« "The Future of Juvenile Appeals in the United States" | Main | Concurrence laments "trend" of federal prosecutors seeking "significantly enhanced terms of imprisonment under the guise of 'relevant conduct'" »

October 3, 2014

SCOTUS preview guest-post: "Measuring the Dangerousness of Felonies for Sentencing Purposes"

ThIn this post I lamented that the Supreme Court this week did not grant cert on any new sentencing cases.  But there is still some sentencing fun on the SCOTUS docket thanks to the Justices seemingly never having enough fun with interpretations of the Armed Career Criminal Act.  Helpfully, Professor Stephen Rushin, who filed in an amicus brief in the latest ACCA case, was kind enough to prepare for posting here a thoughtful preview of a case to be argued to the Justices in early November.

With kudos and thanks to Prof Rushin for this material, here is his preview:


What criminal offenses pose the greatest risk of injury to others? This is the empirical question at issue in a case, Johnson v. United States, before the U.S. Supreme Court this coming term. The case stems from the Armed Career Criminal Act (ACCA), which provides for punishment enhancements for offenders previously convicted of burglary, arson, extortion, use of explosives, and any other felony that presents “serious potential risk of injury to another.”

Since the passage of the ACCA, courts and litigants have struggled to determine which felonies pose such a “serious potential risk of injury to another.” The Court has interpreted this so-called residual clause of the ACCA to cover a range of felonies, including attempted burglary and fleeing from a police officer in a motor vehicle.    

In Johnson, the Court must now decide whether the residual clause also covers the possession of a short-barreled shotgun. So how dangerous is mere possession of an unlawful weapon? Professors Evan Lee, Eric Johnson, and I recently submitted an amicus brief in the Johnson case, arguing that the ACCA ought to cover these sorts of weapons law violations. 

At first, our argument may seem counter-intuitive. How, after all, can mere possession ever pose a “serious potential risk of injury to another?” Well that depends on how you define a “potential risk of injury.” Admittedly, offenses like weapons possession cannot, or usually do not, injure another person directly. But that does not mean that such offenses do not pose “serious potential risk of injury to another.” Congress’s use of the word “potential” in conjunction with the word “risk” suggests that a felony need not be the direct or exclusive source of an injury in order to qualify under the residual clause. We read the ACCA to mean that any offense that facilitates or is otherwise meaningfully associated with highly injury-prone offenses “poses a serious potential risk of injury.”

Of course, this raises the next obvious question—to what extent are weapons law violations, like possession of a short barreled shotgun, associated with injuries to victims? In previous ACCA cases, the Court has turned to a wide range of statistical data to measure the dangerousness of various felony offenses. In each case, the Court has attempted to find accurate statistical measures of how frequently a particular felony offense leads to injuries. The Court then compares this to the approximate injury frequency of injuries stemming from the offenses explicitly enumerated in the ACCA—burglary, arson, extortion, and use of explosives. 

This basic methodology makes perfect sense. Since Congress specifically enumerated a small number of offenses as “violent felonies” in the ACCA, the Court should presume that any offense of equal or greater dangerousness also warrants inclusion under the residual clause. But in employing this methodology, the Court has often relied on weak statistical data.

In entering into this ongoing debate, my coauthors and I make a simple recommendation to the Court in our amicus brief. We suggest that the Court should use the National Incident Based Reporting System (NIBRS) in measuring the dangerousness of offenses under the ACCA residual clause. For the unfamiliar, we have traditionally recorded crime data in the U.S. via the Uniform Crime Reports (UCR), which primarily record aggregate-level information on the prevalence of eight major criminal offenses—homicide, aggravated assault, rape, burglary, larceny, arson, and auto-theft. With the exception of homicides, these UCR records little to no details about the circumstances surrounding each offense. Recently, though, the FBI has begun collecting additional crime data through the database known as NIBRS. This system requests information from local law enforcement agencies on 46 different offense categories. NIBRS also groups together criminal offenses into incident-level data. This means that if an offender commits two different offenses as part of a single criminal incident, NIBRS groups these two offenses together for data analysis purposes. For example, suppose that an offender commits an assault in the course of committing a burglary. Traditionally, the UCR would register that event as two separate criminal events. By contrast, NIBRS groups together these two criminal offenses into a single incident. Police agencies that use NIBRS also report information on the circumstances of each criminal incident, including whether the incident resulted in any physical injuries to victims.

Of course NIBRS is not perfect. The NIBRS database is not perfectly representative of the United States. Although NIBRS greatly expands on the number of offense categories traditionally used in the UCR, it still cannot capture every single offense category. Nevertheless, NIBRS represents perhaps the best statistical resource available for measuring the “potential risk of injury” associated with felony offenses. For one thing, NIBRS represents the largest and most comprehensive database on injuries associated with criminal offenses. In addition, because NIBRS groups together multiple offenses into incidents, it allows researchers to measure more accurately the risk associated with criminal offenses. And NIBRS allows the Court to compare the dangerousness of different felony offenses accurately because it uses a consistent methodology across reporting jurisdictions.

So how do weapons law violations stack up compared to the explicitly enumerated felonies listed in the ACCA? In a previous study, Evan Lee, Lynn Addington, and I found that weapons law violations like possession of a short-barreled shotgun were more frequently associated with injuries than burglaries, arsons, or extortions. 5.36 percent of incidents involving weapons law violations in 2010 led to some type of physical injury to a victim, compared to just 4.41 percent of extortions, 1.11 percent of arsons, and 1.02 percent of burglaries.

Of course, these sorts of statistics alone cannot resolve the question before the Court. But we argue that this data cuts in favor of including weapons law violations under the ACCA residual clause.

October 3, 2014 at 11:28 AM | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference SCOTUS preview guest-post: "Measuring the Dangerousness of Felonies for Sentencing Purposes":


Come on! Crimes are not dangerous. People who commit certain kinds of crimes are dangerous. The problem is people; people who are at risk of committing crimes. The crimes they commit is evidence that they may commit another crime. Some kinds of crimes are more predictive then others. But it takes additional information to accurately asses risk; e.g., current age, family history and so on.

Penalties are set before the fact by lawmakers, at a time when the problem is not fully knowable. They are intended to forestall criminal conduct on the theory that potential criminals will be dissuaded by a potential penalty. As such they are rather crude devices that have some, but limited utility. There is a good deal of research, which indicates that potential criminals are not influenced much by threatened penalties. Ask anyone-what is the penalty for burglary? Who knows?

Posted by: Tom McGee | Oct 3, 2014 11:08:46 PM

Mala prohibita are not dangerous.

Dangerousness should be restricted to past acts that caused serious injury. Serious is defined as an injury requiring professional medical attention. A bruise, not serious. A broken nose, serious.

Then count, 1, 2, 3, dead. Even if the death penalty is falsely applied, one is getting rid of a bad guy, based on count 1 and 2, to everyone's benefit, especially those closest, his family. Start the count at the age of adulthood, 14. No violent criminal sees 18, the beginning of their busiest phase of life.

Assume crime cannot be generally deterred. Only incapacitation remains. There is a very high murder rate in the drug and crime world. The profits and pleasures are so huge and intense, risk of death does not deter.

I come out of prison. I have a choice, $20,000 a year at McDonald's, $2 million a year stealing cars on consignment for body shops, at age 17. Kidding? Going straight is sheer madness. What were the geniuses here doing at age 17? I am referring to a real person with a real income.

Posted by: Supremacy Claus | Oct 4, 2014 12:13:29 AM

In other words, an offense that results in no injury 95% of the time creates a "serious potential risk" of injury. Uh huh.

Posted by: Dubious | Oct 4, 2014 8:59:45 AM

Dubious beat me to it. But the actual percentage, since the calculator is already out of the toolbox, is on average 2.18% violent. That means that these violent crimes are 97.82% nonviolent. Put another way, 97.82% of these felons are presumed to be as violent as the 2.18% that were prima facie violent. A rather sneaky way to sneak in the statistically false presumption of guilty. Another question is what percentage of that 2.18% of violence required medical care? My bet is it would fall to 1% are actually dangerous.

If so, 99% of these presumed "violent" felons pay the penalty for the violence of the 1%.

Where's Occupy when you need them?

Posted by: George | Oct 4, 2014 2:07:41 PM

nothing unusual there. these gov't fuckups or frauds use the same math in the registry scheme.

90% of the studies done in the last 20 years including everyone the DOJ does shows re-offense rates for most sex crimes is 5-15%. but they have no problem fucking up the lives of 1,000,000 individuals because 50 to 100,000 of them MIGHT reoffend.

Posted by: rodsmith | Oct 5, 2014 1:10:59 AM

Post a comment

In the body of your email, please indicate if you are a professor, student, prosecutor, defense attorney, etc. so I can gain a sense of who is reading my blog. Thank you, DAB