Links to Consider, 11/5
Joseph Politano on signs of an economic slowdown; Yascha Mounk talks with Sam Harris; A Debate on Incarceration; Misha Saul on Greelights
aggregate spending growth has tracked gross labor income rather closely throughout 2021/2022 despite the massive influx of cash from federal stimulus programs—something Matt Klein has pointed out several times.
…Aggregate personal spending growth is outpacing aggregate personal income growth, and the value of households’ financial assets has also fallen.
Overall, I wonder if we’re seeing some unsustainable momentum in the economy. Businesses may be hoarding workers because they have a hard time hiring, but they cannot do that indefinitely if they see revenue start to fall. And consumers are maintaining spending a bit beyond their means, but that will not last.
Yascha Mounk talks with Sam Harris, who says,
there's not a lot of active racism keeping qualified black people out of good jobs and educational opportunities and all the rest of it; you can also say that, at the same level of qualification, it is a positive advantage to be black at this point, in almost any part of society that would be truly desirable to a qualified candidate, whether you're looking for an educational opportunity, or you want to work in media, tech or in a Fortune 500 company. Being a black applicant, or being a person of color more generally, at the same level of qualification—America is your oyster at this moment. This is just a fact. You talk to anyone at any of these companies; you talk to anyone in Ivy League admissions, in medical schools, any nonprofit; they are desperate to hire qualified people of color. And if you're white or Asian, you're at a positive disadvantage, generally speaking. It is a fact, and basically everyone knows it.
And so, what's happening on the left is that you have a generation of activists determined to lie about all of that
On Bari Weiss’ substack, Rafael Mangual says,
I believe the bigger issue is under incarceration. That is, those who have come into contact with the criminal justice system and should be incarcerated, but aren’t.
Some respond that it would be better and more humane to invest in rehabilitation so that there are fewer incorrigible offenders. The problem with this is that it presupposes that we know how to rehabilitate criminal offenders at scale. We don’t.
In counterpoint, Lara Bazelon writes,
Does the system make mistakes and let dangerous people out? On occasion, yes. But more often than not, it locks away people for far longer than necessary, and to deleterious effect.
The risk of de-carceration will never be zero. But neither is the cost of over-incarceration. As this Brennan Center report found, by using criteria that separates those most likely to reoffend from the rest of those incarcerated, the U.S. could release about 40 percent of our prisoners without harming public safety.
To me, this sounds like a classic case of Type I and Type II errors. You can make the mistake of keeping a harmless person in prison. You can make the mistake of letting an incorrigible repeat offender go free. Other things equal, if you commit fewer errors of one type, you commit more errors of the other type.
I’ll wager that using artificial intelligence you could reduce both types of errors. That is, an AI that has been trained on data could provide better estimates of the probability that letting someone out of prison will endanger society. But I doubt that the legislators, judges, and lawyers within the criminal justice system would be receptive to such an approach.
The reason McConaughey reads like he’s from an alien culture is because he is. Greenlights is the perfect pairing to David Hackett Fischer’s Albion’s Seed: Four British Folkways in America as a Platonic example of the Appalachian hero. Where Hillbilly Elegy is forceful, resentful and pessimistic, Greenlights is unapologetic, glorious, optimistic. Two opposite perspectives on an Appalachian culture laced with physical violence and substance abuse.
Thanks to a commenter for the pointer.
I'm not sure that the "under vs. overincarceration" debate is correctly reading the facts on the ground within the court and corrections systems. Generally, courts aim to minimize the use of incarceration in large part because the "carrot" for many offenders is more attractive than the stick. This is also not well reflected by the statistics that pundits use, in part because one of the major carrots is the expungement of criminal records, the dropping of charges, and negotiating a lower set of charges in return for good behavior. It'll undercount the reform success stories because their records get wiped and they will be under-charged anyway at the outset of their short and truncated criminal careers. It'll overcount the recidivists because their records will not get wiped and courts will lard on all the charges that they are likely to be eligible for rather than allowing the defendants to plead out of the worst charges.
Judges want to suspend sentences in return for good behavior, and if the offender screws up, the stick comes out. For reasonable offenders, this works fine -- the results can be close to the ideal. However, when this system is taken to the extreme and many of the offenders are not reasonable but are habitual criminal degenerates, the rational approach doesn't work, and no amount of reform will transmute the hardened criminal or lunatic into a reasonable person.
I think it's also wrong to think that the system "invests in rehabilitation" or that it really could do much along those lines. There's not really any rehabilitating going on that does not come from the offender themselves. There are serious limitations to what the state can do on this point; it does a disservice to pretend like the state can do it rather than focusing on what civil society can do to give reasonable convicts a better pathway to productive lives. The whole theory that social workers do anything positive at all doesn't seem well supported by the evidence of the 20-21st centuries.
Further, Bazelon's comparison to other countries that aren't the US makes no sense whatsoever from a legal policy perspective. It's worse than comparing oranges to starfruits. There is almost nothing useful that you can garner from such comparisons because everything about those foreign countries from how the legal systems work at a basic level to demographics to geography etc. are all totally alien. The studies she cites suffers from the same data pollution problem that I mentioned earlier. You can throw 'em all out.
You don't need a complex AI to identify a great many people who are at low risk of reoffending. Some are in prison for purely victimless crimes; some - especially men - are much older than when they committed the crime, and crime is strongly correlated with testosterone levels in men. I wouldn't be too surprised if it were found that you could release almost any male criminal at 30 who committed a crime before 20 with very little recidivism. My cutoffs are fairly arbitrary, of course.
And incarceration rate is almost certainly the wrong variable to look at. Becker's model of criminality has been empirically challenged. Probability of getting caught is generally more important than magnitude of punishment. You could probably clean up the streets a lot just by ensuring that people committing low level offenses - burglary, theft, less severe assaults - get caught, even if they only spend a few months in jail or community service. One underrated explanation for the 90s crime drop is the security hypothesis (https://en.wikipedia.org/wiki/Crime_drop#Security_hypothesis) which more or less implies that serious crimes like murder are much less likely to happen in a society where you can't even get away with robbery or stealing a car, since many killings happen in conjunction with such crimes, though also because fewer people will get on the first ladder of the cascade toward becoming a hardened criminal.