Inmates sit in a discussion group at the Stanislaus County Jail in Modesto, Calif. (Rich Pedroncelli/AP)

Most of the controversy over crime and punishment in the United States has focused on how many people are in prison. You don't hear as much about jails, and yet for most Americans the local jail is where they're likely to experience the justice system.

Far more Americans go to jail in a given year than to prison, although most of them have not been convicted of any crime. Then there are those with mental illnesses who simply don't have other options. And increasingly, jail has become a de facto punishment for poverty, as the poor are forced to remain there in lieu of bail while awaiting trial.

"What's going on with the jail population doesn't get a lot of attention relative to prison incarceration," said Jesse Jannetta, a researcher at the Urban Institute.

Convicts are sent to prison to serve their sentences, but people generally go to jail for short periods while awaiting a hearing or a trial. Alongside the explosion of the population in prison over the past two decades, the number of people spending time in jail has also increased drastically, according to a study published by the Vera Institute of Justice this week.

On any given day in 1983, about 96 people were in jails per 100,000 U.S. residents, according to the report, which compiled data from the Justice Department and other sources. That figure had more than doubled by 2007, to 259. It's decreased slightly since then to 231 in 2013.

By contrast, there were about 478 inmates in state and federal prisons per 100,000 residents in 2013, but because people spend much shorter periods in jail than they do in prison, the jail system reaches far more people. The nation's jails recorded some 11.7 million admissions in 2013, according to the report. (Individuals who went to jail more than once that year likely account for a large fraction of that total, but it's impossible to say exactly how much.) That's an increase from 6 million admissions 30 years previously.

While more and more Americans are going to jail, far fewer of them are committing crimes, as the chart below shows, and police are making fewer arrests. In view of the research showing that the decline in crime is a result of other factors than the increase in incarceration, you'd expect fewer crimes to result in fewer people in custody.

(Vera Institute of Justice)

Of course, jails do serve an important purpose with regard to public safety. They are a place where potentially dangerous people can be held while their cases are dealt with. Unfortunately, the report from the Vera Institute suggests that the increase in jail bookings has a variety of causes, none of which are closely related to keeping the public safe.

As many public mental hospitals closed down beginning four decades ago or so, law enforcement has wound up handling those who would have once been their sickest patients -- the people who don't seek out or remain in treatment, and who inevitably disrupt public order so police are constantly called on to restrain them.

At the same time, the average amount of time that any one person spends in jail has increased from two weeks to more than three, according to the report, suggesting that an overburdened system of courts may be taking longer to process cases. And police are increasingly making the decision to jail arrestees. It used to be that only 51 people were jailed for every 100 arrested, but now, there are nearly as many bookings as arrests. The causes of this increase are unclear.

Yet the most alarming reason that more people are spending time in jail could be that more judges are demanding bail from defendants awaiting trial.

Of every five defendants in felony cases who were released pending trial in 1990, three were let go on their own recognizance or on some other condition that did not involve bail. By 2009, the number had reversed, as three out of five were required to post bail, according to the report. As a result, a majority of those held before their trials in New York City jails in 2013 were there simply because they couldn't come up with $2,500 -- or less, in many cases.

"That's a function of poverty," said Nick Turner, the Vera Institute's director. "Your ticket out depends on your ability to post bail."

The practice might be defensible if these defendants were accused of white-collar fraud, in which case a financial sanction could well discourage them from fleeing. In fact, the courts' reliance on bail has the effect of jailing poor defendants, even if they committed minor crimes and are unlikely to attempt to flee justice, while allowing wealthier defendants to go regardless of the risk they pose to public safety.

And since even a few days in jail can mean a loss of a job for a low-wage worker, jails can have the effect of forcing the poor further into poverty, without contributing to public safety.

"It jeopardizes your employment. It can interfere with your schooling. It can affect your ability to obtain housing, and it has consequences for your family and your community," said Laurie Garduque of the MacArthur Foundation.

In conjunction with the report's release, the MacArthur Foundation announced $75 million in grants to encourage judges, district attorneys and other local authorities to think differently about jails.

Financial security does not predict whether a defendant will flee or offend again before the trial begins, but other factors do, such as defendants' employment statuses, criminal records and relationships with neighbors and family.

In Kentucky, a single agency uses this kind of information to quantitatively evaluate all defendants, and the state is now releasing seven in ten defendants pending trial without requiring bail from them. The report notes that only 8 percent of the defendants Kentucky released under this program were arrested again, compared to 16 percent of those nationally who were released on bail. That discrepancy is a reminder that the ability to post bail reveals more about how wealthy or poor defendants are than about whether they endanger public safety.

John Pfaff, a law professor at Fordham University, said that the equations that agencies such as the one in Kentucky use to evaluate defendants are becoming more reliable, as researchers and statisticians examine more data and refine their techniques. "We actually have increasingly good models of who poses a risk and who doesn't pose a risk," he said, adding that in jailing those who are unlikely to flee or commit additional crimes, "We're spending money that we don't have to spend. We're disrupting lives that we don't have to disrupt."

The Urban Institute's Jannetta has written that these models, if poorly designed, can wrongly give the impression that defendants of color pose a greater risk than white defendants. Blacks may be more likely to be arrested in general, regardless of how many crimes they actually commit. It's a concern that U.S. Attorney General Eric Holder has also expressed.

Yet Jannetta noted that the current system is probably more biased in any case. "The risk assessment tools and the outcomes that they would derive are probably better than subjective judgment, because ultimately that's what you end up using -- or, in the case of bond, it's who can come up with the money," he said.