3 ms·

I think you are applying Goodhearts law both too broadly and not broadly enough at the same time. IMO there is no way being away from the law with it without mo

I think you are applying Goodhearts law both too broadly and not broadly enough at the same time. IMO there is no way being away from the law with it without money being involved. It equally applies in any bureaucratic system where goals get set and achieving them gives you power or any other benefit. If humans are to take action together over a prolonged time the law is gonna get in your way. The best way to mitigate this is to avoid proxy-metrics and measure accomplishing the actual goal. Another issue is that as someone smarter than me said who was quoted by Russ Roberts on econtalk: You are not allowed to ask for better people. The right system must make the wrong people do the right thing. To come back to our concrete subject at hand: Don't penalize the penal institution for reoffense, but reward them if former inmates stay out of jail. I'd make the ongoing pay for inmates being in jail just low enough to not be profitable with performance bonuses set at several intervals after the inmates were released and didn't mess up again. This somewhat opens the door for releasing inmates early. I'd need to contemplate and discuss with experts on how early released currently gets decided on to come up with a good way to address this, but it seems like there are several good-sounding options. This even might be a non-issue if the increased likelihood of a future bonus outweighs the cost of the inmate staying longer in prison.

kerkeslager 3y ago
> IMO there is no way being away from the law with it without money being involved. There are obvious ways, you're just discounting any solution that doesn't involve corporations taking profits. > Another issue is that as someone smarter than me said who was quoted by Russ Roberts on econtalk: You are not allowed to ask for better people. The right system must make the wrong people do the right thing. This argument is obviously wrong because it proves too much. We obviously can limit who we put in charge of solving problems--why not, after all, just lock everyone in a prison with no wardens and let the prisoners sort themselves out? After all our hands are tied, we can't limit who we put in charge! > To come back to our concrete subject at hand: Don't penalize the penal institution for reoffense, but reward them if former inmates stay out of jail. I'd make the ongoing pay for inmates being in jail just low enough to not be profitable with performance bonuses set at several intervals after the inmates were released and didn't mess up again. This somewhat opens the door for releasing inmates early. I'd need to contemplate and discuss with experts on how early released currently gets decided on to come up with a good way to address this, but it seems like there are several good-sounding options. This even might be a non-issue if the increased likelihood of a future bonus outweighs the cost of the inmate staying longer in prison. Obviously this is a better incentive structure than what we have, but why are you insisting putting money as an intermediary measure here? What I mean is, whatever measure of how well a rehabilitation system is working, is already going to be an imperfect model of how well that rehabilitation system is actually working. Adding a payment scheme to this adds a second imperfect model: now in addition to modeling how a rehabilitation system works, you're imperfectly modeling how the payment scheme will affect the achievement of your already-imperfect metrics. This is not necessary, and is going to make the entire system less effective. This is an unnecessary complication which you are insisting upon because of a misguided economic ideology. The same pro-corporate, anti-human ideology that created this problem is not going to be the solution to this problem. Specifically responding to this bit: "I'd make the ongoing pay for inmates being in jail just low enough to not be profitable" This means that all the corporation has to do is find a way to make a thin margin of profit beyond what you're paying them, and the incentive shifts in favor of keeping inmates in jail indefinitely. Sell inmates' private data? Sell prison labor (a.k.a. slavery)? Sell phone calls? Gouge prices on commissary? You know... the things they're already doing. How about we release members of rival gangs in pairs? If one kills the other, that's one reoffense, but the other can't reoffend any more, so that might just be profitable. What if inmates with a high reoffense risk just happened to have medically necessary amputations? It's harder to commit sex crimes if you don't have genitals, and harder to commit violent crimes if you don't have a leg. What if we just release inmates in other jurisdictions so they reoffend away from where reoffense stats are being collected? Remember: if you aren't willing to do it, your competitor will! Lots of different forms of corporate malfeasance have been illegalized in the century and a half since the industrial revolution, and yet, corporations are still harming people for profit because they keep "innovating" new ways to make a profit with no consideration for whether it harms people. It's absurd to keep playing whack-a-mole with the latest corporate "innovations" because you can't imagine any motivation to do anything besides money. And of course we want to avoid government overreach here, right? So we shouldn't regulate against things that we don't know are happening or put too many burdensome auditing requirements to actually find out if bad things are happening. We've got to give corporations a few decades of profits before some good people sacrifice their careers to smuggle out compelling evidence that bad things are happening... only then can we start making laws to protect people--the free market comes before protecting people.
- ajmurmann 3y ago
  In everything you are writing there seems to be a implicit assumption that there are incentives in place already that the monetary reward system would mess with. You made a lot of points I'd wavy to reply to but all of it is secondary to understanding that part better. I'm of the believe that if you don't set a incentive explicitly one will emerge that you didn't set. Ultimately people work in any institution to make a living, so they are also likely to try to get a promotion. What will get them that promotion if we don't set explicit goals and inventives around them? I'd argue that we are seeing that that's often times being personally liked by your coworkers and boss. That's a huge issue we are seeing in law enforcement right now where the main goal seems to be sticking together over actually serving the community. Now I know less about prison guard culture, but I'd bet similar dynamics exist.

kerkeslager 3y ago

> IMO there is no way being away from the law with it without money being involved. There are obvious ways, you're just discounting any solution that doesn't involve corporations taking profits. > Another issue is that as someone smarter than me said who was quoted by Russ Roberts on econtalk: You are not allowed to ask for better people. The right system must make the wrong people do the right thing. This argument is obviously wrong because it proves too much. We obviously can limit who we put in charge of solving problems--why not, after all, just lock everyone in a prison with no wardens and let the prisoners sort themselves out? After all our hands are tied, we can't limit who we put in charge! > To come back to our concrete subject at hand: Don't penalize the penal institution for reoffense, but reward them if former inmates stay out of jail. I'd make the ongoing pay for inmates being in jail just low enough to not be profitable with performance bonuses set at several intervals after the inmates were released and didn't mess up again. This somewhat opens the door for releasing inmates early. I'd need to contemplate and discuss with experts on how early released currently gets decided on to come up with a good way to address this, but it seems like there are several good-sounding options. This even might be a non-issue if the increased likelihood of a future bonus outweighs the cost of the inmate staying longer in prison. Obviously this is a better incentive structure than what we have, but why are you insisting putting money as an intermediary measure here? What I mean is, whatever measure of how well a rehabilitation system is working, is already going to be an imperfect model of how well that rehabilitation system is actually working. Adding a payment scheme to this adds a second imperfect model: now in addition to modeling how a rehabilitation system works, you're imperfectly modeling how the payment scheme will affect the achievement of your already-imperfect metrics. This is not necessary, and is going to make the entire system less effective. This is an unnecessary complication which you are insisting upon because of a misguided economic ideology. The same pro-corporate, anti-human ideology that created this problem is not going to be the solution to this problem. Specifically responding to this bit: "I'd make the ongoing pay for inmates being in jail just low enough to not be profitable" This means that all the corporation has to do is find a way to make a thin margin of profit beyond what you're paying them, and the incentive shifts in favor of keeping inmates in jail indefinitely. Sell inmates' private data? Sell prison labor (a.k.a. slavery)? Sell phone calls? Gouge prices on commissary? You know... the things they're already doing. How about we release members of rival gangs in pairs? If one kills the other, that's one reoffense, but the other can't reoffend any more, so that might just be profitable. What if inmates with a high reoffense risk just happened to have medically necessary amputations? It's harder to commit sex crimes if you don't have genitals, and harder to commit violent crimes if you don't have a leg. What if we just release inmates in other jurisdictions so they reoffend away from where reoffense stats are being collected? Remember: if you aren't willing to do it, your competitor will! Lots of different forms of corporate malfeasance have been illegalized in the century and a half since the industrial revolution, and yet, corporations are still harming people for profit because they keep "innovating" new ways to make a profit with no consideration for whether it harms people. It's absurd to keep playing whack-a-mole with the latest corporate "innovations" because you can't imagine any motivation to do anything besides money. And of course we want to avoid government overreach here, right? So we shouldn't regulate against things that we don't know are happening or put too many burdensome auditing requirements to actually find out if bad things are happening. We've got to give corporations a few decades of profits before some good people sacrifice their careers to smuggle out compelling evidence that bad things are happening... only then can we start making laws to protect people--the free market comes before protecting people.

ajmurmann 3y ago
In everything you are writing there seems to be a implicit assumption that there are incentives in place already that the monetary reward system would mess with. You made a lot of points I'd wavy to reply to but all of it is secondary to understanding that part better. I'm of the believe that if you don't set a incentive explicitly one will emerge that you didn't set. Ultimately people work in any institution to make a living, so they are also likely to try to get a promotion. What will get them that promotion if we don't set explicit goals and inventives around them? I'd argue that we are seeing that that's often times being personally liked by your coworkers and boss. That's a huge issue we are seeing in law enforcement right now where the main goal seems to be sticking together over actually serving the community. Now I know less about prison guard culture, but I'd bet similar dynamics exist.