to ex post. This approach seeks to answer the question of what (risk of) harm to others the community would agree, ex ante, each member should be permitted to impose on the others. If these are really ex ante thought experiments, then actual harm plays no role in answering that question, as it has to be answered before knowing whether actual harm will ensue. Fried suggests that contractualist accounts usually reflect a hybrid of ex ante and ex post perspectives. To the extent that contractualist accounts endow bargainers with ex post knowledge of consequences and permit them to use that knowledge in deciding what they will agree to, the account converts to simple ex post analysis.
Fried examined the first and third approaches. In the first approach, deontologists typically divide conduct into two categories: that which has harmed others, and that which imposes a risk of harm. Conduct which produces harm is judged under deontological norms which presumptively prohibit all such conduct. Conduct which merely imposes a risk of harm creates a potential for a rights violation. But since that potential may never ripen into actual harm, a number of deontologists have concluded that the mere imposition of risk cannot constitute a rights violation. This appears to put morality in an untenable position: we cannot judge the permissibility of risky conduct until it is too late to prevent the harm that results from it. Recognizing that position to be unintuitive and morally untenable, numerous deontologists (Fried mentioned John Goldberg, Benjamin Zipursky, Judith Thomson, Stephen perry, and Claire Finkelstein) have sought to bring risk into the deontological fold. Some have done so by treating imposition of risk as a completed harm (by, for example, causing anxiety). Others have done so by treating the right to be free from "risk of harm" as a secondary right derived from the core interest not to be harmed. Fried suggested that these theories return us to the quandary of condemning all action, because all action poses a risk of harm. She also questioned whether other approaches that suggest, for example, a right not to be subject to conduct that expresses contempt (suggested by Judith Thomson) or a right to self defense (suggested by Dworkin) are able to capture satisfactorily what exactly is wrong with exposing others to a risk of (serious) harm.
The third (contractualist) approach, Fried emphasized, expresses the idea of mutual respect not by prohibiting conduct that will harm others, but by insisting on a procedure for resolving its permissibility that treats everyone equally. The procedure favored by contractualism is to require parties to chose rules to govern future conduct ex ante, before they know exactly who they will be ex post and how they will be affected by the rules. Applied to the problem of harm, such ex ante procedures are likely to produce something very close to what would result from a cost-benefit calculus, for reasons that are very familiar going back to Harsanyl and Rawls. It need not be straight aggregation of costs and benefits; one could weight for risk aversion, and also take certain choices off the table if we think people will systematically regret them ex post. Like a cost-benefit calculus, it also leaves open the question of compensation. Thus, the community could decide, ex ante, that they wanted to authorize a vaccine that would minimize aggregate disease, but also to compensate anyone harmed by the vaccine.
However, in practice, many contractualist accounts of harm to others are not pure ex ante bargains. To the extent accounts also endow parties to the contract with the effective power to block any agreement that will turn out badly for them – either directly, with a veto, or indirectly, by appealing to other people's sense of fairness as to what they would choose if they knew they would be the victim killed by the vaccine – we recreate the problem with the ex post perspective raised by the first approach: Whenever ex post compensation will not actually make the victim whole, in the sense that they would be indifferent ex ante to harm or compensation, the would-be victims will always veto the proposed conduct, making all action impermissible.
Fried offered a provisional conclusion: Non-welfarist accounts have not articulated a coherent alternative to some form of cost-benefit calculus for figuring out what risky conduct should be permitted to go forward in society. She suggested that those approaches that do not simply boil down to a cost-benefit calculus have tried to split the difference between an ex ante and ex post perspective, but that such a compromise is not possible: Everything falls on one side or the other of the critical divide, which is whether, at the moment of hypothetical choice, one does or does not have information about one's expected position ex post that differentiates it from aggregate expected value, and whether one has the power to block that choice, should one wish to.
Finally, she suggested that using non-welfarist accounts to guide ex ante choices and action recreates the paradox of moral luck. In this regard, she noted the possibility that one may believe one acts prudently ex ante and nonetheless regret it ex post (invoking Bernard Williams on this point). She suggested that such inconsistent judgments are irreducibly part of what it is to be human, to live always in the ex post and not in the alternative worlds that might have been. Rather than treating our attachment to the ex post perspective as reflecting cognitive or moral error, she suggested we must recognize it as a psychological/emotional necessity that requires some compromise with rational thought.
Katharine G. Young, Graduate Fellow in Ethics 2007-08