Is Mass Incarceration Inevitable?  Part 2. Much Smaller is Still Very Large

Measuring the size of the problem and the solution.

|The Volokh Conspiracy |

[The discussion in this post is excerpted from Andrew Leipold, Is Mass Incarceration Inevitable? 56 American Criminal Law Review 1579 (2019).]

In a prior post I suggested that there were significant barriers to reducing the size of America's prison and jail populations, a necessary step if we are to overcome what is now routinely called "mass incarceration."  This is the first  of three posts dedicated to addressing those barriers.

But first, it is worth taking a minute to discuss what we mean by mass incarceration.  We know that the label at least means that we have "a lot," and "too many" prisoners, but after that it gets murky.  Given the amount of crime in the US—over 1.2 million violent crimes and almost 7.7 million property offenses in 2017—it is hard to say how much is too much.  As John Pfaff neatly put it, "'mass incarceration' essentially boil down to claims that we have too many people in prison, although we don't really know how many too many; and that we should reduce that number, although we don't really know what the new goal should be."

Nonetheless, we have a couple of measures that might lead us to think that the current rate is too high.  The first is that in the last half century we have been incarcerating at a dramatically higher rate than we did for the half century before that.  In 1925 the prison incarceration rate was about 100 inmates per 100,000 people in this country, a figure that remained steady for about 50 years.  But from the mid-1970s until about 2010, the rate rose about 500%, with the jail population rate (which wasn't reliably measured until the mid-1980s) following a similar path.

U.S. Incarceration Rates, 1925-2016

There actually may be less to these trends than meets the eye.  As Bernard Harcourt has explained, although the prison and jail rates have soared over the last 50 years, the overall institutionalization rate has not.  In the middle part of the century (when the prison rate was low), there were a large number of people who were involuntarily committed to psychiatric institutions—550,000 in 1950, or roughly one-quarter of today's prison and jail population.  By the 1990s, however, when prison populations were climbing, deinstitutionalization of those with mental illness had reduced the numbers of involuntary psychiatric patients to a mere 30,000.

U.S. Institutionalization Rates, 1934-2000.

(Many thanks to Professor Harcourt for giving me permission to reproduce this chart.)

Whether and how this information should change our thinking about mass imprisonment is an intriguing question.  But ultimately, even if the growth in prison and jail inmates is not as straightforward as the numbers suggest, the core point remains, and is perhaps strengthened.  Prison and jails are designed to punish and detain, and their rapid growth is cause for concern, even if it is explained in part by a shift from treating large number of people as criminals rather than patients.  More pointedly, if addressing the prison size requires targeting those who are behind bars unnecessarily, then inmates with (potentially) treatable problems might be an excellent place to start.

There is a second metric that is at least suggestive of a problem:  comparing the U.S. incarceration rate to those of other countries.  The U.S. incarcerates an estimated 655 inmates per 100,000 people (although this is based on 2016 numbers, so the current number will be smaller).  And as we look down the list at the rest of the top five, we do not see countries that we would normally compare ourselves to.  The country that incarcerates at the second highest rate is El Salvador (604 inmates/100k people), followed by Turkmenistan (552), the U.S. Virgin Islands (542), and Thailand (524).

Outside of the U.S. and Thailand, the countries with the highest incarceration rates in the world are quite small.  So perhaps a more useful comparison is between the U.S. and other countries with large economies.  Some are democracies, some are not, but using economic size to define a peer group ensures that we are looking at counties with roughly comparable resources to address social problems.  Making this comparison shows that the U.S. now widens its lead in the incarceration race, and by a hefty amount.

Incarceration Rates for Countries with the Largest Economies.

GDP RankCountryIncarceration Rate
1United States655
5United Kingdom140
Non-U.S. Average144
*There are reasons to be skeptical that the numbers reported by China represent a true apples-to-apples comparison


Even in the absence of a precise definition of "mass incarceration," these numbers strongly suggest that there is something unusual and problematic about our approach to criminal punishment.

Putting ask the political challenges and resource constraints that will hamper the efforts to reduce the incarceration rate (more on these in the next two posts), the sheer size of the task makes it almost inconceivable that the U.S. will either return to rate more in line with our earlier history, or will rejoin the pack of other large-economy countries.

Consider the following.  For the U.S. to reduce its incarceration rate to the average rate of the other top-10 economies (144 per 100,000 people, as shown in the table), it would have to cut its rate by a staggering 75%.  Stated differently, the U.S. could empty every state and federal prison in the country, leaving only inmates in the local jails, and still be above the world average.  (There were roughly 745,000 people in U.S. jails in 2017, for a rate of 229 inmates per 100,000 people—higher than any other top-10 economy other than Russia and Brazil.)

Alternatively, we might set a goal of cutting our incarceration rate to a point that is more in line with our past.  Returning to the rate that existed in the middle part of the last Century, 100 inmates per 100,000 people, seems fanciful.  (Or as Jeff Bellin has put it, "we are closer to landing an astronaut on Pluto than we are returning to 1970s incarceration levels.")  Even a return to 1990s levels, which would require a 20% to 25% reduction in current incarceration rates, would be difficult.  And even if we could achieve this more modest goal, the U.S. would still have the seventh highest incarceration rate in the world (overall) and still have the highest rate among the countries with the largest economies.

In short, if "mass" incarceration is defined in relative terms—how we compare to the rest of the world—or in reference to historical incarceration rates in this country, no realistic amount of sentencing and criminal law reform is going to change the distinctive space the U.S. now occupies.


NEXT: Today in Supreme Court History: October 8, 1888

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. Federal and state prisons are very different. The federal prison population is too large; state prison populations are often roughly the right size. State prisons hold too many drug offenders, and not enough murderers. Federal criminal sentences are too long, and people are there based on statutes that often exceed Congress’s constitutional power under the interstate-commerce clause, meaning that their conduct should only have been regulated by state law, not federal law. State sentences for violent crime are often too short — people are usually surprised when they learn how short sentences for violent crime usually are, as John Pfaff noted to the Cato Institute — and people who commit homicide receive far shorter sentences than people think. Even though there are studies that indicate that harsh penalties deter violent crime more effectively than lenient penalties. And longer sentences incapacitate offenders, especially repeat offenders and younger offenders.

  2. Most prisoners are in state prisons, where the overall size of the prison population is about right, but the composition is not right (state prisons hold too few violent offenders, but too many drug offenders, and the one roughly cancels out the other). State prison populations should not be substantially reduced. By contrast, some reduction of the federal prison population through the First Step Act was appropriate.

  3. We look somewhat better if you normalize those international statistics by demographics. Not great, though, just better.

    But my chief complaint here is treating a heterogeneous country of over 300 million as an undifferentiated mass. Large expanses of America have levels of criminality, and incarceration, no worse than Germany or France. Just looking at states, the rate varies nearly four fold, and the variations within states are even greater.

    You really can’t learn anything if you throw all your data in a blender, and hit “puree”.

    1. Ok, but then you can’t treat Germany and France as undifferentiated masses either. They, like the US, have large urban centers and rural areas.

      1. That’s perfectly true!

        What I’m saying here is that, when you’re analyzing crime data, you really need to avoid aggregating data across large, heterogeneous areas. It hides everything that is driving the variation.

        1. Brett, have you ever considered that crime might occur in some proportion to opportunities for crime? And that those opportunities might vary with the frequency with which people interact? And that a given populated space would increase geometrically the number of human interactions available, as density goes up?

          I mention that, because I suppose factors of that sort might be useful to you if you ever wanted to critique the evident anti-urban bias you bring to crime discussions. It could be, if you looked at it, that Chicago or Baltimore, for instance, might generate less crime per-interaction-opportunity than would some typical rural backwater.

          Maybe it would turn out, by a measure of that sort (crimes per human interaction), that our major urban areas are also our lowest-frequency crime venues, and that the rural areas come in higher. What do you think?

          1. What, you’re proposing a model of crime where criminals lack agency, they just sort of randomly wander around, and commit crimes only when they bump into a potential victim? Like, you only burglarize a house if you find yourself at the doorstep at night, rather than having traveled TO that doorstep with the aim of burglarizing it? Rural folk don’t commit murders because they’re camped out in the middle of corn fields, nobody nearby to murder?

            Well, why don’t you suggest a way to test that hypothesis? You’ll still need the data, and not aggregated.

            1. Don’t like my hypothesis, do you?

              1. I think it’s kind of silly in regards to most crimes, but probably has some validity in regards to a few specific crimes. Certainly looks good for traffic offenses.

                But, suppose that your theory proves out: Murder is a direct consequence of population density. What does this suggest as a policy response? Evacuating cities?

                1. Brett, you have not yet grasped the implications. It could be that murder totals in cities are higher than elsewhere with regard to absolute numbers, but lower in terms of the quantum of human interaction the city provides. If that is so, the higher murder numbers in the city could be a result of more interactions, instead of a result of a higher rate. And that could mean the city rate could even be a bargain, compared to others places elsewhere, featuring lower absolute numbers, but more murders per-interaction.

                  Not saying any of this is actually happening. Just speculating.

                  1. Right, just speculating.

                    And my point is that if you want to know if this is true, (Rather than just speculating.) you need data to test your hypothesis against, and that means “you really need to avoid aggregating data across large, heterogeneous areas.”

                    Nation-wide numbers are basically worthless for any purpose except public relations. They don’t tell you enough to figure out what is driving things, and what policies you might adopt if you DO decide all those gang murders on the South side of Chicago aren’t just a natural consequence of the vibrant city life a bit further North where they aren’t happening. And that you could have the vibrant city life without piling up bodies in windrows.

          2. have you ever considered that crime might occur in some proportion to opportunities for crime?

            Sure, but this is not just about interactions. Auto theft varies according to how easy it is to steal one. Which obviously has a lot to do with technology.

            Also theft will vary acording to the market for the stolen items. It’s really easy to steal garden gnomes, but how much are you going to get from your fence ?

            And that those opportunities might vary with the frequency with which people interact? And that a given populated space would increase geometrically the number of human interactions available, as density goes up?

            The city obviously increases the prospects of valuable wanted interactions – which is why we have cities – but also unwanted interactions, all the way from noisy neighbors and crowded sidewalks, to gangs of muggers.

            But it’s not just a function of the number of interactions. The closer packed rats are, the more angry and aggressive they get. So pack too many in a confined space and not only will you get more interactions but the quality of the interactions will become increasingly aggressive.

            This obviously applies to humans too – as a rush hour commute will confirm.

            1. Rush hour commuting is not a hot bead of crime, I don’t think your pop-sociology is really getting to the bottom of things.

          3. I disagree with the premise that crime goes up due to interaction with other people. There is no evidence anywhere that suggests a law abiding citizen will commit crimes when he is a densely populated environment as opposed to a rural environment. Density doesn’t dictate ethics or morality.
            Poor economic environment is a better indicator of crime.

    2. I’ll agree with Brett on this one.

  4. The institutionalization rate graph is certainly interesting, and it would be interesting to see a crime rate graph superimposed on it.
    Or perhaps a few different crime rate graphs – say homicide, violent crime, all non drug crime, and all crime.

    That dip in the middle from say 1965 to 1995. Hmm.

    Anyway. One other way at looking at the incarceration rate or institutionalization rate is to consider how often you come across someone who you think ought to be in prison, but isn’t. If this happens too frequently, maybe the rates are too low.

    Of course it depends a bit on where you walk around. My feeling is that roughly 2% of the population are lunatics, but only about half of those are dangerous lunatics. So about 1% of the population looks about right. Which is about where we are.

    And this comforts with the international “data.” London is certainly better stocked with weirdos-on-the-loose than New York. It was not always so.

    And their incarceration rate is much lower.

    1. I was surprised at the rate of mental institutionalization too. I knew there was some drop in mental patients around the time the crime rate started rising (California governor Ronald Reagan was famous for it), bu had no idea of the magnitude.

      It would be interesting to see crime rates on that graph. Even more interesting would be to extend the comparison with other countries to include their mental patients. It’s a shame the combined incarceration rate is brought up and then dropped like a hot potato.

  5. In short, if “mass” incarceration is defined in relative terms—how we compare to the rest of the world—or in reference to historical incarceration rates in this country, no realistic amount of sentencing and criminal law reform is going to change the distinctive space the U.S. now occupies.

    The argument thus far does not adequately support that conclusion. It is good to consider reversing the effect of putting mental health cases into the prisons. Correcting that would make a notable difference by itself.

    Another hugely influential factor is length of sentences. One useful measure of incarceration could be the number of persons incarcerated. Another measure could be person-years of incarceration. On that second basis, it is easy to understand that the longest sentences are disproportionately influential in determining the overall incarceration rate. Conversely, if those sentences were found to be uselessly long, reducing them would be disproportionately influential in reducing the rate.

    Other things, many of them having to do with race, would also make a difference. Consider:

    1. Who supposes, controlling for everything, that black communities are not more heavily policed than white communities? By whatever percentage that difference in law enforcement attention increases incarceration, reform of the (controlled for everything) disparity will reduce needless incarceration of blacks.

    2. Who supposes that blacks are not more often convicted than whites, given similar evidence in similar cases? By whatever percentage that difference in conviction rates can be corrected, it would reduce needless incarceration of blacks.

    3. Who supposes that blacks do not receive longer sentences than whites do for similar crimes? Once again, an opportunity to reduce needless incarceration, and this time in an especially sensitive statistical area.

    4. Likewise with regard to both the conviction and sentencing questions, who supposes blacks fare as well as whites in plea bargain situations, either with regard to opportunities to plead to lesser crimes, or with regard to length of sentencing?

    Note also, the relationship between those various factors is not additive, it is compounded. Does disparate police surveillance make a black person twice as likely to be incarcerated for a crime as an equally guilty (or innocent) white person? Is a black person typically sentenced to twice the time a white person would be sentenced to for the same crime? Then the effect of correcting those two factors together would reduce black incarceration to 25% of its existing level, as measured by person-years of incarceration.

    Those numbers were just chosen for convenience, to illustrate the principle. I have no idea what the correct figures should be. The point is, that to reach the conclusion quoted above, someone would have to measure those kinds of numbers with high accuracy. Has that been done? Can it be done?

    Some will notice that I have used as a working premise that the incarceration rate for whites is a baseline for the correct level of incarceration. That is arbitrary, of course, and could be discussed separately, and no doubt should be.

    1. It is arbitrary. Indeed, many criminal laws are effectively less enforced in the black community, depressing the black incarceration rate. A much smaller percentage of violent crimes against blacks are solved: Only 30 percent of all murders involving black victims are cleared, whereas most murders involving white victims are cleared. Because most crimes against blacks are by other blacks (PolitiFact noted that between 2010 and 2013, “92 percent of blacks who were murdered were killed by other blacks”), this underenforcement actually reduces the percentage of incarcerated people who are black. So the existing higher incarceration rate of blacks is actually lower than it would be if society actually effectively protected black people against violent crime. If society actually protected black victims adequately, the black incarceration rate would be higher still than it is now.

      Many studies find that blacks don’t receive longer sentences for the same crime. (See, e.g., Stephen P. Klein, et al., “Race and Imprisonment Decisions in California,” 247 Science 812 (1990)).
      Other studies find some disparity, although none claim that sentencing disparity is the main reason blacks are a disproportionate share of the prison population.

      A 1991 RAND Corporation study of adult robbery and burglary defendants in 14 large U.S. cities found that a defendant’s race or ethnic group bore almost no relation to conviction rates, sentencing severity, or other key measures. In 1994, federal government statistician Patrick A. Langan analyzed data on 42,500 defendants in the nation’s 75 largest counties and found “no evidence that, in the places where blacks in the United States have most of their contacts with the justice system, that system treats them more harshly than whites.” As he noted, “No Racism in the Justice System,” “Many studies have been conducted that show no bias in the arrest, prosecution, adjudication, and sentencing of blacks,” while “other studies show possible evidence of bias.”

      Similarly, the fact that African-Americans are convicted and incarcerated at a higher rate than whites is primarily due to the higher black crime rate, and victim reporting of such crimes when they occur, not police racism. Crime is heavily black-on-black, and black victims of violence crimes disproportionately identify their assailant as black. As the Bureau of Justice Statistics explains, most crimes are committed mostly between members of the same racial group, and this is true for “rape or sexual assault,” “simple assault,” “aggravated assault,” and indeed, “all types of violent crime except robbery,” which is disproportionately committed by blacks against non-blacks. (See Race and Hispanic Origin of Victims and Offenders, 2012-2015).

      1. When I write above, that “many criminal laws are effectively less enforced in the black community,” I am not attributing this to current racism by the police.

        1. Hans, one critique you seem to ignore is that the data you reference to support your conclusions is self-evidently based on the work product of the law enforcement community—which, of course, has been under criticism for long-standing incuriosity about its own performance with regard to race. Not to mention, sometimes cooking the books with regard to race. (Of course I am aware that there are also highly conscientious efforts being made by law enforcement in many venues.)

          From time to time, after extraordinary incidents draw unusual intention from investigators outside law enforcement, investigations surface to show that criticism of racial bias in the system has been richly deserved. So my suggestion is that the reality is unlikely accurately described by a list of citations which suggests it all comes out one way, and everything is fine.

          By the way, you didn’t mention plea bargains. Do you have data to show that black defendants get the same benefit (or suffer no more penalties) from plea bargains than do white defendants? How would that be studied and measured? Given the preponderance of criminal cases settled by plea bargains, it seems like something that ought to be understood accurately. And not‚ by the way, just in the, “75 largest counties.” Might that be a fruitful area for more investigation?

          1. You seem to be ignoring victimization surveys. Unless blacks are deliberately lying about being victimized by blacks, there really isn’t any question that the crime rate in (certain!) black communities is terrifyingly high, and committed by blacks.

            1. Both things can be true, that blacks commit crimes at a higher rate than whites, and that they are punished overly harshly for said crimes compared to their white counterparts.

              In fact, there’s pretty good statistical evidence that this is the case.

              1. Yes, they could both be true, but Hans related evidence that casts doubt on the latter claim.

                1. Other studies find some disparity, although none claim that sentencing disparity is the main reason blacks are a disproportionate share of the prison population.

                  So…not really.

          2. Does anyone have data on plea bargains? This article, at least the abstract, says for misdemeanors there is evidence of racial bias in plea bargaining, but that goes away where “severe felonies” are concerned. It says white plea bargainers with no criminal history do better than black plea bargainers with no criminal history, but white and black plea bargainers with criminal histories do about the same.

            A quick google search reveals dozens of studies on this precise subject.

    2. “By whatever percentage that difference in law enforcement attention increases incarceration, reform of the (controlled for everything) disparity will reduce needless incarceration of blacks.”

      I lived in a high crime, majority-minority neighborhood for more than half my life. And my law abiding minority neighbors sure weren’t looking to have the police presence cut back. The notion that we ought to ration police services to minority communities seems like something out of Jim Crow. Every citizen deserves to be equally defended against predatory criminals.

      1. Not my notion, Absaroka. It seems to me that your comments in reply often rely on cartoon images instead of on careful evaluation of what I say. It gets tiring dealing with the cartoon characters, so I wish you would go back by yourself and try not to read into what I write the views you would most like to criticize.

        Here is a hint in this case. It is perfectly possible that law enforcement in a black neighborhood could be at once notably higher than in a similar white one, and also insufficient in the judgment of the law abiding folks who live there.

        1. Your posts are generally not models of clarity. They tend to be long on vague generalities, leaving your readers unsure what on earth you are actually proposing in any practical sense.

          Here, for example, you want to ‘reform the disparity’ of policing between high and low crime areas. That seems to imply either decreasing the policing in high crime (and generally poor) areas or increasing policing in low crime (generally affluent or rural) areas. Both of those seem like rather odd responses. Do you have some other way of accomplishing your stated objective?

          FWIW, in the jurisdictions when I have lived the police are quite transparent about how they allocate resources, because in general everyone wants more of them[1], and they generally allocate those resources to … where the crime is. The homicide squad doesn’t ride around looking for bodies, they go to where people report murders have occurred. The same is true for burglaries, assaults, car thefts, etc. The police give press conferences explicitly explaining that they have had an upsurge in 911 calls from Precinct X, and are adding officers to cope.

          That’s how it’s done now – by all means explain the specifics of how you’d like to change that allocation of resources.

          [1]one glaring exception is the stop-n-frisk era in NYC, which was certainly outrageous; the residents there had every right to be pissed. Not because the police were devoting resources to high crime areas, but because they were doing suspicion less stops.

  6. “Even in the absence of a precise definition of ‘mass incarceration,’ these numbers strongly suggest that there is something unusual and problematic about our approach to criminal punishment.”

    And I strongly suggest you stop making pre-disposed, subjective judgments on the situation.

    Something unusual? OK.
    Problematic? Where does that come from? What standard(s) are you using to show it’s actually a “problem?”

  7. It’s always people who have absolutely no contact with street criminals who want to release them from prison. After all, it won’t change their lives one bit.

  8. What does the total institutionalization rate look like in other countries?

  9. As Stephen Lathrop mentioned earlier sentence length is an important metric in comparing incarceration rates in various countries. This is from an article entitled “Sentencing and Prison Practices in Germany and the Netherlands (2013)”:

    In 2006 in Germany, 75 percent of prison sentences were for 12 months or less and 92 percent of sentences were for two years or less. In addition, Germany suspended the vast majority of prison sentences that were under two years—in about 75 percent of cases, so only a very small percentage of those sentenced
    ever went to prison [see Figure 3]. Similarly, in the Netherlands in 2012, the vast majority of sentences (91 percent) were for one year or less, going up to 95 percent if sentences of two years or less are included.

    It is difficult to find a side by side comparison of sentence lengths by type of crime but it does seem that in Western Europe sentences for murder, rape, assault, etc. are much lower than in the US. Is this a good thing? I don’t think so. Letting someone out in five or six years for first degree murder seems very wrong to me.

    For other types of crimes, and in particular drug crimes, I am completely in favor of complete decriminalization or at least greatly reducing sentencing.

    1. Legalization, not decriminalization.

      Decriminalization is a halfway step to legalization, which gives you the worst of both worlds: Lots of the formerly fully illegal conduct, but keeping the black market it leads to.

      You need the drug business legal enough that the guy selling the drugs can call the police if somebody robs him, instead of hiring the local gang for security. That’s the only way reforming drug laws reduces REAL crime.

      1. wrt drugs I agree that legalization is the ideal but I don’t agree that decriminalization is the worst of both worlds. It reduces incarcerations and while it doesn’t solve the supply side problem it is easier to sell to the population. Marijuana is one thing but try pushing for legalization of heroin.

        1. I’m less concerned about the welfare of people stupid enough to use drugs, than I am the welfare of sensible people who get caught in the fallout of the war on drugs. Decriminalization may benefit the former, but it is of little benefit to the latter; They get the disadvantages of rampant drug use AND flush black markets.

          1. Once a libertarian always a libertarian I suppose.

        2. “I’m less concerned about the welfare of people stupid enough to use drugs…” In other words until you get your desired outcome, legalization, then you are fine with stupid drug users being put in jail for using drugs. No problem, they’re stupid. It doesn’t seem to occur to you that failing to jail these stupid people is not retarding the march that you apparently see as right-around-the corner legalization.

          I used to think that way when I was about eighteen. I grew out of it by the time I was about twenty.

  10. Even in the absence of a precise definition of “mass incarceration,” these numbers strongly suggest that there is something unusual and problematic about our approach to criminal punishment.

    The fundamental problem with this analysis is that it does not take into account the difference in crime rates between locales. If two locales have the exact same approach to criminal punishment, but one has five times the criminal activity as the other, you would expect to see five times as many people incarcerated in that locale.

    Now, of course, cross-country comparisons of crime rate have many difficulties, like differences in laws and reporting rates. But the least problematic is probably the intentional homicide rate, since it is considered a crime everywhere and is relatively (though certainly not perfectly) resistant to reporting issues.

    So, let us check the reported intentional homicide rates of our eleven largest economies.

    United States — 5.30
    China — 0.60
    Japan — 0.20
    Germany — 1.00
    United Kingdom — 1.20
    India — 3.22
    France — 1.30
    Brazil — 30.50
    Italy — 0.67
    Canada — 1.80
    Russia — 9.20

    Then let’s calculate the incarcerations-per-intentional-homicide rate:

    United States — 655 ÷ 5.30 = 123.6
    China — 118 ÷ 0.60 = 196.7
    Japan — 41 ÷ 0.20 = 205.0
    Germany — 77 ÷ 1.00 = 77.0
    United Kingdom — 140 ÷ 1.20 = 116.7
    India — 33 ÷ 3.22 = 10.2
    France — 106 ÷ 1.30 = 81.5
    Brazil — 348 ÷ 30.50 = 11.4
    Italy — 101 ÷ 0.67 = 150.7
    Canada — 107 ÷ 1.80 = 59.4
    Russia — 373 ÷ 9.20 = 40.5

    Presented this way, the US incarceration rate doesn’t look unusual at all. In the Group of Seven large industrialized countries, the incarceration rate ranges from 59.4 to 205.0 prisoners per intentional homicide, with the US ranking third of the seven in incarcerations-per-homicide, its 123.6 prisoners per homicide victim only slightly above the mean (116.3) and median (116.7) number of prisoners-per-homicide in the G7.

    1. “Intentional homicide is defined by the United Nations Office on Drugs and Crime (UNODC) in its Global Study on Homicide report: Within the broad range of violent deaths, the core element of intentional homicide is the complete liability of the direct perpetrator, which thus excludes killings directly related to war or conflicts, self-inflicted death (suicide), killings due to legal interventions or justifiable killings (such as self-defence), and those deaths caused when the perpetrator was reckless or negligent but did not intend to take a human life (non-intentional homicide).”

      1. Yeah.

        Obviously, country-to-country differences on the elements of the definition can hurt the comparability of those numbers. But other crimes have all the definition problems to a greater extent; even when the nominal definitions are identical, the reporting rates for, say, rapes, are going to vary on cultural factors.

        By the homicide numbers, America is clearly a substantially more violently criminal society per capita than the other G7 nations, which pretty much fits everybody’s qualitative impression. Accordingly, we should entirely expect a higher per capita rate of incarceration in the US than in other G7 nations.

        Given the differences in homicide rates and incarceration rates are remarkably similar, it suggests that US incarceration policy is not particularly unusual or problematic. Rather, American violent crime rates are unusual and problematic. Crime reduction, not penal reform, would be the necessary path to reducing incarceration rates.

  11. Seems to me that while the “per 100,000” is a useful metric, there’s a very different dynamic in managing a large culturally heterogeneous population than a small, largely homogeneous one. Comparing the US to Scandinavia or Japan represents is still not apples-apples.

Please to post comments