Is Mass Incarceration Inevitable? Part 3. What About Violent Crime?

Are we prepared to be more lenient with violent criminals?

|The Volokh Conspiracy |

Reducing the national incarceration rate is a test of our political will.  Keeping a lot fewer people behind bars might well lead to an uptick in crime, especially if we release more inmates without paying for the support they need for successful reintegration once they are released (more on this in the next post).  And the reform efforts to date suggest that our commitment to reducing the prison population may not be matched by our willingness to take the risks necessary to make it happen.  This can been seen most clearly in our treatment of those who commit violent crimes.

Many (although by no means all) of the reform efforts in the states and by the federal government concentrate on reducing the fact of incarceration or the sentence length for those who commit property crimes or drug offenses.  This is hardly surprising.  Violent people scare and enrage us, making the decision to lock them up for long periods more palatable.

For the same reason, proposals to release those convicted of violent crimes earlier, through shorter sentences, easier access to bail, or increased sentence credits, are politically difficult, because the retributivist feelings stirred by these crimes is higher, which in turn makes it harder to argue that it is "unnecessary" to keep these offenders behind bars.  And from a strict numbers standpoint, there are far, far more property and drug crimes committed than violent crimes, so if our goal is inmate reduction generally, it is entirely reasonable to focus on those convicted of non-violent crimes, and to say little about murderers, sex offenders, and armed robbers.

Even if we skip past the problems of drawing lines between violent and non-violent crimes (a frequent and unsatisfying source of litigation), there are two problems with excluding violent criminals from reform efforts.

First, if our goal is to reduce the population of inmates who are being held "unnecessarily," because they present a low statistical risk of reoffending (an idea mentioned in an earlier post), drawing a bright line between violent and non-violent criminals is hard to justify.  In an extensive study of recidivism of state prison inmates, the Bureau of Justice Statistics found:

  • After their release from prison, those who had been incarcerated for committing a violent crime were no more likely to be arrested for a new offense than those who had been imprisoned for any other type of crime.
  • Those imprisoned for a violent crime were only slightly more likely to be re-arrested for a violent crime than those who had been previously imprisoned for a property or public order crime.
  • However, those previously convicted of a violent crime were somewhat more likely to be rearrested for a violent crime that those previously convicted of a drug offense.

It is probably appropriate and inevitable that we exclude some violent criminals from reform efforts on retributivist grounds, regardless of their likelihood of reoffending.  The point here is simply that categorically excluding violent crime from reform proposals may be based on exaggerated assumptions about the risks of doing otherwise.

The second problem with focusing on non-violent inmates is a practical one: the numbers suggest that we are unlikely to escape a world of mass incarceration without doing something about violent offenders.  At the end of 2016, the state U.S. prison population looked like this:

Prison Inmate Population by Most Serious Crime of Conviction

Crime of ConvictionPercent of Inmates
Public Order (e.g., weapons, DUI)12%


As these figures show, excluding violent criminals from reform efforts would make the de-population task even more daunting than it already is.  Leaving the jail population aside, if we want to reduce the prison population by say, 20% we would need to release or divert from prison something like half of the drug offenders and weapons violators, and perhaps a third of the thieves and burglars.  (And even then, we would still have the highest rate of incarceration among the world's top-10 economies.)

The problems of excluding violent-crime inmates from reform efforts gets more complex when we add considerations of race and ethnicity.  As reported by the Bureau of Justice Statistics, African Americans and Hispanics are already over-represented in inmate populations compared to their numbers in the general population, and the numbers get worse if we focus on violent crimes.

Sentenced prisoners under jurisdiction of state correctional authorities, by most serious offense (as of Dec. 31, 2016)

As the chart shows, African Americans and Hispanics are more likely to be imprisoned for violent crimes than for property crimes—in contrast to Whites, where the opposite is true.  So if we decrease the percentage of those in prison for property offenses only, we will necessarily increase the percentage of those behind bars for violent crimes.  As a result, even if the net impact of reform efforts is to reduce the total number of prisoners, excluding violent crimes from these efforts is likely to make the racial and ethnic imbalance worse rather than better.

In his terrific book, John Pfaff warns that "[u]ntil we accept that meaningful prison reform means changing how we punish violent crimes, true reform will not be possible."  He's right, and perhaps the so-far successful efforts at reducing the population of mostly non-violent offenders will pave the way for a second wave of addressing violent offenders.  But we should neither assume nor underestimate the political will needed to take this second step, nor should we be surprised if it turns out that our entirely justified fear of violent crime and our feelings about those who commit them lead to a different calculation about the importance of reducing mass incarceration.


NEXT: Today in Supreme Court History: October 9, 1954

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. The release of violent criminals is a hard sell because of anecdotal, but common, news reports of homicides that often list the accused’s long stream of prior arrests for violent crimes that resulted in only dropped charges, probation, or early parole.

  2. This series reads kind of strange, because it seems to treat the actual purposes of incarceration, retribution and incapacitation, as mere obstacles to the obvious proper goal of incarceration policy, minimizing the number of people incarcerated. If only the public would consent to being victimized more, and to seeing their victimizers go unpunished, it would be easier to do the right thing!

    Might I suggest that the ideal approach to reducing how many people are incarcerated, is focusing on reducing crime rates?

    1. There was a comment thread on this in the first post on this topic, which if you recall you started. In short, doing what you suggest would lead to identifying who commits the most crimes, and then what to do about it. As you can guess, it gets complicated rapidly when people starting making the classic correlation = causation mistake.

    2. the actual purposes of incarceration, retribution and incapacitation

      If these were the only purposes, our criminal laws would look pretty different.

      You don’t think jails are punitive? Or that it’s been a very common policy argument over the past three decades that their conditions must be so, and sentences must be long to discourage criminality?

      1. Incarceration serves three purposes: Deterrence, retribution, (AKA “punishment”) and incapacitation.

        To the extent that deterrence is possible, it’s my expectation that just retribution and incapacitation are probably harsh enough to accomplish it.

        1. To the extent that deterrence is possible, it’s my expectation that just retribution and incapacitation are probably harsh enough to accomplish it.

          I beg to differ. My suspicion is that :

          (1) some crimes and criminals will be significantly deterred by the deterrent effect, even if the incapacitation effect is small
          (2) and vice versa

          Thus you may have some things – eg white collar fraud etc – which you might think have a low reading on the incapacitation index, ie the offender is unlikely to reoffend on release. If you set the punishment to “non custodial” because of this, you will lose the deterrent effect of what for this sort of crime may be quite a high reading on the deterrence index.

      2. Why shouldn’t they be punitive? Confinement is punitive. And you must think that long sentences do not discourage criminality. So if you eliminated the punishment altogether would there be an increase in the crime? I noticed that some prog cities are declining to prosecute what they consider to be minor crimes like shoplifting. Unexpectedly it seems that shoplifting is increasing.

        1. The difficulty is this; incarceration for petty crimes creates a generalized deterrent effect that is unmeasurable. Imagine if there were not police, like there wasn’t until the mid 19th century. Would crime be crazy rampant? No, generally speaking most people are civilized. But if there were no police, a certain subset of people will take the opportunity to commit crimes they wouldn’t have otherwise.

          Likewise, locking people up for shoplifting is expensive and putting people in jail/prison for stealing small items doesn’t do much of anything to reduce the crimes we really care about (rape, robbery, murder, etc.). Some people even argue that it’s unfair, which it’s not, but it certainly is expensive and taxes are high. But if you decline to prosecute, that group of uncivilized people will take advantage.

          1. I don’t think that a utilitarian argument wrt crime prevention and punishment is appropriate. You have a law you enforce it. If the law is worth having, and in my opinion laws against theft are worth having, you enforce it. Cost should not even enter the equation.

            1. First, I’m not making a utilitarian argument, unless by that you mean that that where I imply to only lock the worst offenders up argument is utilitarian. As for punishment, I’m naught but asking you to get you to be specific about a level of punishment.

              Second, lots of laws are on the books that I promise, you don’t think are worth having.

              Third, while I do agree that laws against theft are worth having, cost *always* enters the equation because trade-offs are inevitable with scarce resources. Maybe your local county sheriff would love to do more DUI roadblocks, but his overtime was cut to pay fund a new snowplowing division, and the elected officials didn’t want to raise taxes to pay for both simultaneously.

              1. “Likewise, locking people up for shoplifting is expensive and putting people in jail/prison for stealing small items doesn’t do much of anything to reduce the crimes we really care about (rape, robbery, murder, etc.”

                The first part is a utilitarian argument, the second is just that people shouldn’t be punished for stealing small items. Ok we can agree to disagree on that.

                “Second, lots of laws are on the books that I promise, you don’t think are worth having.”
                See my first comment: “We need fewer laws that lock people up and shorter sentences for some crimes…”

            2. While I agree about the worthiness of the law, prison is not the only option for transgressions of that law.

          2. Except for the very most petty theft, locking people up for theft actually does tend to be cost effective, because thieves typically get so little return on their thefts of articles, (As opposed to money.) that they typically must steal a hundred $K a year just to reach the poverty line. Jailing burglars in particular is wildly cost effective.

      3. Sarcastro (doubting that deterrence, retribution and incapacitation can be the real reasons for imprisoning people)

        You don’t think jails are punitive?

        = retribution

        Or that it’s been a very common policy argument over the past three decades that their conditions must be so, and sentences must be long to discourage criminality?

        = deterrence

        1. Yes, I took criminal law as well. But politics is not sociology.

    3. You’re both wrong. The right approach is focusing on reducing the number of people who haven’t harmed others who become victims, either of crimes that should be crimes, or of imprisonment for crimes that should not be crimes.

    4. This series reads kind of strange, because it seems to treat the actual purposes of incarceration, retribution and incapacitation, as mere obstacles to the obvious proper goal of incarceration policy, minimizing the number of people incarcerated.

      Yes, it reads like a kind of learned, polite, cognitive dissonance. Prof Leipold is intellectually aware of, and honest enough to report, the reasons for why making the reduction of incarceration the central function of the criminal justice system might be a little strange.

      But having mentioned them, he then moves right along, back to his prime objective as if the previous objection had disappeared into the mist.

  3. We need fewer laws that lock people up and shorter sentences for some crimes, but violent crimes don’t belong in either category. The last thing we need is to adopt some Euro standard of a light wrist slap and back on the street for violent criminals.

    1. What do you think would be an appropriate sentence for a violent crime like armed robbery? The thing is, that long sentences aren’t deterrent, nor do they provide any rehabilitation (in fact the opposite). What they do, though, is provide societal retribution and put the offender in a place where they can’t attack anyone but other criminals who are fellow inmates.

      I won’t argue that those latter two are not both acceptable outcomes though, but most people age out of crime after 30 years on this rock, regardless of being in prison or not.

      1. Honestly I’d just as soon have them executed but that isn’t a very popular sentiment these days. If we are just talking about sentence length then twenty years or so is probably adequate. My view is that adults committing violent crime deserve no sympathy.

        1. What types of violent crimes? As the post notes correctly, there is a lot of gray area as to what is, and isn’t, a “violent crime.”

          1. Well you used armed robbery which is what I was responding to. Assault, for instance, encompasses a wide range some of which is not all that serious. Likewise with manslaughter so yes I’m referring to the blacker areas of the gray to black spectrum.

            1. How about armed robbery with a fake gun? Or a calling in a bomb threat but leaving an empty box? Or he said/she said date rape where both parties are intoxicated? See what I’m getting at. So if we are going to use the death penalty, it’s far easier to draw lines for homicide, like we do now.

              1. I see what you’re getting at but I never said that you don’t make distinctions among crimes. For the armed robbery that I described in the anecdote I would cheerfully have pulled the trigger on those perps. As it was we did convict them and it being Georgia they got a nice long sentence in a not nice place. Best we could do.

                1. You suggested in your original reply that a death sentence was an appropriate response for a violent crime because the criminals deserve no sympathy, to which I asked you to get specific, and the response has been to cite an anecdote from anther comment thread?

                  1. How specific do you want me to be? I said that there are distinctions among crimes, that some crimes considered to be violent are not very serious, that there is a spectrum from gray to black, etc. You asked what an appropriate prison sentence was for armed robbery and I said twenty years. I’m not going to waste time parsing the length of sentence for every hypo you dream up.

                    1. I don’t expect you to, but I was hoping that by going through the process of thinking about it via my questions, that you’d see that your off the cuff answer of the death penalty for all violent crimes was misguided at best and downright stupid at worst.

                    2. Stupid is thinking that’s what I said.

                    3. You said, I quote, “I’d just as soon have them executed but that isn’t a very popular sentiment these days.” My condolences for taking you seriously.

          2. There has been a lot of litigation over that exact point with regard to the Federal ACCA. A string of 3 “violent felonies” are a requirement for increased sentences under that act. What is and isn’t violence is a matter of some confusion. I think I recall a recent case where actual or attempted Second Degree Murder was alleged to be non-violent.

            1. Not just alleged: the Ninth Circuit found that second degree murder (in a case where the murderer shot the victim in the head) was not a crime of violence United States v. Begay, 934 F.3d 1033 (9th Cir. 2019).

        2. I’d just as soon have them executed but that isn’t a very popular sentiment these days

          Yes, iIwas thinking that too – purely from the perspective of Prof Leipold’s objective of reducing the prison population. There’s more than one way to leave a prison. Also what about transportation ? I’m sure we could do with a poor country like, say, Angola, to take a few hundred thousand of our prisoners. Say a $10,000 a year fee per prisoner. Good business on both sides.

      2. “but most people age out of crime after 30 years”

        It’s not always easy to take advantage of that, though. If you keep sending a persistent offender to jail until he’s old enough to stop offending (50??) he is unlikely to be able to support himself from age 50 on. We’re going to be supporting him one way or another – minimum security prison, halfway house, or various kinds of welfare, take your pick.

        Note that this is a strictly economic argument. There are moral arguments as well. Those get pretty complicated. There is some appeal to the argument that, say, a mugger who isn’t going to mug anyone again deserves to walk free again. OTOH, when you consider what to do with a murderer who won’t murder again, it’s hard not to remember that the murder victim will never have that privilege.

        1. Well it’s hard for me to forget the mugger’s victim. Mugging someone is not a petty crime; it is violent and quite often life changing for the victim.

        2. Just an anecdote: I served on a jury for an armed robbery charge. The trial was at least two maybe three years after the robbery. When the store clerk testified he couldn’t help shaking violently and several times broke down completely.

          1. No disagreement here. Heck, we were burgled decades ago, and you lose a bit of innocence forever. We never leave the door unlocked, deal with the alarm, etc.

            But I’d rather be burgled than mugged, and rather be mugged than murdered. To whatever degree you recover from a mugging, you recover less from a murder.

          2. I forgot to mention that they had forced him into a back room where they undoubtedly intended to murder him. Fortunately the woman driving the car got spooked and honked a pre-arranged signal and they ran out.

        3. In my view, a community’s need to keep a violent criminal incapacitated may very well require a longer incarceration than is really needed as “society’s revenge.”

          In those cases, after the appropriate “revenge” time has passed, I’d offer the convict early release in exchange for incapacitation. So for instance a convicted murderer might be allowed release after 10 years in exchange for having both legs cut off above the knees (thus making it much easier for potential victims to get away from him).

  4. The OP says: “”As the chart shows, African Americans and Hispanics are more likely to be imprisoned for violent crimes than for property crimes—in contrast to Whites, where the opposite is true.”

    Maybe because Blacks and Hispanics are more likely to commit violent crimes than Whites? Drawing that inference from the chart is more reasonable than what you suggest, which is that Whites aren’t punished as often for violent crimes.

    1. Dr. Martin Luther King Jr. told a congregation in 1961:

      “Do you know that Negroes are 10 percent of the population of St. Louis and are responsible for 58% of its crimes? We’ve got to face that. And we’ve got to do something about our moral standards … We know that there are many things wrong in the white world, but there are many things wrong in the black world, too. We can’t keep on blaming the white man. There are things we must do for ourselves.”

  5. It is very strange that for-profit private prisons and cheap prison labor aren’t part of this discussion, yet. The system is at least as responsive to incentives as individuals; if we want less of something the first step is to stop subsidizing it.

    1. It’s not surprising at all, because the for-profit prison system is created by government looking to lock people up cheaper than they can do it themselves. It is, in short, the private sector responding to government incentives rather than an excess capacity of prisons in the private sector driving incarceration.

      Likewise, cheap prison labor, called the Convict Labor System, while at one time an important part of the South’s Jim Crow system, doesn’t drive incarceration rates because even with virtually unpaid labor, it’s still more way expensive to have a convict make XYZ than paying for a private sector company to do it.

    2. On top of mad_kalak’s comment, privately operated prisons are a small fraction of the incarceration issue. the percentage of the total US Federal and State prisoners housed in privately operated prisons is under 10%.

      To the extent that mass incarceration is driven by lobbying from interested parties, the public sector prison guard unions are likely a much bigger factor.

    3. Unfortunately, the prison guards’ unions, at least in Texas and California, have tons of money to spend at every election to ensure that the governors and legislatures there won’t agree to any reform that would cost a lot of them their jobs. That’s a major part of the problem getting prison reform passed.

  6. Prof Leipold’s recidivism statistics comparing recidivism for violent criminals with non violent criminals overlook three fairly fundamental points :

    (1) expected value – a 2% chance of a murderer reoffending and a 2% chance of a burglar reoffending represent the same probability of reoffending, but do not represent anything like the same expected value of reoffending (in terms of harm to the victim.)

    (2) since on average violent criminals get longer sentences than non violent criminals, on average violent criminals will be older when released. Since age is inversely correlated with reoffending that will tip the statistics in favor of the violent criminals. Consequently you need to correct for that, by measuring reoffending rates for violent and non violent criminals released at the same age.

    (3) and last but not least, the proper comparison to measure the value of the incapacitation effect is to compared the rates of reoffending by released convicts with the rates of offending by unconvicted persons of the same age.

    1. If he’s referring to the data set I think he is, then #2 is already included. I agree that #1 is an issue, but that’s a political judgment, there’s no statistical method that sorts that one out for you (reasonably, anyways, unless you want to enter utilitarian territory).

  7. Of course, if we adopted a sane drug policy, the impact on property and violent crime would be profound, as no small percentage of those are related to prohibition rather than drugs themselves.

    Is that really any less politically impossible than substantially reducing punishment on violent crimes directly?

    1. This is the received opinion, but I have my doubts.

      A “drug criminal” who steals to finance his drug habit will very probably steal less if drugs become cheaper, from decriminalisation. But are such drug criminals – ie users – the great part of those serving time for drugs offences ?

      But a “drug criminal” who is trying to make money from drugs is a businessman trying to do business. If the profits of his business – which depend on its illegality, otherwise he’d be put out of business by Amazon, Target and Walmart – are ripped away by decriminalisation, then he has to earn a living in some other way.

      Will it be by taking a job at the Walmart checkout, or will it be by mugging or food stamps fraud ?

      Although I have nothing against a sensible drugs policy, I’d say hoping for a big crime reduction dividend may be wishful thinking. We’d be more likely to get a crime displacement dividend, and how much are they worth ?

      1. Given that a lot of the current cannibusiness is run by people who seem to have a long prior expertise in its cultivation and sale, I’d say that it’s not unreasonable to think there will be substantial crime reductions. However, if the federal government finally removes some of the financial and legal barriers to engaging in the trade, I’m not entirely certain that the current small-time operators will be able to outcompete big business. So to your point, I think we’ll have to see how it shakes out.

Please to post comments