More Research Shows the Success of Harris County’s Misdemeanor Bail Reform

Paul, I want to delve into the finer points of your research, but first I think it’s worth pointing out that since the ODonnell ruling took effect, fewer people in Harris County are detained pretrial. 

That’s correct. Our analysis shows that the federal injunction alone — which was only one component of the overall reform — increased misdemeanor pretrial release rates by about 12%, which equates to thousands of people each year. Of course, under the old system, we were okay releasing most of these folks if they could fork out $50-$500 in cash, but many could not. It’s hard to make the argument that those released under the ODonnell injunction are a big threat to public safety if we were willing to release them previously so long as they had money to buy their release.

That’s an excellent point. Before we jump into your report, can you explain why you thought it was important to look at bail reform in Harris County? 

Well, there’s an ongoing debate in Harris County about whether misdemeanor bail reform worked as intended. You have some folks, like the district attorney, arguing that releasing more people charged with misdemeanors has fueled an increase in crime. There are other folks like the federal bail monitor who say the reforms have been very successful. [As part of the ODonnell settlement, a federal monitor was appointed to track the reforms for seven years and provide detailed reports to the public.]

These sorts of discussions aren’t just occurring in Harris County. There are many other communities thinking about or trying to implement various types of pretrial reform. This question of whether releasing more people before trial who have been accused of low-level offenses impacts public safety, and in what ways, is an important one.

What approach did you take to assess the ODonnell reforms?

We focused on a key component of the reform — a federal ruling that required the county to release, on an unsecured basis, people who would have been previously detained when unable to pay small amounts of cash as secured bail. This approach was ultimately included as part of the settlement of the ODonnell case.

We took advantage of the fact that this policy was implemented all at once on a particular date — all of a sudden you release a set of people charged with misdemeanors who would have been detained had their cases come up just a few days before. While at the same time, it was basically business as usual in the felony courts. We were able to use that natural experiment of the onset of the injunction to tease out the effect of making more people eligible for release. What we found is that misdemeanor bail reform has been pretty successful.

When you say successful, how do you quantify that?

We see a few things. First of all, the people themselves end up with more lenient sentences — they are less likely to plead guilty, less likely to be sentenced to jail, and more likely to get shorter sentences. The decrease in guilty pleas is particularly important because we know one thing that pretrial detention can do in low level cases is create perverse incentives to plead guilty, even when you are innocent. But even though we increase freedom from custody and ratchet back incarceration, there is no adverse impact on public safety, and that’s true even when we look out many years after the initial case. We do see some evidence that releasing people earlier can reduce the likelihood that their case concludes properly —“failure to appear,” if you will — but this happens in a very small fraction of cases. 

One of the big takeaways for other jurisdictions considering similar bail reforms is that it’s possible to implement such a reform without increasing crime or creating lots of problems.

Yet you have this false narrative both in Harris County and in many places across the country that bail reform is leading to an increase in crime. What do you think is driving that?

I think there are a few things going on. First of all, people often talk about ​“bail reform” without being very precise about what exactly they are referring to. So, for example, we might get Crime Stoppers publishing lists of people who are released multiple times after being charged with felonies, who then go on to commit serious crimes while released. Someone might look at that and say, ​“Well, that just shows that bail reform isn’t working.” But, of course, the ODonnell reforms have nothing to do with people charged with felonies; they only apply to misdemeanors. More broadly, often what happens is that someone sees something wrong with the pretrial system — and I agree, there is a lot that could be improved — but then they conflate that with ​“bail reform,” and conclude it must be the fault of ​“bail reform.” That is a mistake.

Another common problem is that people often claim ​“bail reform” caused this or that, when in reality it isn’t clear whether we are seeing the impact of pretrial reform or some other factor that generally affects crime patterns. For example, Harris County, like many other jurisdictions in the U.S., has seen an uptick in certain categories of crime recently. It seems natural to think that because we’ve had bail reform in the past few years and there’s also been an uptick in crime the two must be linked, and indeed some policymakers in the county have claimed as much. But of course, there have been natural disasters, a pandemic, changes in gun availability — there are all sorts of other things that have changed in the past few years as well that may be driving crime rates.

How do you cut through the noise and those misleading claims to spotlight your research and show these reforms are having a positive impact on communities and public safety as a whole?

One contribution we make — and one reason for doing this research — is we think very carefully about disentangling the effects of expanding misdemeanor pretrial release from other stuff going on in the system. We try to do this in a pretty convincing way, using methods well accepted in the econometrics and statistical community for teasing out cause and effect. 

For example, we know that different types of people are moving through the system at different points in time, so one of the things that we do in our analysis is control for hundreds of contextual variables that capture the underlying characteristics of people, what they are charged with, etc. This helps ensure that we are capturing the true impacts of the injunction, rather than the fact that you have different types of people showing up in, say, 2016 versus 2018 or 2022. As another example, we do a bunch of ​“stress testing” of our results to ensure that we get the same conclusions across different samples, different ways of formulating the analysis, etc.

We try to help people understand in an accessible way how our methodology works, and why it might lead to a different answer than just looking at time trends or looking at samples that have been selected in various ways, as some past faulty research on this topic has done. These are complicated problems with lots of moving parts, and it isn’t easy, but it’s super important to get these narratives right.

It helps that in Harris County we have this abrupt shift, so if you look at the raw data, you can see much of the story there. For example, there’s a clear change in patterns of conviction rates that is specific to people charged with misdemeanors and that starts right when they begin releasing more people pretrial under the injunction.

We spoke with Ames Grawert with the Brennan Center a few weeks back and he had this quote that’s stuck with me: ​“Crime is complex, and policymakers should be wary of simplistic answers.”

I think that’s great. One thing that bothers me about statements like ​“bail reform isn’t working” is that these claims are too simplistic. Bail reform encompasses all sorts of potential interventions, so we need to be very clear about what sort of an intervention we are talking about, and compared to what alternative. It’s very different to ban cash bail, introduce a risk assessment tool, or require judges to consider dangerousness as well as flight risk when making pretrial release decisions, but these are all arguably ​“bail reform.” 

It turns out when we are very careful and specific about what it is we want to measure — for example, when we say, I want to measure the effects of releasing more people charged with misdemeanors — and then apply appropriate empirical methods, it is sometimes possible to get a good answer. Not necessarily a simple answer, but one that reflects the reality of what is happening, and that can offer useful evidence for other jurisdictions considering something similar.

Bail often harms people who are economically disadvantaged and can’t afford to pay for their freedom. How have Harris County’s reforms affected that group of people, and those who come from communities historically burdened by over policing?

As part of our analysis, we looked to see whether the injunction had differential effects for certain groups of people. We found that that the beneficial impacts of the reform were widely felt. If we look at people who come from poorer vs. less poor ZIP codes, people with and without a prior record, or Black or Hispanic people, for all these groups we see that the reform was generally beneficial in terms of facilitating release, reducing conviction rates, and reducing jail. For no group was there any evidence to suggest that the injunction caused an increase in crime.

Speaking of going forward, what’s the next move? There’s a wealth of research coming out about the benefits of misdemeanor bail reform in Harris County, yet so much misinformation to the contrary. 

Well, step one is getting the findings of rigorous, well-designed research studies in the hands of the public, which is hopefully what efforts such as those of NPPJ can help accomplish. Beyond that, there is still much to learn. One issue that hasn’t been addressed much thus far is impacts of these types of reforms beyond the criminal justice system. For example, misdemeanor bail reform may contribute to higher earnings, better job stability, or produce other benefits that make these policies more attractive, but these sorts of outcomes have rarely been studied.

Of course, there are also big concerns about what’s happening to people who are charged with felonies, and there’s ongoing litigation about that in Harris County. As both victim-oriented organizations like Crime Stoppers and more progressive organizations like the Texas Center for Justice and Equity have pointed out, we need better empirical evidence to understand what’s going on among that population. Extending our insights into people who are charged with more serious crimes for Harris County — and for other jurisdictions — is certainly important.

What lessons can be learned from this type of systemic reform by other jurisdictions considering making their own changes?

This is an area where you can start small and move your way up. Our evidence is pretty compelling that detaining people on lower-level offenses and using cash bail doesn’t make a lot of sense. So that might be a good starting place for people who haven’t yet waded into reform.