The frequent detention of these kinds of offenders is also an issue at the federal level. This week, the Department of Justice’s inspector general, Michael Horowitz, released an internal audit that provides concrete evidence of how infrequently federal prosecutors avoid jailing low-level offenders by diverting them or allowing them to await trial at home. “In fact, we found nearly half of the districts rarely used pretrial diversion,” Horowitz said in a taped summary of the findings. “We also determined that there were a significant number of additional low-level and non-violent offenders potentially suitable for pretrial diversion.” Offenders who don’t remain in jail before their trials often end up receiving lesser sentences or no prison time, sometimes even avoiding a conviction altogether. The report estimates that tens of millions of dollars would be saved by incarcerating fewer low-level offenders.

Facing the possibility of spending months in jail waiting for trial, many defendants plead guilty, even if they have to pay hefty fines or endure years of probation. “A lot of people don’t think about or don’t realize some of the collateral consequences of pleading guilty in terms of the effect it has on your criminal record, the ability to find jobs, to get into schools, to get a variety of public benefits,” said Megan Stevenson, an economist and professor at the University of Pennsylvania Law School.

Stevenson conducted a study of the impact of bail fees on residents in Philadelphia which concluded that “those who can’t afford bail are 13 percentmore likely to be convicted and will receive incarceration sentences that are on average five months longer.” In that city, only 51 percent of those charged with bail of $500 or less were able to pay the minimum 10 percent required to go free within three days. Her analysis also provides evidence that “pretrial detention increases the likelihood of being convicted, pleading guilty, being sentenced to incarceration, and being required to pay hundreds or thousands of dollars in court fees.” I recently talked with Stevenson about her work. An edited version of our conversation follows.

Megan Stevenson: We’ve already known that people that are detained are more likely to be found guilty, and are more likely to have unfavorable case outcomes. But those who are detained are systematically different from those who are released. They’re generally facing more serious charges, have long criminal histories, and may be less able to afford a good lawyer. I identified a natural experiment in Philadelphia that enabled me to say definitively that pretrial detention has this causal effect on case outcomes. It causes people to be more likely to plead guilty, and to plead guilty to worse terms.

Juleyka Lantigua-Williams: Only because of the immediacy of being set physically free sooner?

Stevenson: That’s one mechanism: the immediacy of being physically free. Also, some of the most damaging effects of pretrial detention can happen really quickly within the first few days or a week. If you lose your job, if you lose your apartment, if you need to find somebody else to take care of your kids, at that point the cost of future incarceration might not be so high after you’ve already been locked up for a week. It reduces incentive to fight against a plea deal that involves another six months of jail. Basically, if you’re out and you’re at home the idea of pleading guilty to six months of jail might sound a lot worse than if you’re already in jail, and you’ve already lost your job, and you’ve already lost your apartment.

Stevenson: There’s a bail reform movement going on across the country, and that’s fantastic, but I don’t think it’s clear yet how far the reforms are going to go, or what form they’re going to take … I think the more that people understand how detrimental detention is for the case outcomes, I think it builds support for the argument that, to whatever extent we possibly can, we want to reduce any socioeconomic disparities in terms of who gets detained. In Philadelphia all you have to pay is a 10 percent deposit of your bail in order to get out, so if you have a $500 bail all you need to come up with is $50. About half of these people can’t come up that money, or at least can’t come up with it within the first three days. Even with very, very low amounts of bail you’re detaining a lot of people that don’t really need to be detained. These people are poor people, and because they’re detained, they’re more likely to be convicted, they’re more likely to plead guilty.

Something like 30 or 40 percent of shoplifters are detained, making up three percent of arrestees. You don’t need to detain anyone who is just charged with shoplifting. I understand detaining people that are charged with beating their wives, or raping people, or doing really horrible crimes, but shoplifting? What I hope my research would help to do is to spark a conversation about shifting particularly from setting low levels of bail to releasing people on their own recognizance. For matters of bail that are a few hundred dollars or a few thousand dollars, just let them go free, because otherwise you’re detaining people just because they can’t come up with a very small amount of money.

Lantigua-Williams: Anything in your analysis that you were not expecting?

Stevenson: I could answer that in two ways. One is, I didn’t expect the detention rates to be so high for what seemed to me to be pretty trivial crimes like shoplifting. That was surprising for me. I also didn’t expect the detention rates to be so high, for people that have relatively low amounts of bail set. The other thing that was maybe not shocking, but I found interesting, was that these effects are greater for people that have very limited prior experience with the criminal justice system—people who are first- or second-time arrestees. That may be because being in jail is particularly scary or intimidating, or they’re less savvy.