A taxonomy of liability rules

There are several ways in which legal systems can apportion liability be­tween parties. Historically, a broad variety of liability rules has been developed by legal systems. Most early legal systems adopted liability rules that did not depend on the fault of the tortfeasor.

We shall proceed with the presentation of some of the most common liability rules, starting from strict liability to simple negligence and more complex legal regimes. In our analysis we shall utilize the conventional terminology by which the injurer is defined as the individual who does not suffer harm in an accident and the victim as the individual who suffers such harm. In this survey, we shall focus on two-party accidents.

There are two fundamental possibilities in a two-party accident. The first occurs when both parties have to take precaution in order to avoid the acci­dent (bilateralprecaution). The second is given by situations in which either party can take precaution and successfully avoid an accident (alternative precaution). In the second case, there is a waste of precaution cost if both parties take precaution, since one party’s precaution would already have been enough. A particular and common case of alternative precaution is unilateral precaution. As in alternative precaution, one party’s precaution is enough to prevent the accident, but only one party has the actual possibility of avoiding the accident.

Strict rules: no liability and strict liability

Strict liability can be thought of as the mirror image of no liability. A party who occasions harm to another will compensate the victim regardless of who is at fault. This rule is the converse of no liability. No liability can in fact be thought of as strict liability on the part of the victim, in that the victim always bears the loss regardless of the parties’ fault. No liability and strict liability can thus be considered the limit points in the range of possible liability rules. The choice between strict liability and no liability has obvious distributive effects, in that strict liability results in the victim always being compensated, while no liability makes the victim bear all accident costs.

The different allocation of accident costs has clear incentive effects. In a strict liability system, the injurer has to bear both the cost of precaution and the expected accident cost and, hence, he/she will minimize the sum of those costs. This will lead to the efficient level of precaution. On the contrary, a no­liability rule does not achieve an efficient result because the injurer would bear the cost of precaution without internalizing the benefit of such precau­tion. In the absence of liability, the injurer would adopt no precaution at all, which is an inefficient result. On the other hand, if we look at the victim’s incentives to take precaution, we see that the opposite is true. Strict liability creates no incentives for victim precaution, while no liability would shift the entire residual liability on the victim, inducing optimal victim care. It follows that strict liability and no liability can give incentives to take efficient precau­tion only to one party, respectively either the injurer or the victim. Strict liability will fail to produce an efficient outcome when the avoider is the victim, and no liability will fail when the avoider is the injurer.

With respect to alternative precaution, the result is slightly different. In the case of unilateral precaution, if the tort law system fails to target the avoider, he/she will take no precaution at all, while in the case of alternative precau­tion, either party can take precaution; therefore, imposing liability on the party who is not the least-cost avoider will result in suboptimal precaution levels and excessive precaution costs. Strict liability and no liability can thus yield efficient results only in the case of unilateral or alternative precaution, provided that liability is allocated on the least-cost avoider. In the case of bilateral precaution, both strict liability and no liability fail to generate opti­mal incentives, because neither rule can simultaneously threaten both parties with liability in a Nash equilibrium. In bilateral-precaution situations a differ­ent rule is therefore needed to induce both parties to adopt the necessary precautions.

Negligence rules in general

Fault can be seen as a way of creating optimal incentives on both tortfeasors and victims and also of achieving efficiency in the case of bilateral precau­tion. Negligence rules draw a line between liability and no liability by identifying a level of due care and verifying whether the relevant party adopted that level of due care. American case law in a sense anticipated the economic definition of negligence, adopting the simple and formal logic of cost-benefit analysis to adjudicate tort cases. Already in 1947, Judge Learned Hand, in the celebrated decision of United States v. Carroll Towing Co. (159 F.2d 169 (2d Cir. 1947)), clarified the tradeoffs between the costs and benefits of risk and prevention activities using a mathematical formula. This rule became a milestone in the law of torts, and it is now known as the Hand formula of negligence. The formula defines negligence as a function of three variables: (a) the probability of a harmful event occurring (magnitude of risk); (b) the seriousness of the damage that may result from this event (gravity of harm); and (c) the cost of preventing the occurrence of the harm­ful event (burden of prevention).

Although the Hand formula does not directly consider the social value of risk-creating behaviour, it produces the proper incentives for the evaluation of such behaviour. By imposing a balance between risk and prevention, the result in Carroll Towing encourages individuals to weigh the cost of preven­tion against the utility of the behaviour. When deciding whether to engage in an activity, the reasonable person will consider whether the utility derived from the activity justifies the risk of liability and/or the cost of prevention (this is, indeed, the question of the Restatement (Second) of Torts § 291, comment (a), which asks whether ‘the game is worth the candle’). According to this logic, individuals will respond to liability rules by undertaking the socially optimal level of precaution. A vast region of law and economics literature has explored the wisdom of this tort doctrine, often with the use of formal economic models, bringing to light the importance of using marginal (rather than total) values in the assessment of liability. Along similar lines, after establishing a positive economic model of tort law, Landes and Posner (1982) conclude that the Hand formula of negligence, as applied, coincides with the economic model of due care.

Introducing fault means setting a due level of precaution, defined by the legislator or by the judge. The due level of precaution should be set to be equal to the efficient level of precaution. Under any negligence rule the judge has to perform such a test by confronting the level of precaution actually taken by the parties with the due level of precaution.

Among such complexities is the fact that while some forms of precaution are easily observable ex post, others are very difficult, or even impossible to assess and to compare with the legal standard of precaution. In the presence of non-observable precautions, it is clear that individuals would rationally limit their investment to observable precaution to avoid negligence and re­frain from investing in non-observable precautions, since they could not draw much benefit from such investment.

In the law and economics literature the case of non-observable precautions is generally treated under the discussion of care versus activity levels. The most common example of activity level is the repetition of a dangerous action, such as driving. Although courts may occasionally take into account the frequency of an activity in their assessment of negligence, often no threshold of ‘optimal frequency’ can easily be utilized by legal rules as a liability allocation mechanism, given the difficulty of pinpointing a critical value to separate efficient from excessive activity. Since courts cannot be asked to balance unascertainable costs and benefits and cannot be asked to evaluate non-observable precaution levels, it is clear that the types of precau­tions that are evaluated for the finding of negligence are generally confined to care levels, not activity levels. Therefore, the introduction of the criterion of negligence introduces a dichotomy between care-type and activity-type pre­caution investments (Shavell, 1980a). No such distinction between care and activity level is relevant in regimes of strict liability and no liability.

Negligence rules under which the victim is the residual bearer

Hereafter we shall analyse those rules that are generally referred to as negli­gence rules. It will soon be clear that we can think of them as being constructed by adding a negligence defence to a rule of no liability.

Simple negligence Within negligence regimes, the most straightforward rule is simple negligence. Under simple negligence an injurer is liable for dam­ages only if he/she is found negligent. The victim bears the so-called ‘residual liability’, in the sense that he/she has to bear the consequences of the accident if the injurer cannot be blamed for negligence. In this sense, simple negli­gence is analogous to a no-liability rule, because it leaves residual liability on the victim.

With unilateral-precaution accidents, when the victim is the avoider, the injurer cannot be declared negligent since it is not possible for him/her to take effective precautions. Therefore, the victim bears the cost of the accident and he/she will have incentives to take the optimal level of precautions (care and activity level) as under no liability. If the avoider is the injurer, he/she will have to pay only when he/she does not take at least the due level of care. If the injurer is negligent, he/she has to bear the cost of care and the expected accident cost (pay damages to the victim). On the contrary, if the injurer takes due care, he/she avoids liability and bears only the cost of care. If due care is set at the efficient level, the injurer will have incentives to take due care. We can conclude that in unilateral-precaution cases simple negligence produces the right incentive to take optimal care when either the injurer or the victim is the avoider.

However, with respect to activity level, only the victim, as a residual bearer, has incentives to take the optimal level of precaution. In equilib­rium, in fact, the injurer will adopt due care and avoid liability, so that any investment in non-observable precautions would yield him/her no private benefit. Simple negligence thus gives efficient incentives with respect to activity level only to the victim, since he/she bears the full cost of the accident in equilibrium. The same logic allows us to show that also in the case of alternative precaution both parties face incentives to adopt optimal care levels. However, we know that in such situations efficiency requires that only the least-cost avoider to take care. If the other party or both parties adopt precautions there is an inefficient result. One way to avoid obtaining such an inefficient outcome is to formulate the negligence criterion in light of such requirement, so that negligence could be found only when the injurer is the least-cost avoider.

The introduction of the requirement of negligence improves the perform­ance of the rule in bilateral-precaution situations. The negligence criterion makes both parties take the optimal level of care in all situations (unilateral, alternative and bilateral precaution), but gives incentives to choose an opti­mal activity level only to the residual bearer, the victim.

Contributory and comparative negligence Under contributory negligence, the injurer is liable to compensate his/her victim only if he/she was negligent and the victim was careful. In all remaining cases the victim remains the residual bearer and receives no compensation for his/her loss. Therefore, the victim does not have the right to compensation when both were negligent, when both were careful, and obviously when the injurer was careful and the victim negligent. Similar results are reached with a rule of comparative negligence. In a regime of comparative negligence, however, victim negli­gence does not constitute a complete bar to recovery but leads to a reduction of liability in proportion to the parties’ respective levels of negligence.

In both regimes, the injurer can escape liability by taking the due level of care. This creates the appropriate incentive for the injurer to comply with the legal standard of care. Given that the injurer can reasonably expect to bear the entire residual loss, he/she would also face incentives to behave carefully, since he/she would internalize the full benefit of his/her precaution invest­ment. Contributory and comparative negligence thus create efficient care incentives for both parties, but only the victim, as a residual bearer, would have incentives to undertake an optimal activity level. The injurer is able to avoid liability with the adoption of due care and therefore would have no incentive to invest in non-observable precautions. Contributory and compara­tive negligence thus produce the same set of incentives generated by a rule of simple negligence, but they have possible distributive effects, because they would either foreclose or reduce compensation when the victim is found negligent.

Strict liability with negligence defences: the injurer as the residual bearer A negligence rule can also be applied in conjunction with strict liability. In these cases, the residual bearer is the injurer. Under strict liability with the defence of contributory negligence, there is a test on the victim’s fault. If the victim is at fault, he/she is barred from obtaining compensation. When there is no fault on the part of the victim, the injurer is strictly liable, regardless of his/her fault. A regime of strict liability with a defence of dual contributory negligence encompasses a double test on fault; the negligence criterion is applied to both the victim and the injurer. In this case the victim bears the accident loss only if he/she was negligent and the injurer careful: in all the remaining cases he/she is entitled to compensation. The injurer has to pay damages if both were negligent, if only the victim was careful and if both were careful. This rule generates the same incentive effects as strict liability with a defence of contributory negligence.

Under both variants of this rule of strict liability with negligence defences, the victim has incentives to take the due level of care to avoid losing his/her right to compensation in the case of an accident. If the victim is careful, the injurer bears the expected accident cost and will take the level of precaution (care and activity level) that minimizes the total cost of accidents, a level that would correspond to the socially optimal level of precaution.

Under these regimes, it is sufficient for the victim to take due care in order to be compensated for the accident loss, so that he/she does not have any incentive to take (additional) unobservable precaution. On the contrary, since the injurer is the residual bearer, he/she will have incentives to invest in both observable and unobservable precautions.

Comparative causation and loss-sharing rules

Under the liability regimes examined above, if neither party is at fault the loss is either entirely borne by the victim (negligence rules under which the victim is the residual bearer) or is shifted entirely on the tortfeasor (strict liability with negligence defences). These rules lack explicit ways for apportioning the loss between a faultless victim and a faultless tortfeasor. Historically, comparative causation emerges in the midst of legal systems based on negligence, in re­sponse to the conviction that, in the absence of fault, there is no obvious reason to let the loss fall on the innocent victim, just like there is no reason to shift it on the tortfeasor. The criterion of comparative causation allows the spreading of an accident loss among a faultless tortfeasor and an innocent victim on the basis of the relative causal contribution of the parties to the loss. The principle of comparative causation only operates as a residual basis for liability in the presence of faultless parties, avoiding the all-or-nothing allocation of liability generated by traditional rules.

In terms of levels of care, a rule of comparative causation under negligence may induce both victims and tortfeasors to adopt socially optimal levels. Comparative causation differs from traditional regimes in this respect, since both parties face positive shares of the accident loss in equilibrium. This results in the spreading of expected accident loss and activity level incentives between the parties, rather than the concentration of such losses and incen­tives on one or the other party. As a result, under comparative causation the activity level chosen by one party improves at the expense of the other. Thus, neither version of comparative causation dominates traditional negligence and strict liability rules on both activity-level margins.

The loss-sharing and resulting dilution of activity level incentives may or may not increase total net benefits. Loss spreading in equilibrium may pro­mote optimal risk allocation among risk-averse agents when insurance is not readily available. Loss spreading may similarly minimize distortion of incen­tives deriving from truncated liability when tortfeasors face large potential losses. However, comparative causation is also likely to exacerbate adminis­trative costs, given the need to ascertain relative causation and the need to adjudicate cases even in situations where neither party is at fault. This may explain the limited spread of this rule in contemporary legal systems.