On a January morning in Jacksonville, Gary Smith brakes, parks and stares at where a house ought to be.
The Pulaski County sheriff's deputy is looking for a woman who missed a court date after a 2014 arrest.
Normally, Smith would approach the front door, rap his knuckles against wood and wait. But there is no front door, no trace of foundation, just a corner parcel of yellow grass.
"That was easy," Smith says before he tells his accompanying deputy the next address.
Smith, who works in the warrants division, will spend his shift searching for people who failed to appear in Pulaski County Circuit Court.
At initial hearings, when setting bail, district and circuit judges try to determine who won't skip court or re-offend if released before trial.
Right now, judges in Pulaski County often have little to go on: a police report, a glimpse of the defendant's personal history, and that's it. With scant details, they decide who should re-enter the community while waiting for their trials.
And when the defendant doesn't show, paperwork must be filed. Deputies then must track certain people down, further devouring county resources.
Some jurisdictions across the country have forged a different path.
Counties in Kentucky, Arizona and other states use a pretrial planning tool called the Public Safety Assessment. Developed by the Laura and John Arnold Foundation, the system uses data to determine who should be released before trial.
Unlike human judgment, the tool is not swayed by race, class or gender. The method thins jail populations and curtails failure-to-appear rates without a spike in crime, the foundation has said.
Now, Pulaski County could follow suit. A committee of decision-makers, convened by County Judge Barry Hyde, meets once a month to discuss criminal justice changes.
Among the members are Pulaski County Sheriff Doc Holladay, Prosecuting Attorney Larry Jegley and Little Rock Police Chief Kenton Buckner.
The tool's cost is still a question mark. But at the group's December meeting, members indicated that they saw a need for using the method.
"We all are aware that we have a high crime rate, and we have limited resources, and we have overcrowding at our jail. Those are all constants," said Chastity Scifres, chief deputy county attorney.
"So the only question is, how do we help all those things, while at the same time, running the judicial system more effectively?"
Pretrial planning, she said, could be the county's answer.
A COMMON CHARGE
Failing to appear in court is a common charge among Pulaski County jail book-ins, jail data show.
Over five years, 122,120 people were booked into the Roosevelt Road facility on about 405,540 charges.
About 46,015 of those charges were failure-to-appear warrants, issued for 19,520 people, according to data compiled by Maj. Matthew Briggs, who oversees jail operations.
Last year, of the roughly 21,240 people booked into jail, about 5,820 of them faced a failure-to-appear charge. That's 27 percent, or just over one in four.
About 1,580 people -- or more than 7 percent of people booked in -- were arrested on a failure-to-appear warrant alone.
Many people in that subgroup were taken to jail after traffic stops, or during other benign encounters with law enforcement officers, after a police officer discovered the outstanding warrant.
While some people dodge their court dates on purpose, others may be hospitalized, miss their bus or have a death in the family. Sometimes a person moves and doesn't know who to notify about his address change, so he never gets a court notice, Scifres said.
A lot of people just forget or misplace their citation slips, said Little Rock District Judge Mark Leverett, who sits on the coordinating committee.
"You get a piece of paper, you go home. If you aren't fairly organized, you could put it in a drawer and just forget about it," he said.
That's what happened to William Clinton, who missed an afternoon district court date in November.
Clinton said the 1:30 p.m. start time threw him off. He called that day to reschedule his hearing, but the new court notice never arrived in the mail, he said.
Eventually, a failure-to-appear warrant was activated. It took him two months to get back on the court docket and clear up the issue, he said.
Though the initial offense was just a traffic violation and Clinton wouldn't be kept in jail, knowing he had an active warrant troubled him. If he were pulled over, a police officer could take away his driver's license and impound his car, he said.
"The worst case scenario, you could be without a job. You don't have a car, you don't have money to get your car out of impound," Clinton said.
"It makes you nervous every time you're driving to work."
While a missed court date could result from a minor slip in judgment, the consequences are significant.
After a circuit judge issues warrants for people's arrests, the sheriff's office prioritizes the warrants based on the alleged crimes, said Pulaski County Circuit Judge Herb Wright, another committee member.
Sometimes the officials wait until a person comes in contact with law enforcement again, which can take weeks, months or years, he said.
Leverett, who oversees city code violations, said he attaches bail amounts to the failure-to-appear warrants he issues so the person can bond out of jail. And, in general, for lesser charges like misdemeanors, defendants are released and get new court dates.
But if the warrant comes from circuit court, meaning the defendant is accused of a felony, that person must be jailed at least until he sees a judge.
"If I let them go, they'll take me to jail," Briggs said at the December committee meeting.
'MANPOWER, PAPERWORK, TIME'
Failure-to-appear warrants rack up costs because they compound work for public servants.
"It's hard to put a number on that," Scifres said. "But we're talking about manpower, paperwork, time."
In 2017, the Pulaski Circuit/County clerk's office handled more than 3,100 documents related to failure-to-appear warrants, equivalent to eight hours of work a week, on average, Jason Kennedy estimated. He's assistant chief deputy of the clerk's office.
Those warrants are handed to Smith and two other deputies in the warrants division on the bottom floor of the county courthouse.
Smith said that on a slow day he logs a couple of hundred miles in his sport utility vehicle. On Tuesday, a slow day, he had eight warrants to work.
One woman Smith was searching for had died of a drug overdose, he learned. One man had been in jail since before Christmas.
"You get what you can do," Smith said. "Get as many as you can, as fast as you can."
In district court, judges often tell their clerks to put warrants into the Arkansas Crime Information Center database that police officers use while on the job.
After an arrest on a warrant, a law enforcement officer files the return of the warrant. That tells the judge to take action, said Pulaski County Circuit Judge Mary McGowan, who serves on the committee.
The process "sounds probably more cumbersome than it is," McGowan said. "But you can also see the ability to have error because of the number of times it's touched by different individuals."
Those warrants -- and occasional errors -- elongate dockets, pile up paperwork and add to the number of people sheriff's deputies have to transport, all of which costs the county money.
That's not to mention the redundancies that failing to appear for a trial can cause, deputy prosecutor and committee member Barbara Mariani said.
"It just puts everything at a standstill. You get all the way up to trial, once they fail to appear, you start [the case] over again," she said. "It's very costly."
Failing to appear in court also harms victims, Mariani added.
"They're in limbo until the person is found."
Plus, jail is pricey. To keep someone for 24 hours costs $63 on average. Booking in and holding 5,820 people with failure-to-appear warrants last year cost somewhere between $183,300 to $366,600, based on that average.
Nailing down a number is tricky because some people were jailed for only an hour, while others were kept for far longer, Briggs said.
A public safety assessment tool could cut administrative legwork and "increase public safety, safely," Scifres said.
The tool, when integrated with software, would use databases -- likely the Administrative Office of the Courts and the Arkansas Crime Information Center -- to create a score from 1-6 for a person who is arrested.
A score of 1 represents a "low risk" person, who is likely to appear in court and unlikely to commit a new crime. A score of 6 is the greatest risk.
The tool also flags a small number of people deemed likely to commit violent crimes if released.
Nine factors contribute to the score: age, the violence of the offense for which the person was arrested, pending charges, previous convictions on misdemeanors, felonies and violent offenses, failures to appear in the past two years, failures to appear more than two years ago and previous incarceration.
Nothing else -- not a person's race, sex, education, income, home address, drug use history or employment status -- are considered.
The score informs a judge's pretrial decisions, though the judge always has "absolute discretion," Scifres said.
Right now, Pulaski County's six circuit and five district judges who regularly hear criminal cases are presented with far less data.
At an initial hearing, in general, the prosecuting attorney and public defender each recommend bond amounts for the judge to consider.
But those attorneys might not know anything about the defendant's past, Wright said. During Saturday court, a prosecutor might have an arrest report, and that's it, Mariani said.
"I have to make a bond recommendation from one piece of paper from one jurisdiction," she said.
Plus, some records of a defendant's history, like previous incarcerations or pending charges, are scattered among dozens of agencies.
"We don't have a central location where everyone is looking at the same information," Mariani said.
And the public defender often relies on the defendant to be truthful about previous convictions, Leverett said. So that insight is "obviously biased," he said.
"I'm handicapped when I'm trying to make a decision," Leverett said, referring to his shift overseeing Saturday criminal court. "They could have 15 [failure-to-appear warrants], and no one knows. They could have a string of prior criminal convictions, and no one knows."
"We do the best that we can do, but we would be in a much better position with a tool like this," he added.
Jurisdictions that use the Arnold Foundation tool have documented successes, while also noting areas to improve.
Lucas County, Ohio, doubled the number of people released before trial, while decreasing the failure-to-appear rate and saving 1,000 jail bed days.
In Yakima County, Wash., which implemented the tool in 2016, defendants were able to post bond faster, with less racial disparity between those given and refused bail.
Regan Miller, the chief district judge in Mecklenburg County, N.C., which uses the tool, testified to its effectiveness on the foundation's website.
When Miller started on the bench, everyone charged with a certain crime faced a corresponding monetary bond, he said in a video.
"I came to realize that that's not fair," he said.
Miller added that he is a "strong believer that the more information you have, the better your decision-making is.
"In using the risk assessment tool, I feel much more comfortable that I'm making the right decision and engaging in what justice is supposed to be about. And that is, treating each individual who comes in front of me fairly," he said.
These programs aren't happening "in a vacuum," said McGowan, the Pulaski County circuit judge.
"These things are out there, they're collecting data, and if you look at the data, you see, 'Hey, this is probably a good idea,'" she said.
Along with the assessment tool, county officials are considering another layer of pretrial planning: a defendant notification system.
As it stands, defendants get court notices on paper, either handed or mailed to them.
Some places are going paperless, sending timely alerts through phone calls, emails or text messages.
These virtual reminders can decrease failure-to-appear rates and improve perceptions of justice system fairness, according to a brief from the Pretrial Justice Center for Courts.
"We're not looking to enable," Scifres emphasized. "But people receive alerts about things all the time, whether that be doctor appointments or dentist appointments."
"These people have been arrested. They haven't been convicted," she added.
There is no fix-all to keep everyone from missing court, in part because "if they're serial, they won't appear," said Leverett, the Little Rock district judge. "They'll just repopulate my docket."
Pulaski County jail data indicate that some people who miss court the most need other means of help.
Ten people alone are responsible for nearly 600 failure-to-appear warrants. One of those people is a homeless man who suffers from alcohol addiction and has exhibited signs of a mental illness, according to multiple Little Rock police reports.
A pattern unfolds through his arrests. The 42-year-old is charged with a misdemeanor, loitering, public intoxication or criminal trespassing, then arrested again and again for failing to appear in court.
He's been served with 114 failure-to-appear warrants, which means he's been transported and booked into jail 114 separate times on those charges alone.
And all those warrants create a patchwork of fines and jail time: 30 days here, a couple of hundred dollars there.
But unlike those who experience homelessness or habitually skip court, Pulaski County jail data also show that about 10,250 people were arrested only once on a failure-to-appear warrant in the past five years.
Those are the people that pretrial planning could help, Briggs told the committee in December.
They're people who might have slipped up, people like Clinton.
With a risk assessment tool, if Clinton were arrested on a criminal charge, his score would reflect that he usually makes his court appearances. And a phone or text notification would make missing a court date much more difficult, he said.
STEPS THAT REMAIN
Though many committee members have vocalized support for risk assessment, cost is still an issue. The tool is free to use, but automating it and integrating it into the county's existing software is not.
Relying on current employees to manually input all the information is unlikely.
"We don't have the manpower," Scifres said. "And we don't really have the budget to hire that manpower."
Automation costs depend on the number of users, which is still up in the air. Which department would oversee the risk assessment tool is another unknown.
The county also would need permission to access certain databases. How well the tool works depends on the quality of the information it can access.
And some judges might resist using the tool, though Scifres said that can be mitigated through education.
Still, counties across the country, including Sebastian County, are rallying for some form of pretrial planning. Act 423 of 2017, a state criminal justice reform law, strongly recommends that jurisdictions adopt such a tool.
Pulaski County could be an exemplar, Scifres said.
"It's a best practice," she said. "So there is no reason why we shouldn't go ahead and do it."
On Tuesday, the coordinating committee will meet in downtown Little Rock to further discuss the tool. For change to take hold, Leverett said, the right people are in the room.
"We've got shot callers."
SundayMonday on 01/14/2018