I guess "breakage" only happens in extremes. If I'm liable for $1000 * 1000 I might as well be liable for anything. If I'm liable for $1000 * 2 it's different. Same issue as in scaling punishments due to probability of getting caught.
How would you compare this mechanism to betting on the outcome of a trial?
But not as a desirable element! It’s something we accept because there’s not a better way. I don’t see how deliberately introducing randomness into the equation actually makes things better here.
I fall to see the appeal. Perhaps I'm missing something. I definitely want better (Coasian) resolution of small harms, but this way seems flawed. My first objection is that I simply wouldn't trust the randomness function. It seems very easily corrupt able.
More generally—I guess I don't see how this helps the problem. You go to the courthouse, pissed off about a $50 loss, and if you lose the toss, now you're even more pissed off. Why bother to go at all? Unless you're really pissed, I guess, but even then, the uncertainty of the whole thing does not feel like it serves justice. Perhaps I'm missing something.
"In fact, forget expected utility, this is a world of lottery ticket buyers. Joe Gambler will be less scared of being sued when he feels lucky. Sally Stresspill will get ulcers from the fear of losing her house when she bumps into someone. I oppose the motion!" and the fact that utility isnt even. I wouldn't take a 50% bet on money cause risk hurts!
Diminishing marginal returns set in really quickly for compensation payments. People who sue for damages and win get most of their benefit from psychological effects like having the law validate their complaint, and feeling like the outcome was "fair". And since estimating the monetary value of damages tends to be difficult, many different payment levels can be rationalized as fair. (But not no payment.)
Judge Richard Posner was one of my dissertation advisers. He liked my lottery idea but told me it strongly went against legal culture. I asked him if he would ever use lotteries as a judge. He said no because it would get him impeached.
James, right that is the idea. But in my experience once people understand and grant that point, they still have strong reservations about using this new ability.
There is some fixed cost to bringing a lawsuit regardless of the level of damages. The smaller the level of damages the more likely that the cost of using the court system to litigate would be greater than any beneficial deterrence created by the litigation.
Let's say that the social costs of litigation = $10,000 + X(Expected Verdict),while the deterrence benefit of litigation = Y(expected verdict)where X and Y are positive but less than 1.
We only want a trial when $10,000 + X(Expected Verdict)>Y(expected verdict). This won't happen when the expected verdict is small.
But under your lottery system we only have a trial probabilistically so this reduces the expected cost of litigation but it might not reduce the deterrence value of suits. So absent a lottery system I think it best that NEV suits are not brought, but I think you might be right that under a lottery system it could be beneficial to have NEV suits.
Your idea reduces the cost of using the court system. By reducing this cost it becomes socially beneficial to have more harms subject to judicial punishment.
I guess "breakage" only happens in extremes. If I'm liable for $1000 * 1000 I might as well be liable for anything. If I'm liable for $1000 * 2 it's different. Same issue as in scaling punishments due to probability of getting caught.
How would you compare this mechanism to betting on the outcome of a trial?
How?
This may break marginal deterrence
But not as a desirable element! It’s something we accept because there’s not a better way. I don’t see how deliberately introducing randomness into the equation actually makes things better here.
We accept uncertainty and randomness in many aspects of society and justice.
Okay, that answers the first objection, presumably, but not the second. What am I missing here?
It is quite possible to create shared random processes, where we can all check that it is really random.
I fall to see the appeal. Perhaps I'm missing something. I definitely want better (Coasian) resolution of small harms, but this way seems flawed. My first objection is that I simply wouldn't trust the randomness function. It seems very easily corrupt able.
More generally—I guess I don't see how this helps the problem. You go to the courthouse, pissed off about a $50 loss, and if you lose the toss, now you're even more pissed off. Why bother to go at all? Unless you're really pissed, I guess, but even then, the uncertainty of the whole thing does not feel like it serves justice. Perhaps I'm missing something.
> The few reactions I have heard have been surprisingly (to me) negative
That's the first time I read about this, but my first reaction is positive
"In fact, forget expected utility, this is a world of lottery ticket buyers. Joe Gambler will be less scared of being sued when he feels lucky. Sally Stresspill will get ulcers from the fear of losing her house when she bumps into someone. I oppose the motion!" and the fact that utility isnt even. I wouldn't take a 50% bet on money cause risk hurts!
Diminishing marginal returns set in really quickly for compensation payments. People who sue for damages and win get most of their benefit from psychological effects like having the law validate their complaint, and feeling like the outcome was "fair". And since estimating the monetary value of damages tends to be difficult, many different payment levels can be rationalized as fair. (But not no payment.)
This is one of the best ideas I have heard for a long time!
Wait a sec...
"More generally, you could keep doubling your suit"
+
"I'd let the person being sued also risk the suit, double or nothing, as long as they showed they were good for the maximum damage amount."
=
Couldn't a sufficiently rich defendant double it over and over and over again until s/he got "tails" on the coin flip and wiped it out to nothing?
Robin,
Judge Richard Posner was one of my dissertation advisers. He liked my lottery idea but told me it strongly went against legal culture. I asked him if he would ever use lotteries as a judge. He said no because it would get him impeached.
James, right that is the idea. But in my experience once people understand and grant that point, they still have strong reservations about using this new ability.
Robin,
There is some fixed cost to bringing a lawsuit regardless of the level of damages. The smaller the level of damages the more likely that the cost of using the court system to litigate would be greater than any beneficial deterrence created by the litigation.
Let's say that the social costs of litigation = $10,000 + X(Expected Verdict),while the deterrence benefit of litigation = Y(expected verdict)where X and Y are positive but less than 1.
We only want a trial when $10,000 + X(Expected Verdict)>Y(expected verdict). This won't happen when the expected verdict is small.
But under your lottery system we only have a trial probabilistically so this reduces the expected cost of litigation but it might not reduce the deterrence value of suits. So absent a lottery system I think it best that NEV suits are not brought, but I think you might be right that under a lottery system it could be beneficial to have NEV suits.
Your idea reduces the cost of using the court system. By reducing this cost it becomes socially beneficial to have more harms subject to judicial punishment.