Thread by @JohnFPfaff

Looking through my tweets, I've attacked Cotton tweets making this claim in 2016 and 2018. Now 2021. It's like a cycle for him.

Here's the main rebuttal: so small a fraction of ppl committing crimes end up in prison that the problem isn't there. It's earlier.

https://twitter.com/TomCottonAR/status/1379438205704019969

People don't trust the police, or don't think reporting to the police is worth their time, or have other reasons to fear going to the police (like immigration status).

55% of non-lethal violent crimes go unreported, and 65% of property crimes.

Of those crimes that the police hear about or witness themselves, the percent that result in an arrest are below 20% for property crimes, and often below 50% for violent offenses... and that's after those low reporting rates.

Now, to be absolutely fair, clearance is not the only measure of police effectiveness: a cop on the corner may deter the crime in the first place, so no arrest needed.

"Low clearance" thus does not automatically translate into "ineffective policing," but it's concerning.

Regardless, low rates of reporting plus low rates of arrest mean that only about 20% of violent crimes and 5% of property crimes result in an arrest.

THAT'S the problem.

Increased risk of DETECTION, not harsher sanctions imposed whenever randomly arrested, is what deters.

After that, the data gets even messier than the usual level of messiness, but a ... challenging ... dataset suggests that about 30% of felony cases get dropped by the DA or judge (but the data is old, is from urban counties, and is... quirky).

After that? While almost everyone whose case moves ahead is convicted, usually via a guilty plea, about 40% are sent to prison, another 30% to jail.

All told? Under 10% of all violent crimes, and under 1% of prop, go to prison.

Which UNDERMINES Cotton's point.

Now, maybe we are just good at choosing the right 10% of ppl convicted of violence...

But the main filters are not reporting and not arresting--not DAs looking at cases and judges sentences--so that seems v unlikely.

Which suggests changing prison pops likely not doing much.

Anyway, this isn't a serious criminological argument. It's not intended to be one.

Cotton is autocratic at heart. He wanted to deploy elite military units against protestors.

And while prison is ineffective at stopping crime, supporting it is a good for of autocratic-signaling.

And will add, just to be clear: if you really want to reduce crime, and are thinking solely within the conventional criminal justice framework, the data is fairly unambiguous.

More policing deters. More prison does not. It's not a tradeoff.*

* Here, economics owes the world an apology. One of Becker's most seminal papers made this v argument: the risk of detection and the magnitude of the sanction are equal tradeoffs: 50% of 2 year in prison = 5% chance of 20: expected value is 1 year either way.

This is wrong.

The data are clear: 50% chance of 2 years deters far more than 5% chance of 20.

But Econ, and then Law & Econ, have been slow to accept that. And those two movements have def influenced how a lot of lawyers/legal academics think about this.