Tuesday, September 26, 2017

Interpreting new UCR crime data: Context is everything

John Pfaff offers a sensible way to look at recent Uniform Crime Report data: Still near the bottom of a trough on crime rates, despite two years of increase in violent offense totals. Usefully, Pfaff places the much-touted increase in homicides in 2015 and 2016 in context with this graph:

Year-to-date murder totals in 2017 are headed back down, according to published reports, but those mid-year estimates are often wrong. One simply must have patience before declaring a trend either way; data can't be manufactured which doesn't exist yet. Regardless, Grits would consider three years in a row of increases a full-blown trend. If it drops back down, then from a statistical perspective, maybe the increase was just noise and the trend could even continue in the other direction. ¿Quien sabe?

Either way, as Pfaff points, out, reformist policy prescriptions don't really change. We don't prevent violent crime by incarcerating low-level nonviolent offenders, arresting people for traffic-ticket debt, or keeping elderly, parole-eligible inmates incarcerated long after they cease to pose a public safety threat. Indeed, so many criminal-justice methods have been proven ineffective and counterproductive - e.g., evidence now strongly shows that incarcerating low-risk offenders increases their recidivism rates - that, if crime is rising, it becomes more imperative than ever to reform the justice system to focus on people who pose the most serious risk. Resources are scarce in the post-Harvey era and it doesn't make sense to spend them on anachronistic policies that don't work.

RELATED: From the Texas Tribune, "Texas murder rate went up again last year, remains relatively low."

MORE: At Sentencing Law and Policy, Doug Berman offered this interesting observation:
At the risk of seeming a bit too Pollyannaish, I think the FBI report that property crimes in 2016 dropped for the 14th consecutive year is a big piece of the national crime story very much worth celebrating. Though violent crimes rates understandably get the most attention, property crimes impact the most people — there are, roughly speaking, more than five property crimes for every violent crime — so drops [in] property crimes can end up meaning a lot more persons and families experienced a crime-free year even when there are spikes in violent crime.
AND MORE: From Thomas Abt in the New York Times.


Anonymous said...

LOL... your focus on reduced property crimes appears to be Rule 2 of the Police Union playbook, Grits. Come on man! People want to feel safe. Starting to feel a little squeamish about these numbers, eh?

What we're beginning to see here is history repeating itself. We're slowly drifting back to the parole revolving door policies of the 70's and 80's where violent crime rates jumped off the charts. The only thing that's restraining those numbers now is relatively decent economy where jobs are available. Let unemployment begin to rise and you and all the other "smart on crime" liberals are going have to start wiping egg off your face.

Here's some tried and true conventional wisdom that anyone with common sense can relate to: Close prisons=violent crime goes up; Build more prisons=violent crime goes down.

Oh, one other thing. Most people understand that many property crime offenders eventually progress to violent crimes. The more you leave on the street under the auspices of some feel good rehabilitation program, the more likely they create a new violent crime statistic.

Anonymous said...

"...anyone with common sense...", "Most people understand..."

Got any statistics or trustworthy references to go with your equations?

Most people with intelligence could back up their opinions with data, not strawmen or insults.