What's your take on the classic narrative that crime occurs because people are hungry and need the money to provide for their families? In your cases a, b and c it doesn't feel like any of those people do it out of need and desperation, which is the argument you often hear, but more out of choice.
> We domesticated ourselves as a species by deleting [killing] bad men each generation, every generation, for 250,000 years.
The Tyranny of the Cousins only works in Dunbar's group numbers of people, it doesn't work for modern states or cities.
Also, pretty much any weirdo, smartmouth, iconoclast, person who bucks conventional norms, introverted, less-socially-skilled, socially dominant, too-visibly-better than the existing Cousins, and who knows what else, are at life-and-death risk in such a Tyranny. And I think at least one, and usually several, of those descriptors probably applies to at least 80% of the people here.
Also, even if you assume we'd have the scientific method, if you think of most of the people who drove a lot of the technological and economic growth that's given us our modern world and lifted billions out of poverty, a lot of them would have been killed in a Tyranny for those "offenses." Newton. Tesla. Semmelweis. Ford. Zuck. Brin and Page. The list goes on - pretty much any captain of industry or major nerd, would have likely been killed for social reasons.
At absolute minimum before agreeing to such a policy I would want to have an awful lot of crimes removed from the books, or defined far more clearly and narrowly, to make sure it doesn't become possible to judicially murder someone by e.g. tricking them into downloading three copyrighted songs.
Could make a constitutional case that the right to a speedy trial is being systematically violated, and the only appropriate remedy would be for every region with a backlog of cases to hire more judges (or whoever else is needed) until that's no longer the case. If cost is a concern, raise taxes on land value - since, after all, such rights would be worthless without the law defining and enforcing them. Just the actual land, though, not structures.
For pretty minor crimes, that's not a bad description of an English magistrates' court. They are limited to never passing a sentence above six months. Offences are divided into "Crown court only" (indictment), "Magistrates' Court only" (summary) and "Prosecutor's choice" (either way). The magistrates can refer an either-way case to Crown Court if they think it should be a sentence of more than six months (any crime with a max over two years is always indictment).
ETA: UK courts are divided up as follows:
For criminal:
Magistrates' Courts: three lay magistrates with a legal advisor ("Recorder"); the magistrates are trained but not lawyers; they're sort of a permanent jury. Sometimes a district judge sitting alone (wearing blue robes, so a "blue judge")
Crown Court: circuit judge (wearing purple robes, so a "purple judge") and twelve jurors. This is where most serious or violent crimes are heard.
Central Criminal Court (aka "Old Bailey"): hears all the most major criminal cases (the closer to London the crime was committed, the more likely it is to be moved from a normal Crown Court to CCC), High Court judge (wearing red robes, so a "red judge") and twelve jurors.
Red judges can hear major cases in any Crown Court, but in practice, they mostly work in London and all major cases in London are CCC; non-London Crown Courts will hear some murder cases and those do often involve a red judge trekking up for the case.
Appeals from the magistrates' courts are heard by the Crown Court; from the Crown Court or the CCC they go to the Court of Appeal (Criminal Division) where they are heard by a panel of judges ("Lords Justices of Appeal", wearing black robes, not known as "black judges", though). Appeals from CA go to the Supreme Court (aka UKSC).
Civil cases have a similar system, with the small claims court taking the place of the magistrates, the County Courts the place of the Crown Courts and the High Court the place of CCC. Nominally, the small-claims court is just an simplified procedure in the County Court, and they're all heard by purple judges, though civil judges actually wear simple black robes these days, the same as barristers do, not the fancy things that criminal judges wear (note: a judge may be criminal one day and civil the next; the ranks of judges are district, then circuit, then high court, then appeal, then supreme court. In practice, most judges are somewhat specialist)
That would last approximately however long it took for the first black guy to be publicly flogged by a white guy, then we have major riots. Maybe it works in Singapore but I don't think it could work in the US.
This is where technology should come in! Surely it's trivial to build an automatic flogger in this day and age, and you can even find the ghettoest-looking guy to design it for good measure.
I kind of wonder; given what we've read in Scott's post much of the outcry might be coming from leftists rather than people in poor neighborhoods, but of course they are great at stirring up outrage. I suppose you could try to have racially matched floggers, but that gets complicated fast. ("We, uh, need a Cablinasian flogger for this next guy...")
I find this conclusion puzzling up against the data showing increased policing is many more times effective at reducing crime than prison, which is after all a kind of punishment, including that it prevents the victimization in the first place.
It is possible to go to prison for simple possession in some states, but it's not common and certainly the vast majority of drug offenders in prison, especially the ones with very long sentences, were dealing. And at the federal level, pretty close to 100% of prosecutions are for distribution, even though possession is criminal (and in some case, a felony).
The current focus on thoughtcrime imprisonment, particularly in the UK, where violent habitual offenders are being released to make room for those who posted mean tweets, will have some very interesting effects.
In "A Clockwork Orange," the Home Secretary explains that they need to let violent criminals like Alex out to make room for the new political prisoners.
Wasn't there the implication that the party that tried the Ludovico Technique ('rehabilitation') was Labor, and the one that let him out was the Tories (the politician's kind of conservatively dressed)?
If so it would be a funny comment on who's into censorship over 60 years.
That will be a no, then. It's also straight nonsense that people are being imprisoned for "mean tweets". Musk has been ploughing this unfortunate furrow.
Okay, so if I find you a story of a person who was arrested for "mean tweets," would you admit you are wrong? I expect what you will do instead is say either that it's very few people, or that if their tweets were illegal (labeled "hate speech", "incitement" or whatever), then them being imprisoned is GOOD actually.
I'm aware that a disturbing number of people are being questioned by police, arrested, or even prosecuted over tweets and posts in the UK. But none of the cases I am aware of led to actual *imprisonment*, as far as I know.
And umpteen bajillion stories of people getting police visits for the radically Orwellian “non crime hate incidents,” but I’ll leave that googling to you.
I’m pretty sure it was the UK, where I read about an old Jewish dude who likes to haunt anti-Israel protests. He hadn’t done anything much, so as to say, but he made the mistake of looking thoroughly Jewish, apparently.
So “mean tweets” has a physical manifestation. Seen deets?
People who end up in prison tend to have very high time preference. That seems to me like the probable reason that adding more time to an already long sentence will have a very weak deterrent effect. Think about it this way: does the kind of person who would get convicted of two felonies and then go back out and consider committing a third felony seem like someone whose decision-making is driven by what happens to him five years from now?
Unless you're talking about white collar crime sprees, those would be prevented in most cases even by a much lower level of deterrence. Those people have something to lose (namely, a decent career) so even a modest risk of an arrest record will outweigh any economic benefit of property crime.
It takes a long time, lots of support, and most importantly of all, willingness. Some people *do* want to do better, but don't know how/need a ton of help. I have sympathy for them, they often came out of circumstances that were chaotic, they slipped through cracks, and if they get help they will be grateful for it.
And some people, with all the help from all the bleeding hearts in the world, will never change because they have nothing but contempt for ordinary people with ordinary lives and the mores of ordinary society. I have no sympathy for them and no "oh no if they have to spend ten years in jail it will interfere with their enjoyment of life" cares.
Historically, the approach was rather the opposite, with frequent and aggressive use of the death penalty for offenses that we wouldn't today think merit it - that's actually the root meaning of "felony" - a crime subject to the death penalty. And the spectacle and ritual around public executions and torturings had social control functions as well.
I think at least part of the reason for this is that certain types of crime represent an economic niche that *somebody* is going to fill, and it's typically going to be people with the least to lose.
That is true of supplying various vices, but not of normal crimes with victims. There's not an economic niche needing to be filled for people stealing cars, burgling houses, or mugging little old ladies.
If true, this suggests corporal punishment would have a much more significant deterrence effect, since it's delivered all at once in the immediate future. (It likely won't have the incapacitation effect, because it doesn't take that long to heal from lashes.)
FWIW, pickpockets used to operate in the crowd that gathered to watch convicted pickpockets have their hand chopped off. (I think that was Dickens, but it's probably true.)
Sure, because the spectacle of public punishment would have been one of the largest and most crowded reliably-scheduled gatherings of multiple classes on offer at the time - too good a target to pass up.
You're misremembering your history. Pickpockets were sometimes publicly hanged, as stealing enough handkerchiefs could be a capital offence (though many would have had the sentence commuted), but even 100 years earlier than Dicken's time, amputating hands had long since fallen out of use as a punishment.
OTOH, hanging doesn't detract from the main point. Awareness of severe penalties does not, in itself, act as a strong deterrent. You need to combine that with an expectation of being caught , and if you do that you don't need the *severe* penalty. It just has to be more than "the cost of doing business".
I guess the missing control is, how much pickpocketing happens in those crowds compared to other crowds? If it's the same, I guess deterrence do not work but there's a chance that it's significantly lower.
Yeah, I think at least for lunkhead street crime, if the dude committing the crime was thinking a lot about where his current choices would leave him in ten years, he would be on an entirely different life path. That kind of crime selects for people who are giving *very little* weight to possible bad outcomes a decade or more in the future.
Seems like the only real solution is for society to develop a parallel informal justice system based on the intuition that any method of clear self-defense of property and/or body is justified and allowable. Shoplifting rates would drop precipitously if there was a very good chance of being whacked with a baseball bat or even shot by a shopkeeper. Vehicle theft would probably stop if there was an excellent chance of painful booby-trapping. Etc.
Well, certainly the shopkeep has to be careful about not whacking a non-thief with his bat, lest s/he be punished for engaging in *unjustified* violence.
But that's also why I used the words like "clear self defense."
Remember when I said the words like "informal" and "intuition?"
Even monkeys know when something has been stolen from them. People know what's right and wrong. When it's clear and obvious that one person is attempting to wrong another, a good civilization would shrug at whatever immediate consequences happen to the wrong-doer as a result of their wrong-doing.
Risk of prosecution is what is suppressing such shootings and baseball battings right now. Presumably you want to expand the range of circumstances under which it is justifiable. But do you want to expand it so much that loss of *property* can be justifiably met with *fatal* forc e?
Sure. While a *deep* value of property is not intuitive for the economic upper classes, that is not how people much closer to basic survival think of their items. Having a car stolen is at worst an inconvenience for a wealthy person, but for a working poor person, having their car stolen often means losing income, perhaps even their job, and possibly even their home/safety net. If a poor person wants to booby trap their car, I sympathize with the urge.
The standard advice in east Africa is: if someone snatches your bag think before shouting "stop thief" because if you shout that the crowd will kill the thief. I don't know if this is true but I believe it is, and the belief doesn't give me a particularly warm and comfortable feeling. Also, theft is very very common in east Africa.
I don't know. Possibly the mob turns on you if it murders the "thief" and finds no stolen property on the corpse. Being given the warning has had the beneficial effect on me of being extra careful of my possessions so as not to have to initiate the process.
I can't remember the source of the advice but a Google finds this in a guide book (other side of Africa)
"The only thing worse than being robbed is seeing what happens if some poor unfortunate is caught: mob justice looms large on the streets of Freetown, where suspected thieves can be beaten up, belted or worse by angry bands at a moment’s notice."
If you love criminals so much go live in San Francisco. Otherwise keep your soft hearted nonsense to yourself. Normal people don't want to live in the violence and chaos that people like you create. Maybe when you see people you care about being victimized as a consequence of attitudes like yours you'll get some sense.
I actually do. The problem is community punishment is really prone to false-positives (you bump into someone, they think you're stealing from them, and therefore are allowed to beat you up) and differential enforcement.
One of the advantages of having cops is it's their job to do this stuff, and when they screw up we can go after them. Diffusing it onto the general public leads to all kinds of problems that are the reasons we outsource it to cops in the first place.
I agree. It's forgivable when young children get over-excited and carry something out of a store, but the other two examples aren't okay.
I have never, ever, not once in almost 45 years of life, been accused of shoplifting, because I am ostentatiously careful about not giving the impression I could be shoplifting. I don't think I've even even put an item in a pocket when my hands were fuller than I was expecting.
But that's because I care a lot about presenting myself as an honest and trustworthy customer. Someone saying, "but I happen to care about other stuff more, so I shouldn’t be punished for that!" isn't a persuasive argument for me.
The examples are ok, this can happen, and it doesn't really matter.
It matters a lot less than some shoplifters, or some distracted people, or some people seeming like they did shoplift when not, dying because they took a wrong hit.
I accidentally shoplifted not long ago. I bought four yogurts and only ran it through three times. I discovered my error by accident when looking at the receipt to check my fuel points.
Caba lists two cases of accidental shoplifting, one as a kid. Leaving only one as an adult.
N=1 isn't much evidence for anything. Assuming they had 10 years of adult life, even if you use p < 0.05 as a criterion, you can't rule out their actual rate of such mishaps being once in 190 years.
Consequences for violent crime and shoplifting are pretty severe in Singapore, are they not? And personal /property crime is comparatively low, is it not?
Does it really make sense to call that torture? Punishment of any kind has to be unpleasant to be effective, no? I don't know where the line between torture and punishment should be drawn, but Singaporean caning strikes me as one that falls on the P side of it.
Yep, it's amusing how the current western Overton window is squarely at "everything except prison is cruel and unusual". Prison is plenty cruel in most places, and the unusual part is self-fulfilling.
You're one of the few well-functioning dictatorships in the world, and very small in size...and if I can be a little un-PC don't have populations that cause huge problems.
One day drones will do this non-lethally and prison won’t make as much sense for most things. You’ll just shout a drone word and a bunch of them will flock in and take care of the problem.
Laser drones with a 99.99% correct conviction rate will instantly non-damagingly laser the culprit for three city blocks unless they reverse plea bargain into being arrested and tried
Should this approach be reserved for poor people's crimes, or would you condone extending it to upper class crimes as well, such as wage theft, tax evasion and bribery?
I'll go you one further: I'd love to live in the kind of society where someone enacts vengeance on any of the key Purdue Pharma Sacklers, and, after a trial establishing the facts, the jury takes 30 seconds to deliberate before issuing a dismissive shrug as their verdict.
Our justice system has way, way over-corrected in favor of predators like the Sacklers. It'd be great to live in a world where no DA would even bother to bring the murderer of a Sackler Board member to trial because there would be no hope of finding a jury willing to convict them. The Sackler Board members probably wouldn't have been quite so cavalier about pushing an addictive product if they believed there was a good chance of being injured or murdered by grieving family members.
I'd like to rely on the collective intuition of a far-more justice-based society to make those decisions! It'd be great if a sales person way down the hierarchy was just a *little* worried that a grieving parent of a dead overdosed 20 year old might decide to come after *them* for selling to that kid's doctor. But I'd also like that parent to be worried it might not fly with a jury.
I seriously doubt this be the result. Most people wouldn't suddenly start carrying weapons to wave for starts, surely there are only a few who want to, but can't right now. Plus, crime happens mostly because it *isn't* caught in time. The outrageous occasions when someone steals something on broad daylight, simply isn't much of it. It's not like chains and department stores are going to adopt this either - liability for accidents is likely way scarier then some inventory lose. And for detterence? Scott already covered that!
Plus people would make the calculation that it would be OK to rob old or frail people or children or women, because they could count on winning the ensuing fight. Plus a lot of people would get killed defending themselves from the crooks where they might have previously lost a handbag or backpack.
Spoken like a dude who isn't a woman who thwarted TWO separate robbery attempts by strangers merely by indicating she had a firearm and was prepared to use it.
Nice job! (Seriously--I was scared of criminals before I was scared of feminists.) Happy you got out safe...I would have been tempted to fire.
You then get into the whole gun ownership question, which I am always flipping back and forth on. I get the feeling the rationalists have already looked into this but I also get the feeling they are liberal wimps who just don't like the idea (look at the recent Prop 36 thread) . I'm sure someone here can fill me in further with actual data...
Nonsense. This is wildly incorrect. Responsible American law-abiding gun owners are very aware that even in the best-documented, clearest-cut cases of self-defense, they are likely to be hassled tremendously by law enforcement and/or the perpetrator's family. Most are far more afraid of losing their house for righteously shooting a criminal than they are of being in a deadly threat encounter itself. That's why self-defense insurance exists.
Remember, you're a woman and unlikely to be targeted by people looking for thieves. As seen below false positives are a problem. Theft now becomes an excuse for violence as well as self-defense. You can see racial bias erupting (on both sides, though only one will be covered by the media) with false accusations of theft leading to violence. Not to mention booby-traps going off at the wrong time--you're going to be walking by cars and have them blow up because someone set their booby trap too sensitive, or get killed when the swinging anti-theft blade the shopkeeper sets up triggers when you step on the wrong square.
I'm sympathetic to your concerns with high crime rates in liberal cities (I lived through the transition in the 90s and heard about the reverse one in the 2010s), but IMHO we just need to let the cops do their jobs again.
EDIT: I'm disavowing this statement and admitting that I was wrong. It was sexist by progressive standards, argument ad hominem by rationalist standards, and uncharitable by conservative standards. I'm stepping away from ACX for a while. Good luck to all.
My actual views on why they can't are too un-PC to post here and I will respect Scott's desire to avoid trouble with the progressive left so he can post on everything else he finds interesting and, now, not have his children harassed by leftists.
RE accidental shop-lifting: It happened to me once, with a larger (supermarket) purchase, like 100€. (Background: I live in Germany.) My partner and I both thought the other one had paid. We noticed at home, called the super market, and returned another time to pay. We got a bouquet of flowers (I'm not kidding!) and a bottle of something for our honesty. We were told this happens all the time (dunno how pricey the stolen purchases are on average), but this was the first time the long-time employee we contacted has seem people actively come back to pay another time.
This probably won't convince you, but I'm very glad I don't live in a world where a forgetful grandma at the self-checkout in front of me might kiss the floor after getting a dent in her head. It could be your grandparent, for Christ's sake! We seem to inhabit very different mental spaces
I'm not sure if I believe you really don't understand my "mental space." Human beings can tell the difference between thieves' furtive theft or smash-and-dash and a legitimate customer obliviously walking away with an extra yogurt. It should be fine for a motivated shopkeeper to threaten the former group with a baseball bat and politely remind the latter group to come back to pay. They're different people.
That's where we started! Our current system is supposed to be an improvement on it. The norm for most of history was vengeance killing, public stoning and other forms of lynching. Governments hardly ever had the power to deter crime outside of cities. If you want a sense for why the West transitioned to laws & courts, The Oresteia by Aeschylus is a good dramatization.
Yes, and unfortunately the current system has drifted so far away from actual "justice" that it's no longer an "improvement." It's just time to drift back a bit.
Yeah, I kind of agree. Except they have these cycles--we had a similar situation in the 60s and 70s and they all decided to get 'tough on crime' in the 90s and crime fell long enough for liberals to forget why those policies existed in the first place.
My impression is that longer and surer imprisonment can incapacitate criminals and even deter crime, but it can take quite a while for the criminal class to get the message.
For example, America started getting softer on crime maybe in the early 1960s. Then crime started to get notably worse in the mid to later 1960s, and American finally started to get tougher on crime maybe at the end of the 1970s. But then crime remained mostly high, although we did have a nice period of lower crime around 1984 in between the powder cocaine years and the crack cocaine. (I'm using homicide rates, which are fairly reliable, and calling them the "crime" rate.)
But the crack years (roughly 1988-1994) were horrible for the homicide rate. What seemed to be happening was that, as older criminals were increasingly locked up, violent criminals kept getting younger and younger, and they were peculiarly homicidal.
The crack years overwhelmed the prison system. Back during the crack years, a friend of mine who is in the comedy business was talking to a colleague's brother, who was a prison guard in upstate New York. The guard said, "Is there anybody you'd really like to murder? Because if there is, now's the time to do it, because we're so overcrowded they'll let you out after only about eight years." (A prison guard with famous comedian genes can have pretty memorable insights.)
But America built a ton of prisons in the 1990s. And suddenly, in the second half of the 1990s, the incapacitation strategy recommended by James Q. Wilson in the mid-1970s in "Thinking About Crime" started to work. And then it worked surprisingly well. In particular, tough-on-crime New York City became remarkably less lethal.
It worked so well that the Establishment began to forget the hard lessons it had once learned and after a couple of decades started the soft-on-crime cycle all over again, causing homicides to soar due to the Ferguson Effect and especially the Floyd Effect.
Now we may be headed back into a tough-on-crime cycle due to the predictable failures of soft-on-crime policies.
Are you aware of the standard arguments against increased incarceration being responsible for much of the secular crime drop, and do you have responses to them anywhere?
We had decreased incarceration for about a decade and a half in the liberal 1960s and 1970s, during which crime eventually soared. Then we had increased incarceration for about about a decade and a half during which crime remained high and even peaked toward the end of that period in the early 1990s.
A whole lot of college professors in the social sciences got hired during those 30+ years in which everybody who was anybody either was for decreased incarceration or at least admitted increased incarceration wasn't visibly working.
Then, all of a sudden, in the later 1990s, increased incarceration suddenly seemed to start working and kept working for a couple of decades. In fact, a return to decreased incarceration seems to have led to an explosion in the homicide rate in 2015-2021.
Would you suddenly admit that your career was based on a fallacy? Would you turn against your academic mentor who boosted your career?
Scott's one of the very few people who could get away with that, I think, assuming he had enough data to convince his rationalist fanbase it was actually based on dispassionate investigation, etc. He already survived one direct attack from the NYT.
Your lefty academics are definitely a different example. Watching crime fall through the 90s in my hometown was one of the things that (a) permanently made me a a tough-on-crime guy and (b) convinced me the social sciences were FOS. I've softened on the social sciences in general since then, but I think the leftist monoculture in academia is one of its biggest weaknesses.
What is responsible? I know Levitt suggested abortion contributed, but then critics said his results came from a coding error. Kevin Drum has been saying it was lead, but now we have papers indicating that was publication bias. A lot of such papers were written during the crime decline, but we witnessed crime increases after Ferguson & Floyd that can't be explained by lead being re-introduced or abortion going away.
Even if the increase and drop from the 60's to the 90's was largely driven by a couple of causes, it doesn't mean those are the only causes for changes in the crime rate. So increases in crime around 2020 don't necessarily mean anything about earlier changes.
We have had during my lifetime three periods of liberalization of the criminal justice system and the dominance of anti-law and order thinking: the long Warren Court era and the short Ferguson Effect and Floyd Effect eras. All three saw the homicide rate rise sharply.
But maybe if we tried it a fourth time, it would stop happening.
Levitt has addressed the coding error, and fixed it and A) it did not dramatically change the results in his first paper and B) in a follow up paper 20 years after the original, the predictions made in the original were borne out. I have not yet come across a convincing critique of his work
Here is a link to a podcast episode where Levitt discusses all this. It also includes a transcript if you don't like listening to things like this:
The other major change (which is also discussed in that podcast) is lead (I haven't heard the publication bias argument before nor heard a response, but I'm skeptical). Crime is complicated and no one is claiming those are the _only_ two causes (another one, also discussed in that episode, is increases in number of police, which convienently agrees with Scott's post).
But I am relatively convinced that the combination of decreases in lead, decreases in unwanted pregnancies, and increasing police preseance are by far the most effective _and_ most cost effective deterrents to crime.
Great point. A lot of these "but ackually [common sense thing] can't be right because [points to study]" crash into the reality of (a) the existence of places like Singapore and El Salvador, and (b) you-can-see-the-timing-perfectly-on-a-graph homicide surge post-George Floyd / "defund the police" (and Ferguson effects in general). Obviously criminals adjust their behavior to the perceived probability of being caught. They're stupid but they're not *that* stupid.
How do you know that it was the George Floyd stuff, rather than stuff related to the pandemic? The same could be said about Ferguson effects: was it rather something related to the increase in poverty or inequality?
In public policy, there are always so many things changing that observations of correlations usually don't tell you anything. That's not to say that you're wrong that Ferguson effects are causative, but you've not claimed there's any evidence that it's causative (rather than correlation).
This was an incredible summary and very helpful - thank you
This also matches my experience working with some low-income boys as a volunteer. It took me too long to realize how terrible they were at time-discounting and weighing risk. Where I was saying: "this will only hurt a LITTLE but that might RUIN your life," they heard: "this WILL hurt a little but that MIGHT ruin your life." And "will" beats "might" every time. One frustrating kid I dealt with drove without a license (after losing it) several times and drove a little drunk occasionally, despite my warnings that he would get himself in a lot of trouble. He wasn't caught and proudly told me that I was wrong: nothing bad happened, whereas something bad definitely would have happened if he didn't get home after X party. Surprise surprise: two years later he's in jail after drunk driving and having multiple violations of driving without a license.
People, especially young men with little education, are pathologically bad at evaluating risk. If the criminals are risk analysts committing white collar fraud, then sure you can increase the costs of committing a crime and let them factor that into their risk calculations. But for normal criminals: ordinary low-education, risk-blind high-testosterone schmucks, the best thing you can do is increase the probability of getting caught and feeling immediate repercussions, regardless of how small they are.
I moved for a while to a central European country and was impressed with the criminal justice system and the low level of crime/high level of compliance, despite the high amount of migration. The biggest thing I noticed is the certainty of being caught that you mentioned. Sentences were (to my American eyes) relatively short and weak - and more liberal Americans often thought there was less crime because the country was somehow more humane. But the main thing I noticed was: the certainty of being caught was high. Police were everywhere and very responsive. Small fines and one-night-in-jail punishments were everywhere for petty crimes. I think the U.S. would benefit from strong policing with simple, consistent punishments for many small crimes, especially for first time offenders.
One of my favorite lessons I have learned from working in a Corrections-adjacent position the last couple of years is that generally, people who get incarcerated aren’t there because of their great decision-making skills.
That's why, not to be unkind to Scott, I'm smiling wryly at his "imagine your life after ten years in jail".
If the majority of offenders were married, with kids they had within marriage by their spouse, and working jobs, with supportive families... we wouldn't be having a discussion over "ten years in jail, too much?" because they wouldn't be putting themselves into the position of doing crimes worth ten years of jail in the first place.
White collar crime is somewhat different, because you will get idiots who think embezzling half a million from their employer is the perfect crime and they'll never get caught because they're smart, after all they're working an office job and they're not some underclass type.
"Your partner has long since filed for divorce and is happily remarried to someone else. Your kids have long since moved on; if they remember your name at all, it’s as “that guy who was never there for us”
However, all those children are not necessarily by their spouses, if they have spouses in the first place, and if the people I have encountered with kids under 18 were all married to the fathers/mothers of said kids, then polygamy would indeed be legal.
Though I will concede that they are married, according to the link:
"In state prison, the likelihood of being a parent was most common among married inmates (71%), compared to inmates who were separated (64%), divorced (55%), never married (45%), or widowed (36%). Among federal inmates, married inmates were more likely to report being a parent than inmates in all marital statuses except those inmates who were separated from their spouse. There was no difference in the prevalence of being a parent between married federal inmates and separated federal inmates. The likelihood of being a parent varied little by education for both state and federal inmates."
Even if they had kids, the vast majority of them wouldn't have been there for their kids in the first place. Probably not for their romantic partners either, at least not in any good ways.
Interestingly, decision-making skills are literally one of the things they try to teach as part of treatment in Minnesota prisons. Cognitive interventions, that sort of thing.
My sense is that it works great for people who mostly don’t need it.
Putting this in as nice and polite a way as I can: I suspect I am on the low-end for IQ (perfectly high midwit-ish 117) of the average ACX reader, but interestingly wife (135+ish, based off ACT score, STEM PhD from Ivy League uni) is closer to modal for this site. So I get to have the fun experience of interacting with someone a full StDev smarter than me, and one thing I often experience is that smart people tend to underestimate how much harder it is for someone with that difference in raw intellectual capacity to do things they can ("No, love of my life, I am not interested in making a Python model to figure out which healthcare option is the best for us."). This was especially heightened when we were foster parents for awhile, and one of our placement was...whatever the appropriate term is nowadays for someone with a 63 IQ (which we were probably the first people in his family to find out about, since I was told I was the first person to ever show up for one of his school's IEP meetings); living with top and bottom 1% intellectual functioning was a very illustrative experience!
IMHO, a major problem with discussion about crafting programs for effective deterrence is that (especially for violent crimes-Jason Manning's review of Fragging notes how Wolfgang noted the modal murder starts with a very trivial dispute) the people who mostly need to be deterred are frankly too stupid for that to work (executive planning and IQ are decently correlated and breaks down rapidly below a certain threshold, which is higher than the cutoff for legal disability) and the people crafting these program won't get that the processes aren't likely to work with the population they are trying to fix because that population is too different in mental functioning to pay off. Our friend Joe from the Open Thread my intellectually get that he shouldn't murder someone for repeatedly taking fries off his plate (actual case Manning notes) and that he will be punished if he does so, but it will be very hard for that intellectual knowledge to kick in before he has killed the fry taker.
So incapacitation may have more value (especially, as mentioned in the comments, the age factor of incapacitation: the value of keeping people in jail in their criminally-prime years).
The IQ and time preference thing is a big thing; everyone seems to be designing systems based on their own experience, which often breaks down terribly for the left half of the curve. I remember Derbyshire commenting on how nobody ever had any sympathy for the unbookish, or realized how painful learning new things was for people who weren't into school.
As an aside, how did you land her? My impression was those sorts of women were obsessed with finding a guy who was at least as smart as they were.
"As an aside, how did you land her? My impression was those sorts of women were obsessed with finding a guy who was at least as smart as they were."
Careful application of Andrew Tate's lessons, duh.
My serious answer to this would be we both are religious (which really narrowed down her dating pool especially), both of us reaching an age/stage of life where time was a driving factor, me being better than average at knowing and accepting limitations and not caring too much about her being better than me in certain areas, and her note that "smart people are very annoying".
I do think I am an interesting case study in hypergamy, in that at every level of our relationship we were in a situation where she was the more highly educated (and frankly because of that, the higher earner, as well).
Also, I think the way crimes are punished in the US is almost optimized to be confusing and incomprehensible for even normal people, let alone dumb people who stopped going to school when they were 16.
Oh yeah, your drunk-driving kid is a great example. Some people will always be surprised that "if I do X, Y happens? how can that be?" And unhappily some of *those* are "Y happened? That's so unfair! They're picking on me! The fact I did X has nothing to do with it!"
The first set may eventually learn from hard experience and immediate repercussions. The second set will never learn, nor want to learn, because it will always be Somebody Else's Fault.
That fits with what I've always heard. Severe punishment isn't particularly effective. What matters is quick and certain, with quick being more important than certain. This fits with various operant conditioning studies that I ran across.
Here in the US, my husband had the experience of someone who'd committed a crime against him telling him outright, "What are you going to do, call the cops? They're not going to care."
The whole situation would have worked out very differently if there were a high likelihood of SOME consequences, rather than a low likelihood of harsh consequences.
These comparisons are somewhat strained, given that in some European countries, "cops" include people who never leave their office and perform some sort of administrative duty that in other countries isn't performed by police at all.
For example, issuing various permits.
Czechia is notorious for having a large amount of cops nominally, but relatively few beat cops. Most of the difference are bureaucrats formally employed by police.
Interesting. I wonder how much has to do with risk aversion? I'm sort of the pathological opposite--too risk averse. The one time I decided to take more risks in life I sped right into a speeding camera.
Though I can't help but recall a recent Matt Levine column on rating investment strategies. There's an ongoing debate between people who made 20% per year on large cap tech stocks and people who have made ~5% per year with a more diversified strategy who claim their risk adjusted returns are higher.
It is hard to convince the people who made more money that they actually made less.
Similar debates show up throughout various sectors of society.
I don't pay much attention to the FBI's statistics on murder offending (not that many statistics nerds join the FBI), but I do pay a lot of attention to the CDC's cause of death statistics where the cause is homicide.
My impression is that the tough on crime era had a big impact on driving down the homicide victimization rate (and thus presumably also the murder offending rate) among Hispanics.
It appears to me that Latinos both here in America (where Hispanics now number 65 million) and in Latin America (e.g., El Salvador, where murder rates soared then plummeted) are fairly sensitive to the messages about crime sent by the government and the culture.
When the politicians and thought leaders turned anti-law & order after George Floyd's death in late May 2020, blacks immediately started dying more by homicide and traffic fatality as the cops retreated to the donut shop. But Hispanics took longer to loosen up and adjust to less police assertiveness.
Due to their enormous increase in numbers, Latinos are now America's swing voters when it comes to America's crime rate, so criminologists should focus on what interventions persuade Hispanics that the cops are serious about crime.
Similarly, the Hispanic rate of death by motor vehicle accident dropped sharply after 2008.
All the evidence I've seen suggests that both homicide fatalities and traffic fatalities correlate in recent years. They're both related to how scared of the cops people are, so they correlate with whether people speed and whether people carry illegal handguns in their cars.
Sometimes written d'Hondt which I think looks better. There's a d'Hondt method, independently invented by Jefferson, for divvying up parliamentary seats according to vote share.
Jefferson invented it for divvying up congressional seats to states. These are mathematically equivalent problems, but the two separate purposes for which they were being put in the English-speaking USA and the French-speaking France and Belgium meant that it took many decades for the equivalency to be realised.
>This is potentially a good way to study deterrence, because one group of criminals (those with 0-1 strike) will get a normal sentence for their next crime, but another (those with 2 strikes) will get a long sentence.
This is not correct. As the paper says, "a criminal with one strike who is convicted of any subsequent felony (not nec essarily a strike) faces an automatic doubling of the sentence length on that conviction"
Maybe the wrong way to look at it, but it sorta matches both my intuition and some of the points I've seen made on subjects such as capital punishment as well - at some point that's a little lower than people tend to naturally think, we get decently diminishing returns on increasing the *severity* of punishment. My understanding is that the *certainty* of punishment matters meaningfully more after we hit a certain brightline
I don't think this universally holds, but I do think it holds at least a little bit, particularly with the almost-nonexistently-weak deterrence effects. Let's say that the max sentence for a crime is 5 years and not life - anyone thinking deeply about the cost-benefit of prison probably is deterred enough by the smallest prison term necessary to ruin their life. Anyone *not* deterred by that is either
- expecting on some level that they won't get caught, so cranking up punishment severity isn't gonna matter too much when they don't expect to serve it
- they don't care about consequences, in which case the deterrence effects are just not gonna be there no matter what you do
This would seem to support the idea of the extra cops doing more than the actual prisons - really any consequence/engagement with the legal system being certain is better for preventing crime than significant uncertainty of big consequences. But just guesses on that mechanism of why prisons are weakly criminogenic/mostly neutral despite the big positive incapacitation effect
I think there is an error in the "The Swedish data suggested a “three strikes law” should cut crime by 83%" section.
The preceding text says :
"A one-strike law (life in prison after one violent crime conviction) would cut almost 80% of violent crime in Sweden; a two-strikes law would cut 63%, and a three-strikes law would cut 57%."
Only the violent on a first strike basis should be incarcerated ... the rest should be compensating their victims, even if it takes a lifetime of Sundays; and if there are no victims then there is no crime. This way the current prison space can be utilised far more effectively, to keep the violent out of society for longer periods ...
Many criminals are judgement proof, lacking assets to compensate people, and don't really have much in the way of skills to do anything to compensate people.
that there are so many people in such a position strikes me as a far bigger crime than all the property crimes (that people get thrown in jail for, not the white collar stuff like foreclosure fraud that the government shrugs at) put together
> that there are so many people in such a position strikes me as a far bigger crime
Are you using the word "crime" in a literal sense or you just saying it's very bad that this is the case? I also believe the world would be better if everyone had assets and valuable skills. It'd be hard for me to call it a crime without pointing to who the criminal is. Is there a person/class of people you think should be prosecuted for this?
Thank you for putting the time and effort into this.
"How much do we care about drug crime? ... Roodman says not much. It’s either personal use, which is mostly victimless, at least compared to other crimes). Or it’s dealing, which is demand-driven and almost impossible to stop through incarceration - if you lock up every member of Cartel A, that just creates a lucrative niche for Cartel B to step in and fill."
I'm afraid I'm one of the Neanderthals who go "I don't care if it has a deterrent effect or if it's not cost-effective, put these sons-of-guns in prison and get them off the streets" If Cartel B steps in, then just go after Cartel B as well.
Mandatory declaration: I'm aware this story is more anecdotal and it's based on "one guy says", but if it's anyway true, then hell yeah throw the book at Cartels A to Z!
"Young cannabis users are being spiked with heroin and pushed to take cocaine by dealers intent on hooking them to expand their market.
Children as young as 13 are being offered free cocaine samples by dealers when they buy cannabis, according to addiction therapist Michael Guerin.
As the cocaine high is short and the drug is expensive, users can quickly end up in debt and be forced to become mules, carrying large quantities of drugs for organised crime to pay back debt.
David Lane of the HSE’s Drug and Alcohol Taskforce in Cork and Kerry, has also warned of cannabis being laced with potent synthetics and said that is currently a bigger concern than other forms of spiking.
“We have seen synthetic substances adulterating other drugs," Mr Lane said.
“There has been a big change regarding nitazenes [highly potent synthetic opioids] and synthetic drugs, and that is where a lot of our concern is still. They can be more potent and are therefore more likely to cause overdoses."
Heroin use has been decreasing amongst younger age cohorts and new users in Ireland.
But many heroin users who sell small quantities of the drug to fund their own habit, have begun ‘pushing’ it on new users or spiking cannabis with heroin for unsuspecting smokers to hook them and thereby expand their client base.
“The heroin user, because they are so addicted, are under pressure to create new business to sustain their own habit,” Mr Guerin said.
“They’re spiking cannabis users with it to try to get them hooked so that they've got people to sell it to. And they’re encouraging them to try cocaine.
"The rationale behind it is to entrap them more.
“Any serious drug addict, irrespective of what they're taking, unless they have a pocket full of money, which is unusual, are involved in the sale and distribution of drugs purely for economic reasons. They have to be."
He said cannabis users as young as 13 and 14 have told him they have been given free samples of cocaine to try.
"They’re offered to sniff it off the end of a key," he said.
"They would try it and if they started to use it to any degree whatsoever they couldn't afford to pay for it, because cocaine is an expensive pastime. They then became involved in underage moving, distributing, delivering, collecting money all that stuff."
He said it is not unusual for cocaine users to say their habit is costing them between €1,000 and €1,500 a week.
Heroin dealers are also pushing the drug on cocaine users and crack cocaine users to medicate the cocaine ‘come-down’, Mr Guerin said."
Singapore is only techically an island; it's separated from Malaysia by a strait less than a mile wide in most places. Malaysia has similarly strict drug laws, but if deterrence doesn't work, that shouldn't matter.
It's way easier to walk across a border or drive a truck across a border or lead a mule across a border or dig a tunnel under a border than it is to cross a strait less than a mile wide.
I'm not arguing that Singapore doesn't successfully deter crime or that deterrence doesn't work. I'm saying Singapore is starting with a number of advantages. If Malaysia has similarly strict drug laws but Singapore has less of a drug problem than that speaks to something different about Singapore, and one thing that's different about Singapore is that it's easier to handle the supply side, not just the demand side, because it's a lot easier for Singapore to both police its island borders and to inderdict local production.
That sure argues for an approach to cannabis like the recent wave of states legalizing first medical, and then recreational sales and use of marijuana and related items. State-sanctioned pot has standards of regulation and testing that prevent turning it into a "gateway drug" by actually offering access to harder drugs. Likewise, real dispensaries follow regulations to the maximum extent.
The question is, how much of marijuana's attractiveness comes from its biochemical properties and how much is due to other reasons like it being a forbidden fruit or social pressure. It's possible that some spectrum of altered mental states available via legal substances would be satisfying for the vast majority and they would stop seeking illegal ones, but I wouldn't be too surprised if legalizing marijuana would just mostly result in something else mostly taking its niche and this continuing to happen with more and more drugs. Drug dealers must eat every day, after all.
I lean pretty strongly towards "doing drugs is victimless;" that doesn't mean selling it to kids is the same as selling to adults. But as is often the case with "sin crimes", the fact that the business is illegal is what drives a lot of the criminal activity. Compare alcohol during prohibition to now. The people who made, distributed, and sold it were mostly organized gangs who had to protect their property and "turf" with violence, as going to the police wasn't an option. That's not the case anymore. Marijuana being illegal is exactly what causes it to be sold by criminals, who use other crime to support or hide their business. This also makes it harder to regulate--a legitimate business has incentives not to sell to kids and not put cocaine in their weed; the local street corner thug does not.
The other thing you don't acknowledge in this comment is the monetary cost. It's easy to just say "crack down on X, Y, and Z" if it's mostly other people paying for it.
Speaking of rehabilitation, Minnesota is putting into practice an interesting experiment. Incarcerated persons (their term) for most offenses can have up to half their sentence reduced if they voluntarily engage in fairly intense substance abuse treatment and other programming and therapy. They then go out on heavily supervised probation.
The animating principle, as I understand it from Corrections leadership and their on-staff researcher (an acquaintance of mine) is that the one thing that has been shown to reduce recidivism (so ignoring incapacitation effects) is keeping people busy while they are in prison. It almost doesn’t matter with what, though therapy seems preferable. So the idea goes that if you make it voluntary, so IPs have buy-in, you make it therapeutic to help with the substance abuse issues behind a LOT of crime, and you make it intense enough so that IPs have to really want it, you might actually have a successful rehabilitation program of sorts, at least on the margins. Plus, you reduce costs by shortening sentences.
Sadly, they’re likely to f*** the implementation up by not being able to hire enough therapists and case managers to provide programming, enough probation agents to monitor them afterwards, and enough literal space to actually provide treatment at this level. So I’m not expecting much. But I appreciate the effort.
My view on the worst case for this is that the pilot programme works, then the state says "Great, we can cut recidivism! Let's expand this and roll it out to everywhere!" and then "Okay yeah but it can't cost more than X amount", so there aren't enough therapists hired, the after-release support isn't there, etc. due to "are you crazy, we're not gonna spend that much on jailbirds, the voters would riot" effect, then it all fails in the predictable ways and back to square one.
I like your blithe confidence that they’re doing this as a pilot program rather than rolling it out wholesale without a clear funding stream and hoping for the best. It speaks to your generous heart.
Literal words spoken to an audience by the Commissioner of the Department of Corrections: “We wanted to pass the bill [to reduce sentencing] while we had a window for it, so then we could go get the money for it later.”
Clearly I am too chicken-livered for the bold, venturesome local government of Minnesota! Full speed ahead and damn the torpedoes!
Is this the kind of initiative we could have expected from Vice President Walz (ex-governor of Minnesota), had the happy dreams of the nation come to pass? 😁
The Commissioner was his hand-picked guy, and Walz approved the budget, so… yes. Yes it is.
And I say this as a wild-eyed liberal in many respects.
This is likely going to be a disaster and it’s just sad. We’ve already had unrest from the IPs that it hasn’t been implemented yet, so the Department took the bold step of issuing the inmates… a memo telling them to wait.
Everything is for the best in this best of all possible worlds.
Isn't that what happened with deinstitutionalization of schizophrenics? They thought they could put them on all on outpatient meds, they cut the funding for the clinics, and they all wound up homeless or in jail?
> Incarcerated persons (their term) for most offenses can have up to half their sentence reduced if they voluntarily engage in fairly intense substance abuse treatment
OK, I stab a guy in an alley. Why am I supposed to engage in intense substance abuse treatment?
> Sadly, they’re likely to f*** the implementation up by not being able to hire enough therapists and case managers to provide programming, enough probation agents to monitor them afterwards, and enough literal space to actually provide treatment at this level. So I’m not expecting much. But I appreciate the effort.
Not just this, but drug and opiate rehab success rates, for people who voluntarily self-select into the programs, is ~30% for alcohol* and <20% for opiates.**
The success rates for heavily coerced "do this or get years of prison" populations is likely to be considerably lower (it's ~5x lower for alcohol in non-self-selected pops).
This argues that the (millions? Tens of millions?) being spent on these programs for even the poor implementation you expect are almost certainly 80%+ wasted.
* White 2012: "An analysis of reported outcomes in 415 Scientific Reports, 1868-2011"
Combined alcohol and drug abstinence rates average 30.3%, "although remission from opioid addiction appears to be less stable and durable than other patterns of remission."
** The best number for long term opiate relapse is Hser 2007, which has findings from 33 year followups. 18% achieve abstinence "for at least 5 years."
Hser, Y (2007) Predicting long-term stable recovery from heroin addiction: Findings from a 33 year follow up study. Journal of Addictive Diseases. 26(1), 51-60.
"But I notice that the European countries we’re talking about here all have high recent new immigrant populations which on average commit more crimes per person than natives."
Well, yeah, with the vast majority of these crimes being "having immigrated illegally"
I think you're generalizing from the US context, where migrants are less criminal than natives. In most European countries, the opposite is true - for example, https://www.researchgate.net/publication/338563093_Migrants_and_Crime_in_Sweden_in_the_Twenty-First_Century finds that 73% of Swedish murders are committed by immigrants. Immigrants are about 20% of Swedes, but most of those are Finns, and I think the 73% of murders are probably mostly committed by the 2 - 5% of the population who are Middle Eastern immigrants.
I stopped and doubled-checked what I thought I knew based on what reads like a general statement about recent immigrants, you might clarify this in the text.
I know it's true that Middle Eastern immigrants are disproportionately criminal in Europe, but does anyone know how this fact can coexist with the extremely low (peacetime) murder rates in the actual Middle East?
Perhaps it tracks with the extremely strict punishments for minor crimes that the Quran recommends? Even if most middle eastern countries aren't cutting off hands for petty theft, whatever their current policies are probably downstream of that.
European cultures are generally unwilling to go to such harsh measures for small crimes, mostly because western laws are downstream of Christian morality that emphasizes forgiveness and mercy.
I don't think anyone would dispute that cutting off the hand of anyone caught shoplifting wouldn't be incredibly effective at reducing shoplifting. All of a sudden shopkeepers have an extremely obvious sign to look out for, shoplifting literally becomes more difficult with only one hand, and generally I'd imagine such a deterrence is a lot more effective than only taking away your freedom.
Can this be because these people were torn from their native culture and locales where simple "neighborhood watch" prevented them from ever trying anything and put into great big cities where nobody knows them, in a culture where many things they thought forbidden are permitted, and without much money, work, or support network?
Also, there is a small matter of Quran allowing Muslim to (at least) "violate contracts" with non-believers. I'm not sure if this "believer vs. non-believer" really matters or not, but this point is often used in anti-immigration discussions, that for Muslims, Christians aren't "real people". Kind of like a pogrom against a Jew settlement used to be a "normal thing", but the same people would balk at visiting same violence at their Christian neighbours.
I was going to say "WEIRD (enforcement of norms/laws by neutral, impartial arbiters in support of a neutral, impartial ideal-"justice") vs non-WEIRD (enforcement of norms/laws by clan warfare and threat of massive blood feuds) legal systems", but Papua New Guinea has that too and it is emphatically not a place that can be described as having low murder rates, so Islam has to play some factor here (possibly by limiting access to intoxicants?)
Although one might be tempted to note that when murder does tend to happen in the Middle East, it has a nasty tendency to rapidly turn into Civil War.
That (the first paragraph) is the explanation I heard for high crime rates in Africa. Reportedly, traditional African tribes have strict discipline systems that cause their members to behave well. But when urbanization came to Africa - more suddenly than anywhere else due to the extreme cultural gradient between Africans and colonizers - all of a sudden lots of young African men were left without any of the tribal discipline, and many turned to crime. In a sense, just like Native Americans were decimated by diseases for which they had developed no immunity, Africans were socially "decimated" by social "diseases" for which they developed no immunity.
Probably "us vs them" mentality could explain some of the discrepancy.
Also a disproportionately large part of immigrants must come from the worst places, including literal refuges or refuge-adjacent people likely having shitty circumstances, sometimes trauma and firsthand experience with violence, both as victims and as perpetrators.
Middle Eastern countries have moderately high, not low, crime rates. Look at: https://data.worldbank.org/indicator/VC.IHR.PSRC.MA.P5?locations=ZQ. The murder rate averages around 4 per 100k. A bit lower than America's in the first term of Obama but much higher than Europe's.
The source countries of middle eastern immigrants in Germany tend to be the most dysfunctional ones, who sadly don't have reliable statistics. That said, I suspect that you're right and the crime rates of middle easterners in Europe are much higher than in the Middle East by reasons that can not be entirely explained by skewing male and young. They may be cultural effects like the interaction of their honor culture with Teutonic WEIRDness.
"So it’s possible that Europe is still adjusting to being a high-crime continent, whereas America has already adjusted by raising incarceration rates."
The thing is, I can't find any evidence of this really being the case. Overall crime levels seem to be stable or dropping in Europe fairly consistently over the past ten years - https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Crime_statistics - with no sign of the big increase you'd expect given the large influx of migrants. There's an increase between 2021 and 2022 but that seems quite easy to explain by the end of lockdowns.
I can't claim to have reviewed them all in detail or to have looked in the bias, but the relationship between immigration and crime really doesn't seem to be so clear (and certainly doesn't seem to have pushed up crime rates overall). And having visited both major European cities and American cities in the past year, European ones really do feel safer, even in large migrant areas.
Apparently sociologists say the best way to assess crime is to look at the homicide rate. In most European countries, it’s in the 1-2 range for 100k inhabitants while it has been between 4.5 and 6 in the US in the last decade. Depending on what feels like “the right amount of crime” (as much as this can make sense, at least), the trade-off can just appear different.
I think generalizing about Europe like this seems like a mistake. E.g. Germany has a homicide rate of 5 per million people per year while Russia has one of 50, and those are just the two most populous countries in Europe.
similarly, the disparity in homicide rates between and even *within* US states is tremendous. Detroit had a homicide rate in 2023 of ~41/100k. Michigan (which contains Detroit) had a homicide rate of 8.6/100k. Many states are in the 2-ish/100k range, with NH at 1.8/100k.
Yeah, it is overgeneralizing to say that Europe is now a high crime continent. We do however see pretty large increases in eg gun homicides in countries like Sweden that took in a large proportion of immigrants from war-torn middle-eastern countries (murders up 50% since 2007 and gun homicides up like 3X since 2012). But I think when you aggregate over all of Europe these trends are drowned out by mostly decreasing violent crime in much larger native European populations.
This just isn't true - the very large majority are legal immigrants. Sure, a lot of them lied about their rights to refugee status, and it would be sensible to check this a lot harder, but they are in their countries legally.
"Take a moment to imagine how your own life would change if you spent ten years in prison. Seriously, spend a literal few seconds thinking about your specific, personal life."
The problem is, some/many/lots of the people going to jail don't have spouses, jobs, etc. in the first place. Ten years in jail and then coming out to all the consequences afterwards is their life *already*, after six months in jail/multiple 'second chances' every time they came up before the court.
I've mentioned this one example before from my personal experience over my work life, but I'll repeat it once again: a girl that I first encountered as an early school leaver, and because she was engaging with various services that I coincidentally happened to work for, I could track her career over the years.
So she went from vulnerable teenager with disrupted home life, to dropping out of school, to not engaging with the support services for early school leavers, to what used to be called taking up with a bad crowd, but that's probably a disfavoured term nowadays because it's discriminatory and judgemental of persons with societal difficulties (aka being petty scumbags), to (of course) weed and other 'soft' drugs to picking up a heroin habit. Along the way she had a kid with, of course, no husband and the boyfriend flitted in and flitted out.
End result, she stabbed another young woman in the stomach at a house party (drugs and drink present, of course) and ended up with a prison sentence while her mother took care of the child.
She never *had* the "my ordinary normal life of a job and family and the rest of it will be disrupted by going to jail!" In fact, I think prison probably was *better* for her as it was a chance to break those exact connections that were dragging her down, and letting her get sober and clean and maybe, who knows, a miracle, getting her head together and getting skills and help for post-prison life.
There are other people who die early and I see the names in the local paper and I go "Oh yeah" because again, I know the background, and I'm not surprised because this is all the inevitable ending to the life they lived.
I've encountered some of the early school leavers on the programme for such, and I've forecast even at that age (late teens) they're gonna end up in jail, because the little idiots don't give a flying fuck about learning any kinds of skills leading to getting a job and a normal life. That's for the squares and the fools, they want weed, easy money, and free time. They're on the path to petty crime if not already started on it, and one day they'll go too far and either end up stabbed by a bigger, badder criminal or they'll finally run out of "second chances" and have to do time.
And it'll be the best place for them, because they *will* be off the streets and not interfering with ordinary people. They're not interested in rehab even in or after coming out of jail, they'll fall back into the same old lifestyle because all they are interested in is drugs, easy money, and free time, and six months or ten years is all the same.
Re: the shoplifting, I think *all* the commentators are right; Graham about the view of the cops on the ground, Andrew about the "scared straight" but I imagine that for first time offenders or people just starting on the petty crime path, and CJW for how the sausage is made in the courts and legal systems.
I recall there being lots of fears that the lack of fathers around due to overincarceration was going to make their sons more likely to be criminals. Then a study showed it actually made them less likely because these fathers in particular were such bad influences. The average criminal is just not the average person.
I've been listening to A History of Rock Music in 500 Songs and it's striking how many of the early rock stars (it's only up to 1968-9 so we're not really into later ones) come from exactly this sort of chaotic background. One reason for the "27 Club" type early deaths is that if you take someone like this and throw enormous amounts of money and fame at them, they end up taking huge quantities of drugs and dying. Sometimes you get a Keith Richards who eventually does straighten up his act and lives.
Ironically, near the other end of society, I'd be most concerned about contracting HIV in jail through rape--without having to pay for anything else I could just leave the stocks growing in the brokerage account. When they let me out I'd just retire.
You wonder if some calculation like this is driving wealthy criminals like Weinstein.
My mom worked as a teacher in a juvenile prison (for serious offenders). She and the rest of the teachers all honestly wanted to help those kids, but also expected that nearly all of them would just return to a life of crime when they turned 18 and were released.
Incapacitation effects depend upon the birth rate. If there are lots more young people than older people, as there was in 1970 when the criminology professors' orthodoxy hardened, then incapacitation won't do much good. However, in 2024, incapacitation is pretty effective.
I keep hiring about all the crimes committed by people on parole/probation, so I'm surprised to hear "over half of people incarcerated aren't there for any crime at all". Shouldn't there be lots of low-hanging fruit of people to lock up because they're actually committing crimes while out under supposed supervision?
You allude to it twice but then pass over it "for simplicity": None of these studies measure within-prison crime, which is largely not counted, tracked, tried, or prosecuted. It is not only plausible but likely that "incapacitation" is at best simply relocation of crimes. And even _that_ is leaving aside the regular drumbeat of deeply sadistic crimes by guards, which are even less counted, tracked, tried, or prosecuted.
The quality of statistics on crimes within prisons is very poor, of course; ttbomk it is not practically possible to know if the average prisoner commits crimes at a greater rate while incarcerated than while outside, or likewise the rate at which people in prison are crime victims versus those outside.
If your value function is some sort of utilitarian sum over society, you have to count those, or else your utilitarianism is just gerrymandering people into and out of the boundaries of the utility summation (in which case you can optimize utility much more simply than by incarcerating; simple Schmittianism will do the job more easily, and it's free since you don't count the cost).
Agree, I think it's not a valid simplifying assumption to ignore crime in prison when assessing incapacitation, good statistics or not. Incarceration obviously incapacitates some serious (and non-serious) property crimes but makes it a lot easier to commit assault. There seems to be some good data about sexual assault, which seems to be higher frequency in prisons than outside, but not finding it for the (much more common) normal assaults, attacks with shanks, etc.
As I said on a previous post of yours, I don't think the tough on crime crowd realizes how many resources the US system is *obligated* to spend in the case of a criminal prosecution (and you can't have jail or prison without a criminal prosecution).
If you charge someone with shoplifting and insist on 30 days in jail, the defendant can obligate the system to provide
- a lawyer
- a jury of 6 (in minnesota) or 12 (if you're going to impose prison time)
- a judge
- a court reporter
- a prosecutor
- a cop witness
...several of these for multiple hearings (you'll appear before a judge using your lawyer several times in the course of a criminal case)
additionally, that same defendant can obligate the *victim* to provide
- footage of the incident
- a store witness
all for a $5 pack of hamburger patty (yes, I have seen a shoplifting trial on a $5 pack of hamburger patty)
...and at the end of all this, even if the conviction sticks, you still have to sell a judge on sending some poor guy to jail for *thirty days*. Not that it can't happen, but it's such an immense expense of time and effort for everyone involved that it's usually much easier to cut a deal.
We can have juries of smaller than 12 for misdemeanors but a criminal case always entitles you to a jury (and a lawyer).
There are a lot of other points of friction in a courtroom too: to prove a claim of shoplifting, the prosecutor can't just wheel out footage of the incident and show that footage to the jury and say "enough said, please convict." The rules of evidence require things like videotapes to enter through a witness, that witness has to be subject to cross examination, and so on. A shoplifting prosecution would take an afternoon even if the lawyer for the defendant wasn't incentivized to draw the process out.
Where I live, there has been an extended effort to let into the country members of those groups that are relatively higher in the commission of crime, for the past - let's see, when was Reagan - forty years? Against this headwind, how or why would you expect to be able to reduce crime via - prison or any other means?
I was referring to illegal immigrants specifically in terms of their crime rate (other than, obvs, being “illegal immigrants”) being lower than native born people
I had originally thought that the impression that immigrants commit more crime than natives was because white people were using their own neighborhoods as a calibration point, and immigrants commit more crimes than whites, but less crime than the (white + black + etc) national average. But https://siepr.stanford.edu/news/mythical-tie-between-immigration-and-crime says that immigrants even commit less crime than the average white!
There is one major confounder in those studies, which I haven't seen any of the authors try to address (though it's been a year or two since I looked closely at it): a high percentage of criminal immigrants are removed from the country after committing sufficiently serious crimes, while U.S. citizens obviously aren't. So we'd expect immigrants to have a lot fewer opportunities to commit crimes, and American criminals to have a lot more chances to end up going to jail (which, if I remember correctly, is what the studies were measuring).
In a nutshell, it's hard to say for sure whether illegal immigrants (LEGAL immigrant crime rates may be easier to track) commit more or less crime than natives, because studies only count those who've been processed, and a criminal who commits multiple crimes only counts as one.
I'm blown away by the thoroughness and tenacity with which answers to fundamental societal problems are pursued here. This may be a long and dense read, but everyone with an opinion on incarceration should make time for it.
> This may be a long and dense read, but everyone with an opinion on incarceration should make time for it.
Why? That seems like the least likely audience for the piece; people who are interested in the issue will already be familiar with everything this covers.
Lots of people have opinions on topics on which they are uninformed. Plus, the exhaustive nature of this particular piece likely has new information even for active consumers of incarceration articles. It's hard to imagine a person (apart maybe from the researchers themselves) who looks at this and thinks, "I already knew all that."
"People with opinions on incarceration" likely includes a lot of people whose current information about the issue was curated by people less evenhanded than Scott.
What if the question isn’t how to reduce crime, it’s how to reduce the impact of crime on productive members of society? If so, punishment is unlikely to be the most effective approach.
Focusing on reducing impact opens the conversation up. Take the reconciliation practices used in some places. Those have nothing to do with punishment. They are about healing and moving on.
There isn’t healing from petty theft but we could offer people more support. Take a bike theft as a trivial example, nobody gives help the victim. They spend resources to hunt the criminal, if anything, then more resources trying to punish them. The victim is left with zero. As a society is this the right outcome? I am not convinced.
Some baseline crime will always exist in large societies. I’m merely questioning that resources could be focused on helping the productive more directly.
That would incentivize people to pretend to be the victim of some crime for financial gain
Which would have to be gated by spending resources on a process of "approving" someone as a victim, but then you're spending resources on something which isn't reconciliation for the victims, and isn't disincentivizing anyone from offending (if you care about disincentivizing anyone from offending)
---
If you're trying to spare the criminal, I think it's possible for crimes to be neutral or a net negative to the offender (instead of being zero-sum), in which they make gains too if crime is disincentivized
& Even when offenders are more harmed than helped by measures taken, I would think those measures taken (often punishment) and the resources they cost would be justifiable because of the amount of people casually benefiting from living in a safer and more secure world.
The current baseline crime in modern, large societies is what it is partly because of measures taken to catch criminals; and personally, I think there is a limit to how high of a crime rate I'd be willing to tolerate, and once over it I'd find the world a very undesirable place to live in
My read of the stats and studying drug rates particular few people are truly bad, most prefer to live a better life but don’t see it as possible. Crime happens for a bunch of reasons, I think lack of community attachment is a good place to start. That is not solved by the stick/trying to disincentivize.
We already spend money evaluating victims, that’s the bit about investigating the crime then deciding to pursue perpetrators.
I am not expert here, rather the policy discussion circles around levels of punishment with very little discussion of novel ideas.
It’s disappointing and revealing to me that in your list of costs and benefits you seem to forget to even mention the victims of violent crime, who get really messed up for a long time when it happens to them. Or worse, dead.
The costs to the victims was the primary analysis:
> Roodman uses two methods: first, he values a crime at the average damages that courts award to victims, including emotional damages. Second, he values it at what people will pay - how much money would you accept to get assaulted one extra time in your life?
The list of costs and benefits you're looking at specifically only includes the remaining things people care about beyond the cost to victims themselves.
"Second, he values it at what people will pay - how much money would you accept to get assaulted one extra time in your life?"
This is a question so abstract as to be meaningless. It's no different than "How much would you have to be paid to step into the ring with an in-his-prime Mike Tyson?" It's a very easy question to answer, but it's the answer to a proposal for a voluntary transaction, and the answer is less than significant for people who have never had the shit beaten out of them and have only watched crime on television.
"How much money would you accept to have the shit beaten out of you and raped and left bruised and cut up in an alley behind the Howard Johnson's?" is a question that can only be meaningfully answered by people who have had that experience.
1- Make (as in, with a law or something) certain percentage of the population should be incarcerated, no less and no more; let's say 0.4%.
2- Make a huge excel sheet of every criminal in the country with all the original sentence amount totals.
3- Sort by sentence totals
4- Put in jail the amount of criminals that comes up to 0.4% of the total population.
If the total population of the country drops by, say, 1000 people; release the 4 people with the least amount of total sentence from the prison. If population increases, say, 1000 people by immigration or birth, catch and imprison the next 4 people in the list and imprison.
When somebody completes their sentence and leave prison, or die in prison, arrest and jail the next person who is out on the sorted excel list (in fact, maybe tail the top 100 or so people in the list which are just out of the top 0.4% because A- you might need to put them in soon when somebody leaves or the population of the country increases B- they are probably the most likely people to commit crimes and climb in the top 0.4%)
In this way, the "toughness" of the prison sentences becomes a mostly moot point and only relevant in relation to each other. I think this would also be good media! TV shows about the people about to get released whenever somebody on the outside do some more crime and push them off the list; or the coverage of the manhunt when somebody is released and the next person in line have to be caught and put into prison. That would be great entertainment for all!
Lastly, an unrelated idea about the shoplifter issue. If the problem is the court sources being exhausted for this kind of minor crimes that happen often; why not combine all of them into one (or a couple of) huge organized crime trials? Instead of trying to try 300 shoplifters one by one; see if they have some common people that they sell their stolen merchandise to; make a relation map and in one big organized crime trial, try all of them as one monolithic gang. Use the court resources smart! Added benefit is, organized crime probably has a lot longer sentence than simple shoplifting so they'll be out of the community for longer!
God I hope I was the person who was in charge of deciding everything for the world.
This likely will alter the cost equation - need to factor in cost of expanding capacity. In the case of NY, I’d be shocked if there wasn’t capacity to incarcerate the 327 shoplifters, but maybe applied more broadly it’s a factor in choosing the types of crimes to allocate prison capacity towards
Overcrowded / resource-strapped prisons also probably have a lower impact on recidivism
"why we can’t punish repeat shoplifters at all, with short sentences or long sentences." The impediments to arresting and convicting criminals are intended to prevent Type I errors--convicting the innocent. But we probably could identify procedures that have almost no marginal impact on reducing type I errors but result in very large marginal increases in Type II errors--letting incorrigible criminals go. Eliminating such procedures would be a win.
To adjust my views on prison, I'd need to see how it compares to other methods of reducing crime. As a start, what's the marginal cost/benefit of the things you hear from your liberal theater aunt? In order of squishiness/relevance: how relatively effective would it be to increase access to jobs, make restorative justice more available to juveniles, invest in rehab centers, create after-prison reintegration programs, incentivize early parenting education (e.g., NFP in NY), build more food banks, deliver on UBI, remove the millions of service lines that still have lead pipes, give mandatory therapy to non-violent criminals, plant more trees in poor neighbourhoods, ship criminals to Australia, and make them watch Turin Horse... you get the idea. The last three are a joke, I think.
What do you really mean by "increase access to jobs," though? If you're forcing major employers to take on more ex-cons, they'll see that as a cost and take steps to minimize it, probably involving treating their employees worse than before.
You didn't list "hire more cops" and "spend more on the court" in your list. They might not be things your liberal aunt says but that's what Scott recommends instead of increased sentencing (and the commensurate cost).
I'm in favour of whatever is likely to address the issue in the least harmful way. It's just a matter of putting everything side-by-side and making a reasonable assessment based on the available data, and I'll put my thumb on whatever scale works (within reason). I'm not against hiring more cops or spending more on the court if that meaningfully reduces violent crime or helps deal with the few NY kleptos, etc. But say someone is doing a deep-dive comparison between amputation and bloodletting for gangrene, and it turns out antibiotics can solve the issue in most cases, I'd sure want to know! I'm not suggesting the analogy is appropriate here, but if I'm going to argue for something like having more guns on the street or locking people behind bars, I'd want to know how effective they are compared to other methods that are on the table. Just basic utilitarian thinking. (NOTE: I enjoyed the article and am in no way criticizing the framing here, I'm just thinking out loud about how these solutions proposed compare to other ones. I appreciate your question.)
I think the poll is done the wrong way around, in a way that should be economically neutral but probably psychologically is not.
Instead of asking: "How much money would you accept to be assaulted an additional time."
Ask: "You have already been assaulted. You live in Ancapistan, so you have to fund the imprisonment of your attacker. How much of your money do you dedicate to keeping them in prison?"
I think this works less well because most victims of crime are poor people who don't have money. Also, keeping your specific assailant in prison doesn't reverse your assault! I would spend lots of money not to be murdered, but keeping a murderer in prison would be a social (not personal) good and get much less money from me.
> I would spend lots of money not to be murdered, but keeping a murderer in prison would be a social (not personal) good and get much less money from me.
You might be surprised. I predict that, despite what you say here, that would actually depend on who the murderer had murdered.
This is way off. I would spend vastly more to avoid being a victim of crime than to imprison the perpetrator after the fact- that does me no good. In fact in many scenarios I would spend literally nothing to prosecute or imprison the perpetrator (why should I?- locking them up doesn't benefit me unless I think there's a material chance of the same offender going after me again).
Except in a few cases (domestic abuse springs to mind), this will be the normal state of affairs. Punishing a criminal has benefits, but usually almost none of them acrue to the victim of the crime; they are distributed across society.
Right. If some dude breaks into my house while I'm gone and steals my stuff, the benefit to me of putting him in prison for a few years is very small--it's not likely he was going to hit my house again. But it's still a net win overall--everyone has their probability of getting burglarized go down a bit.
I get what you're saying about psychological non-neutrality, but your proposed reformulation seems economically incoherent. The amount I'm willing to pay to keep someone in prison is almost certainly much less than the pain-and-suffering cost I sustained from the crime itself; I'm not going to devote a huge amount of my own resources to getting revenge, or to protecting other people from a criminal who isn't especially likely to re-victimize me personally.
One other comment I want to throw in here. Hang around for a while with correctional workers, at least in Minnesota, and you will eventually hear someone talk about “evidence-based practices,” usually right before they spit at the ground. Coming from secondhand medical background, this confused the hell out of me for a solid six months. A) In medicine, EBP is a set of practices designed to make sure that commonly used interventions, you know, work. B) Evidence seems… good? Not here. A new director might be described as “super EBP” with a roll of the eyes.
I eventually worked out that in practice, EBP represents a rehabilitative, “soft on consequences” approach in prisons that a lot of front-line workers, especially older ones, think is ineffective at best or just plain too kind to offenders who deserve worse. So there’s a major divide between Central Office people who are trying to rehabilitate the inmates based on “best evidence,” and the guards and case managers who think that time should be hard. It’s a pretty good microcosm of the current epidemic crisis, actually.
There's this disconnect in general where people who've never in their lives seen a hardened criminal up close get to decide policies, and even the whole philosophy of dealing with crime, which ends up in a predictable clusterfuck. (There's a similar issue in psychiatry, from Scott's telling.) Essentially this means that society is still too immature to understand these issues on the institutional level.
I think this generalizes to "people who've never in their lives seen X up close get to decide policies...which ends up in a predictable mess" for most values of X. Education--often the people deciding the curriculum and policies aren't the line teachers who have to deal with the effects. Business--the higher-ups make policy and buy software, the grunts on the ground have to implement it. Basically, as soon as the decision-makers are separated from the doers by too many layers, bad things happen.
I think this view of deterrence is fatally flawed.
When a young man decides to go to dental school, he doesn’t do a careful EV calculation of the RoI of that choice over his anticipated working years. He just notices that dentists seem to have it pretty good, so maybe he should go for that.
Suppose tomorrow there is a subtle change to dental insurance law that causes dentists to start making about 10% less per hour worked. Is this kid going to immediately notice and change his mind about dental school? That seems very doubtful to me. He’ll still be a dentist, but maybe he’ll end up regretting the choice, and maybe 20 years from now, kids will have noticed all the dentists with regrets, and over time you’ll gradually start to see falling enrollments at dental schools.
I’d guess deterrent effects on would-be career criminals will actually mostly work via the same slow and indirect mechanisms, and none of the studies you described would be able to pick up on these effects at all.
I think this is true in the sense that criminals just go off a general gestalt feeling of what the punishment is, rather than the actual punishment. But I think this is relevant to the question we're asking - if the government increased sentences, criminals wouldn't know, and it wouldn't deter them.
We agree that would-be criminals mostly don’t know the laws on the books. But they do know about their cousin Joe. They know Joe has poor impulse control and commits lots of crimes when he’s out of prison. They know roughly what fraction of his life Joe spends behind bars. Suppose Joe’s life outcomes affect their assessment of whether the criminal lifestyle in general is likely to be a rewarding one.
Is this something that could be found in the studies you are using to assess deterrence? I think not, and I also think this kind of folk wisdom mediated deterrence is likely to be the most important kind of deterrence.
That would mean that increasing sentences would work, just with a delay, and not in a way that you know how to measure.
It didn't take long to liberalize the system in the 1960s toward more crime, but it took the 1980s and part of the 1990s to change the culture back toward less crime. This would suggest to me that elites should be very cautious before deciding that we have too much law and order.
Or, how’s this one: Suppose some Brazilian farmers don’t understand that their traditional cassava preparation techniques eliminate cyanogenic glycosides. Then, we could logically conclude that giving them a new strain of cassava without cyanogenic glycosides won't affect how they prepare cassava tomorrow.
If you largely agree with this, then Steve Sailer's point above in the thread about the 10-20 year lag in "soft on crime society" vs. "tough on crime society" permeating culture should also become more plausible.
The last sentence here is wrong, and I think "splendric" is spot on (I was waiting to see if someone would make this point).
Take a step back and look at the result that "deterrence effects are small" estimated (in the good papers at least) from local variation between two sets of would-be offenders. A common interpretation of "deterrence effects are small" is that it stems from "the average criminal is not exactly a genius and does not logically account for long term implications of their behaviour / think about the exact consequences of law breaking" - you can't expect to deter them by threatening consequences because they don't think about them. You or I are much more capable of making this sort of calculation / consideration of future consequences, but we're not the ones committing common crime.
But if that is true, it basically means they are particularly unlikely to attend to institutional quirks (of the legal system etc) that all of these studies exploit to estimate deterrence. E.g. if someone with two strikes faces much higher punishment for an additional offense than someone otherwise similar with one strike, this distinction won't achieve very much, because the would-be-criminals won't attend very much to it (If at all).
But it's much easier to attend to "the general vibe of a thing" than to "precise institutional details". Do criminals deeply consider how many strikes they are sitting on, or changes in sentencing laws, before deciding whether to commit crime? Perhaps not. But changes in laws / becoming more/less "tough on crime" should (often with lag) change the general gestalt feeling of "is society cracking down on crime", both because vibes are rapidly transmitted in the zeitgeist plus there is social learning. And it is much easier to respond to this, and in fact people surely do (e.g. when I decided what degree to study, I had a rough understanding that my field pays well and this presumably informed my decision, rather than knowing "the exact causal effect of doing major x instead of major y on lifetime earnings for someone with my demographics"). Which in precise academic speak means there will probably be a violation of SUTVA - you introduce a law that affects group A and not group B (or effects them differentially), but because of the change in the zeitgeist, both groups change their behavior. Thus studies that isolate differences between group A and group B will underestimate the effects of said laws. And the more you think "criminals are too dumb / incapable of attending to complicated incentives to change their behavior in response" the greater the fraction of the overall effect you should think happens through the zeitgeist channel, and thus the more you should think that these studies can't capture the true effects. And thus the more credence you should give to the possibility that the big crackdowns on crime in the 80s did produce the cratering of crime rates from the early 90s etc. The post-GF "defund the police" homicide spike is a great example of this - as BLM apologists like to point out (in trying to argue that Ferguson effects can't be real), not a lot of defunding actually happened - certainly not enough to explain a 50% increase in the Black homicide rate. Yet we saw a huge spike in the crime rate because overnight the message was communicated that "it's open season to commit crime and police should be afraid to step in" (In a slightly different "Ferguson effect" example, Roland Fryer tells a story of gangs getting junior members to openly break gun laws in front of police to verify that they won't step in anymore, before the gang would launch a gang war).
When society eases up on law and order, criminals notice quickly. E.g., Chicago's worst day ever for murder with 18 homicides was six days after George Floyd died. When society toughens up, it can take years for criminals to get the message. Implication: be very prudent about loosening up.
> But I think this is relevant to the question we're asking - if the government increased sentences, criminals wouldn't know, and it wouldn't deter them.
That's only relevant if the policy change you're considering is "we make sentencing harsher for a week, and then we undo that change and put it right back to the original level of harshness".
If you want to know what effect a permanent change will have, then you can't assume criminals won't know about the change, because they will know. When sentencing changes, the gestalt feeling of what sentencing is like will also change.
Three times in my lifetime, the zeitgeist shifted toward less law and order (1960s-70s, Ferguson, George Floyd) and crime rose. The first was more pervasive and took longer to roll back because it involved more changes, such as shorter sentences. The latter two could be reversed in a few years because they mostly involved less policing, so the politicians telling the cops they were back on their side did the trick.
Also, while criminals tend to be pretty dim, cops tend to be of average intelligence. Cops pay attention to the news so the Ferguson Effect and the Floyd Effect seem to have been due in sizable part to cops noticing that politicians, the press, and other influential members of society wanted them to police less aggressively. So they retreated to the donut shop.
Thinking about it from first principles, it seems intuitive to me that deterrence effect of longer incarcerations (for first and repeat offenders alike) would be negligible, due to scope insensitivity: that the brain just can't multiply. People can kinda visualize what a night in prison would feel like, but why, among all the other domains where research shows people cannot, would they be able to correctly weigh the disutility of of 30 or 365 or 3650 nights in prison? You'd be lucky if they can evaluate it logarithmically.
Which kinda gestures towards a class of alternative solutions: what ARE the sort of things people can tangibly feel? How about corporal punishment (or, let's not beat around the bush: torture)? I don't know if there's scholarly consensus among biologists, but from what I gather, pain (emotional pain would also suffice, but is more difficult to administer) is an adaptation to efficiently train individuals to avoid harmful stimuli, and imagining the peak intensity of various forms of torture would be very vivid in the minds of potential offender (while the intensity of imprisonment is similar to waiting for a delayed bus, only suffered through prolonged stretches of time), and so, from first principles, I would expect it to have maximum deterrence effect. Moreover, unlike imprisonment, it does not destroy individuals' social lives, and just like one can try to estimate the disutility of e.g. robbery at $25,000 as was done in the article, one can try to estimate the disutility of torture, and I would wager criminals themselves would prefer torture to imprisonment.
Is my intuition completely wrong with this? When (Western) societies abandoned corporal punishments, had they concluded that these punishments don't, as a matter of actual fact, work? Or was the decision made on purely philosophical grounds of rights-based rather than duty-based moral theories, and aversion to any form of coercion and violence (which nowadays manifests itself in effective permission to commit atrocities ranging from schoolyard bullying to megadeathcrimes, as relevant authorities tie themselves to knots avoiding any actions that would Actually Work).
I'm all for corporal punishment, but it lacks the incapacitative effect of imprisonment, which seems to be the main positive effect (on the margin, at least). My preferred solution to that is the death penalty for anyone who needs to be incapacitated, but that seems to be pretty far outside the Overton window.
Seemingly, incarceration in a low-cost prison (e.g. more or less permanent solitary confinement so little guarding needed) would be vastly superior to just killing criminals of all crime levels.
Death penalty requires a higher standard of proof (rightly so, I think, given how much harder a wrongful conviction is to fix) and as a result is in many cases far more expensive overall than just locking someone up indefinitely.
I could support such a system only if the prisoners were given the option of euthanasia. Forcing them to live indefinitely in such a state seems unreasonably cruel.
The very fact that it's over with quickly might make it more appealing to prospective plea bargainers - and arresting officers - than a drawn-out trial and appeal process, thereby reducing prosecutor caseload, thus improving swiftness and surety, which could plausibly be even more important than the incapacitative effect.
> it lacks the incapacitative effect of imprisonment, which seems to be the main positive effect (on the margin, at least).
But it also lacks the "prison as a crime school" effect, which is apparently pretty strong for beginner criminals. If the first couple of sentences would be low, replacing them with (reliably guaranteed) caning might be a serious improvement, on the current margins.
I agree. If someone is unlikely to reoffend after a caning, they don't need to be incapacitated. Probably many of them should be monitored via ankle bracelets, however.
Considering how criminality varies by sex and age, it also seems plausible to me that hormonal treatment could have a rehabilitative effect, so that might be an alternative to incapacitation.
> Which kinda gestures towards a class of alternative solutions: what ARE the sort of things people can tangibly feel?
Probaby: "I just saw a cop car roll by here five minutes ago. Better not do crimes for a while." Which isn't all that expensive an experience to create, compared to many of the alternatives on the table.
The age-crime curve is a big reason to think that the criminogenic effect of prison is to reduce crime. It's roughly as simple to understand as the incapacitation effect.
Well, if a 20-year-old kills someone, and you let him out of jail when he's 60, you get 40 years of incapacitation. Most of the crime prevented by that stretch of incapacitation will have occurred during the first 20 years, for age reasons.
But also, after you let him out, he's unlikely to commit crimes because he's old. It seems difficult to count this as "incapacitation", because he's not in jail. But it's very relevant to the question of "after this guy gets out of jail, will he be more or less likely to commit more crimes?". He'll be less likely.
That's what _I_ thought we were talking about with aftereffects, at least.
"An optimistic school of thought hopes that stricter sentences will decrease crime after release. After all, the longer you’re in prison, the longer you have to be scared straight, or find God, or get rehabilitated.
A pessimistic school of thought suspects that stricter sentences will increase crime after release. I find this one more plausible."
What does seem to happen reasonably often is that maybe sometimes in their forties, a lot of career criminals get tired of this shit and yo-yo:ing between prison and a criminal life on the outside. It seems to be neither being scared or being rehabilitated, and instead just realizing that this is a crappy life, actually worse than having a boring 9-to-5 job.
An amount of prison is probably important to make this happen. It also means that a lot of the time, you don't have to lock people up for life - only until 45.
If we start to kill people on the grounds that they have a net negative contribution to society (economic or utilitarian), we will get a _very_ different society, and probably not for the better.
Wrongfully convicted folks definitely don't win in that system relative to the current one, and there'd be a lot more temptation to frame someone (especially when already "in for a penny" - I mean, what's a judge gonna do if he catches on to the evidence tampering, have you killed twice?) so the rate of wrongful convictions would presumably increase.
I know you don't like to apply your own conflict theory and ascribe self serving motives to large groups <snark>other than AI risk doubters</snark>, but it's high time you start considering whether there are whole fields of research dedicated to reinforcing pet conclusions about major issues.
We know for damned sure that this is possible. The cigarette industry did *fifty years* of research proving that cigarettes were safe. We now know that that research was all bunk, but for a very long time, it was treated as reasonable, well meaning, rigorous work.
I won't say which side of this particular discussion might have such a self serving bias, but I think you should consider whether there is such a side and then chase it down in your customary thorough and detailed manner.
I would love love love to see an ACX post titled "Is One Side of the Crime Debate the New Tobacco Industry Science?"
I went through lots of different studies, compared readings by people on opposite sides, and solicited feedback from people on both sides. I discussed which parts everyone agrees with and which parts are controversial. I also included the cost-benefit analysis which lets you do a sort of sensitivity analysis if you want. I don't know what else you want.
I want you to consider whether one or the other side is entirely constructed and framed to service a desired answer. This is related to but different from attempting to find the true statements inside each side's frame.
Consider the possible conclusions of what you've done: "finding A seems substantially true". "finding B seems directionally true but I have concerns about the methodology". "I'm not convinced that finding C was reached fairly". All good conclusions.
Compare the possible conclusions of what I'm asking: "I have looking into the community and history of sides A and B and cannot really find anything that makes me think they were or are anything but honest academic communities with reasonable self admitted biases". "After looking into the history of side A, I believe the academic community around it was built in a very self serving and motivated manner, ala Phillip Morris, and while there are many well meaning people on side A, they are working in a structure that's built intentionally to serve a specific outcome". "I am surprised to find that the academic communities around both side A and side B were built dishonestly".
To the modern ear, the latter set of conclusions sound harsh, maybe even paranoid. But there are real examples of motivated research communities constructed to reach motivated conclusions. Foods. Pollutants. Politicized departments in scary 20th century countries.
I am aware of the theory that the only reason anyone is ever tough-on-crime is because they're racist and the whole field of criminology was founded on racism from the get-go - but I'm just so sick of this kind of argument. "Oh, I can think of one reason the other side might be biased, therefore they're all intellectually bankrupt and you can safely ignore all the evidence they produce." I'm just sick of this. Everyone on both sides always believes something like this and they're always trying to convince me I'm dumb for looking at the evidence and taking it seriously, and I've stopped caring and I'll stick to that.
There are a couple of issues where I genuinely believe one side is operating in bad faith - tobacco companies, race/gender, etc - but I try to have a really high threshold for believing it so that people aren't constantly sniping me with "but the other side is evil so you have to dismiss them out of hand right now!"
Cards on the table, I actually believe the *soft-on-crime* side was wrong from the get go.
And maybe, maybe to a lesser extent the tough-on-crime side was rotten, but if so, a far lesser extent.
I haven't done the work, and won't because it's not my job or passion, but I think if you did the work, you would find that the soft-on-crime side was in the dead center bullseye of the motivated, flawed psychology research criticized by John Ioannidis, Richard Feynman and others.
A more interesting question perhaps is, why does one believe one side is operating in bad faith? Why do I believe soft-on-crime is done in bad faith? Why do you believe race/gender is done in bad faith?
I believe soft-on-crime was in bad faith for two reasons. First, I've lived in a lot of places (at least 9 US, plus 2 international), and like anyone else I have friends and family from a few others, and anecdotally I see a clear picture: the scary draconian places have women walking around at night, and the cool progressive places have women not doing that.
California in particular is the worst offender. My personal experience includes SF knife wielder chasing a friend, SF random serious head wound guy, SF car breakin, LA public masturbator, LA car breakin, San Jose light arson, San Jose car breakin, San Jose deadly overdose on playground, San Jose overdose in apartment hallway, confrontation with SF sword wielder, confrontation with SF car kicker, and eight bike thefts. Mountain View and Palo Alto, which are markedly more straight laced and somewhat more policey, are commensurately nicer to walk around at night. El Paso downtown is extremely nice to walk around at night, and they have policemen shooing away skateboarders. Beijing is extremely nice to walk around at night, and there are men with sticks ready to drag you away if you are a public nuisance.
The second reason I think soft-on-crime is in bad faith is that I came from a strong and rigorous research background (math, with a side of chemistry and a family in medicine and pharma), and I bummed around in psychology research for a while, because it was interesting, but it was *bad*, it was really badly conducted, it was cargo cult in the way Feynman claimed, it was corrupt in the way Arielly was eventually outed, it was feckless in the idiots-with-fMRI-machines way. So I don't really trust when it conflicts with my anecdotal experience.
Again, that's just my experience motivating my feeling about crime. If you don't share it, the claim won't appeal to you, and that's very understandable.
It seems to me that's exactly what Scott did. He didn't find the smoking gun you want, but also he wasn't looking at research published by the private prison industry association.
It seems like we need to move the Overton window to allow for much higher levels of incarceration for the incapacitation benefits. This would require a seemingly insurmountable political effort to convince our fellow citizens that it would improve society and the effects would be concentrated amongst our most vulnerable and criminal adjacent populations. We could layer in all the type of rehabilitative efforts and strategies that the soft on crime folks want, but the main goal is to get criminals off the street including the zombies on Methadone mile in Bston, Kensington in Philly, etc.
One very deep area of needed analysis and change is the effect of the Warren court on criminal justice. The history of Miranda, Gideon, Terry, etc. was really the first "defund the police" movement that we all view as needed a heroic civil rights movement against the police state. In reality they were a burden to law enforcement that created this procedural nightmare for our criminal justice system that not only lead to a spike in crime, but is now used as a reason to not imprison criminals (It's just so expensive!).
Thomas Sowell outlines this well (along with how the sex education revolution arguably caused a spike in teen pregnancy and VD) in his book Social Justice Fallacies. I'm sure there are plenty of holes in his argument but even after a vicious attack I think it's mostly right.
The modern history of cities and towns ravaged by devastating crime is one where cops know who is doing it, where they are doing it, but are impotent in making a meaningful difference. All our talk on solving this issue has to work around that reality. I wish we could change that.
It seems prima facie really unlikely that prison would have _no_ effect, in either direction, on later criminal activity. After all, it's a massive intervention in someone's life - what are the odds that it wouldn't matter _at all_?
Instead, it seems a lot more likely that the data and research tools available don't produce any results, and that this lack of results gets erroneously transformed into a "result" that there are no effects.
I think the way I would look at it is that if prison didn't have enough of an effect to be detectable with current methods, you would get better methods. If no methods are up to detecting the effect, then the effect must be very small and you can round it off to zero.
Or the optimistic worldview: societies are so well-calibrated that they adjust the level of incarceration to the exact spot where the marginal benefits are zero. As a perfect society should.
I'm joking, but only halfway: there is some feedback loop, even if it operates on the level of decades.
A perfectly calibrated society would not have zero marginal benefits from incarceration; it would have zero difference between marginal benefits and marginal costs. The actual marginal benefits from something as expensive as incarceration should be easily large enough to detect, unless the measurement methods are extremely bad.
I thought what research shows is that it has a large incapacitation effect and a small to potentially large aftereffect in the opposite direction, with deterrence at current levels of incarceration being minimal. That effects cancel each other out indicates the intervention does have big effects, just in mutually canceling directions.
A note on shoplifting. If you can't punish shoplifters because you don't have the resources to try all of them, then you have two main options to change that.
(1) Fund some kind of "shoplifting task force" (or petty crimes task force if you want broader reach, but this will spread your efforts). Hire more prosecutors, judges, public defenders, bailiffs, court reporters, etc. and try more cases. If it's constitutional, you could target this effort at people with multiple prior recent arrests first, although the population will get the idea that everybody gets a few free misdemeanor tickets.
(2) Increase the difference between the minimum sentence that you get if you go to trial vs. pleading out. In other words, create "first degree," "second degree" etc. shoplifting and make the penalty for the higher degrees painful enough that people will plead to third degree shoplifting and do a year in prison rather than risk a conviction for second degree and five years.
#2 is inhumane and has been reasonably compared to medieval practices of confession under torture, but is also the way we manage our trial dockets in the US, unfortunately.
> the population will get the idea that everybody gets a few free misdemeanor tickets.
That seems like a fairly acceptable outcome. Ceiling would continue falling at a hard-to-predict rate, so anyone smart enough to do the math probably wouldn't want to get too close to the de facto limit - and if there's a broader societal sense that something is being done about the problem, even first-time offenders would face a higher risk of being a) reported at all, b) while the prosecutor happens to be in a bad mood, or having a slow day.
It's been noted elsewhere in several places, but I'll note here too that another "main" option to change that situation is to lower the amount of resources you spend to try each shoplifter.
Scott did you find anyone who even tried to answer the seemingly hyperimportant question of 'how do we reduce the impact on recidivism from gang activity in prisons'?
I could imagine it being much more cost-effective for the government to directly pay some entity to hire former convicts, with incentives related to what fraction of them reoffend, as an example.
Tons of complicated incentives to navigate for sure, but seems a lot worse than the default of prison being training camp for more crime?
Also why can't we have corporal punishment/labor camps offered as an alternative.
I.e. 'you can have the current system, or you can opt for 1/4th the duration at hard, barely productive but somewhat productive labor' and see if that reduces recidivism.
i posted elsewhere about Skarbek's book, but based on reading that book, it seems to me you could put a very big dent in prison gangs just by making prisons very safe places where the prisoners don't need to fear violence from other prisoners. this would also rob the prison gangs of the power they exercise over gangs and criminals on the outside.
Let any prisoner spend as long as they like in solitary - in or out at mealtimes, maybe - with a cheap laptop and an internet connection, so long as they don't physically break said laptop (allowing some slack for accidents, repair / replacement costs) or use it to commit additional significant crimes. Gang recruiters would need one heck of a sales pitch if the BATNA is "left alone to watch cat videos!" Room for further improvement from there, of course, but you might as well grab the low-hanging fruit first.
> Let any prisoner spend as long as they like in solitary - in or out at mealtimes, maybe - with a cheap laptop and an internet connection
We don't even let them have internet access after they leave prison, and you want them to have it while they're in there?
(Last I heard, some judges were starting to be uncomfortable with cutting people off from the internet on the grounds that it's functionally equivalent to casting them out of society itself, but I suspect progress on that front is slow.)
There are far more plausible ways to rape or murder somebody through a foot-thick concrete wall, than there are via TCP/IP - assuming reasonable levels of monitoring to rule out e.g. drone weapon platforms. If reducing homicide and sexual assault within prisons is a desirable goal, that seems like a plausible candidate for cheapest AND surest way to make it happen.
Probably at least a few prisoners will lack technical savvy about exactly how much spyware can be crammed into a state-issued laptop, and thereby end up exchanging emails which score some lucky prosecutor a pile of easy wins.
To sell it to judges, consider principles of self-defense. Almost always better to retreat than fight back, given the option, right? So give 'em the option. But long-term sensory deprivation is too torturous to be a viable alternative to violence - sorta like expecting somebody to retreat from their own home. Obsolete laptops can often be obtained at scrap prices, while prisoners who earned some degree of trust could practice valuable trade skills by fixing up busted ones, meaning costs stay low and there are plausible rehabilitative benefits even if the internet access itself is considered detrimental.
I feel like this is going to work a lot better on ACX readers than actual criminals. Lots of people *like* to fight, especially in jail, and lots of men want to be a member of a group that has a strong identity and fights with other groups.
Conversely, many ACX readers may find solitary confinement with a laptop to not only be preferable to being in a gang in prison, but to normal life *out* of prison.
Lots of people like to *win* fights. I'd guess not as many enjoy losing. With a line of retreat available, some will take it - reducing available targets for the worst bullies - and many more will think it over enough to undermine their own courage, instead of benefiting from the clarity of desperation. Fewer, fairer, more fully consensual fights still feels like it'd be useful progress.
Strong shared identity and frequent inter-tribal physical conflict, but unable to coerce folks into joining, because anyone who doesn't like the vibe can just leave without first needing to negotiate for another group's protection, seems like it could shift that overall social dynamic away from "agoge," https://acoup.blog/2019/08/16/collections-this-isnt-sparta-part-i-spartan-school/ more toward something acceptably similar to "sports team."
This isn't data, but at least one cop told me that what is needed is not longer sentences, but swift and sure sentences.
The prison time can be short but you need to know that you are likely to be caught and sentenced soon after committing a crime, and you will serve your entire sentence.
Not sure whether that is reflected in the data.
Sort of like toilet-training a puppy. You gotta catch them in the act to get the training to work. Going back two minuites afterwards has no effect; you might as well be talking about ancient history to Fido.
I have a fair amount of experience in Europe. Europeans are all about social norms and The Done Thing.
That said, I was in Brussels during the Marc Dutroux Fiasco, and a woman screamed at me that in Texas, this guy would be getting lethal injection already. I told her that she should not get her hopes up, he'd be on his fifth criminal appeal or so.
Friend of the Comment Feed @grahamfactor had a post noting significant differences in US and Euro policing norms, highlighting how Euro cops are allowed significantly more leeway and freedom in investigating people and seizing evidence
Also, it's hard to make punishment swift and certain while simultaneously providing robust procedural safeguards against railroading innocent defendants.
If it's swift it can be argued that the damage would be more minimal. And innocents are less likely to be repeat offender (?), which seems to be the main topic here.
Procedural safeguards become less important when the cost to a genuine innocent is less severe, which becomes feasible if punishment per offense is progressively stepped down as response times get tighter. Tomorrow in jail (but out by tuesday) for a misstep today scares a high-time-preference criminal more, *and unlucky good citizens less,* than twenty-to-life starting eight months later.
In the case of honest mistakes, I am tentatively inclined to agree with you, if we're talking about modest fines, a day or two in jail, or a few weeks' probation.
But I am worried about wrongful punishments imposed by police and magistrates with arbitrary, venial, or malicious motives. E.g. paperwork says "disorderly conduct", but the actual offense is "failure to bribe" or "contempt of cop" or "being a visible member of a disfavored minority". Likewise, wrongful punishments could result from a malicious accuser who is sincerely believed by a well-meaning magistrate.
The lower the standard of proof and the fewer procedural safeguards, the more latitude those proposing and approving punishments have to abuse the system, and the more abuse I would expect to be attempted.
At a minimum, there needs to be a hearing with a moderate burden of proof prior to punishment, plus an opportunity to appeal after the fact and get compensation and a clean record if exonerating, and also some kind of systematic oversight looking for abuse of the system.
My takeaway is that we need to work much harder on deterring crime with police and other measures, both to prevent the crime itself and to change the political calculus on prison. If crime were less of a problem, there would be less knee-jerk desire to increase imprisonment.
I don't think it makes sense to talk about deterrence as "% decrease in crime per extra year in prison". Shouldn't we talk about "% decrease in crime per % change in prison time" instead? Making the average murderer spend 19 years in prison instead of 18 years probably doesn't reduce murder much. But making the average junkie spend 15 months in prison than 3 months seems like it would be much more of a deterrent. If this is true, than deterrent effects would be much more powerful for smaller crimes, and could even become large enough to be worth considering for those?
I was a bit confused about the age thing until I realized that age isn't an aftereffect for being in prison (free people also age) so we shouldn't include it as a prison aftereffect.
I am now sold on the Australian model as a solution to basically all crime.
No, not modern Australia, which sits comfortably above the European system and below the American one in terms of incarceration rates, and doesn't seem to be doing anything novel, but the Australian system as it was for the United Kingdom a few centuries ago. Penal colonies.
If the problem with increasing prison sentences to the point we eliminate 90% of the crime is that it would cost insane amounts of money, then shipping those prisoners to a completely separate society somewhere far away would have the incapacitation effect (which seems to be by far the most important factor), without any of the significant costs. A one-way flight ticket is maybe a thousand dollars? Pretty much anywhere in the world.
If you think penal colonies are cruel, present the option to the criminal to join one, or our current system. If 5+ years of imprisonment essentially breaks all your social bonds anyways, it seems a much preferable alternative to have freedom in a new, if harsh, land, rather than be imprisoned in a literal prison near to home. If you're worried that penal colonies will be dangerous and deadly, as they're full of criminals, that's sort of what prisons are already. Except with a colony you can go out to the sticks and rarely interact with other people, and can have the means to readily defend yourself.
Unfortunately it doesn't look like there's anywhere left that could function as a modern day penal colony without being cruel. The whole idea is incapacitation, so it doesn't work if we just ship them to Montana or something and call it a day, as they could easily take a bus out of there. It would have to be separated enough that returning to society without permission was sufficiently difficult. Even a large island (let's say we sacrificed Hawaii's big island for this purpose) can easily be escaped if not for significant monitoring.
I think Robert Heinlein's, The Moon Is A Harsh Mistress should be our model for an actual solution. The moon is far enough away that it would be quite difficult to get back, and the long term effects of microgravity might make returning to earth difficult, painful, or straight up impossible. For those of you that don't know about the book; the moon was essentially one big penal colony. There were guards, but they didn't actually police what you did so long as you weren't actively sabotaging the operations of the colony. Mars could theoretically function as the same sort of thing.
Could this be the most compelling argument for Mars colonization? Could this be how colonizing Mars actually makes sense for improving life here on earth? Probably not. But it's fun to think outside the box.
Somebody who burrows their way back up to the surface in Ohio can find edible wild plants, liquid water, and hitchhike to anywhere in North America. Somebody who does so in Antarctica simply freezes to death. Maybe, if they're resourceful and lucky, they make it far enough to build a boat (though surely not from local trees...), avoid being eaten by leopard seals, and start a new life in South Africa, Tierra del Fuego, or get shot by whoever's guarding the fence at McMurdo AFB.
But, I don’t think it would be a good thing for US soft power to be sending raving bands of armed criminals to conquer and occupy neighboring countries.
A one way flight ticket + a rifle and ammunition would definitely be cheaper than whatever we’re doing now.
I think anyone who would be comfortable with something like that would also be comfortable with executing the prisoners who are repeat offenders by firing squad though, which is much simpler. Most people who fall in the Overton window of thought on the issue probably don’t have any reason to want to conquer and occupy Afghanistan or Iraq.
Loss of exile as a punishment is related to enabling internal mobility for the law-abiding.
When criminals were exiled, new entrants to a community weren't exactly welcomed with open arms.
(Also: exile, the severity of which is graded by distance from home, is a really common punishment in traditional China. But it's described in more detail in Water Margin (all the characters are criminals), and it involves substantially more than exile - in Water Margin, once you reach the place to which you've been exiled, you're imprisoned there! You're also branded, or tattooed - it's some kind of permanent mark on the face - but that doesn't surprise me as much as learning that "exile" means "being held in prison, but farther away from your home than usual".)
Cost per kilogram to orbit would need to drop by quite a few orders of magnitude for outer space penal colonies to make any kind of sense, even if you're not sending much life support gear along with them - which seems like the kind of thing there'd be protests over.
Note also that the total population in prison in the US is under two million, and this is considered a large number in an international context, so everyone will fit comfortably in one city. It wouldn't even be a large city.
You’re right, and if it was to be done it would have to be a deliberate use of our controlled empty land.
The idea in my head at least, was that they end up so far away, and outside much if any state capacity, that you really didn’t have to worry about doing any work (or more like spending any money) to keep them contained.
The necessities of practicality always seem to beat any speculative ideas I have. 😔
Actually, I'd be completely on board with a scheme like:
- We carve out a modest region which is now The Prison of the United States
- We put up customs barriers around it so prisoners can't leave
- Non-prisoners are free to go in or out
- We don't try to administrate the people inside; they govern themselves.
Heck, we could extend this to allowing anyone from anywhere in the world to immigrate to The Prison, gaining the same "imprisoned and can't leave" status. Or, allow them to leave only by being deported.
After surveying the evidence I am 100% on “team police” wrt crime. The most “equitable” way to address crime is deterrence rather than punishment, and the most effective way to deter crime is with police. If one young man shoots another, then _two_ lives are ruined. If a police officer standing on a street prevents that, it’s a huge win for society. That’s why the progressive prosecutor movement failed: it was being run at the same time as the movement for police abolition. If you want to decrease people’s appetite for incarceration, which would be a win for society, you need to actually increase police presence to counteract the decrease in incapacitation. IMO this is a significant part of why Europe has both less crime and less incarceration: they have a much higher rate of police per capita, and they are stricter about laws we barely enforce like requiring people to pay for public transit.
I would guess the higher amount of guns per capita in the US should mean we should expect the US needs to higher spending on police relative to all European countries. Not just equipment but training - probably hard to train someone to distinguish in a split second between a cell phone and a gun being pulled out?
I'm somewhat confused about American police. Before BLM, there were more than 10 years of articles about how American police is turning into a powerful army in itself, buying armoured vehicles, riot guns etc. These things are expensive, even if bought at an army sale. So apparently American police was very well funded during this period. Shouldn't there have been a huge decrease in crime then? Or funding was there, but spent badly (on shiny toys instead of more officers)?
In almost any organization salaries are much more expensive than equipment since they’re a recurring cost, and it’s often a different pool of money. I would 100% agree that fancy toys don’t help with crime but it’s probably more the fault of the federal government to earmark money for equipment rather than salaries (I think what you’re probably referring to is Homeland Security money that was spent on fancy chembio gear or whatever).
A lot of the military grade equipment and/or snow cone machines were funded through grants from the DHS in the wake of 9/11. Lots of federal funds to buy machine guns and ammo and armored vehicles for "counterterrorism". Not so much county and state funds for the actual day to day running of police operations.
> These things are expensive, even if bought at an army sale. So apparently American police was very well funded during this period. Shouldn't there have been a huge decrease in crime then?
American police spend approximately all police-hours on traffic tickets, rather than solving actual crimes. And this is with homicide closure rates of ~50% and rape ~30% and property crimes ~10%.
It's a bit of math to get there, but my full argument is here:
Bottom line, if you want to increase police funding, you need to legislate at a high level that they CANNOT use the extra funds / police-hours for traffic tickets, but instead must use them on solving actual crimes.
Otherwise, we're just going to get more traffic tickets and similarly laughable closure rates of actual crimes.
Maybe vehicle theft and larceny are substitutes? You know you have two strikes, so you're not about to steal a car, but now you don't have the money you would've sold the car for, so you just have to earn your living the hard way: stuffing your sweatpants with toiletries and walking out of CVS looking like the Michelin Man.
> On the other hand, people in the field keep saying there’s no relationship.
The field doesn't want prison to reduce crime, therefore the research will say it doesn't. The last guy who violated that rule in the general field of criminology was John Lott, and look what happened to him. Lysenkoism requires that the outcome be known *before* the research is started.
> And they’d already gotten caught twice, which suggests they’re either worse at crime than average, or operating in areas with better-than-average police coverage.
Or do a lot more crime than average.
> not being too picky about whether the people they imprison were guilty of any specific crime or just kind of gang-adjacent.
Gang-adjacent, like an auxiliary that can be called to maintain gang control once the "official" gang leaders are in prison? You can't kill a Tree of Heaven by cutting it down, you have to uproot the entire plant. Same for gangs.
I was going to continue in this vein, but then I realized something: No one involved in studying the issue seems to understand how the criminal justice system works. They all seem to think that every defendant gets a fair trial, where all of the charges are put to a jury, who then convict (or not, based on the evidence) and pass the case to the judge for sentencing.
Which, lmao.
What actually happens is that a guy is arrested for, as a demonstrative of the principles involved, armed robbery with a handgun - he knocked over a gas station. Let's assume the charges are as follows:
- Armed robbery (Forcing the clerk to turn over the money)
- Assault with a deadly weapon (Pointing the gun at the clerk)
- Battery (Shoving someone during the robbery)
- Witness Intimidation (Threatening the clerk not to call the cops or he'll come back)
All told, he's facing 60 years all-in. But he knows that if he gets caught he'll be out in no more than 2. How?
1) He's in a district with a Soros-model DA, who drops most of the charges and negotiates a plea deal based exclusively on the robbery, drops reference to the weapon, and agrees to a suspended sentence of 6 months. (A suspended sentence is one that is not served unless the defendant is arrested on a different charge while on probation)
2) He's in a district with a normal DA, who drops the witness intimidation and assault charges, negotiates on the armed robbery and assault, and drops the assault to agree on an 18-month sentence in exchange for saving the DA time and effort.
Plus, judges only see criminals when they are in full suck-up mode, and not during their criminal activity, so a lot of judges end up assuming that criminals are much better than they actually are.
The guy in the first part of the video is apparently honestly pleading his case and then the judge says he's going to jail and the guy, well, you see.
But most judges just see "suck-up" and hand out light sentences, and almost all judges will accept a plea deal.
So we see a Red Queen's Race between state legislators trying to impose meaningful sentences for crimes, and DAs and judges trying to keep the plea-deal machine oiled. As a result, you see a bigger and bigger gap between what sentences could be and what people are actually sentenced because of those pressures.
Which is way tagging the analysis to the passing of "tough on crime" bills is either a complete waste of time or deliberate malfeasance. Just because a bill passes doesn't mean the system changes.
But if passing laws in general doesn't lend evidence to the effect of incarceration, doesn't that exactly cancel out the supposed benefit of passing laws causing less crime?
Because the latter question is what people are using this information to act on. What achievable interventions do you think are effective, given your worldview?
> But if passing laws in general doesn't lend evidence to the effect of incarceration, doesn't that exactly cancel out the supposed benefit of passing laws causing less crime?
My point is that the US data isn't meaningful because the laws are being actively subverted. Soros DAs are the perfect example of this concept on a broader scale - many of them simply refused to prosecute certain offenses, which meant that the law had no impact. That's why law enforcement swear an oath to enforce the law. But Progressives and Socialists see everything as an exercise of pure power, unbound by duty or honor, so they will take the oath and openly violate it the same day.
> What achievable interventions do you think are effective, given your worldview?
Barring plea deals absent some concrete benefit to the state beyond expediency would solve a whole host of problems in the justice system, including this one.
> Barring plea deals absent some concrete benefit to the state beyond expediency would solve a whole host of problems in the justice system, including this one
I still don't get it. You're talking about extrajudicial forces working to subvert the law. Why wouldn't those forces equally subvert barring plea deals by instead calling them dee peals? Isn't that just putting aside all of your hardheaded realism about how things actually work in the real world in exchange for a fantasy of justice?
> You're talking about extrajudicial forces working to subvert the law.
There's nothing extrajudicial about judges and prosecutors.
> Why wouldn't those forces equally subvert barring plea deals by instead calling them dee peals?
Because the law solved the semantics bullshit scam a long time ago. I know it's a favorite play of rationalists to say "Well what if we just change the meaning of words, make up new ones, or decide they don't mean anything at all?" but the law doesn't allow that. Check any state or Federal code. The first section will be a section of definitions. Does your conduct match the definition? Then we don't care what name you give it, it counts.
> Isn't that just putting aside all of your hardheaded realism about how things actually work in the real world in exchange for a fantasy of justice?
No, because I've actually thought about and studied this for longer than you have been aware of the problem. Eliminating plea deals breaks a ton of incentives in the "Speedy Conviction Assembly Line" the modern justice system has devolved to.
> The guy in the first part of the video is apparently honestly pleading his case and then the judge says he's going to jail and the guy, well, you see.
Interesting to see random-guy-in-a-suit beating the hell out of the defendant. He (the defendant) gets restrained by one guy who is obviously security and two more guys who obviously are not.
Why does the audio keep cutting out in the video? It's not to obscure swearing.
>Maybe it’s hard to deter violent crimes (because these are executed in a fit of passion), it’s hard to deter property crimes (because criminals need the money), but it’s easier to deter drug crimes (because people make more rational decisions about whether or not to use drugs)?
My default assumption is that it's easier to *not get caught* for drug crimes if you are being careful, since taking drugs is usually a victimless crime, and if you carefully do it in private at the right times there will be no criminal complaint for police to investigate. So increased caution for the same amount of crimes is especially effective for this category.
Skarbek's book "The Social Order of the Underworld: How Prison Gangs Govern the American Penal System" is a great read.
The key point is that, because most criminals expect to go to prison, they know they will someday be at the mercy of prison gangs who have the power to inflict violence on the inside. This gives prison gangs enormous power outside of the prison walls; non-prison gangs usually pay tribute to a given prison gang.
(This incidentally is what worries me about El Salvador. Having all the gang members in prison right now might be like having guerrillas driven up to the mountains; the state may be ceding its monopoly on violence in a certain part of the country and this could eventually be a problem even if the rest of the country is safe in the meantime.)
Putting a dent in the power of prison gangs would be intrinsically good and probably help with crime on the outside.
Unfortunately solutions to this are probably unpalatable. Removing the gangs' power to commit violence and smuggle drugs would look like making conditions better and safer for prisoners (undesirable to the right), but would require large increases in funding for prisons alongside stricter security (undesirable to the left).
However this is an issue that's nicely orthogonal to "turning the knob" of more/less sentencing, so I figure it's worth mentioning.
Any automatic restriction also risks cutting off positive social connections, and has costs in terms of duplicated facilities, more guards to watch all those separate spaces, etc. Wouldn't stop gangs, either - only need one guy per pod with a shiv and a phone. If you want everybody in the big house to be able to avoid bad influences, better plan would be to just say prisoners can't be put in the same room together without mutual assent - anybody wants to get away from their current cellmate, or a particular crowd at the cafeteria, or anything like that, it happens ASAP, with no questions about why.
I didn't read his book, but I listened to his econtalk episode when it came out, and as I remember his story was that prison gangs had actually made prisons safer on the inside. But of course they have other deleterious effects.
They arose in response to the state being unable to maintain prisoner safety, sort of like warlords or mafiosi bringing back some order when the state is too weak to do it.
If I was a victim of a non violent crime, I would be willing to let the criminal go free if I could be rewarded with some of the money that would otherwise be spent on the trial and the incarceration.
You are not going to be compensated for more than the cost of the crime. For the state to do this would be a very bad idea, we’d effectively be paying criminals to be criminal. You might also be underestimating the emotional distress of being mugged, which can’t be compensated.
I think these things are impossible to measure. I don’t see a contradiction though. Let’s say that instead of imprisoning Bob the mugger we compensate the victims for the price of the items stolen. And let’s say that’s easy to do. Ignoring fraud for a moment, Bob is free to mug as he will, and the State intervenes not to lock him up but to pay the victim off. This becomes a job, not really a criminal activity.
Fraud would happen of course: “I mug you, you get compensated for the phone and I get the phone”.
Retribution is indeed the primary reason for imprisonment or incapacitation, and it rarely is mentioned in these debates. The alternative would be not sending people to prison if they were very very sorry indeed, and could prove it.
When I say the primary aim of the system is retribution that’s an actual description of reality.
Take Norway as an example -largely seen as rehabilitative . While they focus on humane prisons and rehabilitation, crimes like murder can see offenders locked away for up to 21 years, extendable indefinitely if they’re deemed a danger. That’s partly retributive, ensuring the punishment reflects the severity of the crime, but also acts as a deterrent to others.
However deterrence isn’t necessarily effective - read the piece again - and may not work at all for mafia types and hardened criminals for whom going to jail is just part of the lifestyle. If the law doesn’t work to deter, it can work to satisfy justice.
It is true that many humans think retribution is inherently good. But many other humans don't. For example, I don't care about retribution as a terminal goal pretty much at all. Bad things happening to bad people is a negative for me. I want good things to happen to all people. I care about criminal justice only in so far as it helps increase that expectation value.
I don’t think it’s always the primary aim, but the concept is important in many situations. “Retribution" sounds harsh, but most people ("society" or "the public") expect their criminal justice system to include an element of punishment apart from deterrence and incapacity for at least the most violent, horrific, and pain-causing crimes. It's not entirely clear that the desire for retribution in such cases is immoral, such that the criminal justice system should ignore the will of the majority if that were possible. Even very progressive people tend to back away from the most extreme consequences of non-retributive justice when confronted with examples. A professor in a criminal law class liked to walk people through hypotheticals based on real crimes to show that almost everyone expects and even desires some form of retributive punishment for certain classes of crimes. I don't recall all of the examples, but two I remember were someone throwing acid into the face of someone who politely turned down a request for a date, leading to permanent and horrific disfigurement, and someone who tortured and killed small children to "get back" at an ex. Imagine these crimes were committed against you or a close friend. Even if you assume that the perpetrators would never commit another crime again and that nobody would ever hear what punishment the perpetrators received (i.e., no punishment at all), very few people are willing to accept that those committing such crimes should go free and unpunished. Often, people want to fight the hypothetical - that you can never *really know* whether a criminal will offend again, for example, and so retribution is "wrapped up" in other reasons for the sentence.
I wonder are there any useful studies on the effects on corporal punishment like caning/ whipping etc compared to prison time. Historically this was by far the most common form of punishment, receiving support from a wide group like George Washington, Thomas Jefferson, Abraham Lincoln to Theodore Roosevelt. Then it suddenly it disappeared all across Europe in 1940s.
Intuitively a corporal punishment is all deterrence and no in-capacitance and so by your analysis its effect on crime rates should be small (but it’s also much less expensive than prisons). This doesn’t sit right with me though, I speculate there is something in the nature of corporal punishments that makes it a much more effective deterrent than prison.
Also in many cases I would personally prefer corporal punishment as the more human option over prison. If I was caught shop lifting, 30 whips seems like a better deal, than a waste of 6 months to 3 years of my life. With a corporal punishment, you can return to your life quickly, with the hope being that the punishment instilled enough fear in you that you wouldn’t want to get caught doing crime again.
One argument in favor of corporal punishment over prison is that the punishment accrues sooner, so it will be a better deterrent for people with high time preference.
Middle class guys probably would suffer worse, they take fewer hard knocks during everyday life.
I do wonder if it's going to increase crime among masochists. Presumably you wouldn't have hot women doing the caning. Where's that guy Giles who posts here sometimes?
"Small increases in incarceration cause small decreases in crime. Large increases in incarceration cause large increases in crime." I think you mean "large DECREASES in crime" in the second sentence.
"While criminals are in prison, they can’t commit more crimes (except within the prison walls; we ignore these for sake of simplicity)"
I agree ignoring crimes in prisons makes the analysis simpler, but it also puts your thumb on the scale in the pro (additional marginal) incarceration direction.
Note that Roodman also makes this assumption (p.10, Section 2.2.2):
"Because prison crime is so rarely studied, and because crime outside of prison has particular political salience, 'incapacitation; in this review will refer to crime outside of prison."
This seems in tension with identifying Roodman as skewed in an anti-incarceration direction.
Separately, I'd be interested to see an analysis on jails, where people are incarcerated who are still legally presumed innocent and one expects have a higher rate of actual innocence. This is particularly relevant for bail reform debates
"a robber may steal $100 worth of goods, but valuing his crime at $100 in costs ignores the disutility of (eg) living in fear"
But if the robber steals from a store, the store is out its cost of those goods, and has to make it up in its profit on other sales. Net profit per sale depends on the actual goods sold, but it usually ends up being about 1% or so when you count that more expensive goods which give you a higher margin also sell less often.
In any case, the robber gets $100 (retail price) of goods, and maybe sells it on for $10-25. Cost of the goods to the store from the wholesaler might be $50-60. My own experience was with books, and there it was $60.
The store then has to sell $6000 or so worth of goods (retail price) to have a net profit of $50 and be able to rebuy that $50, because it's a straight loss. And the store's cost isn't just cost of goods - there's staff salary, rent or mortgage for the store, etc.
Insurance may pay that $60, but then will make it up in increased premiums, or refuse altogether if it keeps happening. And then the store likely closes.
So even ignoring the living in fear part (and for robbery it is fear, since the crime requires a threat of violence), it's still not just $100 in cost.
None of this cost shows up in government statistics of course, since none of it is the government's loss to bear.
This is like when a person sues to recover money. The court awards the money. Then there's court costs, which the court may or may not award.
Then there's the years of your life it took to sue the other party in the first place. The times you had to be in court instead of doing what you would normally be doing.
Mostly that's just not counted.
It also assumes that anything at all can be recovered from the robber. Which is not usually the case.
And then there's the cost of the loss prevention measures stores turn to in an attempt to stop the bleeding. Locking up more and more of their products, for instance, which leads to some combination of higher staffing needs, higher employee workloads, and poorer customer service.
"So maybe Roodman is right about shorter sentences, and everyone else is right about longer sentences. Going from a month to a year in prison is so disruptive and criminogenic that it risks canceling the benefits of eleven extra months of incapacitation. But going from ten years to eleven years mostly just gives you the incapacitation."
This makes a lot of sense to me. I would guess that the post-prison criminogenic effects of prison are heavily front-loaded i.e. heavily sublinear in the length of the prison term, whereas the incapacitation effect is close to linear in the length of the prison term (unless it's a really long term, so that aging out effects start to be important).
I agree with this. Anecdotally, fear of prison is the biggest reason why I don't commit crimes. Even bad ones like murder - I would rather commit murder and get away with it than go to prison for a murder I didn't commit.
Is that relevant for longer vs shorter sentences though? Like I agree enforcement vs not can have a big effect here. But are you more likely to commit crimes if you have to stay in prison for 6 months vs 1 year?
If not I think it reinforced Scott's point - increase certainty of enforcement first before worrying about longer sentences.
Has anyone looked into prison sentences whose duration inversely correlates with the age of the offender? Criminality (and political activism, and plenty of other things) peak in the late teens/twenties, so maybe focusing sentences which keep offenders in this age group behind bars until they've "mellowed out" would go much farther than comparable sentences for e.g. 50 year olds.
This is sort of the flip side of the push for harsher sentences for repeat offenders, since repeat offending is undoubtedly positively correlated with age (since you've have more time to re-offend).
In your costs/benefits of prison section, shouldn't you include the management/prison costs of the counterfactual prevented crimes as a benefit. I.e. how much police/courts/prison time is used up for an additional 3.9 cases of Larceny etc. It seems like it would be a high cost unless you want to do away with the justice system entirely.
I think you made a minor mistake in this section: “This broadly agrees with our numbers from Sweden, California, and El Salvador above. Small increases in incarceration cause small decreases in crime. Large increases in incarceration cause large increases in crime. If you doubled the incarceration rate, locking up an extra million people, then crime would decrease ~30% at current US margins (maybe less, because you’re shifting the margin and getting diminishing returns).”
Specifically you say “Large increases in incarceration cause large increases in crime” which I think is supposed to be “decreases in crime”
Given my extensive years in crimlaw, I have a lot of thoughts about this. A recurring problem with discourse in this area is that the practical realities involve an intersection point between criminals (a highly irrational subculture) and bureaucrats (a class of people who are given idealistic speeches at conferences and then go back to a job where there's bizarre incentives to say 2+2=5 if it clears their desk briefly). So not only are the decisions of the actors irrational and arbitrary, even identifying WHAT happened is inscrutable to people who don't practice crimlaw. I don't have time to react to everything here, but just a few points on the non-linearity of how criminals see sentences.
"So most deterrence will look more like the Proposition 36 proposal we discussed last month, which increases shoplifting sentences from six months to three years. If we use H&T’s numbers (probably inappropriate since there may be nonlinear effects), we would expect that section of Prop 36 to deter crime by 2%."
Correct about the expected nonlinear effect. I believe the Proposition would be ineffective because practical constraints would force prosecutors to amend charges rather than seek the heightened sentences required. But if somehow the sentences actually came to pass, the criminal class sees this VERY differently. A misdemeanor with a 6 month jail max probably means that at worst you get arrested, are held in some county facility, can't post bond, your lawyer screws around with discovery and motions to make themselves feel like they did something, and then you plead out to time served at some point. A 3 year sentence isn't just about the duration, it has to be served in a prison, which is a very different world than county jail.
Many lower-level criminals who will happily sit in jail for 6 months would push back strongly at having to enter the prison system even if you told them to expect parole at the same 6 month mark. Whether or not it deters the crime, it certainly deters pleading guilty as charged and greatly increases the chance of the defendant snap-accepting any misdemeanor plea bargain offer. Jail vs prison punishments are different in kind, not merely in duration. (There are large metros with jails that behave more like prisons with stable inmate populations serving sentences, as opposed to keeping "sentenced to jail" and "awaiting trial, couldn't post bail" people together, but as the "sentenced" group would be mostly low-level offenders and the "awaiting trial" group would have included more serious offenders, I wouldn't be surprised if these facilities were actually preferable to ordinary jail.)
"Ross 1982 investigated the effects of some widely publicized laws that increased penalties for drunk driving. They found that it decreased drunk driving 66% (!) in the year the law came out, but this effect gradually faded to zero over the next three years. The most likely explanation is that the publicity around the new law got people excited, but nobody really knows the laws around drunk driving anyway (do you know what the range of sentences for DUI in your jurisdiction is?) and so once the excitement of “law is stricter now!” faded from memory, the law had no effect."
Find this strange, as drunk drivers are uniquely deterrable among all criminals. This is because drunk driving is one of the few crimes likely to have a high prevalence among the middle class and important community figures. An esteemed local businessman cannot go to jail, if the DUI has a mandatory jail penalty that is a huge problem. In my state first offense DUIs typically result in probation and some annoying classes before you can drive again, whereas 2nd offenses have mandatory 5 days jail and 3rd offenses are felonies. And what's more unique, MADD actually keeps stats on this stuff and will go after you if you show weakness on DUIs, so judges and prosecutors here will in fact do the labor of trying a case before they are willing to reduce charges on a DUI. You may not know the penalties for DUI in your jurisdiction, but I guarantee you that the people inclined to drive drunk know them. Heavy drinkers all know someone who's gotten a DUI, and who got this lecture from his lawyer, and you will notice in practice altered behavior by prior DUI offenders.
If the deterrence effect lowered that much, my guess would be that the legal system was not imposing the heightened penalties -- a problem that would not show in the data unless you knew A) the NCIC charge code entered by the cop for the arrest itself, B) the charge code on the original felony complaint, and C) the charge code on the disposed case. And even if you did, you're likely to fail at point (A) because -- unless a background check shows enough priors to make it a felony and therefore something he'd seek bail for -- the cop may not have collected enhancement data from the suspect and just figured he'd enter it as a 1st offense generic DUI at the station and then fix it up on the probable cause affidavit. Determining whether people who should get increased penalties as the law is designed are actually facing them is pretty hard.
***************************
There are also threshold points even along a spectrum of years that are given irrational weight during plea negotiations. Numerous repeat offenders in my state were far more likely to accept a deal of 7 yrs than 8, this break point was consistently noted in negotiations. The only explanation we had was that historically C felonies maxed out at 7 and were treated more leniently by parole board than A/B felonies-- but if you had committed a B felony and were subject to the higher range, then the difference between 7 and 8 was no different than the difference between 8 and 9, getting 7 didn't make them treat you as if you were a lower class offender. But they acted as if it mattered to them greatly.
Which is all to say that you can't really ever boil this down to each year deters X, because there are both real and imagined thresholds that affect different social classes and types of offender in differing ways that are both hard to locate in data and defy rational behavioral analysis.
The whole jail/prison thing suggests to me that a remand system (where people awaiting trial are remanded to prison if they aren't granted bail) would have an enormous impact.
This is certainly how the UK works. There are cells in police stations, but they're only used for people being held for questioning (which has the usual Habeas Corpus limits of a few days - once charges are issued, they go to a remand hearing and then either to remand or bail).
Remand prisoners are held separately from convicts in prison, but that's just a case of wings being designated differently - if there are more remand prisoners and fewer convicts, then the same cells get redesignated.
I expect the problem in the US is get another example of too many levels of government: jails are county and prisons are state, yeah?
I've heard that it is often the case that county jails are worse places to be than prisons. Is this a common opinion in your experience? For example that seems to be the consensus as well as common knowledge in this thread:
It depends on the inmate's background and personality type. In prison it's easier to get things, the social environment is more stable, career criminals who've been to prison before usually prefer it over jail. I think the problem with jails is that there's a lot of uncertainty which leads to psychological instability, new arrivals can be disruptive and unstable, trials are still upcoming and people are nervous. It's also usually true that the jail facilities themselves are worse than a prison's, the food is worse, the bunking situation is chaotic, etc. Inmates at prisons get iPads now, with limited function but it beats crowding around the one TV to watch Maury Povich at the jail.
Nevertheless, if you had never been to prison before, it would be a much bigger shock to your system than jail is. If you were middle class or above, and you only were doing a short stint, I think you'd prefer the jail. If you were gonna be there long enough to learn the system and get comfortable, and thought you were capable of having the social IQ to deal with it, prison is materially better. So a lot of it depends on your background and why you ended up there.
Another factor is that it's much easier to hide what you're in for at prison than at jail, because the jail inmates all go over for the same big criminal law dockets as you do. If you're charged with child molestation, even if the lawyer tries to be careful about it (waives reading of charges at arraignment, for example) eventually there will be some legal argument at the bench and the other guys will put it together that you're in for that and everyone at the jail will gang up on you. (Which even if you're the "they deserve it" type, is not what you want happening for people who haven't been convicted of anything yet, and is also a pain in the butt for jailors because they can't just let that happen or there'd be problems, possibly expensive ones.)
> There’s not enough justice system resources (public defenders, courtrooms, etc) to give shoplifters trials, so the prosecutors have no bargaining power, so they accept plea bargains for minimal punishment.
I take the opposite lesson from this, actually: the decision to obligate the state to spend more on trials is the problem. (Or, at least, it hasn't passed a cost-benefit analysis, yet.)
Like, the naive argument for the death penalty is "it's cheaper than keeping them in prison", because you don't have to house or feed corpses. But this isn't actually true in America, because we've made trials for death penalty cases substantially more expensive than trials for prison-in-life cases, and the increased expense of the trial is larger than than the decreased expense on prison. But this is a policy choice! We don't have to have the appeals system we have, or the standards for trials that we have.
Similarly, we simply haven't checked whether we need juries to be twelve people. (Secretly, 14, because you need alternates.)
Smaller juries would have hard-to-measure effects on the wrongful conviction rate, and probably not save much time or expense. Better approach might be to dig into where the procedural bottlenecks are, what sort of motions get used as stalling tactics.
Also, ought to let juries access the official transcript during their deliberations rather than relying on memory. Not directly relevant to speeding things up, but it's widely agreed to be dead stupid, might as well fix that too as long as we're meddling in there anyway.
On a related note: don't let the jury see the trial live. Record the trial, edit out the parts where someone says or does something that has to be stricken from the record, and give the jury the edited recording.
> But here we’re interested in its potential to confound studies. A 20 year old who gets 5 years in prison is released at 25 - still young! - but a 20 year old who gets 10 years in prison is released at 30 - too old to be leaping on rooftops and running from cops. The National Sentencing Commission understands this problem, and matches the experimental and control groups by age at release. But this introduces a new bias - now they’re different ages when they start committing crimes. Might a person who starts crime at 15 be a more disturbed and committed criminal than one who starts at 20? Seems plausible. I think this might be responsible for a lot of the seemingly positive effect of sentences > 5 years.
This is an issue that comes up in several contexts. For example, what happens if you try to hold your child back from entering school if their birthday is near the cutoff, so they're the oldest in their class instead of the youngest? No matter what identification strategy you use, true age will be a confounder to measuring the effect of relative age compared to peers. I've been told that this means the underlying question ("what is the impact of holding your child back?") is fundamentally unanswerable, although I think the real conclusion is "increased age is one of the effects, which you should account for when determining the effects." So in this case, the fact that a longer sentence means the prisoner is older when they're released is just part of the true effect on recidivism, and it might not be possible to separate out how much effect from a longer sentence is due to the direct effect of being in prison longer and how much is because the person is older upon release, but this doesn't really matter for a CBA of longer sentences.
Feels like a re-run of the homelessness essay. Prison abolitionist activists intentionally drive up the cost/friction in arresting/imprisoning people. Research shows massive value in imprisoning people, but wait, its so expensive to imprison them, so even with the significant reduction in crime it is not 'cost efficient'.
The real solution is just, making it efficient again. This also applies to the justice system more broadly. It actually shouldn't cost the US government tens of thousands of dollars to arrest and charge somebody on their 20th count of shop lifting.
How do you propose we do that? The main costs are providing a lawyer (constitutionally required if jail is a possibility), providing a jury trial (required in many states if jail is a possibility, and required everywhere for a sentence over 6 months) and providing constitutionally-required procedural protections.
Yup totally, repealing the sixth amendment is literally the only way to change any aspect of the system and it is not politically likely, totally agree, its actually impossible to change, wow.
Why would you need to repeal it? The current interpretation of the "right to have the assistance of counsel" is quite obviously unrelated to the meaning that was written down; all you have to do is say "let's use the original meaning of this law".
Examine and recompile the existing procedures - in particular, statistical analysis of how they're typically used and abused - with an eye toward providing the same level of protection for actual innocents while reducing unnecessary delays.
Make "one night in jail and a GPS tracking anklet for a year" the default sentence for nuisance-level offenses, with automatic logs of the tracker's position, so repeat offenders with no fixed address are easy to find for follow-up questions.
1. I’m not sure “study the problem and figure out the best solution” is actually much of a proposal. And can you even sketch out how you could get the data you’d need to perform that “statistical analysis”?
2. I’m not sure how your GPS proposal offers much help in incapacitating people for minor offenses, and it also seems like it doesn’t have much bearing on the topic here.
Lot of court proceedings are a matter of public record. That's most of what's needed, it's just not organized for the right sort of quick reference. Grab a prosecutor and a public defender, have them lay out types of cases as game-theory decision graphs, plug in empirical values for the various branches, use that to run cost-benefit analyses on specific procedural safeguards.
"Better that a thousand guilty men go free than one innocent wrongly punished," right? But at larger ratios, it eventually stops being better, and we do seem to have a non-negligible number of innocents being wrongly punished anyway. Whose job is it to find Pareto improvements on that curve? Did somebody 250 years ago write "temporary I hope hope hope" in the margin of a sketchy legislative proposal that's since come to be treated more like holy scripture?
Regarding the second point, massive routine use of those could reduce the cost / friction of arresting people (a prerequisite to most other punishment) by making them easier to physically find. Instead of putting out a passive bench warrant and waiting for the person of interest to catch an officer's attention for some other reason, track them down and sort it out same-day.
yes, it should be the more strikes you have the faster you can get some sanction (GPS ankle monitor, mandatory check-ins, community service, jail, prison, exiled to Ohio, death by nitrogen)
it's completely ridiculous how many discontinuities are in these processes
"the longer you’re in prison, the longer you have to be scared straight, or find God, or get rehabilitated." I think this is missing one factor. If a criminal has anything like the same life arc as I do at 77, then the longer time in prison means when you get out, you are too damned tired all the time to do anything, much less commit violent crime.
I visited El Salvador early this year and think the visible presence of the military throughout the country is probably also deterring violent crime. In San Salvador, towns, and even along the highway it's a common sight to see groups of 5 - 20 armed, camouflaged troops standing or walking around. My tour guide told me the troops were mostly looking for gang members who evaded the recent mass arrests.
Note that that the much-publicised mass incarceration began in 2022, but the decline in crime (which I agree looks pretty miraculous) began in 2015, and if anything it slowed after 2022, so we can be pretty confident that the two aren't connected.
To me, this all (which I don't have methodological issues with) tells me that "defund the police"[1] is exactly wrong-headed. Instead, we should have *more* police presence (increasing certainty of intervention).
It seems that most people want to have *outcomes*, which are some complex function of *both* severity of punishment and certainty of punishment. So if certainty goes down, people start insisting on more severity. Even though that's less effective. It's sorta like how behavior is a function of both intrinsic motivation and extrinsic motivation (threats of punishment, external enforcement). But people tend to seek "homeostasis" to some degree, decreasing intrinsic motivation as extrinsic motivation gets bigger. But then if you relax the extrinsic pressure...intrinsic does not increase. Leading to reduced overall behavior.
[1] whether that's *actually* defunding the police partially or entirely or just shifting resources to non-punitive measures like social workers.
I was all set to write a comment lecturing about how this deterrent and incapacitation effects measures were happening at the margin, but then I scrolled down and you had a whole section explaining this, so muy bueno.
Maybe what we need to do is reduce the costs of incarceration, then. Why does it need to cost so much to keep someone in prison? Bring back 19th century criminal justice: prison ships, stockades, flogging, and sentencing people to transportation (this time to, I dunno...the Aleutians?)
Punishment is the primary stated purpose of incarceration, and directly bears on the societal question of how much incarceration there should be.
This analysis of marginal costs and benefits of incarceration addresses the question it poses, "Do longer prison sentences reduce crime?" But the answer, and the broader economic analysis at the end would, I think, be viewed by many or most members of society through this lens:
1. What is the appropriate punishment for the crime? (An economist would include Roodman's costs to the prisoner in answering this question.)
2. Can society afford the cost of such punishment?
If there is a net benefit to society, the answer is obviously yes. Adding to the severity of punishment until the marginal cost/benefit is zero is morally dubious.
If there is a net cost, then at some point the cost might be large enough to merit compromising on the appropriate punishment. But that's still a compromise between two competing concerns, and the balancing point will have a nonzero and possibly substantial net cost to society.
That's why prison guards reject the evidence-based approaches: they're there to enforce punishment, which is omitted from the evidence-based calculus.
P.S. I presently don't seem to have the ability to type replies of more than a few lines on a web browser without the comment window expanding enough that the "Post" button gets hidden beneath the next comment below and can't be clicked. So if anybody chooses to engage in conversation with me on this post, my responses will necessarily be short and pithy.
Substack breaks their comment section in new and innovative ways every iteration. That problem is on the web. On the app - where I now am - I can’t select text to answer as doing that collapses the comment. Both are new bugs.
Exactly! People think they're enlightened and humanitarian when they dismiss punishment in favor of deterrence, but if you are only concerned with deterrence then the best strategy for deterrence may be out of proportion to the crime. Locking up all shoplifters for 20 years would deter a lot of crime (through incapacatance), yet it would be an unjust punishment. They would be punished more than they deserved!
Also, "enlightened and humanitarian" applies when talking about future jobs for convicted felons. Through the punishment lens, if you've "paid your price to society", you're ready to work again. Neat and transactional. Through the deterrence lens, you get to worry about recidivism rates and all that, and refuse to hire.
(I get to write somewhat longer replies if I hold my phone sideways...)
I think the section on the cost/benefit of imprisonment leads to odd conclusions the way it is structured. If it costs $15k/yr to imprison someone in Alabama but $120k/yr in California... everyone from jaywalkers on up should be imprisoned in Alabama and only the murderers and rapists are worth it in California. Is the implication that the value to society of imprisoning dangerous criminals directly relates to the cost of running the prison system in a given state? Further, if the cost of prison could be radically lowered, say a few thousand per year, is it then worthwhile to imprison people for minor infractions like speeding tickets?
Also, there has to be some fun utilitarian calculation where it isn't ethical to imprison anyone because the cost of suffering in prison outweighs the suffering of the victims of the convicts.
Almost 50 thousand people a year are killed in automobile accidents every year in the US. I would argue that it is not categorically wrong to imprison people for traffic violations.
You mentioned some potential explanations as to why homicides in El Salvador fell by 95%, but you missed my preferred explanation: stopping organized crime provides a much greater benefit in reducing crime than stopping individuals do.
Like, if you arrest a single murderer, or a single shoplifter, you haven't really reduced the chances of another, unconnected person committing those same crimes. However, if you arrest an entire gang, those gang members no longer have the power and influence to commit crimes like they did before, so it actually does prevent future criminals from joining those gains and committing more crimes.
So basically, I think that the situation in El Salvador was relatively unique, as their society was being terrorized by organized crime. If you quadrupled the incarceration rates in America, I feel as though it would be extremely unlikely that homicide rates would see the same drop.
> People first organised higher order societies in order to protect against barbarian marauders.
The theory is not uncommon in studies of the near east (same region as the middle east, but earlier in time) that people first organized in order to handle temple construction and temple administration.
However, I vaguely recall reading about the discovery of a very large prehistoric battlefield where analysis of the remains showed that they were drawn from a very wide geographic area, implying that in fact people were organized on large scales long before we had previously imagined was true.
On "after effects," skipping all the way to length of incarceration misses a lot of what's going on for first time offenders. As your own examples note, many of the problems making it harder to make an honest living accrue from being a felon as such, and not from jail time. I'll add on that merely being *arrested* triggers a similar cascade of consequences. Your average person will be fired from their job very quickly after arrest -- either because of the arrest itself or just because it leads to missing work. And an arrest on its own, without a conviction, is sufficient to make it hard to get hired on in a lot of jobs.
But this is all specific to first-time offenders; going from someone presumed to be a law-abiding citizen to being a felon changes everything. An incremental felony after that doesn't change much. From your description, it seems like the studies are blurring this together? I'd be interested to see evidence focused specifically on first time offenders.
My guess would be that there's a large, negative effect of your first felony (all else equal) but not much effect at all of subsequent felonies. In that case, there's a very different calculus for the incapacitation vs. recidivism tradeoff for different, identifiable groups.
It's amusing that there isn't a big organized push from the "social justice" crowd to make discrimination against the arrested/convicted to be illegal. I suppose that pretending their client minorities are saintly is more expedient than actually trying to improve their lot.
You're missing the deeper issue - they'd be shooting themselves in the foot, since aggressively discriminating against people who've done certain sorts of crimes (or things they wish were crimes) is itself the social justice crowd's core business model.
Fantastic article. I’m curious about any evidence to support more benefit to ankle monitor tracking over incarceration for nonviolent criminals? Seems like it could be cost effective and potentially less recidivism since their life wouldn’t be as disrupted.
Thanks for posting this – it was a really interesting read. I have a lot of thoughts, some of which I’ll try to share:
1. Crime data is very, very bad, which makes drawing conclusions based on it even more difficult than in other areas. The two main sources of data (in the U.S.) are the Uniform Crime Reports (a collection of data submitted to the FBI from police departments) and the National Crime Victimization Survey (where DOJ researchers ask people what crimes they were the victims of). Obviously we’d expect the numbers to be different since they’re measuring related but different things, but they’re not even close. For instance, in 2019 (this was the last year before the FBI switched from issuing actual reports to using a “data viewer” that I find annoying to use, otherwise it’s arbitrary), the UCR lists about 1.2 milliion violent crimes, while the NCVS lists 5.8 million (and obviously, that doesn’t include murder!). The conventional wisdom is that vehicle theft data is unusually reliable, because even victims who otherwise don’t care usually need to make a police report for insurance purposes. But the UCR has 721,885 vehicle thefts, while the NCVS lists 495,670. With that kind of disparity, it’s hard to say we’ve even figured out how to tell how much crime there really with any real certainty. That doesn’t make empirical analysis useless, but I do think it needs to be heavily discounted.
2. The three strikes laws are designed to increase punishments for what would otherwise be relatively minor crimes, so intuitively it’s not necessarily surprising that the deterrent effect would be more significant on those crimes than more serious ones, which would typically carry a significant punishment either way.
As far as European data is concerned, I would expect that their ability to apply criminal punishment more quickly would also have a meaningful effect on deterrence (which would carry over into the aftereffects part as well, I suppose).
3. “(How much do we care about drug crime? Tabarrok says somewhat, because drugs can be a “gateway crime” that gets people into the criminal lifestyle. Roodman says not much. It’s either personal use, which is mostly victimless, at least compared to other crimes).”
Overlooked, I think, by people who aren’t directly involved working in or otherwise seeing the criminal justice system (I certainly didn’t know it until I got involved), is the extent to which virtually all criminal activity except for sex crimes, domestic violence, and certain complex financial fraud is motivated by drug addiction. People don’t shoplift, steal cars, break into houses, or rob liquor stores because they want merchandise, cars, personal property, or money for its own sake: they do it so they can use the stuff they steal to get drugs. (Yes, it’s not literally everyone. But if we could eliminate all crimes that weren’t motivated by drugs, we’d be in pretty good shape.) So even if you don’t care about people abusing drugs per se, that abuse leads directly to the crimes that you almost certainly do care about, and there’s a real public interest in stopping them. (That’s not to say that our current way to serve that interest, of course.)
With respect to dealers, my impression (from seeing data, but I don’t have it in front of me) is that drug use in America is meaningfully down since we began aggressively prosecuting traffickers ~35-40 years ago. And intuitively, it seems hard to believe that drugs would be harder to get if they were being sold by Phillip Morris or Pfizer or whatever, or that you wouldn’t see more people using them if they were available that way. But again, I don’t have hard numbers to back this up (could be another interesting project!).
4. I’ve read a lot of the “aftereffects” literature trying to correlate prison term length to negative post-release outcomes, and I’ve yet to see one that was very persuasive in making the case for causation. Imagine two defendants with similar (or identical!) criminal records who are both charged with the lowest degree of assault – but the first barely injured the victim at all, while the second injured them so severely as to almost cross the line to the higher-level offense. I haven’t seen a study set up to capture that difference (much less control for it) – but that is, in my experience, precisely the kind of things that tends to lead to a higher sentence, and it’s at least plausible that the person who would do something like that would also be a higher risk of reoffending, independent of the sentence length.
5. I think you’re substantially underestimating both the cost of crime, and the frequency per criminal (even if you ignore drugs – after all, most drug dealers and users commit drug crimes weekly if not daily).
The NCVS for the last few years has had ~18 million violent and property crimes a year. If we estimate (based on inmate numbers) something like 2 million people in the criminal population, that puts 7 as the low end, not the high end.
On the other side, court-awarded restitution is a terrible measure of the actual cost of crime, even in purely economic terms, and certainly if you’re interested in measuring utility (which I assume we are). Roodman puts it pretty well himself: “This approach has generally failed to include intangible costs of crime for communities, as distinct from thevictims and perhaps their families. Almost everyone in a neighborhood feels some harm, for example, when assaults and robberies become more common.”
Excellent post thank you. After reading this I’m convinced that our main priority should be hiring a lot more police. The stats are clear, it makes sense, and you see it in the relative policing and crime rates of the US and Europe . Also it’s hard to cut through all of the statistics and adjustments and variations but it seems clear to me that violent repeat offenders should be getting very long prison sentences for incapacitation and so they age out of violence. Commit multiple violent crimes and you just give up the right to live free in society, we just shouldn’t tolerate that. I think there’s an underestimation of the trauma of shootings and armed robbery/carjacking and what it does to the psyche of people that live in environments where such things are common and the costs are higher than estimated.
Ok, so longer prison sentences and increased incarceration and three-strikes laws do decrease crime -- primarily because they remove the criminal from the environment where he can commit crimes. The problem is that for this method to work, we'd need to drastically increase the number of prisons, and doing so is prohibitively expensive. Thus, here is my Modest Proposal (tm): we institute a three-strikes law, where Strike 3 is summary execution. Now we don't need nearly as many prisons, and we get the maximum possible incapacitation and minimum possible aftereffects. What could possibly go wrong ?
I would object on the grounds that execution is a punishment that is out of proportion to the crime: but I have been told that punishment is a barbaric goal for a justice system, and that we should focus on deterrence instead. Your proposal passes the deterrence requirement with flying colors!
If we substitute execution with citizen revokement I think it may actually be a workable idea. Although making someone having no citizenship has already been banned by UN.
"Here's the deal. Option A, you get to die, maybe of old age, in a fancy concrete box with some other very unpleasant people. Option B, we file some paperwork at an embassy, in a week you get a Mongolian passport with special notes saying you're not welcome here, or in Russia or China. Then you get set up with a cargo-ship job to start off your new life, and if you set foot in this country again we'll shoot you, but otherwise it's all somebody else's problem. Which do you prefer?"
See discussion above on how California's Three Strikes Law decreased crime much less than hoped. I don't see why this would be any better - in fact, it would probably be worse, given that resistance to executing people for minor crimes would limit it to the most severe criminals.
The marginal prisoner in Massachusetts may be a much badder dude than the marginal prisoner in Louisiana. But here is anecdotal reason to believe he is not:
Last week in New Orleans, there were three mass shootings. Two of them happened at the same second line in New Orleans East, 45 minutes apart. In two separate incidents, a gunman opened fire at a large outdoor gathering. In total, two people died, and ten were wounded. This is not national news. No arrests have been made.
The third shooting took place in the French Quarter four days later. One of the three shooters was caught immediately. Turns out, he had already served seven years in prison for armed robbery. Last year, he was arrested again and charged with domestic battery, child endangerment, and possession of a firearm by a convicted felon. A plea deal got him out on parole, the terms of which he violated near daily according to his ankle monitor's logs. He was wearing the ankle monitor at the time he opened fire on a French Quarter street, wounding three people and killing another.
For added spice: at a second line the weekend after all this, NOPD cracked down and arrested seven people for illegally carrying firearms. One had (allegedly) brought both his gun and his drugs to the parade. One was a felon (allegedly) caught with a stolen firearm.
It appears that Louisianans walk around committing felonies like all the time, and it's totally normal, and even after you're convicted of multiple violent crimes, you can keep violating your parole and it's no biggie until you commit a mass shooting.
Really excellent post. I loved this. The only downside to crime posts: the inevitable authoritarian halfwits in the comments trying to outdo each other.
"No one wants to give life sentences to serial shoplifters more then I do!"
"Shut up, Carl. I hate shoplifters MORE. That's why I'm going to imply they should be killed, and then when somebody objects, I'm going to act innocent and defensive because I used weasel words and didn't explicitly say what I meant!"
Even if some values fundamentally differ, presumably an acceptable compromise can still usually be reached? That's the premise of liberalism anyway, and it doesn't seem like there's anything better to replace it with yet.
These people are entirely real. The vigilantists and extremists have made their presence very well known in these comments, right on schedule. Perhaps you ought to read further. I feel no need to hold your hand by screenshotting the offenders. The comments are right there for you to read, if you are so inclined.
Fascinatingly, Joseph Smith's 1844 presidential campaign platform included prison abolishment and using the buildings as "seminaries of learning". He wrote,
"Petition your State legislatures to pardon every convict in their several penitentiaries, blessing them as they go, and saying to them, in the name of the Lord, Go thy way, and sin no more. Advise your legislators, when they make laws for larceny, burglary, or any felony, to make the penalty applicable to work upon roads, public works, or any place where the culprit can be taught more wisdom and more virtue, and become more enlightened."
The new DA from LA in an interview cited some truly ridiculous level of benefit of reintegration programs for former prisoners so that may also be part of a comprehensive solution: More cops, more courts, more reintegration, and with those things the overall incarceration rate will naturally go down over time as crimes are prevented in the first place. Someone with experience in the field told me there are some dumb problems with hiring cops as well, like requiring everyone who works in law enforcement start as a beat cop, which creates unnecessary disincentives for people who are physically unsuited to working as beat cops or overqualified for the position.
I've heard of the "detective fast track hiring" idea before, and while I understand the appeal, I don't think it's a good one. Increasing the segregation between investigation and patrol is likely to result in problems for both.
What's absurd is that it's never even tried. If it were tried there would be data on what its effects are. Instead there's this religion of how things must be done even though it has obvious downsides.
Because trying and failing means more dead people, unsolved crimes, increased crime, disorder, and public dissatisfaction. Trying and failing, when a lot of other people within the profession share my skepticism, means heads roll both literally and figuratively.
A patrol officer tends to be a jack of all trades-conducting investigations, tactics, traffic control, surveillance, gathering intel, people skills, interview and interrogation, identifying and gathering evidence, warrant service, obtaining warrants, presenting cases and providing testimony in court, etc. You tend to find out what you're drawn to and/or good at as a part of the job, and lean in that direction, both officially and unofficially. Got a passion for arresting DUI's? Maybe get transferred to the DUI task force, or get some extra training, equipment, or assignments for DUI investigations. Enjoy finding wanted people? Get assigned to the Marshal Task Force, get some extra tactical and intel training, maybe get a K9. Find yourself enjoying everything surrounding investigations? Become a detective.
The problem with skipping all of that is threefold. First, people don't actually know whether they would be interested in that side of the job until they have some exposure to the reality. Second, all of these skills need to be practiced to become proficient. It's why so many jobs still use some form of apprenticeship. If I make a mistake on a shoplifting call, that's a lesson learned; a mistake that tanks a homicide investigation is the wrong time to learn that lesson. And third, the type of person who claims they could be a detective but don't go into police work because they would have to spend time as a patrol officer would not be a good detective. Sometimes being a detective means sifting through 40 GB of porn and three thousand text messages from someone's phone. Sometimes it means cracking jokes with a child molester. Often, it means being woken up at 3 in the morning because someone is dead or dying. Most police officers see the person who wants to be a TV detective but not a patrol officer in the same way as veterans look at someone who says they didn't enlist because they "would punch a drill instructor in the face."
You also benefit from the connections between patrol officers and detectives. Detectives need to be fluent in "cop" in order to properly comprehend police reports. Patrol officers need to be aware of evidence and investigation procedures in order to secure a crime scene, perform initial interviews, and write the initial report. Detectives then need to know what knowledge patrol will and will not have access to in order to reach out for further information. Identifying suspects from property crimes based on a picture, having a working knowledge of aliases or pseudonyms, or knowing which houses in an area are infamous for which events. This requires people move back and forth between patrol and investigation.
Then, you have the need for detectives to be functional as police officers. Those other skills, the tactical awareness, the basic understanding of how to act on patrol, are learned through experience more than they are taught. You can't invite a bunch of detectives onto a crime scene without the basic understanding of how to act on patrol without a lot of babysitting. There's also the flexibility requirement involved. Think of how Marines say "every marine a rifleman." For a lot of departments, the ability to mobilize every available resource in an oh-shit scenario-ranging from natural disasters, to critical incidents, to mass rioting-requires every (or at least almost every) sworn officer to be proficient in response to those scenarios.
In short, it's not a "religion of how things must be done," it's a serious of best practices as prescribed by decades of practical experience. To relate this to the engineering world, OceanGate gave us plenty of data on what corners you shouldn't cut in submersible design; that only cost the world five deaths. Crowdstrike taught us how not to release a software patch, and we still don't know exactly how much it cost, in money or lives.
Automatically excluding qualified candidates from detective jobs means more dead people, unsolved crimes, increased crime, disorder, and public dissatisfaction. Certainly the job of policing is important, but that's not a reason to shut down discussion about how to make it better, especially when your argument is based on tradition and intuition instead of data and there's a compelling counterargument.
At the risk of sounding perverse, I'm curious about other potential deterrents. How effective is say, Sharia law, at deterring violent crime or petty theft? I imagine that robbing a convenience store becomes very un-attractive if there's a good chance I'll lose my right hand as a result (beyond the pain, the social exile and bodily incapacitation don't sound like a good deal) . There must be research on the effectiveness of capital punishment as a deterrent, no? I imagine that's the best proxy we'll get for some of the more extreme forms of justice.
It's still a pretty poor proxy. "All things being equal" it would work, but in a lot of axis there's pretty sizable difference between countries with Sharia law or similar (including Europe not too long ago). Socioeconomic and cultural aspects aside, as has been pointed out, the certainty of getting caught for instance seems to matter more than the harshness of the punishments itself. And I personally suspect it's not the harshness in itself but emotional immediacy (which, if true, might imply corporal punishments would have effects that longer prison sentences wouldn't).
I think the better proxy would be 19th century when Europe/North America stopped using corporal punishments, especially if there are countries in which the punishments were common until they were banned (c.f. countries in which corporal punishments were banned because public opinion had already turned against them and the practice had already largely ceased).
Of course if one was certain that they would be caught, and lucid enough to care, they wouldn't commit the crime at all (given a severe enough punishment). It's just not clear to me that many of these shoplifters, for example, are worried about getting caught. It appears habitual for a large number of them, and they have not yet encountered a drastic enough punishment to stop doing it. They just don't care about being caught (or at least not enough). I think the majority of violent crimes are similar in that they are crimes of passion (while not habitual, not rationalizable either under this paradigm of being certain of being caught).
I like the point on "emotional immediacy" though. The ancient Greeks used exile as an alternative to death and it appears to have been deemed a worse fate. Probably because it was also a death sentence for the vast majority, but also because of the more potent emotional effect.
The severity of the punishment does seem to dissuade crime in wealthy eastern states like Singapore and Qatar, those are potential proxies if the data exists.
They're not worried about being caught because they've seen other people do it and not be meaningfully caught - it's not that they don't care at all, just a more-or-less rational assessment that the odds seem low enough to be acceptable. If they saw it happen reliably, they'd start worrying about it happening to them.
I don't fully buy this. Sure you can probably get away with petty theft without being caught on any independent offense. But how do we explain the presence of serial offenders -- ones that offend, get caught, receive a slap on the wrist, are released, and repeat? These serial offenders are disproportionately represented in the statistics. In these cases, the "rational" assessment is probably something like "when I've been caught before I received X, X isn't so bad". I don't think there's anything more sophisticated going on in terms of expected value analysis. If a thief was truly rational, then the way to dissuade them is to make it so (catch rate) * X outweighs (1-catch rate) * Y, where Y is the value of the stolen goods.
Note that the "Bloody Code" (execution for lots of non-murder crimes) was passed in England when urbanisation drove a huge increase in violent crime, and didn't much lower crime rates (executions were mostly committed to transportation to Australia for crimes that didn't kill anyone) and was then abolished accompanied by a massive drop in the crime rate - which happened at the same time as the introduction of the Metropolitan Police, who started catching far more criminals.
Interesting point! I don't disagree that the catch rate of criminals is a significant parameter. Increasing the catch rate when it's near zero, is probably the single most valuable thing a society can do to dissuade crime. The value on the margins seems slim however, how much more bang for your buck do you get at the 1,000,000th police officer or resource poured into increasing the catch rate of criminals? At that point you have to play around with other parameters, if you want to further dissuade crime.
Absolutely. The question is "Which margin is the cheapest to move?" Which is the sort of question that these giant meta-analyses are much better at answering than the sorts of argument we're engaging in in the comments.
Great article on an important topic. A judge once said (paraphrasing here) “There’s only two reasons to send somebody to prison. Either we’re scared of ‘em or we’re mad at ‘em.”
I am a felony prosecutor in a major US metro area—I should probably be prosecuting crimes right now rather than writing this—but I wanted to add two thoughts to the conversation. First, that I think that folksy judge was right. Retribution, which only got a bullet point in this article, shouldn’t been seen as seedy or as a somehow illegitimate end of the criminal justice system; it is an essential part of the glue that holds society together. A community where people don’t expect criminal-level misbehavior to be fairly and reliably (albeit not perfectly) punished is a lower trust community, with more brazen criminality, and just generally a worse place to live.
Second, it is quite correct to hold that the next marginal anti-crime dollar would be better spent on policing than on incarceration. However, that also isn’t an easy solution. Before you can actually do that in a meaningful way, you need to address what has become a generational challenge in much of the country: staffing police departments. The job has always been dangerous and (comparatively) poorly paid. Add in the decline in public esteem for the job—particularly post-George Floyd—and the rate limiting bottleneck to crime reduction in most metropolitan areas has become recruiting enough talented people willing to do that job. We just don’t have enough and you can’t manufacture them with a 5% pay raise or some other token effort.
Actually, one last thought on the article: deterrence has a qualitative aspect that is often overlooked or at least understated.
How about merging police unions and teacher's unions, forking over a set fraction of land-value-tax revenue from the region they're responsible for, then giving the union broad latitude how to handle quantity vs. quality tradeoffs of salaries within that budget, with outside oversight mostly limited to checking that announced and implemented policies match, and removing the parties responsible when there's some egregious misconduct?
How about it? Or, more to the point, can you elaborate on why you think that would help address the issues that TheAnswerIsAWall was mentioning? It strikes me as orthogonal at best.
Teachers and police officers have societal functions which often overlap: being the officially-designated "adult in the room" around a lot of people with poor impulse control, teaching those people the lessons they need in order to become functional members of society. Both groups are, in the US at least, understaffed, underpaid, overburdened with useless paperwork, and widely despised. Teachers are often hindered in their ability to help the best students by lack of authority to remove the most disruptive ones; wouldn't surprise me if cops also have major problems they could solve more efficiently by handing out academic exercises.
Whole point of having a trade union in the first place is for people doing similar work to get together, pool resources, and push for their collective class interests, rather than competing - dumping problems on each other - to mutual detriment. People talk about a "school-to-prison pipeline," which sounds to me like teachers sending off problem students to be dealt with by the cops, while "ignorance is no excuse" is law enforcement trying to shift blame back the other way. Get 'em all in the same room together, with a simple incentive structure that rewards overall success at making the community a place where more people want to live (rather than bureaucratic ass-covering), hopefully they'll hammer out better solutions.
Great post! Coincidentally, it agrees with my priors. :)
Even with footnote 3) I was chafing at the conflation of arrest and conviction for the majority of the piece - the philosophical distinction alone is large, but you addressed that the critical component is the process and expense involved with actually securing that conviction. My understanding is that's where most of the resource constrains show up, and was ultimately satisfied by that treatment here.
I was disappointed that the deterrence effects of certainty and swiftness were explicitly outside the scope of this post. That's a reasonable approach to the topic for now, but what I have seen agrees with the later aside that certainty at least is a major driver of deterrence effects. It's also even harder to study than severity, so understandably tough sell. Swiftness might be worth a look though, anyone happen to have a link to good research?
"But Italy has a lower incarceration rate than the US, potentially meaning that the average criminal they incarcerate is worse"
Maybe they just have fewer criminals?
It seems like you're assuming there's a fixed amount of criminality available in society. (I vaguely remember you mentioning some historical precedent involving Quakers that supported this on SSC a long time ago.)
Even if you grant that there's a fixed amount behavior that society can criminalize, that amount might vary between societies. It may be that Italy just has fewer criminals.
>Might a person who starts crime at 15 be a more disturbed and committed criminal than one who starts at 20? Seems plausible. I think this might be responsible for a lot of the seemingly positive effect of sentences > 5 years.
Isnt it the opposite? If theyre the same age at release, then those with the long sentence were imprisoned younger, and so started crime younger, so this should make the long sentences look worse.
>My strategy was to use the pro-incarceration and anti-incarceration reviews to determine a sort of “Overton window” of what smart people thought they could get away with claiming. Then, if the window was very wide and I couldn’t figure out where I landed, I used the apparently-unbiased Nagin as a tiebreaker.
My thought on reading this was that the non-Nagin studies will contribute very little to the final conclusion, and it seems that was right.
Great piece. My left-wing brain is puzzled by one seemingly glaring omission though: someone help me understand the focus on recidivism as the primary "aftereffect" worth exploring. Why should I care about recidivism in particular? I care about crime rates. My only question is "do crime rates go up or down with increased prison sentencing?". There are many narratives which suggest that crime rates would go up with increased prison sentencing. For example, "destabilization of communities with high incarceration rates means kids growing up poorer and without fathers which means increased propensity to grow up and become criminals". This could be true even if every ex-convict became a saint upon release. This seems a critical piece of the analysis of the value of prison, yet one that was entirely left out. Perhaps it just wasn't in Scott's scope?
The El Salvador data, at least, addresses this, by showing that murder rates plummeted after they increased incarceration. Of course, we'll have to see if these results persist over a generation or if this is just a band-aid solution.
If every ex-con became a saint upon release, presumably some of those saints would step up to care for the children of the not-yet-saints who were behind bars at any given time.
> For example, "destabilization of communities with high incarceration rates means kids growing up poorer and without fathers which means increased propensity to grow up and become criminals".
That's not a good example; the relationship between being raised without a father and criminality is completely confounded by the fact that those missing fathers are mostly criminals.
If you separate out children who are raised without fathers because the father chose to leave or was jailed, and children who are raised without fathers because the father died, the boost to criminality is entirely absent from the second group.
That’s very interesting, thank you for sharing! Do you have a source for this? I assume it’s harder to prove that in a counterfactual world where the criminal parent is present (having never been prosecuted), the kid has the same likelihood of becoming a criminal.
That said, when you call it a bad example, it seems you misunderstand me, because it wasn’t meant to be obviously correct, merely plausible, to illustrate that a post like this should address a broader idea of aftereffects, like you just did.
It seems odd to me that turning innocent people wrongly imprisoned into criminals isn't taken into account at all in the "Aftereffects" discussion. As Scott says later, anywhere from 2-5% of people in prison are actually innocent, and the recidivism rate is about 80%, so prison is turning about 1-4% of it's population from innocent law-followers to criminal law-breakers. It's not huge numbers, but at the margins we're considering, it seems like it's at least worth talking about when considering criminogenic effects?
I agree that the actual recidivism rate will be different for the guilty vs. the innocent, and I actually started to include a note in the comment about that, but when it got longer than the actual comment I decided to just cut it. In the end, the same forces that prevent most ex-convicts from living a normal life will apply to both groups, and I have no way of knowing the actual rate, so 80% will have to do.
On the other hand, the claim that "large fraction of them are guilty of other crimes" does seem completely unjustified. What are you basing that on? When you look at actual wrongful convictions that have been overturned, most of them are just people in the wrong place at the wrong time. I'm sure some of them are guilty of other crimes because that's basically true of literally any group, but I can't see any reason to believe it's a "large fraction".
"When you look at actual wrongful convictions that have been overturned"
How much of this confounded by the ones that are publicised and the ones where the Innocence Project and the like are investigating are the ones where it's "wrong place, wrong time" and the ones where the police picked up a usual suspect who had done ten burglaries that week but not the one they convicted him for are not really of interest?
The idea that the more people incarcerated already the less there is to gain from incarcerating more people is obvious enough, but Scott’s clever use of the extra guy being arrested in a country with 50% of the population already in jail vs the first guy in a country which has arrested nobody is a clever debating tactic, which I will use. Also I will use the word crimogenic as if I’ve always known it.
But two points.
1) What’s missing from this deterrence story is the heretofore honest citizen who isn’t registered as criminal yet. Jailing nobody would deter none of us from speeding, drinking, drink driving, or shop lifting (etc) and even murder. Why not kill the wife if she’s pissing you off? I’d be inclined to grab a few iPads myself if all it entailed was a nice smile from the shop manager, and a mere finger wagging from the local constabulary.
But in many ways the debate is sterile. We arrest people for deterrence and hope for rehabilitation, but mostly we do it for retribution. A man who kills his wife in a fit of passion but is genuinely remorseful is probably no future threat to society. He’s likely not a serial wife killer. He may find prison crimogenic (a word I’ve always known) but what of it? Send him to jail anyway because he killed his wife.
If we don’t do that, state retribution will be replaced by personal revenge. That was the point of law to begin with, to replace revenge and feuds with process.
I think that’s partially true but revenge is innate anyway. Even if you thought that killing the guy who killed your son would not deter his clan - but rather encourage them - to attack your clan, you would do it anyway.
Civilisation is about replacing arbitrary and personal conflict with laws and process, replacing chaos with order.
Under an evolved-for-deterrence theory, the point of attacking your son's killers isn't to discourage future attacks on surviving members of your clan by the specific parties responsible, it's to retain the credibility of the original signal.
Is this a theory that you just made up and have created a compound word for? Kudos but also - What’s this - German?
Anyway I don’t think revenge is about deterrence in the most part. Deterrence is something that a group wants, not the individual. Revenge is personal.
Crime is cultural. It is a set of behaviors people learn from those around them. It cannot be understood in terms of individual incentives and individual decisions without understanding how the cultural context shapes those incentives and decisions. An analysis which looks at crime on a single person basis will *always* miss this force, which is the most important one out there. Criminality is self-perpetuating the same way that any other habit or career is, and changing it is like any other cultural change -- very hard!
>In the five years following the passage of the law in 1994, crime dropped 33% in California compared to 26% in the rest of the country, meaning only 7% that you could optimistically attribute to Three Strikes.
Hold on, should we really be looking at the short term? During the section on deterrence, you describe the California three strikes as guaranteeing at least a twenty-year sentence. In that section, looking at immediate effects is fine, since we were talking about potential criminals imagining the consequences and choosing not to crime. But if your theory is that they can't commit crimes because they are in prison, excluding the vast majority of the extra time the spend in prison means excel most of the effect. I'm pretty sure our courts won't let the legislature retroactively increase the sentence of prisoners partway through their sentence, much less rearrest released prisoners, so the effect of the three strikes law is at most (not including deterrence) to reduce the number of arrests per criminal (within a certain subpopulation) to one. You don't specify what the expected rate of arrests was before the three strikes law, but since you talk about how the law was tailored for the worst of the worst, I would guess that the incapacitation effect was already significant over the five-year period. If the law is turning five-year sentences into twenty-year sentences, the expected effect over the next five years (again excluding deterrence) is literally zero, since everyone sentenced under the new lay would have been locked up for the rest of the study period even without the new law. If the typical turnaround between arrests is around 2.5 years (counting not just time in jail, but also time between release and reoffending and getting caught) then you can prevent the need for a second arrest for a 50% reduction in crime (among the affected subpopulation) compared to an 87.5% reduction in crime (same subpopulation) over 20 years. Obviously looking further into the future creates more potential for other events to mess with the data, but unless the previous laws were very soft the short time period is cutting off a huge portion of the effect, and scaling up to account for the fact that you are only looking at the tip of the iceberg doesn't work well when low numbers will get rounded to zero and assumed to be the result of confounders.
Editing to add a nitpick:
>That would mean that, if prisoners commit 10 crimes/year after release, the same people, given a noncustodial sentence, would commit 8 crimes/year.
>In order to neutralize the effect of one year imprisonment (-10 crimes), the negative effects of incarceration (+2 crimes/year) would have to continue for five years. But they probably won’t, because we know that most of these people get rearrested sooner than that anyway.
One year of imprisonment is only preventing eight crimes, since that is how many they would have committed if they weren't imprisoned. So the increase in crimes would only have to last four years after release on average to balance out. It doesn't seem to meet that bar either, so the conclusion is unchanged, but it bugs me that you are giving imprisonment credit for prevent hypothetical crimes that were only on the table due to imprisonment.
>>"Aftereffects must vary by place, time, and person."
This is by very far key to all this, surely?
This post is a sterling effort (as all "Much More Than ..." write ups, but probably the least ultimately satisfying of the ones I've read. I think partially because it doesn't make enough of a point of saying IT'S ABOUT THE US, and while it tends to stick to the American data/analyses it also makes visits to other countries' data (suitably caveated).
In the UK criminology I believe the received wisdom is that short sentences (not a day or there, I don't think anyone gets three days in jail here ever) are the most damaging, because the have disruptive to destructive effects on the prisoner's life while giving no chance for any rehabilitation to take place (leaving aside the fact that neither of the UK countries is exactly a beacon for prisoner rehabilitation anyway).
But it feels to me still that discussing prison effects without taking into account the specific prison environment and what happens there apart from being locked up is too general.
As to the social desire for "justice to be done", again, I feel that when people talk about crime concerns they really don't mean someone walking out if Tesco with a packet of cheddar and a bottle of wine (one case I recall from my last visit to court) or a drunk neighbour yelling abuse at his equally drunk wife (another case) or someone caught with a baggie of weed and another one of blow (again, same court).
That's an excellent point. The US is sort of in a unique situation with one of the higher crime and incarceration rates in the world. Countries elsewhere along the curve may have different solutions.
"Prison is less cost-effective than other methods of decreasing crime at most current margins. If people weren’t attracted by the emotional punch of how “tough-on-crime” it feels, they would probably want to divert justice system resources away from prisons into other things like police and courts."
Why not both? I imagine most pro-prison people are also pro-police, just as people who favour spending more on healthcare believe in spending more on both diabetes and cancer.
I remember reading about crime and punishment in 18th-century England, when hanging was the scheduled penalty for everything from shoplifting to murder. It was also the early days of 'criminology,' and thoughtful observers couldn't help noticing that the death penalty seemed to have no deterrent effect (although it certainly had 'incapacitation' and 'recidivism' covered). Nor did the prospect of going to hell after being hung. What was the answer? Fortunately, a real-world option presented itself around the time that America refused to take British convicts: go to the southern hemisphere - which must have been like going to Mars - settle New South Wales with convicted felons (bonus: claim the land before the French did, probably with better outcomes for the indigenous people), and start a new country. And on the whole, it's worked out well. Not every jurisdiction has that option, though.
Strange statement about the French. By and large Anglo imperialism was worse than other European powers in its treatment of the natives, and Australia was very very bad - classifying aborigines under flora and fauna until 1967.
The flora/fauna thing is completely false. https://aiatsis.gov.au/explore/right-vote which explains that many Aboriginal people had had voting rights since the 19th century. Also, France had slavery in its colonies until 1847 (after a brief hiatus 1794-1802), so I'm not sure why its imperialism is considered better.
Many is clearly untrue. Some. And no voting rights for most until 1967. Whether or not the French were always great, the idea that Australia would have been worse for natives seems odd, as it really couldn’t have been much worse.
I don't mean to defend colonisation in any way, or get drawn into a comparison with the French, but thinking that it couldn't get any worse is a failure of imagination.
What if the Tasmanian genocide had been 100% of the population instead of 20%? What if the token efforts to stop blackbirding hadn't happened at all? What if blackbirding had been on the same scale as the transatlantic slave trade, or under conditions even closer to those the Belgians implemented in the Congo? What if the Stolen Generation had all been sent to death camps instead?
True. As you've probably read, it actually increased murder, because if someone caught you stealing and the punishment for both theft and murder was hanging...well, better not to leave any witnesses.
Probably the most extreme example I can think of is Imperial China, where several revolts, including one leading to the establishment of a new dynasty, resulted from people making a similar calculation.
"Hey Liu Bang, the prisoners escaped!"
"Crap! What's the punishment for letting prisoners escape?"
I disagree with posts like these that talk solely about the criminal's perspective and center the crime they committed without contextualizing with the victim's perspective.
The point of punishing criminals is not just about them--it's about showing their victims that society gives a shit about what happened to them. Maybe stealing $250 worth of stuff sounds like a minor crime to the people on these and the EA blogs, but that's enough to ruin a poor person's life for a while, depending on what was stolen. Ruining someone's life and destroying their trust in society deserves punishment, period, even if their stolen goods aren't fancy enough to matter to rich people.
So... if I'm understanding correctly, your counterargument to the comment I was replying to is that society does not, and should not, care what happens to poor people? "Screw you, Jack, I got mine," more or less, as anyone who hasn't already pulled themselves up by their bootstraps must, ipso facto, be morally unfit?
If that's your position, I think our respective core values and assumptions may be too far divergent for much further productive discussion to be possible, though one way or another I'd be happy to be proven wrong.
That doesn't solve the problem of broken social trust. Also, why would you be okay with a system that essentially gives a free $250 to people as a reward for doing something bad?
If everybody's getting UBI that's adequate for typical subsistence needs, there'll probably be less theft in the first place, when somebody does it anyway they can be prosecuted vigorously (in fully justified confidence it wasn't for bread to feed their starving family), and the victim is in a better position to focus on obtaining longer-term justice since they're less immediately busy with damage control.
I am deeply sceptical that a significant part of the decline in homicide in El Salvador was due to mass incarceration, unless the flow of time and causation has reversed itself.
That graph of homicides per year you posted has a really sharp peak in 2015, at around 100 per 100,000 people per year. Thereafter, it starts to plummet dramatically every year.
So El Salvador clearly got /something/ right, and that something happened around 2015.
The thing is that mass incarceration in El Salvador started in 2022, by which time the homicide rate had already fallen massively, to about 10 per 100,000, and was still falling rapidly. That improvement didn't accelerate in 2022 - yes, it continued, but if anything it slowed.
So I think this is almost certainly correlation not causation - unless falling homicide in El Salvador somehow caused mass incarceration!
This doesn't quantify by far the most important effect of prison: The reduction in time available for increasing the size of the next generation of high-probability criminals.
The well documented population criminality difference exists because of thousands of years of cities enacting population-altering punishments.
This effect has a compounding interest which completely overwhelms all other effects. It's utterly negligent to claim to comprehensively cover this issue if one isn't willing to address this factor.
One thing that worries me about the incapacitation argument is the apparent assumption that being a high-volume criminal is an innate trait of an individual, rather than a positional role in society.
Like: I don't personally know how you turn stolen bikes and cosmetics into cash, but assuming there's some black market for those things, does removing a bunch of existing suppliers not just increase the incentives for new suppliers to get into the game, like any other market?
I don't really know how drug dealing works, but if you arrest all the people selling the drugs, do the people making the drugs not just recruit new people to sell them?
If there's some cultural conglomerate that produces criminality (like a gang/other criminal friendgroup of some kind, or maybe a really poor neighborhood where the cops never help and everyone assumes it's them vs the world and doesn't feel a moral obligation to follow the law, or w/e), does removing some members of that group create opening that new people enter?
Etc. I don't know enough about how the people who commit lots of crimes get into that position, but without a really exhaustive understanding of it I would not natively assume that those people are not replaced when they get removed. In the same way that I would assume that shutting down the top 3 shoe retailers in the country would not decrease the total number of shoes being sold 5 years later.
Sure, you'd expect to use standard microeconomic tools to model a market for stolen goods or illegal drugs. Disincentivizing people from supplying them (by creating a risk that they'll go to prison for a long time if they do) would be expected to shift the supply curve left, increasing supply and reducing the quantity produced. The latter, of course, is what we want to achieve.
What's your take on the classic narrative that crime occurs because people are hungry and need the money to provide for their families? In your cases a, b and c it doesn't feel like any of those people do it out of need and desperation, which is the argument you often hear, but more out of choice.
The classic narrative doesn't seem very robust, to put it lightly: https://www.astralcodexten.com/p/prison-and-crime-much-more-than-you/comment/79269336.
> We domesticated ourselves as a species by deleting [killing] bad men each generation, every generation, for 250,000 years.
The Tyranny of the Cousins only works in Dunbar's group numbers of people, it doesn't work for modern states or cities.
Also, pretty much any weirdo, smartmouth, iconoclast, person who bucks conventional norms, introverted, less-socially-skilled, socially dominant, too-visibly-better than the existing Cousins, and who knows what else, are at life-and-death risk in such a Tyranny. And I think at least one, and usually several, of those descriptors probably applies to at least 80% of the people here.
Also, even if you assume we'd have the scientific method, if you think of most of the people who drove a lot of the technological and economic growth that's given us our modern world and lifted billions out of poverty, a lot of them would have been killed in a Tyranny for those "offenses." Newton. Tesla. Semmelweis. Ford. Zuck. Brin and Page. The list goes on - pretty much any captain of industry or major nerd, would have likely been killed for social reasons.
Is that really the world you want to live in?
At absolute minimum before agreeing to such a policy I would want to have an awful lot of crimes removed from the books, or defined far more clearly and narrowly, to make sure it doesn't become possible to judicially murder someone by e.g. tricking them into downloading three copyrighted songs.
Could make a constitutional case that the right to a speedy trial is being systematically violated, and the only appropriate remedy would be for every region with a backlog of cases to hire more judges (or whoever else is needed) until that's no longer the case. If cost is a concern, raise taxes on land value - since, after all, such rights would be worthless without the law defining and enforcing them. Just the actual land, though, not structures.
For pretty minor crimes, that's not a bad description of an English magistrates' court. They are limited to never passing a sentence above six months. Offences are divided into "Crown court only" (indictment), "Magistrates' Court only" (summary) and "Prosecutor's choice" (either way). The magistrates can refer an either-way case to Crown Court if they think it should be a sentence of more than six months (any crime with a max over two years is always indictment).
ETA: UK courts are divided up as follows:
For criminal:
Magistrates' Courts: three lay magistrates with a legal advisor ("Recorder"); the magistrates are trained but not lawyers; they're sort of a permanent jury. Sometimes a district judge sitting alone (wearing blue robes, so a "blue judge")
Crown Court: circuit judge (wearing purple robes, so a "purple judge") and twelve jurors. This is where most serious or violent crimes are heard.
Central Criminal Court (aka "Old Bailey"): hears all the most major criminal cases (the closer to London the crime was committed, the more likely it is to be moved from a normal Crown Court to CCC), High Court judge (wearing red robes, so a "red judge") and twelve jurors.
Red judges can hear major cases in any Crown Court, but in practice, they mostly work in London and all major cases in London are CCC; non-London Crown Courts will hear some murder cases and those do often involve a red judge trekking up for the case.
Appeals from the magistrates' courts are heard by the Crown Court; from the Crown Court or the CCC they go to the Court of Appeal (Criminal Division) where they are heard by a panel of judges ("Lords Justices of Appeal", wearing black robes, not known as "black judges", though). Appeals from CA go to the Supreme Court (aka UKSC).
Civil cases have a similar system, with the small claims court taking the place of the magistrates, the County Courts the place of the Crown Courts and the High Court the place of CCC. Nominally, the small-claims court is just an simplified procedure in the County Court, and they're all heard by purple judges, though civil judges actually wear simple black robes these days, the same as barristers do, not the fancy things that criminal judges wear (note: a judge may be criminal one day and civil the next; the ranks of judges are district, then circuit, then high court, then appeal, then supreme court. In practice, most judges are somewhat specialist)
That would last approximately however long it took for the first black guy to be publicly flogged by a white guy, then we have major riots. Maybe it works in Singapore but I don't think it could work in the US.
This is where technology should come in! Surely it's trivial to build an automatic flogger in this day and age, and you can even find the ghettoest-looking guy to design it for good measure.
Brb registering a patent
I kind of wonder; given what we've read in Scott's post much of the outcry might be coming from leftists rather than people in poor neighborhoods, but of course they are great at stirring up outrage. I suppose you could try to have racially matched floggers, but that gets complicated fast. ("We, uh, need a Cablinasian flogger for this next guy...")
I find this conclusion puzzling up against the data showing increased policing is many more times effective at reducing crime than prison, which is after all a kind of punishment, including that it prevents the victimization in the first place.
It is possible to go to prison for simple possession in some states, but it's not common and certainly the vast majority of drug offenders in prison, especially the ones with very long sentences, were dealing. And at the federal level, pretty close to 100% of prosecutions are for distribution, even though possession is criminal (and in some case, a felony).
How many instances of what crime besides murder would you be in favor of execution?
Let's say... Vehicle theft. How many times would someone do that before you would suggest killing them?
Great research here
The current focus on thoughtcrime imprisonment, particularly in the UK, where violent habitual offenders are being released to make room for those who posted mean tweets, will have some very interesting effects.
In "A Clockwork Orange," the Home Secretary explains that they need to let violent criminals like Alex out to make room for the new political prisoners.
Wasn't there the implication that the party that tried the Ludovico Technique ('rehabilitation') was Labor, and the one that let him out was the Tories (the politician's kind of conservatively dressed)?
If so it would be a funny comment on who's into censorship over 60 years.
Do you have a link on the violent offenders being released point?
https://www.youtube.com/watch?v=kPlA8EYv4mI
I don't see anything about violent offenders in the video or the related news articles.
It’s not specifically violent only, but they’re in the mix for early release
That will be a no, then. It's also straight nonsense that people are being imprisoned for "mean tweets". Musk has been ploughing this unfortunate furrow.
Okay, so if I find you a story of a person who was arrested for "mean tweets," would you admit you are wrong? I expect what you will do instead is say either that it's very few people, or that if their tweets were illegal (labeled "hate speech", "incitement" or whatever), then them being imprisoned is GOOD actually.
Instead of talking hypotheticals, it would be preferable to just find such a case.
It's even simpler than that: just find a law that makes "mean" tweets illegal.
I'm aware that a disturbing number of people are being questioned by police, arrested, or even prosecuted over tweets and posts in the UK. But none of the cases I am aware of led to actual *imprisonment*, as far as I know.
The legal standard of incitement is much, much, much lower in the UK, the American mind cannot fathom how low: https://www.bbc.com/news/articles/cy76dxkpjpjo
And umpteen bajillion stories of people getting police visits for the radically Orwellian “non crime hate incidents,” but I’ll leave that googling to you.
I’m pretty sure it was the UK, where I read about an old Jewish dude who likes to haunt anti-Israel protests. He hadn’t done anything much, so as to say, but he made the mistake of looking thoroughly Jewish, apparently.
So “mean tweets” has a physical manifestation. Seen deets?
Someone called for violence: someone else committed the violence they called for. Of course that's incitement. What other standard could there be?
If you don't want to be held legally responsible for violence, then don't call for it.
https://www.theguardian.com/uk-news/2022/jun/14/ex-police-officer-jailed-for-20-weeks-over-racist-whatsapp-messages
In other news, Australia just approved s law that makes social media off-limits to children under 16.
Neatly avoids all vague thought police free speech issues!
Courtesy of the BBC, "more than a third" of early release offenders had violent offenses: https://www.bbc.com/news/articles/cd0594gx71xo
At least one early release offender went on to commit sexual assault literally the same day as his release and was sent back: https://www.the-independent.com/news/uk/crime/prisoner-early-release-scheme-amari-ward-assault-b2612921.html
Further comments that sexual offenders are not supposed to be eligible and that the vetting may be inadequate: https://www.the-independent.com/news/uk/politics/prisoners-early-release-sex-offenders-priti-patel-b2632979.html
People who end up in prison tend to have very high time preference. That seems to me like the probable reason that adding more time to an already long sentence will have a very weak deterrent effect. Think about it this way: does the kind of person who would get convicted of two felonies and then go back out and consider committing a third felony seem like someone whose decision-making is driven by what happens to him five years from now?
Unless you're talking about white collar crime sprees, those would be prevented in most cases even by a much lower level of deterrence. Those people have something to lose (namely, a decent career) so even a modest risk of an arrest record will outweigh any economic benefit of property crime.
Over a long enough time you can change the culture even of the underclass to assume that a life of crime is a bad idea. But it takes a long time.
It takes a long time, lots of support, and most importantly of all, willingness. Some people *do* want to do better, but don't know how/need a ton of help. I have sympathy for them, they often came out of circumstances that were chaotic, they slipped through cracks, and if they get help they will be grateful for it.
And some people, with all the help from all the bleeding hearts in the world, will never change because they have nothing but contempt for ordinary people with ordinary lives and the mores of ordinary society. I have no sympathy for them and no "oh no if they have to spend ten years in jail it will interfere with their enjoyment of life" cares.
Historically, the approach was rather the opposite, with frequent and aggressive use of the death penalty for offenses that we wouldn't today think merit it - that's actually the root meaning of "felony" - a crime subject to the death penalty. And the spectacle and ritual around public executions and torturings had social control functions as well.
I think at least part of the reason for this is that certain types of crime represent an economic niche that *somebody* is going to fill, and it's typically going to be people with the least to lose.
That is true of supplying various vices, but not of normal crimes with victims. There's not an economic niche needing to be filled for people stealing cars, burgling houses, or mugging little old ladies.
If true, this suggests corporal punishment would have a much more significant deterrence effect, since it's delivered all at once in the immediate future. (It likely won't have the incapacitation effect, because it doesn't take that long to heal from lashes.)
FWIW, pickpockets used to operate in the crowd that gathered to watch convicted pickpockets have their hand chopped off. (I think that was Dickens, but it's probably true.)
Sure, because the spectacle of public punishment would have been one of the largest and most crowded reliably-scheduled gatherings of multiple classes on offer at the time - too good a target to pass up.
You're misremembering your history. Pickpockets were sometimes publicly hanged, as stealing enough handkerchiefs could be a capital offence (though many would have had the sentence commuted), but even 100 years earlier than Dicken's time, amputating hands had long since fallen out of use as a punishment.
OTOH, hanging doesn't detract from the main point. Awareness of severe penalties does not, in itself, act as a strong deterrent. You need to combine that with an expectation of being caught , and if you do that you don't need the *severe* penalty. It just has to be more than "the cost of doing business".
I guess the missing control is, how much pickpocketing happens in those crowds compared to other crowds? If it's the same, I guess deterrence do not work but there's a chance that it's significantly lower.
By "immediate future" you mean after months of trials in court, like any other punishment?
https://www.astralcodexten.com/p/prison-and-crime-much-more-than-you/comment/79108162
You might be interested in my other comment, about how most of the logistics here is a policy choice, which we could choose differently.
Yeah, I think at least for lunkhead street crime, if the dude committing the crime was thinking a lot about where his current choices would leave him in ten years, he would be on an entirely different life path. That kind of crime selects for people who are giving *very little* weight to possible bad outcomes a decade or more in the future.
Seems like the only real solution is for society to develop a parallel informal justice system based on the intuition that any method of clear self-defense of property and/or body is justified and allowable. Shoplifting rates would drop precipitously if there was a very good chance of being whacked with a baseball bat or even shot by a shopkeeper. Vehicle theft would probably stop if there was an excellent chance of painful booby-trapping. Etc.
Man, I wish I could live in that utopia.
Well, certainly the shopkeep has to be careful about not whacking a non-thief with his bat, lest s/he be punished for engaging in *unjustified* violence.
But that's also why I used the words like "clear self defense."
Nah.
What's the standard for proof? Who gets to decide?
Seems like a *really* bad approach.
Remember when I said the words like "informal" and "intuition?"
Even monkeys know when something has been stolen from them. People know what's right and wrong. When it's clear and obvious that one person is attempting to wrong another, a good civilization would shrug at whatever immediate consequences happen to the wrong-doer as a result of their wrong-doing.
Risk of prosecution is what is suppressing such shootings and baseball battings right now. Presumably you want to expand the range of circumstances under which it is justifiable. But do you want to expand it so much that loss of *property* can be justifiably met with *fatal* forc e?
Sure. While a *deep* value of property is not intuitive for the economic upper classes, that is not how people much closer to basic survival think of their items. Having a car stolen is at worst an inconvenience for a wealthy person, but for a working poor person, having their car stolen often means losing income, perhaps even their job, and possibly even their home/safety net. If a poor person wants to booby trap their car, I sympathize with the urge.
How does the old joke go?
"Do you really value your *property* more than his *life*?"
"The fact he's trying to steal my property is proof he values my property more than his life, too!"
The value of property is defined socially. So is the value of life. It's not just A against B.
The standard advice in east Africa is: if someone snatches your bag think before shouting "stop thief" because if you shout that the crowd will kill the thief. I don't know if this is true but I believe it is, and the belief doesn't give me a particularly warm and comfortable feeling. Also, theft is very very common in east Africa.
Can you summon a mob to kill anyone you don't like by shouting this at them?
Yes, and it is one of the major reasons support for mob violence is actually supported by more women than men in many countries (men fear they could be falsely accused). There are also several cases where mobs are used to kill out-group members with disabilities etc: https://www.amnesty.org/en/latest/news/2024/10/nigeria-escalation-of-mob-violence-emboldens-impunity/
I don't know. Possibly the mob turns on you if it murders the "thief" and finds no stolen property on the corpse. Being given the warning has had the beneficial effect on me of being extra careful of my possessions so as not to have to initiate the process.
I can't remember the source of the advice but a Google finds this in a guide book (other side of Africa)
https://www.bradtguides.com/destinations/africa/sierra-leone/
"The only thing worse than being robbed is seeing what happens if some poor unfortunate is caught: mob justice looms large on the streets of Freetown, where suspected thieves can be beaten up, belted or worse by angry bands at a moment’s notice."
Only if your target fails a saving throw.
It's a real problem in my country, and a reason why "mob justice" is heavily discouraged by government. There's even a similar one recently that makes national news (Indonesian language link) https://www.liputan6.com/regional/read/6146524/dituduh-intel-mahasiswa-di-makassar-dikeroyok-massa-hingga-meninggal
Consider directing your overactive empathy towards the actual victims of crime, rather than the perpetrators.
Have a think about what crimes are being perpetrated and by whom, in this scenario
If you love criminals so much go live in San Francisco. Otherwise keep your soft hearted nonsense to yourself. Normal people don't want to live in the violence and chaos that people like you create. Maybe when you see people you care about being victimized as a consequence of attitudes like yours you'll get some sense.
Thank you
I actually do. The problem is community punishment is really prone to false-positives (you bump into someone, they think you're stealing from them, and therefore are allowed to beat you up) and differential enforcement.
One of the advantages of having cops is it's their job to do this stuff, and when they screw up we can go after them. Diffusing it onto the general public leads to all kinds of problems that are the reasons we outsource it to cops in the first place.
Ehhh, call me a wimp, but that's not the culture I want to live in.
"Shoplifting rates would drop precipitously if there was a very good chance of being whacked with a baseball bat or even shot by a shopkeeper."
Multiple times in my life I shoplifted accidentally and came across as a thief.
I'm glad I didn't get whacked with a baseball bat, or shot!
I wrote about it in another thread:
https://open.substack.com/pub/astralcodexten/p/open-thread-354?r=1dtkhh&utm_campaign=comment-list-share-cta&utm_medium=web&comments=true&commentId=76683665
Also relevant: https://scottaaronson.blog/?p=3903
"Surely living like this would have massive, life-ruining consequences in other contexts?"
Maybe if I were a surgeon, which I am not.
I agree. It's forgivable when young children get over-excited and carry something out of a store, but the other two examples aren't okay.
I have never, ever, not once in almost 45 years of life, been accused of shoplifting, because I am ostentatiously careful about not giving the impression I could be shoplifting. I don't think I've even even put an item in a pocket when my hands were fuller than I was expecting.
But that's because I care a lot about presenting myself as an honest and trustworthy customer. Someone saying, "but I happen to care about other stuff more, so I shouldn’t be punished for that!" isn't a persuasive argument for me.
> I agree. It's forgivable when young children get over-excited and carry something out of a store, but the other two examples aren't okay.
Really? Example #3 is "I paid for a bunch of bananas, stuffed them into my jacket, and left the store". How could that not be OK?
The examples are ok, this can happen, and it doesn't really matter.
It matters a lot less than some shoplifters, or some distracted people, or some people seeming like they did shoplift when not, dying because they took a wrong hit.
I accidentally shoplifted not long ago. I bought four yogurts and only ran it through three times. I discovered my error by accident when looking at the receipt to check my fuel points.
Caba lists two cases of accidental shoplifting, one as a kid. Leaving only one as an adult.
N=1 isn't much evidence for anything. Assuming they had 10 years of adult life, even if you use p < 0.05 as a criterion, you can't rule out their actual rate of such mishaps being once in 190 years.
Probably you would have been more careful not to accidentally shopliftt if there were dire consequences attached, no?
No kidding. Also, multiple times? Like, more than 3? At some point, it stops being accidental.
Even so, is severe injury/brain damage/dying a reasonable thing to experience for a (potentially one-time!) genuine mistake?
Let's try it for a while and see how often people are injured for genuine mistakes!
Where I live you'd have a hard time finding anyone who agrees it's a reasonable punishment even for actual, intentional shoplifting.
Maybe. But most people who shoplift once or twice aren’t doing it because they’re thinking clearly about the consequences.
Why is that the only real solution?
We don't have such crazy crime problems here in Singapore, and we certainly didn't use your 'only real solution'.
Consequences for violent crime and shoplifting are pretty severe in Singapore, are they not? And personal /property crime is comparatively low, is it not?
If so, great, that works, too!
Singapore canes people. Caning is incredibly painful! Effectively, the Singaporean solution to crime is torture.
https://x.com/cremieuxrecueil/status/1858978956982022350
Does it really make sense to call that torture? Punishment of any kind has to be unpleasant to be effective, no? I don't know where the line between torture and punishment should be drawn, but Singaporean caning strikes me as one that falls on the P side of it.
How would you define torture in a way that excludes caning, but doesn't exclude a bunch of things that you'd consider torture?
If I had to chose between being caned, or going to prison for a year, I would choose caning every time.
Yep, it's amusing how the current western Overton window is squarely at "everything except prison is cruel and unusual". Prison is plenty cruel in most places, and the unusual part is self-fulfilling.
Prison is stronger torture than caning.
Singapore's solution of torture ("corporal punishment") would work too.
You're one of the few well-functioning dictatorships in the world, and very small in size...and if I can be a little un-PC don't have populations that cause huge problems.
One day drones will do this non-lethally and prison won’t make as much sense for most things. You’ll just shout a drone word and a bunch of them will flock in and take care of the problem.
Laser drones with a 99.99% correct conviction rate will instantly non-damagingly laser the culprit for three city blocks unless they reverse plea bargain into being arrested and tried
Should this approach be reserved for poor people's crimes, or would you condone extending it to upper class crimes as well, such as wage theft, tax evasion and bribery?
Of course I would.
I'll go you one further: I'd love to live in the kind of society where someone enacts vengeance on any of the key Purdue Pharma Sacklers, and, after a trial establishing the facts, the jury takes 30 seconds to deliberate before issuing a dismissive shrug as their verdict.
Our justice system has way, way over-corrected in favor of predators like the Sacklers. It'd be great to live in a world where no DA would even bother to bring the murderer of a Sackler Board member to trial because there would be no hope of finding a jury willing to convict them. The Sackler Board members probably wouldn't have been quite so cavalier about pushing an addictive product if they believed there was a good chance of being injured or murdered by grieving family members.
They already more or less do that with pedophiles.
How far do you extend culpability? Board members? Executives? Managers? Truck drivers? The guy who did Purdue's IT support?
I always kind of wondered how much worse the Sacklers were than every other pharm company selling opioids, which have legitimate medical uses.
I'd like to rely on the collective intuition of a far-more justice-based society to make those decisions! It'd be great if a sales person way down the hierarchy was just a *little* worried that a grieving parent of a dead overdosed 20 year old might decide to come after *them* for selling to that kid's doctor. But I'd also like that parent to be worried it might not fly with a jury.
I seriously doubt this be the result. Most people wouldn't suddenly start carrying weapons to wave for starts, surely there are only a few who want to, but can't right now. Plus, crime happens mostly because it *isn't* caught in time. The outrageous occasions when someone steals something on broad daylight, simply isn't much of it. It's not like chains and department stores are going to adopt this either - liability for accidents is likely way scarier then some inventory lose. And for detterence? Scott already covered that!
Plus people would make the calculation that it would be OK to rob old or frail people or children or women, because they could count on winning the ensuing fight. Plus a lot of people would get killed defending themselves from the crooks where they might have previously lost a handbag or backpack.
Spoken like a dude who isn't a woman who thwarted TWO separate robbery attempts by strangers merely by indicating she had a firearm and was prepared to use it.
Nice job! (Seriously--I was scared of criminals before I was scared of feminists.) Happy you got out safe...I would have been tempted to fire.
You then get into the whole gun ownership question, which I am always flipping back and forth on. I get the feeling the rationalists have already looked into this but I also get the feeling they are liberal wimps who just don't like the idea (look at the recent Prop 36 thread) . I'm sure someone here can fill me in further with actual data...
Nonsense. This is wildly incorrect. Responsible American law-abiding gun owners are very aware that even in the best-documented, clearest-cut cases of self-defense, they are likely to be hassled tremendously by law enforcement and/or the perpetrator's family. Most are far more afraid of losing their house for righteously shooting a criminal than they are of being in a deadly threat encounter itself. That's why self-defense insurance exists.
Remember, you're a woman and unlikely to be targeted by people looking for thieves. As seen below false positives are a problem. Theft now becomes an excuse for violence as well as self-defense. You can see racial bias erupting (on both sides, though only one will be covered by the media) with false accusations of theft leading to violence. Not to mention booby-traps going off at the wrong time--you're going to be walking by cars and have them blow up because someone set their booby trap too sensitive, or get killed when the swinging anti-theft blade the shopkeeper sets up triggers when you step on the wrong square.
I'm sympathetic to your concerns with high crime rates in liberal cities (I lived through the transition in the 90s and heard about the reverse one in the 2010s), but IMHO we just need to let the cops do their jobs again.
EDIT: I'm disavowing this statement and admitting that I was wrong. It was sexist by progressive standards, argument ad hominem by rationalist standards, and uncharitable by conservative standards. I'm stepping away from ACX for a while. Good luck to all.
I think Scott's entire post was about why cops and detention facilities can't do their jobs? It's due to a kind of cost-disease in the justice system?
My actual views on why they can't are too un-PC to post here and I will respect Scott's desire to avoid trouble with the progressive left so he can post on everything else he finds interesting and, now, not have his children harassed by leftists.
That worked out really well for all the black victims of white lynch mobs.
Take your strawman elsewhere, please.
RE accidental shop-lifting: It happened to me once, with a larger (supermarket) purchase, like 100€. (Background: I live in Germany.) My partner and I both thought the other one had paid. We noticed at home, called the super market, and returned another time to pay. We got a bouquet of flowers (I'm not kidding!) and a bottle of something for our honesty. We were told this happens all the time (dunno how pricey the stolen purchases are on average), but this was the first time the long-time employee we contacted has seem people actively come back to pay another time.
This probably won't convince you, but I'm very glad I don't live in a world where a forgetful grandma at the self-checkout in front of me might kiss the floor after getting a dent in her head. It could be your grandparent, for Christ's sake! We seem to inhabit very different mental spaces
I'm not sure if I believe you really don't understand my "mental space." Human beings can tell the difference between thieves' furtive theft or smash-and-dash and a legitimate customer obliviously walking away with an extra yogurt. It should be fine for a motivated shopkeeper to threaten the former group with a baseball bat and politely remind the latter group to come back to pay. They're different people.
Fair enough, I suppose we just have a different intuition for the false positive rate and/or the cost of said false positives.
And the people who trespass or shoplift accidentally?
Reread my comment, thanks!
You mean the claim.that ordinary people have an infallible intuition of guilt?
I didn't use those words, so maybe reread my comment again, thanks!
That's where we started! Our current system is supposed to be an improvement on it. The norm for most of history was vengeance killing, public stoning and other forms of lynching. Governments hardly ever had the power to deter crime outside of cities. If you want a sense for why the West transitioned to laws & courts, The Oresteia by Aeschylus is a good dramatization.
Yes, and unfortunately the current system has drifted so far away from actual "justice" that it's no longer an "improvement." It's just time to drift back a bit.
Yeah, I kind of agree. Except they have these cycles--we had a similar situation in the 60s and 70s and they all decided to get 'tough on crime' in the 90s and crime fell long enough for liberals to forget why those policies existed in the first place.
My impression is that longer and surer imprisonment can incapacitate criminals and even deter crime, but it can take quite a while for the criminal class to get the message.
For example, America started getting softer on crime maybe in the early 1960s. Then crime started to get notably worse in the mid to later 1960s, and American finally started to get tougher on crime maybe at the end of the 1970s. But then crime remained mostly high, although we did have a nice period of lower crime around 1984 in between the powder cocaine years and the crack cocaine. (I'm using homicide rates, which are fairly reliable, and calling them the "crime" rate.)
But the crack years (roughly 1988-1994) were horrible for the homicide rate. What seemed to be happening was that, as older criminals were increasingly locked up, violent criminals kept getting younger and younger, and they were peculiarly homicidal.
The crack years overwhelmed the prison system. Back during the crack years, a friend of mine who is in the comedy business was talking to a colleague's brother, who was a prison guard in upstate New York. The guard said, "Is there anybody you'd really like to murder? Because if there is, now's the time to do it, because we're so overcrowded they'll let you out after only about eight years." (A prison guard with famous comedian genes can have pretty memorable insights.)
But America built a ton of prisons in the 1990s. And suddenly, in the second half of the 1990s, the incapacitation strategy recommended by James Q. Wilson in the mid-1970s in "Thinking About Crime" started to work. And then it worked surprisingly well. In particular, tough-on-crime New York City became remarkably less lethal.
It worked so well that the Establishment began to forget the hard lessons it had once learned and after a couple of decades started the soft-on-crime cycle all over again, causing homicides to soar due to the Ferguson Effect and especially the Floyd Effect.
Now we may be headed back into a tough-on-crime cycle due to the predictable failures of soft-on-crime policies.
Are you aware of the standard arguments against increased incarceration being responsible for much of the secular crime drop, and do you have responses to them anywhere?
We had decreased incarceration for about a decade and a half in the liberal 1960s and 1970s, during which crime eventually soared. Then we had increased incarceration for about about a decade and a half during which crime remained high and even peaked toward the end of that period in the early 1990s.
A whole lot of college professors in the social sciences got hired during those 30+ years in which everybody who was anybody either was for decreased incarceration or at least admitted increased incarceration wasn't visibly working.
Then, all of a sudden, in the later 1990s, increased incarceration suddenly seemed to start working and kept working for a couple of decades. In fact, a return to decreased incarceration seems to have led to an explosion in the homicide rate in 2015-2021.
Would you suddenly admit that your career was based on a fallacy? Would you turn against your academic mentor who boosted your career?
"Would you suddenly admit that your career was based on a fallacy? Would you turn against your academic mentor who boosted your career?"
Or worse, and risk being called a racist or similar terms and risk your career itself?
Scott's one of the very few people who could get away with that, I think, assuming he had enough data to convince his rationalist fanbase it was actually based on dispassionate investigation, etc. He already survived one direct attack from the NYT.
Your lefty academics are definitely a different example. Watching crime fall through the 90s in my hometown was one of the things that (a) permanently made me a a tough-on-crime guy and (b) convinced me the social sciences were FOS. I've softened on the social sciences in general since then, but I think the leftist monoculture in academia is one of its biggest weaknesses.
Didn't Canadian cities have a similar pattern of crime rates rising and then falling, but without the big increase in incarceration?
In general, social science is where nice,comprehensible one-parameter models of complicated social phenomena go to die.
It did sound like a just-so story, but I've heard a significant reason for falling crime rates in the 90s was Roe v Wade.
What is responsible? I know Levitt suggested abortion contributed, but then critics said his results came from a coding error. Kevin Drum has been saying it was lead, but now we have papers indicating that was publication bias. A lot of such papers were written during the crime decline, but we witnessed crime increases after Ferguson & Floyd that can't be explained by lead being re-introduced or abortion going away.
Even if the increase and drop from the 60's to the 90's was largely driven by a couple of causes, it doesn't mean those are the only causes for changes in the crime rate. So increases in crime around 2020 don't necessarily mean anything about earlier changes.
We have had during my lifetime three periods of liberalization of the criminal justice system and the dominance of anti-law and order thinking: the long Warren Court era and the short Ferguson Effect and Floyd Effect eras. All three saw the homicide rate rise sharply.
But maybe if we tried it a fourth time, it would stop happening.
Levitt has addressed the coding error, and fixed it and A) it did not dramatically change the results in his first paper and B) in a follow up paper 20 years after the original, the predictions made in the original were borne out. I have not yet come across a convincing critique of his work
Here is a link to a podcast episode where Levitt discusses all this. It also includes a transcript if you don't like listening to things like this:
https://freakonomics.com/podcast/abortion-and-crime-revisited-update-2/
The other major change (which is also discussed in that podcast) is lead (I haven't heard the publication bias argument before nor heard a response, but I'm skeptical). Crime is complicated and no one is claiming those are the _only_ two causes (another one, also discussed in that episode, is increases in number of police, which convienently agrees with Scott's post).
But I am relatively convinced that the combination of decreases in lead, decreases in unwanted pregnancies, and increasing police preseance are by far the most effective _and_ most cost effective deterrents to crime.
A sidenote, but the Levitt "I fixed my error" thing really shows researcher degrees of freedom. Iirc, the sequence goes as follows:
Levitt: When I use method x, I get result A
Comment: Levitt's code actually uses (inappropriate) method y. If you actually use x, you get ~A.
Levitt: Well, if I use method z instead, result A (approximately) returns.
Great point. A lot of these "but ackually [common sense thing] can't be right because [points to study]" crash into the reality of (a) the existence of places like Singapore and El Salvador, and (b) you-can-see-the-timing-perfectly-on-a-graph homicide surge post-George Floyd / "defund the police" (and Ferguson effects in general). Obviously criminals adjust their behavior to the perceived probability of being caught. They're stupid but they're not *that* stupid.
How do you know that it was the George Floyd stuff, rather than stuff related to the pandemic? The same could be said about Ferguson effects: was it rather something related to the increase in poverty or inequality?
In public policy, there are always so many things changing that observations of correlations usually don't tell you anything. That's not to say that you're wrong that Ferguson effects are causative, but you've not claimed there's any evidence that it's causative (rather than correlation).
Do you have a citation for the impact of lead on crime being inflated by publication bias? That sounds very interesting- and also totally plasible
https://www.sciencedirect.com/science/article/pii/S0166046222000667
This was an incredible summary and very helpful - thank you
This also matches my experience working with some low-income boys as a volunteer. It took me too long to realize how terrible they were at time-discounting and weighing risk. Where I was saying: "this will only hurt a LITTLE but that might RUIN your life," they heard: "this WILL hurt a little but that MIGHT ruin your life." And "will" beats "might" every time. One frustrating kid I dealt with drove without a license (after losing it) several times and drove a little drunk occasionally, despite my warnings that he would get himself in a lot of trouble. He wasn't caught and proudly told me that I was wrong: nothing bad happened, whereas something bad definitely would have happened if he didn't get home after X party. Surprise surprise: two years later he's in jail after drunk driving and having multiple violations of driving without a license.
People, especially young men with little education, are pathologically bad at evaluating risk. If the criminals are risk analysts committing white collar fraud, then sure you can increase the costs of committing a crime and let them factor that into their risk calculations. But for normal criminals: ordinary low-education, risk-blind high-testosterone schmucks, the best thing you can do is increase the probability of getting caught and feeling immediate repercussions, regardless of how small they are.
I moved for a while to a central European country and was impressed with the criminal justice system and the low level of crime/high level of compliance, despite the high amount of migration. The biggest thing I noticed is the certainty of being caught that you mentioned. Sentences were (to my American eyes) relatively short and weak - and more liberal Americans often thought there was less crime because the country was somehow more humane. But the main thing I noticed was: the certainty of being caught was high. Police were everywhere and very responsive. Small fines and one-night-in-jail punishments were everywhere for petty crimes. I think the U.S. would benefit from strong policing with simple, consistent punishments for many small crimes, especially for first time offenders.
One of my favorite lessons I have learned from working in a Corrections-adjacent position the last couple of years is that generally, people who get incarcerated aren’t there because of their great decision-making skills.
That's why, not to be unkind to Scott, I'm smiling wryly at his "imagine your life after ten years in jail".
If the majority of offenders were married, with kids they had within marriage by their spouse, and working jobs, with supportive families... we wouldn't be having a discussion over "ten years in jail, too much?" because they wouldn't be putting themselves into the position of doing crimes worth ten years of jail in the first place.
White collar crime is somewhat different, because you will get idiots who think embezzling half a million from their employer is the perfect crime and they'll never get caught because they're smart, after all they're working an office job and they're not some underclass type.
More than half of prisoners have children under 18! https://bjs.ojp.gov/content/pub/pdf/pptmc.pdf
Believe me, I am unhappily aware of that.
"Your partner has long since filed for divorce and is happily remarried to someone else. Your kids have long since moved on; if they remember your name at all, it’s as “that guy who was never there for us”
However, all those children are not necessarily by their spouses, if they have spouses in the first place, and if the people I have encountered with kids under 18 were all married to the fathers/mothers of said kids, then polygamy would indeed be legal.
Though I will concede that they are married, according to the link:
"In state prison, the likelihood of being a parent was most common among married inmates (71%), compared to inmates who were separated (64%), divorced (55%), never married (45%), or widowed (36%). Among federal inmates, married inmates were more likely to report being a parent than inmates in all marital statuses except those inmates who were separated from their spouse. There was no difference in the prevalence of being a parent between married federal inmates and separated federal inmates. The likelihood of being a parent varied little by education for both state and federal inmates."
This does not refute Deisach's point, and even you should realize that.
Even if they had kids, the vast majority of them wouldn't have been there for their kids in the first place. Probably not for their romantic partners either, at least not in any good ways.
Interestingly, decision-making skills are literally one of the things they try to teach as part of treatment in Minnesota prisons. Cognitive interventions, that sort of thing.
My sense is that it works great for people who mostly don’t need it.
"My sense is that it works great for people who mostly don’t need it."
Like all the most successful interventions, then?
(Gosh, I'm cynical today).
Putting this in as nice and polite a way as I can: I suspect I am on the low-end for IQ (perfectly high midwit-ish 117) of the average ACX reader, but interestingly wife (135+ish, based off ACT score, STEM PhD from Ivy League uni) is closer to modal for this site. So I get to have the fun experience of interacting with someone a full StDev smarter than me, and one thing I often experience is that smart people tend to underestimate how much harder it is for someone with that difference in raw intellectual capacity to do things they can ("No, love of my life, I am not interested in making a Python model to figure out which healthcare option is the best for us."). This was especially heightened when we were foster parents for awhile, and one of our placement was...whatever the appropriate term is nowadays for someone with a 63 IQ (which we were probably the first people in his family to find out about, since I was told I was the first person to ever show up for one of his school's IEP meetings); living with top and bottom 1% intellectual functioning was a very illustrative experience!
IMHO, a major problem with discussion about crafting programs for effective deterrence is that (especially for violent crimes-Jason Manning's review of Fragging notes how Wolfgang noted the modal murder starts with a very trivial dispute) the people who mostly need to be deterred are frankly too stupid for that to work (executive planning and IQ are decently correlated and breaks down rapidly below a certain threshold, which is higher than the cutoff for legal disability) and the people crafting these program won't get that the processes aren't likely to work with the population they are trying to fix because that population is too different in mental functioning to pay off. Our friend Joe from the Open Thread my intellectually get that he shouldn't murder someone for repeatedly taking fries off his plate (actual case Manning notes) and that he will be punished if he does so, but it will be very hard for that intellectual knowledge to kick in before he has killed the fry taker.
So incapacitation may have more value (especially, as mentioned in the comments, the age factor of incapacitation: the value of keeping people in jail in their criminally-prime years).
The IQ and time preference thing is a big thing; everyone seems to be designing systems based on their own experience, which often breaks down terribly for the left half of the curve. I remember Derbyshire commenting on how nobody ever had any sympathy for the unbookish, or realized how painful learning new things was for people who weren't into school.
As an aside, how did you land her? My impression was those sorts of women were obsessed with finding a guy who was at least as smart as they were.
"As an aside, how did you land her? My impression was those sorts of women were obsessed with finding a guy who was at least as smart as they were."
Careful application of Andrew Tate's lessons, duh.
My serious answer to this would be we both are religious (which really narrowed down her dating pool especially), both of us reaching an age/stage of life where time was a driving factor, me being better than average at knowing and accepting limitations and not caring too much about her being better than me in certain areas, and her note that "smart people are very annoying".
I do think I am an interesting case study in hypergamy, in that at every level of our relationship we were in a situation where she was the more highly educated (and frankly because of that, the higher earner, as well).
+1
Also, I think the way crimes are punished in the US is almost optimized to be confusing and incomprehensible for even normal people, let alone dumb people who stopped going to school when they were 16.
Oh yeah, your drunk-driving kid is a great example. Some people will always be surprised that "if I do X, Y happens? how can that be?" And unhappily some of *those* are "Y happened? That's so unfair! They're picking on me! The fact I did X has nothing to do with it!"
The first set may eventually learn from hard experience and immediate repercussions. The second set will never learn, nor want to learn, because it will always be Somebody Else's Fault.
You might be overestimating how much considering goes on.
Thanks, this really brings the point home.
That fits with what I've always heard. Severe punishment isn't particularly effective. What matters is quick and certain, with quick being more important than certain. This fits with various operant conditioning studies that I ran across.
Anyone who has ever house-trained a puppy should know what effective training-by-punishment looks like.
Here in the US, my husband had the experience of someone who'd committed a crime against him telling him outright, "What are you going to do, call the cops? They're not going to care."
The whole situation would have worked out very differently if there were a high likelihood of SOME consequences, rather than a low likelihood of harsh consequences.
The US has far fewer cops than most European countries per capita.
The US has about 242 / 100k (FBI figures, 2019)
Just picking a few examples, Belgium has 334, Germany 349, Hungary 367, France 422, Italy 456, Spain 534
The European jurisdictions that are lower than the US are: England-and-Wales, Sweden, Iceland, Denmark and Norway.
These comparisons are somewhat strained, given that in some European countries, "cops" include people who never leave their office and perform some sort of administrative duty that in other countries isn't performed by police at all.
For example, issuing various permits.
Czechia is notorious for having a large amount of cops nominally, but relatively few beat cops. Most of the difference are bureaucrats formally employed by police.
Interesting. I wonder how much has to do with risk aversion? I'm sort of the pathological opposite--too risk averse. The one time I decided to take more risks in life I sped right into a speeding camera.
The will/might distinction is insightful.
Though I can't help but recall a recent Matt Levine column on rating investment strategies. There's an ongoing debate between people who made 20% per year on large cap tech stocks and people who have made ~5% per year with a more diversified strategy who claim their risk adjusted returns are higher.
It is hard to convince the people who made more money that they actually made less.
Similar debates show up throughout various sectors of society.
I don't pay much attention to the FBI's statistics on murder offending (not that many statistics nerds join the FBI), but I do pay a lot of attention to the CDC's cause of death statistics where the cause is homicide.
My impression is that the tough on crime era had a big impact on driving down the homicide victimization rate (and thus presumably also the murder offending rate) among Hispanics.
It appears to me that Latinos both here in America (where Hispanics now number 65 million) and in Latin America (e.g., El Salvador, where murder rates soared then plummeted) are fairly sensitive to the messages about crime sent by the government and the culture.
When the politicians and thought leaders turned anti-law & order after George Floyd's death in late May 2020, blacks immediately started dying more by homicide and traffic fatality as the cops retreated to the donut shop. But Hispanics took longer to loosen up and adjust to less police assertiveness.
Due to their enormous increase in numbers, Latinos are now America's swing voters when it comes to America's crime rate, so criminologists should focus on what interventions persuade Hispanics that the cops are serious about crime.
Similarly, the Hispanic rate of death by motor vehicle accident dropped sharply after 2008.
All the evidence I've seen suggests that both homicide fatalities and traffic fatalities correlate in recent years. They're both related to how scared of the cops people are, so they correlate with whether people speed and whether people carry illegal handguns in their cars.
Typos
incarcerating [has the] potential to confound studies
Dhodnt should be Dhondt
"Dhondt"
What an unfortunate name to have from the perspective of auto-correcting
Sometimes written d'Hondt which I think looks better. There's a d'Hondt method, independently invented by Jefferson, for divvying up parliamentary seats according to vote share.
Jefferson invented it for divvying up congressional seats to states. These are mathematically equivalent problems, but the two separate purposes for which they were being put in the English-speaking USA and the French-speaking France and Belgium meant that it took many decades for the equivalency to be realised.
>This is potentially a good way to study deterrence, because one group of criminals (those with 0-1 strike) will get a normal sentence for their next crime, but another (those with 2 strikes) will get a long sentence.
This is not correct. As the paper says, "a criminal with one strike who is convicted of any subsequent felony (not nec essarily a strike) faces an automatic doubling of the sentence length on that conviction"
Great piece.
Maybe the wrong way to look at it, but it sorta matches both my intuition and some of the points I've seen made on subjects such as capital punishment as well - at some point that's a little lower than people tend to naturally think, we get decently diminishing returns on increasing the *severity* of punishment. My understanding is that the *certainty* of punishment matters meaningfully more after we hit a certain brightline
I don't think this universally holds, but I do think it holds at least a little bit, particularly with the almost-nonexistently-weak deterrence effects. Let's say that the max sentence for a crime is 5 years and not life - anyone thinking deeply about the cost-benefit of prison probably is deterred enough by the smallest prison term necessary to ruin their life. Anyone *not* deterred by that is either
- expecting on some level that they won't get caught, so cranking up punishment severity isn't gonna matter too much when they don't expect to serve it
- they don't care about consequences, in which case the deterrence effects are just not gonna be there no matter what you do
This would seem to support the idea of the extra cops doing more than the actual prisons - really any consequence/engagement with the legal system being certain is better for preventing crime than significant uncertainty of big consequences. But just guesses on that mechanism of why prisons are weakly criminogenic/mostly neutral despite the big positive incapacitation effect
I think there is an error in the "The Swedish data suggested a “three strikes law” should cut crime by 83%" section.
The preceding text says :
"A one-strike law (life in prison after one violent crime conviction) would cut almost 80% of violent crime in Sweden; a two-strikes law would cut 63%, and a three-strikes law would cut 57%."
So why 83% for 3 strikes?
Thanks, fixed. I was mixing up the Swedish number (83% for 1, 57% for 3) with an American number (80% for 3).
The paragraph's headline is still "Why Can’t Sweden Cut Crime 85% With A Three Strikes Law?", though.
Only the violent on a first strike basis should be incarcerated ... the rest should be compensating their victims, even if it takes a lifetime of Sundays; and if there are no victims then there is no crime. This way the current prison space can be utilised far more effectively, to keep the violent out of society for longer periods ...
Many criminals are judgement proof, lacking assets to compensate people, and don't really have much in the way of skills to do anything to compensate people.
that there are so many people in such a position strikes me as a far bigger crime than all the property crimes (that people get thrown in jail for, not the white collar stuff like foreclosure fraud that the government shrugs at) put together
> that there are so many people in such a position strikes me as a far bigger crime
Are you using the word "crime" in a literal sense or you just saying it's very bad that this is the case? I also believe the world would be better if everyone had assets and valuable skills. It'd be hard for me to call it a crime without pointing to who the criminal is. Is there a person/class of people you think should be prosecuted for this?
>Is there a person/class of people you think should be prosecuted for this?
the upper class
Give people some universal income (we can use land value tax to finance it), it will probably reduce crime in itself.
Then you can use a part of the universal income to compensate victims.
We've experimented with UBI grants. It made people poorer. https://www.youtube.com/watch?v=oyoMgGiWgJQ
... This is quite unfortunate.
Maybe we should kill the evil gods, so things can have normal consequences (like, giving money to people make them richer, not poorer)
Thank you for putting the time and effort into this.
"How much do we care about drug crime? ... Roodman says not much. It’s either personal use, which is mostly victimless, at least compared to other crimes). Or it’s dealing, which is demand-driven and almost impossible to stop through incarceration - if you lock up every member of Cartel A, that just creates a lucrative niche for Cartel B to step in and fill."
I'm afraid I'm one of the Neanderthals who go "I don't care if it has a deterrent effect or if it's not cost-effective, put these sons-of-guns in prison and get them off the streets" If Cartel B steps in, then just go after Cartel B as well.
Mandatory declaration: I'm aware this story is more anecdotal and it's based on "one guy says", but if it's anyway true, then hell yeah throw the book at Cartels A to Z!
https://www.irishexaminer.com/news/arid-41518547.html
"Young cannabis users are being spiked with heroin and pushed to take cocaine by dealers intent on hooking them to expand their market.
Children as young as 13 are being offered free cocaine samples by dealers when they buy cannabis, according to addiction therapist Michael Guerin.
As the cocaine high is short and the drug is expensive, users can quickly end up in debt and be forced to become mules, carrying large quantities of drugs for organised crime to pay back debt.
David Lane of the HSE’s Drug and Alcohol Taskforce in Cork and Kerry, has also warned of cannabis being laced with potent synthetics and said that is currently a bigger concern than other forms of spiking.
“We have seen synthetic substances adulterating other drugs," Mr Lane said.
“There has been a big change regarding nitazenes [highly potent synthetic opioids] and synthetic drugs, and that is where a lot of our concern is still. They can be more potent and are therefore more likely to cause overdoses."
Heroin use has been decreasing amongst younger age cohorts and new users in Ireland.
But many heroin users who sell small quantities of the drug to fund their own habit, have begun ‘pushing’ it on new users or spiking cannabis with heroin for unsuspecting smokers to hook them and thereby expand their client base.
“The heroin user, because they are so addicted, are under pressure to create new business to sustain their own habit,” Mr Guerin said.
“They’re spiking cannabis users with it to try to get them hooked so that they've got people to sell it to. And they’re encouraging them to try cocaine.
"The rationale behind it is to entrap them more.
“Any serious drug addict, irrespective of what they're taking, unless they have a pocket full of money, which is unusual, are involved in the sale and distribution of drugs purely for economic reasons. They have to be."
He said cannabis users as young as 13 and 14 have told him they have been given free samples of cocaine to try.
"They’re offered to sniff it off the end of a key," he said.
"They would try it and if they started to use it to any degree whatsoever they couldn't afford to pay for it, because cocaine is an expensive pastime. They then became involved in underage moving, distributing, delivering, collecting money all that stuff."
He said it is not unusual for cocaine users to say their habit is costing them between €1,000 and €1,500 a week.
Heroin dealers are also pushing the drug on cocaine users and crack cocaine users to medicate the cocaine ‘come-down’, Mr Guerin said."
Singapore seems able to crack down on drugs enough to actually reduce it. But other countries are different.
An island microstate has a number of inherent advantages in that regard.
Singapore is only techically an island; it's separated from Malaysia by a strait less than a mile wide in most places. Malaysia has similarly strict drug laws, but if deterrence doesn't work, that shouldn't matter.
It's way easier to walk across a border or drive a truck across a border or lead a mule across a border or dig a tunnel under a border than it is to cross a strait less than a mile wide.
I'm not arguing that Singapore doesn't successfully deter crime or that deterrence doesn't work. I'm saying Singapore is starting with a number of advantages. If Malaysia has similarly strict drug laws but Singapore has less of a drug problem than that speaks to something different about Singapore, and one thing that's different about Singapore is that it's easier to handle the supply side, not just the demand side, because it's a lot easier for Singapore to both police its island borders and to inderdict local production.
That sure argues for an approach to cannabis like the recent wave of states legalizing first medical, and then recreational sales and use of marijuana and related items. State-sanctioned pot has standards of regulation and testing that prevent turning it into a "gateway drug" by actually offering access to harder drugs. Likewise, real dispensaries follow regulations to the maximum extent.
The question is, how much of marijuana's attractiveness comes from its biochemical properties and how much is due to other reasons like it being a forbidden fruit or social pressure. It's possible that some spectrum of altered mental states available via legal substances would be satisfying for the vast majority and they would stop seeking illegal ones, but I wouldn't be too surprised if legalizing marijuana would just mostly result in something else mostly taking its niche and this continuing to happen with more and more drugs. Drug dealers must eat every day, after all.
I lean pretty strongly towards "doing drugs is victimless;" that doesn't mean selling it to kids is the same as selling to adults. But as is often the case with "sin crimes", the fact that the business is illegal is what drives a lot of the criminal activity. Compare alcohol during prohibition to now. The people who made, distributed, and sold it were mostly organized gangs who had to protect their property and "turf" with violence, as going to the police wasn't an option. That's not the case anymore. Marijuana being illegal is exactly what causes it to be sold by criminals, who use other crime to support or hide their business. This also makes it harder to regulate--a legitimate business has incentives not to sell to kids and not put cocaine in their weed; the local street corner thug does not.
The other thing you don't acknowledge in this comment is the monetary cost. It's easy to just say "crack down on X, Y, and Z" if it's mostly other people paying for it.
Speaking of rehabilitation, Minnesota is putting into practice an interesting experiment. Incarcerated persons (their term) for most offenses can have up to half their sentence reduced if they voluntarily engage in fairly intense substance abuse treatment and other programming and therapy. They then go out on heavily supervised probation.
The animating principle, as I understand it from Corrections leadership and their on-staff researcher (an acquaintance of mine) is that the one thing that has been shown to reduce recidivism (so ignoring incapacitation effects) is keeping people busy while they are in prison. It almost doesn’t matter with what, though therapy seems preferable. So the idea goes that if you make it voluntary, so IPs have buy-in, you make it therapeutic to help with the substance abuse issues behind a LOT of crime, and you make it intense enough so that IPs have to really want it, you might actually have a successful rehabilitation program of sorts, at least on the margins. Plus, you reduce costs by shortening sentences.
Sadly, they’re likely to f*** the implementation up by not being able to hire enough therapists and case managers to provide programming, enough probation agents to monitor them afterwards, and enough literal space to actually provide treatment at this level. So I’m not expecting much. But I appreciate the effort.
My view on the worst case for this is that the pilot programme works, then the state says "Great, we can cut recidivism! Let's expand this and roll it out to everywhere!" and then "Okay yeah but it can't cost more than X amount", so there aren't enough therapists hired, the after-release support isn't there, etc. due to "are you crazy, we're not gonna spend that much on jailbirds, the voters would riot" effect, then it all fails in the predictable ways and back to square one.
I like your blithe confidence that they’re doing this as a pilot program rather than rolling it out wholesale without a clear funding stream and hoping for the best. It speaks to your generous heart.
Literal words spoken to an audience by the Commissioner of the Department of Corrections: “We wanted to pass the bill [to reduce sentencing] while we had a window for it, so then we could go get the money for it later.”
Clearly I am too chicken-livered for the bold, venturesome local government of Minnesota! Full speed ahead and damn the torpedoes!
Is this the kind of initiative we could have expected from Vice President Walz (ex-governor of Minnesota), had the happy dreams of the nation come to pass? 😁
The Commissioner was his hand-picked guy, and Walz approved the budget, so… yes. Yes it is.
And I say this as a wild-eyed liberal in many respects.
This is likely going to be a disaster and it’s just sad. We’ve already had unrest from the IPs that it hasn’t been implemented yet, so the Department took the bold step of issuing the inmates… a memo telling them to wait.
Everything is for the best in this best of all possible worlds.
Isn't that what happened with deinstitutionalization of schizophrenics? They thought they could put them on all on outpatient meds, they cut the funding for the clinics, and they all wound up homeless or in jail?
> Incarcerated persons (their term) for most offenses can have up to half their sentence reduced if they voluntarily engage in fairly intense substance abuse treatment
OK, I stab a guy in an alley. Why am I supposed to engage in intense substance abuse treatment?
That was just one example. There’s cognitive training like Decisions Points, plain old therapy, restorative justice work, to name a few.
But an AWFUL LOT of people in prison have a substance abuse problem.
> Sadly, they’re likely to f*** the implementation up by not being able to hire enough therapists and case managers to provide programming, enough probation agents to monitor them afterwards, and enough literal space to actually provide treatment at this level. So I’m not expecting much. But I appreciate the effort.
Not just this, but drug and opiate rehab success rates, for people who voluntarily self-select into the programs, is ~30% for alcohol* and <20% for opiates.**
The success rates for heavily coerced "do this or get years of prison" populations is likely to be considerably lower (it's ~5x lower for alcohol in non-self-selected pops).
This argues that the (millions? Tens of millions?) being spent on these programs for even the poor implementation you expect are almost certainly 80%+ wasted.
* White 2012: "An analysis of reported outcomes in 415 Scientific Reports, 1868-2011"
Combined alcohol and drug abstinence rates average 30.3%, "although remission from opioid addiction appears to be less stable and durable than other patterns of remission."
https://imgur.com/a/MwtqFEf
** The best number for long term opiate relapse is Hser 2007, which has findings from 33 year followups. 18% achieve abstinence "for at least 5 years."
Hser, Y (2007) Predicting long-term stable recovery from heroin addiction: Findings from a 33 year follow up study. Journal of Addictive Diseases. 26(1), 51-60.
"But I notice that the European countries we’re talking about here all have high recent new immigrant populations which on average commit more crimes per person than natives."
Well, yeah, with the vast majority of these crimes being "having immigrated illegally"
I think you're generalizing from the US context, where migrants are less criminal than natives. In most European countries, the opposite is true - for example, https://www.researchgate.net/publication/338563093_Migrants_and_Crime_in_Sweden_in_the_Twenty-First_Century finds that 73% of Swedish murders are committed by immigrants. Immigrants are about 20% of Swedes, but most of those are Finns, and I think the 73% of murders are probably mostly committed by the 2 - 5% of the population who are Middle Eastern immigrants.
I stopped and doubled-checked what I thought I knew based on what reads like a general statement about recent immigrants, you might clarify this in the text.
I know it's true that Middle Eastern immigrants are disproportionately criminal in Europe, but does anyone know how this fact can coexist with the extremely low (peacetime) murder rates in the actual Middle East?
Perhaps it tracks with the extremely strict punishments for minor crimes that the Quran recommends? Even if most middle eastern countries aren't cutting off hands for petty theft, whatever their current policies are probably downstream of that.
European cultures are generally unwilling to go to such harsh measures for small crimes, mostly because western laws are downstream of Christian morality that emphasizes forgiveness and mercy.
I don't think anyone would dispute that cutting off the hand of anyone caught shoplifting wouldn't be incredibly effective at reducing shoplifting. All of a sudden shopkeepers have an extremely obvious sign to look out for, shoplifting literally becomes more difficult with only one hand, and generally I'd imagine such a deterrence is a lot more effective than only taking away your freedom.
I think it's more wealth than Christianity--medieval Europe was Christian and was pretty brutal.
Can this be because these people were torn from their native culture and locales where simple "neighborhood watch" prevented them from ever trying anything and put into great big cities where nobody knows them, in a culture where many things they thought forbidden are permitted, and without much money, work, or support network?
Also, there is a small matter of Quran allowing Muslim to (at least) "violate contracts" with non-believers. I'm not sure if this "believer vs. non-believer" really matters or not, but this point is often used in anti-immigration discussions, that for Muslims, Christians aren't "real people". Kind of like a pogrom against a Jew settlement used to be a "normal thing", but the same people would balk at visiting same violence at their Christian neighbours.
I was going to say "WEIRD (enforcement of norms/laws by neutral, impartial arbiters in support of a neutral, impartial ideal-"justice") vs non-WEIRD (enforcement of norms/laws by clan warfare and threat of massive blood feuds) legal systems", but Papua New Guinea has that too and it is emphatically not a place that can be described as having low murder rates, so Islam has to play some factor here (possibly by limiting access to intoxicants?)
Although one might be tempted to note that when murder does tend to happen in the Middle East, it has a nasty tendency to rapidly turn into Civil War.
"torn"
That ol' passive voice again.
That (the first paragraph) is the explanation I heard for high crime rates in Africa. Reportedly, traditional African tribes have strict discipline systems that cause their members to behave well. But when urbanization came to Africa - more suddenly than anywhere else due to the extreme cultural gradient between Africans and colonizers - all of a sudden lots of young African men were left without any of the tribal discipline, and many turned to crime. In a sense, just like Native Americans were decimated by diseases for which they had developed no immunity, Africans were socially "decimated" by social "diseases" for which they developed no immunity.
You should probably add "reported" to the murder rates. Honor killings are more or less accepted and hence won't be counted.
Probably "us vs them" mentality could explain some of the discrepancy.
Also a disproportionately large part of immigrants must come from the worst places, including literal refuges or refuge-adjacent people likely having shitty circumstances, sometimes trauma and firsthand experience with violence, both as victims and as perpetrators.
Middle Eastern countries have moderately high, not low, crime rates. Look at: https://data.worldbank.org/indicator/VC.IHR.PSRC.MA.P5?locations=ZQ. The murder rate averages around 4 per 100k. A bit lower than America's in the first term of Obama but much higher than Europe's.
The source countries of middle eastern immigrants in Germany tend to be the most dysfunctional ones, who sadly don't have reliable statistics. That said, I suspect that you're right and the crime rates of middle easterners in Europe are much higher than in the Middle East by reasons that can not be entirely explained by skewing male and young. They may be cultural effects like the interaction of their honor culture with Teutonic WEIRDness.
You went on to say this:
"So it’s possible that Europe is still adjusting to being a high-crime continent, whereas America has already adjusted by raising incarceration rates."
The thing is, I can't find any evidence of this really being the case. Overall crime levels seem to be stable or dropping in Europe fairly consistently over the past ten years - https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Crime_statistics - with no sign of the big increase you'd expect given the large influx of migrants. There's an increase between 2021 and 2022 but that seems quite easy to explain by the end of lockdowns.
There's also quite a few studies which show opposite results to the one you cited: https://migrant-integration.ec.europa.eu/library-document/crime-and-immigration-evidence-large-immigrant-waves_en, https://www.sciencedirect.com/science/article/abs/pii/S0167268123001713, or https://docs.iza.org/dp14647.pdf
I can't claim to have reviewed them all in detail or to have looked in the bias, but the relationship between immigration and crime really doesn't seem to be so clear (and certainly doesn't seem to have pushed up crime rates overall). And having visited both major European cities and American cities in the past year, European ones really do feel safer, even in large migrant areas.
Apparently sociologists say the best way to assess crime is to look at the homicide rate. In most European countries, it’s in the 1-2 range for 100k inhabitants while it has been between 4.5 and 6 in the US in the last decade. Depending on what feels like “the right amount of crime” (as much as this can make sense, at least), the trade-off can just appear different.
I think generalizing about Europe like this seems like a mistake. E.g. Germany has a homicide rate of 5 per million people per year while Russia has one of 50, and those are just the two most populous countries in Europe.
similarly, the disparity in homicide rates between and even *within* US states is tremendous. Detroit had a homicide rate in 2023 of ~41/100k. Michigan (which contains Detroit) had a homicide rate of 8.6/100k. Many states are in the 2-ish/100k range, with NH at 1.8/100k.
Yeah, it is overgeneralizing to say that Europe is now a high crime continent. We do however see pretty large increases in eg gun homicides in countries like Sweden that took in a large proportion of immigrants from war-torn middle-eastern countries (murders up 50% since 2007 and gun homicides up like 3X since 2012). But I think when you aggregate over all of Europe these trends are drowned out by mostly decreasing violent crime in much larger native European populations.
This just isn't true - the very large majority are legal immigrants. Sure, a lot of them lied about their rights to refugee status, and it would be sensible to check this a lot harder, but they are in their countries legally.
"Take a moment to imagine how your own life would change if you spent ten years in prison. Seriously, spend a literal few seconds thinking about your specific, personal life."
The problem is, some/many/lots of the people going to jail don't have spouses, jobs, etc. in the first place. Ten years in jail and then coming out to all the consequences afterwards is their life *already*, after six months in jail/multiple 'second chances' every time they came up before the court.
I've mentioned this one example before from my personal experience over my work life, but I'll repeat it once again: a girl that I first encountered as an early school leaver, and because she was engaging with various services that I coincidentally happened to work for, I could track her career over the years.
So she went from vulnerable teenager with disrupted home life, to dropping out of school, to not engaging with the support services for early school leavers, to what used to be called taking up with a bad crowd, but that's probably a disfavoured term nowadays because it's discriminatory and judgemental of persons with societal difficulties (aka being petty scumbags), to (of course) weed and other 'soft' drugs to picking up a heroin habit. Along the way she had a kid with, of course, no husband and the boyfriend flitted in and flitted out.
End result, she stabbed another young woman in the stomach at a house party (drugs and drink present, of course) and ended up with a prison sentence while her mother took care of the child.
She never *had* the "my ordinary normal life of a job and family and the rest of it will be disrupted by going to jail!" In fact, I think prison probably was *better* for her as it was a chance to break those exact connections that were dragging her down, and letting her get sober and clean and maybe, who knows, a miracle, getting her head together and getting skills and help for post-prison life.
There are other people who die early and I see the names in the local paper and I go "Oh yeah" because again, I know the background, and I'm not surprised because this is all the inevitable ending to the life they lived.
I've encountered some of the early school leavers on the programme for such, and I've forecast even at that age (late teens) they're gonna end up in jail, because the little idiots don't give a flying fuck about learning any kinds of skills leading to getting a job and a normal life. That's for the squares and the fools, they want weed, easy money, and free time. They're on the path to petty crime if not already started on it, and one day they'll go too far and either end up stabbed by a bigger, badder criminal or they'll finally run out of "second chances" and have to do time.
And it'll be the best place for them, because they *will* be off the streets and not interfering with ordinary people. They're not interested in rehab even in or after coming out of jail, they'll fall back into the same old lifestyle because all they are interested in is drugs, easy money, and free time, and six months or ten years is all the same.
Re: the shoplifting, I think *all* the commentators are right; Graham about the view of the cops on the ground, Andrew about the "scared straight" but I imagine that for first time offenders or people just starting on the petty crime path, and CJW for how the sausage is made in the courts and legal systems.
I recall there being lots of fears that the lack of fathers around due to overincarceration was going to make their sons more likely to be criminals. Then a study showed it actually made them less likely because these fathers in particular were such bad influences. The average criminal is just not the average person.
I've been listening to A History of Rock Music in 500 Songs and it's striking how many of the early rock stars (it's only up to 1968-9 so we're not really into later ones) come from exactly this sort of chaotic background. One reason for the "27 Club" type early deaths is that if you take someone like this and throw enormous amounts of money and fame at them, they end up taking huge quantities of drugs and dying. Sometimes you get a Keith Richards who eventually does straighten up his act and lives.
> Sometimes you get a Keith Richards who eventually does straighten up his act and lives.
And I've personally always wondered - why aren't we cloning Keith Richards' and Ozzy Osbourne's livers and kidneys and studying them?
Or cloning them en masse and transplanting them into everybody needing transplants? They pretty obviously have some robust hardware there.
They could probably be persuaded by the humanitarian mission aspects of it.
Ironically, near the other end of society, I'd be most concerned about contracting HIV in jail through rape--without having to pay for anything else I could just leave the stocks growing in the brokerage account. When they let me out I'd just retire.
You wonder if some calculation like this is driving wealthy criminals like Weinstein.
My mom worked as a teacher in a juvenile prison (for serious offenders). She and the rest of the teachers all honestly wanted to help those kids, but also expected that nearly all of them would just return to a life of crime when they turned 18 and were released.
Incapacitation effects depend upon the birth rate. If there are lots more young people than older people, as there was in 1970 when the criminology professors' orthodoxy hardened, then incapacitation won't do much good. However, in 2024, incapacitation is pretty effective.
Seems unlikely. The ratio of let's say ~25 year olds (peak criminals) to ~35 year olds (average employed person) wasn't really so different in 1970.
This was excellently done, thank you. Overall I am not sure why GPS monitoring hasn't yet taken more market share from prison sentences.
I keep hiring about all the crimes committed by people on parole/probation, so I'm surprised to hear "over half of people incarcerated aren't there for any crime at all". Shouldn't there be lots of low-hanging fruit of people to lock up because they're actually committing crimes while out under supposed supervision?
This system is so messed-up
It sounds like we've made trials way too expensive if we have people racking up so many arrests. https://marginalrevolution.com/marginalrevolution/2023/04/misdemeanor-bail.html https://x.com/robinhanson/status/1116753961329905665
thank you for this very interesting comment
You allude to it twice but then pass over it "for simplicity": None of these studies measure within-prison crime, which is largely not counted, tracked, tried, or prosecuted. It is not only plausible but likely that "incapacitation" is at best simply relocation of crimes. And even _that_ is leaving aside the regular drumbeat of deeply sadistic crimes by guards, which are even less counted, tracked, tried, or prosecuted.
The quality of statistics on crimes within prisons is very poor, of course; ttbomk it is not practically possible to know if the average prisoner commits crimes at a greater rate while incarcerated than while outside, or likewise the rate at which people in prison are crime victims versus those outside.
If your value function is some sort of utilitarian sum over society, you have to count those, or else your utilitarianism is just gerrymandering people into and out of the boundaries of the utility summation (in which case you can optimize utility much more simply than by incarcerating; simple Schmittianism will do the job more easily, and it's free since you don't count the cost).
Surely we can at least observe the homicide rate in prison?
Agree, I think it's not a valid simplifying assumption to ignore crime in prison when assessing incapacitation, good statistics or not. Incarceration obviously incapacitates some serious (and non-serious) property crimes but makes it a lot easier to commit assault. There seems to be some good data about sexual assault, which seems to be higher frequency in prisons than outside, but not finding it for the (much more common) normal assaults, attacks with shanks, etc.
But those victims are also convicts, so we don't care.
Hello. Criminal defense attorney here.
As I said on a previous post of yours, I don't think the tough on crime crowd realizes how many resources the US system is *obligated* to spend in the case of a criminal prosecution (and you can't have jail or prison without a criminal prosecution).
If you charge someone with shoplifting and insist on 30 days in jail, the defendant can obligate the system to provide
- a lawyer
- a jury of 6 (in minnesota) or 12 (if you're going to impose prison time)
- a judge
- a court reporter
- a prosecutor
- a cop witness
...several of these for multiple hearings (you'll appear before a judge using your lawyer several times in the course of a criminal case)
additionally, that same defendant can obligate the *victim* to provide
- footage of the incident
- a store witness
all for a $5 pack of hamburger patty (yes, I have seen a shoplifting trial on a $5 pack of hamburger patty)
...and at the end of all this, even if the conviction sticks, you still have to sell a judge on sending some poor guy to jail for *thirty days*. Not that it can't happen, but it's such an immense expense of time and effort for everyone involved that it's usually much easier to cut a deal.
We can have juries of smaller than 12 for misdemeanors but a criminal case always entitles you to a jury (and a lawyer).
There are a lot of other points of friction in a courtroom too: to prove a claim of shoplifting, the prosecutor can't just wheel out footage of the incident and show that footage to the jury and say "enough said, please convict." The rules of evidence require things like videotapes to enter through a witness, that witness has to be subject to cross examination, and so on. A shoplifting prosecution would take an afternoon even if the lawyer for the defendant wasn't incentivized to draw the process out.
Where I live, there has been an extended effort to let into the country members of those groups that are relatively higher in the commission of crime, for the past - let's see, when was Reagan - forty years? Against this headwind, how or why would you expect to be able to reduce crime via - prison or any other means?
Excellent point
My understanding is that immigrants commit crimes at a lower rate than the native population, so I think your premise is just plain wrong here.
I was referring to illegal immigrants specifically in terms of their crime rate (other than, obvs, being “illegal immigrants”) being lower than native born people
Native born does most of the work in that sentence, given the birth rate of certain immigrants.
Of course, the fact that police nowhere ask about citizenship does the rest.
Are you in the US? Don't immigrants commit less crime on average than natives here? (see https://www.americanimmigrationcouncil.org/research/debunking-myth-immigrants-and-crime)
I had originally thought that the impression that immigrants commit more crime than natives was because white people were using their own neighborhoods as a calibration point, and immigrants commit more crimes than whites, but less crime than the (white + black + etc) national average. But https://siepr.stanford.edu/news/mythical-tie-between-immigration-and-crime says that immigrants even commit less crime than the average white!
There is one major confounder in those studies, which I haven't seen any of the authors try to address (though it's been a year or two since I looked closely at it): a high percentage of criminal immigrants are removed from the country after committing sufficiently serious crimes, while U.S. citizens obviously aren't. So we'd expect immigrants to have a lot fewer opportunities to commit crimes, and American criminals to have a lot more chances to end up going to jail (which, if I remember correctly, is what the studies were measuring).
Median age of an immigrant to the US is 28. Not many 28 years olds want to join a youth street gang.
Counterpoints:
https://www.realclearinvestigations.com/articles/2024/11/25/illegal_migrants_less_likely_to_commit_crime_guess_again_1074276.html
https://cis.org/Report/Misuse-Texas-Data-Understates-Illegal-Immigrant-Criminality
In a nutshell, it's hard to say for sure whether illegal immigrants (LEGAL immigrant crime rates may be easier to track) commit more or less crime than natives, because studies only count those who've been processed, and a criminal who commits multiple crimes only counts as one.
I'm blown away by the thoroughness and tenacity with which answers to fundamental societal problems are pursued here. This may be a long and dense read, but everyone with an opinion on incarceration should make time for it.
One has to be a particular king person to read this article. A super majority of people are not that kind of person.
> This may be a long and dense read, but everyone with an opinion on incarceration should make time for it.
Why? That seems like the least likely audience for the piece; people who are interested in the issue will already be familiar with everything this covers.
Lots of people have opinions on topics on which they are uninformed. Plus, the exhaustive nature of this particular piece likely has new information even for active consumers of incarceration articles. It's hard to imagine a person (apart maybe from the researchers themselves) who looks at this and thinks, "I already knew all that."
> Lots of people have opinions on topics on which they are uninformed.
Those people are especially unlikely to get anything out of reading this piece, even if you could somehow compel them to.
"People with opinions on incarceration" likely includes a lot of people whose current information about the issue was curated by people less evenhanded than Scott.
And that audience often leaves killer comments that will get engagement from Scott and the rest of us, to everyone's edification.
Yeah, welcome to the club!
I'd recommend going through his back catalog on Slate Star Codex, he's a little busy with the kids right now.
What if the question isn’t how to reduce crime, it’s how to reduce the impact of crime on productive members of society? If so, punishment is unlikely to be the most effective approach.
Why? I'm not following.
What’s the outcome we want?
Focusing on reducing impact opens the conversation up. Take the reconciliation practices used in some places. Those have nothing to do with punishment. They are about healing and moving on.
There isn’t healing from petty theft but we could offer people more support. Take a bike theft as a trivial example, nobody gives help the victim. They spend resources to hunt the criminal, if anything, then more resources trying to punish them. The victim is left with zero. As a society is this the right outcome? I am not convinced.
Some baseline crime will always exist in large societies. I’m merely questioning that resources could be focused on helping the productive more directly.
That would incentivize people to pretend to be the victim of some crime for financial gain
Which would have to be gated by spending resources on a process of "approving" someone as a victim, but then you're spending resources on something which isn't reconciliation for the victims, and isn't disincentivizing anyone from offending (if you care about disincentivizing anyone from offending)
---
If you're trying to spare the criminal, I think it's possible for crimes to be neutral or a net negative to the offender (instead of being zero-sum), in which they make gains too if crime is disincentivized
& Even when offenders are more harmed than helped by measures taken, I would think those measures taken (often punishment) and the resources they cost would be justifiable because of the amount of people casually benefiting from living in a safer and more secure world.
The current baseline crime in modern, large societies is what it is partly because of measures taken to catch criminals; and personally, I think there is a limit to how high of a crime rate I'd be willing to tolerate, and once over it I'd find the world a very undesirable place to live in
I follow you but think you’ve got a few bits off.
My read of the stats and studying drug rates particular few people are truly bad, most prefer to live a better life but don’t see it as possible. Crime happens for a bunch of reasons, I think lack of community attachment is a good place to start. That is not solved by the stick/trying to disincentivize.
We already spend money evaluating victims, that’s the bit about investigating the crime then deciding to pursue perpetrators.
I am not expert here, rather the policy discussion circles around levels of punishment with very little discussion of novel ideas.
It’s disappointing and revealing to me that in your list of costs and benefits you seem to forget to even mention the victims of violent crime, who get really messed up for a long time when it happens to them. Or worse, dead.
The costs to the victims was the primary analysis:
> Roodman uses two methods: first, he values a crime at the average damages that courts award to victims, including emotional damages. Second, he values it at what people will pay - how much money would you accept to get assaulted one extra time in your life?
The list of costs and benefits you're looking at specifically only includes the remaining things people care about beyond the cost to victims themselves.
"Second, he values it at what people will pay - how much money would you accept to get assaulted one extra time in your life?"
This is a question so abstract as to be meaningless. It's no different than "How much would you have to be paid to step into the ring with an in-his-prime Mike Tyson?" It's a very easy question to answer, but it's the answer to a proposal for a voluntary transaction, and the answer is less than significant for people who have never had the shit beaten out of them and have only watched crime on television.
"How much money would you accept to have the shit beaten out of you and raped and left bruised and cut up in an alley behind the Howard Johnson's?" is a question that can only be meaningfully answered by people who have had that experience.
This is hard to evaluate, but it would not be fair to say he didn't even mention the victims of violent crimes.
Also it seems way more important to put resources on that, than shoplifting or theft.
And the bottom line is ....?
See the section at the end titled "Summary"
My idea after reading this:
1- Make (as in, with a law or something) certain percentage of the population should be incarcerated, no less and no more; let's say 0.4%.
2- Make a huge excel sheet of every criminal in the country with all the original sentence amount totals.
3- Sort by sentence totals
4- Put in jail the amount of criminals that comes up to 0.4% of the total population.
If the total population of the country drops by, say, 1000 people; release the 4 people with the least amount of total sentence from the prison. If population increases, say, 1000 people by immigration or birth, catch and imprison the next 4 people in the list and imprison.
When somebody completes their sentence and leave prison, or die in prison, arrest and jail the next person who is out on the sorted excel list (in fact, maybe tail the top 100 or so people in the list which are just out of the top 0.4% because A- you might need to put them in soon when somebody leaves or the population of the country increases B- they are probably the most likely people to commit crimes and climb in the top 0.4%)
In this way, the "toughness" of the prison sentences becomes a mostly moot point and only relevant in relation to each other. I think this would also be good media! TV shows about the people about to get released whenever somebody on the outside do some more crime and push them off the list; or the coverage of the manhunt when somebody is released and the next person in line have to be caught and put into prison. That would be great entertainment for all!
Lastly, an unrelated idea about the shoplifter issue. If the problem is the court sources being exhausted for this kind of minor crimes that happen often; why not combine all of them into one (or a couple of) huge organized crime trials? Instead of trying to try 300 shoplifters one by one; see if they have some common people that they sell their stolen merchandise to; make a relation map and in one big organized crime trial, try all of them as one monolithic gang. Use the court resources smart! Added benefit is, organized crime probably has a lot longer sentence than simple shoplifting so they'll be out of the community for longer!
God I hope I was the person who was in charge of deciding everything for the world.
This is terrible, inhumane and would make for a great reality show.
gotta keep the masses entertained, right? maybe that's a big enough plus point to make up for all the inhumanity of the proposition :)
Seems like prison capacity / overcrowding is an issue too: https://en.m.wikipedia.org/wiki/Prison_overcrowding_in_the_United_States
This likely will alter the cost equation - need to factor in cost of expanding capacity. In the case of NY, I’d be shocked if there wasn’t capacity to incarcerate the 327 shoplifters, but maybe applied more broadly it’s a factor in choosing the types of crimes to allocate prison capacity towards
Overcrowded / resource-strapped prisons also probably have a lower impact on recidivism
Something's gone wrong here:
> (maybe “life sentences” should end at 65, since incarcerating potential to confound studies
Thanks, fixed.
my major takeaway: we need more controlled experiments.
So i vote for making "enforced participation in a controlled experiment" one of the possible punishments.
"why we can’t punish repeat shoplifters at all, with short sentences or long sentences." The impediments to arresting and convicting criminals are intended to prevent Type I errors--convicting the innocent. But we probably could identify procedures that have almost no marginal impact on reducing type I errors but result in very large marginal increases in Type II errors--letting incorrigible criminals go. Eliminating such procedures would be a win.
To adjust my views on prison, I'd need to see how it compares to other methods of reducing crime. As a start, what's the marginal cost/benefit of the things you hear from your liberal theater aunt? In order of squishiness/relevance: how relatively effective would it be to increase access to jobs, make restorative justice more available to juveniles, invest in rehab centers, create after-prison reintegration programs, incentivize early parenting education (e.g., NFP in NY), build more food banks, deliver on UBI, remove the millions of service lines that still have lead pipes, give mandatory therapy to non-violent criminals, plant more trees in poor neighbourhoods, ship criminals to Australia, and make them watch Turin Horse... you get the idea. The last three are a joke, I think.
Let us know what you find out.
What do you really mean by "increase access to jobs," though? If you're forcing major employers to take on more ex-cons, they'll see that as a cost and take steps to minimize it, probably involving treating their employees worse than before.
You didn't list "hire more cops" and "spend more on the court" in your list. They might not be things your liberal aunt says but that's what Scott recommends instead of increased sentencing (and the commensurate cost).
What is your reaction to those two things?
I'm in favour of whatever is likely to address the issue in the least harmful way. It's just a matter of putting everything side-by-side and making a reasonable assessment based on the available data, and I'll put my thumb on whatever scale works (within reason). I'm not against hiring more cops or spending more on the court if that meaningfully reduces violent crime or helps deal with the few NY kleptos, etc. But say someone is doing a deep-dive comparison between amputation and bloodletting for gangrene, and it turns out antibiotics can solve the issue in most cases, I'd sure want to know! I'm not suggesting the analogy is appropriate here, but if I'm going to argue for something like having more guns on the street or locking people behind bars, I'd want to know how effective they are compared to other methods that are on the table. Just basic utilitarian thinking. (NOTE: I enjoyed the article and am in no way criticizing the framing here, I'm just thinking out loud about how these solutions proposed compare to other ones. I appreciate your question.)
I think the poll is done the wrong way around, in a way that should be economically neutral but probably psychologically is not.
Instead of asking: "How much money would you accept to be assaulted an additional time."
Ask: "You have already been assaulted. You live in Ancapistan, so you have to fund the imprisonment of your attacker. How much of your money do you dedicate to keeping them in prison?"
I think this works less well because most victims of crime are poor people who don't have money. Also, keeping your specific assailant in prison doesn't reverse your assault! I would spend lots of money not to be murdered, but keeping a murderer in prison would be a social (not personal) good and get much less money from me.
Hm, now I'm curious. Maybe it's worth to ask both.
> I would spend lots of money not to be murdered, but keeping a murderer in prison would be a social (not personal) good and get much less money from me.
You might be surprised. I predict that, despite what you say here, that would actually depend on who the murderer had murdered.
This is way off. I would spend vastly more to avoid being a victim of crime than to imprison the perpetrator after the fact- that does me no good. In fact in many scenarios I would spend literally nothing to prosecute or imprison the perpetrator (why should I?- locking them up doesn't benefit me unless I think there's a material chance of the same offender going after me again).
Except in a few cases (domestic abuse springs to mind), this will be the normal state of affairs. Punishing a criminal has benefits, but usually almost none of them acrue to the victim of the crime; they are distributed across society.
Right. If some dude breaks into my house while I'm gone and steals my stuff, the benefit to me of putting him in prison for a few years is very small--it's not likely he was going to hit my house again. But it's still a net win overall--everyone has their probability of getting burglarized go down a bit.
I get what you're saying about psychological non-neutrality, but your proposed reformulation seems economically incoherent. The amount I'm willing to pay to keep someone in prison is almost certainly much less than the pain-and-suffering cost I sustained from the crime itself; I'm not going to devote a huge amount of my own resources to getting revenge, or to protecting other people from a criminal who isn't especially likely to re-victimize me personally.
One other comment I want to throw in here. Hang around for a while with correctional workers, at least in Minnesota, and you will eventually hear someone talk about “evidence-based practices,” usually right before they spit at the ground. Coming from secondhand medical background, this confused the hell out of me for a solid six months. A) In medicine, EBP is a set of practices designed to make sure that commonly used interventions, you know, work. B) Evidence seems… good? Not here. A new director might be described as “super EBP” with a roll of the eyes.
I eventually worked out that in practice, EBP represents a rehabilitative, “soft on consequences” approach in prisons that a lot of front-line workers, especially older ones, think is ineffective at best or just plain too kind to offenders who deserve worse. So there’s a major divide between Central Office people who are trying to rehabilitate the inmates based on “best evidence,” and the guards and case managers who think that time should be hard. It’s a pretty good microcosm of the current epidemic crisis, actually.
There's this disconnect in general where people who've never in their lives seen a hardened criminal up close get to decide policies, and even the whole philosophy of dealing with crime, which ends up in a predictable clusterfuck. (There's a similar issue in psychiatry, from Scott's telling.) Essentially this means that society is still too immature to understand these issues on the institutional level.
I think this generalizes to "people who've never in their lives seen X up close get to decide policies...which ends up in a predictable mess" for most values of X. Education--often the people deciding the curriculum and policies aren't the line teachers who have to deal with the effects. Business--the higher-ups make policy and buy software, the grunts on the ground have to implement it. Basically, as soon as the decision-makers are separated from the doers by too many layers, bad things happen.
I think this view of deterrence is fatally flawed.
When a young man decides to go to dental school, he doesn’t do a careful EV calculation of the RoI of that choice over his anticipated working years. He just notices that dentists seem to have it pretty good, so maybe he should go for that.
Suppose tomorrow there is a subtle change to dental insurance law that causes dentists to start making about 10% less per hour worked. Is this kid going to immediately notice and change his mind about dental school? That seems very doubtful to me. He’ll still be a dentist, but maybe he’ll end up regretting the choice, and maybe 20 years from now, kids will have noticed all the dentists with regrets, and over time you’ll gradually start to see falling enrollments at dental schools.
I’d guess deterrent effects on would-be career criminals will actually mostly work via the same slow and indirect mechanisms, and none of the studies you described would be able to pick up on these effects at all.
I think this is true in the sense that criminals just go off a general gestalt feeling of what the punishment is, rather than the actual punishment. But I think this is relevant to the question we're asking - if the government increased sentences, criminals wouldn't know, and it wouldn't deter them.
We agree that would-be criminals mostly don’t know the laws on the books. But they do know about their cousin Joe. They know Joe has poor impulse control and commits lots of crimes when he’s out of prison. They know roughly what fraction of his life Joe spends behind bars. Suppose Joe’s life outcomes affect their assessment of whether the criminal lifestyle in general is likely to be a rewarding one.
Is this something that could be found in the studies you are using to assess deterrence? I think not, and I also think this kind of folk wisdom mediated deterrence is likely to be the most important kind of deterrence.
That would mean that increasing sentences would work, just with a delay, and not in a way that you know how to measure.
It didn't take long to liberalize the system in the 1960s toward more crime, but it took the 1980s and part of the 1990s to change the culture back toward less crime. This would suggest to me that elites should be very cautious before deciding that we have too much law and order.
Or, how’s this one: Suppose some Brazilian farmers don’t understand that their traditional cassava preparation techniques eliminate cyanogenic glycosides. Then, we could logically conclude that giving them a new strain of cassava without cyanogenic glycosides won't affect how they prepare cassava tomorrow.
Fair enough for tomorrow, but long-term?
If you largely agree with this, then Steve Sailer's point above in the thread about the 10-20 year lag in "soft on crime society" vs. "tough on crime society" permeating culture should also become more plausible.
The last sentence here is wrong, and I think "splendric" is spot on (I was waiting to see if someone would make this point).
Take a step back and look at the result that "deterrence effects are small" estimated (in the good papers at least) from local variation between two sets of would-be offenders. A common interpretation of "deterrence effects are small" is that it stems from "the average criminal is not exactly a genius and does not logically account for long term implications of their behaviour / think about the exact consequences of law breaking" - you can't expect to deter them by threatening consequences because they don't think about them. You or I are much more capable of making this sort of calculation / consideration of future consequences, but we're not the ones committing common crime.
But if that is true, it basically means they are particularly unlikely to attend to institutional quirks (of the legal system etc) that all of these studies exploit to estimate deterrence. E.g. if someone with two strikes faces much higher punishment for an additional offense than someone otherwise similar with one strike, this distinction won't achieve very much, because the would-be-criminals won't attend very much to it (If at all).
But it's much easier to attend to "the general vibe of a thing" than to "precise institutional details". Do criminals deeply consider how many strikes they are sitting on, or changes in sentencing laws, before deciding whether to commit crime? Perhaps not. But changes in laws / becoming more/less "tough on crime" should (often with lag) change the general gestalt feeling of "is society cracking down on crime", both because vibes are rapidly transmitted in the zeitgeist plus there is social learning. And it is much easier to respond to this, and in fact people surely do (e.g. when I decided what degree to study, I had a rough understanding that my field pays well and this presumably informed my decision, rather than knowing "the exact causal effect of doing major x instead of major y on lifetime earnings for someone with my demographics"). Which in precise academic speak means there will probably be a violation of SUTVA - you introduce a law that affects group A and not group B (or effects them differentially), but because of the change in the zeitgeist, both groups change their behavior. Thus studies that isolate differences between group A and group B will underestimate the effects of said laws. And the more you think "criminals are too dumb / incapable of attending to complicated incentives to change their behavior in response" the greater the fraction of the overall effect you should think happens through the zeitgeist channel, and thus the more you should think that these studies can't capture the true effects. And thus the more credence you should give to the possibility that the big crackdowns on crime in the 80s did produce the cratering of crime rates from the early 90s etc. The post-GF "defund the police" homicide spike is a great example of this - as BLM apologists like to point out (in trying to argue that Ferguson effects can't be real), not a lot of defunding actually happened - certainly not enough to explain a 50% increase in the Black homicide rate. Yet we saw a huge spike in the crime rate because overnight the message was communicated that "it's open season to commit crime and police should be afraid to step in" (In a slightly different "Ferguson effect" example, Roland Fryer tells a story of gangs getting junior members to openly break gun laws in front of police to verify that they won't step in anymore, before the gang would launch a gang war).
When society eases up on law and order, criminals notice quickly. E.g., Chicago's worst day ever for murder with 18 homicides was six days after George Floyd died. When society toughens up, it can take years for criminals to get the message. Implication: be very prudent about loosening up.
> But I think this is relevant to the question we're asking - if the government increased sentences, criminals wouldn't know, and it wouldn't deter them.
That's only relevant if the policy change you're considering is "we make sentencing harsher for a week, and then we undo that change and put it right back to the original level of harshness".
If you want to know what effect a permanent change will have, then you can't assume criminals won't know about the change, because they will know. When sentencing changes, the gestalt feeling of what sentencing is like will also change.
Three times in my lifetime, the zeitgeist shifted toward less law and order (1960s-70s, Ferguson, George Floyd) and crime rose. The first was more pervasive and took longer to roll back because it involved more changes, such as shorter sentences. The latter two could be reversed in a few years because they mostly involved less policing, so the politicians telling the cops they were back on their side did the trick.
Also, while criminals tend to be pretty dim, cops tend to be of average intelligence. Cops pay attention to the news so the Ferguson Effect and the Floyd Effect seem to have been due in sizable part to cops noticing that politicians, the press, and other influential members of society wanted them to police less aggressively. So they retreated to the donut shop.
I sorta feel like 'dental school students' and 'criminals' are two very different populations overall.
Thinking about it from first principles, it seems intuitive to me that deterrence effect of longer incarcerations (for first and repeat offenders alike) would be negligible, due to scope insensitivity: that the brain just can't multiply. People can kinda visualize what a night in prison would feel like, but why, among all the other domains where research shows people cannot, would they be able to correctly weigh the disutility of of 30 or 365 or 3650 nights in prison? You'd be lucky if they can evaluate it logarithmically.
Which kinda gestures towards a class of alternative solutions: what ARE the sort of things people can tangibly feel? How about corporal punishment (or, let's not beat around the bush: torture)? I don't know if there's scholarly consensus among biologists, but from what I gather, pain (emotional pain would also suffice, but is more difficult to administer) is an adaptation to efficiently train individuals to avoid harmful stimuli, and imagining the peak intensity of various forms of torture would be very vivid in the minds of potential offender (while the intensity of imprisonment is similar to waiting for a delayed bus, only suffered through prolonged stretches of time), and so, from first principles, I would expect it to have maximum deterrence effect. Moreover, unlike imprisonment, it does not destroy individuals' social lives, and just like one can try to estimate the disutility of e.g. robbery at $25,000 as was done in the article, one can try to estimate the disutility of torture, and I would wager criminals themselves would prefer torture to imprisonment.
Is my intuition completely wrong with this? When (Western) societies abandoned corporal punishments, had they concluded that these punishments don't, as a matter of actual fact, work? Or was the decision made on purely philosophical grounds of rights-based rather than duty-based moral theories, and aversion to any form of coercion and violence (which nowadays manifests itself in effective permission to commit atrocities ranging from schoolyard bullying to megadeathcrimes, as relevant authorities tie themselves to knots avoiding any actions that would Actually Work).
I'm all for corporal punishment, but it lacks the incapacitative effect of imprisonment, which seems to be the main positive effect (on the margin, at least). My preferred solution to that is the death penalty for anyone who needs to be incapacitated, but that seems to be pretty far outside the Overton window.
Seemingly, incarceration in a low-cost prison (e.g. more or less permanent solitary confinement so little guarding needed) would be vastly superior to just killing criminals of all crime levels.
Death penalty requires a higher standard of proof (rightly so, I think, given how much harder a wrongful conviction is to fix) and as a result is in many cases far more expensive overall than just locking someone up indefinitely.
I could support such a system only if the prisoners were given the option of euthanasia. Forcing them to live indefinitely in such a state seems unreasonably cruel.
The very fact that it's over with quickly might make it more appealing to prospective plea bargainers - and arresting officers - than a drawn-out trial and appeal process, thereby reducing prosecutor caseload, thus improving swiftness and surety, which could plausibly be even more important than the incapacitative effect.
> it lacks the incapacitative effect of imprisonment, which seems to be the main positive effect (on the margin, at least).
But it also lacks the "prison as a crime school" effect, which is apparently pretty strong for beginner criminals. If the first couple of sentences would be low, replacing them with (reliably guaranteed) caning might be a serious improvement, on the current margins.
I agree. If someone is unlikely to reoffend after a caning, they don't need to be incapacitated. Probably many of them should be monitored via ankle bracelets, however.
Considering how criminality varies by sex and age, it also seems plausible to me that hormonal treatment could have a rehabilitative effect, so that might be an alternative to incapacitation.
> Which kinda gestures towards a class of alternative solutions: what ARE the sort of things people can tangibly feel?
Probaby: "I just saw a cop car roll by here five minutes ago. Better not do crimes for a while." Which isn't all that expensive an experience to create, compared to many of the alternatives on the table.
The age-crime curve is a big reason to think that the criminogenic effect of prison is to reduce crime. It's roughly as simple to understand as the incapacitation effect.
If we used crime vouchers, insurers would figure out the most efficient way to reduce crime. https://www.overcomingbias.com/p/who-vouches-for-youhtml
I think that gets counted under incapacitation.
Well, if a 20-year-old kills someone, and you let him out of jail when he's 60, you get 40 years of incapacitation. Most of the crime prevented by that stretch of incapacitation will have occurred during the first 20 years, for age reasons.
But also, after you let him out, he's unlikely to commit crimes because he's old. It seems difficult to count this as "incapacitation", because he's not in jail. But it's very relevant to the question of "after this guy gets out of jail, will he be more or less likely to commit more crimes?". He'll be less likely.
That's what _I_ thought we were talking about with aftereffects, at least.
"An optimistic school of thought hopes that stricter sentences will decrease crime after release. After all, the longer you’re in prison, the longer you have to be scared straight, or find God, or get rehabilitated.
A pessimistic school of thought suspects that stricter sentences will increase crime after release. I find this one more plausible."
What does seem to happen reasonably often is that maybe sometimes in their forties, a lot of career criminals get tired of this shit and yo-yo:ing between prison and a criminal life on the outside. It seems to be neither being scared or being rehabilitated, and instead just realizing that this is a crappy life, actually worse than having a boring 9-to-5 job.
An amount of prison is probably important to make this happen. It also means that a lot of the time, you don't have to lock people up for life - only until 45.
If we start to kill people on the grounds that they have a net negative contribution to society (economic or utilitarian), we will get a _very_ different society, and probably not for the better.
Wrongfully convicted folks definitely don't win in that system relative to the current one, and there'd be a lot more temptation to frame someone (especially when already "in for a penny" - I mean, what's a judge gonna do if he catches on to the evidence tampering, have you killed twice?) so the rate of wrongful convictions would presumably increase.
I know you don't like to apply your own conflict theory and ascribe self serving motives to large groups <snark>other than AI risk doubters</snark>, but it's high time you start considering whether there are whole fields of research dedicated to reinforcing pet conclusions about major issues.
We know for damned sure that this is possible. The cigarette industry did *fifty years* of research proving that cigarettes were safe. We now know that that research was all bunk, but for a very long time, it was treated as reasonable, well meaning, rigorous work.
I won't say which side of this particular discussion might have such a self serving bias, but I think you should consider whether there is such a side and then chase it down in your customary thorough and detailed manner.
I would love love love to see an ACX post titled "Is One Side of the Crime Debate the New Tobacco Industry Science?"
(For those who don't know Scott's conflict theory vs mistake theory, it's at the old blog: https://slatestarcodex.com/2018/01/24/conflict-vs-mistake/ )
I went through lots of different studies, compared readings by people on opposite sides, and solicited feedback from people on both sides. I discussed which parts everyone agrees with and which parts are controversial. I also included the cost-benefit analysis which lets you do a sort of sensitivity analysis if you want. I don't know what else you want.
See also https://slatestarcodex.com/2019/07/17/caution-on-bias-arguments/
I want you to consider whether one or the other side is entirely constructed and framed to service a desired answer. This is related to but different from attempting to find the true statements inside each side's frame.
Consider the possible conclusions of what you've done: "finding A seems substantially true". "finding B seems directionally true but I have concerns about the methodology". "I'm not convinced that finding C was reached fairly". All good conclusions.
Compare the possible conclusions of what I'm asking: "I have looking into the community and history of sides A and B and cannot really find anything that makes me think they were or are anything but honest academic communities with reasonable self admitted biases". "After looking into the history of side A, I believe the academic community around it was built in a very self serving and motivated manner, ala Phillip Morris, and while there are many well meaning people on side A, they are working in a structure that's built intentionally to serve a specific outcome". "I am surprised to find that the academic communities around both side A and side B were built dishonestly".
To the modern ear, the latter set of conclusions sound harsh, maybe even paranoid. But there are real examples of motivated research communities constructed to reach motivated conclusions. Foods. Pollutants. Politicized departments in scary 20th century countries.
I am aware of the theory that the only reason anyone is ever tough-on-crime is because they're racist and the whole field of criminology was founded on racism from the get-go - but I'm just so sick of this kind of argument. "Oh, I can think of one reason the other side might be biased, therefore they're all intellectually bankrupt and you can safely ignore all the evidence they produce." I'm just sick of this. Everyone on both sides always believes something like this and they're always trying to convince me I'm dumb for looking at the evidence and taking it seriously, and I've stopped caring and I'll stick to that.
There are a couple of issues where I genuinely believe one side is operating in bad faith - tobacco companies, race/gender, etc - but I try to have a really high threshold for believing it so that people aren't constantly sniping me with "but the other side is evil so you have to dismiss them out of hand right now!"
Cards on the table, I actually believe the *soft-on-crime* side was wrong from the get go.
And maybe, maybe to a lesser extent the tough-on-crime side was rotten, but if so, a far lesser extent.
I haven't done the work, and won't because it's not my job or passion, but I think if you did the work, you would find that the soft-on-crime side was in the dead center bullseye of the motivated, flawed psychology research criticized by John Ioannidis, Richard Feynman and others.
A more interesting question perhaps is, why does one believe one side is operating in bad faith? Why do I believe soft-on-crime is done in bad faith? Why do you believe race/gender is done in bad faith?
I believe soft-on-crime was in bad faith for two reasons. First, I've lived in a lot of places (at least 9 US, plus 2 international), and like anyone else I have friends and family from a few others, and anecdotally I see a clear picture: the scary draconian places have women walking around at night, and the cool progressive places have women not doing that.
California in particular is the worst offender. My personal experience includes SF knife wielder chasing a friend, SF random serious head wound guy, SF car breakin, LA public masturbator, LA car breakin, San Jose light arson, San Jose car breakin, San Jose deadly overdose on playground, San Jose overdose in apartment hallway, confrontation with SF sword wielder, confrontation with SF car kicker, and eight bike thefts. Mountain View and Palo Alto, which are markedly more straight laced and somewhat more policey, are commensurately nicer to walk around at night. El Paso downtown is extremely nice to walk around at night, and they have policemen shooing away skateboarders. Beijing is extremely nice to walk around at night, and there are men with sticks ready to drag you away if you are a public nuisance.
The second reason I think soft-on-crime is in bad faith is that I came from a strong and rigorous research background (math, with a side of chemistry and a family in medicine and pharma), and I bummed around in psychology research for a while, because it was interesting, but it was *bad*, it was really badly conducted, it was cargo cult in the way Feynman claimed, it was corrupt in the way Arielly was eventually outed, it was feckless in the idiots-with-fMRI-machines way. So I don't really trust when it conflicts with my anecdotal experience.
Again, that's just my experience motivating my feeling about crime. If you don't share it, the claim won't appeal to you, and that's very understandable.
It seems to me that's exactly what Scott did. He didn't find the smoking gun you want, but also he wasn't looking at research published by the private prison industry association.
See my reply to him. He did something related, but different.
Thanks Scott. I was eagerly anticipating this post, and it didn't dissapoint.
Yes, it was really clear and interesting.
"Much more than you wanted to know" is some of my most favourite series
It seems like we need to move the Overton window to allow for much higher levels of incarceration for the incapacitation benefits. This would require a seemingly insurmountable political effort to convince our fellow citizens that it would improve society and the effects would be concentrated amongst our most vulnerable and criminal adjacent populations. We could layer in all the type of rehabilitative efforts and strategies that the soft on crime folks want, but the main goal is to get criminals off the street including the zombies on Methadone mile in Bston, Kensington in Philly, etc.
One very deep area of needed analysis and change is the effect of the Warren court on criminal justice. The history of Miranda, Gideon, Terry, etc. was really the first "defund the police" movement that we all view as needed a heroic civil rights movement against the police state. In reality they were a burden to law enforcement that created this procedural nightmare for our criminal justice system that not only lead to a spike in crime, but is now used as a reason to not imprison criminals (It's just so expensive!).
Thomas Sowell outlines this well (along with how the sex education revolution arguably caused a spike in teen pregnancy and VD) in his book Social Justice Fallacies. I'm sure there are plenty of holes in his argument but even after a vicious attack I think it's mostly right.
The modern history of cities and towns ravaged by devastating crime is one where cops know who is doing it, where they are doing it, but are impotent in making a meaningful difference. All our talk on solving this issue has to work around that reality. I wish we could change that.
It seems prima facie really unlikely that prison would have _no_ effect, in either direction, on later criminal activity. After all, it's a massive intervention in someone's life - what are the odds that it wouldn't matter _at all_?
Instead, it seems a lot more likely that the data and research tools available don't produce any results, and that this lack of results gets erroneously transformed into a "result" that there are no effects.
I think the way I would look at it is that if prison didn't have enough of an effect to be detectable with current methods, you would get better methods. If no methods are up to detecting the effect, then the effect must be very small and you can round it off to zero.
Or the optimistic worldview: societies are so well-calibrated that they adjust the level of incarceration to the exact spot where the marginal benefits are zero. As a perfect society should.
I'm joking, but only halfway: there is some feedback loop, even if it operates on the level of decades.
A perfectly calibrated society would not have zero marginal benefits from incarceration; it would have zero difference between marginal benefits and marginal costs. The actual marginal benefits from something as expensive as incarceration should be easily large enough to detect, unless the measurement methods are extremely bad.
I thought what research shows is that it has a large incapacitation effect and a small to potentially large aftereffect in the opposite direction, with deterrence at current levels of incarceration being minimal. That effects cancel each other out indicates the intervention does have big effects, just in mutually canceling directions.
@Scott Grammar Error
"Large increases in incarceration cause large increases in crime." should be flipped. Please delete this comment once you fix, or determine I'm wrong.
The correction should change it to "large decreases in crime"
A note on shoplifting. If you can't punish shoplifters because you don't have the resources to try all of them, then you have two main options to change that.
(1) Fund some kind of "shoplifting task force" (or petty crimes task force if you want broader reach, but this will spread your efforts). Hire more prosecutors, judges, public defenders, bailiffs, court reporters, etc. and try more cases. If it's constitutional, you could target this effort at people with multiple prior recent arrests first, although the population will get the idea that everybody gets a few free misdemeanor tickets.
(2) Increase the difference between the minimum sentence that you get if you go to trial vs. pleading out. In other words, create "first degree," "second degree" etc. shoplifting and make the penalty for the higher degrees painful enough that people will plead to third degree shoplifting and do a year in prison rather than risk a conviction for second degree and five years.
#2 is inhumane and has been reasonably compared to medieval practices of confession under torture, but is also the way we manage our trial dockets in the US, unfortunately.
Which of prosecuter time, jury, judge, public defender, and physical court space do you think we can reduce?
> the population will get the idea that everybody gets a few free misdemeanor tickets.
That seems like a fairly acceptable outcome. Ceiling would continue falling at a hard-to-predict rate, so anyone smart enough to do the math probably wouldn't want to get too close to the de facto limit - and if there's a broader societal sense that something is being done about the problem, even first-time offenders would face a higher risk of being a) reported at all, b) while the prosecutor happens to be in a bad mood, or having a slow day.
It's been noted elsewhere in several places, but I'll note here too that another "main" option to change that situation is to lower the amount of resources you spend to try each shoplifter.
Scott did you find anyone who even tried to answer the seemingly hyperimportant question of 'how do we reduce the impact on recidivism from gang activity in prisons'?
I could imagine it being much more cost-effective for the government to directly pay some entity to hire former convicts, with incentives related to what fraction of them reoffend, as an example.
Tons of complicated incentives to navigate for sure, but seems a lot worse than the default of prison being training camp for more crime?
Also why can't we have corporal punishment/labor camps offered as an alternative.
I.e. 'you can have the current system, or you can opt for 1/4th the duration at hard, barely productive but somewhat productive labor' and see if that reduces recidivism.
i posted elsewhere about Skarbek's book, but based on reading that book, it seems to me you could put a very big dent in prison gangs just by making prisons very safe places where the prisoners don't need to fear violence from other prisoners. this would also rob the prison gangs of the power they exercise over gangs and criminals on the outside.
Let any prisoner spend as long as they like in solitary - in or out at mealtimes, maybe - with a cheap laptop and an internet connection, so long as they don't physically break said laptop (allowing some slack for accidents, repair / replacement costs) or use it to commit additional significant crimes. Gang recruiters would need one heck of a sales pitch if the BATNA is "left alone to watch cat videos!" Room for further improvement from there, of course, but you might as well grab the low-hanging fruit first.
> Let any prisoner spend as long as they like in solitary - in or out at mealtimes, maybe - with a cheap laptop and an internet connection
We don't even let them have internet access after they leave prison, and you want them to have it while they're in there?
(Last I heard, some judges were starting to be uncomfortable with cutting people off from the internet on the grounds that it's functionally equivalent to casting them out of society itself, but I suspect progress on that front is slow.)
There are far more plausible ways to rape or murder somebody through a foot-thick concrete wall, than there are via TCP/IP - assuming reasonable levels of monitoring to rule out e.g. drone weapon platforms. If reducing homicide and sexual assault within prisons is a desirable goal, that seems like a plausible candidate for cheapest AND surest way to make it happen.
Probably at least a few prisoners will lack technical savvy about exactly how much spyware can be crammed into a state-issued laptop, and thereby end up exchanging emails which score some lucky prosecutor a pile of easy wins.
To sell it to judges, consider principles of self-defense. Almost always better to retreat than fight back, given the option, right? So give 'em the option. But long-term sensory deprivation is too torturous to be a viable alternative to violence - sorta like expecting somebody to retreat from their own home. Obsolete laptops can often be obtained at scrap prices, while prisoners who earned some degree of trust could practice valuable trade skills by fixing up busted ones, meaning costs stay low and there are plausible rehabilitative benefits even if the internet access itself is considered detrimental.
I feel like this is going to work a lot better on ACX readers than actual criminals. Lots of people *like* to fight, especially in jail, and lots of men want to be a member of a group that has a strong identity and fights with other groups.
Conversely, many ACX readers may find solitary confinement with a laptop to not only be preferable to being in a gang in prison, but to normal life *out* of prison.
Lots of people like to *win* fights. I'd guess not as many enjoy losing. With a line of retreat available, some will take it - reducing available targets for the worst bullies - and many more will think it over enough to undermine their own courage, instead of benefiting from the clarity of desperation. Fewer, fairer, more fully consensual fights still feels like it'd be useful progress.
Strong shared identity and frequent inter-tribal physical conflict, but unable to coerce folks into joining, because anyone who doesn't like the vibe can just leave without first needing to negotiate for another group's protection, seems like it could shift that overall social dynamic away from "agoge," https://acoup.blog/2019/08/16/collections-this-isnt-sparta-part-i-spartan-school/ more toward something acceptably similar to "sports team."
“Large increases in incarceration cause large increases in crime” -> I think you mean decrease here.
This isn't data, but at least one cop told me that what is needed is not longer sentences, but swift and sure sentences.
The prison time can be short but you need to know that you are likely to be caught and sentenced soon after committing a crime, and you will serve your entire sentence.
Not sure whether that is reflected in the data.
Sort of like toilet-training a puppy. You gotta catch them in the act to get the training to work. Going back two minuites afterwards has no effect; you might as well be talking about ancient history to Fido.
I have a fair amount of experience in Europe. Europeans are all about social norms and The Done Thing.
That said, I was in Brussels during the Marc Dutroux Fiasco, and a woman screamed at me that in Texas, this guy would be getting lethal injection already. I told her that she should not get her hopes up, he'd be on his fifth criminal appeal or so.
Friend of the Comment Feed @grahamfactor had a post noting significant differences in US and Euro policing norms, highlighting how Euro cops are allowed significantly more leeway and freedom in investigating people and seizing evidence
https://grahamfactor.substack.com/p/earl-warrens-greatest-mistake
Yes everyone agrees certainty of punishment way more important. Is a consensus across academy and practitioners. Just mechanically resource intensive.
Also, it's hard to make punishment swift and certain while simultaneously providing robust procedural safeguards against railroading innocent defendants.
No argument there. Like anything else, it's a tradeoff.
If it's swift it can be argued that the damage would be more minimal. And innocents are less likely to be repeat offender (?), which seems to be the main topic here.
Procedural safeguards become less important when the cost to a genuine innocent is less severe, which becomes feasible if punishment per offense is progressively stepped down as response times get tighter. Tomorrow in jail (but out by tuesday) for a misstep today scares a high-time-preference criminal more, *and unlucky good citizens less,* than twenty-to-life starting eight months later.
In the case of honest mistakes, I am tentatively inclined to agree with you, if we're talking about modest fines, a day or two in jail, or a few weeks' probation.
But I am worried about wrongful punishments imposed by police and magistrates with arbitrary, venial, or malicious motives. E.g. paperwork says "disorderly conduct", but the actual offense is "failure to bribe" or "contempt of cop" or "being a visible member of a disfavored minority". Likewise, wrongful punishments could result from a malicious accuser who is sincerely believed by a well-meaning magistrate.
The lower the standard of proof and the fewer procedural safeguards, the more latitude those proposing and approving punishments have to abuse the system, and the more abuse I would expect to be attempted.
At a minimum, there needs to be a hearing with a moderate burden of proof prior to punishment, plus an opportunity to appeal after the fact and get compensation and a clean record if exonerating, and also some kind of systematic oversight looking for abuse of the system.
My takeaway is that we need to work much harder on deterring crime with police and other measures, both to prevent the crime itself and to change the political calculus on prison. If crime were less of a problem, there would be less knee-jerk desire to increase imprisonment.
Great post!
I don't think it makes sense to talk about deterrence as "% decrease in crime per extra year in prison". Shouldn't we talk about "% decrease in crime per % change in prison time" instead? Making the average murderer spend 19 years in prison instead of 18 years probably doesn't reduce murder much. But making the average junkie spend 15 months in prison than 3 months seems like it would be much more of a deterrent. If this is true, than deterrent effects would be much more powerful for smaller crimes, and could even become large enough to be worth considering for those?
I was a bit confused about the age thing until I realized that age isn't an aftereffect for being in prison (free people also age) so we shouldn't include it as a prison aftereffect.
I am now sold on the Australian model as a solution to basically all crime.
No, not modern Australia, which sits comfortably above the European system and below the American one in terms of incarceration rates, and doesn't seem to be doing anything novel, but the Australian system as it was for the United Kingdom a few centuries ago. Penal colonies.
If the problem with increasing prison sentences to the point we eliminate 90% of the crime is that it would cost insane amounts of money, then shipping those prisoners to a completely separate society somewhere far away would have the incapacitation effect (which seems to be by far the most important factor), without any of the significant costs. A one-way flight ticket is maybe a thousand dollars? Pretty much anywhere in the world.
If you think penal colonies are cruel, present the option to the criminal to join one, or our current system. If 5+ years of imprisonment essentially breaks all your social bonds anyways, it seems a much preferable alternative to have freedom in a new, if harsh, land, rather than be imprisoned in a literal prison near to home. If you're worried that penal colonies will be dangerous and deadly, as they're full of criminals, that's sort of what prisons are already. Except with a colony you can go out to the sticks and rarely interact with other people, and can have the means to readily defend yourself.
Unfortunately it doesn't look like there's anywhere left that could function as a modern day penal colony without being cruel. The whole idea is incapacitation, so it doesn't work if we just ship them to Montana or something and call it a day, as they could easily take a bus out of there. It would have to be separated enough that returning to society without permission was sufficiently difficult. Even a large island (let's say we sacrificed Hawaii's big island for this purpose) can easily be escaped if not for significant monitoring.
I think Robert Heinlein's, The Moon Is A Harsh Mistress should be our model for an actual solution. The moon is far enough away that it would be quite difficult to get back, and the long term effects of microgravity might make returning to earth difficult, painful, or straight up impossible. For those of you that don't know about the book; the moon was essentially one big penal colony. There were guards, but they didn't actually police what you did so long as you weren't actively sabotaging the operations of the colony. Mars could theoretically function as the same sort of thing.
Could this be the most compelling argument for Mars colonization? Could this be how colonizing Mars actually makes sense for improving life here on earth? Probably not. But it's fun to think outside the box.
Wouldn't an underground colony in Antarctica be far more economical?
Somebody who burrows their way back up to the surface in Ohio can find edible wild plants, liquid water, and hitchhike to anywhere in North America. Somebody who does so in Antarctica simply freezes to death. Maybe, if they're resourceful and lucky, they make it far enough to build a boat (though surely not from local trees...), avoid being eaten by leopard seals, and start a new life in South Africa, Tierra del Fuego, or get shot by whoever's guarding the fence at McMurdo AFB.
I think the US can do far better than this: arm them to the gills and use them conquer (and subsequently occupy) territory.
I imagine that would work.
But, I don’t think it would be a good thing for US soft power to be sending raving bands of armed criminals to conquer and occupy neighboring countries.
I was thinking more Iraq and Afghanistan than Mexico and Canada.
A one way flight ticket + a rifle and ammunition would definitely be cheaper than whatever we’re doing now.
I think anyone who would be comfortable with something like that would also be comfortable with executing the prisoners who are repeat offenders by firing squad though, which is much simpler. Most people who fall in the Overton window of thought on the issue probably don’t have any reason to want to conquer and occupy Afghanistan or Iraq.
Raping a bunch of Iraqi women under the US flag wouldn't go over well.
#MakeBashiBazouksGreatAgain
The Russians make it work, but they’re Russians.
For "year three of the one-week operation, begging for North Korea's scraps" values of "work," sure. I like to think we should do better.
In the last few decades we seem to lost the option to exile, which is the default punishment for millenia
its still routine with non-citizen immigrants in a lot of places
That seems like the only avenue left for exile indeed. Now I wonder if someone's done a research about hypothetically exiling citizens
Everywhere's been taken, sadly. Gotta wait until they can be shipped to Mars.
Really? There are loads of uninhabited islands. Mostly cold, mind you...
Loss of exile as a punishment is related to enabling internal mobility for the law-abiding.
When criminals were exiled, new entrants to a community weren't exactly welcomed with open arms.
(Also: exile, the severity of which is graded by distance from home, is a really common punishment in traditional China. But it's described in more detail in Water Margin (all the characters are criminals), and it involves substantially more than exile - in Water Margin, once you reach the place to which you've been exiled, you're imprisoned there! You're also branded, or tattooed - it's some kind of permanent mark on the face - but that doesn't surprise me as much as learning that "exile" means "being held in prison, but farther away from your home than usual".)
Cost per kilogram to orbit would need to drop by quite a few orders of magnitude for outer space penal colonies to make any kind of sense, even if you're not sending much life support gear along with them - which seems like the kind of thing there'd be protests over.
> Unfortunately it doesn't look like there's anywhere left that could function as a modern day penal colony without being cruel.
The USA is packed to the gills with empty land. There's no shortage. https://en.wikipedia.org/wiki/Federal_lands
Note also that the total population in prison in the US is under two million, and this is considered a large number in an international context, so everyone will fit comfortably in one city. It wouldn't even be a large city.
You’re right, and if it was to be done it would have to be a deliberate use of our controlled empty land.
The idea in my head at least, was that they end up so far away, and outside much if any state capacity, that you really didn’t have to worry about doing any work (or more like spending any money) to keep them contained.
The necessities of practicality always seem to beat any speculative ideas I have. 😔
Actually, I'd be completely on board with a scheme like:
- We carve out a modest region which is now The Prison of the United States
- We put up customs barriers around it so prisoners can't leave
- Non-prisoners are free to go in or out
- We don't try to administrate the people inside; they govern themselves.
Heck, we could extend this to allowing anyone from anywhere in the world to immigrate to The Prison, gaining the same "imprisoned and can't leave" status. Or, allow them to leave only by being deported.
It'd be interesting to see what happened.
Ah, so this is why rich people are behind Elon.
After surveying the evidence I am 100% on “team police” wrt crime. The most “equitable” way to address crime is deterrence rather than punishment, and the most effective way to deter crime is with police. If one young man shoots another, then _two_ lives are ruined. If a police officer standing on a street prevents that, it’s a huge win for society. That’s why the progressive prosecutor movement failed: it was being run at the same time as the movement for police abolition. If you want to decrease people’s appetite for incarceration, which would be a win for society, you need to actually increase police presence to counteract the decrease in incapacitation. IMO this is a significant part of why Europe has both less crime and less incarceration: they have a much higher rate of police per capita, and they are stricter about laws we barely enforce like requiring people to pay for public transit.
I would guess the higher amount of guns per capita in the US should mean we should expect the US needs to higher spending on police relative to all European countries. Not just equipment but training - probably hard to train someone to distinguish in a split second between a cell phone and a gun being pulled out?
Also social skills in de-escalation, to solve that sort of problem before it comes down to a matter of split-second reflexes.
I'm somewhat confused about American police. Before BLM, there were more than 10 years of articles about how American police is turning into a powerful army in itself, buying armoured vehicles, riot guns etc. These things are expensive, even if bought at an army sale. So apparently American police was very well funded during this period. Shouldn't there have been a huge decrease in crime then? Or funding was there, but spent badly (on shiny toys instead of more officers)?
In almost any organization salaries are much more expensive than equipment since they’re a recurring cost, and it’s often a different pool of money. I would 100% agree that fancy toys don’t help with crime but it’s probably more the fault of the federal government to earmark money for equipment rather than salaries (I think what you’re probably referring to is Homeland Security money that was spent on fancy chembio gear or whatever).
A lot of the military grade equipment and/or snow cone machines were funded through grants from the DHS in the wake of 9/11. Lots of federal funds to buy machine guns and ammo and armored vehicles for "counterterrorism". Not so much county and state funds for the actual day to day running of police operations.
> These things are expensive, even if bought at an army sale. So apparently American police was very well funded during this period. Shouldn't there have been a huge decrease in crime then?
American police spend approximately all police-hours on traffic tickets, rather than solving actual crimes. And this is with homicide closure rates of ~50% and rape ~30% and property crimes ~10%.
It's a bit of math to get there, but my full argument is here:
https://performativebafflement.substack.com/p/more-than-80-of-police-hours-are
Bottom line, if you want to increase police funding, you need to legislate at a high level that they CANNOT use the extra funds / police-hours for traffic tickets, but instead must use them on solving actual crimes.
Otherwise, we're just going to get more traffic tickets and similarly laughable closure rates of actual crimes.
Maybe vehicle theft and larceny are substitutes? You know you have two strikes, so you're not about to steal a car, but now you don't have the money you would've sold the car for, so you just have to earn your living the hard way: stuffing your sweatpants with toiletries and walking out of CVS looking like the Michelin Man.
> Do longer prison sentences reduce crime?
Yes.
> On the other hand, people in the field keep saying there’s no relationship.
The field doesn't want prison to reduce crime, therefore the research will say it doesn't. The last guy who violated that rule in the general field of criminology was John Lott, and look what happened to him. Lysenkoism requires that the outcome be known *before* the research is started.
> And they’d already gotten caught twice, which suggests they’re either worse at crime than average, or operating in areas with better-than-average police coverage.
Or do a lot more crime than average.
> not being too picky about whether the people they imprison were guilty of any specific crime or just kind of gang-adjacent.
Gang-adjacent, like an auxiliary that can be called to maintain gang control once the "official" gang leaders are in prison? You can't kill a Tree of Heaven by cutting it down, you have to uproot the entire plant. Same for gangs.
I was going to continue in this vein, but then I realized something: No one involved in studying the issue seems to understand how the criminal justice system works. They all seem to think that every defendant gets a fair trial, where all of the charges are put to a jury, who then convict (or not, based on the evidence) and pass the case to the judge for sentencing.
Which, lmao.
What actually happens is that a guy is arrested for, as a demonstrative of the principles involved, armed robbery with a handgun - he knocked over a gas station. Let's assume the charges are as follows:
- Armed robbery (Forcing the clerk to turn over the money)
- Assault with a deadly weapon (Pointing the gun at the clerk)
- Battery (Shoving someone during the robbery)
- Witness Intimidation (Threatening the clerk not to call the cops or he'll come back)
All told, he's facing 60 years all-in. But he knows that if he gets caught he'll be out in no more than 2. How?
1) He's in a district with a Soros-model DA, who drops most of the charges and negotiates a plea deal based exclusively on the robbery, drops reference to the weapon, and agrees to a suspended sentence of 6 months. (A suspended sentence is one that is not served unless the defendant is arrested on a different charge while on probation)
2) He's in a district with a normal DA, who drops the witness intimidation and assault charges, negotiates on the armed robbery and assault, and drops the assault to agree on an 18-month sentence in exchange for saving the DA time and effort.
Plus, judges only see criminals when they are in full suck-up mode, and not during their criminal activity, so a lot of judges end up assuming that criminals are much better than they actually are.
For a truly brilliant example of the difference between "suck-up" and "normal", https://www.youtube.com/watch?v=Q24L2CfGHIc
The guy in the first part of the video is apparently honestly pleading his case and then the judge says he's going to jail and the guy, well, you see.
But most judges just see "suck-up" and hand out light sentences, and almost all judges will accept a plea deal.
So we see a Red Queen's Race between state legislators trying to impose meaningful sentences for crimes, and DAs and judges trying to keep the plea-deal machine oiled. As a result, you see a bigger and bigger gap between what sentences could be and what people are actually sentenced because of those pressures.
Which is way tagging the analysis to the passing of "tough on crime" bills is either a complete waste of time or deliberate malfeasance. Just because a bill passes doesn't mean the system changes.
But if passing laws in general doesn't lend evidence to the effect of incarceration, doesn't that exactly cancel out the supposed benefit of passing laws causing less crime?
Because the latter question is what people are using this information to act on. What achievable interventions do you think are effective, given your worldview?
> But if passing laws in general doesn't lend evidence to the effect of incarceration, doesn't that exactly cancel out the supposed benefit of passing laws causing less crime?
My point is that the US data isn't meaningful because the laws are being actively subverted. Soros DAs are the perfect example of this concept on a broader scale - many of them simply refused to prosecute certain offenses, which meant that the law had no impact. That's why law enforcement swear an oath to enforce the law. But Progressives and Socialists see everything as an exercise of pure power, unbound by duty or honor, so they will take the oath and openly violate it the same day.
> What achievable interventions do you think are effective, given your worldview?
Barring plea deals absent some concrete benefit to the state beyond expediency would solve a whole host of problems in the justice system, including this one.
> Barring plea deals absent some concrete benefit to the state beyond expediency would solve a whole host of problems in the justice system, including this one
I still don't get it. You're talking about extrajudicial forces working to subvert the law. Why wouldn't those forces equally subvert barring plea deals by instead calling them dee peals? Isn't that just putting aside all of your hardheaded realism about how things actually work in the real world in exchange for a fantasy of justice?
> You're talking about extrajudicial forces working to subvert the law.
There's nothing extrajudicial about judges and prosecutors.
> Why wouldn't those forces equally subvert barring plea deals by instead calling them dee peals?
Because the law solved the semantics bullshit scam a long time ago. I know it's a favorite play of rationalists to say "Well what if we just change the meaning of words, make up new ones, or decide they don't mean anything at all?" but the law doesn't allow that. Check any state or Federal code. The first section will be a section of definitions. Does your conduct match the definition? Then we don't care what name you give it, it counts.
> Isn't that just putting aside all of your hardheaded realism about how things actually work in the real world in exchange for a fantasy of justice?
No, because I've actually thought about and studied this for longer than you have been aware of the problem. Eliminating plea deals breaks a ton of incentives in the "Speedy Conviction Assembly Line" the modern justice system has devolved to.
> The guy in the first part of the video is apparently honestly pleading his case and then the judge says he's going to jail and the guy, well, you see.
Interesting to see random-guy-in-a-suit beating the hell out of the defendant. He (the defendant) gets restrained by one guy who is obviously security and two more guys who obviously are not.
Why does the audio keep cutting out in the video? It's not to obscure swearing.
Names, probably.
Almost certainly. A lot of government-released audio will have identifiers removed.
I believe the random guy is the judge's clerk.
I think this is a typo: “Large increases in incarceration cause large increases in crime.”
>Maybe it’s hard to deter violent crimes (because these are executed in a fit of passion), it’s hard to deter property crimes (because criminals need the money), but it’s easier to deter drug crimes (because people make more rational decisions about whether or not to use drugs)?
My default assumption is that it's easier to *not get caught* for drug crimes if you are being careful, since taking drugs is usually a victimless crime, and if you carefully do it in private at the right times there will be no criminal complaint for police to investigate. So increased caution for the same amount of crimes is especially effective for this category.
Skarbek's book "The Social Order of the Underworld: How Prison Gangs Govern the American Penal System" is a great read.
The key point is that, because most criminals expect to go to prison, they know they will someday be at the mercy of prison gangs who have the power to inflict violence on the inside. This gives prison gangs enormous power outside of the prison walls; non-prison gangs usually pay tribute to a given prison gang.
(This incidentally is what worries me about El Salvador. Having all the gang members in prison right now might be like having guerrillas driven up to the mountains; the state may be ceding its monopoly on violence in a certain part of the country and this could eventually be a problem even if the rest of the country is safe in the meantime.)
Putting a dent in the power of prison gangs would be intrinsically good and probably help with crime on the outside.
Unfortunately solutions to this are probably unpalatable. Removing the gangs' power to commit violence and smuggle drugs would look like making conditions better and safer for prisoners (undesirable to the right), but would require large increases in funding for prisons alongside stricter security (undesirable to the left).
However this is an issue that's nicely orthogonal to "turning the knob" of more/less sentencing, so I figure it's worth mentioning.
Any automatic restriction also risks cutting off positive social connections, and has costs in terms of duplicated facilities, more guards to watch all those separate spaces, etc. Wouldn't stop gangs, either - only need one guy per pod with a shiv and a phone. If you want everybody in the big house to be able to avoid bad influences, better plan would be to just say prisoners can't be put in the same room together without mutual assent - anybody wants to get away from their current cellmate, or a particular crowd at the cafeteria, or anything like that, it happens ASAP, with no questions about why.
I didn't read his book, but I listened to his econtalk episode when it came out, and as I remember his story was that prison gangs had actually made prisons safer on the inside. But of course they have other deleterious effects.
They arose in response to the state being unable to maintain prisoner safety, sort of like warlords or mafiosi bringing back some order when the state is too weak to do it.
This analysis ignores that punishment is a good in itself. The victims of crimes deserve retribution against the people who have wronged them.
If I was a victim of a non violent crime, I would be willing to let the criminal go free if I could be rewarded with some of the money that would otherwise be spent on the trial and the incarceration.
You are not going to be compensated for more than the cost of the crime. For the state to do this would be a very bad idea, we’d effectively be paying criminals to be criminal. You might also be underestimating the emotional distress of being mugged, which can’t be compensated.
Worry not - your two objections work in opposite directions!
If the emotional distress of being mugged is so high, why wouldn't Kristian be compensated for that too?
I think these things are impossible to measure. I don’t see a contradiction though. Let’s say that instead of imprisoning Bob the mugger we compensate the victims for the price of the items stolen. And let’s say that’s easy to do. Ignoring fraud for a moment, Bob is free to mug as he will, and the State intervenes not to lock him up but to pay the victim off. This becomes a job, not really a criminal activity.
Fraud would happen of course: “I mug you, you get compensated for the phone and I get the phone”.
Retribution is indeed the primary reason for imprisonment or incapacitation, and it rarely is mentioned in these debates. The alternative would be not sending people to prison if they were very very sorry indeed, and could prove it.
I'm pretty sure most people would say that deterrence, not retribution, is why we should imprison even the very sorry ones.
When I say the primary aim of the system is retribution that’s an actual description of reality.
Take Norway as an example -largely seen as rehabilitative . While they focus on humane prisons and rehabilitation, crimes like murder can see offenders locked away for up to 21 years, extendable indefinitely if they’re deemed a danger. That’s partly retributive, ensuring the punishment reflects the severity of the crime, but also acts as a deterrent to others.
However deterrence isn’t necessarily effective - read the piece again - and may not work at all for mafia types and hardened criminals for whom going to jail is just part of the lifestyle. If the law doesn’t work to deter, it can work to satisfy justice.
See the incapacitation part.
It is true that many humans think retribution is inherently good. But many other humans don't. For example, I don't care about retribution as a terminal goal pretty much at all. Bad things happening to bad people is a negative for me. I want good things to happen to all people. I care about criminal justice only in so far as it helps increase that expectation value.
I don’t think it’s always the primary aim, but the concept is important in many situations. “Retribution" sounds harsh, but most people ("society" or "the public") expect their criminal justice system to include an element of punishment apart from deterrence and incapacity for at least the most violent, horrific, and pain-causing crimes. It's not entirely clear that the desire for retribution in such cases is immoral, such that the criminal justice system should ignore the will of the majority if that were possible. Even very progressive people tend to back away from the most extreme consequences of non-retributive justice when confronted with examples. A professor in a criminal law class liked to walk people through hypotheticals based on real crimes to show that almost everyone expects and even desires some form of retributive punishment for certain classes of crimes. I don't recall all of the examples, but two I remember were someone throwing acid into the face of someone who politely turned down a request for a date, leading to permanent and horrific disfigurement, and someone who tortured and killed small children to "get back" at an ex. Imagine these crimes were committed against you or a close friend. Even if you assume that the perpetrators would never commit another crime again and that nobody would ever hear what punishment the perpetrators received (i.e., no punishment at all), very few people are willing to accept that those committing such crimes should go free and unpunished. Often, people want to fight the hypothetical - that you can never *really know* whether a criminal will offend again, for example, and so retribution is "wrapped up" in other reasons for the sentence.
I wonder are there any useful studies on the effects on corporal punishment like caning/ whipping etc compared to prison time. Historically this was by far the most common form of punishment, receiving support from a wide group like George Washington, Thomas Jefferson, Abraham Lincoln to Theodore Roosevelt. Then it suddenly it disappeared all across Europe in 1940s.
Intuitively a corporal punishment is all deterrence and no in-capacitance and so by your analysis its effect on crime rates should be small (but it’s also much less expensive than prisons). This doesn’t sit right with me though, I speculate there is something in the nature of corporal punishments that makes it a much more effective deterrent than prison.
Also in many cases I would personally prefer corporal punishment as the more human option over prison. If I was caught shop lifting, 30 whips seems like a better deal, than a waste of 6 months to 3 years of my life. With a corporal punishment, you can return to your life quickly, with the hope being that the punishment instilled enough fear in you that you wouldn’t want to get caught doing crime again.
One argument in favor of corporal punishment over prison is that the punishment accrues sooner, so it will be a better deterrent for people with high time preference.
Middle class guys probably would suffer worse, they take fewer hard knocks during everyday life.
I do wonder if it's going to increase crime among masochists. Presumably you wouldn't have hot women doing the caning. Where's that guy Giles who posts here sometimes?
I've thought that as well. I don't think we're going anywhere with it though. Too far beyond the pale these days.
"Small increases in incarceration cause small decreases in crime. Large increases in incarceration cause large increases in crime." I think you mean "large DECREASES in crime" in the second sentence.
"While criminals are in prison, they can’t commit more crimes (except within the prison walls; we ignore these for sake of simplicity)"
I agree ignoring crimes in prisons makes the analysis simpler, but it also puts your thumb on the scale in the pro (additional marginal) incarceration direction.
Note that Roodman also makes this assumption (p.10, Section 2.2.2):
"Because prison crime is so rarely studied, and because crime outside of prison has particular political salience, 'incapacitation; in this review will refer to crime outside of prison."
This seems in tension with identifying Roodman as skewed in an anti-incarceration direction.
Separately, I'd be interested to see an analysis on jails, where people are incarcerated who are still legally presumed innocent and one expects have a higher rate of actual innocence. This is particularly relevant for bail reform debates
"a robber may steal $100 worth of goods, but valuing his crime at $100 in costs ignores the disutility of (eg) living in fear"
But if the robber steals from a store, the store is out its cost of those goods, and has to make it up in its profit on other sales. Net profit per sale depends on the actual goods sold, but it usually ends up being about 1% or so when you count that more expensive goods which give you a higher margin also sell less often.
In any case, the robber gets $100 (retail price) of goods, and maybe sells it on for $10-25. Cost of the goods to the store from the wholesaler might be $50-60. My own experience was with books, and there it was $60.
The store then has to sell $6000 or so worth of goods (retail price) to have a net profit of $50 and be able to rebuy that $50, because it's a straight loss. And the store's cost isn't just cost of goods - there's staff salary, rent or mortgage for the store, etc.
Insurance may pay that $60, but then will make it up in increased premiums, or refuse altogether if it keeps happening. And then the store likely closes.
So even ignoring the living in fear part (and for robbery it is fear, since the crime requires a threat of violence), it's still not just $100 in cost.
None of this cost shows up in government statistics of course, since none of it is the government's loss to bear.
it’s $100 in cost by definition because if someone gave $100 to the store it would completely restore the situation to what it was before
No, if someone gave $60 to the store it would completely restore the situation.
You're ignoring the time it takes to reorder stuff for sale, wait for it to be sold, etc.
Assuming no cost to the store to get the $60 (have you ever dealt with an insurance company?), and no time involved (so no time cost for that money).
Also no damage to other things or people involved in the theft. Which is incorrect, because it's robbery not theft.
This is like when a person sues to recover money. The court awards the money. Then there's court costs, which the court may or may not award.
Then there's the years of your life it took to sue the other party in the first place. The times you had to be in court instead of doing what you would normally be doing.
Mostly that's just not counted.
It also assumes that anything at all can be recovered from the robber. Which is not usually the case.
And then there's the cost of the loss prevention measures stores turn to in an attempt to stop the bleeding. Locking up more and more of their products, for instance, which leads to some combination of higher staffing needs, higher employee workloads, and poorer customer service.
"So maybe Roodman is right about shorter sentences, and everyone else is right about longer sentences. Going from a month to a year in prison is so disruptive and criminogenic that it risks canceling the benefits of eleven extra months of incapacitation. But going from ten years to eleven years mostly just gives you the incapacitation."
This makes a lot of sense to me. I would guess that the post-prison criminogenic effects of prison are heavily front-loaded i.e. heavily sublinear in the length of the prison term, whereas the incapacitation effect is close to linear in the length of the prison term (unless it's a really long term, so that aging out effects start to be important).
biggest omission: deterrence effects on people who are non-criminals just because they are deterred.
I agree with this. Anecdotally, fear of prison is the biggest reason why I don't commit crimes. Even bad ones like murder - I would rather commit murder and get away with it than go to prison for a murder I didn't commit.
Is that relevant for longer vs shorter sentences though? Like I agree enforcement vs not can have a big effect here. But are you more likely to commit crimes if you have to stay in prison for 6 months vs 1 year?
If not I think it reinforced Scott's point - increase certainty of enforcement first before worrying about longer sentences.
Has anyone looked into prison sentences whose duration inversely correlates with the age of the offender? Criminality (and political activism, and plenty of other things) peak in the late teens/twenties, so maybe focusing sentences which keep offenders in this age group behind bars until they've "mellowed out" would go much farther than comparable sentences for e.g. 50 year olds.
This is sort of the flip side of the push for harsher sentences for repeat offenders, since repeat offending is undoubtedly positively correlated with age (since you've have more time to re-offend).
In your costs/benefits of prison section, shouldn't you include the management/prison costs of the counterfactual prevented crimes as a benefit. I.e. how much police/courts/prison time is used up for an additional 3.9 cases of Larceny etc. It seems like it would be a high cost unless you want to do away with the justice system entirely.
I think you made a minor mistake in this section: “This broadly agrees with our numbers from Sweden, California, and El Salvador above. Small increases in incarceration cause small decreases in crime. Large increases in incarceration cause large increases in crime. If you doubled the incarceration rate, locking up an extra million people, then crime would decrease ~30% at current US margins (maybe less, because you’re shifting the margin and getting diminishing returns).”
Specifically you say “Large increases in incarceration cause large increases in crime” which I think is supposed to be “decreases in crime”
Given my extensive years in crimlaw, I have a lot of thoughts about this. A recurring problem with discourse in this area is that the practical realities involve an intersection point between criminals (a highly irrational subculture) and bureaucrats (a class of people who are given idealistic speeches at conferences and then go back to a job where there's bizarre incentives to say 2+2=5 if it clears their desk briefly). So not only are the decisions of the actors irrational and arbitrary, even identifying WHAT happened is inscrutable to people who don't practice crimlaw. I don't have time to react to everything here, but just a few points on the non-linearity of how criminals see sentences.
"So most deterrence will look more like the Proposition 36 proposal we discussed last month, which increases shoplifting sentences from six months to three years. If we use H&T’s numbers (probably inappropriate since there may be nonlinear effects), we would expect that section of Prop 36 to deter crime by 2%."
Correct about the expected nonlinear effect. I believe the Proposition would be ineffective because practical constraints would force prosecutors to amend charges rather than seek the heightened sentences required. But if somehow the sentences actually came to pass, the criminal class sees this VERY differently. A misdemeanor with a 6 month jail max probably means that at worst you get arrested, are held in some county facility, can't post bond, your lawyer screws around with discovery and motions to make themselves feel like they did something, and then you plead out to time served at some point. A 3 year sentence isn't just about the duration, it has to be served in a prison, which is a very different world than county jail.
Many lower-level criminals who will happily sit in jail for 6 months would push back strongly at having to enter the prison system even if you told them to expect parole at the same 6 month mark. Whether or not it deters the crime, it certainly deters pleading guilty as charged and greatly increases the chance of the defendant snap-accepting any misdemeanor plea bargain offer. Jail vs prison punishments are different in kind, not merely in duration. (There are large metros with jails that behave more like prisons with stable inmate populations serving sentences, as opposed to keeping "sentenced to jail" and "awaiting trial, couldn't post bail" people together, but as the "sentenced" group would be mostly low-level offenders and the "awaiting trial" group would have included more serious offenders, I wouldn't be surprised if these facilities were actually preferable to ordinary jail.)
"Ross 1982 investigated the effects of some widely publicized laws that increased penalties for drunk driving. They found that it decreased drunk driving 66% (!) in the year the law came out, but this effect gradually faded to zero over the next three years. The most likely explanation is that the publicity around the new law got people excited, but nobody really knows the laws around drunk driving anyway (do you know what the range of sentences for DUI in your jurisdiction is?) and so once the excitement of “law is stricter now!” faded from memory, the law had no effect."
Find this strange, as drunk drivers are uniquely deterrable among all criminals. This is because drunk driving is one of the few crimes likely to have a high prevalence among the middle class and important community figures. An esteemed local businessman cannot go to jail, if the DUI has a mandatory jail penalty that is a huge problem. In my state first offense DUIs typically result in probation and some annoying classes before you can drive again, whereas 2nd offenses have mandatory 5 days jail and 3rd offenses are felonies. And what's more unique, MADD actually keeps stats on this stuff and will go after you if you show weakness on DUIs, so judges and prosecutors here will in fact do the labor of trying a case before they are willing to reduce charges on a DUI. You may not know the penalties for DUI in your jurisdiction, but I guarantee you that the people inclined to drive drunk know them. Heavy drinkers all know someone who's gotten a DUI, and who got this lecture from his lawyer, and you will notice in practice altered behavior by prior DUI offenders.
If the deterrence effect lowered that much, my guess would be that the legal system was not imposing the heightened penalties -- a problem that would not show in the data unless you knew A) the NCIC charge code entered by the cop for the arrest itself, B) the charge code on the original felony complaint, and C) the charge code on the disposed case. And even if you did, you're likely to fail at point (A) because -- unless a background check shows enough priors to make it a felony and therefore something he'd seek bail for -- the cop may not have collected enhancement data from the suspect and just figured he'd enter it as a 1st offense generic DUI at the station and then fix it up on the probable cause affidavit. Determining whether people who should get increased penalties as the law is designed are actually facing them is pretty hard.
***************************
There are also threshold points even along a spectrum of years that are given irrational weight during plea negotiations. Numerous repeat offenders in my state were far more likely to accept a deal of 7 yrs than 8, this break point was consistently noted in negotiations. The only explanation we had was that historically C felonies maxed out at 7 and were treated more leniently by parole board than A/B felonies-- but if you had committed a B felony and were subject to the higher range, then the difference between 7 and 8 was no different than the difference between 8 and 9, getting 7 didn't make them treat you as if you were a lower class offender. But they acted as if it mattered to them greatly.
Which is all to say that you can't really ever boil this down to each year deters X, because there are both real and imagined thresholds that affect different social classes and types of offender in differing ways that are both hard to locate in data and defy rational behavioral analysis.
The whole jail/prison thing suggests to me that a remand system (where people awaiting trial are remanded to prison if they aren't granted bail) would have an enormous impact.
This is certainly how the UK works. There are cells in police stations, but they're only used for people being held for questioning (which has the usual Habeas Corpus limits of a few days - once charges are issued, they go to a remand hearing and then either to remand or bail).
Remand prisoners are held separately from convicts in prison, but that's just a case of wings being designated differently - if there are more remand prisoners and fewer convicts, then the same cells get redesignated.
I expect the problem in the US is get another example of too many levels of government: jails are county and prisons are state, yeah?
I've heard that it is often the case that county jails are worse places to be than prisons. Is this a common opinion in your experience? For example that seems to be the consensus as well as common knowledge in this thread:
https://www.reddit.com/r/Prison/comments/1bvqzcm/which_is_worse_prison_or_jail/
It depends on the inmate's background and personality type. In prison it's easier to get things, the social environment is more stable, career criminals who've been to prison before usually prefer it over jail. I think the problem with jails is that there's a lot of uncertainty which leads to psychological instability, new arrivals can be disruptive and unstable, trials are still upcoming and people are nervous. It's also usually true that the jail facilities themselves are worse than a prison's, the food is worse, the bunking situation is chaotic, etc. Inmates at prisons get iPads now, with limited function but it beats crowding around the one TV to watch Maury Povich at the jail.
Nevertheless, if you had never been to prison before, it would be a much bigger shock to your system than jail is. If you were middle class or above, and you only were doing a short stint, I think you'd prefer the jail. If you were gonna be there long enough to learn the system and get comfortable, and thought you were capable of having the social IQ to deal with it, prison is materially better. So a lot of it depends on your background and why you ended up there.
Another factor is that it's much easier to hide what you're in for at prison than at jail, because the jail inmates all go over for the same big criminal law dockets as you do. If you're charged with child molestation, even if the lawyer tries to be careful about it (waives reading of charges at arraignment, for example) eventually there will be some legal argument at the bench and the other guys will put it together that you're in for that and everyone at the jail will gang up on you. (Which even if you're the "they deserve it" type, is not what you want happening for people who haven't been convicted of anything yet, and is also a pain in the butt for jailors because they can't just let that happen or there'd be problems, possibly expensive ones.)
> There’s not enough justice system resources (public defenders, courtrooms, etc) to give shoplifters trials, so the prosecutors have no bargaining power, so they accept plea bargains for minimal punishment.
I take the opposite lesson from this, actually: the decision to obligate the state to spend more on trials is the problem. (Or, at least, it hasn't passed a cost-benefit analysis, yet.)
Like, the naive argument for the death penalty is "it's cheaper than keeping them in prison", because you don't have to house or feed corpses. But this isn't actually true in America, because we've made trials for death penalty cases substantially more expensive than trials for prison-in-life cases, and the increased expense of the trial is larger than than the decreased expense on prison. But this is a policy choice! We don't have to have the appeals system we have, or the standards for trials that we have.
Similarly, we simply haven't checked whether we need juries to be twelve people. (Secretly, 14, because you need alternates.)
Smaller juries would have hard-to-measure effects on the wrongful conviction rate, and probably not save much time or expense. Better approach might be to dig into where the procedural bottlenecks are, what sort of motions get used as stalling tactics.
Also, ought to let juries access the official transcript during their deliberations rather than relying on memory. Not directly relevant to speeding things up, but it's widely agreed to be dead stupid, might as well fix that too as long as we're meddling in there anyway.
On a related note: don't let the jury see the trial live. Record the trial, edit out the parts where someone says or does something that has to be stricken from the record, and give the jury the edited recording.
> But here we’re interested in its potential to confound studies. A 20 year old who gets 5 years in prison is released at 25 - still young! - but a 20 year old who gets 10 years in prison is released at 30 - too old to be leaping on rooftops and running from cops. The National Sentencing Commission understands this problem, and matches the experimental and control groups by age at release. But this introduces a new bias - now they’re different ages when they start committing crimes. Might a person who starts crime at 15 be a more disturbed and committed criminal than one who starts at 20? Seems plausible. I think this might be responsible for a lot of the seemingly positive effect of sentences > 5 years.
This is an issue that comes up in several contexts. For example, what happens if you try to hold your child back from entering school if their birthday is near the cutoff, so they're the oldest in their class instead of the youngest? No matter what identification strategy you use, true age will be a confounder to measuring the effect of relative age compared to peers. I've been told that this means the underlying question ("what is the impact of holding your child back?") is fundamentally unanswerable, although I think the real conclusion is "increased age is one of the effects, which you should account for when determining the effects." So in this case, the fact that a longer sentence means the prisoner is older when they're released is just part of the true effect on recidivism, and it might not be possible to separate out how much effect from a longer sentence is due to the direct effect of being in prison longer and how much is because the person is older upon release, but this doesn't really matter for a CBA of longer sentences.
> but a 20 year old who gets 10 years in prison is released at 30 - too old to be leaping on rooftops and running from cops.
...how old are you? Yes, you decay over time, but not _that_ quickly.
That part is a quote from the original post, so you should ask Scott.
Fair enough.
Feels like a re-run of the homelessness essay. Prison abolitionist activists intentionally drive up the cost/friction in arresting/imprisoning people. Research shows massive value in imprisoning people, but wait, its so expensive to imprison them, so even with the significant reduction in crime it is not 'cost efficient'.
The real solution is just, making it efficient again. This also applies to the justice system more broadly. It actually shouldn't cost the US government tens of thousands of dollars to arrest and charge somebody on their 20th count of shop lifting.
"If you find yourself in a hole, stop digging."
How do you propose we do that? The main costs are providing a lawyer (constitutionally required if jail is a possibility), providing a jury trial (required in many states if jail is a possibility, and required everywhere for a sentence over 6 months) and providing constitutionally-required procedural protections.
Oh, it is constitutionally-required, well then I guess it is impossible to change.
I don’t think a movement to repeal the sixth amendment is going to get a lot of traction.
Yup totally, repealing the sixth amendment is literally the only way to change any aspect of the system and it is not politically likely, totally agree, its actually impossible to change, wow.
Why would you need to repeal it? The current interpretation of the "right to have the assistance of counsel" is quite obviously unrelated to the meaning that was written down; all you have to do is say "let's use the original meaning of this law".
Examine and recompile the existing procedures - in particular, statistical analysis of how they're typically used and abused - with an eye toward providing the same level of protection for actual innocents while reducing unnecessary delays.
Make "one night in jail and a GPS tracking anklet for a year" the default sentence for nuisance-level offenses, with automatic logs of the tracker's position, so repeat offenders with no fixed address are easy to find for follow-up questions.
1. I’m not sure “study the problem and figure out the best solution” is actually much of a proposal. And can you even sketch out how you could get the data you’d need to perform that “statistical analysis”?
2. I’m not sure how your GPS proposal offers much help in incapacitating people for minor offenses, and it also seems like it doesn’t have much bearing on the topic here.
Lot of court proceedings are a matter of public record. That's most of what's needed, it's just not organized for the right sort of quick reference. Grab a prosecutor and a public defender, have them lay out types of cases as game-theory decision graphs, plug in empirical values for the various branches, use that to run cost-benefit analyses on specific procedural safeguards.
"Better that a thousand guilty men go free than one innocent wrongly punished," right? But at larger ratios, it eventually stops being better, and we do seem to have a non-negligible number of innocents being wrongly punished anyway. Whose job is it to find Pareto improvements on that curve? Did somebody 250 years ago write "temporary I hope hope hope" in the margin of a sketchy legislative proposal that's since come to be treated more like holy scripture?
Regarding the second point, massive routine use of those could reduce the cost / friction of arresting people (a prerequisite to most other punishment) by making them easier to physically find. Instead of putting out a passive bench warrant and waiting for the person of interest to catch an officer's attention for some other reason, track them down and sort it out same-day.
yes, it should be the more strikes you have the faster you can get some sanction (GPS ankle monitor, mandatory check-ins, community service, jail, prison, exiled to Ohio, death by nitrogen)
it's completely ridiculous how many discontinuities are in these processes
"the longer you’re in prison, the longer you have to be scared straight, or find God, or get rehabilitated." I think this is missing one factor. If a criminal has anything like the same life arc as I do at 77, then the longer time in prison means when you get out, you are too damned tired all the time to do anything, much less commit violent crime.
I visited El Salvador early this year and think the visible presence of the military throughout the country is probably also deterring violent crime. In San Salvador, towns, and even along the highway it's a common sight to see groups of 5 - 20 armed, camouflaged troops standing or walking around. My tour guide told me the troops were mostly looking for gang members who evaded the recent mass arrests.
Note that that the much-publicised mass incarceration began in 2022, but the decline in crime (which I agree looks pretty miraculous) began in 2015, and if anything it slowed after 2022, so we can be pretty confident that the two aren't connected.
To me, this all (which I don't have methodological issues with) tells me that "defund the police"[1] is exactly wrong-headed. Instead, we should have *more* police presence (increasing certainty of intervention).
It seems that most people want to have *outcomes*, which are some complex function of *both* severity of punishment and certainty of punishment. So if certainty goes down, people start insisting on more severity. Even though that's less effective. It's sorta like how behavior is a function of both intrinsic motivation and extrinsic motivation (threats of punishment, external enforcement). But people tend to seek "homeostasis" to some degree, decreasing intrinsic motivation as extrinsic motivation gets bigger. But then if you relax the extrinsic pressure...intrinsic does not increase. Leading to reduced overall behavior.
[1] whether that's *actually* defunding the police partially or entirely or just shifting resources to non-punitive measures like social workers.
I was all set to write a comment lecturing about how this deterrent and incapacitation effects measures were happening at the margin, but then I scrolled down and you had a whole section explaining this, so muy bueno.
Maybe what we need to do is reduce the costs of incarceration, then. Why does it need to cost so much to keep someone in prison? Bring back 19th century criminal justice: prison ships, stockades, flogging, and sentencing people to transportation (this time to, I dunno...the Aleutians?)
Punishment is the primary stated purpose of incarceration, and directly bears on the societal question of how much incarceration there should be.
This analysis of marginal costs and benefits of incarceration addresses the question it poses, "Do longer prison sentences reduce crime?" But the answer, and the broader economic analysis at the end would, I think, be viewed by many or most members of society through this lens:
1. What is the appropriate punishment for the crime? (An economist would include Roodman's costs to the prisoner in answering this question.)
2. Can society afford the cost of such punishment?
If there is a net benefit to society, the answer is obviously yes. Adding to the severity of punishment until the marginal cost/benefit is zero is morally dubious.
If there is a net cost, then at some point the cost might be large enough to merit compromising on the appropriate punishment. But that's still a compromise between two competing concerns, and the balancing point will have a nonzero and possibly substantial net cost to society.
That's why prison guards reject the evidence-based approaches: they're there to enforce punishment, which is omitted from the evidence-based calculus.
P.S. I presently don't seem to have the ability to type replies of more than a few lines on a web browser without the comment window expanding enough that the "Post" button gets hidden beneath the next comment below and can't be clicked. So if anybody chooses to engage in conversation with me on this post, my responses will necessarily be short and pithy.
Thanks.! I'm going to think of this as the Wizard of Oz solution.
Substack breaks their comment section in new and innovative ways every iteration. That problem is on the web. On the app - where I now am - I can’t select text to answer as doing that collapses the comment. Both are new bugs.
Exactly! People think they're enlightened and humanitarian when they dismiss punishment in favor of deterrence, but if you are only concerned with deterrence then the best strategy for deterrence may be out of proportion to the crime. Locking up all shoplifters for 20 years would deter a lot of crime (through incapacatance), yet it would be an unjust punishment. They would be punished more than they deserved!
Also, "enlightened and humanitarian" applies when talking about future jobs for convicted felons. Through the punishment lens, if you've "paid your price to society", you're ready to work again. Neat and transactional. Through the deterrence lens, you get to worry about recidivism rates and all that, and refuse to hire.
(I get to write somewhat longer replies if I hold my phone sideways...)
I'm having the same bug - you can work around it by posting a one-line comment and then editing it.
I think the section on the cost/benefit of imprisonment leads to odd conclusions the way it is structured. If it costs $15k/yr to imprison someone in Alabama but $120k/yr in California... everyone from jaywalkers on up should be imprisoned in Alabama and only the murderers and rapists are worth it in California. Is the implication that the value to society of imprisoning dangerous criminals directly relates to the cost of running the prison system in a given state? Further, if the cost of prison could be radically lowered, say a few thousand per year, is it then worthwhile to imprison people for minor infractions like speeding tickets?
Also, there has to be some fun utilitarian calculation where it isn't ethical to imprison anyone because the cost of suffering in prison outweighs the suffering of the victims of the convicts.
Almost 50 thousand people a year are killed in automobile accidents every year in the US. I would argue that it is not categorically wrong to imprison people for traffic violations.
Which we actually do, for DUI cases.
You mentioned some potential explanations as to why homicides in El Salvador fell by 95%, but you missed my preferred explanation: stopping organized crime provides a much greater benefit in reducing crime than stopping individuals do.
Like, if you arrest a single murderer, or a single shoplifter, you haven't really reduced the chances of another, unconnected person committing those same crimes. However, if you arrest an entire gang, those gang members no longer have the power and influence to commit crimes like they did before, so it actually does prevent future criminals from joining those gains and committing more crimes.
So basically, I think that the situation in El Salvador was relatively unique, as their society was being terrorized by organized crime. If you quadrupled the incarceration rates in America, I feel as though it would be extremely unlikely that homicide rates would see the same drop.
> People first organised higher order societies in order to protect against barbarian marauders.
The theory is not uncommon in studies of the near east (same region as the middle east, but earlier in time) that people first organized in order to handle temple construction and temple administration.
However, I vaguely recall reading about the discovery of a very large prehistoric battlefield where analysis of the remains showed that they were drawn from a very wide geographic area, implying that in fact people were organized on large scales long before we had previously imagined was true.
Some thoughts as a former defense attorney.
On "after effects," skipping all the way to length of incarceration misses a lot of what's going on for first time offenders. As your own examples note, many of the problems making it harder to make an honest living accrue from being a felon as such, and not from jail time. I'll add on that merely being *arrested* triggers a similar cascade of consequences. Your average person will be fired from their job very quickly after arrest -- either because of the arrest itself or just because it leads to missing work. And an arrest on its own, without a conviction, is sufficient to make it hard to get hired on in a lot of jobs.
But this is all specific to first-time offenders; going from someone presumed to be a law-abiding citizen to being a felon changes everything. An incremental felony after that doesn't change much. From your description, it seems like the studies are blurring this together? I'd be interested to see evidence focused specifically on first time offenders.
My guess would be that there's a large, negative effect of your first felony (all else equal) but not much effect at all of subsequent felonies. In that case, there's a very different calculus for the incapacitation vs. recidivism tradeoff for different, identifiable groups.
It's amusing that there isn't a big organized push from the "social justice" crowd to make discrimination against the arrested/convicted to be illegal. I suppose that pretending their client minorities are saintly is more expedient than actually trying to improve their lot.
You're missing the deeper issue - they'd be shooting themselves in the foot, since aggressively discriminating against people who've done certain sorts of crimes (or things they wish were crimes) is itself the social justice crowd's core business model.
There absolutely is such a movement, including from left-leaning states and the Biden administration.
There is. It's illegal to discriminate based on criminal background unless you can prove that it's relevant to the job. I talked a bit about this at https://www.astralcodexten.com/p/book-review-the-origins-of-woke
Fantastic article. I’m curious about any evidence to support more benefit to ankle monitor tracking over incarceration for nonviolent criminals? Seems like it could be cost effective and potentially less recidivism since their life wouldn’t be as disrupted.
Thanks for posting this – it was a really interesting read. I have a lot of thoughts, some of which I’ll try to share:
1. Crime data is very, very bad, which makes drawing conclusions based on it even more difficult than in other areas. The two main sources of data (in the U.S.) are the Uniform Crime Reports (a collection of data submitted to the FBI from police departments) and the National Crime Victimization Survey (where DOJ researchers ask people what crimes they were the victims of). Obviously we’d expect the numbers to be different since they’re measuring related but different things, but they’re not even close. For instance, in 2019 (this was the last year before the FBI switched from issuing actual reports to using a “data viewer” that I find annoying to use, otherwise it’s arbitrary), the UCR lists about 1.2 milliion violent crimes, while the NCVS lists 5.8 million (and obviously, that doesn’t include murder!). The conventional wisdom is that vehicle theft data is unusually reliable, because even victims who otherwise don’t care usually need to make a police report for insurance purposes. But the UCR has 721,885 vehicle thefts, while the NCVS lists 495,670. With that kind of disparity, it’s hard to say we’ve even figured out how to tell how much crime there really with any real certainty. That doesn’t make empirical analysis useless, but I do think it needs to be heavily discounted.
2. The three strikes laws are designed to increase punishments for what would otherwise be relatively minor crimes, so intuitively it’s not necessarily surprising that the deterrent effect would be more significant on those crimes than more serious ones, which would typically carry a significant punishment either way.
As far as European data is concerned, I would expect that their ability to apply criminal punishment more quickly would also have a meaningful effect on deterrence (which would carry over into the aftereffects part as well, I suppose).
3. “(How much do we care about drug crime? Tabarrok says somewhat, because drugs can be a “gateway crime” that gets people into the criminal lifestyle. Roodman says not much. It’s either personal use, which is mostly victimless, at least compared to other crimes).”
Overlooked, I think, by people who aren’t directly involved working in or otherwise seeing the criminal justice system (I certainly didn’t know it until I got involved), is the extent to which virtually all criminal activity except for sex crimes, domestic violence, and certain complex financial fraud is motivated by drug addiction. People don’t shoplift, steal cars, break into houses, or rob liquor stores because they want merchandise, cars, personal property, or money for its own sake: they do it so they can use the stuff they steal to get drugs. (Yes, it’s not literally everyone. But if we could eliminate all crimes that weren’t motivated by drugs, we’d be in pretty good shape.) So even if you don’t care about people abusing drugs per se, that abuse leads directly to the crimes that you almost certainly do care about, and there’s a real public interest in stopping them. (That’s not to say that our current way to serve that interest, of course.)
With respect to dealers, my impression (from seeing data, but I don’t have it in front of me) is that drug use in America is meaningfully down since we began aggressively prosecuting traffickers ~35-40 years ago. And intuitively, it seems hard to believe that drugs would be harder to get if they were being sold by Phillip Morris or Pfizer or whatever, or that you wouldn’t see more people using them if they were available that way. But again, I don’t have hard numbers to back this up (could be another interesting project!).
4. I’ve read a lot of the “aftereffects” literature trying to correlate prison term length to negative post-release outcomes, and I’ve yet to see one that was very persuasive in making the case for causation. Imagine two defendants with similar (or identical!) criminal records who are both charged with the lowest degree of assault – but the first barely injured the victim at all, while the second injured them so severely as to almost cross the line to the higher-level offense. I haven’t seen a study set up to capture that difference (much less control for it) – but that is, in my experience, precisely the kind of things that tends to lead to a higher sentence, and it’s at least plausible that the person who would do something like that would also be a higher risk of reoffending, independent of the sentence length.
5. I think you’re substantially underestimating both the cost of crime, and the frequency per criminal (even if you ignore drugs – after all, most drug dealers and users commit drug crimes weekly if not daily).
The NCVS for the last few years has had ~18 million violent and property crimes a year. If we estimate (based on inmate numbers) something like 2 million people in the criminal population, that puts 7 as the low end, not the high end.
On the other side, court-awarded restitution is a terrible measure of the actual cost of crime, even in purely economic terms, and certainly if you’re interested in measuring utility (which I assume we are). Roodman puts it pretty well himself: “This approach has generally failed to include intangible costs of crime for communities, as distinct from thevictims and perhaps their families. Almost everyone in a neighborhood feels some harm, for example, when assaults and robberies become more common.”
Excellent post thank you. After reading this I’m convinced that our main priority should be hiring a lot more police. The stats are clear, it makes sense, and you see it in the relative policing and crime rates of the US and Europe . Also it’s hard to cut through all of the statistics and adjustments and variations but it seems clear to me that violent repeat offenders should be getting very long prison sentences for incapacitation and so they age out of violence. Commit multiple violent crimes and you just give up the right to live free in society, we just shouldn’t tolerate that. I think there’s an underestimation of the trauma of shootings and armed robbery/carjacking and what it does to the psyche of people that live in environments where such things are common and the costs are higher than estimated.
Ok, so longer prison sentences and increased incarceration and three-strikes laws do decrease crime -- primarily because they remove the criminal from the environment where he can commit crimes. The problem is that for this method to work, we'd need to drastically increase the number of prisons, and doing so is prohibitively expensive. Thus, here is my Modest Proposal (tm): we institute a three-strikes law, where Strike 3 is summary execution. Now we don't need nearly as many prisons, and we get the maximum possible incapacitation and minimum possible aftereffects. What could possibly go wrong ?
I would object on the grounds that execution is a punishment that is out of proportion to the crime: but I have been told that punishment is a barbaric goal for a justice system, and that we should focus on deterrence instead. Your proposal passes the deterrence requirement with flying colors!
If we substitute execution with citizen revokement I think it may actually be a workable idea. Although making someone having no citizenship has already been banned by UN.
"Here's the deal. Option A, you get to die, maybe of old age, in a fancy concrete box with some other very unpleasant people. Option B, we file some paperwork at an embassy, in a week you get a Mongolian passport with special notes saying you're not welcome here, or in Russia or China. Then you get set up with a cargo-ship job to start off your new life, and if you set foot in this country again we'll shoot you, but otherwise it's all somebody else's problem. Which do you prefer?"
The Supreme Court has said it’s also unconstitutional. Trop v. Dulles, 356 U.S. 86 (1958).
See discussion above on how California's Three Strikes Law decreased crime much less than hoped. I don't see why this would be any better - in fact, it would probably be worse, given that resistance to executing people for minor crimes would limit it to the most severe criminals.
The marginal prisoner in Massachusetts may be a much badder dude than the marginal prisoner in Louisiana. But here is anecdotal reason to believe he is not:
Last week in New Orleans, there were three mass shootings. Two of them happened at the same second line in New Orleans East, 45 minutes apart. In two separate incidents, a gunman opened fire at a large outdoor gathering. In total, two people died, and ten were wounded. This is not national news. No arrests have been made.
The third shooting took place in the French Quarter four days later. One of the three shooters was caught immediately. Turns out, he had already served seven years in prison for armed robbery. Last year, he was arrested again and charged with domestic battery, child endangerment, and possession of a firearm by a convicted felon. A plea deal got him out on parole, the terms of which he violated near daily according to his ankle monitor's logs. He was wearing the ankle monitor at the time he opened fire on a French Quarter street, wounding three people and killing another.
For added spice: at a second line the weekend after all this, NOPD cracked down and arrested seven people for illegally carrying firearms. One had (allegedly) brought both his gun and his drugs to the parade. One was a felon (allegedly) caught with a stolen firearm.
It appears that Louisianans walk around committing felonies like all the time, and it's totally normal, and even after you're convicted of multiple violent crimes, you can keep violating your parole and it's no biggie until you commit a mass shooting.
The culture around both alcohol and firearms is very different in New Orleans vs. Boston, as well.
Really excellent post. I loved this. The only downside to crime posts: the inevitable authoritarian halfwits in the comments trying to outdo each other.
"No one wants to give life sentences to serial shoplifters more then I do!"
"Shut up, Carl. I hate shoplifters MORE. That's why I'm going to imply they should be killed, and then when somebody objects, I'm going to act innocent and defensive because I used weasel words and didn't explicitly say what I meant!"
You should probably engage with those fantasy posters where you see them rather than the straw men you’ve concocted in your head.
Even if some values fundamentally differ, presumably an acceptable compromise can still usually be reached? That's the premise of liberalism anyway, and it doesn't seem like there's anything better to replace it with yet.
These people are entirely real. The vigilantists and extremists have made their presence very well known in these comments, right on schedule. Perhaps you ought to read further. I feel no need to hold your hand by screenshotting the offenders. The comments are right there for you to read, if you are so inclined.
So engage with the actual comments.
That would be the rhetorical equivalent of banging your head against a locked door in an attempt to seek entrance.
Fascinatingly, Joseph Smith's 1844 presidential campaign platform included prison abolishment and using the buildings as "seminaries of learning". He wrote,
"Petition your State legislatures to pardon every convict in their several penitentiaries, blessing them as they go, and saying to them, in the name of the Lord, Go thy way, and sin no more. Advise your legislators, when they make laws for larceny, burglary, or any felony, to make the penalty applicable to work upon roads, public works, or any place where the culprit can be taught more wisdom and more virtue, and become more enlightened."
The new DA from LA in an interview cited some truly ridiculous level of benefit of reintegration programs for former prisoners so that may also be part of a comprehensive solution: More cops, more courts, more reintegration, and with those things the overall incarceration rate will naturally go down over time as crimes are prevented in the first place. Someone with experience in the field told me there are some dumb problems with hiring cops as well, like requiring everyone who works in law enforcement start as a beat cop, which creates unnecessary disincentives for people who are physically unsuited to working as beat cops or overqualified for the position.
I've heard of the "detective fast track hiring" idea before, and while I understand the appeal, I don't think it's a good one. Increasing the segregation between investigation and patrol is likely to result in problems for both.
What's absurd is that it's never even tried. If it were tried there would be data on what its effects are. Instead there's this religion of how things must be done even though it has obvious downsides.
Because trying and failing means more dead people, unsolved crimes, increased crime, disorder, and public dissatisfaction. Trying and failing, when a lot of other people within the profession share my skepticism, means heads roll both literally and figuratively.
A patrol officer tends to be a jack of all trades-conducting investigations, tactics, traffic control, surveillance, gathering intel, people skills, interview and interrogation, identifying and gathering evidence, warrant service, obtaining warrants, presenting cases and providing testimony in court, etc. You tend to find out what you're drawn to and/or good at as a part of the job, and lean in that direction, both officially and unofficially. Got a passion for arresting DUI's? Maybe get transferred to the DUI task force, or get some extra training, equipment, or assignments for DUI investigations. Enjoy finding wanted people? Get assigned to the Marshal Task Force, get some extra tactical and intel training, maybe get a K9. Find yourself enjoying everything surrounding investigations? Become a detective.
The problem with skipping all of that is threefold. First, people don't actually know whether they would be interested in that side of the job until they have some exposure to the reality. Second, all of these skills need to be practiced to become proficient. It's why so many jobs still use some form of apprenticeship. If I make a mistake on a shoplifting call, that's a lesson learned; a mistake that tanks a homicide investigation is the wrong time to learn that lesson. And third, the type of person who claims they could be a detective but don't go into police work because they would have to spend time as a patrol officer would not be a good detective. Sometimes being a detective means sifting through 40 GB of porn and three thousand text messages from someone's phone. Sometimes it means cracking jokes with a child molester. Often, it means being woken up at 3 in the morning because someone is dead or dying. Most police officers see the person who wants to be a TV detective but not a patrol officer in the same way as veterans look at someone who says they didn't enlist because they "would punch a drill instructor in the face."
You also benefit from the connections between patrol officers and detectives. Detectives need to be fluent in "cop" in order to properly comprehend police reports. Patrol officers need to be aware of evidence and investigation procedures in order to secure a crime scene, perform initial interviews, and write the initial report. Detectives then need to know what knowledge patrol will and will not have access to in order to reach out for further information. Identifying suspects from property crimes based on a picture, having a working knowledge of aliases or pseudonyms, or knowing which houses in an area are infamous for which events. This requires people move back and forth between patrol and investigation.
Then, you have the need for detectives to be functional as police officers. Those other skills, the tactical awareness, the basic understanding of how to act on patrol, are learned through experience more than they are taught. You can't invite a bunch of detectives onto a crime scene without the basic understanding of how to act on patrol without a lot of babysitting. There's also the flexibility requirement involved. Think of how Marines say "every marine a rifleman." For a lot of departments, the ability to mobilize every available resource in an oh-shit scenario-ranging from natural disasters, to critical incidents, to mass rioting-requires every (or at least almost every) sworn officer to be proficient in response to those scenarios.
In short, it's not a "religion of how things must be done," it's a serious of best practices as prescribed by decades of practical experience. To relate this to the engineering world, OceanGate gave us plenty of data on what corners you shouldn't cut in submersible design; that only cost the world five deaths. Crowdstrike taught us how not to release a software patch, and we still don't know exactly how much it cost, in money or lives.
Automatically excluding qualified candidates from detective jobs means more dead people, unsolved crimes, increased crime, disorder, and public dissatisfaction. Certainly the job of policing is important, but that's not a reason to shut down discussion about how to make it better, especially when your argument is based on tradition and intuition instead of data and there's a compelling counterargument.
At the risk of sounding perverse, I'm curious about other potential deterrents. How effective is say, Sharia law, at deterring violent crime or petty theft? I imagine that robbing a convenience store becomes very un-attractive if there's a good chance I'll lose my right hand as a result (beyond the pain, the social exile and bodily incapacitation don't sound like a good deal) . There must be research on the effectiveness of capital punishment as a deterrent, no? I imagine that's the best proxy we'll get for some of the more extreme forms of justice.
It's still a pretty poor proxy. "All things being equal" it would work, but in a lot of axis there's pretty sizable difference between countries with Sharia law or similar (including Europe not too long ago). Socioeconomic and cultural aspects aside, as has been pointed out, the certainty of getting caught for instance seems to matter more than the harshness of the punishments itself. And I personally suspect it's not the harshness in itself but emotional immediacy (which, if true, might imply corporal punishments would have effects that longer prison sentences wouldn't).
I think the better proxy would be 19th century when Europe/North America stopped using corporal punishments, especially if there are countries in which the punishments were common until they were banned (c.f. countries in which corporal punishments were banned because public opinion had already turned against them and the practice had already largely ceased).
Of course if one was certain that they would be caught, and lucid enough to care, they wouldn't commit the crime at all (given a severe enough punishment). It's just not clear to me that many of these shoplifters, for example, are worried about getting caught. It appears habitual for a large number of them, and they have not yet encountered a drastic enough punishment to stop doing it. They just don't care about being caught (or at least not enough). I think the majority of violent crimes are similar in that they are crimes of passion (while not habitual, not rationalizable either under this paradigm of being certain of being caught).
I like the point on "emotional immediacy" though. The ancient Greeks used exile as an alternative to death and it appears to have been deemed a worse fate. Probably because it was also a death sentence for the vast majority, but also because of the more potent emotional effect.
The severity of the punishment does seem to dissuade crime in wealthy eastern states like Singapore and Qatar, those are potential proxies if the data exists.
They're not worried about being caught because they've seen other people do it and not be meaningfully caught - it's not that they don't care at all, just a more-or-less rational assessment that the odds seem low enough to be acceptable. If they saw it happen reliably, they'd start worrying about it happening to them.
I don't fully buy this. Sure you can probably get away with petty theft without being caught on any independent offense. But how do we explain the presence of serial offenders -- ones that offend, get caught, receive a slap on the wrist, are released, and repeat? These serial offenders are disproportionately represented in the statistics. In these cases, the "rational" assessment is probably something like "when I've been caught before I received X, X isn't so bad". I don't think there's anything more sophisticated going on in terms of expected value analysis. If a thief was truly rational, then the way to dissuade them is to make it so (catch rate) * X outweighs (1-catch rate) * Y, where Y is the value of the stolen goods.
Note that the "Bloody Code" (execution for lots of non-murder crimes) was passed in England when urbanisation drove a huge increase in violent crime, and didn't much lower crime rates (executions were mostly committed to transportation to Australia for crimes that didn't kill anyone) and was then abolished accompanied by a massive drop in the crime rate - which happened at the same time as the introduction of the Metropolitan Police, who started catching far more criminals.
Interesting point! I don't disagree that the catch rate of criminals is a significant parameter. Increasing the catch rate when it's near zero, is probably the single most valuable thing a society can do to dissuade crime. The value on the margins seems slim however, how much more bang for your buck do you get at the 1,000,000th police officer or resource poured into increasing the catch rate of criminals? At that point you have to play around with other parameters, if you want to further dissuade crime.
Absolutely. The question is "Which margin is the cheapest to move?" Which is the sort of question that these giant meta-analyses are much better at answering than the sorts of argument we're engaging in in the comments.
Great article on an important topic. A judge once said (paraphrasing here) “There’s only two reasons to send somebody to prison. Either we’re scared of ‘em or we’re mad at ‘em.”
I am a felony prosecutor in a major US metro area—I should probably be prosecuting crimes right now rather than writing this—but I wanted to add two thoughts to the conversation. First, that I think that folksy judge was right. Retribution, which only got a bullet point in this article, shouldn’t been seen as seedy or as a somehow illegitimate end of the criminal justice system; it is an essential part of the glue that holds society together. A community where people don’t expect criminal-level misbehavior to be fairly and reliably (albeit not perfectly) punished is a lower trust community, with more brazen criminality, and just generally a worse place to live.
Second, it is quite correct to hold that the next marginal anti-crime dollar would be better spent on policing than on incarceration. However, that also isn’t an easy solution. Before you can actually do that in a meaningful way, you need to address what has become a generational challenge in much of the country: staffing police departments. The job has always been dangerous and (comparatively) poorly paid. Add in the decline in public esteem for the job—particularly post-George Floyd—and the rate limiting bottleneck to crime reduction in most metropolitan areas has become recruiting enough talented people willing to do that job. We just don’t have enough and you can’t manufacture them with a 5% pay raise or some other token effort.
Actually, one last thought on the article: deterrence has a qualitative aspect that is often overlooked or at least understated.
How about merging police unions and teacher's unions, forking over a set fraction of land-value-tax revenue from the region they're responsible for, then giving the union broad latitude how to handle quantity vs. quality tradeoffs of salaries within that budget, with outside oversight mostly limited to checking that announced and implemented policies match, and removing the parties responsible when there's some egregious misconduct?
How about it? Or, more to the point, can you elaborate on why you think that would help address the issues that TheAnswerIsAWall was mentioning? It strikes me as orthogonal at best.
Teachers and police officers have societal functions which often overlap: being the officially-designated "adult in the room" around a lot of people with poor impulse control, teaching those people the lessons they need in order to become functional members of society. Both groups are, in the US at least, understaffed, underpaid, overburdened with useless paperwork, and widely despised. Teachers are often hindered in their ability to help the best students by lack of authority to remove the most disruptive ones; wouldn't surprise me if cops also have major problems they could solve more efficiently by handing out academic exercises.
Whole point of having a trade union in the first place is for people doing similar work to get together, pool resources, and push for their collective class interests, rather than competing - dumping problems on each other - to mutual detriment. People talk about a "school-to-prison pipeline," which sounds to me like teachers sending off problem students to be dealt with by the cops, while "ignorance is no excuse" is law enforcement trying to shift blame back the other way. Get 'em all in the same room together, with a simple incentive structure that rewards overall success at making the community a place where more people want to live (rather than bureaucratic ass-covering), hopefully they'll hammer out better solutions.
Great post! Coincidentally, it agrees with my priors. :)
Even with footnote 3) I was chafing at the conflation of arrest and conviction for the majority of the piece - the philosophical distinction alone is large, but you addressed that the critical component is the process and expense involved with actually securing that conviction. My understanding is that's where most of the resource constrains show up, and was ultimately satisfied by that treatment here.
I was disappointed that the deterrence effects of certainty and swiftness were explicitly outside the scope of this post. That's a reasonable approach to the topic for now, but what I have seen agrees with the later aside that certainty at least is a major driver of deterrence effects. It's also even harder to study than severity, so understandably tough sell. Swiftness might be worth a look though, anyone happen to have a link to good research?
"But Italy has a lower incarceration rate than the US, potentially meaning that the average criminal they incarcerate is worse"
Maybe they just have fewer criminals?
It seems like you're assuming there's a fixed amount of criminality available in society. (I vaguely remember you mentioning some historical precedent involving Quakers that supported this on SSC a long time ago.)
Even if you grant that there's a fixed amount behavior that society can criminalize, that amount might vary between societies. It may be that Italy just has fewer criminals.
>Might a person who starts crime at 15 be a more disturbed and committed criminal than one who starts at 20? Seems plausible. I think this might be responsible for a lot of the seemingly positive effect of sentences > 5 years.
Isnt it the opposite? If theyre the same age at release, then those with the long sentence were imprisoned younger, and so started crime younger, so this should make the long sentences look worse.
>My strategy was to use the pro-incarceration and anti-incarceration reviews to determine a sort of “Overton window” of what smart people thought they could get away with claiming. Then, if the window was very wide and I couldn’t figure out where I landed, I used the apparently-unbiased Nagin as a tiebreaker.
My thought on reading this was that the non-Nagin studies will contribute very little to the final conclusion, and it seems that was right.
Great piece. My left-wing brain is puzzled by one seemingly glaring omission though: someone help me understand the focus on recidivism as the primary "aftereffect" worth exploring. Why should I care about recidivism in particular? I care about crime rates. My only question is "do crime rates go up or down with increased prison sentencing?". There are many narratives which suggest that crime rates would go up with increased prison sentencing. For example, "destabilization of communities with high incarceration rates means kids growing up poorer and without fathers which means increased propensity to grow up and become criminals". This could be true even if every ex-convict became a saint upon release. This seems a critical piece of the analysis of the value of prison, yet one that was entirely left out. Perhaps it just wasn't in Scott's scope?
The El Salvador data, at least, addresses this, by showing that murder rates plummeted after they increased incarceration. Of course, we'll have to see if these results persist over a generation or if this is just a band-aid solution.
If every ex-con became a saint upon release, presumably some of those saints would step up to care for the children of the not-yet-saints who were behind bars at any given time.
> For example, "destabilization of communities with high incarceration rates means kids growing up poorer and without fathers which means increased propensity to grow up and become criminals".
That's not a good example; the relationship between being raised without a father and criminality is completely confounded by the fact that those missing fathers are mostly criminals.
If you separate out children who are raised without fathers because the father chose to leave or was jailed, and children who are raised without fathers because the father died, the boost to criminality is entirely absent from the second group.
That’s very interesting, thank you for sharing! Do you have a source for this? I assume it’s harder to prove that in a counterfactual world where the criminal parent is present (having never been prosecuted), the kid has the same likelihood of becoming a criminal.
That said, when you call it a bad example, it seems you misunderstand me, because it wasn’t meant to be obviously correct, merely plausible, to illustrate that a post like this should address a broader idea of aftereffects, like you just did.
It seems odd to me that turning innocent people wrongly imprisoned into criminals isn't taken into account at all in the "Aftereffects" discussion. As Scott says later, anywhere from 2-5% of people in prison are actually innocent, and the recidivism rate is about 80%, so prison is turning about 1-4% of it's population from innocent law-followers to criminal law-breakers. It's not huge numbers, but at the margins we're considering, it seems like it's at least worth talking about when considering criminogenic effects?
I agree that the actual recidivism rate will be different for the guilty vs. the innocent, and I actually started to include a note in the comment about that, but when it got longer than the actual comment I decided to just cut it. In the end, the same forces that prevent most ex-convicts from living a normal life will apply to both groups, and I have no way of knowing the actual rate, so 80% will have to do.
On the other hand, the claim that "large fraction of them are guilty of other crimes" does seem completely unjustified. What are you basing that on? When you look at actual wrongful convictions that have been overturned, most of them are just people in the wrong place at the wrong time. I'm sure some of them are guilty of other crimes because that's basically true of literally any group, but I can't see any reason to believe it's a "large fraction".
"When you look at actual wrongful convictions that have been overturned"
How much of this confounded by the ones that are publicised and the ones where the Innocence Project and the like are investigating are the ones where it's "wrong place, wrong time" and the ones where the police picked up a usual suspect who had done ten burglaries that week but not the one they convicted him for are not really of interest?
As Melvin said, it's completely nuts to think that the 80% recidivism rate applies equally to the rightly convicted and the wrongly convicted.
Great article, Roodman does not come off so well.
The idea that the more people incarcerated already the less there is to gain from incarcerating more people is obvious enough, but Scott’s clever use of the extra guy being arrested in a country with 50% of the population already in jail vs the first guy in a country which has arrested nobody is a clever debating tactic, which I will use. Also I will use the word crimogenic as if I’ve always known it.
But two points.
1) What’s missing from this deterrence story is the heretofore honest citizen who isn’t registered as criminal yet. Jailing nobody would deter none of us from speeding, drinking, drink driving, or shop lifting (etc) and even murder. Why not kill the wife if she’s pissing you off? I’d be inclined to grab a few iPads myself if all it entailed was a nice smile from the shop manager, and a mere finger wagging from the local constabulary.
But in many ways the debate is sterile. We arrest people for deterrence and hope for rehabilitation, but mostly we do it for retribution. A man who kills his wife in a fit of passion but is genuinely remorseful is probably no future threat to society. He’s likely not a serial wife killer. He may find prison crimogenic (a word I’ve always known) but what of it? Send him to jail anyway because he killed his wife.
If we don’t do that, state retribution will be replaced by personal revenge. That was the point of law to begin with, to replace revenge and feuds with process.
Urge to get personal revenge probably evolved for deterrence-related reasons, though.
I think that’s partially true but revenge is innate anyway. Even if you thought that killing the guy who killed your son would not deter his clan - but rather encourage them - to attack your clan, you would do it anyway.
Civilisation is about replacing arbitrary and personal conflict with laws and process, replacing chaos with order.
Under an evolved-for-deterrence theory, the point of attacking your son's killers isn't to discourage future attacks on surviving members of your clan by the specific parties responsible, it's to retain the credibility of the original signal.
Is this a theory that you just made up and have created a compound word for? Kudos but also - What’s this - German?
Anyway I don’t think revenge is about deterrence in the most part. Deterrence is something that a group wants, not the individual. Revenge is personal.
Good work. Thanks.
Crime is cultural. It is a set of behaviors people learn from those around them. It cannot be understood in terms of individual incentives and individual decisions without understanding how the cultural context shapes those incentives and decisions. An analysis which looks at crime on a single person basis will *always* miss this force, which is the most important one out there. Criminality is self-perpetuating the same way that any other habit or career is, and changing it is like any other cultural change -- very hard!
From the section on super offenders:
>In the five years following the passage of the law in 1994, crime dropped 33% in California compared to 26% in the rest of the country, meaning only 7% that you could optimistically attribute to Three Strikes.
Hold on, should we really be looking at the short term? During the section on deterrence, you describe the California three strikes as guaranteeing at least a twenty-year sentence. In that section, looking at immediate effects is fine, since we were talking about potential criminals imagining the consequences and choosing not to crime. But if your theory is that they can't commit crimes because they are in prison, excluding the vast majority of the extra time the spend in prison means excel most of the effect. I'm pretty sure our courts won't let the legislature retroactively increase the sentence of prisoners partway through their sentence, much less rearrest released prisoners, so the effect of the three strikes law is at most (not including deterrence) to reduce the number of arrests per criminal (within a certain subpopulation) to one. You don't specify what the expected rate of arrests was before the three strikes law, but since you talk about how the law was tailored for the worst of the worst, I would guess that the incapacitation effect was already significant over the five-year period. If the law is turning five-year sentences into twenty-year sentences, the expected effect over the next five years (again excluding deterrence) is literally zero, since everyone sentenced under the new lay would have been locked up for the rest of the study period even without the new law. If the typical turnaround between arrests is around 2.5 years (counting not just time in jail, but also time between release and reoffending and getting caught) then you can prevent the need for a second arrest for a 50% reduction in crime (among the affected subpopulation) compared to an 87.5% reduction in crime (same subpopulation) over 20 years. Obviously looking further into the future creates more potential for other events to mess with the data, but unless the previous laws were very soft the short time period is cutting off a huge portion of the effect, and scaling up to account for the fact that you are only looking at the tip of the iceberg doesn't work well when low numbers will get rounded to zero and assumed to be the result of confounders.
Editing to add a nitpick:
>That would mean that, if prisoners commit 10 crimes/year after release, the same people, given a noncustodial sentence, would commit 8 crimes/year.
>In order to neutralize the effect of one year imprisonment (-10 crimes), the negative effects of incarceration (+2 crimes/year) would have to continue for five years. But they probably won’t, because we know that most of these people get rearrested sooner than that anyway.
One year of imprisonment is only preventing eight crimes, since that is how many they would have committed if they weren't imprisoned. So the increase in crimes would only have to last four years after release on average to balance out. It doesn't seem to meet that bar either, so the conclusion is unchanged, but it bugs me that you are giving imprisonment credit for prevent hypothetical crimes that were only on the table due to imprisonment.
>>"Aftereffects must vary by place, time, and person."
This is by very far key to all this, surely?
This post is a sterling effort (as all "Much More Than ..." write ups, but probably the least ultimately satisfying of the ones I've read. I think partially because it doesn't make enough of a point of saying IT'S ABOUT THE US, and while it tends to stick to the American data/analyses it also makes visits to other countries' data (suitably caveated).
In the UK criminology I believe the received wisdom is that short sentences (not a day or there, I don't think anyone gets three days in jail here ever) are the most damaging, because the have disruptive to destructive effects on the prisoner's life while giving no chance for any rehabilitation to take place (leaving aside the fact that neither of the UK countries is exactly a beacon for prisoner rehabilitation anyway).
But it feels to me still that discussing prison effects without taking into account the specific prison environment and what happens there apart from being locked up is too general.
As to the social desire for "justice to be done", again, I feel that when people talk about crime concerns they really don't mean someone walking out if Tesco with a packet of cheddar and a bottle of wine (one case I recall from my last visit to court) or a drunk neighbour yelling abuse at his equally drunk wife (another case) or someone caught with a baggie of weed and another one of blow (again, same court).
That's an excellent point. The US is sort of in a unique situation with one of the higher crime and incarceration rates in the world. Countries elsewhere along the curve may have different solutions.
"Prison is less cost-effective than other methods of decreasing crime at most current margins. If people weren’t attracted by the emotional punch of how “tough-on-crime” it feels, they would probably want to divert justice system resources away from prisons into other things like police and courts."
Why not both? I imagine most pro-prison people are also pro-police, just as people who favour spending more on healthcare believe in spending more on both diabetes and cancer.
I remember reading about crime and punishment in 18th-century England, when hanging was the scheduled penalty for everything from shoplifting to murder. It was also the early days of 'criminology,' and thoughtful observers couldn't help noticing that the death penalty seemed to have no deterrent effect (although it certainly had 'incapacitation' and 'recidivism' covered). Nor did the prospect of going to hell after being hung. What was the answer? Fortunately, a real-world option presented itself around the time that America refused to take British convicts: go to the southern hemisphere - which must have been like going to Mars - settle New South Wales with convicted felons (bonus: claim the land before the French did, probably with better outcomes for the indigenous people), and start a new country. And on the whole, it's worked out well. Not every jurisdiction has that option, though.
Strange statement about the French. By and large Anglo imperialism was worse than other European powers in its treatment of the natives, and Australia was very very bad - classifying aborigines under flora and fauna until 1967.
The flora/fauna thing is completely false. https://aiatsis.gov.au/explore/right-vote which explains that many Aboriginal people had had voting rights since the 19th century. Also, France had slavery in its colonies until 1847 (after a brief hiatus 1794-1802), so I'm not sure why its imperialism is considered better.
Many is clearly untrue. Some. And no voting rights for most until 1967. Whether or not the French were always great, the idea that Australia would have been worse for natives seems odd, as it really couldn’t have been much worse.
I don't mean to defend colonisation in any way, or get drawn into a comparison with the French, but thinking that it couldn't get any worse is a failure of imagination.
What if the Tasmanian genocide had been 100% of the population instead of 20%? What if the token efforts to stop blackbirding hadn't happened at all? What if blackbirding had been on the same scale as the transatlantic slave trade, or under conditions even closer to those the Belgians implemented in the Congo? What if the Stolen Generation had all been sent to death camps instead?
True. As you've probably read, it actually increased murder, because if someone caught you stealing and the punishment for both theft and murder was hanging...well, better not to leave any witnesses.
Probably the most extreme example I can think of is Imperial China, where several revolts, including one leading to the establishment of a new dynasty, resulted from people making a similar calculation.
"Hey Liu Bang, the prisoners escaped!"
"Crap! What's the punishment for letting prisoners escape?"
"Death!"
"What's the punishment for rebellion?"
"Death!"
"Well then..."
The eventual solution was to replace transportation with prisons, massively reduce sentences ... and create the Metropolitan Police.
> Large increases in incarceration cause large increases in crime.
Typo, I think "increases" should be "decreases".
> making some good points bout ways
Another typo(?)
The answer to the crime problem is finding a less expensive but just as incapacitating alternative to prison. I recommend exile
I disagree with posts like these that talk solely about the criminal's perspective and center the crime they committed without contextualizing with the victim's perspective.
The point of punishing criminals is not just about them--it's about showing their victims that society gives a shit about what happened to them. Maybe stealing $250 worth of stuff sounds like a minor crime to the people on these and the EA blogs, but that's enough to ruin a poor person's life for a while, depending on what was stolen. Ruining someone's life and destroying their trust in society deserves punishment, period, even if their stolen goods aren't fancy enough to matter to rich people.
As an alternative possibility, how about giving everyone enough stuff so they can afford to lose $250 of it without that being life-ruining?
So... if I'm understanding correctly, your counterargument to the comment I was replying to is that society does not, and should not, care what happens to poor people? "Screw you, Jack, I got mine," more or less, as anyone who hasn't already pulled themselves up by their bootstraps must, ipso facto, be morally unfit?
If that's your position, I think our respective core values and assumptions may be too far divergent for much further productive discussion to be possible, though one way or another I'd be happy to be proven wrong.
That doesn't solve the problem of broken social trust. Also, why would you be okay with a system that essentially gives a free $250 to people as a reward for doing something bad?
If everybody's getting UBI that's adequate for typical subsistence needs, there'll probably be less theft in the first place, when somebody does it anyway they can be prosecuted vigorously (in fully justified confidence it wasn't for bread to feed their starving family), and the victim is in a better position to focus on obtaining longer-term justice since they're less immediately busy with damage control.
Where are you getting the idea that people are shoplifting to feed their families?
I am deeply sceptical that a significant part of the decline in homicide in El Salvador was due to mass incarceration, unless the flow of time and causation has reversed itself.
That graph of homicides per year you posted has a really sharp peak in 2015, at around 100 per 100,000 people per year. Thereafter, it starts to plummet dramatically every year.
So El Salvador clearly got /something/ right, and that something happened around 2015.
The thing is that mass incarceration in El Salvador started in 2022, by which time the homicide rate had already fallen massively, to about 10 per 100,000, and was still falling rapidly. That improvement didn't accelerate in 2022 - yes, it continued, but if anything it slowed.
So I think this is almost certainly correlation not causation - unless falling homicide in El Salvador somehow caused mass incarceration!
Falling homicide may well have caused mass incarceration, if a decrease in gang power made it politically possible to go after the gangs.
Previous presidents bribed the gangs to spare civilians, and it sorta worked for a few years.
There even is a historic equivalent, known as Danegeld.
This doesn't quantify by far the most important effect of prison: The reduction in time available for increasing the size of the next generation of high-probability criminals.
The well documented population criminality difference exists because of thousands of years of cities enacting population-altering punishments.
This effect has a compounding interest which completely overwhelms all other effects. It's utterly negligent to claim to comprehensively cover this issue if one isn't willing to address this factor.
One thing that worries me about the incapacitation argument is the apparent assumption that being a high-volume criminal is an innate trait of an individual, rather than a positional role in society.
Like: I don't personally know how you turn stolen bikes and cosmetics into cash, but assuming there's some black market for those things, does removing a bunch of existing suppliers not just increase the incentives for new suppliers to get into the game, like any other market?
I don't really know how drug dealing works, but if you arrest all the people selling the drugs, do the people making the drugs not just recruit new people to sell them?
If there's some cultural conglomerate that produces criminality (like a gang/other criminal friendgroup of some kind, or maybe a really poor neighborhood where the cops never help and everyone assumes it's them vs the world and doesn't feel a moral obligation to follow the law, or w/e), does removing some members of that group create opening that new people enter?
Etc. I don't know enough about how the people who commit lots of crimes get into that position, but without a really exhaustive understanding of it I would not natively assume that those people are not replaced when they get removed. In the same way that I would assume that shutting down the top 3 shoe retailers in the country would not decrease the total number of shoes being sold 5 years later.
Sure, you'd expect to use standard microeconomic tools to model a market for stolen goods or illegal drugs. Disincentivizing people from supplying them (by creating a risk that they'll go to prison for a long time if they do) would be expected to shift the supply curve left, increasing supply and reducing the quantity produced. The latter, of course, is what we want to achieve.
Right, it would shift it, but probably not 100%.