“1) Punishments get scaled by the judged likelihood of guilt, i.e. judge says there’s a 65% chance Bill is the killer, Bill gets 65% of the punishment.”
Almost. There is a level of probability below which there is “reasonable doubt”. Let’s say that would be 70%. So if the probability of me being guilty is below that I go free. If the probability is above that I get the punishment in relation to the level that the probability is above it. So if my probability of being guilty is 80% I’d get 1⁄3 of the punishment.
It’s extremely important that the conditional probability of being punished, given innocence, is very low. Most people’s decision-making process “overweights”(*) the importance of small probabilities. For example, they prefer a sure $100 over a 0.95 chance of $110 and a 0.05 chance of -$10, even when by most measures their utility of money is essentially linear in this range.
The deterrent effect of the law, therefore, drops substantially if people start thinking “even if I commit no crime, I could very well be convicted anyway.” The psychological distance between expectations of punishment conditional on crime vs innocence can diminish by a lot, if the chance of punishment of the innocent rises just a little.
(*) Why the scare-quotes? By standard theories of utility, which I regard with some suspicion, it is wrong to weight possible outcomes non-linearly with their probability.
The deterrent effect of the law, therefore, drops substantially if people start thinking “even if I commit no crime, I could very well be convicted anyway.”
Rather tangential, but it seems to me that this could be highly relevant to a nasty spiral in certain marginalized groups that (probably correctly) perceive themselves as being unusually likely to be falsely convicted.
“1) Punishments get scaled by the judged likelihood of guilt, i.e. judge says there’s a 65% chance Bill is the killer, Bill gets 65% of the punishment.”
Almost. There is a level of probability below which there is “reasonable doubt”. Let’s say that would be 70%. So if the probability of me being guilty is below that I go free. If the probability is above that I get the punishment in relation to the level that the probability is above it. So if my probability of being guilty is 80% I’d get 1⁄3 of the punishment.
It’s extremely important that the conditional probability of being punished, given innocence, is very low. Most people’s decision-making process “overweights”(*) the importance of small probabilities. For example, they prefer a sure $100 over a 0.95 chance of $110 and a 0.05 chance of -$10, even when by most measures their utility of money is essentially linear in this range.
The deterrent effect of the law, therefore, drops substantially if people start thinking “even if I commit no crime, I could very well be convicted anyway.” The psychological distance between expectations of punishment conditional on crime vs innocence can diminish by a lot, if the chance of punishment of the innocent rises just a little.
(*) Why the scare-quotes? By standard theories of utility, which I regard with some suspicion, it is wrong to weight possible outcomes non-linearly with their probability.
Rather tangential, but it seems to me that this could be highly relevant to a nasty spiral in certain marginalized groups that (probably correctly) perceive themselves as being unusually likely to be falsely convicted.