Digital

Exclusive online content

Crime and Punishment

By: Joseph M. Bessette
February 7, 2018

n the Summer 2017 CRB, Joseph M. Bessette reviewed Locked In: The True Causes of Mass Incarceration and How to Achieve Real Reform, by John Pfaff. We're glad they have agreed to peruse questions about the effectiveness of U.S. incarceration policy in reducing crime. John Pfaff is a Professor of Law at Fordham Law School. Joseph Bessette, a long-time CRB contributor, is the Alice Tweed Tuohy Professor of Government and Ethics at Claremont McKenna College.

* * *   

John: First, thank you to Joseph Bessette for his thoughtful comments on Locked In. Obviously, like any author, I hope a review will start with “never has there been a book better than this” and only improve from there. But critical reviews such as this one always help refine and improve my ideas—which is, I suppose, a nobler goal than stoking my own ego. And thank you to the Claremont Review for providing this forum to engage with each other. With that, I want to jump right in.

Broadly speaking, Bessette’s criticisms of my book fall into two general categories, one practical and one philosophical. The practical complaint is that we need prisons to fight crime, and violent crime in particular: as he puts it, we have a “crime problem” more than a “prison problem,” and thus discussions like mine about how to cut prison populations are misguided. The philosophical objection is that by glossing over retributivist justifications for prison, I’m advocating for a criminal justice system devoid of justice: were policymakers to take the path I suggest, he states, “what we used to call `criminal justice’ will have ceased to be about substantive justice at all.”

In this post, I want to address the practical claim, that we need prison to deal with our crime problem. I have much to say about the “substantive justice” point as well—especially the challenging issue of whose definition of substantive justice a fractured and pluralistic society like ours should adopt—but I will save that for a second round of discussions.

Summarizing quickly, Bessette’s pragmatic argument for prisons rests on a few, straight-forward points: (1) we have a lot of crime, violent crime in particular; (2) our clearance rates—the rates at which people are arrested for crimes—are low, requiring us to be harsh to those we do punish; and (3) the recidivism rates of those we release from prison are high, demonstrating the need to keep them locked up for longer periods. Taken together, he sees these points as undermining my claim that we need to lock fewer people in prison and for less time, including those convicted of crimes of violence.

Each of his points is fairly sound empirically, although each is also somewhat more complicated than he suggests when we dig a little deeper into the numbers. These points, however, do not entail the conclusion he draws from them, that we need to rely on prison to (at least roughly) the extent that we do today. 

Let’s start with each stylized empirical claim. Bessette provides some striking numbers on victimization: that in 2015 the US experienced 15,000 homicides, 327,000 robberies, and so on. These are large numbers, but then we have a lot of people. How do our crime rates compare to, say, Western European countries, which have far lower incarceration rates? International comparisons are always fraught with definitional peril, but roughly speaking we don’t look that different.

Crime and Incarceration Rates per 100,000 People (2015)

Country (prison/jail)

Homicide

Assault

Theft

Incarceration

US

4.4

238

1776

666

France

1.6

367

1846

103

England/Wales

1.2

744

2216

146

Denmark

1.0

26

3436

59

Germany

0.8

157*

1647

77

Sweden

0.9

48*

3828

53

Outside of homicide—an “except for” on the order of “other than that, Mrs. Lincoln…,” and one I’ll turn to soon—we’re not really an outlier when it comes to crime. This is all the more clear when we realize that Germany and Sweden both redefined assault in the early 2000s in a way that caused big drops in official statistics (to 1/4 its prior level for Germany, and to 1/20 for Sweden).

Domestic comparisons are equally puzzling. Our crime rate today is about what it was in 1970, but our incarceration rate is about five times higher. Unless you think the average crime-aged American is much more violent today than in the 1970s—and I think it’s pretty clear that the Millenials are distinctly less violent than the similarly-aged Boomers were in the 1970s—then the scale of incarceration is again hard to justify purely on crime-rate terms.

Even our elevated murder rate can’t explain our incarceration rate difference. About 189,000 people were in prison for murder or manslaughter in 2014—about 14% of all state inmates. These inmates also make up nearly 2/3 of all people serving more than 10 years in prison. This is not a trivially-sized population. But, at the same time, that means that our combined prison and jail incarceration rate for non-homicide offenses is still in the 500s, still way above anything comparable to Europe.

So a simple mechanistic more-crime-leads-to-more-prisoners relationship can’t explain our differences with Europe. That said, our elevated murder rate might play a role here, not in terms of the number of people in prison for murder, but what it does to the politics of crime more generally. As the criminologist Franklin Zimring has argued, perhaps we are punitive in part because we always fear that any crime could result in death. The chances of that are actually fairly low, but much higher than in Europe, and our sensationalist media, news and fiction shows alike, only heightens people’s fear.

Less a crime problem, then, than a fear-of-crime problem, where the fear isn’t entirely tethered to real, underlying risk. (Tellingly, Americans consistently think that crime has risen ever year since 1991, despite steady and dramatic declines since then. The public is truly poorly informed about the real risk of crime in the US.)

When it comes to clearance rates, Bessette’s argument is on its strongest empirical footing. Our clearance rates are low, often appallingly so. Only about 60% of all murders result in an arrest, and in her book Ghettoside Jill Leovy shows that while the clearance rates for murders in LA County is about 60% overall, when the victim is a black male it drops to about 30%. For nearly 2/3 of all black men murdered in Los Angeles, the LAPD makes no arrest. That’s unacceptable.

Yet this point in no way justifies the use of incarceration. If we underpolice—or, more accurately, mis-police, focusing too much on lower-level crimes while major offenses like homicide and rape go under-enforced—then the solution should be to improve policing, not to rely more on prisons. The evidence is clear that if anything deters, it is the risk of detection and arrest, not the threat of long, far-off punishments that are imposed only on the off-chance that the offender gets arrested in the first place.

This shouldn’t surprise us. There’s long been evidence that those who tend to engage in criminal and anti-social behavior are very present-minded—partly because they tend to be young, but this seems to be true even when controlling for age. The immediacy of the officer down the block arresting them is always going to be a more forceful deterrent than some sort of delayed punishment. And it certainly means that making sentences longer than they already are will barely budge whatever deterrent effect the original sanctions had.

In the end, policing is clearly much more efficient than incarceration. Steve Levitt, for one, wrote that the evidence suggests a dollar spent on policing goes about 20% further than a dollar spent on incarceration. There are a host of street-level interventions that show promise in stemming violent crime, from Ceasefire to Cure Violence to cognitive therapy; we should fund these far more than we fund prison. (Don’t, however, overstate the ease of doing this: criminal justice is plagued with misaligned incentives that encourage state and local governments to rationally fund less effective approaches at the expense of more promising options, a critical digression I have to leave aside for now.)

If not deterrence, then, what about incapacitation? I argue that prison is a bad bet from an incapacitation perspective as well. People age into and age out of offending—a wealth of evidence demonstrates that people’s start criminal careers in their early teens with property crimes, turn increasingly violent in their latter teens and early twenties, and then generally subside in the late twenties or early thirties. So long sentences needlessly detain people.

Worse than that, we rarely impose long sentences on people until they are in their thirties—just as they are likely to start aging out of crime. The average age in California of a person receiving a third strike sentence, for example, is forty-three! Tellingly, when California recently granted over 2,000 third-strike inmates early release, their recidivism rate was one-tenth the state average (5%, vs. 50%)—in no small part because they were older.

Bessette, however, points to a study from the BJS indicating that in a nationwide sample, 69% of all those released from prison over the age of 40 are re-arrested within five years, a rearrest rate not much different than that for those much younger (there isn’t room to go into the tension between the BJS and California results). This leads him to conclude that perhaps those we send to prison don’t actually age out of offending, even if people generally do. I disagree.

To start, for reasons too technical to delve into here, the BJS’s methodology overstates the risk that someone released from prison returns to prison, perhaps by as much as a third—so it’s not so much that 50% return to prison (as that BJS report states) but 33%; the over-estimate is likely even more pronounced for arrests. Furthermore, looking to arrest rates is tricky, since people on parole can be arrested for a lot of trivial things, or perhaps even arrested incorrectly. Unfortunately, the BJS study doesn’t provide results for new convictions or prison readmissions by age, which are more reliable—though still flawed—measures of reoffending.

Also, here Bessette makes what is likely the only clear error in his review. He says that of that 69% of older people who are rearrested, we don’t know how many are rearrested for violence. But we do: the BJS reports that 14% of those over 40 are rearrested for violence, compared to 28% for those 24 and under (see Table 4 here). Even those who have been to prison age out of crime, and out of violent crime in particular.

But there is a deeper conceptual issues to address. Bessette may have the causation backwards: even if he’s right that those leaving prison when old reoffend at high rates, perhaps that is because they were in prison. Prison can be traumatic both mentally and physically, it can sever social ties needed to successfully re-enter the outside world, and people’s skills and talents and health all often decline while locked up. Beyond biology, two major reasons why older people offend less is that they are more likely to be married and to have a job, both of which appear to reduce the risk of offending; time in prison clearly interferes and undermines both pathways out of crime. And research indicates that rehabilitation programs consistently work less well inside prisons than outside their walls.

This causal misstep is common in criminal justice policy: we impose tough sanctions, observe a person reoffend, and immediately assume it must mean we aren’t being tough enough. Perhaps, though, it means we are being too tough, or at least ineptly tough.

Now, to be clear, my points here do not compel someone to be an abolitionist, even from a purely public-safety perspective; I myself am not one. There are some people who pose serious, on-going risks that require us to confine them (although we could, and I believe should, do so far more humanely). And some evidence suggests that short, quickly-impose sentences (more often in local jails than state prisons) can have powerful deterrent effects for specific problems related to drug abuse.

But if public safety is the goal, none of the numbers that Bessette raises require us to turn to the prison as the primary solution, or even a central one. The best estimates suggest that the crime-reducing impact of prisons today is close to (though not at) zero. There are other approaches that not only work better, but are more humane, less costly, and impose fewer social costs on the communities that bear the brunt of enforcement.

What, then, about retribution? Is prison still justifiable, perhaps even required, for the criminal justice system to be, well… just? Bessette suggests that it is. In later posts, I’ll argue that the issue is actually far more complicated.

* * *

Joseph: Let me begin by returning the compliment to John Pfaff for his reasoned response to my somewhat critical review of his work Locked In: The True Causes of Mass Incarceration and How to Achieve Real Reform. He makes a variety of interesting points, drawing on issues debated within the criminal justice literature. Unfortunately, he attributes to me positions regarding clearance rates and recidivism that I did not take in my review and he then devotes much of his response to attempting to refute these positions. More significantly, he avoids the central element of my critique by failing to detail the kinds of large-scale reductions in punishment that he calls for in his book for both nonviolent and violent offenders.

Pfaff describes my position as follows: “Summarizing quickly, Bessette’s pragmatic argument for prisons rests on a few, straight-forward points: (1) we have a lot of crime, violent crime in particular; (2) our clearance rates—the rates at which people are arrested for crimes—are low, requiring us to be harsh to those we do punish; and (3) the recidivism rates of those we release from prison are high, demonstrating the need to keep them locked up for longer periods.” While Pfaff is correct on point (1), I did not take the positions he attributes to me in points (2) and (3).

Let’s start with his second point on “clearance rates.” Clearance rates refer to the percentage of crimes that result in an arrest or are “cleared” in some other way, such the death of the offender in a gunfight with the police. If the police have “solved” the crime—even if they arrested someone who is later acquitted—the crime is considered “cleared” in the FBI statistics. Here is what I said about arrest rates, or clearance rates, in my review: “Of course, not all criminals are arrested. For some of these serious crimes arrest rates are shockingly low: just 29% for robbery and 13% for burglary.” That’s it. But for this brief mention, my entire argument focused not on those who get away with their crimes, but on those whom the police arrest and courts convict. Indeed, my very next sentence was “Yet, altogether, police made over 10 million arrests in 2015, 1.5 million for violent crimes (including misdemeanor assaults).”

Nowhere in the review did I argue that low clearance rates call for “harsh” punishment. (Indeed, I did not advocate “harsh” punishment on any grounds.) I do not believe that the burglars we catch should be punished more harshly because so many others get away with the same crime. (Note that although no one is arrested in 87% of the burglaries in the United States, it does not follow that 87% of burglars are never arrested for burglary. Burglary is a classic recidivist crime, and the more burglaries a criminal commits the more likely he is eventually to be arrested.) The burglars we do arrest and convict should be punished as much as they deserve for the crime(s) for which they were convicted adjusted by considerations of prior record.

It is, rather, the pure deterrence theorist who might argue that low clearance rates should be counteracted by longer sentences. After all, a rational burglar will compare the likely take from a burglary against the average “cost” he might pay for his crime, such as days incarcerated in jail or prison. Pfaff’s book dismisses the deterrent effect of longer sentences, but surely at some point sentence length matters if criminal punishment is to deter at all. If, say, a state reduced the maximum sentence for burglary from several years in state prison to a few weekends in local jail, is it plausible that burglaries would not increase? Prosecutors like Mike Hestrin in California’s Riverside County have been arguing that this is precisely the effect that new laws in California have had in sending convicted offenders, who used to go to state prison for at least a year, to local jails instead, where overcrowding results in just weeks or even days behind bars. He tells the story of “a known thief who had been stealing from the merchant’s store was caught stealing again and working a calculator on his phone. The arresting officers found out the thief had been adding up what he was stealing to make sure he was not taking more than $900 in merchandise so, if caught, he would only be charged with a misdemeanor, no matter how [many] times he was caught.” The result: a few days in jail and virtual impunity to steal again. As we will see, Pfaff’s proposal to bring U.S. incarceration rates more in line with those in Europe will effectively sweep all non-violent offenders from our prisons and jails, as well as many violent offenders.

What, then, of Pfaff’s third point on recidivism? Once again, my discussion was very brief and did not draw the conclusion that high recidivism rates “demonstrat[e] . . . the need to keep [prisoners] locked up for longer periods.” I introduced recidivism rates to raise a caution about a key argument in Pfaff’s book: “For almost all people who commit violent crimes, . . . violence is not a defining trait but a transitory state that they age out of. They are not violent people; they are simply going through a violent phase.” Now, since the context of the discussion is all about incarceration, Pfaff seems here clearly to be discussing violent offenders in prison – that is, the one-half of all state prisoners who were convicted of murder, rape, robbery, or aggravated assault. Unfortunately, Pfaff does not tell us what he means by “defining trait,” what period of time constitutes a “transitory state” (5 years, 10, 15, 20?), nor at what age “almost all people who commit violent crimes” cease to become dangerous. It is no surprise to anyone that criminals with a history of violence are less dangerous in their 60s and 70s than they were in their 20s. But this is not Pfaff’s point; for he gives us some indication of when the aging effect kicks in: “someone who acts violently when he’s eighteen years old may very well be substantially calmer by the time he’s thirty-five.”

Using this as a guide, I presented in my review the best recidivism data we have for inmates released from state prisons. And what I reported was that “fully 69% of those at least 40 years old at the time of release were arrested for a felony or serious misdemeanor within five years—not much lower than the 77% for all offenders. The aging effect, apparently, does not turn most hardened criminals into law-abiding citizens, though it may slow them down a bit.” In his response to my review Pfaff rightly notes that I overlooked a table early in the detailed recidivism report by the Bureau of Justice Statistics (BJS) (tracking state inmates released in 2005) that presents age-related data on rearrests for a violent crime. I thank Pfaff for correcting my oversight since the data I did not report strengthen my case. Here are the 3-year arrest rates for a violent crime by age of the inmate when released:

                          % arrested for a violent crime
Age at release               within 3 years

24 or younger

28.6%

25–29

24.8%

30–34

20.1%

35–39

19.5%

40 or older

14.3%

Note that these data cover rearrests only for a 3-year period, not the full 5 years covered in most of the BJS report. We know from other data in the report that the 5-year recidivism rates are about 13% (not percentage points) higher than the 3-year rates. So, if we had 5-year data for arrests for a violent crime, the numbers would be higher than those reported here for 3 years. Moreover, the reported data almost certainly understate how many inmates commit a violent crime after release from prison since we know from police data that in 2005 arrests did not occur in 38% of murders, 59% of rapes, 75% of robberies, and 45% of aggravated assaults. (FBI, Crime in the United States, 2005, Table 25.) Although it is true that not all of those arrested for a violent crime actually committed the offense, this factor is almost certainly swamped by the large number of violent crimes that never result in an arrest at all.

Now, as the BJS data show, there is, indeed, an aging effect. No one who studies criminal justice would deny that when criminals age beyond their 20s their criminal activity decreases. But do the reported data show that “almost all” violent offenders in prison age out of violence by their mid-30s or even their 40s? Surely not. Note, for example, that while a fourth of those released from prison in their late 20s were arrested for a violent crime within 3 years, so were a fifth of those in their late 30s. While this is progress, it hardly shows the dramatic aging-out effect that is so important to Pfaff’s argument. Even those 40 and over (which, of course, would include some in their 50s and 60s) had an arrest rate for violent crimes equal to 58% of the rate for those in their late 20s. Put another way, BJS found that fully 1 in 7 of the oldest released inmates (40 and older) were arrested for a violent crime within 3 years. It seems likely that the true violent offending figure over a 5-year period (recognizing that about half of violent crimes never result in an arrest) is more like 1 in 5 (or possibly higher). So, we ought to be very cautious about claims that “almost all” violent offenders age out of violent crime (or at least age out of violent crime by their mid-30s or 40s – I have little doubt that “almost all” do if they reach 80 or 90). And since we should care not just about violent crime – we expect government to protect our property as well as our persons – it bears repeating that according to the BJS report fully 69% of those at least 40 years old at the time of release were rearrested for a new crime within five years—not much lower than the 77% for all offenders.

Relatedly, Pfaff maintains in his response that “for reasons too technical to delve into here, the BJS’s methodology overstates the risk that someone released from prison returns to prison, perhaps by as much as a third—so it’s not so much that 50% return to prison (as that BJS report states) but 33%; the over-estimate is likely even more pronounced for arrests.” The reader, unfamiliar with the technical issues at hand, would likely get the impression from Pfaff’s response that the BJS statisticians made some error in how they designed their study, which would be surprising given that BJS’s national recidivism studies, which go back more than three decades, are widely considered the gold standard in such investigations. As best I can tell from Pfaff’s longer discussion in his book and the study that he links in his response, the issue is not whether BJS properly collected and analyzed the data but rather how to describe the findings. The various BJS recidivism studies begin with those released from prison and track them for 3-5 years, compiling state and federal records of their arrests, convictions, and returns to prison. In its latest and most comprehensive study, BJS identified more than 400,000 inmates released from prisons in 30 cooperating states in 2005. It then tracked a sample of some 69,000 for five years and reported the results. One key result: 76.6% of those released were arrested for a new crime within 5 years and 55.4% were convicted of a new crime. It appears from his book (though not from his response) that Pfaff is not actually challenging the accuracy of these findings. Rather, his point in the book is that knowing the recidivism rate of those released from prison is not the same thing as knowing the eventual recidivism rate for those admitted to prison. Compared to those admitted to prison in any one year, those released in a single year will somewhat over-represent repeat offenders, and repeat offenders by definition are those more likely to reoffend. I take it that this is Pfaff’s point in his book. But, of course, this does not mean that the BJS data on those released from prison are in any way inaccurate or, as Pfaff maintains in his response, that there is some problem with BJS’s methodology. To the best of my knowledge, no BJS recidivism report has every held that a study of the eventual recidivism rate for those admitted to prison would produce the same results as their studies measuring recidivism of those released from prison. The fact is that a large majority of those we release from state prisons get in trouble with the law within 3-5 years and about half get convicted of a new crime. Moreover, when you adjust for the length of coverage, these results have held up for more than three decades. High recidivism is a depressing fact. Considerations of both prudence and humanity dictate that we do whatever we can to turn offenders away from a life of crime. But I see no evidence in the national data of the past three decades that we have gotten better at doing just that.

As noted above, Pfaff had attributed to me the following position: “the recidivism rates of those we release from prison are high, demonstrating the need to keep them locked up for longer periods.” Longer than what? Pfaff does not say. If he means longer than the much reduced sentences he would have to advocate to bring U.S. incarceration rates into line with those in Europe, then, yes, prison sentences should be much longer than this new dispensation; for high recidivism rates are powerful evidence of the folly of releasing many additional hundreds of thousands of prison and jail inmates into our communities. But if Pfaff means that I argued in my CRB review that high recidivism rates call for lengthening existing sentences, then, no, I made no such argument. The point of my review was to highlight just how dangerous Pfaff’s proposal was, not to set out my own policies and criteria for guiding the American criminal justice system. Nonetheless, the reader will have noted that I am primarily a retributivist, not a utilitarian. That is to say, I would keep the focus on what convicted offenders deserve, while also recognizing that no responsible policymaker should ignore the deterrent and incapacitative effects of prison sentences. In a word, if we do justice by seeing to it that convicted offenders who deserve time in prison receive it, we will simultaneously deter some others from offending (and perhaps deter the instant offender from re-offending) and we will incapacitate this particular offender from committing crimes in the community while he serves (though he may remain a threat to other inmates and prison officials).

To support his case for punishing even serious violent offenders less than we now do, Pfaff begins his book on p. 1 with the “shocking” comparison of U.S. incarceration rates with those of other liberal democracies: “We have more total prisoners than any other country in the world, and we have the world’s highest incarceration rate, one that is four to eight times higher than those in other liberal democracies, including Canada, England, and Germany.” Then, in his response to my review he presents data from 2015 on three types of crime in six Western European countries, holding that “roughly speaking we don’t look that different.” Yet, despite a similar crime problem, the U.S. has an incarceration rate that is vastly higher than those in the six European countries. If these countries can manage their similar crime problems with just a fraction of the U.S. incarceration rate, why can’t we? 

Pfaff does not tell us why he picked these six countries and these three crimes. Why, for example, include Denmark and Sweden with three of the most populous nations of Western Europe? Why include assault, but not rape and robbery? So, what I have done using Pfaff’s source for crime data in 2015 and the standard source for total incarceration rates around the world—The World Prison Population List—is to reconstruct Pfaff’s table with (1) data from the United States and the five most populous Western European nations and (2) the five most serious crimes for which the FBI reports incidence data for the United States, adding data on residential burglary. 

2015

Country

Homicide

Rape

Robbery

Assault

Burglary

Burglary of residence

Incarceration rate (number)

U.S.

4.9

39

102

238

491

352*

698 (2,217,000)

Germany

0.8

9

55

157

571

206

78 (63,628)

France

1.5

19

158

367

574

351

95 (60,896)

Italy

0.8

--

58

105

--

386

86 (52,434)

England/Wales

0.9

62

88

744

694

336

148 (85,843)

Spain

0.7

2.7

139

63

425

195

136 (63,025)

 Average for   European countries

0.94

23.2

100

287

566

295

109

* 71.6% of burglaries were of residences according to Crime in the U.S., 2015, Table 23.

The U.S. total incarceration rate is 6.4 times higher than the average for the five most populous Western European countries. Put another way, the average rate for the five European countries is only 15.6% of the U.S. rate. Thus for the United States to have the same incarceration rate as Europe, the total number incarcerated would have to be 345,852 (15.6% of 2,217,000). This would require emptying our jails and prisons of some 1,871,148 inmates (2,217,000 – 345,852). Let’s call it 1,871,000.

Now, Pfaff himself concedes that murder rates in the United States are about 4 times higher than in Western Europe. In his response he writes that: “Even our elevated murder rate can’t explain our incarceration rate difference. About 189,000 people were in prison for murder or manslaughter in 2014—about 14% of all state inmates. . . . This is not a trivially-sized population. But, at the same time, that means that our combined prison and jail incarceration rate for non-homicide offenses is still in the 500s, still way above anything comparable to Europe.” So, Pfaff will allow a somewhat higher incarceration rate in the United States for the simple reason that murders rates are so much higher here. So, let’s say that about three-quarters of the 189,000 criminals now serving time in state prisons in the United States for murder or manslaughter (equal to 141,000 inmates) should not be counted if we want to equalize U.S. and European incarceration rates. That reduces the number we must remove from jails and prisons to 1,730,000 of the 2,217,000 now incarcerated. Yet if we (1) closed down every local jail in the country (no more incarceration for thieves, many burglars, and those who assault others short of great bodily injury); (2) removed about half of all federal prisoners (which would have to include some number of drug traffickers, serious white-collar offenders, and violent offenders); (3) removed from our state prisons every single inmate convicted of a property offense (including all the burglars), a drug offense (including all drug traffickers), and a public order offense (including all those convicted of the unlawful sale, distribution, or possession of weapons; driving under the influence; obstructing justice; bribery; and contributing to the delinquency of a minor), we would still be about 290,000 inmates short of our goal. But all that are left in our state prisons are those serving time for murder (171,000), involuntary manslaughter (17,000), rape or other sexual assault (163,000), robbery (169,000), assault (mainly aggravated assault) (134,000), and other violent offenses (42,000). This is a total of 696,000 violent offenders. More than two-fifths of them will have to be freed if we are to reach European level incarceration rates (even controlling for the difference in murder rates).

So what is Pfaff’s plan for getting us into European incarceration rate territory? If our crime problem is about the same (murder excepted), why can’t we get down to something close to Germany’s rate of only 78 persons behind bars per 100,000 residents (just 11% of the U.S rate); or France’s rate of 95 (14% of the U.S. rate), or Italy’s rate of 86 (12% of the U.S. rate), or even England and Wales’ rate of 148 (21% of the U.S. rate)? Unfortunately, nowhere in his book, nor in his response to my review, does Pfaff show us just how he would get U.S. incarceration rates in line with European ones. As this simple exercise shows, it cannot be done without effectively ending anything like genuine punishment for most serious crimes in the United States, including perhaps most serious violent crimes.

So, why then are U.S. incarceration rates so much higher than European ones, even adjusting for differences in murder rates? Much of the answer lies in the simple fact that even apart from murder, the United States remains a more violent place than your typical European country. Here Pfaff implicitly denies what many criminologists have long maintained: that a community’s murder rate is a kind of gauge of its broader violent crime problem. We would expect that if one country had a murder rate four times that of another, it would also be a more violent society in other ways as well. 

Consider rape, for example, a crime missing from Pfaff’s table in his response to my review. As my table above shows, the reported rape rate in the United States in 2015 was 14 times higher than in Spain, more than 4 times than in Germany, and more than twice as high as in France. England/Wales was the outlier, with a rape rate about 60% higher than in the U.S.  (As Pfaff rightly notes, “international comparisons are always fraught with definitional peril”; so caution is always called for in these comparisons.)

As the table above shows, as one moves down the list of crimes from most serious to the less serious, the statistical differences between the U.S. and the most populous European nations seem to disappear. Yet, even here there can be real differences in crime seriousness concealed by the overall data. Consider robbery. The reported U.S. rate is about the same as the average for the other countries: almost twice as high as in Germany and Italy but lower than in France and Spain. Yet another source – the Fifth Edition of the European Sourcebook of Crime and Criminal Justice Statistics, Table 1.2.1.14 (downloadable here)—shows for 2011 the rates for robbery with a firearm, widely recognized by policymakers, judges, and others as a more serious crime than unarmed robbery or robbery with a less lethal weapon. FBI data from Crime in the United States, 2011, Table 19, allow a comparison with U.S. rates. Here are the rates (number of crimes reported per 100,000 residents) for robbery with a firearm for 2011: United States, 45.8; Germany, 4.3; France, 10.0; England/Wales, 4.5; and Spain, 4.3. (Italy did not report.) Thus, the U.S. rate was more than ten times higher than that in three of the four reporting European countries, and more than four times higher than in the other. So, just comparing overall robbery rates between the U.S. and European countries would obscure the fact that gun robberies are much more common in the United States. Thus it would not be surprising if compared to European countries the United States had a higher incarceration rate for robbery.

We know then that the United States has many more murders, rapes, and gun robberies per person than European nations. It likely also have more gun assaults. The FBI reports that in 2015 guns were used in 24% of aggravated assaults in the United States. Given how rare gun robberies are in Europe, it would not be surprising if gun assaults are also much less common there than in the United States.

These are compelling reasons why incarceration rates in the United States are considerably higher than in Europe. Nonetheless, it may well be that crime-for-crime, offenders in the United States are somewhat more likely to be sentenced to incarceration or to serve longer behind bars. Yet it is not obvious what the relevance of such a difference would be to American criminal justice policies. Europe has abolished the death penalty, but 31 American states retain it. But this hardly proves that Europe is right and the 31 American states are wrong. More than a few European nations punish murderers much less severely than does the United States. Anders Breivik killed 77 in Norway in 2011 and received a 21-year sentence, which will expire when he is 52 (though a special preventive detention provision of Norwegian law may allow his indefinite confinement if he is judged to be a continuing danger). Volkert van der Graaf assassinated the Dutch politician Pim Fortuyn during a political campaign in 2002 and though found to me of sound mind was sentenced to just 16 years in prison and was then freed after 12. He now lives as a free man in the Netherlands. Such punishments for mass murder and political assassination are inconceivable in the United States, even in the states without the death penalty. I have not heard any responsible public figure in the United States argue that we should take our direction in punishing murderers from Norway and the Netherlands.

If Europe is to be Pfaff’s guide, he owes us some account of just how he would bring U.S. incarceration rates down to European levels without utterly undermining the deterrent and incapacitative effects of our criminal laws, and without destroying public confidence in our governing institutions for failing to justly punish the guilty.