Sex Crimes

The 'Frightening and High' Factoid About Sex Offender Recidivism Still Stalks Courts Across the Land, Completely Untethered From Actual Numbers

Justice Anthony Kennedy's bogus 2002 claim figures prominently in defenses of an Arizona bail ban.



Last May the Arizona Supreme Court overturned a state ban on bail for people accused of sexual assault "when the proof is evident or the presumption great," concluding that the categorical exclusion violated the constitutional right to due process. Critics of that decision are urging the U.S. Supreme Court to take up the case, Arizona v. Goodman, and their arguments highlight the continuing influence of misconceptions about the "frightening and high risk of recidivism" among sex offenders.

That phrase comes from Justice Anthony Kennedy's plurality opinion in the 2002 Supreme Court case McKune v. Lile, where he claimed "the rate of recidivism of untreated offenders has been estimated to be as high as 80%." Kennedy's characterization of the recidivism risk as "frightening and high" has been echoed in scores of decisions upholding restrictions on sex offenders. But the original source for his recidivism estimate was an uncorroborated assertion in a 1986 Psychology Today article by a therapist who has repudiated the number, saying he is "appalled" at its lingering impact.

An example of that impact can be found in a brief from Arizona Voice for Crime Victims (AVCV), a group that wants the Supreme Court to hear Arizona v. Goodman and uphold the state's bail ban. The brief, which was written by AVCV lawyer Steven Twist, Dallas attorney Allyson Ho, and University of Utah law professor (and Volokh Conspiracy contributor) Paul Cassell, mentions "frightening and high" recidivism rates three times; it also refers to "frighteningly high," "alarmingly high," and "extremely high" recidivism rates. Yet the numbers cited in the brief fall far short of the discredited estimate that Kennedy cited in McKune:

  • "Studies have found that 17 percent of sex offenders were convicted of another sex offense within five years of release—with 21 percent reconvicted within ten years." That is based on an analysis of studies that included rapists and child molesters but not other kinds of sex offenders (such as people caught with child pornography).
  • "One study shows that over a five-year period, 21.4 percent of sex offenders were rearrested for violent offenses—nearly identical to the 21.7 percent of homicide convicts who were rearrested for violent offenses during that same period." That number is for people convicted of rape or sexual assault. If anything, it indicates that sex offenders don't have unusually high recidivism rates.
  • "A Department of Justice study found that a significant number of sex offenders—14 percent—not only reoffend, but also do so while out on bail." That figure applies to people charged with rape who were arrested while on bail; only 3 percent were charged with felonies, and it's not clear how many of those were sex offenses.

Even the highest recidivism rate cited by AVCV is 73 percent lower than the estimate on which Kennedy relied. And while it may be appropriate in this case to cite figures for rapists, it's important to keep in mind that sex offenders are a diverse group, many of whom have never assaulted anyone. Furthermore, some subgroups of sex offenders have much higher recidivism rates than others.

A 2014 meta-analyis covering almost 8,000 sex offenders found a five-year recidivism rate of about 20 percent among "high-risk" offenders but less than 3 percent among the rest. The recidivism rate for the high-risk offenders rose to 32 percent after 15 years but remained flat thereafter, indicating that even offenders initially classified as "high risk" pose little or no continuing threat if they go 15 years without committing another sex crime. The 15-year recidivism rate for low-to-moderate-risk offenders was just 5 percent.

The AVCV brief quotes a 2013 Supreme Court decision that said "there is evidence that recidivism rates among sex offenders are higher than the average for other types of criminals." Yet data cited by the brief contradict that contention.

"An analysis of 400,000 state prisoners," AVCV says, "found that 21 percent of sex offenders were rearrested for a crime within six months of release, 31 percent were rearrested within one year, 44 percent within two years, 51 percent within three years, and 60 percent within five years." Those numbers, which apply to people convicted of rape or sexual assault, are for any type of new crime, as opposed to a sex offense specifically. Even so, the five-year recidivism rate for these sex offenders was lower than the rates for people convicted of robbery, assault, burglary, larceny, fraud, drug offenses, and "public order" offenses. The only group with a lower five-year recidivism rate (51 percent vs. 60 percent) was people convicted of homicide.

The total recidivism rate highlighted by AVCV is more than 10 times as high as an arguably more relevant figure from the same study: Less than 6 percent of rapists were arrested for a new rape within five years of their release. Even more relevant: The figures for defendants on bail, the focus of this case, indicate that only 3 percent of people charged with rape were arrested for any sort of new felony. By comparison, 6 percent of defendants charged with assault, 11 percent of defendants charged with robbery, and 12 percent of defendants charged with burglary were arrested for felonies while on bail.

The fact that 97 percent of the people charged with rape or sexual assault were not arrested for a new felony while free on bail certainly seems worth considering if you are trying to assess the case for denying such defendants bail as a group. So does the fact that other kinds of defendants—defendants who are not subject to a categorical bail ban in Arizona—are more likely to be arrested for felonies while on bail.

Even if you think Arizona's bail ban is good policy (or at least constitutional), you should be troubled by the continued judicial reliance on repeatedly refuted claims about sex offenders that were erroneously endorsed by the Supreme Court 16 years ago. Recidivism rates for sex offenders are not "higher than the average for other types of criminals" (with the exception of murderers), and they are nowhere near as high as Justice Kennedy claimed in the ruling that everyone keeps quoting. In cases that hinge on the danger posed by sex offenders, those facts should matter.