EFFICIENCY OF LIABILITY RULES WITH MULTIPLE VICTIMS

. This paper investigates the structure of liability rules from the efficiency perspective when there are multiple victims. It is shown that, when there is one injurer and multiple victims, there is no liability rule with the property of invariably yielding efficient outcomes. The fact that there is no rule which is efficient for all applications of course does not in any way preclude the possibility of a rule being efficient with respect to some subclass of applications which may be of interest. We consider in this paper the important subclass of applications (′) which are such that the expected loss of a victim depends only on the care level taken by that victim and the care level taken by the injurer. It is shown that a sufficient condition for a one‐injurer multiple‐victim liability rule to be efficient with respect to the above subclass of applications ′ is that its structure be such that: (i) whenever the injurer is negligent and a particular victim is non‐negligent, the entire loss incurred by that victim must be borne by the injurer; and (ii) whenever a particular victim is negligent and the injurer is non‐negligent, the entire loss incurred by that victim must be borne by the victim himself. In fact, for an important subclass of one‐injurer multiple‐victim liability rules, characterized by the condition that the proportions in which the loss incurred by a particular victim is to be borne by the injurer and that victim must depend only on the non‐negligence proportions of the injurer and that victim, the above condition is both necessary and sufficient for efficiency with respect to the restricted subclass of applications ′.