THREAD: I have some thoughts on SB10, which was just signed today by Governor Brown.
If this is all that gets accomplished, it's not enough. There is much, much more to do when it comes to reforming pretrial release/detention practices in California (and they should be, overwhelmingly, release practices)
So if SB10 is the finish line, I'm disappointed. If it's the starting line, OK. In the interest of full disclosure, I pitched @robbonta on the importance of bail reform a couple of years ago and have worked with the coalition that supported (then didn't support) SB 10.
I completely understand why progressive support for SB10 dried up after the amendments, and I share all of those concerns. What initially excited me about bail/pretrial reform was the idea that we could change the presumption from detention to release.
I think, following @Pretrial ED Cherise Fanno Burdeen's suggestion, that we need to make pretrial operate like this: the prosecution needs to do something to keep someone in. Today, it's that the defense needs to do something to get someone out.
Make the prosecution do the work to detain someone. Make them prove that someone is dangerous. Don't make the defense prove they aren't dangerous.
The harms to detention are certain: You can't work. Your family income suffers, your kids and dependents suffer, you can't help build a defense. You plead out to time served because who wouldn't?
The harms to release are speculative. What if this guy goes out and commits a crime? We overweight the speculative harms and ignore the certain harms. We detain most because one person might be risky.
RA's can help avoid this. They don't say "1 in 100" means you should detain all 100. But they do have problems, of course.
Our initial draft used risk assessment tools to get people out only--if you scored low, you were released. If you scored high, you got kicked into a full, adversarial proceeding (at which the RA score was not evidence), with a lawyer, right to confront/call witnesses, etc.
We also said RA's had to be "trained" on open data sets, that they could not be proprietary/black box (you have to know how the score was calculated), they have to be validated, etc. We still need those safeguards.
We need to make sure RA's are actually accurate. Don't take something on faith just because it has math.
While RA's have problems, so do judges--and you can't audit a judge's "gut" the way you can actuarial tools. Judge's aren't necessarily good at prediction. They only see the failures return to court. Successes don't come back.
The virtue of discretion depends on whether you're the one exercising it or whether it benefits you. If a judge makes a decision, we should be able to know why--and, more importantly, what their track record is.
Justice Alito is always writing these doomsday predictions about what will happen if we do anything remotely progressive in CJ--see, e.g., his dissent in Plata--but he's been wrong about all of it!
There are very real concerns about disparate impact from race/SES. But that's because we live in a world with structural inequities.
POC are more likely to be arrested than I, a white man, am because of structural racism. Any RA is going to reflect that reality. So "risk" of rearrest is just that. It's not a proxy for risk of reoffending.
Scores have to be contextual. What are we measuring, and is it what we want to measure?
My race makes me less likely to be arrested, and an RA will reflect that. What it _doesn't_ mean is that I or other white people are less likely to commit crimes. An RA doesn't say why the risks are different. It just says "this is what the system is likely to produce."
Most people would agree with this. Who is more likely to be pulled over--a white person or a black person? A risk assessment tool would say the same thing. The distinction is that we would know why the former assessment was made.
So if we move toward RA's, we are going to have to really consider how we want them to operate, what we think they are telling us (versus what they _are_ telling us), and what we do with that information. Like, maybe we start policing white neighborhoods more heavily!
You can always get rid of racial inequities by leveling up--and then we can really see how much popular support there is for being "tough on crime." I'll believe that @jeffsessions is tough on drugs the day white fraternities at 'Bama get raided--but that's another story.
So we need to move forward with our discussion about and implementation of RA's. I want state standards, and I want the above points to be where we start. Other than that, we need to make sure that the problem doesn't just shift.
I think we need to watch carefully to make sure that the bail bond industry doesn't move towards other profit-seeking nodes in the CJ system (like, say, electronic monitoring with costs paid by those monitored). Rent-seeking distorts decisionmaking!
And I'm sure the bail bond industry will fight this in court and via lobbying. They've made tons of money with basically no risk--and with all costs of enforcement privatized. That's a sweetheart deal for them--and a ripoff for us. /end
