Facts on recidivism undermine case for leniency legislation, Part Two

Taylor Millard, one of the excellent writers at Hot Air, has taken issue with an article by Daniel Horowitz on recidivism as it relates to sentencing reform, including the FIRST STEP legislation (which is back door sentencing reform). I cited Horowitz’s article in my post, “Cold Facts on Recidivism Undermine Case for Leniency Legislation.” Thus, I want to bring Millard’s arguments to the attention of readers interested in the issue and try to explain why Horowitz’s points (and mine) withstand them.

Millard made three main points. First, Millard criticized Horowitz from relying on recidivism numbers based on re-arrest rates because people are sometimes arrested mistakenly. Second, Millard claimed that the recidivism numbers Horowitz used from Texas, where rehabilitation programs supposedly have cut recidivism, predate the implementation of the rehabilitation programs. Third, Millard argued that Texas’ crime rate numbers, which have been going down, are more relevant than recidivism rates to the debate over criminal justice reform.

As to the first point, Millard misunderstands Horowitz’s argument. His point (and mine) wasn’t that re-arrest rates are a perfect indicator of recidivism. Indeed, I expressly acknowledged that “not everyone arrested has committed a crime.”

Horowitz’s point was that those who tout Texas’ rehabilitation program do so by comparing Texas re-incarceration rates with federal re-arrest rates. But a fair comparison must be “apples-to-apples” — either federal re-arrests to Texas re-arrests or federal re-incarcerations to Texas re-incarcerations.

While we’re on the subject, though, I’ll argue that re-arrests are a much better indicator of recidivism than re-incarcerations. Sure, not everyone who’s arrest committed a crime. Law enforcement makes mistakes.

It’s also the case, however, that not everyone who commits a crime gets arrested. In fact, according to the Pew Research Center, hardly a conservative source, “most crimes are not reported to police, and most reported crimes are not solved.” Thus, arrests reflect less than a quarter of total crimes. Unless one believes that three-quarters of arrests are of the wrong person, the arrest rate understates the recidivism rate.

Millard’s second point is that the recidivism numbers Horowitz used from Texas, where rehabilitation programs supposedly have slashed recidivism, predate the implementation of the rehabilitation programs. Millard points out that the adjusted Bureau of Justice numbers cited throughout Horowitz’s piece (and mine) looked at prisoners released in 2005, two years before those who underwent Texas’ new rehabilitation program re-entered society.

But Horowitz didn’t cite the Bureau of Justice numbers in discussing Texas recidivism. Instead, he cited the National Association of Assistant U.S. Attorneys.

What data set did that organization rely on? I’m not certain, but it appears to be data published by the State of Texas in 2017 — data that comes from the period after Texas instituted its correction reforms. (See page 2, Figure 2, in which the 2011 number for re-arrests among those released from prison (which is almost the same for 2012 and 2013) matches the number Horowitz cited).

Millard’s third point is that we should focus on crime rates, not recidivism rates. The crime rate in Texas has been falling since the state adopted corrections reform, Millard observes.

But for purposes of discussing the sentencing and release of prisoners, it’s the recidivism rate that should guide us. The overall crime rate tells us the extent to which the entire population elects to engage in crime. The recidivism rate tells us the extent to which the previously incarcerated population elects to do so.

Only the incarcerated population is the subject of sentencing reform. It’s the decisions its members make after release that should guide us when considering such reform.

Unfortunately, the formerly incarcerated, including Texans, elect to commit crime at an astoundingly high rate — far too high to make sentencing them leniently or releasing them early a good idea, if we’re at all concerned about public safety.

Notice: All comments are subject to moderation. Our comments are intended to be a forum for civil discourse bearing on the subject under discussion. Commenters who stray beyond the bounds of civility or employ what we deem gratuitous vulgarity in a comment — including, but not limited to, “s***,” “f***,” “a*******,” or one of their many variants — will be banned without further notice in the sole discretion of the site moderator.