Monday, February 1, 2010

Guest Blogger Richard Epstein on "Multiple Standards of Liability in Tort Law: Of context and categories"

At the past meeting of the American Association of Law Schools, the Torts & Compensation Systems Section, chaired by John Goldberg, devoted its session to an examination of my academic writings on the law of tort.  Over the years, I have written on many discrete topics in tort law, but the general focus of that panel was on the first work that I wrote on the subject, “A Theory of Strict Liability,” 2 J. Legal Studies 151 (1973), which was written as a self-conscious response to Richard A. Posner’s highly influential article, A Theory of Negligence, 1 J. Legal Stud. 29 (1972),

At stake in my original debate with Judge Posner was a question that is critical to the articulation of a sensible system of tort law, but of somewhat narrower scope that I had suspected at that time.  More specifically, the first and easiest class of tort cases to organize are situations that involve the physical harms that one individual commits to a stranger.  The simplest versions of this wrong are ordinary trespasses that involve the application of force by the defendant, or by instruments under his control, to the person or property of the plaintiff.  The more complex causal chains of indirect causation involve, for example, the setting of traps that are set off by the conduct of a plaintiff who, typically, is ignorant of the peril that has been set in his path.

My position was, and remains, that the simple rules of push/pull causation do better than any complex negligence inquiry that seeks to compare in familiar fashion three quantities on which it is always difficult to get a handle:  the burden of precautions as measured against the expected magnitude of the loss, which is captured by knowing the probability of harm and its severity in all alternative states of the world.  The fewer the hypothetical questions, and the simpler the causal arguments, the more reliable the judicial determinations.  In this case, it is not possible to show any reason why the use of a simple rule of these injuries leads to systematic distortions in the care levels that people will take in response to the situations that they face.

What is striking to me today, however, is not the debate over negligence and strict liability in these harm-to-stranger cases, but how both the negligence and strict liability rules do a very poor job of tracking any sensible system of liability in a raft of other contexts.  The first of these involve the occurrence of harm that arises out of some kind of consensual arrangement.  Against the general drift of modern law, I usually prefer that the tort rules in these cases be regarded as default principles that can be varied by private agreement that better reflects the joint intention of the parties.  Thinking of these tort rules as default rules raises the question of whether anyone in his right mind would want either the strict  liability or negligence rule in many discrete contexts. 

It turns out in medical malpractice, strict liability is a dead loser because the rate of accidents caused by medical intervention is high, but set off by the possibility of real benefits which provide full compensation from the ex ante perspective.  Forcing liability under strict liability shuts down the business. But the objective negligence standard, unconstrained by reference to custom, is in general a loser as well, for the same reason: it ignores the ex ante compensation from successful treatment. So the hard question is figuring out what custom entails in areas where there is a strong diversity of medical opinion.  The right way to do this is hard to guess, but the one clear loser is to carry over notions of what counts as care to avoid contact to cases of determining care levels in giving aid and comfort.  The Hand formula is useless there.

It is also quite beside the point in other contexts.  That formula maps very poorly into many occupier’s liability cases, where, especially with licensees, the question of whether a defect is known or latent has some real punch.  That same distinction should govern most product liability cases, I believe, and the modern tendency to use the risk utility test in this area almost always leads to the worst form of Monday-morning quarterbacking by judges and juries alike. A far better conception of the subject puts direct cost/benefit analysis on the back burner and asks this one question:  has the upstream provider of a good put the downstream user in a position that he or she has the needed information to make responsible decisions about its use? Avoid the traps, and let responsible parties then decide what risks to take in light of their own objectives.  Do not organize the system around the paternalist view that manufacturers have to control the behavior of downstream persons once they have lost control of their goods.

Both strict liability and ordinary negligence theories also tend to fare poorly in those cases where harms are inflicted by third parties on strangers.  One such context involves physical assaults on the defendant’s premises.  A second arises when one defendant operates as an inspector of the activities of a second responsible actor.  In both these contexts, the sensible approach is to impose general liability on the wrongful actor, without worrying about the position of, say, an insurer or inspector.  But if the wrongdoer is for some reason out of the picture, the tendency to reach out to the third party is great, if the alternative is no liability at all. 

The risk here of course is that strict liability is far too stringent.  Ironically, the same critique can be leveled against negligence liability under a protean cost/benefit analysis, which will also drive the insurers from the field when their assistance is needed most.  Yet at the same time, it is not wise to allow these persons to knowingly rubberstamp dangerous behavior of the very person over which they are supposed to exercise some independent oversight.  In the end, the best standard tries to use a recklessness standard to weave between the two extremes, buttressed in some cases by charges of “negligence plus,” which cover cases of egregious conduct that didn’t just miss some hidden danger, but which magnified the underlying risks.  These cases are critical where not only the injured party is in privity with the underlying actor, but of even greater relevance, where that party is not, as in cases of pollution or explosions, the one that hurt third persons. 

A short column cannot explore all the complex variations.  But it can point them out, as a stubborn reminder that context matters in tort, especially in those cases with strong contractual overtones.  Keeping all the balls in the air at the same time is a real challenge to any tort theorist, both within and outside of the law and economics tradition.  My sense of the field has surely evolved since I first wrote back in 1973.

- Richard A. Epstein
James Parker Hall Distinguished Service Professor of Law
University of Chicago

Guest Blogger, Scholarship | Permalink

TrackBack URL for this entry:

Listed below are links to weblogs that reference Guest Blogger Richard Epstein on "Multiple Standards of Liability in Tort Law: Of context and categories":


Prof. Epstein,
For many years, I've been curious about footnote 1 on page 331 of your Takings book. In speaking about benefits of limited liability, it reads in relevant part:

"Tort claims are a very different matter. Here the legitimation of the corporate form is best understood as a set of complex forced exchanges. As a quid pro quo for limited liability, a corporation can be properly compelled to have liability insurance to meet its anticipated risks.

Does this mean that a state can justifiably require an LLC or corporation to purchase some minimal level of general liability insurance? I'm not sure I'm willing to go that far - but I'd be curious to hear your explanation. Has your view evolved on this issue?

Posted by: Greg | Feb 1, 2010 1:57:51 PM

No. I have not retreated too much from this point of view. The whole point here is to recognize that forced exchanges have to work for the mutual benefit of all, and that cannot happen if the outsiders are forced to take large uncompensated risks. With large corporations and routine accidents, the aggregation of capital eliminates the need. So for these companies it does not matter. They are answerable anyhow. But with those risks, like with nuclear power, that are low frequency and high severity, the optimal corporate strategy could be to let it happen without insurance. That externality is one worth guarding against.


Posted by: Richard Epstein | Feb 1, 2010 4:55:52 PM

Post a comment