35 Comments

It internalizes considerably more than the entire harm. First (which I haven't remarked on yet) the victims pay for their own damages even if their harm was not their fault at all. (So, they internalize harm they did not cause.) This is probably a more significant point than my comparable one about contributory negligence, but together with the double payment for harms caused by contributory negligence, it shows that the best estimate of the amount of harm with same available information is a proper fraction of the payments you propose.

Expand full comment

"make all responsible parties pay the damage levels of all other parties hurt by the accident" when we add a payment for those other damages to one's own damage, one does internalize the entire harm.

Expand full comment

I guess that what you're saying is that trying to fine-tune the optimal levels of care and activity (as in debates over negligence versus strict liability) is less important for efficiency than is reducing litigation costs. I was looking at Posner's text, and litigation costs are implicitly treated as less important than optimizing care and activity. (I have no idea personally about the relative importance of these costs.)But to evaluate this proposal on its own terms: it seems wrongheaded to set the amount paid out to the government by those who cause the accident the amount of damage they participated. Given that economic efficiency is maximized when persons causing the accident are compelled to internalize the harm they cause, why not set the amount to be paid to the fraction of the total damage in which they participated that best approximates (over all accident perpetrators in general) the amount of damage? (Otherwise, you are causing too much care and too little activity, when this is correctable on average by charging a fraction of the participant damages.) Applying the fraction doesn't make the procedure any more complicated.

Expand full comment

"optimal levels of care and activity … simple system that gets pretty much all of these things right"

Expand full comment

Are you familiar with New Zealand's system? At the risk of over simplifying: it's like an expanded workers compensation scheme - government pays all damages regardless of identification of responsible party according to a schedule of damage payments, but also fines responsible parties on a negligence basis. No one can sue anyone.

This type of system has taken over workplace accident liability pretty much throughout the western world,. Workplaces used to be the primary source of accidents, and maybe still are but not to the degree they once were. Hundreds of workers used to die on major construction projects. The system seems to work better than any type of civil system.

Expand full comment

Today, I’ll post on a more efficient system of accident liability, which similarly raises more revenue at a lower cost.

I was thinking, did Robin even give an argument for the efficiency of his system? Rereading, I find it very strange that raising more revenue at lower cost is - in an accident liability system - the criterion for efficiency.

[Anyway, what are the limits on how much revenue can be raised by an accident liability system? If that were really the object, there would be obviously more "efficient" systems" (like making the accident perpetrators pay a still higher amount to the government).]

This is so obviously ridiculous, that I'm inclined to think I must be misunderstanding you. Well, I put the blame on the exposition. ;)

Expand full comment

I didn't specify how "court must determine responsibility, i.e., who caused the accident." There are obvious tradeoffs of saying more vs less people responsible.

Expand full comment

It seems suspicious that the system is "discontinuous" - if you're tangentially involved in an accident, you assume as much liability as someone greatly involved. Do you have in mind that the courts would cut off under some threshold?

Expand full comment

I think you are assuming strict liability, but that is not our usual liability rule.

Expand full comment

Your idea is only an incentive improvement to the extent that accident victims are not taking enough care to prevent their own victimization due to the possibility of getting liability payment. Seems like exactly the same idea?

Having the possibility of wergild taken away would have been a significant additional incentive to be take care to prevent one's own murder, for example by being more careful about antagonizing others. The middle ages was a much less civilized time, where disagreements and disputes spiraled into violence more often than today. With the change, if you're murdered, there's now a higher chance of your surviving family falling into poverty, and that consideration would have helped to moderate people's reactions to perceived slights and injustices.

Expand full comment

That's only distantly related. Only an incentive improvement to the extend that murder victims were not taking enough care to prevent their own murder due to the possibility of gaining wergild.

Expand full comment

Something I just read in Steven Pinker's The Better Angels of Our Nature: Why Violence Has Declined seems to be a historical instance of this idea:

>Feuding among knights and peasants was not just a nuisance but a lost opportunity. During Norman rule in England, some genius recognized the lucrative possibilities in nationalizing justice. For centuries the legal system had treated homicide as a tort: in lieu of vengeance, the victim’s family would demand a payment from the killer’s family, known as blood money or wergild (“man-payment”; the wer is the same prefix as in werewolf, “man-wolf”). King Henry I redefined homicide as an offense against the state and its metonym, the crown. Murder cases were no longer John Doe vs. Richard Roe, but The Crown vs. John Doe (or later, in the United States, The People vs. John Doe or The State of Michigan vs. John Doe). The brilliance of the plan was that the wergild (often the offender’s entire assets, together with additional money rounded up from his family) went to the king instead of to the family of the victim. Justice was administered by roving courts that would periodically visit a locale and hear the accumulated cases. To ensure that all homicides were presented to the courts, each death was investigated by a local agent of the crown: the coroner.

Expand full comment

The mainstream position in law and economics (to my knowledge) is that the common law was subject to intense pressure to deliver economically efficient results. If the common law got it as wrong as Robin claims, this is prima facie surprising.

Expand full comment

And the passengers have to get separate insurance to get their medical bills paid, so this one accident would now have like a half-dozen different payments (people paying the full cost to the government, insurance companies paying the full cost? back to the victims).

Expand full comment

It seems to me the over-deterrence occurs at the level of individual incentives, rather straightforwardly if I understand correctly. On an efficient scheme, I have to make good on the costs of my negligence. But here I have to make good on the effects of others' negligence too - because both pay for the overlapping negligence.

An example to test if my understanding is correct. A and B have a collision where they are both contributorily negligent. A has several passengers. Since A and B's negligence are both implicated, both A and B must pay to the government the damages to the three passengers.

In practical terms, you are over-deterring the driving of automobiles.

Expand full comment

Are there no payouts in this system? Liability isn't just about discouraging bad behavior, it's about making the victims of that bad behavior whole, or at least ameliorating their suffering. You'd need a much more robust social insurance system (should be easy to fund given the new revenue raised) along with the new liability rules to make up for losing that aspect.

Expand full comment