On the night of November 2, 2015, Maria Soto’s 18-year-old son Daniel went out with friends and did not come home. At 1:30 p.m. the next day, Maria finally got a call: Daniel had been stabbed and was in the hospital—and was under arrest.
A man had accosted Daniel and his friends outside of a restaurant. They had fought, and the man pulled a knife. Cut and bleeding, Daniel staggered up to a police officer, who called an ambulance and arrested him. Apparently, the man with the knife had gotten to the officer first.
Once he arrived at the hospital, Daniel received minimal medical treatment—Advil for pain and occasional new dressings for his wound. On November 10, he was taken to court, where he pled “not guilty” to a felony assault charge. The judge set bail at $30,000.
Maria, a single mother who worked as a stenographer, made enough to pay rent and bills for herself and her two sons, but had no savings and no property to sell or use as collateral. No bail bondsmen would give her a payment plan she could afford.
Maria felt horrible, knowing her son was hurting, locked up in jail, and there was nothing she could do to help him. “It was terrible. He’s my son. I wasn’t eating. I wasn’t sleeping. I just worried about him.”
Meanwhile, Daniel also could not sleep, due to the pain from his injury and the hard jail bed. He was assigned a top bunk and struggled to climb up to it. Sometimes pus would ooze from his wound due to the exertion. He asked his mother to bail him out, but understood she could not come up with the money. “I just had to ride it out,” Daniel said. He missed school and slipped behind in his studies. On Thanksgiving, Maria and the rest of the family ate their meal without him.
Finally, on December 17, over six weeks after his arrest, Daniel had his preliminary hearing—the first opportunity in court for the judge to hear proof of the crime. The judge dismissed the case, saying there was no evidence he committed a crime. Daniel was able to go home, but he had lost a semester of school and a month-and-a-half of his life to jail for a crime he did not commit, all because his family did not have money to pay for his freedom.
Tens of thousands of people arrested for a wide range of crimes spend time locked up in jail because they do not post bail. Nearly every offense in California is bail-eligible, yet many defendants cannot afford to pay. In California, the majority of county jail prisoners have not been sentenced, but are serving time because they are unable to pay for pretrial release.
This report concludes that California’s system of pretrial detention keeps people in jail who are never found guilty of any crime. The state jails large numbers of people for hours and days against whom prosecutors never even file criminal charges. People accused of crimes but unable to afford bail give up their constitutional right to fight the charges because a plea will get them out of jail and back to work and their families. Judges and prosecutors use custody status as leverage to pressure guilty pleas. As one Californian who went into debt to pay fees on $325,000 bail for a loved one who was acquitted said, the actors in California’s bail system are “not in it for justice.”
Those locked up pretrial are overwhelmingly poor, working class, and from racial and ethnic minorities. California’s median bail rate is five times higher than that of the rest of the country. There is a clear correlation between the poverty rate and the unsentenced pretrial detention rate at the county level in California. The state is also plagued by profound racial disparities in pretrial detention rates due to racial disparities in arrest and booking rates. The rate at which black people are booked into California jails is many times higher than for white people—for example, it is nine times higher in San Francisco.
Bail and pretrial detention in California subject arrestees to unfair treatment, arbitrary detention, wealth discrimination, and other violations of their basic rights. People unable to pay bail remain in jail regardless of guilt or innocence. Poor and middle-income people incur debilitating debt to gain the advantages to fighting their cases that pretrial freedom bestows.
There is an alternative to California’s system of money bail and pretrial detention. Given the large numbers of people locked up in California despite never being charged with an offense, as well as the large numbers of low-level offenders who are jailed, the best reform would divert the great majority of defendants out of custody through extensive use of release with citations. The remainder would have detailed, individualized hearings before a court could order pretrial detention.
This alternative to money bail as the determinant for custody would reject the current trend of using profile-based statistical predications of risk instead of money bail as the basis for pretrial detention or supervision decisions. Instead, it would rely on detailed, individualized hearings to determine whether any pretrial defendant may be deprived of their liberty.
Wrongful Pretrial Detention
From 2011-2015, police in California made almost 1.5 million felony arrests. Of those, nearly one in three, close to half-a-million people, like Daniel Soto, were arrested and jailed, but never found to be guilty of any crime. Some spent hours or days behind bars. Some spent weeks; others, months and even years. The cost to taxpayers of this pretrial punishment is staggering: each day a person is held in custody costs an average of $114. In six California counties examined in detail in this report (Alameda, Fresno, Orange, Sacramento, San Bernardino, and San Francisco), the total cost of jailing people whom the prosecutor never charged or who had charges dropped or dismissed was $37.5 million over two years.
Over a quarter-of-a-million people sat in jail for as long as five days, accused of felonies for which evidence was so lacking prosecutors could not bring a case. Many were victims of baseless arrests; others, mistakes of judgment or misunderstandings of the law. The remainder had cases filed, but lacked sufficient proof of guilt, resulting in eventual dismissal or acquittal after weeks and months in jail. A large percentage of these not guilty people either had to pay bail, often plunging themselves or their families into crushing debt, or had to contest their cases while locked up in county jails.
These nearly half-a-million people spent time in jail at taxpayer’s expense, missing work, not picking their children up at school, not caring for elderly parents, missing classes, and subject to violence and miserable conditions, because they did not post bail. They were punished for crimes they did not commit, not because they were too dangerous to release, but because they could not come up with money to pay for their release, in cases where the criminal justice system ultimately found them not guilty.
Coerced Guilty Pleas
Many Californians accused of crimes, but unable to afford bail, give up their constitutional rights to fight the charges because a guilty plea will get them out of jail. Prosecutors often argue for high bail because a defendant is “too dangerous to let out” before trial, then offer the same “dangerous” person a time-served, go home sentence in exchange for a guilty plea. Some judges set bail a defendant cannot possibly pay, to encourage guilty pleas for the sake of rapid processing of cases.
Pretrial detention causes higher conviction rates mainly by coercing people to plead guilty in order to get out of jail sooner. In the six counties analyzed from 2014-2015, 71-91 percent of misdemeanor and 77-91 percent of felony defendants who stayed in jail until they received their sentence were released before the earliest possible trial date. They all pled out before they had a chance to assert their innocence. Pretrial detention allows courts to process cases more quickly, but distorts justice by coercing guilty pleas.
A Discriminatory System
California’s system of money bail and pretrial detention discriminates based on wealth. Rich people simply pay bail and buy their freedom. People of more modest means sometimes can cobble together the money to pay a bondsman the 8-10 percent non-refundable fees normally charged to secure their release. In the six counties examined in detail, 70-80 percent of arrestees could not, or did not, pay bail. Those who did not pay were either eventually released from jail in other ways, such as on their own recognizance or by court orders, or stayed in jail until they were sentenced. People at liberty can help with their defense; they can go to work, go to school, attend a drug rehabilitation program or enroll in psychological counselling, all of which can show the judge there is no need to punish harshly; they appear in court showered and groomed, in their own clothes, not jail uniforms.
People who cannot afford bail have none of these advantages. They have barriers communicating with their lawyers; cannot help locate witnesses and evidence; cannot participate in programs to improve themselves and make themselves look better in the court’s eyes; and cannot earn money. They sit in jail, surrounded by misery, feeling stress about the case, unable to get calm advice from family and friends. They cannot sleep well. They will look like criminals when they appear in court, shackled or behind a glass partition. Many judges are likely to see them as just another defendant to process.
The case of Daria Morrison and Sarah Jackson illustrates the income-based discrimination in California’s money bail system. Both women were arrested together and charged with a robbery; neither had a prior criminal record. Yet their fates were very different. Daria had sufficient help to pay the bondsman’s fee, was released from custody, and offered a reduced charge that will result in a dismissal in one year by the prosecutor. Her co-defendant, Sarah, equally culpable for the crime, remained behind bars, unable to pay for bail. She ended up pleading guilty to two serious felony charges.
The bail system is also racially discriminatory. Though violent crime has dropped steadily since the early 1990s, California continues to put people in jails and prisons in massive numbers. On a single day in 2015, California had 201,000 people behind bars, with 1.15 million arrests throughout the year, causing many thousands more to cycle through the jails during the year. This high rate of incarceration disproportionately affects black people, who are over 6.5 times as likely as white people to be locked up. Data analyzed by Human Rights Watch from a variety of California counties shows jail booking rates for black people are significantly higher than for white and Latino people.
In this time of increasing incarceration, the use of pretrial detention has also increased dramatically. In California, consistently over 63 percent of prisoners in county jails have not been sentenced, but are serving time because they cannot afford to pay bail. Studies have calculated California’s median bail as being five times greater than that for the rest of the country.
California law does not require a judge to inquire into a defendant’s ability to pay, and judges rarely do when setting bail amounts. Instead, they rely on arbitrarily determined bail schedules that set amounts to coincide with the level of the charge. While judges have discretion to depart from them, they tend to treat the schedules as mechanical formulas to apply in most cases. Experts and advocates―and even some judges―told Human Rights Watch that bail in California is set to keep people in jail, coerce guilty pleas, and make court machinery move more rapidly.
Most defendants rely on bail bondsmen to get out of detention. Bondsmen charge a fee of up to 10 percent of the actual bail amount, which is not refundable, even if the case is dismissed or charges are not filed. Bondsmen charge as much down-payment as they can, sometimes the full amount of the fee, or work out payment plans that they enforce with the threat of revoking the bond and sending the accused back to jail.
This system often means that poor and middle-income families must borrow from friends and family, raid retirement plans, cut back on food, bills, and holiday presents, miss rent payments, and sell personal property to pay for their loved one’s freedom.
While the numbers are staggering, the true measure of the harm caused by California’s system of money bail is in the stories of the people who have been through this system:
Bail: An Ineffective Tool
The stated purposes of setting bail are to protect public safety by preventing potentially dangerous people from causing harm before their cases are adjudicated and to prevent people from fleeing the jurisdiction or otherwise evading their obligation to go to court.
But bail is not a particularly effective tool to meet these goals. Lack of in-depth, individualized hearings means judges do not have sufficient knowledge to assess risks with accuracy, defaulting to bail schedules and overusing detention. Vast numbers of people are jailed pretrial due to “dangerousness,” while only a tiny percentage actually commit violent crimes while awaiting trial. People with money pay for release regardless of how dangerous they are.
Few people actively evade court. Most who fail to appear do so due to negligence or error, homelessness or mental disabilities, or because they cannot miss work or find child care. Many who miss appearances eventually return to court on their own. Imposing bail improves court appearance rates in moderate amounts, but detains many more people than is necessary. Other pretrial services, like reminder calls, are proven to reduce missed court dates without incurring the costs of locking people in jail.
International human rights law permits the use of pretrial detention and money bail, but only if they are limited and are consistent with the right to liberty, the presumption of innocence, and the right to equality under the law. A person’s liberty may not be curtailed through arbitrary laws or the arbitrary enforcement of law in a given case. International human rights law condemns discrimination based on race, ethnicity, gender, and wealth. Decisions about pretrial detention must be grounded in reasoning that contains specific individualized facts and circumstances, and not by reference to simple formulas, patterns, or stereotypes.
Profile-Based Risk Assessment Tools
Many who seek to reform California’s system of money bail and pretrial detention are turning to profile-based risk assessment tools. These take information about the accused, compare it to known behaviors of other people with similar characteristics, and generate a prediction about risk of future criminal conduct or missed court appearances. The predictions are statistical estimates based on a profile.
On the surface, these tools claim to avoid human biases and facilitate release of more people from pretrial detention, while promising rapid decision-making.
But these tools risk being a sophisticated form of racial profiling that produce biased outcomes because they ask questions implicating race, and because the underlying information evaluated, based on policing and law enforcement, reflects a system that is itself riddled with racial bias. If arrest and conviction data is racially biased, the tools that use this data to make decisions about who stays in jail and who gets released will generate racially biased outcomes.
The tools provide only statistical predictions based on non-contextual information and do not allow for explanation of prior criminal history. For example, a person who missed a court date because their return slip had the wrong day but came to court two days later would get the same negative score for failing to appear as someone who fled the country to avoid court. The profiles may miss specific, serious threats that do not appear on the surface of the criminal history, as someone with a minimal criminal record may represent an extreme danger in the given circumstances.
Despite the veneer of objectivity, the risk scores are subjectively defined and can be manipulated to direct fewer or greater numbers of people into custody or under supervision, depending on the needs of those administering the tools. For example, in Santa Cruz County, the tool was adjusted to double the number of people released under conditions of supervision.
While jail overcrowding provides incentive to use the tools to reduce pretrial detention, given the massive amount of jail construction going on in California, the tools may be used to increase detention in the future. A risk assessment tool can put people under increased levels of supervision or fill jails as easily as it can facilitate release.
Reform Requires Individualized Procedures
Instead of profiling and risk assessment by statistical prediction, or jailing people based on their wealth, California should adopt a system that favors release and assesses the risk of danger in an individualized, contextual way. As a default rule, only those accused of serious felonies should merit consideration for pretrial detention in the first place. The rest, with narrow exceptions, should be released from custody at the arrest stage and issued a citation requiring them to appear in court on a particular date. Cite and release would vastly reduce the number of people jailed without having charges filed against them.
The few who do stay in custody should have a full adversarial hearing, with an enforceable legal presumption of release absent proof by the prosecutor of a specific need to detain. Defendants should have capable legal representation when they get to court. The hearing should include testimony about the actual crime, so the judge can evaluate its seriousness and the likelihood of eventual conviction, an ability to pay hearing, and an opportunity to present individualized evidence favoring release or detention based on specific risk of pretrial harm.
This proposed system would involve significant changes in California courts’ approach to administering justice, and would be challenging to implement. But the advantages are essential. These changes would:
- Prioritize public safety by causing courts and prosecutors to focus on those defendants who truly pose a danger, while releasing those who do not.
- Decrease the harm suffered by families when their loved ones are jailed, and limit financial burdens placed on poor people who pay for their freedom.
- Mitigate the income-based discrimination of the current money bail system.
- Decrease the number of people, particularly innocent people, coerced into pleading guilty because of their custody status.
- Save the public money by cutting jail costs.
- Honor the presumption of innocence and treat people in court as human beings, not numbers.
Above all, it would increase the quality of justice in California.
- Expand legal requirements for law enforcement to cite and release without arrest to include all misdemeanor and all non-violent/non-serious felony suspects, with narrow exceptions, thus limiting the number of people placed in pretrial custody at all.
- Establish enforceable standards for setting bail or detaining pretrial, requiring release absent significant proof of a specific danger to the community or specific risk of evasion of court process.
- Establish procedures for meaningful hearings on pretrial detention and bail setting, including a testimonial probable cause determination and an ability to pay hearing, as well as opportunity to present mitigating and aggravating factors, while providing sufficient resources for appointed counsel to research, investigate, and conduct these hearings.
- Reject the use of statistical predictions of the likelihood of pretrial misconduct as a basis for or factor in setting bail or pretrial detention.
This report is based on research conducted from September 2015-January 2017. Findings are based on 151 interviews.
Eighty-six interviews were with criminal justice professionals, including judges, district attorneys and other prosecutors, defense lawyers, including public defenders, probation officers and administrators, pretrial services personnel, academic experts, court administrators, policy analysts, law enforcement personnel, and court administration consultants.
Sixty-seven interviews were with people who had direct personal experience with pretrial detention in California as arrestees, prisoners, or immediate family members or partners of an arrestee or prisoner.
We also spoke to 21 attorneys and investigators who described the experiences of specific clients, and community organizers who work with people involved in the criminal system and their families.
The interviews in total cover experiences in 14 counties in the state. Just over 50 percent of the interviews of people with personal experience of being detained involved cases from Los Angeles County, as it is by far the county with the largest jail and court system. Berkeley Law students conducted 30 of the interviews contained in this report.
Of those who personally faced imprisonment pretrial whose stories we heard either directly or from a family member or an attorney, fifty-five were male and ten were female. Thirty-two were Latino, twenty-four were black, and nine were white. Some had significant criminal records; others did not. Some were convicted of some crime following their detention; many others were not.
Human Rights Watch identified people who had experiences with the pretrial detention system via several sources, including criminal defense attorneys who referred us to their former and current clients, and community organizations that work with people who have contact with the criminal system. Researchers spent time in court observing proceedings, including bail and detention hearings, and spoke to people they met in court.
Interviews were semi-structured and covered a range of topics, including description of the trajectory of the criminal case, efforts to pay bail, impact of detention on the individual and the family, impact of paying bail on the individual and the family, conditions of custody, and impact of custody status on the ability to contest the charges.
The interviews sought to determine if the pretrial detention system caused financial, physical, psychological, and/or penal harm. To the greatest extent possible, researchers reviewed court and attorney files, other court records, jail records, news accounts, and other independent sources of information to verify the case descriptions. Supporting documents are on file at Human Rights Watch.
Human Rights Watch uses pseudonyms for the individuals interviewed and their family members to respect their privacy, minimize the impact of revealing an encounter with the criminal system, including arrest or conviction, and to protect those who are vulnerable. Some of the people we spoke to are in jail or prison, on probation, or live on the streets where they may be subject to retaliation for speaking out about an injustice within the system. We have also disguised the names of lawyers who spoke about their clients to keep their clients’ identities hidden, and of criminal system professionals requesting anonymity so they could be more forthright in discussing the system and the actions or perspectives of colleagues and superiors.
All documents cited are publicly available or are on file with Human Rights Watch.
The Policy Advocacy Clinic at U.C. Berkeley School of Law provided outstanding assistance to Human Rights Watch on this report. Working under the supervision of Clinic Director Jeff Selbin and Teaching Fellow Stephanie Campos-Bui, law students Danica Rodarmel, Da Hae Kim and Mel Gonzalez prepared a background research memo about money bail and pretrial detention in California, nationally and internationally. The students compiled a list of suggested experts and other stakeholders in the California bail system, including judges, prosecutors, defense attorneys, law enforcement and non-profit organizations. After training from Human Rights Watch, the students conducted 30 of the interviews contained in this report.
To conduct data analysis for this report, Human Rights Watch requested data covering everyone booked into jail in 2014 and 2015 from every county in California. The responses from counties varied greatly as did the quality of the data provided. Many counties were unable to provide data at all, especially the smaller ones. In total, twenty counties throughout the state provided some sort of data. Different counties kept track of different things, and tracked similar things differently. For example, some counties carefully tracked bail amounts, while others did not. Some counties changed bail amounts to zero when the prisoner posted bond. For inclusion in the analyses, a county must have included data indicating whether there was a no bail hold flagged for each detainee. Otherwise, it is impossible to determine whether a detainee likely had bail set.
Each county provided descriptive “booking type” and “release reason” categorical variables using unique codes. Each county coded bookings and releases differently and no county could provide a manual detailing how specific types of bookings or releases should be coded by staff. Human Rights Watch recoded all booking and release types into new, coherent categories to our best ability, informed by conversations with sheriff’s department staff. Booking types typically fell into categories such as street arrests, en route bookings (bookings coming from or held for other jurisdictions), warrant bookings, parole or probation violations, or re-arrests. For each analysis in the report, notes indicate which types of bookings were included.
Counties provided information about all initial booking charges, and for some counties, conviction charges, per person. For counties that provided additional post-booking charges, only the initial booking charges were used. Offenses were coded as infractions, misdemeanors, non-serious felonies, and serious felonies, as defined in California Penal Code section 1192.7(c). The most serious crime for each person was identified by first ranking the charges by level of crime and then selecting the first crime listed in the database under the highest ranked level of crime.
Our analysis is limited by the data provided by counties, and therefore presents Human Rights Watch’s best estimates for describing jail bookings, bail, and releases in the counties included in the report. Those counties were selected because they provided data that contained enough variables and seemingly accurate data to provide estimates for specific research questions.
In addition to the county-level jail booking data, Human Rights Watch analyzed data from county bail schedules, the California Board of State and Community Corrections, the California Department of Justice, and the Bureau of Justice Statistics’ State Court Processing Statistics.
Pretrial detention, as with all aspects of the criminal system, is highly localized, with differences from county to county, courthouse to courthouse, and courtroom to courtroom. Surveying the practices of each of California’s 58 counties and of the hundreds of individual courtrooms is beyond the scope of this report.
Pretrial Detention in the Context of Over-Incarceration
Pretrial detention in California, and throughout the country, is a significant part of a criminal system that incarcerates too many people, including people innocent of any crime; discriminates against racial minorities and poor people; and imprisons people for too long.
At the end of 2015:
- There were approximately 2,173,800 people in prisons and local jails throughout the United States.
- The national rate of incarceration was 870 per 100,000 adults.
- 6,741,000 adults were under correctional supervision, including parole and probation, a rate of 1 in every 37.
These figures make the US the world leader in imprisonment, significantly outstripping overtly authoritarian countries like China, Russia, and Iran. California had the second highest total number of prisoners in the country, behind only Texas, with 550,600 people under correctional supervision, including 201,000 in jail or prison.
Rates of imprisonment increased dramatically from the late 1970s until just a few years ago, though violent crime rates have fallen steadily since their peak in 1992, from 1,055.3 per 100,000 to 426.4 per 100,000 in 2015.
The racial and economic class dimensions are inescapable. The incarceration rate for white people, based on 2010 census data, is 450 per 100,000; 831 per 100,000 for Latino people; and 2,306 per 100,000 for black people. The same study revealed a rate of 3,036 per 100,000 for black people in California, compared with 453 per 100,000 for white people.
Nationally, prisoners overwhelmingly come from the poorest economic class. One study showed the median pre-incarceration income for all male prisoners was 52 percent less than the median income of non-incarcerated men. The rate for incarcerated women was 42 percent less.
As rates of imprisonment have increased dramatically, so too has the practice of pretrial detention. Nationally, from 1990 to 2009, the use of money bail increased from 37-61 percent. During this time, the percentage of people detained pretrial grew considerably.
Pretrial Detention in California
California counties detain pretrial at a far higher rate than the rest of the country. In recent years, around 63 percent of prisoners in California jails have not been convicted or pled guilty.
As with nearly all aspects of the criminal system, these figures are subject to local variations among counties. Inyo, for example has a pretrial detention rate of just less than 40 percent, while Siskiyou’s rate is 87 percent. Of the larger counties, Los Angeles and Sacramento’s rates are just over 50 percent; Alameda, San Bernardino, and San Francisco’s are over 75 percent; and Riverside and Santa Clara’s rates are in the high 60s-low 70s percent range.
The total numbers of people detained pretrial in California at any given point in time varies, ranging between 52,000 and 42,000 from January 2014-January 2016. Jails range from having a small number of pretrial prisoners, to housing thousands.
Human Rights Watch analysis of data from six California counties (see Table 1, below) finds that just 20-30 percent of bail eligible prisoners ended up posting bond. The failure to post bond comes at a cost to California’s taxpayers: In Sacramento County, the cost of detaining people who were bail-eligible but who did not pay bail was over $44.3 million
There were wide differences between counties in how prisoners who did not post bond were ultimately released from jail:
- People jailed in San Francisco County were more regularly released on their own recognizance or under other pretrial release programs;
- Orange County released very few people on their own recognizance or under other pretrial release programs;
- In Alameda County, nearly 40 percent of people booked into jail remained in custody until dismissal or were released with no charges filed;
- In San Bernardino County, one of every three people booked into jail was released due to court orders to reduce the jail population;
- In Orange and Sacramento Counties, higher percentages stayed in custody until their sentences were complete.
In each county analyzed, black people were booked into jails at a much higher rate than white people. In San Francisco County, the ratio was nine to one, when controlling for population size. Because of the high rates of black people booked into custody, the problems of the bail system have a disproportionate impact and contribute to racial bias in the overall criminal system.
II. Pretrial Detention in California
Pretrial Detention Process
Different authorities use their discretion, guided by certain rules, to make crucial pretrial custody decisions at a series of distinct stages in California’s criminal justice system.
The police officer in the field decides whether to arrest or simply issue a citation; the supervisor at the station decides whether or not to set a bail; the prosecutor decides to file, reject, or delay the case; the prosecutor in court decides to request bail or agree to own recognizance release; and the judge decides what amount of bail to set. Additionally, the accused is sometimes able to make a decision whether or not to pay the bail―depending on wealth, family and community support, and willingness to make other financial sacrifices. Finally, the bail bondsman decides whether or not to offer terms that the accused and their family or supporters can meet.
Step One: Police Deployment and Enforcement Choices
One set of crucial decisions made long before anyone is arrested relates to police deployment. Police departments have limited resources and make choices about where to concentrate patrols and what enforcement priorities to emphasize. These choices, in an aggregate sense, determine who gets arrested and with what frequency.
Nathan Ramos lived in an encampment of homeless people in the Skid Row section of downtown Los Angeles. Because the Los Angeles Police Department (LAPD) had chosen to deploy large numbers of officers to the area to enforce “quality of life” crimes, like sleeping on the sidewalk, Ramos had frequent contacts with officers.
In early 2012, officers arrested him for having his tent on the sidewalk. They took him to the Central Station lock-up and booked him, ignoring his requests for medical attention, and placed him in a holding cell with just a concrete bench, a sink, and a toilet, for over twelve hours. He received no food while there. Eventually they moved him to the main LAPD jail where they put him in a cell that felt “like an ice box.” After two days in custody, police moved Ramos to the lock-up at the 77th Street Station, and released him a day later. The prosecutor never filed charges against him.
Step Two: Police Decide to Arrest or Release
A police officer with probable cause to believe someone has committed a crime, through observation or witness report, has authority to arrest. Police may arrest at the request of a private person, and may also use their discretion to issue a warning for certain violations.
For misdemeanor violations, California Penal Code section 853.6 requires police to issue a citation, with a signed promise to appear in court, and release the person without arrest. This rule exempts certain stalking, domestic violence, and restraining order violations. However, the law also allows a series of general exceptions that give police officers nearly unlimited discretion to arrest instead of release.
The exceptions include permitting arrest if: “prosecution of the offense … would be jeopardized by immediate release of the person arrested”; “there was a reasonable likelihood that the offense or offenses would continue or resume, or that the safety of persons or property would be imminently endangered by release”; or “there is reason to believe that the person would not appear at the time and place specified in the notice.”
While these provisions sound appropriate, they are vague, set no standard or oversight for the reasonableness of the officer’s determination, and are open to interpretation. In practice, officers can always articulate some reason to believe the crime will resume or the suspect is dangerous or will not appear in court. In practice, Penal Code section 853.6 barely constrains officers from arresting people on misdemeanor charges, instead of citing them with a signed promise to appear.
If the officer decides to cite and release, the suspect signs a “promise to appear,” and receives a ticket with the court date, time and location, and the nature of the charges. The person receiving the ticket must appear in court to face the charges, or the judge will issue a “bench warrant,” authorizing subsequent arrest. If an officer detains someone and determines they have an outstanding warrant, the officer retains the discretion to arrest, issue a separate citation to appear on the warrant, or simply give the person a warning.
People who are cited usually remain out of custody throughout the pretrial period, while those who are arrested and remain in custody have a much greater chance of having a bail set. The initial decision to make the arrest instead of cite and release can have profound consequences for those arrested and their families.
Michelle Roberts’ boyfriend was arrested for driving under the influence of alcohol. The officer took him to the station for a breathalyzer test, where he blew .081, just over the legal limit. Instead of giving him a citation and allowing him to call Michelle or a cab for a ride home, the officer chose to book him into the Santa Rosa City Jail. Given his low blood alcohol concentration, the officer could not justify refusing release based on intoxication. It does not appear that any of the other exemptions in Penal Code section 853.6 reasonably should have applied. Still, he remained in jail. Michelle had to contact a bondsman and pay a $500 non-refundable premium to get her boyfriend released. He vowed to pay her back, but had financial troubles. The debt, Michelle said, was a “weight” on the relationship, which ended soon afterward. In this case, the officer had a legal reason to cite and release, but chose not to, although other officers may have used their discretion differently.
There is no presumption in favor of citation and release in felony offenses. Police must arrest all felony suspects, whether or not they are dangerous or likely to go to court.
Step Three: Police Station Officers Decision to Set Bail
When police arrest a suspect, they put him or her through the booking process at the station, including taking photographs and fingerprints, checking for outstanding warrants, and filling out various forms. The arresting officer prepares a report describing the offense, including any evidence, witness statements and contact information, and statements by the accused. The officer’s supervisor must review the report and determine whether it describes conduct amounting to a crime and what the crime is.
If the supervisor determines there is no crime or that further investigation is needed to come to a conclusion, they release the arrestee. If the supervisor determines there is a crime, they have discretion to release after booking or to detain until the first court appearance. Each county has its own policies governing jail releases. For misdemeanors, Penal Code section 853.6 sets a presumption in favor of release; however, as with the officer in the field, the in-station supervisor has wide discretion to keep the accused in custody.
Penal Code section 1269b(a) authorizes the officer in charge of the jail to set an initial bail amount for an arrestee held in the jail immediately after booking. The officer sets bail according to the county’s bail schedule, which has a standardized amount based on the charge. The arrestee may then post the bail by depositing money at the jail.
Some counties have judicial officers on duty who will review each arrest and decide whether to order the arrestee released on a promise to appear or to set bail. In Santa Clara County, for example, a magistrate automatically assesses the arrestee’s suitability for own recognizance release.
In Los Angeles County, a government official told Human Rights Watch that a bench officer is assigned to review requests for own recognizance release pre-arraignment. Prisoners call a division within the probation department to request release, which provides a brief evaluation of the prisoner for the on-duty judge. Only one judge reviews applications at any given time. According to the official, the duty judges are generally inexperienced and have little information on which to base decisions, are risk adverse, and do not hear from any advocates in this process.
Step Four: Prosecutor’s Decision to File Criminal Charges
After an arrest, the police officer submits their report to the prosecutor for filing consideration. The prosecutor reviews the report and may reject the case outright, file a different or reduced charge, file the charge recommended by the police, or request further investigation. If the accused is in custody, the prosecutor has 48 hours to file the case from the time of arrest, excluding weekends and holidays. People arrested on Thursdays and Fridays usually spend the weekend in jail before seeing a judge.
Often people will sit in custody, only to be released with no filing. In May 2011, police arrested Jose Alvarez after he participated in a protest at Los Angeles City Hall and accused him of a felony. They booked him at the station and set a bail he could not afford. He did not have money to pay for his release and sat in a police station cell from Friday afternoon until the following Tuesday morning, when they took him to court. Alvarez sat in a crowded holding cell all morning before the deputy district attorney notified his lawyer they were not filing charges. It took them until late evening to process his release and let him go.
Step Five: Setting Bail in Court
If the prosecutor does file charges, the arrested person must appear in court. At the first court appearance, called the arraignment, the accused is assigned an appointed lawyer if they do not hire their own; receives the complaint, which details the charges; and receives the crime report and a copy of their rap sheet. The accused enters a plea, generally after consulting their lawyer and sometimes after evaluation of a settlement offer.
After the accused enters a “not guilty” plea, the judge addresses pretrial detention. If the accused seeks an own recognizance release, their attorney will make the request. If the prosecutor wants a bail set, they will ask the judge to do so. Often the arresting officer or prosecutor will fill out a bail request attached to the complaint submitted to the court.
After the request for bail or own recognizance release, the judge conducts a hearing and decides. The judge may release the accused, with or without conditions, or set a bail. Release conditions that a judge may impose due to concern for public safety or to ensure appearance in court may include requirements to “stay away” from a person or location, attend Alcoholics Anonymous meetings, submit to house arrest, or electronic monitoring. If an individual fails to adhere to these release conditions, a warrant will be issued for their arrest.
Usually, if a person appears in court in response to a citation or a summons, the judge will continue the own recognizance release. The judge may also set bail, usually in accordance with the set bail schedule.
Step Six: Obtaining Bail
If the judge sets bail, the prisoner must decide whether to pay for their release. For many, the decision is simple—the bail is too high. For a homeless person living on General Relief in Los Angeles County, even a bail of a couple hundred dollars is out of reach.
A person who can afford to pay full bail deposits it with the court clerk or law enforcement and immediately secures the prisoner’s release. Assuming the accused returns to court and does not miss future appearances, the person who put up the money will get it all back once the case is resolved and the bond exonerated. However, few people pay the full amount.
Those who cannot pay the full amount may use a bondsman, who charges up to a 10 percent fee and puts up a bond promising to pay the full bail amount if the defendant does not appear in court. The fee is not refundable, regardless of the case’s outcome.
Katherine Gibson’s Case
Katherine Gibson had some drinks with friends after work one Sunday afternoon, then had a minor traffic collision while driving home. Police arrested her, took her to the station, took a blood sample, and booked her into custody.
Katherine had never been in trouble with the law. In her mid-twenties, she had recently moved to Los Angeles from a mid-Western town and set up a small business caring for and walking dogs. But she had begun to have health problems, including a wrist injury and a cancer diagnosis, and she took medications for anxiety.
At the station, handcuffed to her seat, Katherine heard officers making crass comments about another female arrestee. Eventually, they put her in a filthy holding cell, where she sat for several hours on a concrete bench. She was then moved to another cell with bunks and an exposed toilet, which flooded during the night. Terrified of the police and her fellow prisoners, in pain and missing essential medical treatment, Katherine had an anxiety attack, hyperventilating and yelling for help.
Usually, a first time driving under the influence charge results in release from custody after no more than a few hours to get sober and a citation to appear in court. Even a guilty plea for a first offense driving under the influence almost invariably involves probation and a fine, but no jail time. But the police would not release Katherine. They set a bail of $100,000.
At the first opportunity, Katherine called her father for help. He called bondsmen, who offered to post the bail in return for a non-refundable 10 percent fee. Her father did not have the money, but was able to borrow it from a relative. At 9 p.m. Monday night, police released her with an order to appear in court to answer to felony driving under the influence charges.
Katherine had no prior convictions, and there was no evidence anyone was injured badly enough to merit the more serious charge. But the arresting officer had characterized the violation as a felony in his report, so station officers assigned the felony bail level. Had the officer not called it a felony, Katherine would likely have been cited out on her own recognizance, or would have had to pay a $2,000 fee to the bondsman, not $10,000.
“I know I messed up. I know there should be consequences,” Katherine said. But she feels she was set up to fail. The experience has left her discouraged: “I can completely understand why people can’t get out of the system.”
On her court date, Katherine learned the district attorney did not file the felony and that she faced a misdemeanor charge. Despite the reduced charge, Katherine could not get her money back from the bondsman. She wanted to fight the case, but did not have enough money to pay her lawyer to go to trial. So she pled guilty for probation, a fine, community service, and classes. She now cannot afford car insurance, limiting her ability to work, and struggles to pay rent.
How Judges Set Bail
Fixing bail is a serious exercise of judicial discretion that is often done in haste … without the full inquiry and consideration which the matter deserves.
—Stack v. Boyle, 342 U.S. 1, 11 (1951) (J. Jackson, concurring opinion)
Hearings to decide pretrial release status and to set bail amounts in California are generally extremely fast and often involve minimal argument. Judges have imprecise guidelines to direct their discretion, and almost no meaningful oversight. A defendant is entitled to review the bail order within five days, but the practical likelihood of changing the original judge’s decision is very small. The original judge who set bail at arraignment sometimes conducts the review.
In setting bail or granting release, the judge engages in an assessment of risk—primarily related to community safety. They also assess the probability of the defendant not appearing in court. In doing so, the judge considers the seriousness of the charged offense, the defendant’s prior criminal history, and prior missed court dates. The judge may consider mitigating factors about the defendant, including work and schooling, ties to the community, and other factors that counsel may present to the court.
The judge evaluates the seriousness of the offense based on reading the police report; there is no evidentiary hearing with live testimony about what really happened. Counsel may, but often does not, have the time or resources to present additional argument, based on statements, declarations, letters, documents, and representations.
A public defender who handles a high volume of arraignments and/or bail hearings in one Southern California court described having a short time to talk to the prisoner, review the facts of the case, get some mitigating information about employment and community ties, make calls to verify the information, then argue for release or low bail in court. He said that if he had a paralegal or investigator or more attorney assistance at this stage of the case, he would have more success securing release for his clients.
Common court practice is not to put great effort into bail hearings. In Alameda County courts, there is often no attorney appointed for the initial bail hearing. One Los Angeles County Superior Court judge has criticized public defenders for not fighting to get their clients out of jail at arraignment. According to retired San Diego County Judge Lisa Foster:
To be perfectly honest, I didn’t think much about bail, and to the best of my recollection, neither did anyone else—not my colleagues on the bench, not the prosecutors, not the public defenders.
Another Los Angeles County judge observed lawyers do not strenuously litigate bail, and that high bail is a part of court culture. One reason defense lawyers cite for not fighting bail hearings more strenuously, in addition to lacking time and resources to make effective presentations, is that judges tend to avoid making individualized decisions by automatically applying the bail schedule amount based on the charge.
California Bail Schedules Mean High Bail
Each California county sets its own bail schedule according to its own procedures. Usually, the judges meet annually to prepare, adapt, and revise a uniform schedule for all crimes. The law gives no guidance beyond commanding them to consider the seriousness of the charge. One judge from Contra Costa County acknowledged that judges did not base their bail schedule decisions on actual data. The public defender from Contra Costa County, who sends a representative to the judges’ meeting to set the schedules, said bail amounts had “no correlation to public safety or the risk of failure to return to court. They appear to be pulled out of thin air.” A Central California judge who was on his county’s bail schedule committee described receiving a circulated copy of the schedule, reviewing it for a few minutes, then voting to approve it.
Bail schedules vary drastically from county to county, without apparent correlation to crime rates, income levels, or even regional preferences. Though the bail levels may differ by county, overall, they are extremely high. The median bail amount in California ($50,000) is over five times that of the rest of the country. Overall bail levels increased in California by an average of 22 percent from 2003-2013, though some individual counties have reduced their bail levels.
The stated purpose of the bail schedule is to provide a bail amount for law enforcement officers to set after booking an arrestee and determining not to release that person with a citation. The judge is supposed to make an individualized decision about the amount once the defendant comes to court, and only needs to justify departing from the schedule if the offense is a “serious” or “violent” felony or for certain other specified offenses.
However, despite the high levels of bail proscribed by the schedules and the lack of careful planning in creating those schedules, judges across the state tend to use them reflexively instead of making an individualized decision.
Contra Costa County Chief Public Defender Robin Lipetzky told the Little Hoover Commission Regarding Bail Reform and Pretrial Detention:
Unfortunately, what I have seen in Contra Costa is that judges are loath to deviate from the bail schedule regardless of circumstances of the individual charged. In essence, the preset bail schedule has become a presumptive bail for each and every defendant. Blind adherence to a bail schedule has become the default; it is expedient, it requires no independent thought, and it provides easy cover for judges….
Judge Eskin of Santa Barbara County echoed Lipetzky’s assessment, saying that judges set bail on schedule because it is easy and expedient, as they only have a few minutes per case, and using the schedule facilitates getting through the calendar.
The American Bar Association condemns the use of bail schedules, calling them “arbitrary and inflexible” and warns they “inevitably lead to detention of people who pose little danger of re-offending or not appearing in court, while facilitating the release of wealthy dangerous people.”
Many judges prefer the bright line rules that the bail schedules provide. Some are concerned they will be blamed if they release someone from custody with a low bail, and that person commits a future crime; many prefer defendants to be in custody. Using the bail schedule allows judges a quick method of setting bail levels high enough to keep most people in custody without appearing to be especially harsh.
Judges’ deference to bail schedules concentrates power in the hands of prosecutors, who can dictate the amount of bail by what charges they choose to file and how many counts and enhancements they add.
One San Francisco judge related the story of a defendant arrested for statutory rape, an offense punishable as a misdemeanor or a felony. The district attorney filed it as a misdemeanor; the judge set bail pursuant to the misdemeanor bail schedule. The defendant’s boss paid for his bond. After he bailed out, the prosecutor re-filed the case as a felony and requested an increase to the felony bail schedule level. This judge noted that the conduct was no different, nor was the danger to the public and risk of failure to appear, and so refused to increase bail. Other judges may have acquiesced.
Not surprisingly, prosecutors tend to strongly support the use of bail schedules. Alameda County District Attorney Nancy O’Malley told Human Rights Watch that she saw them as a good starting point, though noted prosecutors can ask for increases. Los Angeles County District Attorney Jackie Lacey said that she liked the “consistency” that bail schedules provide. Deputy District Attorney Larry Droeger, representing the Los Angeles County District Attorney’s office at a meeting on Los Angeles County bail reform, expressed his office’s support for using schedules, as they tie the bail amount to the seriousness of the crime and, as a practical risk assessment tool, they believe the schedules work.
The director of pretrial services for one Central California county disagreed with Droeger’s premise, warning it is a mistake to equate risk with the seriousness of the charge.
Daria Morrison’s case provides a good example of the charge not correlating to the actual risk level. Prosecutors charged her with three counts of robbery, and the judge set bail at the scheduled amount of $150,000. The judge did not account for her lack of any criminal record, her role as caretaker for her mother, or that she was working two jobs and going to school. The court eventually heard the evidence during the preliminary hearing, and learned that she had been a passenger in the car and not involved in the robbery itself. She ultimately pled to a much-reduced charge with a community service punishment, but not until her family went into debt paying her bail.
One analyst has determined that lowering bail schedules by 10 percent would reduce the percentage of pretrial detainees by 4 percent. The unaffordable bail amounts in the current schedules keep large percentages of people in what essentially amounts to preventive detention.
Preventive detention means holding a defendant in custody pretrial without any opportunity for release, and prevents the accused from absconding or being a danger to the community.
The California Constitution makes preventive detention extremely rare. Article 1, Section 12 guarantees all defendants the right to pretrial release “on bail by sufficient sureties,” unless they are accused of a capital crime, a violent crime, or felony sexual assault when there is “clear and convincing” evidence that their release will entail a substantial likelihood of serious injury to another person, or any felony when there is “clear and convincing” evidence the defendant threatened to cause serious injury to another and is likely to carry out that threat. Before ordering “no bail” or preventive detention, the judge must find “the facts are evident or the presumption great” that the accused is guilty.
However, judges can and often do avoid the constitutional requirements of formal preventive detention by simply setting a bail amount too high for the accused to pay. Chief Justice of the California Supreme Court Tani Cantil-Sakauye told Human Rights Watch that imposing bail results in preventive detention.
Several other judges also acknowledged this. A former Santa Barbara County judge said, “We set bail at an amount to keep the defendant in jail.” A pretrial services official for a Central California county told Human Rights Watch that a judge from Fresno told him that he used bail as preventive detention. When a judge follows the schedule and sets a $5,000 bail for a homeless person, he knows he may as well have ordered a “no bail” detention.
At least one California appellate decision has said: “… [T]he Court may neither deny bail nor set it in a sum that is the functional equivalent of no bail.” This statement may not have the practical force of law. Though some judges may account for a defendant’s ability to pay, most refuse to consider it. Some judges have an understanding of a defendant’s ability to pay, and deliberately set bail above that. While the California and Federal constitutions forbid “excessive” bail, neither require affordable bail.
Of course, using bail as a replacement for preventive detention does not necessarily advance the cause of public safety, as some released people may commit new crimes regardless of socioeconomic status.
Bail bond industry representatives describe money bail as “a liberty-promoting institution” and cite its ability to allow defendants freedom without major costs to taxpayers. But it can involve significant costs and financial harm to defendants and their families. Fees paid to bail bondsmen are not refunded regardless of the outcome of the case.
After bail is set at the police station or in court, defendants or their supporters may go to bondsmen who then decide whether to accept the bond. Bondsmen look at a variety of risk factors about the accused to decide if they should insure the appearance. One crucial factor they look at is how much money the defendant can pay toward the fee.
An employee of Bail Hotline in Sacramento said charging fees is done “case by case”:
Competition among different bond agencies means they will often make deals, including reducing their fee to 8 percent, sometimes lower. They frequently offer payment plans, sometimes agreeing to down payments as low as 1 percent, along with monthly payments.
Matthew Dixon told Human Rights Watch he spent a week in the Alameda County Jail with a $180,000 bail set before a friend could find a deal from a bondsman. His friend paid $1,500 up front on a $15,000 premium, and Matthew now pays $250 each month to the bondsman, who constantly pressures him to make payments.
Paul Fowler described how his son was arrested and held in Contra Costa County Jail with a $250,000 bail before his court appearance. The bondsman pressured him to pay immediately in case the prosecutor added more charges. Fowler waited. At the arraignment, the judge reduced the bail to $30,000. He paid $1,500 down and set up $300 per month payments on a $3,000 premium.
The American Bar Association, recommending abolition of for-profit bail bonding, decried this discretion in the hands of private, minimally regulated, profit-motivated actors:
Several people whom Human Rights Watch interviewed complained about bail bondsmen taking advantage of their lack of knowledge of the system to get them to pay, or otherwise manipulating them. One person described how a bondsman convinced her mother, diagnosed with a mental illness, to pay a non-refundable fee, when the daughter could have deposited the full bail amount.
Hutch Harutyunyan, of Gotham Bail Bonds, said that, by contract, bondsmen have earned their fees when police release the prisoner. If the case does not get filed, the person paying the fee still owes the money under any agreed upon payment plan. If the prosecutor decides to file the case at some future date, after the court has exonerated the original bond, and the judge sets a new bail, the defendant must pay a completely new fee to obtain bail.
Kevin Ocampo in Alameda County paid a 6 percent fee on a $250,000 bail to get his cousin out of jail. When he went to court, the judge raised the bail to $325,000. The bondsman would not apply the amount already paid to the new bond. Instead, Kevin had to pay a new premium of 8 percent on the new amount.
Henry Anderson said he paid a fee to a bondsman. He went to court, and his case was dismissed. The district attorney later re-filed the charges, the court set a new bail, and Anderson had to pay a whole new fee to secure his release.
The US and the Philippines are the only countries in the world with private, for-profit bail bond industries. Many other countries and some states use financial bail, but require payment directly to a government agency. In Illinois, defendants pay 10 percent of the bail directly to the court clerk. If they make their court dates, the clerk returns their money minus a maximum $100 processing fee. The disadvantage of this type of system for people seeking pretrial release is that they must pay the full 10 percent amount up front. Bondsmen in California allow many people to buy freedom with a low down-payment and installments when they might otherwise not be able to pay. Daria Morrison is still making payments on the bond her father got for her after she spent three weeks in jail on a robbery charge in Los Angeles County, though she is grateful to the bondsman for helping her out of jail.
II. Bail Leads to Jailing People Who Are Not Guilty
One of the most harmful aspects of California’s bail system is that it results in the pretrial incarceration of hundreds of thousands of people without proof they committed any crime.
From 2011-2015, police in California made 1,451,441 felony arrests of individuals, all but a small fraction of whom had bail set for some period of time. Of those, 459,847 were arrested and held in jail, but never found guilty of any crime. Prosecutors did not even file charges against 273,899 of those people.
In other words, over a quarter-of-a million Californians sat in jail for up to five days, accused of felonies for which evidence was so lacking prosecutors could not bring a case.
The others had cases filed, but lacked sufficient proof of guilt, resulting in eventual dismissal or acquittal after weeks and months in jail. Many of these people were victims of baseless arrests; others, mistakes of judgment, or misunderstandings of the law.
These people spent days, weeks, and months in jail while waiting for trial, serving out sentences for crimes they did not commit, losing jobs, missing their families, having to drop out of school, suffering the misery of being locked up. By setting bail that people
cannot afford, the pretrial detention system punishes people without proving their guilt.
The cost to taxpayers of this senseless pretrial punishment is staggering. Each day a person is held in custody costs an average of $113.87. Human Rights Watch analyzed all bookings into jails in Alameda, Fresno, Orange, Sacramento, San Francisco, and San Bernardino Counties for 2014 and 2015. The total cost of jailing people, never found guilty of any crime, just in these counties, was about $37.5 million over the two years.
Bail Keeps People in Jail After Arrest without Basis
Jason Miller is in his mid-forties and lives on the streets in the Skid Row section of downtown Los Angeles. Because of his homeless status, and because the Los Angeles Police Department saturates the neighborhood with officers, he has had many encounters with police in recent years. He has been off probation since 2013, but counts 10-15 arrests since then. As he has no money to pay bail, his arrests mean he goes to jail.
In the summer of 2016, Jason told Human Rights Watch he and an officer had an argument about his dog. Jason demanded to speak to a sergeant, but instead, a lieutenant came and ordered the officers to arrest him. The reason they gave: he possessed narcotics.
They took him to the police station, booked him, and held him under the felony bail schedule amount of $10,000. Jason told Human Rights Watch he had not possessed drugs, but he had no money to get out. He stayed in the station jail, unable to sleep due to the noise, with no books, television, or anyone to talk to. On the third day, he went to court where he was packed into a cell with close to 40 other prisoners, many of whom were starting fights. Finally, at about 4:30 p.m., deputies at the lock-up told him the case was a DA reject—no filing. It took him two more days and $122 to get his dog out of the pound. All his property, including tent, clothing, toiletries, and medications were gone.
Jailing People Who Are Not Guilty
David Gonzalez, 19, was looking forward to going to college in the fall. He had just graduated from high school, lined up his classes for August at Santa Ana College, and was looking for a job when police arrested him and locked him up in the Orange County Jail.
David told Human Rights Watch that the next day, July 7, 2016, police brought him to court for his arraignment. He learned he was accused of raping an unconscious person, a crime punishable by up to eight years in prison. The judge set bail at $100,000.
David had had sex with the girl who had been raped, but had been away in school at the time of the alleged rape. Still, they took him back to jail where he would have to stay unless his family could get the money together.
David’s father worked at a restaurant, making minimum wage. His siblings had no extra money. The family home was about to be foreclosed on, so they could not borrow money against the property.
In jail, David tried to stay out of trouble. A couple of the older guys, seeing he was just a kid, looked out for him a bit and advised him to lay low. Still, he had a cellmate who gave him problems. He had to fight to protect himself a couple of times. Otherwise, there was nothing to do but sit and wait, and hope the truth would emerge.
David’s sister Nina would miss work once a week to drive from San Bernardino to visit him and try to keep his spirits up. Still, he would break down in tears when she saw him. Nina was able to get his school records to help establish that he was in school that day. The case was based on DNA evidence, but the prosecutor had not spoken to the victim about David. Eventually, David’s lawyer located the victim. She confirmed that she had been with David the weekend before the rape, but he had nothing to do with the crime.
On September 30, 2016, the prosecutor spoke to the victim and agreed to dismiss the case. David had spent nearly three months in jail for a crime he did not commit, because bail was so high his family could not afford to pay. He missed his first semester of college.
Bail Keeps People in Jail Who Never Have Charges Filed
Human Rights Watch analyzed Alameda County 2014-2015 jail bookings and release data to determine how many people were released, under what circumstances, and how long they spent in custody.
The data reveals another 3,353 people whose cases were dismissed, but who still spent an average of 27.8 days in jail, probably costing the county more than $10.5 million.
The data shows a relatively small number of people given own recognizance release, but indicates it often took the courts a long time to come to that decision. Own recognizance release primarily occurred in the first week, but sometimes took weeks and months. 3,848 people were released this way, but they averaged 14.9 days in jail. Had they been cited and released by the arresting officer, the county would have saved around $6.5 million.
Of the 12,166 people who posted bond, most did so within the first day.
Figure 3 shows a very small number of police releases compared to “no filings.” This comparison raises a question about the judgment of police supervisors keeping people detained whose cases will not be filed. It points to the potential danger of a book and release program dependent on the station supervisor’s discretion as compared with a rule requiring cite and release instead of arrest for most cases.
Human Rights Watch similarly analyzed booking and release data from five other counties. Sacramento County, which jailed a similar number of people as Alameda County in 2014-2015, provides a comparison. In Sacramento, 5,094 people stayed in custody an average of 3.2 days with no charges filed. Sacramento booked and released about 19 percent of arrestees, over 10,000 people, within a day of arrest, likely reducing the number of “no filing” releases.
- 10,459 people stayed in custody an average of 4.3 days before bailing out, though most were out in about a day. It is unclear how many of those who had to pay for their freedom ended up with no charges filed.
- 953 stayed in custody until their cases were dismissed or a jury acquitted them.
- 4,316 people were not cited out, but got own recognizance release orders from the judge. They stayed in custody an average of 8.2 days; the mean was 2 days.
- Just under 12,000 were released after finishing their sentence, spending an average of 25.4 days in custody.
While not as dramatic as Alameda County’s figures, data from Sacramento shows a substantial number of people in custody with no charges ever filed. Many of those arrestees whose cases did not result in filing had already paid non-refundable bail fees.
The CEO of San Francisco’s Pretrial Diversion Project, Will Leong, said that the district attorney in his county also rejects a large number of cases.
Frank Robinson had a good job with the local transit service. He was arrested in Alameda County on December 23, 2015, on a domestic violence warrant. The police set bail at $130,000. His mother co-signed for the bond and paid $1,000 with an agreement to make payments for the rest of the fee. He got out the next day. The prosecutor did not file criminal charges against him. Frank said:
And now that I am out of jail, I have to pay $200 a month to the bail bond agent. I don’t understand why I have to pay something when the charge was dropped. My family is stunned that this happened to me.
Nancy Wilson described being arrested several times by Oakland police on drug related charges, borrowing money to pay bail, only to have no charges filed. Brandon Watkins had a similar experience, also in Alameda County. Police arrested him, claiming he had committed a battery. They set a $15,000 bail. His parents went to a bond agent, paid $1,500, and secured his release. When Brandon went to court, he learned that there was no case filed against him.
India Fuller was arrested in Sonoma County when her son’s ex-girlfriend accused her of assault. The police set a bail of $265,000. Various family members contributed to her bail fund, gathering $3,000 to give to the bondsman. India had been in jail for four days. The prosecutor dropped the charges, but she still pays $350 per month to the bondsman. She has fallen behind in her car payments. If she loses her car, she will lose her job as a driver.
Replacing arrests with non-custody citations would save police processing costs, reduce jail overcrowding, diminish the harms associated with even short periods of time in jail, like lost jobs and lost property, improve community relations with police, and limit police uses of force associated with “hands on” arrests.
III. Bail and Jail Result in an Unfair Justice System
I’ve seen it. A time served offer on a custody defendant on a low-level charge, all they think about is, “Do I get out today? Can I get out today?” We have to take a look at whether we are contributing to the problem.
—Chief Justice Tani Cantil-Sakauye, California Supreme Court, March 12, 2016
The DA’s objective in making the bail so high and then raising it again when we came up with the original amount was solely to force a plea bargain. Then they kept dragging it out. They were not in it for justice, they were in it for statistics.
—Kevin Ocampo, Alameda County resident, who posted bond for his cousin, May 27, 2016
It’s like someone walks up and puts a gun to your head and says, “Hey, give me your money.”
—Oscar De La Torre, executive director of Pico Youth and Family Center, November 21, 2016
Bail Coerces People into Giving Up the Right to Trial
According to the latest available data:
- 80.8 percent of California filed felony cases resolve through guilty pleas;
- 16.7 percent are dismissed or transferred to another jurisdiction;
- 2.5 percent go to trial.
For non-traffic misdemeanors:
- 62.3 percent result in guilty pleas;
- 35.8 percent are dismissed or diverted;
- Just under 1 percent go to trial.
In felony cases, of those that go to trial:
- 81 percent result in some felony guilty verdict;
- 3 percent result in a reduced misdemeanor verdict;
- 14 percent result in acquittal or dismissal.
Though impossible to quantify, a large number of the guilty pleas represented in the above statistics happen because defendants are detained pretrial and see pleading guilty as their quickest way out of jail. Pretrial prisoners know they will have to wait approximately 30 days on a misdemeanor case and 90 days on a felony case before they go to trial. Prosecutors and judges often offer settlement terms that result in the defendant getting out sooner than it would take to get to trial. Defendants feel this coercion acutely and often give up their right to trial in order to be released.
Studies in different jurisdictions nationwide have found a correlation between pretrial detention and likelihood of conviction, as well as likelihood of a custody sentence and the length of that sentence.
Researchers have attempted to determine whether or not pretrial detention actually causes those negative consequences, by controlling for factors like severity of crime and criminal history that would otherwise affect the results. One study looked at all cases in Philadelphia’s criminal courts and found being in pretrial detention increased likelihood of conviction by 13 percent, primarily through an increase in guilty pleas. On average, those detained received jail or prison sentences five months greater than those fighting their cases from outside. They paid significantly more in court fees. The effect was 17 percent larger for first and second time offenders.
Another significant finding of the Philadelphia study was the distinction drawn between “strong evidence” cases and “weak evidence” cases. “Strong evidence” cases were those like drug or gun possession, in which police most likely found the person with the contraband, or driving under the influence cases, in which a blood-alcohol test objectively measures intoxication. These are cases in which guilt is not easily disputed, so high rates of guilty pleas are expected.
“Weak evidence” cases like assault or burglary or robbery, in which there is often an eyewitness identification question or a self-defense issue, are more readily contested. All else equal, “weak evidence” cases should be more likely to go to trial, as they are harder to prove, and easier to defend. The study found the impact of pretrial detention on guilty pleas much more pronounced for “weak evidence” cases, indicating that pretrial detention was pressuring people who should be expected to fight their cases to plead guilty. The Philadelphia study found similar results for black as for white defendants. However, in a jurisdiction that incarcerates a vastly greater proportion of black people, the negative consequence has racial impact. Other studies in Texas, Philadelphia, New York, and Miami have reached similar conclusions.
The criminal system in California may not follow the exact patterns of the jurisdictions studied by the researchers cited above. However, the findings from these jurisdictions are consistent, the framework of their court systems and pretrial detention decision-making processes are not significantly different from those in California, and no study in any California county has revealed contrary findings.
One judge in Los Angeles County told Human Rights Watch a supervising judge had said to him lower bail would reduce the number of people pleading on terms prosecutors favor. George Gascon, district attorney for San Francisco, said that bail results in pretrial detention, which leads to faster guilty pleas, including by innocent people who want to get out of jail. Deputy Public Defenders Brian Bloom and Rodney Brooks said a significant purpose of setting bail is to get defendants to plead to a less favorable disposition.
Human Rights Watch spoke to several people who described cases in which they, loved ones, or clients pled guilty to get out of jail, though they had a case they thought they should fight.
Carlos Garcia’s case is a classic example of an individual remaining in jail because he cannot afford bail, and entering a guilty plea to get out of jail. Carlos told Human Rights Watch that on September 9, 2010, his parents were attending a meeting of public housing tenants. Carlos stopped by to give his father a set of keys. A Housing Authority police officer stopped him, telling him the meeting was closed to the public. Carlos argued. The argument escalated. According to several witnesses, several officers grabbed Carlos, threw him down, and piled on him, cutting his lip and hurting his back.
The officers arrested Carlos, booked him at the station jail, accused him of “resisting arrest,” and set a $10,000 bail according to the schedule. No one in his family had the money to pay. Carlos told Human Rights Watch that he sat in jail in pain from the beating. He said he looked around the cell and recognized people from his neighborhood who might be dangerous to him. So he sat with his back against the wall, staying awake all night.
The next morning, Carlos went to the San Fernando courthouse. He was not sure if there were witnesses that would testify to his innocence, but he knew if he was out, he could talk to people at the meeting and be better placed to defend himself against the charges.
He also knew he would have to wait at least 30 days before he could go to trial—30 days of poor food, crowded jails cells, stress from other prisoners, and not taking his father to his dialysis appointments. He would lose his job, and miss taking his daughter to school.
The prosecutor offered him time served. Pick up trash on the side of the freeway for 15 days on his weekends. Take some anger management classes. Pay some money. Three years of probation. Get out of jail. Carlos took the deal.
Several hours later, he walked out of the courthouse lock-up with a new criminal conviction on his record, owing money, having lost the opportunity to clear his name. A couple of months later, when he learned that the Housing Authority was using his conviction as a reason to evict his parents, he tried to withdraw his plea and set his case for trial. The judge denied his request.
Data from Jail Releases Indicates Prisoners Plead to Get Out of Jail
Human Rights Watch’s analysis of county booking data tracks the length of time people stayed in jail before being released following sentencing. A pretrial prisoner has a right to go to trial within 30 days of arraignment for a misdemeanor and within about 90 days of arraignment for a felony, which typically occurs within one to three days of arrest. In each of the counties Human Rights Watch examined, the vast majority of people released from jail as “sentenced” were released before the earliest possible date they could have gone to trial. In other words, to assert their innocence at trial, they would have had to stay in jail longer than they did by pleading guilty. In fact, they would have had to reject a plea deal offering them their freedom.
Plea Deals Show Custody Decisions Are Not About Danger
Arthur Charles pled “not guilty” to a misdemeanor domestic battery charge, intending to assert his innocence in trial. His lawyer asked for an own recognizance release, as it was his first offense and he had a job and place to live, away from the complaining witness. The prosecutor opposed release, telling the judge Arthur was too dangerous to be free, even with a court-imposed stay away order. The judge set bail Arthur could not afford. So, he accepted the prosecutor’s settlement offer: plead guilty, get out of jail within a day or two, accept probation with various conditions including expensive classes and fines, and agree to the judge’s order to stay away from his partner. He was too dangerous to release—until he gave up his right to trial.
Robin Lipetzky, chief public defender for Contra Costa County, said that situations like Arthur’s are commonplace. Kenneth Clayman, interim public defender for Santa Barbara County, agreed that “time served” was a common case disposition. In other words, courts and prosecutors are agreeing that public safety is served by releasing people on probation, yet at the outset take the contradictory position that the person is too dangerous to be released on their own recognizance or under pretrial supervision.
Life in Installments
Various studies show that pretrial detention results in more jail time on average following a guilty plea. What the studies do not show is the future impact for those who could have fought their cases but instead accepted probation and a conviction.
From 2010-2014, approximately 75 percent of all felony convictions resulted in probation sentences. Brian Bloom and Rodney Brooks, public defenders in Alameda County, describe Alameda as a “probation county.” They describe how people accused of crimes can only get out by pleading guilty and accepting probation. Even on felonies, they may serve a short jail sentence, but probation can be as long as five years.
Defendants have no leverage to contest any subsequent charges of violating probation as they do not have a right to a jury trial and the standard of proof for the prosecutor is easy to meet. Once on probation, it is easy to get sent back to jail again and again. Bloom and Brooks both referred to it as “doing life on the installment plan.”
Pretrial Detention Strengthens the Prosecution
Prosecutors have an institutional incentive to secure pretrial detention, regardless of public safety concerns, because they can usually resolve cases faster and on terms they prefer if they are negotiating with a defendant in custody. Prosecutors understand that defendants who are not in custody feel less pressure to plead guilty and have many more advantages fighting their case than in-custody defendants.
Aaron Jansen of the Los Angeles County Public Defenders told Human Rights Watch of a client arrested for possessing a knife, a charge that could be a felony or misdemeanor. The prosecutor charged the felony and the judge set a $20,000 bail. The settlement offer
was to serve 16 months in jail. The man rejected it and managed to get bailed out. After he was out of custody, the offer changed to probation and community labor.
Giovanni Giordani, chief trial deputy for the Santa Barbara County Public Defender’s Office, said defendants in custody get worse deals than those on the streets. Defense lawyers told Human Rights Watch numerous stories of clients pleading guilty to get out of custody rather than litigate their cases and assert valid defenses.
Court administrators in one Bay Area county told Human Rights Watch that the district attorney for their county has privately expressed dislike for widespread pretrial release because it makes obtaining guilty pleas more difficult. Contra Costa County Probation Manager Russ Miller observed that prosecutors, just like judges and defense counsel, get caught up in their workload. Since detained defendants plead more easily, pretrial release would slow prosecutors down.
Judiciary’s Institutional Interest in Pretrial Detention
A Contra Costa County judge said the biggest challenge to reforming the bail system would come from judges and prosecutors. Alameda County Deputy Public Defenders Brian Bloom and Rodney Brooks said only a tiny percentage of judges will reduce bail on a large scale or extend own recognizance release; most feel it is safer to keep people in custody.
Judges resist pretrial release for various reasons. One is fear that defendants will not return to court. This concern has some legitimacy, as missed court dates do cost the court in time and money. But most people do return to court; few actively evade prosecution.
Using fear of missed court dates as a blanket excuse to detain risks vast over-incarceration and the many societal and fiscal harms detailed in this report.
The other reason that judges emphasize opposing release is fear that someone they release will commit a crime during the pretrial period, and the judge will be blamed. While there is some validity to this concern, it is one that underlies any judicial decision within the criminal system, particularly sentencing, probation, and pretrial decisions. This concern could be used as an excuse to refuse any risk and always default to incarceration. Or it could lead judges to hold very detailed hearings, gathering and considering as much information as possible to have a more informed judgment of the risk. Such hearings take more time than courts commonly devote to considering pretrial release decisions.
Underlying the judiciary’s overall objection to more widespread pretrial release is a belief that it would hinder the efficient processing of criminal cases through the courts. One crucial imperative judges perceive under the current system is to process high volumes of cases as quickly as possible. Alameda County public defenders said setting lower bail or granting own recognizance release takes more work and time than following the bail schedule.
A former Alameda County courts administrator explained that many judges resist pretrial release because they are concerned that out of custody defendants will clog their calendars. They believe many more defendants will litigate their cases, and that the number of trials will greatly increase. Judges fear the court system would not be able to handle an increase in the percentage of cases being litigated fully instead of simply pled out.
Human Rights Watch analyzed data from Sacramento County that illustrates this point. The median date of guilty plea for defendants on misdemeanor and felonies who stayed in custody was 20 days (mean of 52); 100 days for those out on bail (mean of 122); and 7o days for those on own recognizance release (mean of 97). These statistics show that in-custody defendants pled guilty quickly, while those out of custody litigated their cases for a much longer period of time. Presumably, this added litigation time allowed them to more fully develop defenses and mitigation evidence and obtain better results.
Some Los Angeles County judges told Human Rights Watch that other judges, including supervising judges, had warned them not to release too many pretrial detainees because people in custody plead more readily. One even said that pretrial release should be avoided because it would diminish the prosecutor’s advantage in plea negotiations.
An Alameda County judge said judges who go against the mainstream and the district attorney may develop a “pro-defendant” reputation, and suffer consequences.
She said that the presiding judge may move such a judge to an unfavorable assignment. A pro-defendant reputation can also harm a judge during an election campaign if well organized law enforcement groups line up in opposition. The former Alameda County courts administrator said the judiciary and district attorneys tend to have a close relationship, as many judges are former deputy district attorneys. Because prosecutors oppose pretrial release, judges tend to go along with them.
More pretrial release would result in cases moving more slowly through the courts. It would probably result in more motions and litigation, as opposed to early pleas. More defendants would have a better chance of having their cases dismissed or settled with less emphasis on jail as a punishment.
However, removing the pressure to plea unrelated to guilt or innocence caused by pretrial detention will improve the quality of justice. More in-depth litigation should result in fairer, more accurate outcomes, which would improve public confidence in the court system.
A slower moving court system that does not coerce pleas through pretrial detention may also give prosecutors and judges incentive to make more precise decisions about who to prosecute and incarcerate. It may encourage local governments and law enforcement to reconsider priorities, including finding ways to address drug use, homelessness, and mental illness without criminalization.
Dante Johnson’s Story
On June 28, 2005, Dante Johnson was visiting friends in Palmdale, a town about two hours from his home in Inglewood, and did not come back home until the next day.
Meanwhile, that afternoon, two Inglewood Police officers on patrol saw a young man wearing gang colors riding his bicycle on the sidewalk. When they pulled up within 50 feet of him, he got off his bike and ran down an alley. The officers chased but could not catch the man, who scaled a fence, dropping a gun as he did so.
The officers picked up the gun and later wrote a report. They said they knew the person who ran, but did not identify him by name or in any other way. They gave a fairly generic description—young, black, male, average size and weight. The only distinguishing feature they described was a severe case of facial acne.
The next day, Dante, 18, was taking a walk outside his home when the same officers came up and arrested him, claiming he was the man who had run from them. Dante had been convicted of a 2nd degree burglary when he was 14, and placed on probation. He was arrested for a fight when he was 11. As an adult, he had gotten a ticket for loitering. He was a member of the local gang, as were many people in his family and his neighborhood.
On his first court date, his arraignment, Dante told his lawyer they had the wrong person and described where he had been. His lawyer noticed Dante had a clear complexion and that, based on the report and the officers’ testimony at the preliminary hearing, they did not have a good chance to see the face of the person who ran. The lawyer became convinced Dante was not guilty, and that he could win at trial. They agreed to fight the case.
The prosecutor had filed serious charges against Dante. The base charges were felony possession of a concealed weapon and possession of a loaded firearm. They added a gang enhancement, claiming that the gun possession was for the benefit of the gang. If Dante were to be convicted, in addition to potentially serving eight years in prison, the charge was also a “strike,” exposing him to a doubling of his sentence for any future felonies. Because of the seriousness of the potential consequences and the strength of his defense, he and his lawyer agreed that he must fight the case. The judge set bail at $50,000. Dante’s mother, who worked at a nursing home, could not pay.
So Dante sat in jail while his attorney worked on the case. His attorney appointed a fingerprint expert to examine the gun and had an investigator go to Palmdale to locate the alibi witnesses and find the register Dante had signed at the swimming pool he had gone to during his trip to Palmdale. He identified other people who had negative experiences with these officers, who might testify to their dishonesty. The process was time-consuming. Investigations were slow. The prosecutor was not turning over documents Dante’s lawyer was requesting. Witnesses were hard to find. After 60 days in custody, Dante was desperate.
Dante was moved back and forth between Men’s Central Jail in downtown Los Angeles and North County Correctional Facility, near Valencia. Each time, sheriffs would wake him up at 4 a.m., and stick him in packed crowded holding tanks for hours, before cramming him, shackled, onto the bus. Then he would have to wait in other holding tanks, sometimes all day, fending off other prisoners, sitting on concrete benches, waiting for his court appearance. He would spend hours in the holding cell at the jail, waiting to be processed back to his cell. There were constantly new cell mates, always a concern given the gang and racial tension in the jail, and new deputies, who hit prisoners with flashlights or threw them on the floor. Six men were in a twelve by twelve-foot cell with a toilet in the open. There were constant fights, usually between prisoners of different races.
The cells had little room for exercise, and they were frequently on “lock-down.” Prisoners would play dominos or cards, watch television, or read. Deputies let them outside for three hours, one day a week. At Men’s Central Jail, Dante said, they often lost their outdoor time.
“I’ll plead for time served,” Dante insisted to his lawyer on their pretrial date on September 1, sixty-five days after his arrest. His lawyer argued with him through the glass partition in the attorney interview cell next to the courtroom that he had to fight the case. Dante, wearing his blue jumpsuit with “County Jail” on the back, told his lawyer, “I have to get out of here.”
There was no chance the judge would give Dante an own recognizance release. But the prosecutor knew the weakness of his case and that a plea deal would be an easy way to get a conviction. He agreed: “Plead guilty to the gun charge and the gang enhancement, and he can go home today.” Dante took the deal. Dante plead “no contest” in open court, though his lawyer refused to join in the plea or even sign the waivers. The judge placed Dante on probation, with a condition that he could not be around any gang members, including those in his own family. Dante ate a home-cooked meal that night.
IV. Bail Devastates Poor and Middle-Income Defendants and Households
Wealthy people can afford to pay for their pretrial freedom if the judge sets bail, avoiding the misery of jail and gaining the benefit of being out of custody while fighting their case.
Poor and middle-income people, on the other hand, face a dilemma. For some, the dilemma is easily, if unhappily, resolved, because they simply have no money or chance of obtaining enough money to bail out. Others must decide: stay in jail until the case is resolved, often by an early guilty plea, or incur crushing debt.
Poor living conditions, including over-crowding, unhealthy food, lack of medical treatment, violence from other prisoners and guards, the desire and need to work, stay in school, care for dependents, and pay rent motivate many to accept the financial burden of paying bail.
People also put themselves in financial peril to pay bail because they know that being out of custody vastly improves their chances of successfully resolving their case.
The benefit of being out of custody is known to everyone involved in the system. Victor Lawrence, arrested in Alameda County on a misdemeanor charge, bailed out immediately:
The only thing you can do is bail out. All the young men in there knew, if you go see the judge on your own recognizance, you’re going to be better off.
Judges, lawyers, and court administrators who spoke to Human Rights Watch acknowledged the benefit of being out of custody when fighting a case, citing, among other things, the logistical difficulties of assisting in a defense from jail. Andres Del Alcazar, Santa Clara Deputy public defender, told Human Rights Watch that an out of custody defendant will generally resolve their case for a lighter sentence than one in custody.
In Los Angeles County, an out of custody defendant is usually offered a no-jail plea deal at the “early disposition” hearing, while an in-custody defendant most likely will have to serve further jail time. Giovanni Giordani, chief trial deputy for the Santa Barbara County Public Defender’s Office, told Human Rights Watch that almost invariably in-custody defendants get offered worse plea deals than those out of custody.
Statistical analysis supports the perceived benefit of being out of jail. One frequently quoted study, sponsored by the Laura and John Arnold Foundation (“Arnold”), looking at pretrial populations in Kentucky, found dramatic differences in the outcomes of cases based on whether the accused was in custody.
The Arnold study found that detained defendants were 4.44 times more likely to get sentenced to jail and 3.32 times more likely to get sentenced to prison than those out of custody. It found that detained defendants served 2.78 times longer jail sentences and 2.36 times longer prison sentences. The Arnold study does not purport to show that pretrial detention causes the disparity in outcomes, and critics have argued that their study does not sufficiently control for seriousness of offense.
Human Rights Watch analysis of Sacramento County data supports the perception that being out of jail leads to better results in court, as the longer time to resolve the case generally means a better opportunity to prepare and present a defense.
There are a variety of reasons why being out of custody helps improve a person’s chances of getting a good result in their case. Being out allows more productive attorney-client communications. It allows the accused to work, earn money, and pay for better representation, including experts and investigators. It allows the accused to enroll in drug treatment programs or psychological counselling, attend school, maintain or gain employment, pay restitution, or do other activities that will show the judge and prosecutor that they deserve lenience. It allows the accused to litigate the case and fully develop defenses, including locating hard to find witnesses or pursuing discovery that the prosecutor is not disclosing, without the time pressure imposed by the hardship of being in jail. Being out of jail allows the accused to help locate evidence and witnesses that might be impossible to find without their participation. It is also true that being at liberty can allow a defendant the opportunity to intimidate a witness or destroy evidence.
Daria Morrison and Sarah Jackson’s story illustrates the advantage of getting out of custody. Daria was a 20-year-old college student in Los Angeles, hoping to be a lawyer. She had a good job as a waitress and was saving some money. She cared for her unemployed mother. She had no prior criminal record. Sarah was also 20 and without prior convictions. She and Daria were arrested for a series of robberies in which—according to testimony at the preliminary hearing—neither actively participated, although they were in the car at the scene.
Both were jailed, with bail set, as their case went through the court process. Daria said she was locked in her cell 23 hours a day, with no exercise and nothing to do but read or sleep. She was surrounded by people who were constantly fighting.
After three weeks in custody, her father, who owned a small business, found a bondsman who accepted a 6 percent fee on her $150,000 bail, with $1,500 paid up front and $250 monthly payments, which drained her savings. Daria lost her job. She had to withdraw from school for the semester. She had to check in with the bondsman every week, or they would revoke her bail. But at least she was out.
The benefits of being out of custody were immediate. Daria was able to visit her lawyer easily, more frequently, and in a calm, comfortable setting, unlike the jail. She was able to speak to him freely about her case and herself. Being out with her family helped her get better advice on how to handle her case, and allowed her to think about it with less of a sense of desperation. When she went to court, she wore her own clothes, instead of “looking like a criminal like everybody else” in their jail jumpsuits.
Her lawyer described the benefits of having Daria out of custody: “[It] was indispensable in preparing the defense.… [W]hen interviewing clients in a custody setting, they can be tense and distracted. In my office, they can have coffee, snacks, and there is an opportunity to relax and develop a rapport. This was especially important because [Daria] was hesitant to tell me things she thought might make her look bad in my eyes, but in reality were helpful to our defense.”
Having her out of custody gave Daria’s lawyer time to work out a good deal for her. She pled to a less serious charge with an opportunity to have it dismissed altogether if she completed her community service requirement.
Sarah’s family was not able to bail her out of jail. After three months, she pled guilty for a time-served sentence, but had to accept convictions on two serious felonies that will greatly enhance her sentence on any future crime. She accepted five years of probation, with a seven-year suspended prison sentence, meaning any rule violation could send her to prison for a long time. Unlike Daria, she will not have the charges dismissed and the conviction removed from her record.
In general, prisons and jails house poor people. The inability of poor people to pay for their pretrial release, and all the disadvantages that go along with fighting a case from jail, certainly contributes to their overrepresentation in the prison and jail populations.
Paying Bail Overburdens Poor and Middle-Income Households with Debt
Carlos Garcia, arrested for felony assault, was held in jail on $75,000 bail. The prosecutor offered him a plea deal for two years in prison. He resigned himself to taking the deal and doing the time. However, Carlos’ 13-year old son, who lived with Carlos’ ex-girlfriend, Marta Lopez, missed his father and convinced her to help. Marta was not wealthy, but had some income. She and Carlos’ mother found a bail bondsman who charged them a $6,000 premium with a $1,500 down-payment, and monthly charges of $150. Out of custody, Carlos could work more closely with his attorney, who negotiated a probation sentence, with only a brief return to jail and a possible reduction of the charge to a misdemeanor.
However, it cost him and his family considerably. Carlos gets disability payments that he contributes to his parents’ household expenses. With the added cost of the monthly bail payments, they have fallen behind on the electricity bill, and they struggle to buy food. His mother has postponed dental work, and become depressed due to the financial strain. Marta also contributes to the payments, which has made it harder for her to pay her other bills and support her own children.
Wealthy people can quickly pay to get out of jail. People with a lot of money or with very low bail amounts can pay the full cash amount. Nationally, only about 5 percent of all pretrial detainees pay the full amount. Poor and middle-income people rely on bail bonds. Often it takes time to gather money or to find a bondsman willing to agree to low enough payment terms. People borrow from neighbors, friends, family, even ex-girlfriends, often resulting in stress and conflict among the people who put up money. Marta’s husband is extremely angry she contributed family money to help Carlos.
A recent national Federal Reserve Board study measuring financial health revealed that 47 percent of Americans would either have to borrow money, sell property, or would simply be unable to come up with $400 to cover an emergency expense. A Bankrate survey showed that only 38 percent of Americans could cover a $1,000 emergency room visit. A 2011 National Bureau of Economic Research study found that 50 percent of Americans are “financially fragile,” and that 40 percent would either not be able to raise $2,000 in a month, or would have to sell or pawn property or go into debt to do so. Most people would not be able to pay even a low down-payment on the fee for a typical California bail bond without incurring significant debt. Subsequent payments ensure the debt endures.
According to data collection from 2000-2009, only 20.6 percent of felony defendants in California received “non-financial” release, primarily own recognizance or citation release, while 23.8 percent made bail. The rest stayed in custody. The probability of posting bail is associated with the amount of bail set, as defendants with bail set below $50,000 were four times more likely to post it.
One approach to pretrial detention is to simply lower bail amounts to make them more affordable. In fact, the bail bond industry strongly supports lowering overall bail amounts. Of course, bondsmen frequently offer 1 percent down-payments and have to expend energy to collect the rest. If overall bail amounts are lowered, their liability is reduced, and more people come to them thinking they can pay. One study determined that lowering average bail amounts by $10,000 would gain a 4 percentage point reduction in the share of unsentenced prisoners. Lowering levels of bail would allow more people to get out of custody, while increasing the bail industry’s profitability, but would not change the system’s fundamental wealth-based discrimination.
California’s poverty rate in 2015 was 15.3 percent, higher than the overall rate for the rest of the country, and 20 percent if adjusted for cost of living in the state. Another 20 percent of Californians are on the threshold of poverty. Black and Latino Californians have much higher poverty rates than white Californians.
At the macro level, there is a clear correlation between the poverty rate and the unsentenced jail population. California counties with a higher proportion of their population under the poverty line generally have a higher proportion of their population in jail unsentenced. The correlation is strongest in larger and medium-sized counties and slightly weaker in the smallest counties. The unsentenced population rate also correlates strongly with the unemployment rate.
Nationally, most people who cannot pay bail are in the poorest third of the population, and have a pre-incarceration median income of less than half the median income of non-incarcerated people of similar age. This difference in income is more pronounced for black men and women.
Cara Esparza told Human Rights Watch that her son Sean Brown was arrested and accused of a felony assault on November 2, 2015, in Long Beach. At the police station, they set a $30,000 bail pursuant to the bail schedule. Cara was scared for her teenaged son, who has been diagnosed with bipolar disorder and would be vulnerable if he stayed in jail.
But Cara, whose only employment was as her son’s caregiver, did not have money. To pay for the bond, she knew she would have to go into debt.
She found a bondsman who accepted a 1 percent down-payment of $300 and $150 monthly payments on a $3,500 premium. A family member loaned her the down-payment. Sean got out of jail after three days. A month and a half later, he pled “no contest” to a greatly reduced charge, a misdemeanor battery with a community labor sentence.
Cara told Human Rights Watch that she was slowly paying back the relative who loaned her the money. She gave cash when she could, but often the payment was in the form of baked goods or doing work. The bondsman only took cash, so she had to make considerable cut-backs. She paid less of her monthly gas bill, leaving her owing and in danger of having her gas cut. She bought less food for herself and her son, and reduced her phone plan. They had no money for going to the movies, a Thanksgiving turkey, or Christmas presents. Once she finished paying the bondsman, she and her son would still have to deal with the court fees he owed.
Nationally, bail amounts have been increasing steadily. From 1992-2006, median bail amounts rose by $15,000, while the percentage of cases with bail amounts below $5,000 dropped from around 30 to 20. The percentage of defendants given own recognizance releases has gone down considerably in that time. Los Angeles County Deputy Public Defender Nick Stewart-Oaten observed that for a large portion of his caseload, the bail amount does not even matter, as many are too poor to make any payment.
Jane Meyers lives in Tucson, Arizona, and works in the food service industry. Her son was arrested and faced felony charges in Los Angeles, with a $60,000 bail. Jane attempted to bail out her son, with the plan that he would come live with her and work a restaurant job that she had lined up for him. A bondsman agreed to a relatively low 7 percent premium, without requiring property to secure the bond. Jane simply could not pay. Her son stayed in jail.
Felix Ayala was arrested in Alameda County on a vandalism charge, with a high bail. He had some money for a down-payment, but now struggles with the monthly payments. He has a job, but the added expense forces him to cut back on spending for food and to delay paying utility bills.
Alicia Wright was arrested in West Oakland and charged with possession of narcotics with intent to sell. It was her first offense, but the judge set a high bail, possibly because she was new to the state. Her boyfriend borrowed $2,000 for a down-payment on the premium, and she pays $100 a month toward the balance. The friend who loaned the money is upset, but Alicia does not have money to pay him back. She worries that she will not be able to pay her rent and will lose her home.
Jordan Davis was arrested for possession of hash. At the police station, bail was set at $30,000. He bailed out the next morning, agreeing to make $300 monthly payments. He later negotiated a reduction in the payments to $100 per month. But because Jordan could not pay his other bills, he faced eviction and had to move to a cheaper apartment in a worse neighborhood, buy less healthy foods, and pawn personal property.
Marcus Garza was arrested just before Thanksgiving in 2015, and held with a $35,000 bail. He knew if he did not pay right away, he would spend the long holiday weekend in jail, and he would have to neglect his dry-cleaning business. He had enough savings to pay half of the $3,500 premium, and agreed to $200 monthly payments. The payments and lost savings caused his family considerable financial hardship. They had to cut back on expenses, and cancel a family vacation. It took many months to recover.
Hayward Police officers arrested Victor Lawrence when he refused to show them his identification at a DUI checkpoint on the Friday evening before Memorial Day in 2015. They took him to jail, where the watch commander set bail at $5,485. Inside the police station jail, Victor saw advertisements for various bail agents. There was no other information provided about the process of detention and release. His wife paid the non-refundable premium in cash from savings. If she had not, he knew he would have been stuck in jail until the next Tuesday. Unfortunately, the lost savings came at a bad time: Victor’s teaching contract was set to expire and his wife had just graduated from school. He had planned to spend the money on his three children that summer, but instead had to find ways to save. The prosecutor never filed charges, but Lawrence could not get his money back.
Some in law enforcement deny that bail harms poor people. In an August 2016 blog posting for the Association of Deputy District Attorneys, Michele Hanisee, president of the Association of Los Angeles Deputy District Attorneys, said: “Some could argue that the current bail system does not penalize the poor, it targets the rich.” She said judges sometimes raise bail amounts for wealthy defendants. Marc Debbaudt, president of the Association of Deputy District Attorneys, accepts it as a reasonable part of our system:
The inescapable reality of life, that some have more means than others, and that wealth has advantages over poverty, has never risen to the point where it constitutes a violation of our Constitution. Is wealth unfair? There is no mandate in the Constitution that all citizens must have equal financial status, or must enjoy the same opportunities that can be obtained through their financial abilities.
One California County district attorney agreed that the use of money bail punishes poor people. He said, “There are inequalities everywhere, but government shouldn’t use penal power in ways that discriminate on the basis of wealth.”
The pretrial detention system and the requirement of bail harms even those middle-income people who can manage to pay to avoid punishment before trial.
Kevin Ocampo’s cousin’s ex-wife accused his cousin of serious acts of domestic violence, leading to his arrest and prosecution for felony charges in Alameda County in February 2011. The ex-wife had made false accusations against a previous boyfriend, and had threatened to accuse the cousin of spousal battery if he filed for divorce. Kevin believed his cousin was innocent. The initial bail was $250,000. The bondsman charged Kevin and his cousin’s mother a 6 percent premium. But when they got to court, without an evidentiary hearing to test the truth of the accusations, the judge raised bail to $325,000, forcing them to gather a new premium.
It took them two months to gather the money, borrowing from Kevin’s wife’s 401K retirement plan and from friends and family, taking out a home equity loan, and putting payments on their credit cards. They cancelled plans for a new car and vacations.
The case dragged on for months, then years. The prosecutor offered the cousin a misdemeanor plea, but he refused. Finally, in January, 2014, the case went to trial. The ex-wife’s story, and the case, fell apart. The jury gave a “not guilty” verdict. But Kevin still owed payments to the bail bondsman. “Now you can see why I’ve lost faith in the justice system,” he said.
V. Does Bail in California Serve the Legitimate Purposes of Pretrial Detention?
The stated justification for pretrial custody is to protect the public and to prevent people from willfully refusing to come to court. These are legitimate goals. The question remains as to whether California’s pretrial detention system effectively achieves these goals, and whether they can be achieved without violating human rights.
Protecting Public Safety
San Francisco District Attorney George Gascon told Human Rights Watch money bail does not make the community safer, because people with means just pay and walk out. Rich people, people involved in organized crime, and high ranking gang members, all have access to money to get out of jail. For many in organized crime, bail is just another business cost.
Contra Costa Public Defender Robin Lipetzky wrote:
I have seen over and over again in Contra Costa County the stark reality that persons who are dangerous and wealthy are released pretrial with no conditions (by posting bond) while the poor who pose no danger remain locked up while their case is pending.
She cited an example of a wealthy man charged with murder following a petty argument, in which he was caught on videotape, who was able to bail out.
Since most people bailing out of jail are paying a non-refundable premium without putting up collateral and could not possibly pay the full bail amount anyway, they do not necessarily have a significant financial motive to avoid further crime, even assuming concern about future payment actually serves as a deterrent.
In a 2013 study, Dr. Michael Jones, from the Pretrial Justice Institute, compared defendants released on secured bonds (paid through a bondsman) with defendants released on unsecured bonds (own recognizance release, with a promise to pay if there is a violation of terms) in Colorado. The study controlled for perceived risk level of defendants and found no difference between them in public safety outcomes, defined as new crimes charged.
The bail industry disputes these conclusions. Melanie Ledgerwood, director of government relations for Accredited Surety and Casualty Company, Inc., an insurance company specializing in underwriting bail bonds, said: “The recidivism rate is almost twice as high for unsecured release vs. commercial bail.” Ledgerwood based this on data from a 2007 study using numbers from 1990-2004. However, the Bureau of Justice Statistics acknowledged data in this report is incomplete and risked being interpreted to draw unsupported conclusions. The report did compare re-arrest rates for people released on their own recognizance, unsecured bonds, and secured bonds, finding them very similar.
The California Constitution mandates “public safety” as the primary factor to be considered in the pretrial detention decision. It does not define “public safety,” or how it is to be considered, leaving judges almost entirely to their discretion—a discretion often influenced by institutional pressures to process cases rapidly. A person who commits a theft, uses drugs, or commits vandalism may be causing personal and societal harm, but generally is not threatening public safety. The public safety consideration should assess the likelihood of a person committing a violent crime while out of custody awaiting resolution of their case, not simply the possibility they may break a law.
Statistics from large urban counties in California show the likelihood of violent crime during pretrial release is extremely small:
- Only 1.2 percent of all people charged with felonies commit violent felonies during pretrial release.
- Those accused of violent crimes initially are most likely to commit a new violent offense, at a 2.9 percent rate, though least likely overall to commit a new felony.
- The overall rate of new felony offenses committed pretrial in California is 11.1 percent, slightly higher than the 10.1 percent rate for the rest of the country, which detains a lower percentage of people pretrial.
- Of the 11.1 percent of new felonies, over half (5.7 percent) are new drug offenses.
While proponents of the current system may argue the low numbers of violent crimes committed by people out of custody pending resolution of their case show the system effectively addresses public safety, there is no empirical proof. The lack of careful consideration of bail and release decisions in court means it is likely that courts release people who have a high risk of causing harm, while detaining vast numbers who do not.
The current system, relying almost entirely on the charge to determine bail amount, and the wealth of defendants to determine release or custody, does not target its assessment of risk with specificity. It generally defaults toward incarceration, punishing people for the possibility they will commit some future crime without specific evidence of a threat.
Failures to Appear
The other justification for setting bail and detaining people pretrial is to prevent them from missing court dates. Courts operate less efficiently if people do not appear for their proceedings. When people fail to appear for court, the judge issues a warrant, which requires some cost in processing and sometimes results in their arrest.
It is important to distinguish between two types of “failures to appear.” The first is the defendant who flees the jurisdiction or otherwise actively avoids appearing in court to escape possible consequences; the second is the person who fails to appear due to negligence or an excuse or situation, like being homeless or having a mental condition.
While it is valid to detain a person who is known to be an actual flight risk, it is problematic to lock someone up based on a probability they will miss a court date.
One judge from Alameda County, a former prosecutor, acknowledged the distinction, observing that the problem of missed court dates is not so severe. “They aren’t on the plane to Rio,” she said. Los Angeles County Deputy Public Defender Nick Stewart-Oaten said frequently defendants would miss misdemeanor court in the morning, but show up by the afternoon, after the judge has issued the warrant and entered “failure to appear” on their record. He described one man who had a warrant issued who came to court a few days later. His ticket cited him to the day he appeared. The court had mistakenly calendared him a few days earlier. Stewart-Oaten said some judges, in his experience, do not remove a mistaken “failure to appear” from a person’s record.
Edwin Molina’s Story
Edwin Molina was arrested for misdemeanor spousal battery. He stayed in jail for several days before his brother was able to help bail him out. He used all of his savings on the down-payment and was going into debt to pay the rest. But he got out and did not lose his job. When he left jail, the bondsman handed him a slip of paper with his court date, time, and location.
Edwin arrived at Division 12 of the Compton Courthouse on November 9, 2016 at 8:30 a.m., just as the paper instructed. He sat in the courtroom all morning, waiting for his name to be called. Meanwhile, in Division 48 of the downtown Criminal Court Building, no one answered when the judge called out Edwin’s name. The judge issued a warrant, forfeiting the bond.
When Edwin returned to Division 12 after the lunch break, he asked the bailiff when his case would be called. The bailiff and clerk figured out where he was supposed to be, and Edwin immediately got on the freeway and raced to the downtown courthouse. Luckily, the bailiff had not yet locked the courtroom door. Edwin was able to explain the situation to the judge, who recalled the warrant and reinstated the bond.
Human error causes a certain number of missed court appearances. Police officers writing citations, court employees processing the case, bondsmen, lawyers, bailiffs, and clerks can all make mistakes leading a defendant to miss a court date. Defendants themselves frequently make mistakes—they mis-schedule, lose return slips, or simply forget. San Francisco Deputy Public Defender Chesa Boudin told Human Rights Watch a single case can require many appearances, increasing the chances people may miss one. Often the problem is communication, since many defendants do not have phones.
People living in poverty are more likely to miss court dates. Homeless people have a variety of barriers that can increase the number of missed court dates. For example, they do not have a place to put personal property when they go to court, are more likely to lose or destroy citations or reminder slips, and generally live difficult and disorganized lives. Many are likely to have health problems, including mental health conditions, which prevent them from making appointments, including court appearances.
Captain Gary Newton of the Los Angeles Police Department (LAPD) Office of Special Operations estimated 40 percent of detainees in LAPD custody have a psychosocial disability. Derek Bercher from the Orange County Alternate Public Defender said a large percentage of “failures to appear” involve homeless people. Representatives of the Probation Department from Santa Barbara County said much of their pretrial jail population is homeless, have psychosocial disabilities, and have little chance of making bail.
While many people who miss court dates either come to court on their own within a short time or are quickly picked up by police, there are some who do not return to court. The Bureau of Justice Statistics report attempted to quantify this distinction by measuring “failure to appear” rates and rates at which people who did not appear did not return to court in one year.
The overall rate from 1990-2004 stayed between 20-25 percent, trending downward in the later years. Data shows people released on their own recognizance missed court dates at a rate of 26 percent; those released on bond missed court dates at 18 percent. The one-year rate was between 5-8 percent; 8 percent for own recognizance release; and 3 percent for bond.
Another study, looking at Bureau of Justice Statistics data for certain counties in California from 1990-2000, found bail bonds to be considerably more effective in guaranteeing return to court. However, this study did not account for the facts that bondsmen screen their customers for likelihood of return, including family support and pressure; make reminder calls and request people check in with them regularly; and often require the released person to submit to electronic monitoring or other surveillance.
Research shows reminder calls from the court or a pretrial services agency can have a big impact on reducing missed court dates. A study in one Oregon county showed a 41 percent decrease in “failures to appear” after a program along these lines was implemented; reminder calls in Coconino County in Arizona dropped “failures to appear” from 25 to below 13 percent; Fulton County, Colorado, reported increasing appearance rates from 79-88 percent. More substantive reminders, including explanations of the harms of missing a court date, have an even greater effect on increasing appearance rates.
A recent study analyzing costs and benefits to the pretrial detention system in two large urban counties found the most empirically relevant cost to pretrial release is “increased flight.” However, balanced against the impact of pretrial detention on employment and ability to generate income, as well as the public costs of detention, the study estimated a significant overall net benefit of pretrial release.
There is every reason to believe that if local and state governments invested in positive solutions, such as community-based support and services for people who are homeless and have mental health conditions, instead of using predominantly law enforcement solutions, the costs to courts of missed court dates would dramatically decrease.
If pretrial services oriented toward helping people get to court through reminders and assistance removing barriers to attendance, the net cost of “failures to appear” would likely go down.
There will always be a certain number of people who will not come back to court. However, enhanced criminal penalties exist to deter or punish those people. Pretrial detention punishes others just for the possibility they might miss a court date. As a rule, courts should strive to reserve pretrial detention only for those who have given serious reason to believe they will deliberately abscond. Courts should find alternative means of increasing appearance rates for individuals who are likely to fail to appear for other reasons.
Nelson Perez’ Story
Before October 2013, Nelson Perez was doing pretty well. He had worked in construction since he was 19 years old. He had saved his money, married, and had children. In 2012, he bought a five-bedroom house for his family.
Then trouble began. Nelson’s wife started using drugs and lost her job. She and Nelson split. In the fall of 2013, she told police that Nelson had raped her. He was 29, with no criminal record. When he heard police were looking for him, he went to the station to clear his name. When he got there, they arrested him.
Nelson spent his first night in Glen County Jail, terrified. When he appeared in court, he listened to the judge read the charges, shaking his head in disbelief. The judge set a high bail. Nelson did not have money to pay the fee on the bail. He lacked sufficient equity in his house. His family was poor and could not help him. His friends tried to work with a bondsman and to raise money to pay the fee, but the amount was too much.
Nelson thought the accusations were so clearly untrue that he would get out right away. He soon learned how wrong he was. The case was delayed for months. Meanwhile, he sat in a jail cell, waiting.
Meanwhile, his son lived in four different foster homes. He could not be with his two teenage stepchildren. He lost his home, truck, and other possessions.
Finally, after nearly two years, the case was set for trial. If Nelson were to go to trial and lose, he understood his sentence would be around 40 years in prison. As his trial date approached, the district attorney made an offer: time served, go home with probation and a conviction. It was tempting to take the deal, get out of jail, and not risk spending most of the rest of his life there. The district attorney, who had argued that Nelson was so dangerous that only an unattainable bail could protect the public, was offering to let him out in exchange for a guilty plea.
Nelson refused. He would not plead guilty to a crime he did not commit. He was willing to risk losing and getting a 40-year sentence. On the day of trial, the prosecutor dismissed the case. The district attorney said that not enough evidence existed to present the case to a jury.
When Nelson got out, he had nothing. His credit was ruined. He could not even get a cell phone plan. “I was proud of myself. I built my own house at the age of 29. Now, I have nothing,” he said.
VI. Profile-Based Risk Assessment
The decision to detain a person accused of a crime requires an exercise of discretion by a person in authority. Police officers, prosecutors, judges, even bondsmen, all make decisions that can impact whether the accused is detained or released.
Each decision maker evaluates or assesses the risk the accused will disobey the rules of release by either missing court or committing a new crime. Whether the system that guides this discretion is fair and respects the human rights of defendants depends on three crucial questions:
- Does the system of risk assessment require careful consideration of the individual’s circumstances?
- Is the system biased against one population or group of people?
- Does the system contribute to over-incarceration by pushing too many people into custody, supervision, or some form of surveillance?
The current system of risk assessment, with minimal hearings, vast judicial discretion based on little information, and using the offense charged as the primary proxy for risk through default to the bail schedule, fails on all three points.
Some judges find this system efficient, and many prosecutors enjoy the leverage custody gives them in court; bail bondsmen profit financially. But there is a growing number of stakeholders in the US and California criminal systems who agree the current regime is unfair and harmful. Many are proposing alternatives to replace or modify the bail regime.
Predicting the Statistical Likelihood of Pretrial Misconduct
The most prevalent alternative to money bail, promoted by court administrators, proponents of pretrial services agencies, academics, criminal justice think tanks and foundations, some advocates and lawyers, and others working in the courts, is to switch to the use of “actuarial” or profile-based risk assessment tools to help decision-makers determine who should be released and who should be detained pretrial.
These tools take data about the accused, feed it into a computerized algorithm, and generate a prediction of the statistical probability the person will commit some future misconduct, particularly a new crime or missed court appearance. The statistical probability is based on data about how other people with similar characteristics as the accused have behaved in the past. The tool usually expresses this statistical probability in a number score corresponding to a level of risk—low, high, or medium—and each score or category provides a recommendation to the judge: release, detain, or release with conditions, including pretrial supervision and monitoring.
A law enforcement staff person or official within a pretrial services agency or probation department generally inputs the information about the accused, which comes from court documents, criminal histories, and, depending on the tool used, personal interviews.
For the commonly used “Virginia” tool, the information includes name; race; sex; date of birth; charges; bond amount; charge type (misdemeanor or felony); other pending charges; whether there is an outstanding warrant; past criminal history; two or more “failure to appear” convictions; two or more violent convictions; whether they have lived one year or more at their current residence; employment status; and drug abuse history.
The “Ohio” tool considers the following factors: age at first arrest; two or more prior “failure to appear” warrants; number of “failure to appear” warrants in the past two years; three or more prior jail incarcerations; the number of prison incarcerations; employment status, including public welfare; drug use in the past six months; severe drug problems; and whether they have lived six months or more at their current residence.
A newer and increasingly used tool developed by the Laura and John Arnold Foundation (“Arnold”) does not require a personal interview and takes information entirely from the accused’s criminal history. It inputs age at current arrest; if the current offense is violent; if the current offense is violent and the accused is 20 years old or younger; if the accused has another pending charge; a prior misdemeanor conviction; a prior felony conviction; a prior violent conviction; a pretrial failure to appear in the past two years; a pretrial failure to appear older than two years; and a prior sentence to incarceration.
Some risk assessment tools claim to be “validated,” meaning they have been tested for accuracy, by comparing risk scores to actual results. Experts who have studied risk assessment tools say validation must be ongoing, as accuracy can change. The tools must be validated according to local conditions and local laws. But once a tool is implemented, it is difficult to validate since, for example, most high-risk defendants are detained and their probability of misconduct on pretrial release cannot be evaluated.
Where courts use risk assessment tools, judges maintain discretion to insert their own judgment and disregard the recommendation of the algorithm. Judges may choose to default to the recommendations of the tools, or to override them.
Use of profile-based risk assessment tools is increasing across the country. By May 2015, fifteen states had legislation authorizing courts to use them; six states required their use for all defendants. In California, according to a survey published in August 2015, 42 out of 58 counties used some kind of risk assessment tool, including the Virginia, Ohio, and Arnold tools described. There have been calls to make use of these tools mandatory throughout the state. The California Judicial Council has awarded grants to numerous California counties to study and implement them as part of pretrial services programs.
Proponents of profile-based risk assessment argue they will increase the number of people released from pretrial custody. Several studies analyzing pretrial outcomes in jurisdictions using profile-based risk assessments to guide release and detention decisions have concluded they release more people without increasing failures to appear and pretrial crime rates.
Arnold released a summary of data from the first six months of implementation of its tool in Kentucky, saying it increased the percentage of defendants released from 68-70 percent, and reduced the pretrial arrest rate from 10-8.5 percent. These are modest improvements, to the extent they are statistically significant, though Arnold billed the pretrial arrest rate as a 15 percent fall. The summary did not give figures related to failures to appear. Allegheny County reported a 30 percent fall in the number of defendants admitted into its jail after arraignment in the first 30 days after implementing a series of reforms, including a risk assessment tool.
Arnold issued a press release in August 2016 about implementation of its tool in Lucas County, Ohio, stating “the percentage of pretrial defendants released by the court on their own recognizance … jumped from 14 percent before the county began using [the Arnold tool] to almost 28 percent today.” However, overall release percentages during that time period actually dropped from 76.3 percent before implementation, to 64.3 percent after implementation; detentions increased from 15.7-19.1 percent; and guilty pleas on the first court appearance increased from 8-16.1 percent.
Some California county officials have reported improvements correlating to the use of risk assessment tools. Santa Cruz County reported significant savings in jail bed days, though some of those savings were offset by the cost of pretrial supervision, including electronic monitoring.
San Francisco adopted Arnold’s risk assessment tool in 2016. Prior to adopting the tool, its pretrial release programs were fairly successful, reporting over a 90 percent appearance and safety rate. Following implementation, San Francisco’s average daily jail population did decline, though Human Rights Watch found no publicly available evidence the Arnold tool was the cause. There have been similar declines nearly every year since 2007.
Proponents of profile-based risk assessment tools believe they produce more consistent and accurate risk scores than unaided human judgment. They argue using data analysis will help to ensure more dangerous people will stay in custody and less dangerous people will be released. Some assert their use will overcome judges’ personal biases and overt or unconsciously discriminatory decision-making. A further selling point, say proponents, particularly for tools not requiring personal interviews, is how quickly and efficiently they provide information. It is unclear whether these tools live up to such claims; many who have studied them point to their ineffectiveness and lack of fairness.
Judges Disregard Tools
In California, there is evidence that judges use profile-based risk assessment tools to support setting bail, but often disregard the tools when they recommend release.
The Santa Cruz County Probation Department reported judges followed the profile-based risk assessment tool’s recommendation in 68 percent of cases in 2015. Judges agreed with 84 percent of the “detain” recommendations, but just 47 percent of “release” recommendations. Concurrence discrepancies of this magnitude defeat the stated purpose of using the tools to decrease pretrial incarceration.
This pattern of judges overriding release recommendations is common. A court administrator from Alameda County said judges disregard release recommendations, setting bail for as much as 75 percent of all defendants determined to be “low-risk.”
Some proponents of profile-based risk assessment believe judges will get used to the recommendations and can be trained to follow them in ordering release. Some advocate for rules requiring judges to follow the recommendations, but this would deprive judges of the discretion they need to spare people from a bail recommendation when it is manifestly inappropriate. Judges are likely to resist surrendering their discretion.
Profile-Based Risk Assessment Tools Are Not Individualized
Profile-based risk assessments promise rapid decision-making, but provide only statistical predictions based on non-contextual information.
The tools analyze data in a binary or digital way. Most questionnaires ask “yes” or “no,” if the accused has a prior conviction or arrest, without details: were there mitigating circumstances? Was the crime situational or pathological? Did it happen in the remote past or as a recent pattern? Was there an extrinsic cause, like economic desperation? The questionnaires ask “yes” or “no,” if the accused has a missed court date on their record, without distinguishing between someone deliberately fleeing the jurisdiction or missing court because, for example, they could not find childcare.
One judge from Alameda County expressed her concern that risk assessment tools were too “pro forma” and lacked individual analysis. She said, as a judge, she wants to know about the individual, but acknowledges individualized assessment takes time.
Edwin Molina is a good example of the potential for injustice with non-contextual risk assessment. He went to the wrong court because his bondsman gave him the wrong information. He got to the right court the same day, but a failure to appear already had been entered on his record. The non-contextual tool will score this failure to appear the same as the failure to appear of Max Factor heir Andrew Luster, who famously skipped bail and fled to Mexico during a break in his Ventura County rape trial.
Kim McGill of the Youth Justice Coalition in Los Angeles, California, criticizes non-contextual, profile-based risk assessment as focusing on people’s pathologies (criminal histories, missed court dates), while excluding their strengths and needs. As such, she argues, the tools enable courts to treat individuals as commodities to be processed, without concern for the human and financial impact that the courts’ recommendations of jail or freedom have on individuals, families, and communities.
Because of their inability to factor context, these tools potentially jeopardize public safety. They do not account for specific dangers that a given defendant might present that are not reflected in the “yes” or “no” answers, and may not show specific threats made to a witness, escalating violence, or a personal vendetta.
Prosecutors and defense lawyers can add context to the risk assessment scores, as long as no rules require judges to obey the scores. However, given the pressure to rapidly process cases, particularly at the arraignment stage, it is more likely the use of risk assessment tools will make pretrial detention and bail hearings less substantive and more hurried than they are under the existing system. As judges tend to default to the bail schedule now, they will likely default to detention recommendations made by the tools.
Racial Bias and Profile-Based Risk Assessments
A major selling point for profile-based risk assessment tools is the belief that they are objective and remove judges’ personal biases from the pretrial detention decision. However, there is a growing criticism that these tools have their own implicit racial bias.
Racially Skewed Categories
Using data reflecting existing racial discrimination—including differences in education, employment, housing, and other socioeconomic factors—bias risk scores against black and Latino people. For example, if black people face discrimination in hiring and firing and have much higher unemployment rates than white people, a risk assessment tool that weighs employment status will score black people as higher risk. University of Minnesota professor of law and public policy, Michael Tonry, said:
[The] use of marital status, employment, education, family status, and residential stability as factors in prediction instruments systematically disadvantages minority defendants. The social and economic disadvantages that disproportionately afflict blacks and Hispanics in America are partly the products of historic and ongoing discrimination and bias. It should be at least discomforting that the use of socio-economic status factors in prediction instruments exacerbates those disadvantages.
One prominent study of a widely used profile-based risk assessment tool found it racially biased. The ProPublica study analyzed risk scores generated by the COMPAS tool for over 7,000 people in Broward County, Florida, in 2013 and 2014. This tool uses questions taken from the defendant’s criminal history and from a personal interview, including questions like, “Was one of your parents ever sent to jail or prison?” The study compared the scores to the numbers in each risk category who were re-arrested.
In addition to a high level of inaccuracy, the study described “significant racial disparities.” ProPublica researchers, controlling for criminal history, found COMPAS labeled black people as “high risk” at almost twice the rate as white people, while labeling white people as “low risk” more often than black people. The researchers said “high risk” white people did not reoffend 23.5 percent of the time, while “high risk” black people did not reoffend 44.9 percent of the time. They said “low risk” white people did reoffend 47.7 percent of the time, while “low risk” black people did reoffend only 28 percent.
Northpointe, the company that created and distributes COMPAS, disputes the findings of the ProPublica researchers and denies racial bias in its scores and questions.
A responsive study, led by criminologist Anthony Flores of California State University, Bakersfield, criticized the ProPublica methodology and conclusions. The study found the tool predicted re-arrest about as accurately for white people as for black people within their risk category. It reported black people having a significantly higher recidivism base rate: “Racial differences in failure rates across race describe the behavior of defendants and the criminal justice system, not assessment bias.”
Can Assessments Eliminate Their Biased Prediction?
Many proponents and critics of profile-based risk assessment acknowledge that any given tool is probably not capable of entirely removing bias from the system. Proponents of risk assessment argue that they can at least help reduce the influence of such bias.
Michael Jones of the Pretrial Justice Institute, an organization supporting what he calls “legal and evidence-based pretrial practices, such as the use of risk assessment tools,” told Human Rights Watch “that proper development and ongoing testing of pretrial risk tools can help assure they do not exacerbate racial disparities that may exist in the criminal justice system.”
Complicating efforts to assess discriminatory effect is that these tools use proprietary formulas, not open to independent scrutiny. For example, the ProPublica study noted that Northpointe shared a basic description of its formula, but did not publicly disclose specific calculations, making it impossible to determine the cause of any racial disparities.
Reinforcing Racial Discrimination in the Criminal System: Criminal History in Risk Assessment Reflects Racial Bias in Law Enforcement
Even without analyzing the specific mechanics of a given tool, a serious racial critique of profile-based risk assessment remains. Some people see the primary problem as being bias in the criminal system, which the assessment tools merely reflect. “Perhaps, what looks to be bias is not in the tool—it’s in the system,” criminologist Anthony Flores told the Washington Post. Others see the tools entrenching and enhancing that bias, while covering it with a veneer of scientific objectivity.
Professor Bernard Harcourt of Columbia University School of Law has studied the historical and current use of profile-based risk assessment tools in the US criminal system. “This is not the first time we have been tempted to use a metric of dangerousness as a way to empty [custodial institutions],” he notes. He describes how these risk assessment tools have now evolved away from explicitly racial questions, while reducing the number of factors considered, and focusing increasingly on criminal history. The Arnold tool is a perfect example of this evolution, as it asks a small number of exclusively criminal history-related questions.
However, as Harcourt notes, this focus on criminal history is driven to a large degree by racial bias in law enforcement and the criminal justice system as a whole: “Unfortunately, reliance on criminal history has proven devastating to African American communities and can only continue to have disproportionate impacts in the future.” He cites the steadily increasing proportion of non-white prisoners in jails and prisons throughout the country.
Police enforcement strategies and deployment policies, the crucial first decision point in the pretrial detention system, greatly impact who is arrested, which in turn impacts who is convicted and establishes the criminal history that affects future risk assessment scores.
Implementation of New York City’s “stop and frisk” policy illustrates the racial component. From 2002-2013, the New York Police Department engaged in a strategy of detaining massive numbers of people and using those stops to generate reasons to search for weapons or drugs. The vast majority detained had not committed a crime. Over 80 percent of them were people of color; over 50 percent were black. Because police resources focused on black people disproportionately, they were exposed to higher incidence of arrest and conviction.
Racially Biased Data Generates Racially Biased Outcomes
If arrest and conviction data is racially biased, then the profile-based risk assessment tools that use the data to make “evidence-based” decisions about who stays in jail and who gets released will generate racially biased outcomes.
Racial bias in the system, which makes black and Latino people more likely to be arrested, will tend to confirm the initial biased decision because statistically more black and Latino people identified as high risk will suffer subsequent arrest. According to Laurel Eckhouse of the Human Rights Data Analysis Group,
Thus, the tool will falsely appear to predict dangerousness effectively, because the entire process is circular; racial disparities in arrests biases both the prediction and the justification for those predictions.
Even if these profile-based risk assessment tools reduced the number of people held in pretrial detention—an unproven hypothesis—overreliance on them could seriously damage the legitimacy of California’s criminal system by reinforcing existing racial biases.
Profile-Based Risk Assessment Tools Can Be Used to Increase Jail and Supervision Populations
Defining Risk is a Policy Decision
Santa Cruz County adopted the Arnold tool, known as PSA-Court, in July 2014. The tool recommended release, detention, or release on supervision based on risk scores. The pretrial services section of the Probation Department administered both the implementation of the tool and the supervision of those released. Supervision included electronic monitoring and alcohol monitoring. In its reports, the Probation Department labelled people on pretrial supervision as within the “average daily population,” or “ADP.”
The report for 2015 detailed a decrease in the ADP in the first half of the year, followed by a modest increase in the next half. However, “following modifications of the PSA-Court decision making framework, in the first quarter of CY2016 saw a dramatic rise of the ADP—almost double of previous years.”
They adjusted the “decision making framework,” which is their “guidelines” for using the risk scores to make the release or detain decision in order to greatly increase the number of people placed into supervision. The potential for this kind of tinkering greatly undermines risk assessment tools’ pretense of rigor and objectivity.
Defining risk is subjective. Policy makers decide to define a certain level of risk as high or low. The county implementing the tool must decide if these risks justify imposing bail or other conditions, and at what level. Additionally, counties can add overrides to the tools, saying, for example, that certain crimes will not be subject to the assessment at all and will be flagged for a detention recommendation only. Who makes these decisions, and how transparent and inclusive of community stakeholders the process is, will influence whether they can be made fairly.
It is easy to envision a policymaker deciding the tool is not recommending enough people be placed on supervision or detained, and adjusting it to assign more people to a higher risk category. Risk assessment tools, like any other tool, depend on the will and needs of the person operating them.
A risk assessment tool can add to the numbers of people on conditions of supervision or jailed as easily as it can increase releases from detention.
Many in the movement to reform the pretrial detention and bail systems see increased pretrial supervision as the answer. Pretrial Services Agencies, often branches of local probation departments, provide various supervision options, including simple reminder calls; recommendations for helpful services, like drug programs and mental health treatment; severely intrusive oversight, like electronic monitoring; and probation-like reporting requirements. Usually, the monitoring comes at a financial cost to the supervised defendant and is often administered by private, profit-motivated companies. Some studies tout the effectiveness of supervision in reducing pretrial crime and increasing court return rates. Other studies have found supervision makes little difference in pretrial success rates, particularly for lower risk people, or in preventing new pretrial crimes. In one study, electronic monitoring did not improve results for high-risk people and correlated to significantly increased failure rates for low-risk people.
Adding supervision and conditions of pretrial release, especially electronic monitoring, may make judges more comfortable with ordering release and have an immediate effect of lowering pretrial jail populations. However, excessively stringent release conditions can result in large numbers of technical violations, which lead people back into custody. There is a danger that judges will default to the use of electronic monitoring, using it even in cases for which they might otherwise release without conditions. Replacing pretrial incarceration with electronic monitoring may still result in significant infringements on liberty, particularly in minority communities that receive disproportionate police enforcement.
Profile-Based Risk Assessment Tools Can Be Adapted to Fill or Empty Jails
Much of the impetus behind the current move to reform California’s pretrial detention and money bail system appears to be the administrative necessity to manage overcrowded jails.
As of 2001, 19 California counties had court orders capping the populations of their local jails, including Los Angeles, San Bernardino, Sacramento, Fresno, San Diego, and many smaller counties. As of 2014, 39 of 119 jails, housing 65 percent of all jail prisoners throughout the state, were under these caps, with more lawsuits expected.
In 2011, the state legislature passed AB109, “Public Safety Realignment,” in response to a court order to lower state prison populations. Realignment shifted prisoners serving time for non-violent, non-serious, non-sex related felony offenses from penitentiaries to county jails and shifted responsibility for technical parole violations to counties. This shift exacerbated the already bad overcrowding in California county jails. Statewide jail populations rose from 71,293 in the quarter before implementation to 82,527 by the first quarter of 2014. Still, pretrial prisoners have remained over 60 percent of the total jail population.
Counties have looked to pretrial release in response to jail overcrowding, particularly since Realignment.
Another response to overcrowding, despite an over 20-year trend of declining crime rates, is to build more jails. In 2007, well before Realignment, the California legislature passed AB900, which allocated $1.2 billion for new jail construction; in 2012, it passed SB1022, which added an additional $500 million.
Between these two allocations, the state funded 11,989 new jail beds for 36 counties. Since then, the state has passed SB863 (2014) and SB844 (2016), which have added a combined $770 million for jail construction projects. Excluding SB844 projects, 42 of California’s 58 counties are in the planning or construction phase of jail building or have just completed projects. Local jurisdictions are adding funding to jail construction.
This massive investment in new jails will greatly expand capacity to imprison people during the pretrial period. The danger of profile-based risk assessment tools is that they project the appearance of being objective and scientific―“evidence-based”―but can be used to meet the political and administrative needs of those who control them, without regard to whether or not their use achieves justice. If, with jail expansion, those needs shift to having more people in custody, profile-based risk assessment will serve efficiently.
VII. A Better Way: Increased Cite and Release and Individualized Risk Assessment
The California Constitution, international human rights law, and commonly held values of fairness all hold that government should not take away a person’s freedom without due process, including a fair hearing.
The current system of pretrial detention does not provide that fair hearing. Any change that does not fix this fundamental flaw is not true reform. Though the current reform debate centers on either keeping the old system of money bail or moving to profile-based risk assessments, neither address the lack of a fair hearing. But they are not the only options.
California needs to limit the number of defendants in jail who should never have been there in the first place. As discussed, the state jails people who are innocent, who never have charges filed against them, and who are too poor to afford bail. This needs to change.
A first step is to reduce the total number of accused persons who may be held in pretrial custody, so that the system can focus its resources on individualized risk assessment for those who truly need careful assessment. This could be achieved by requiring cite and release for all misdemeanor and non-serious/non-violent felony arrests, with only a few narrowly tailored exceptions required for public safety.
For the much-reduced number of accused persons who might still warrant pretrial detention, courts can turn to an individualized risk assessment that takes into account the context of the alleged crime and the defendant’s personal circumstances. Federal courts already use many basic elements of this individualized context-based approach.
The centerpiece of the individualized, context-based approach is the requirement for what Justice Rehnquist called in US v. Salerno a “full blown adversary hearing.” The hearing requires the right to counsel, which should begin immediately upon arrest. The accused’s attorney would gather information about their circumstances through privileged, confidential conversations, thus promoting thorough and open inquiry rather than data-driven profile information gathering. Defense attorneys would need to invest time and energy in these hearings. Public defenders, who handle most criminal cases, would need more resources to properly defend the hearings, including gathering and verifying mitigating information.
In some jurisdictions, public defender offices are already experimenting with “bail units,” in which they assign additional attorneys, sometimes adding social workers and investigators, to prepare mitigation information for the bail hearings. Given the proven difference being out of custody makes in the future results of a criminal case, this effort at the beginning of the case should greatly improve overall representation, particularly considering the need to do mitigation preparation at some point in the case anyway, as it is essential to successful plea negotiations or advocacy at sentencing hearings.
The district attorney should have to affirmatively initiate the bail hearing through a written notice advising the accused of the evidence to be used against them. Besides the fairness of knowing evidence against the defendant, the notice requirement and subsequent hearing would compel prosecutors to prioritize defendants they believe need to be detained before trial.
The procedure should include an evidentiary hearing about the facts of the case, from which the judge would determine if there is sufficient evidence of a crime to justify pretrial detention through setting bail, and which would allow the judge to weigh the actual seriousness of the crime and the strength of evidence against the accused.
While this hearing would not be dispositive of the case (even a failure of proof would not result in dismissal as it would in the subsequent preliminary hearing), it would give the court and the parties a better sense of the case for settlement purposes. Federal courts already hold these probable cause detention hearings, as do California juvenile courts.
The adversarial hearing would allow the accused the right to testify, even if they might rarely exercise it, the right to present witnesses and evidence, including the right to introduce evidence by proffer, and the right to cross-examine opposing witnesses. If necessary, the accused would have an opportunity to explain the circumstances of any missed court date or any prior criminal convictions, in contrast to the profile-based risk assessment tool that merely scores the presence of these failures. The accused would get to present personal mitigating facts about their history, circumstances, strengths, and needs; the prosecutor would be able to present actual, not statistical, aggravating factors about the case, the individual, and any threat to safety or of actual flight they may present.
There should be a presumption of release and, as in the federal context, a requirement the prosecutor show “clear and convincing evidence.” Unlike profile-based risk assessment, which focuses on statistical likelihoods, individualized context-based risk assessment would require “evidence that an arrestee presents an identifiable and articulable threat to an individual or the community.”The standard should distinguish between actual risk of flight or evasion of justice, as opposed to a statistical likelihood of missing a court date.
If the evidence overcomes the presumption of release, the judge should be required to generate a non-boilerplate, written record explaining the specific danger and evidence supporting the conclusion of setting bail. The record, which should be immediately available, must be sufficient for review by a different court. That review should occur within a few days of the initial decision. This requirement of notice of the reasons for the decision contrasts with the lack of transparency of profile-based risk assessment tools.
Profile-Based Risk Assessment Should Not Be Part of the Decision
Some within the movement to reform bail suggest a limited role for profile-based risk assessment tools within the context of a more extensive pretrial detention hearing. Rather than relying strictly on recommendations of the tool, courts would use it as one piece of data to add to their consideration in the individualized, adversarial proceeding. Lawyers on each side would add context; the judge would take all factors into consideration.
There are four major problems with this approach:
- First, to the extent the profile-based risk assessment tools are racially biased, they should not be a factor at all.
- Second, given that the tools simply produce a score without explanation of how it was arrived at, it would be impossible for judges to understand how the contextual information that lawyers provided related to the score itself.
- Third, even if judges are supposed to consider other factors, their most likely course will be to default to the risk assessment score, particularly when it points to detention due to the pressure to process cases quickly. Many judges, striving for efficiency and the natural desire to get their work done, coupled with institutional pressure to favor pretrial detention, are likely to shorten hearings and fall back on the risk score to justify their decisions.
- Fourth, defense attorneys could resign themselves to the futility of opposing the risk score, as they do now under the schedule, and litigate the hearings less and less. Judges will follow the current trend of disregarding release recommendations.
While the compromise of a limited role for these tools sounds appealing in theory, the reality of the current court system that values rapid processing of cases over careful, detailed decision-making, could result in the tools simply replacing bail schedules as an efficient justification for detaining people pretrial.
The system Human Rights Watch proposes admittedly involves significant changes in California courts’ approach to administering justice. The detailed, lengthy hearing process would either require a massive infusion of resources or would slow the judicial process down considerably, if grafted onto the current system in piecemeal fashion. To work effectively, the hearing process depends on the widespread use of “cite and release,” in the first instance, to greatly limit the numbers of in-custody defendants who require such hearings. Prosecutors and judges would have to re-orient their use of pretrial custody so that it strictly addresses community safety and identifiable flight risk, and avoids using large-scale detention as leverage to pressure rapid guilty pleas. A system with such added individualized procedures may also point the way toward re-orienting law enforcement priorities away from criminalization of various social problems like drug use and mental illness.
While this proposed system is more resource-intensive per individual than the current approach to bail-setting or the proposed use of profile-based risk assessment tools, it will mitigate the income-based discrimination that exists in the current system of money bail. It will decrease the number of people, particularly innocent people, coerced into pleading guilty because of their custody status. It will save the public money by cutting jail costs. It will honor the presumption of innocence and treat people going through the courts as human beings, not numbers. It will increase the quality of justice in California.
IX. International Human Rights Law
International human rights law permits the use of pretrial release with conditions, including money bail. However, any pretrial restrictions must be consistent with the right to liberty, the presumption of innocence, and the right to equality under the law. California’s system of money bail does not meet these conditions.
Article 9(1) of the International Covenant on Civil and Political Rights (ICCPR), which has been binding on the United States since 1992, says: "Everyone has the right to liberty and security of person." A person’s liberty may not be curtailed arbitrarily, either through arbitrary laws or through the arbitrary enforcement of the law in a given case. To comply with Article 9, "deprivation of liberty must be authorized by law" and "must not be manifestly unproportional, unjust or unpredictable."
The UN Human Rights Committee, the expert body charged with interpreting the ICCPR, has rightly observed that "'[A]rbitrariness’ is not to be equated with ‘against the law,’ but must be interpreted more broadly to include elements of inappropriateness, injustice and lack of predictability. This means that remand in custody pursuant to lawful arrest not only must be lawful but also reasonable in the circumstances." In Human Rights Watch’s view, some of the practices described in this report—the deliberate use of prohibitively high bail to help coerce guilty pleas and the reflexive use of bail schedules that make it difficult for low income defendants to secure their release— constitute arbitrary detention.
Pretrial detention also implicates the presumption of innocence, affirmed in Article 14 of the ICCPR as one of the necessary guarantees for a fair trial. This principle is given added practical resonance by the extensive evidence, discussed above, that pretrial release dramatically enhances a defendant’s ability to prepare a competent defense.
Article 9(3) of the ICCPR explicitly addresses pretrial detention, saying: “It shall not be the general rule.” Article 9(3) authorizes pretrial release subject to guarantees, which may be in the form of money bail or other assurances. When concerns about flight risk or safety require some conditions on pretrial release, non-custodial measures should be used when possible rather than pretrial detention, which should be "a means of last resort."
International treaty bodies and authoritative interpretations of article 9(3) are uniform in the view that, while pretrial detention is permissible under certain circumstances, it should be an exception and as short as possible, and should be proportionate to the maximum potential sentence. Of critical importance is the limitation on imposing pretrial detention for offenses that are not punished with custodial sentences. "If imprisonment is not to be expected as punishment for a crime, every effort should be made to avoid pre-trial detention."
Seriousness of a crime is not in and of itself justification for pretrial detention. According to authoritative interpretations of these standards, pretrial detention should be limited "to essential reasons, such as danger of suppression of evidence, repetition of the offence and absconding…." The apparently widespread practice of California prosecutors arguing that bail is necessary because a defendant is dangerous, only to then press them to accept a plea deal that results in immediate release, contravenes this idea and works hand-in-hand with efforts to use pretrial detention to coerce defendants into pleading guilty.
The Inter-American Commission on Human Rights (IACHR), the body responsible for interpreting the American Convention on Human Rights, which the United States signed in 1977, maintains that authorities making bail decisions or otherwise deciding on the pretrial custody of individuals should not make such decisions automatically by reference to simple formulas, patterns, or stereotypes; rather, the IACHR argues, pretrial detention must be grounded in reasoning that contains specific, individualized facts and circumstances justifying such detention. The IACHR also maintains that careful deliberation must also be practiced when reviewing specific requests for bail reduction and release, and that the lack of such deliberation can render a detention arbitrary.
The reliance of California’s police, prosecutors, and courts on charge-based bail schedules is too automatic and too often without careful deliberation. Similarly, the growing reliance in California on profile-based risk assessment risks reducing the decision-making process to simple formulas or stereotypes, without looking at specific facts and circumstances.
Lack of meaningful effort to tailor the bail amount to a defendant’s financial circumstances means that California’s system of money bail and pretrial detention discriminates against low-income individuals, including people of color. Equality among people may be "the most important principle imbuing and inspiring the concept of human rights." Article 26 of the ICCPR establishes that all persons are equal before the law and entitled to equal protection of the law. ICCPR article 2(1) prohibits discrimination in the context of the rights and freedoms enumerated in the convention, such as the right to liberty.
The monitoring committee of the International Convention on the Elimination of All Forms of Racial Discrimination (ICERD), which the United States ratified in 1994, urges state parties to ensure that the "requirement to deposit a guarantee or financial security in order to obtain release pending trial is applied in a manner appropriate to the situation of persons in vulnerable groups, who are often in straitened economic circumstances, so as to prevent the requirement from leading to discrimination against such persons."
As demonstrated by the data analysis provided in this report, California’s system of pretrial detention has the effect of detaining large numbers of people of color, disproportionate to their population in the state. This data is particularly troubling given that racial groups in the United States have relatively equal rates of criminality.
Prohibited discrimination includes any rule, regulation, or distinction that has the purpose or effect of impairing equal enjoyment of any rights or freedoms. A law or practice may not have a discriminatory intent, but will violate human rights law if it has an unjustifiable or unreasonable disparate impact against a protected group.
To ensure that California’s system of pretrial detention reduces the number of people needlessly held in custody while increasing fairness and respect for the rights of pretrial defendants, Human Rights Watch offers the following recommendations.
To California Lawmakers
- Enact legislation that limits the use of pretrial detention by amending the California Penal Code to require “cite and release,” instead of arrests, for all misdemeanors and non-serious/non-violent felonies, with only narrow exceptions. This modification would include amending the existing “cite and release” law to remove vague and expansive exemptions and replace them with specific requirements of actual danger or demonstrably high flight risk to justify arrest.
- Enact legislation that requires courts to engage in a fact- and context-based, individualized system of risk and needs assessment, with robust procedural protections to guarantee fair hearings and a better understanding of the facts and circumstances surrounding each case.
- Enact legislation establishing enforceable standards for setting bail or detaining pretrial, requiring release absent significant proof of a specific danger to the community or specific risk of evasion of court process.
To California Lawmakers and County and City Officials
- Dedicate greater resources to addressing homelessness, drug abuse, and mental health conditions, while developing effective community-based services and responses to these issues. Consider whether law enforcement resources currently dedicated to addressing these issues could be more appropriately deployed.
- Dedicate greater resources to public defense to enable defenders of indigent clients to adequately and capably represent them at all phases of their cases, particularly at the crucial stage of detention and bail determination. Resources should account for support staff, including paralegals and social workers.
An Individualized System of Risk and Needs Assessment
- In order to enact a fact- and context-based, individualized system of risk and needs assessment, Human Rights Watch recommends California lawmakers enact legislation enabling an adversarial pretrial detention hearing system, with effective representation, procedural safeguards, a presumption for release, and judicial accountability.
California Lawmakers Should Enact Legislation Providing for:
- Presumption of release for all cases: The court should start with the presumption the accused will be released without conditions, unless and until the prosecutor presents sufficient evidence to warrant setting bail or imposing other conditions.
- Notice requirement to initiate the hearing: The prosecutor must file a formal, non-boilerplate written notice, including a description of evidence supporting the request to set bail or detain.
- Time limits for detention hearings: Detention hearings must occur within the time limits of the arraignment, unless the defendant requests a continuance.
- Probable cause hearing: The detention hearing must include a brief probable cause hearing with taking of testimony and cross-examination so the judge understands the seriousness of the offense and defendant’s level of culpability.
- Specific and known risk standard for setting bail: The legal standard required to set bail or detain should be that there exists a specific and known risk the defendant will cause harm or will attempt to evade the court’s authority. Absent proof of such specific risk, no bail or detention should be allowed. Judges should not rely on statistical probabilities or vague generalities about potential danger or flight risk to justify the imposition of bail or other conditions. Disability should not be considered a risk factor. Potential to commit a non-violent offense ordinarily should not constitute a specific risk of harm, absent aggravating circumstances.
- Evaluation by courts of the accused’s needs and abilities: Judges should consider public safety first, but should include in the public safety calculation the needs of the defendant and their dependents, and the defendant’s individual circumstances.
- Ability to pay hearing: Part of the hearing should include an assessment of the defendant’s financial resources and ability to pay an amount of bail, so the judge can carefully determine the least amount required to ensure appearance in court. Courts may not set bail in an amount above a person’s ability to pay.
- Written record supporting imposition of bail: The judge should produce a non-boilerplate written record explaining the reasons for setting bail and justifying the chosen amount of bail, so the order will be understandable to a reviewing court.
- De novo review by a different court: Within a reasonably brief time after an order setting bail, the defendant should have the right to review the order based on evaluation of the original judge’s written record.
- Adversarial hearings: All defendants should have the right to capable counsel, including appointed counsel, immediately after arrest and for any hearing related to pretrial detention. These hearings must be truly adversarial, allowing both sides the right to present and impeach evidence.
- Limitation of restrictive release conditions: Any order for supervised release that substantially restricts a person’s liberty, like electronic monitoring, drug testing, or reporting requirements, should be justified by a specific factual finding of the need for that restriction, made on the record and reviewable.
- Ensuring release conditions are not a financial burden: Release with conditions like electronic monitoring or probation-like supervision, to the limited extent they are used, should not be contingent on the defendant’s ability to pay for them.
- Fair allocation of resources for hearings: The state should sufficiently fund public defenders and other appointed counsel to provide capable representation in these hearings.
California Lawmakers Should Not Promote the Use of Statistical Predictions
- Do not replace the system of scheduled money bail with a profile-based algorithmic system of risk assessment that relies largely on statistical probabilities to make detention or bail decisions.
To Public Defenders’ Offices
- Dedicate sufficient resources to litigating detention and bail hearings, including providing training, creating specialized units, assigning additional attorneys, investigators, paralegals, and social workers to fully prepare for and present individualized mitigation information at the hearing. State and county governments must fund these allocations of resources sufficiently to guarantee capable representation.
Absent implementation of the more structural changes outlined above, Human Rights Watch makes the following interim recommendations.
To Police Departments
- Ensure that every person arrested has easily understandable information about their rights within the criminal system, particularly concerning pretrial detention and release. This should explain procedures by which a person can be detained or released.
- Establish a procedure by which defendants and their families can make a refundable 10 percent deposit for their bail to secure release. This deposit would avoid the bail bondsman altogether. The court may take a reasonable processing fee.
- Refrain as much as possible from reliance on the bail schedules and instead tailor bail amounts to individual circumstances and ability to pay of the accused.
- Allow defendants a brief grace period before issuing warrants for failures to appear in court.
To State and Local Government Officials
- Engage in uniform and systematic collection of data concerning detention and release decisions and court outcomes, to better understand the relative fairness of the court system and guard against racial, economic, and other biases.
- Ensure that pretrial services departments are not part of their probation departments and are not tasked with supervising and surveilling people released pretrial. Pretrial services should be oriented toward helping people get to court and access other useful, voluntary services, like mental health counselling or drug treatment.
- Ensure that bail schedules are limited to use as a guideline for setting bail at the police station for the few arrestees not cited and released, and not as a guideline for judges to default to during court hearings.
- Ensure the creation of bail schedules is transparent and allows input from a variety of stakeholders, including defense lawyers and representatives of communities most impacted by the criminal system. Create standards requiring the amounts be calculated in a way to guarantee court appearance without being prohibitively high.
- Lower the bail amounts in bail schedules. Judges setting amounts based on the hearings and standards described above should not use their discretion to set unaffordable bail.
- Ensure people with disabilities, including mental health conditions, have access to reasonable accommodation at every stage of the process. This accommodation should be individually tailored to the specific requirements of the person concerned.
- Enact careful oversight of the bail industry to ensure bondsmen are not engaging in abusive or predatory practices.
While Human Rights Watch opposes the use of statistical predictions and profile-based risk assessment to guide pretrial detention decisions, we recognize that many jurisdictions inside California as well as other states have already implemented these tools with little to no regulation on their use. Therefore, in order to mitigate the harms that may come from their use, and until they can be discontinued and replaced with detailed, individualized, adversarial hearings, Human Rights Watch recommends the following limitations on risk assessment tools.
California Lawmakers Should Enact Legislation:
- Requiring complete public transparency as to any algorithm or formula used, and detailed information on the sources of underlying data used to make all risk assessment calculations.
- Prohibiting counties or the state from using any tool that claims to be proprietary and refuses to disclose information on that basis.
- Requiring local jurisdictions implementing these tools to create a mechanism for public oversight of any calibrations or adjustments to the valuation of risk, including complete public transparency for any decisions about calibration and adjustment.
- Mandating frequent, comprehensive, and publicly transparent audits of the use of risk assessment tools to ensure they are not increasing or exacerbating racial, ethnic, gender, or any other bias within the criminal justice system.
- Mandating discontinuing the use of risk assessment tools if they are found to be increasing or exacerbating bias.
- Providing a legal mechanism for detainees to challenge the individual findings of the risk assessment tool, including challenging the data input, the profile data relied on to make the prediction, and the formula used to make the prediction. Legal challenge must be available within a short period of time, with a remedy of release from custody if the tool’s prediction is unsupported. Additionally, detainees should be provided a cause of action to sue the local jurisdiction and any private company administering or providing the tools for damages for wrongful detention based on an unsupported prediction.
This report is based on investigations and research conducted at Human Rights Watch by senior advisor Jamie Fellner and senior researcher John Raphling. Brian Root, quantitative analyst at Human Rights Watch, conducted all data analyses. Alison Parker, US Program co-director, and Danielle Haas, senior editor for the Program Department, edited the report. Christopher Albin-Lackey, senior legal advisor, provided legal review and edits. Maya Goldman, US Program associate, provided research and production assistance. Danica Rodarmel, from the University of California Berkeley School of Law, and Jacob Denney, Human Rights Watch graduate intern, provided research assistance. Fitzroy Hepkins, administrative manager, Jose Martinez, administrative senior coordinator, and Olivia Hunter, photography and publications associate, coordinated layout and production.
Human Rights Watch would like to thank the Policy Advocacy Clinic at Berkeley, directed by Jeffrey Selbin, and particularly law students Danica Rodarmel, Mel Gonzalez, and Da Hae Kim, for their research assistance and for their interviews of people impacted by the pretrial detention and bail system.
Human Rights Watch is grateful to the many professionals and experts, including defense lawyers, prosecutors, members of law enforcement, judges, probation and pretrial services officials, courts administrators, academics, policy analysts, bail bondsmen, and community organizers who spoke to us about the pretrial detention and bail system. We would like to thank the members of the defense bar and community organizations, particularly Los Angeles Community Action Network (LA CAN), Youth Justice Coalition (YJC), and People Organized for Westside Renewal (POWER), who introduced us to people directly impacted by California’s system of money bail and pretrial detention and who helped us understand the ways that the system impacts their communities and clients and what solutions might work. Human Rights Watch especially would like to thank the many individuals who courageously shared their experiences with bail and pretrial detention with us.
 Danielle Kaeble and Lauren Glaze, “Correctional Populations in the United States, 2015,” US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics, December 2016, http://ift.tt/2imogOb (accessed March 28, 2017), p. 2.
 Ibid., p. 4.
 Ibid., p. 1.
 Institute for Criminal Policy Research and Birbeck University of London, “World Prison Brief,” undated, http://ift.tt/1povA4N (accessed March 28, 2017). The US is first in overall population, leading China, its next closest competitor, by over half-a-million prisoners. However, the data quality on prison populations from China is highly questionable. It is in second place for rate of imprisonment behind only the Seychelles. This website had a much lower prisoner population rate per 100,000 than the BJS number, which would have put the US in first place for rate of imprisonment.
 Danielle Kaeble and Lauren Glaze, “Correctional Populations in the United States, 2015,” p. 12.
 Crime in California, 2015, California Department of Justice, California Justice Information Services Division Bureau of Criminal Information and Analysis, Criminal Justice Statistics Center, p. 6. Property crime rates declined similarly over this same time period. The violent crime rate in 2015 is a slight increase from the 2014 low of 393.3.
 Leah Sakala, Breaking Down Mass Incarceration in the 2010 Census: State-by-State Incarceration Rates by Race/Ethnicity, Prison Policy Initiative, May 28, 2014, http://ift.tt/2gWfxy8 (accessed March 28, 2017).
 According to Department of Justice statistics, black people are arrested at a much higher rate than those of other races. Bernadette Rabuy and Daniel Kopf, Prisons of Poverty: Uncovering the pre-incarceration incomes of the imprisoned, Prison Policy Initiative, July 9, 2015, http://ift.tt/2jA0ZFj (accessed March 28, 2017). Other research shows black people subjected to traffic stops more frequently, or searched more frequently with no more likelihood of finding contraband, and booked into jail more often. Stephen Bingham et. al, Stopped, Fined, Arrested: Racial Bias in Policing and Traffic Courts in California, East Bay Community Law Center, April 2016, http://ift.tt/2aTQ7Ta (accessed March 28, 2017); Clifton B. Parker, “Stanford big data study finds racial disparities in Oakland, Calif., police behavior, offers solutions,” Stanford News, June 15, 2016, http://ift.tt/1PuEJpJ (accessed March 28, 2017).
 Bernadette Rabuy and Daniel Kopf, Prisons of Poverty: Uncovering the pre-incarceration incomes of the imprisoned.
 Brian A. Reaves, “Felony Defendants in Large Urban Counties, 2009 – Statistical Tables,” US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics, December 2013, http://ift.tt/2oyyOw7 (accessed March 28, 2017), pp. 1, 14.
 Human Rights Watch analysis of US Department of Justice, Bureau of Justice Statistics, “State Court Processing Statistics: Felony Defendants in Large Urban Counties,” 2000-2009, http://ift.tt/2oyiOKQ (accessed March 28, 2017).
 Human Rights Watch analysis of California Board of State and Community Corrections (BSCC) data. BSCC publishes data from its monthly surveys. Counties provide the average daily populations for the month for sentenced and unsentenced jail populations. Human Rights Watch analyzed data from all 58 counties for 2014 and 2015.
 Human Rights Watch analysis of Sacramento County Sheriff’s Department data. The calculation used the $113.87 per day cost estimate from the Public Policy Institute of California. The estimate of $44.3 million is an underestimate because it only includes arrests without warrants, violations, or holds. Sacramento County is the only county that provided data on conviction dates which allow for the calculation of time in jail from arrest to release for non-sentenced releases and arrest to conviction for sentenced releases.
 Human Rights Watch telephone interview with Nathan Ramos, Los Angeles, October 2016.
 The LAPD jail next to its main headquarter is referred to as “the Glass House.” Police hold arrestees there until they bail out or are taken to court. If still in custody after the court appearance, prisoners are sent to the county jail.
 California Penal Code sec. 836 authorizes police to arrest for any misdemeanor or felony committed in their presence. It also allows for arrest if the officer has probable cause to believe the person has committed a felony or a criminal act of family violence, even if not occurring in the officer’s presence. If the offense is a misdemeanor, not occurring in the officer’s presence, the officer must have a civilian witness sign a “private person’s” arrest form.
 California Penal Code sec. 834.
 California Penal Code sec. 853.6(a)(1).
 Ibid., sec. 853.6(a)(2).
 California Penal Code sec. 853.6(i).
 Human Rights Watch telephone interview with Eric Aries, director, Los Angeles Community Action Network’s Homeless Citation Clinic, Los Angeles, January 30, 2016.
 California Criminal Law: Practice and Procedure (Oakland: Continuing Education of the Bar, 2016), p. 100.
 Human Rights Watch telephone interview with Michelle Roberts, Rohnert Park, April 30, 2016.
 A 160-pound male would reach this amount after just over two beers. State of California Department of Motor Vehicles, “California Driver Handbook – Alcohol and Drugs,” 2017, http://ift.tt/2kGJTWI (accessed March 28, 2017).
 California Penal Code sec. 853.6(i)(2).
 Crime reports sometimes leave out details, including witness information and evidence helpful to the accused.
 During the booking process, police will attempt to locate any warrants, including out of county and out of state warrants. They will also check on whether there is a hold from the probation department, parole agent, or immigration department. If there is some other hold, they will not release the arrestee until they resolve that hold. Police are generally capable of checking for warrants and other holds while in the field giving citations.
 For example, in Santa Clara County, with some exceptions, those with a bail of $5,000 or less were released on a jail citation with a promise to appear. County of Santa Clara Bail and Release Work Group, Consensus Report on Optimal Pretrial Justice (draft), February 17, 2016, http://ift.tt/2oyhRlB (accessed March 28, 2017), p. 18. In Kings County, all individuals with bail amounts below $10,000 are cited out either in the field or after booking. Memorandum from David Robinson, Sheriff, Kings County, Pretrial Summit—Alternatives to Bail Options, November 2, 2015 (on file at Human Rights Watch).
 California Penal Code sec. 1269b(b).
 California Penal Code sec. 1269b(a).
 Human Rights Watch telephone interview with [name withheld], Los Angeles County official, Los Angeles, October 3, 2016.
 In some jurisdictions, there are two different prosecutorial agencies: a county level district attorney, who prosecutes all felonies; and a city attorney, responsible for misdemeanors. In other jurisdictions, the district attorney handles all cases.
 California Penal Code sec. 825.
 California Penal Code sec. 405 (“The taking by means of a riot of any person from the lawful custody of any police officer is lynching.”).
 The “rap” is a printout of the defendant’s history of arrests and convictions.
 In misdemeanor cases, prosecutors almost always make settlement offers at the arraignment. In felony cases, defendants rarely plead guilty on the first court date. Prosecutors usually do not make settlement offers on the first appearance. The case is generally set for a preliminary hearing in two weeks, or for settlement conference before the preliminary hearing. Prosecutors generally convey settlement offers at this second court appearance.
 The defendant generally must pay for electronic monitoring, which can be extremely expensive and so often unavailable to poor people. The jail administrator may release people on electronic monitoring instead of bail, if certain conditions are met. California Penal Code sec. 1203.018.
 Approximately 24 percent of California’s pretrial prisoners pay bail to get released. Human Rights Watch analysis of US Department of Justice, Bureau of Justice Statistics, “State Court Processing Statistics: Felony Defendants in Large Urban Counties,” 2000-2009. SCPS data is occasionally referred to in this report. This data is intended to provide information about how felony defendants flow through the court system. It is a sample of all felony cases that occurred in May of each sampled year in each sampled county. The 75 largest counties in the country are sampled. The following data is not “representative” of all California felony cases in the given years. The data has not been collected since 2009, yet this is the most recent data of its kind available. In analyses of the SCPS dataset, Human Rights Watch used the unweighted data from the years 2000 to 2009 to minimize bias toward large southern California counties because they are the only counties available in the most recent years and have higher weights. Interpretation of the analyses using the SPCS data should be done with caution, as it is impossible to explain any causal associations between the patterns reported, and the data is not sufficient to make evaluative statements about effectiveness of different forms of pretrial release. In other words, the data only describes what occurred in California but does not explain why. This analysis also does not take into account changes in California criminal justice policy and practice since the early 2000s.
 A $221-per-month payment from the county for eligible people with no assets or income.
 California Criminal Law: Practice and Procedure, p. 118.
 California Penal Code sec. 1297.
 According to one national study of large urban counties, only about 5 percent of all who bail out pay the full amount. Brian A. Reaves, “Felony Defendants in Large Urban Counties, 2009 – Statistical Tables,” US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics, December 2013, pp. 18-20.
 “Bail, Bonds and Beyond,” presentation by Hutch Harutyunyan, Gotham Bail Bonds, ICDA Fall Seminar 2016, September 17, 2016.
 Human Rights Watch telephone interview with Katherine Gibson, Los Angeles, October 19, 2016; interview and email exchanges with [name withheld], Katherine’s attorney; review of Katherine’s court file.
 California Criminal Law: Practice and Procedure, p. 1973 (“Generally, the arrestee is held until he or she has sobered up. A 4- to 12-hour hold is the norm. The arrestee is then released with a citation and promise to appear (i.e., on O.R.).”).
 Conditions of probation on a first time driving under the influence (California Penal Code sec. 23152) in Los Angeles County include a fine, a suspended license, and DUI classes. An aggravated case, for example one with an accident, might require some additional community service work.
 This bail amount was according the Los Angeles County bail schedule for a felony driving under the influence charge. California Vehicle Code sec. 23153.
 California Vehicle Code secs. 23153, 23554.
 E-mail to Human Rights Watch from [name withheld], Katherine’s attorney, December 2, 2016. The lawyer Katherine eventually hired saw no basis for the felony charge. He reviewed the reports and said they documented “no apparent significant injuries.”
 Human Rights Watch telephone interview with Katherine Gibson, Los Angeles, October 19, 2016; interview and email exchanges with [name withheld], Katherine’s attorney; review of Katherine’s court file.
 California Penal Code sec. 1270.2.
 California Penal Code sec. 1275, California Constitution art. 1, sec. 12.
 Human Rights Watch interview with [name withheld], deputy public defender, Los Angeles, November 2016.
 Human Rights Watch interview with Brian Bloom and Rodney Brooks, Alameda County deputy public defenders, Oakland, March 2016.
 Human Rights Watch interview with Judge [name withheld], Los Angeles County Superior Court, Los Angeles, March 29, 2016.
 Lisa Foster, Director of the Office of Access to Justice of the Department of Justice and former San Diego Superior Court judge, Remarks at ABA’s 11th Annual Summit on Public Defense, February 6, 2016, http://ift.tt/2oyqTz8 (accessed March 28, 2017).
 Human Rights Watch interview with Judge [name withheld], Los Angeles County Superior Court, March 30, 2016.
 Human Rights Watch interview of Baker Ostrin, Los Angeles County Deputy public defender, March 2016; Human Rights Watch interview with Jeff Adachi, public defender, San Francisco County, San Francisco, October 6, 2015.
 For example, Riverside County schedules a bail of $10,000 for all crimes with a maximum sentence of three years and $25,000 for all crimes with a maximum sentence of four years. Neighboring San Bernardino County schedules by the particular crime. Penal Code section 69 (resisting an officer) carries a three-year maximum sentence and is scheduled for $50,000 bail; Penal Code section 118 (perjury) carries a four-year maximum sentence and is also scheduled for $50,000 bail.
 For example, Riverside County schedules an additional $20,000 for each state prison prior alleged pursuant to Penal Code section 667.5 or $75,000 for use of a firearm pursuant to Penal Code section 12022.5.
 California Penal Code sec. 1269b(d).
 California Penal Code sec. 1269b(c). Judges do not create schedules for Vehicle Code infractions. Instead, the Judicial Council does.
 California Penal Code sec. 1269b(e); Human Rights Watch interviews with Sonya Tafoya, Public Policy Institute of California, August 18, 2015, October 5, 2015, and March 14, 2016.
 Human Rights Watch interview with Judge [name withheld], Contra Costa Superior Court, Martinez, March 2016.
 Human Rights Watch interview with Robin Lipetzky, public defender, Contra Costa County, Martinez, March 7, 2016.
 Human Rights Watch interview with Judge [name withheld], [name withheld] County Superior Court, March 2016.
 Sarah Lawrence, “Managing Jail Populations to Enhance Public Safety: Assessing and Managing Risk in the Post-Realignment Era,” Stanford Criminal Justice Center, June 10, 2013, http://ift.tt/2oyktzI (accessed March 28, 2017), p. 11.
 Sonya M. Tafoya, “Assessing the Impact of Bail on California’s Jail Population,” Public Policy Institute of California, p. 2.
 California Penal Code secs. 1269b(a), (b).
 California Penal Code section 1270.1.
 Human Rights Watch interviews with Judge [name withheld], San Francisco Superior Court, San Francisco, March 15, 2016 (who said they normally use the bail schedule); Molly O’Neal, public defender, Santa Clara County, San Jose, March 17, 2016; Andres del Alcazar, deputy public defender, Santa Clara County, San Jose, March 17, 2016; Baker Ostrin, deputy public defender, Los Angeles County, Los Angeles, March 2016; Judge [name withheld], Alameda County Superior Court, Oakland, March 16, 2016 (who said that the culture here is to use the schedule).
 Robin Lipetzky, Chief Public Defender, Contra Costa County, “Written Testimony Presented to the Little Hoover Commission Regarding Bail Reform and Pretrial Detention,” March 21, 2013, http://ift.tt/2oyeFqc (accessed March 28, 2017), p. 3.
 Human Rights Watch interview with Judge Eskin, formerly of Santa Barbara Superior Court, Montecito, March 22, 2016.
 ABA Pretrial Standards, Standard 10-53(e) and commentary. The standard itself says: “Financial Conditions should be the result of an individualized decision taking into account the special circumstances of each defendant, the defendant’s ability to meet the financial conditions and the defendant’s flight risk, and should never be set by reference to a predetermined schedule of amounts fixed according to the nature of the charge.”
Human Rights Watch interview with Judge [name withheld], Santa Barbara Superior Court, Santa Barbara, March 22, 2016. Some judges that spoke with Human Rights Watch indicated they would depart from bail schedules regularly. However, the majority of system professionals agreed that most judges stuck closely to them.
 Human Rights Watch interview with Judge [name withheld], Contra Costa Superior Court, Martinez, March 2016.
 California Penal Code sec. 261.5(c). Statutory rape under this section is sex with a minor, over the age of 16, by someone more than three years older.
 Human Rights Watch interview with Judge [name withheld], San Francisco Superior Court, San Francisco, March 15, 2016.
 Human Rights Watch interview with Nancy O’Malley, district attorney, Alameda County, Oakland, October 6, 2015.
 Human Rights Watch interview with Jackie Lacey, district attorney, Los Angeles County, Los Angeles, March 29, 2016.
 Los Angeles County criminal justice stakeholders meeting convened by Supervisor Sheila Kuehl, Los Angeles, December 15, 2016.
 Human Rights Watch interview with Garry Herceg, deputy county executive and former director of Pretrial Services for Santa Clara County, San Jose, April 9, 2016; email from Garry Herceg to Human Rights Watch, February 1, 2017 (“The Deputy DA in LA County is essentially saying that bail schedules work because they detain people with high bail amounts who have serious charges. This is ironic because money bail schedules, and bail in general, are not intended to detain people, and it certainly is not individualized as required in Stack v. Boyle or Salerno cases. Even if someone has a serious charge and high bail amount, they can still get out by posting a small amount to bail bondsman and be freed, how is that good public safety?”).
 Human Rights Watch telephone interview with Daria Morrison, Los Angeles, November 12, 2016; email and telephone communications with Daria’s attorney; review of court file and preliminary hearing transcript for Daria Morrison and Sarah Jackson’s case.
 Sonya M. Tafoya, “Assessing the Impact of Bail on California’s Jail Population,” Public Policy Institute of California, June 2013, p. 11.
 California Constitution art. 1, sec. 12. Judges may also order “no bail” holds on defendants accused of probation violations. Defendants in custody may also have holds placed on them through the custodial officers that prevent their release on bail because of parole violations, out of jurisdiction warrants, and immigration holds. Our analysis does not address these extrinsic holds.
 Human Rights Watch telephone interview with Chief Justice Tani Cantil-Sakauye, California Supreme Court, Sacramento, March 18, 2016.
 Human Rights Watch interviews with Judge [name withheld], Los Angeles County Superior Court, Los Angeles, March 29, 2016 (who said judges are risk averse in their bail setting; they don’t want to take chances); Judge [name withheld], Santa Barbara County Superior Court, Santa Barbara, March 2016 (who said judges set bail at an amount to keep the defendant in jail); Judge [name withheld], Alameda County Superior Court, Oakland, March 14, 2016 (who said the purpose of bail is jail).
 Human Rights Watch interview with Judge [name withheld], Santa Barbara County Superior Court, Santa Barbara, March 2016.
 Human Rights Watch interview with [name withheld], Pretrial Services for [name withheld] County, April 2016.
 In re Christie, 92 Cal. App. 4th 1105, 1109-10 (2001).
 This line is dicta and not the holding of the case, and does not cite to any other holding, though it would seem to describe a basic, common sense principle of law. The case held that the trial court must generate a specific factual record to explain its deviation from the bail schedule in order to facilitate appellate review. In this case, the judge set bail at 10 times the scheduled amount and needed to explain his reasoning.
 Human Rights Watch interview with Judge [name withheld] Los Angeles County Superior Court, Pasadena, March 30, 2016.
 Human Rights Watch interview with Chesa Boudin, deputy public defender, San Francisco County, San Francisco, March 8, 2016.
 In re Burnette (1939) 35 Cal. App. 2d 358, 360.
 Human Rights Watch interview with George Gascon, district attorney, San Francisco County, San Francisco, March 10, 2016.
 Albert W. Ramirez, Counsel, Golden State Bail Agents Association, “Written Testimony for the Little Hoover Commission,” November 27, 2012, http://ift.tt/2oohWGh (accessed March 28, 2017), pp. 6-7.
 Human Rights Watch interview of [name withheld], Bail Hotline, Oakland, April 25, 2016.
 “Bail, Bonds and Beyond,” presentation by Hutch Harutyunyan, Gotham Bail Bonds, ICDA Fall Seminar 2016, September 17, 2016.
 Human Rights Watch telephone interview with Matthew Dixon, Oakland, March 17, 2016.
 Human Rights Watch telephone interview with Paul Fowler, Richmond, April 15, 2016.
 American Bar Association, Standards for Criminal Justice: Pretrial Release (Third Edition) (Washington, DC: American Bar Association, 2007), p. 45.
 Human Rights Watch telephone interviews with Jessie Green, Oakland, March 26, 2016; Stanley Carter, East Oakland, May 4, 2016; and Terrence Rios, San Francisco, April 15, 2016.
 Human Rights Watch telephone interview Molly Harris, Monterrey, May 1, 2016.
 “Bail, Bonds and Beyond,” presentation by Hutch Harutyunyan, Gotham Bail Bonds, ICDA Fall Seminar 2016, September 17, 2016.
 California Penal Code sec. 1297. When the case is over, either because it is resolved or because it is not even filed, the judge exonerates the bond and orders it returned to the person who paid, usually the bail bondsman.
 Human Rights Watch telephone interview with Kevin Ocampo, San Lorenzo, March 21, 2016.
 Human Rights Watch interview with Henry Anderson, March 15, 2016.
 Timothy Schnacke, “Money as a Criminal Justice Stakeholder: The Judge’s Decision to Release or Detain a Defendant Pretrial,” National Institute of Corrections, September 2014, http://ift.tt/2oybwGG (accessed March 28, 2017), p. 33.
 Human Rights Watch telephone interview with Sharone Mitchell, Illinois Justice Project, Chicago, October 27, 2016.
 Human Rights Watch telephone interview with Daria Morrison, Los Angeles, November 12, 2012.
 Crime in California, 2015, California Department of Justice, California Justice Information Services Division Bureau of Criminal Information and Analysis, Criminal Justice Statistics Center, p. 49. During these years, 68.3 percent of felony arrests resulted in some conviction. 3.3 percent of those arrests were rejected at the station by the supervisor; 15.6 percent were rejected for filing by the prosecutor; and 12.8 percent were either dismissals or acquittals in court. These statistics do not include misdemeanor arrests, which accounted for 72.1 percent of all arrests in 2015. Adding misdemeanor arrests would show the number of people detained pretrial and ultimately never convicted of any crime is significantly greater than the felony arrests alone show.
 Analysis includes every jail booking that was released as a “dismissal” or “no charges filed” or an equivalent description. The actual number of days each of these people were held is multiplied against the Public Policy Institute of California per day cost estimate of $113.87 per day to estimate the total cost of these bookings per county.
 UC Irvine, Institute for Policing in Society, “Race and Policing: Defining the Problem and Developing Solutions,” video livestream, October 7, 2016, http://ift.tt/2oyhSWH (accessed March 28, 2017), comments of Sgt. Cheryl Dorsey, LAPD, ret.
 Human Rights Watch interview with Jason Miller, Los Angeles, November 16, 2016.
 California Health and Safety Code sec. 11350(a).
 Human Rights Watch telephone interview with David Gonzalez and Nina Gonzalez, Orange County, November 2, 2016; review of information from David’s court file.
 Less than 8 percent of those released with no complaint later had charges filed by the prosecutor.
 It is not possible to tell from the data if any of the people released from custody with no filing later had charges filed. Even for those who did, the initial incarceration was unnecessary and wasteful, as the case was not important enough to merit immediate attention or an effort by the prosecutor to maintain custody status.
 This figure uses the statewide average cost for a day in jail of $113.87. Brandon Martin and Ryken Grattet, “Alternatives to Incarceration in California,” Public Policy Institute of California, April 2015.
 Another 3,100 arrestees, mainly for drunk in public violations, were released within a few hours pursuant to California Penal Code section 849.
 Human Rights Watch interview of Will Leong, CEO, San Francisco Pretrial Diversion Project, San Francisco, October 9, 2016.
 Human Rights Watch interview with Frank Robinson, Oakland, April 25, 2016.
 Human Rights Watch interview with Nancy Wilson, Oakland, April 29, 2016.
 Human Rights Watch interview with Brandon Watkins, Oakland, April 25, 2016.
 California Penal Code sec. 242.
 Human Rights Watch interview with India Fuller, San Francisco, March 15, 2016.
 International Association of Chiefs of Police, Citation in Lieu of Arrest: Examining Law Enforcement’s Use of Citations Across the United States, April 2016, http://ift.tt/2oopLfkCitation%20Final%20Report%202016.pdf (accessed March 29, 2017).
 Quoted in “Bail, the Next Frontier of Criminal Justice Reform,” Sacramento Bee, March 25, 2016, http://ift.tt/1LQnV05 (accessed March 29, 2017). Chief Justice Cantil-Sakauye is a former trial court judge and prosecutor.
 Email from Kevin Ocampo, on file with Human Rights Watch, May 27, 2016.
 Human Rights Watch telephone interview with Oscar De La Torre, director, Pico Youth and Family Center, Santa Monica, November 21, 2016. Pico Youth and Family Center is a non-profit organization in Santa Monica, CA, committed to preventing youth violence. It provides various services and case management for hundreds of young people, many of whom have contact with the criminal justice system.
 Over 90 percent of those are dismissals or acquittals. Judicial Council of California, “2015 Court Statistics Report: Statewide Caseload Trends, 2004-2005 Through 2013-2014,” 2015, http://ift.tt/2ooqkFK (accessed March 29, 2017), p. 118.
 Judicial Council of California, “2015 Court Statistics Report: Statewide Caseload Trends, 2004-2005 Through 2013-2014,” 2015.
 California judges, defense lawyers, and even prosecutors told Human Rights Watch that the desire to get out jail often pressured pretrial prisoners to plead guilty regardless of the strength of their case.
 California Penal Code sec. 1382.
 In California, a judge can take an “open” plea over the prosecutor’s objection, and sentence as they deem appropriate. They may not reduce or dismiss a charge.
 Meghan Sacks and Alissa R. Ackerman, “Bail and Sentencing: Does Pretrial Detention Lead to Harsher Punishment?,” Criminal Justice Policy Review, vol. 25 (2012), p. 62. This study looked at defendants in New Jersey and found that pretrial detention resulted in longer sentences on average.
 Megan Stevenson, “Distortion of Justice: How the Inability to Pay Bail Affects Case Outcomes,” University of Pennsylvania Law School, November 8, 2016, http://ift.tt/2oytM31 (accessed March 29, 2017), p. 1.
 Ibid., p. 4.
 Ibid., p. 3.
 A study of misdemeanor defendants in Harris County, Texas came to similar conclusions as the Philadelphia study, drawing a causal connection between pretrial detention and guilty pleas. In this study, the researchers observed that similarly situated detained defendants were 25 percent more likely to plead guilty than out of custody defendants, were 43 percent more likely to get a jail sentence, and served more than twice as long in jail on average. They controlled for other factors that might influence these disparities, like criminal history, demographics and offense types. Paul Heaton, Sandra Mayson, and Megan Stevenson, “The Downstream Consequences of Misdemeanor Pretrial Detention,” University of Pennsylvania Law School, July 2016, http://ift.tt/2oopOrw (accessed March 29, 2017). In one study of pretrial detention in Philadelphia and Miami, the researchers said: “… [W]e find that pre-trial detention significantly increases the probability of conviction, primarily through an increase in guilty pleas…. These results are consistent with … pre-trial detention weakening defendants’ bargaining positions during plea negotiations….” Will Dobbie, Jacob Goldin, and Crystal Yang, “The Effects of Pre-Trial Detention on Conviction, Future Crime and Employment: Evidence from Randomly Assigned Judges,” August 2016, http://ift.tt/2oyyRIj (accessed March 29, 2017). In a study examining data on close to one million cases in New York City from 2009 through 2013, the researchers said: “We find that being detained increases the probability of conviction by over seven percentage points by causing individuals to plead guilty more often. Because pretrial detention is driven by failure to post bail, these adverse effect [sic] disproportionately hurt low-income individuals.” Emily Leslie and Nolan G. Pope, “The Unintended Impact of Pretrial Detention on Case Outcomes: Evidence from NYC Arraignments,” November 9, 2016, http://ift.tt/2oopPf4 (accessed March 29, 2017), p. 1.
 Human Rights Watch interview with Judge [name withheld], Los Angeles County Superior Court, Los Angeles, March 29, 2016.
 Human Rights Watch interview with George Gascon, district attorney, San Francisco County, San Francisco, March 10, 2016.
 Human Rights Watch interview with Brian Bloom and Rodney Brooks, deputy public defenders, Alameda County, Oakland, March 2016.
 Human Rights Watch telephone interview with Olivia Allen, Los Angeles, November 7, 2016; review of Olivia’s case court file; Human Rights Watch telephone interviews with Laura Kyle, Los Angeles, October 19, 2016, and [name withheld], attorney for Anthony Martin, Los Angeles, November 22, 2016; Human Rights Watch email correspondence and telephone interview with [name withheld], attorney for Justin Lee, November 2016.
 Human Rights Watch interview with Carlos Garcia, Los Angeles, October 30, 2016; review of Carlos’ case file, including crime reports and witness statements.
 Human Rights Watch review of witness statements in Carlos Garcia’s case file.
 California Penal Code sec. 148(a).
 Human Rights Watch interview with Carlos Garcia, Los Angeles, October 30, 2016; review of Carlos’ case file, including crime reports and witness statements.
 Public Housing tenants face extremely strict rules concerning criminal convictions. The Housing Authority can evict for almost any kind of criminal conviction the tenant or their family member or guest gets. 24 CFR 982.310. Eviction from Public Housing and loss of public benefits is a common negative consequence for people who plead guilty in order to get out of pretrial custody. Carlos’ parents were eventually able to negotiate with the Housing Authority to keep their apartment.
 California Penal Code sec. 1382.
 California Penal Code sec. 273.5.
 Human Rights Watch telephone interview with [name withheld], attorney for Arthur Charles, Los Angeles, November 16, 2016.
 Robin Lipetzky, Chief Public Defender, Contra Costa County, “Written Testimony Presented to the Little Hoover Commission Regarding Bail Reform and Pretrial Detention,” March 21, 2013, http://ift.tt/2oyeFqc (accessed March 28, 2017), p. 2.
 Human Rights Watch interview with Robin Lipetzky, chief public defender, Contra Costa County, Martinez, March 7, 2016. Prosecutors may argue that the difference is that the defendant is released on probation, which provides protection. However, own recognizance release can and usually does come with the same court orders as probation, including stay away from the complaining witness, surrender any weapons, attend AA classes, and others. A violation of pretrial own recognizance conditions exposes the defendant to further punishment, just as a violation of probation does.
 Crime in California, 2015, California Department of Justice, California Justice Information Services Division Bureau of Criminal Information and Analysis, Criminal Justice Statistics Center, p. 53.
 Human Rights Watch interview with Brian Bloom and Rodney Brooks, deputy public defenders, Alameda County, Oakland, March 2016.
 California Criminal Law: Practice and Procedure (Oakland: Continuing Education of the Bar, 2016), ch. 46.
 “Life,” meaning a life sentence.
 Human Rights Watch interview with Aaron Jansen, deputy public defender, Los Angeles County, Los Angeles, May 3, 2016.
 The sentence would technically be a prison sentence, but would be served in the county jail pursuant to California Penal Code section 1170(h).
 Human Rights Watch interview with Giovanni Giordani, chief trial deputy, Santa Barbara County Public Defenders, Santa Maria, March 25, 2016.
 Human Rights Watch telephone interviews with Nick Stewart-Oaten, deputy public defender, Los Angeles County, Los Angeles, December 12, 2016; [name withheld], deputy public defender, Los Angeles County, Los Angeles, November 2016; and [name withheld], deputy public defender, Orange County, Santa Ana, November 2016.
 Human Rights Watch telephone interviews with [name withheld], Alameda County court official, October 2015, and April 2016.
 Human Rights Watch interview with Russ Miller, probation manager, Contra Costa County, Martinez, March 18, 2016.
 Human Rights Watch interview with Judge [name withheld], Contra Costa County Superior Court, Martinez, March 2016.
 Human Rights Watch interview with Brian Bloom and Rodney Brooks, Alameda County, Oakland, March 2016.
 See Section V: “Does Bail in California Serve the Legitimate Purposes of Pretrial Detention?”
 Human Rights Watch interview with Judge [name withheld], Contra Costa County Superior Court, Martinez, March 2016.
 See Section V: “Does Bail in California Serve the Legitimate Purposes of Pretrial Detention?”
 Human Rights Watch interview with Brian Bloom and Rodney Brooks, Alameda County, Oakland, March 2016.
 Human Rights Watch interview with [name withheld], former courts administrator, Alameda County, Oakland, April 2016.
 Human Rights Watch analysis of Sacramento County booking and release data. Sacramento County provided date of conviction or plea data, which made it possible to understand the relative speed of guilty pleas for in and out of custody defendants.
 Human Rights Watch interviews with Judge [name withheld], Los Angeles County Superior Court, Los Angeles, March 29, 2016; and Judge [name withheld], Los Angeles County Superior Court, Pasadena, March 30, 2016.
 Human Rights Watch interview with Judge [name withheld], Alameda County Superior Court, Oakland, March 14, 2016.
 Attorneys in Alameda County courts independently confirmed this judge’s assessment of the pressures on judges not to release too many defendants pretrial. Human Rights Watch interview with Brian Bloom and Rodney Brooks, Alameda County, Oakland, March 2016.
 Human Rights Watch interview with [name withheld], former courts administrator, Alameda County, Oakland, April 2016. The California judiciary has become significantly more diverse in the past ten years as Governor Schwarzenegger and then Governor Brown have appointed more criminal defense attorneys and civil lawyers to the bench, after many years of domination by prosecutors. Many ex-defense attorneys have a better sense of the impact of pretrial detention when they become judges than ex-prosecutor colleagues. However, many former defense lawyers told Human Rights Watch that they worry about appearing soft on crime and are harsh on pretrial release and sentencing.
 Human Rights Watch interview with Dante Johnson, Soledad, December 12, 2016; review of court file and transcripts; review of attorney’s files; recollections from attorney [name withheld].
 California Penal Code sec. 459.
 California Penal Code sec. 647(h).
 Contrary to common perception, it is rare to find identifiable fingerprints on a gun. Still, a competent investigator should check. Dante’s lawyer was impressed by the confidence with which he demanded that the gun be checked for prints, as it showed that he was not afraid his prints would be on it, therefore being a sign of his innocence.
 California Penal Code sec. 12025(a)(2).
 California Penal Code sec. 12031(a)(1).
 California Penal Code sec. 186.22(b)(1)(B). The gang officer would later testify that gang members only carry guns to commit violent crimes and denied the possibility that a gang member might carry a gun to protect himself. In fact, gang members and other people living in neighborhoods with high rates of violence are often more afraid to be without a weapon for protection than they are afraid of penal consequences if caught with a gun. William Lee, “Top Cop: ‘Chicago is not out of control’ – just parts of it are violent,” Chicago Tribune, December 6, 2016, http://ift.tt/2ghmP1x (accessed March 29, 2017).
 California Penal Code secs. 1170.12 and 667(b)-(i) define the offenses that make up California’s “Three Strikes” law. Passed in 1994, the laws drastically increased prison sentences for anyone convicted of any felony if they had a “serious” or “violent” felony prior. The law has changed recently to enhance sentences only if the current or new crime is “serious” or “violent.”
 Citizens’ Commission on Jail Violence, Report of the Citizens’ Commission on Jail Violence, September 28, 2012, http://ift.tt/2g2PAgS (accessed March 29, 2017); Sarah Liebowitz et al., Cruel and Unusual Punishment: How a Savage Gang of Deputies Controls LA County Jails, ACLU of Southern California and ACLU National Prison Project, September 2011, http://ift.tt/1A8D57T (accessed March 29, 2017).
 Transcript of the plea, on file with Human Rights Watch.
 Human Rights Watch telephone interview with Victor Lawrence, Oakland, April 19, 2016.
 Human Rights Watch interviews with Judge [name withheld], Santa Barbara Superior Court, Santa Barbara, March 22, 2016; and Russ Miller, probation manager, Contra Costa County, Martinez, March 18, 2016; Human Rights Watch telephone interview with Nick Stewart-Oaten, deputy public defender, Los Angeles County, Los Angeles, December 2, 2016.
 Human Rights Watch interview with Andres del Alcazar, deputy public defender, Santa Clara County, San Jose, March 17, 2016.
 Human Rights Watch telephone interview with Nick Stewart-Oaten, deputy public defender, Los Angeles County, Los Angeles, December 2, 2016. For felony cases in Los Angeles, some defendants may appear at a hearing to attempt to settle the case before their preliminary hearing.
 Human Rights Watch interview with Giovanni Giordani, chief trial deputy, Santa Barbara County Public Defender, Santa Maria, March 25, 2016.
 Christopher T. Lowenkamp, Marie VanNostrand, and Alexander Holsinger, Investigating the Impact of Pretrial Detention on Sentencing Outcomes, Laura and John Arnold Foundation, November 2013, http://ift.tt/1IK7puN (accessed March 29, 2017).
 Paul Heaton, Sandra Mayson, and Megan Stevenson, “The Downstream Consequences of Misdemeanor Pretrial Detention,” University of Pennsylvania Law School, July 2016. This study critiques the Arnold study for not distinguishing sufficiently between the types of crimes for which people are detained and subsequently punished. For example, the Arnold study puts “violent” crimes together, but that category includes simple battery, a misdemeanor that is more likely to result in an own recognizance release, with murders and rapes, that never will result in an own recognizance release. It is expected, independent of custodial status, that someone with a murder conviction will receive a much longer sentence than someone with a misdemeanor battery conviction.
 See Section III: “Bail and Jail Result in an Unfair Justice System.”
 Paul Heaton, Sandra Mayson, and Megan Stevenson, “The Downstream Consequences of Misdemeanor Pretrial Detention,” University of Pennsylvania Law School, July 2016, p. 6.
 “Good behavior in the community prior to trial suggests to a judge that a defendant does not pose a danger and will make their scheduled court appearances. These defendants can keep their jobs or find new employment. They are able to continue attending school and therefore demonstrate ties to the community. Taken together, these factors may contribute to a judge’s perception that defendants who are released on bail prior to trial are worthy of more lenient sentences, such as community supervision or other non-custodial sentences. A defendant who is detained pretrial will obviously not have the same opportunity to demonstrate a network of community ties and a pattern of good conduct in the community. Therefore, defendants who are released on bail have a distinct advantage at sentencing.” Meghan Sacks and Alissa R. Ackerman, “Bail and Sentencing: Does Pretrial Detention Lead to Harsher Punishment?,” Criminal Justice Policy Review, vol. 25 (2012), p. 71.
 Several people described being unable to help conduct investigations because they were in custody, leading them to feel they were unable to mount a successful defense, contributing to their decisions to plead guilty. Human Rights Watch interviews with Jeremy Uribe, Los Angeles, November 21, 2016; Nelson Perez, Chico, March 8, 2016; and Dante Johnson, Soledad, December 12, 2016; Human Rights Watch telephone interview with Bill Williams, San Diego, December 15, 2016.
 Intimidating witnesses or destroying evidence may constitute new substantive crimes and expose the perpetrator to further punishment, if caught. An imperfect but effective safeguard to prevent such misconduct is to allow the prosecution to demonstrate specific dangers of a defendant committing this misconduct through an evidentiary hearing, and allowing the court to order detention based on that proof. Imposing financial bail conditions does not necessarily prevent witness intimidation or evidence tampering, as those with sufficient money to bail out are equally if not more likely to commit these acts. The most frequent types of cases where witness intimidation occurs involve domestic violence or gangs. Witness intimidation in gang cases is not dependent on the custody status of the accused, as fellow gang members are as likely to threaten witnesses as the actual person accused. Domestic violence cases present extremely complex questions about custody status. While there is often potential for witness intimidation, equally if not more often, the complaining witness needs the accused out of custody in order to work and support the family. These questions should be addressed through very fact-specific analysis, preferably through an evidentiary hearing.
 Human Rights Watch telephone interview of Daria Morrison, Los Angeles, November 12, 2016; conversations and email exchanges with Daria’s attorney; review of court file and preliminary hearing transcript, on file at Human Rights Watch.
 Email to Human Rights Watch from attorney [name withheld].
 Email to Human Rights Watch from attorney [name withheld], November 11, 2016.
 Bernadette Rabuy and Daniel Kopf, Prisons of Poverty: Uncovering the pre-incarceration incomes of the imprisoned, Prison Policy Initiative, July 9, 2015, http://ift.tt/2jA0ZFj (accessed March 28, 2017).
 Human Rights Watch interview with Carlos Garcia, Los Angeles, October 30, 2016; Human Rights Watch telephone interviews with attorney [name withheld], Van Nuys, November 10, 2016; and Marta Lopez, San Fernando, November 7, 2016; email communication with attorney [name withheld], November 11, 2016; review of Carlos’ court file.
 Brian A. Reaves, “Felony Defendants in Large Urban Counties, 2009 – Statistical Tables,” US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics, December 2013, pp. 18-20. These statistics are based on a survey of 65 of the largest urban counties in the US, over a period from 1990-2009. The amount remained fairly steady over that time. California’s numbers may be lower due to its generally higher bail amounts.
 Human Rights Watch telephone interview with Marta Lopez, San Fernando, November 7, 2016.
 Annamaria Lusardi, Daniel J. Schneider, and Peter Tufano, “Financially Fragile Households: Evidence and Implications,” NBER Working Paper No. 17072, May 2011.
 Human Rights Watch analysis US Department of Justice, Bureau of Justice Statistics, “State Court Processing Statistics: Felony Defendants in Large Urban Counties,” 2000-2009, http://ift.tt/2oyiOKQ (accessed March 28, 2017). See Methodology section.
 Albert W. Ramirez, Counsel, Golden State Bail Agents Association, “Written Testimony for the Little Hoover Commission,” November 27, 2012, http://ift.tt/2oohWGh (accessed March 28, 2017), pp. 1, 5-6.
 Ryan Woolsey, “Poverty in California: Recently Released Census Data,” California State Legislature, Legislative Analyst’s Office, September 13, 2016, http://ift.tt/2k3wLOh (accessed March 29, 2017).
 Sarah Bohn and Caroline Danielson, “Poverty in California.”
 Ibid. Poverty rates for Latinos are 28.8 percent; for black people, 20.2 percent; for white people, 14 percent.
 Correlation r=.606. Human Rights Watch analysis US Department of Justice, Bureau of Justice Statistics, “State Court Processing Statistics: Felony Defendants in Large Urban Counties,” 2000-2009. See Methodology section.
 Correlation r=.566. Ibid.
 Bernadette Rabuy and Daniel Kopf, “Detaining the Poor: How money bail perpetuates an endless cycle of poverty and jail time,” Prison Policy Initiative, http://ift.tt/2jmq0mQ (accessed March 29, 2017).
 Human Rights Watch telephone interview with Cara Esparza, Long Beach, November 14, 2016.
 Had the initial charge been the misdemeanor, he likely would have been granted an own recognizance release. If set, his bail pursuant to the schedule would have been $20,000.
 State Court Processing Statistics data as retrieved from the Felony Defendants in Large Urban Counties reports, 1992-2006. This data is analyzed in Justice Policy Institute, Bail Fail: Why the U.S. Should End the Practice of Using Money for Bail, September, 2012, http://ift.tt/1EoY8nG (accessed March 29, 2017), pp. 10-11.
 Human Rights Watch telephone interview with Nick Stewart-Oaten, deputy public defender, Los Angeles County, Los Angeles, December 2, 2016.
 Human Rights Watch telephone interview with Jane Meyers, Tucson, November 17, 2016.
 Human Rights Watch interview with Felix Ayala, Oakland, April 25, 2016.
 Human Rights Watch interview with Alicia Wright, Oakland, April 5, 2016.
 Human Rights Watch telephone interview with Jordan Davis, Los Angeles, October 11, 2016.
 Human Rights Watch telephone interview with Marcus Garza, Los Angeles, October 20, 2016.
 Human Rights Watch telephone interview with Victor Lawrence, Oakland, April 19, 2016.
 Marc Debbaudt, “The Assault on Safety Continues: Eliminating Bail is the Latest Target,” Association of Deputy District Attorneys, January 27, 2016, http://ift.tt/2oytrNE (accessed March 29, 2017).
 Human Rights Watch interview with [name withheld], county district attorney, March 2016.
 Human Rights Watch telephone interview with Kevin Ocampo, San Lorenzo, March 21, 2016; email from Kevin Ocampo, May 27, 2016.
 Email to Human Rights Watch from Kevin Ocampo, May 27, 2016.
 Human Rights Watch interview with George Gascon, district attorney, San Francisco County, San Francisco, March 10, 2016.
 Robin Lipetzky, Chief Public Defender, Contra Costa County, “Written Testimony Presented to the Little Hoover Commission Regarding Bail Reform and Pretrial Detention,” March 21, 2013, http://ift.tt/2oyeFqc (accessed March 28, 2017), p. 2.
 “Bail, Bonds and Beyond,” presentation by Hutch Harutyunyan, Gotham Bail Bonds, ICDA Fall Seminar 2016, September 17, 2016.
 Thomas H. Cohen and Brian A. Reaves, “Pretrial Release of Felony Defendants in State Courts,” US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics, November 2007, http://ift.tt/2oyiVpK (accessed March 29, 2017).
 Thomas Cohen and Tracey Kyckelhahn, “State Court Processing Statistics Data Limitations,” US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics, March 2010, http://ift.tt/2oyeIlS (accessed March 29, 2017).
 Thomas H. Cohen and Brian A. Reaves, “Pretrial Release of Felony Defendants in State Courts,” p. 9. Own recognizance release showed a 17 percent re-arrest rate; unsecured bond, a 14 percent rate; and secured bond, a 16 percent rate. Unlike the Jones study, cited above, these numbers do not control for factors related to risk for the people within these categories.
 California Constitution, art. 1, sec. 12.
 Human Rights Watch analysis of US Department of Justice, Bureau of Justice Statistics, “State Court Processing Statistics: Felony Defendants in Large Urban Counties,” 2000-2009, http://ift.tt/2oyiOKQ (accessed March 28, 2017). These statistics do not account for the effects of Proposition 47, passed in 2014, which reduced drug possession and several theft-related offenses from felonies to misdemeanors. Many of the types of crimes included in this dataset would not be included in data collected after implementation of Proposition 47. The earlier data shows that pretrial misconduct is higher among people charged with drug offenses, so rates of felony re-arrest may have decreased, though not necessarily total amounts of pretrial misconduct.
 While not conclusive, as many other factors may account for this difference, this statistic does tend to show that California’s pretrial detention system, with its high rate of detainees, is not particularly effective in reducing pretrial crime.
 Human Rights Watch interview with Judge [name withheld], Alameda County Superior Court, Oakland, March 14, 2016.
 Human Rights Watch telephone interview with Nick Stewart-Oaten, deputy public defender, Los Angeles County, Los Angeles, December 2, 2016.
 California Penal Code sec. 273.5.
 Human Rights Watch interview with Edwin Molina, Los Angeles, November 10, 2016.
 California Penal Code sec. 1305. If the defendant does not show up, the court declares the bond forfeit, meaning the court will keep the money deposited. However, there are a variety of time limitations and procedures the court must go through before keeping the money that allow the bond company sufficient opportunity to get their money back.
 Human Rights Watch interview with Chesa Boudin, deputy public defender, San Francisco County, San Francisco, March 8, 2016.
 Haley R. Zettler and Robert G. Morris, “An Exploratory Assessment of Race and Gender-Specific Predictors of Failure to Appear in Court Among Defendants Released via a Pretrial Services Agency,” Criminal Justice Review, vol. 40 (2015), pp. 417-430.
 Human Rights Watch interview with Will Leong, CEO, San Francisco Pretrial Diversion Project, San Francisco, October 9, 2016.
 See Lavan v. City of Los Angeles, 693 F.3d 1022 (9th Cir. 2012).
 Human Rights Watch email correspondence with Eric Aries, director, Los Angeles Community Action Networks Skid Row Homeless Citation Clinic, Los Angeles, January 30, 2017.
 Human Rights Watch interview with Cpt. Gary Newton, Los Angeles Police Department, Los Angeles, March 2016.
 Human Rights Watch interview with Derek Bercher, assistant alternate public defender, Orange County Alternate Public Defender, Santa Ana, October 12, 2016. Bercher noted that there appear to be less missed court dates in the specialized homeless courts that offer services for these individuals and a chance for case dismissals.
 Human Rights Watch interview with Guadalupe Rabago, chief probation officer, and Tanja Heitman, deputy chief probation officer, Santa Barbara, March 23, 2016.
 Human Rights Watch interview with Will Leong, CEO, San Francisco Pretrial Diversion Project, San Francisco, October 9, 2016. He estimated the rate of failure to appear as about 6-9 percent, and that half eventually come to court.
 Thomas H. Cohen and Brian A. Reaves, “Pretrial Release of Felony Defendants in State Courts,” US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics, November 2007, p. 8. This measure is somewhat flawed. Many people who miss a court date, but come back to court soon after, do so on their own; others come back because police have arrested them and brought them back to court. Some who are still unaccounted for after one year have left the jurisdiction or are in hiding; others simply have not had contacts with police, demonstrating a greater likelihood of living a law-abiding life.
 “Bail, Bonds and Beyond,” presentation by Hutch Harutyunyan, Gotham Bail Bonds, ICDA Fall Seminar 2016, September 17, 2016.
 Wendy F. White, “Court Hearing Call Notification Project,” Criminal Justice Coordinating Council & Flagstaff Justice Court, Coconino County, May 17, 2006, http://ift.tt/2oopQzyCounty%20AZ%20Court%20Hearing%20Notification%20Project%20(2006).pdf (accessed March 29, 2017).
 Timothy R. Schnacke, Michael R. Jones, and Dorian M. Wilderman, “Increasing Court-Appearance Rates and Other Benefits of Live-Caller Telephone Court-Date Reminders: The Jefferson County, Colorado FTA Pilot Project and Resulting Court Date Notification Program,” Court Review, vol. 48 (2012): 86-95.
 Brian H. Bornstein, Alan J. Tomkins, and Elizabeth M. Neeley, Reducing Courts’ Failure to Appear Rate: A Procedural Justice Approach, May 2011, http://ift.tt/2oyatqi (accessed March 29, 2017). This study used reminder postcards and concluded such a system would reduce failure to appear in a cost-effective manner. It also concluded that a personal reminder system would be even more effective.
 Will Dobbie, Jacob Goldin, and Crystal Yang, “The Effects of Pre-Trial Detention on Conviction, Future Crime and Employment: Evidence from Randomly Assigned Judges,” August 2016, http://ift.tt/2oyyRIj (accessed March 29, 2017), pp. 2-3. The authors observed that, while pretrial release added a certain increase in crime committed by those released, the increase is offset because, in the long run, people who avoid pretrial custody maintain employment better and avoid committing future crimes. The study says that doing time in jail is “crimogenic,” or leads to more crime, and acquiring a criminal conviction, made more likely by pretrial detention, hurts future job prospects.
 Ibid., p. 3.
 Los Angeles City invested $100,000,000 to address homelessness, 87 percent of which was for law enforcement an emergency services. Gale Holland, “L.A. spends $100 million a year on homelessness, city report finds,” Los Angeles Times, April 16, 2015, http://ift.tt/1E7swGL (accessed M arch 29, 2017).
 News article on file with Human Rights Watch.
 Human Rights Watch telephone interview with Nelson Perez, Chico, March 8, 2016; review of court documents.
 Marie VanNostrand, “Pretrial Risk Assessment – Perpetuating or Disrupting Racial Bias?,” Pretrial Justice Institute, December 6, 2016, http://ift.tt/2oyasme (accessed March 29, 2017). Dr. VanNostrand uses the term “resource-based vs. risk-based” to describe the difference between the current money bail system and the use of statistical probabilities to determine risk. This characterization is only partially accurate, as the current overall pretrial detention system involves uses of discretion unrelated to the individual’s financial resources, while the setting of bail also involves a calculation of risk, and the statistical tools may be influenced by factors separate from risk, like racial background and law enforcement priorities, that will be discussed below. The term also implies these two options are the only ones available.
 Most of the tools appear to generalize misconduct as any kind of future crime, not distinguishing between crimes that are actually dangerous to the community. Some tools do have specific flags for “violent” crime, though it is unclear how judges evaluate the distinction. To the extent the tools equate any criminal conduct with dangerousness, they are highly flawed.
 Marie VanNostrand and Kenneth J. Rose, Pretrial Risk Assessment in Virginia, Virginia Department of Criminal Justice Services, May 1, 2009, http://ift.tt/2ooBTgo (accessed March 29, 2017), pp. 4, 22.
 It is not clear whether race is factored toward the score on this tool. However, it is part of the demographic information requested on the form.
 Virginia Pretrial Risk Assessment Tool. See Sarah Lawrence, “Managing Jail Populations to Enhance Public Safety: Assessing and Managing Risk in the Post-Realignment Era,” Stanford Criminal Justice Center, June 10, 2013, http://ift.tt/2oyktzI (accessed March 28, 2017), p. 25.
 Ohio Risk Assessment System: Pretrial Assessment Tool. Sarah Lawrence, “Managing Jail Populations to Enhance Public Safety: Assessing and Managing Risk in the Post-Realignment Era,” p. 26. Both the Virginia and Ohio tools are used in various jurisdictions nationwide. Often they are modified by the local jurisdiction.
 Zach Dal Pra, “LJAF Public Safety Assessment – PSA,” Laura and John Arnold Foundation, slides 42-43. The scoring sheet does indicate race in its demographic information, but Arnold says that race is not input in generating the score.
 Timothy P. Cadigan and Christopher Lowenkamp, “Implementing Risk Assessment in the Federal Pretrial Services System,” Federal Probation, vol. 75 (2011), p. 31.
 Laura and John Arnold Foundation, “New Data: Pretrial risk assessment tool works to reduce crime, increase court appearances,” August 8, 2016, http://ift.tt/2cGpXQ6 (accessed March 29, 2017), p. 4 (“Though these neutral factors can help judges gauge the risk that a defendant poses, they do not impede a judge’s discretion or authority in any way. The decision about whether to release or detain a defendant always rests with the judge regardless of the scores produced by the risk assessment.”).
 Amber Widgery, “Guidance for Setting Pretrial Conditions,” National Conference of State Legislatures, May 13, 2015, http://ift.tt/2ooorJc (accessed March 29, 2017). Discretionary states are Colorado, Connecticut, Delaware, Hawaii, Illinois, Kansas, Kentucky, Louisiana, Maine, New Jersey, Oklahoma, South Carolina, Vermont, Virginia, and West Virginia; mandatory states are Delaware, Colorado, Kentucky, New Jersey, South Carolina, and West Virginia.
 Lisa Brooks, Meghan Guevara, and Sam Packard, Pretrial Progress: A Survey of Pretrial Practices and Services in California, Californians for Safety and Justice, August 2015, http://ift.tt/2oooViw (accessed March 29, 2017), p. 7.
 Shima Baradaran and Frank L. McIntyre, “Predicting Violence,” Texas Law Review, vol. 90 (2012), pp. 497, 553. This study assumed judges would follow the recommendations of their risk assessment. The numbers also depend on policy decisions about what level of risk would trigger detention: “We can also choose to release all of those who are less than 30 percent likely to commit a crime pretrial to see what the impact would be.”
 Laura and John Arnold Foundation, “Results from the First Six Months of the Public Safety Assessment-Court in Kentucky,” July 2014, http://ift.tt/1IDmWYP (accessed march 29, 2017), p. 2. The analysis was conducted by Dr. Marie VanNostrand, who contracts with Arnold for much of their research. It is worth noting, however, according to statistics compiled by the Legislative Research Commission for the state of Kentucky, pretrial release rates were at 69 percent and the rate of new pretrial arrests was 6 percent in 2011 following implementation of other criminal reform legislation, raising questions about data collection or the magnitude of improvement from the Arnold tool. Legislative Research Commission, Report of the 2011 Task Force on the Penal Code and Controlled Substances Act, December 2011, http://ift.tt/2oot8Tu (accessed March 29, 2017), p. 5.
 Pretrial Justice Institute, The Transformation of Pretrial Services in Allegheny County, Pennsylvania: Development of Best Practices and Validation of Risk Assessment, October 9, 2007, http://ift.tt/2oot6uQ (accessed March 29, 2017), p. vii.
 Laura and John Arnold Foundation, “New Data: Pretrial risk assessment tool works to reduce crime, increase court appearances,” August 8, 2016.
 “NOTICE OF FILING COPY OF PRESENTATION ASSESSING IMPACT OF PUBLIC SAFETY ASSESSMENT,” Jones v. Wittenberg, Case No. C70-388, U.S.D.C. ND Ohio (January 9, 2017). The document is included in Marie VanNostrand, Assessing the Impact of the Public Safety Assessment: Public Safety, Court Appearance and Jail Population, Lucas County, Ohio. Dr. VanNostrand’s report is consistent with the Arnold press release in claiming a reduction in pretrial crime rates and increase in appearance rates. However, the data used compares one year of use of the Arnold tool with three undifferentiated years before its use, making it impossible to determine if the outcomes reflected pre-existing trends. The report does not explain the dramatic increase in early guilty pleas, which are often a result of pretrial detention.
 Human Rights Watch interviews with Guadalupe Rabago, chief probation officer, Santa Barbara County, Santa Barbara, March 23, 2016; and Garry Herceg, deputy county executive and former director of pretrial services, Santa Clara County, San Jose, April 9, 2016; email from Garry Herceg to Human Rights Watch, February 1, 2017. Herceg told Human Rights Watch there were significant increases in the number of people released, with a 1-2 percent increase in failures to appear, which he attributed to the release of more homeless people through use of the risk assessment tool.
 Sarah Fletcher, Alternatives to Custody Report 2015, Santa Cruz County Probation Department, April 2016, http://ift.tt/2oytRUn (accessed March 29, 2017), p. 10. The report does not make clear how the probation department calculated “bed days saved,” or what is the cost and scope of supervision. It does not include data from before implementation of the risk assessment tool. Even with these savings, Santa Cruz County still has a greater percentage of unsentenced prisoners in its jail than the statewide average, and has the highest rate of incarcerating pretrial misdemeanor defendants. See Human Rights Watch analysis of US Department of Justice, Bureau of Justice Statistics, “State Court Processing Statistics: Felony Defendants in Large Urban Counties,” 2000-2009, http://ift.tt/2oyiOKQ (accessed March 28, 2017). These facts indicate that the risk assessment tool has not solved this county’s problem.
 Shima Baradaran and Frank L. McIntyre, “Predicting Violence,” Texas Law Review, vol. 90 (2012), pp. 497, 553.
 Samuel R. Wiseman, Fixing Bail, The George Washington Law Review, vol. 84 (2016), pp. 417, 467.
 Human Rights Watch telephone interview with Edwin Monteagudo, director, Pretrial Services Division, Los Angeles County Probation Department, Los Angeles, April 5, 2016; Human Rights Watch interview with Russ Miller, probation manager, Contra Costa County, Martinez, March 18, 2016.
 Human Rights Watch interviews with [name withheld], courts administrator, Alameda County, Oakland, October 2015, and April 2016. An Alameda County judge who spoke with Human Rights Watch, though not for attribution, said judges did grant own recognizance release to low risk defendants. He said judges sense the risk assessment data is unreliable and think they can make better judgments on their own. Human Rights Watch interview with Judge [name withheld], Alameda County Superior Court, Oakland, March 16, 2016.
 Interview with Judge [name withheld], Alameda County Superior Court, Oakland, March 14, 2016.
 Human Rights Watch interview with Kim McGill, organizer, Youth Justice Coalition, Los Angeles, November 2016. The coalition is an organization led by youth and formerly incarcerated people who frequently come in contact with law enforcement and who advocate for respectful treatment of all people in the criminal system.
 Human Rights Watch interview with Joyce Dudley, district attorney, Santa Barbara County, Santa Barbara, March 21, 2016.
 Michael Tonry, Legal and Ethical Issues in the Prediction of Recidivism, Federal Sentencing Reporter, vol. 26 (2014), 167-176.
 Anthony W. Flores, Christopher T. Lowenkamp, and Kristin Bechtel, “False Positives, False Negatives, and False Analyses: A Rejoinder to ‘Machine Bias: There’s Software Used Across the Country to Predict Future Criminals. And it’s Biased Against Blacks,’” Federal Probation Journal, vol. 80 (2016), 38-46.
 Human Rights Watch telephone interview with Michael Jones, Pretrial Justice Institute, March 16, 2016; email from Michael Jones to Human Rights Watch, February 1, 2017.
 Julia Angwin et. al, “Machine Bias,” ProPublica, May 23, 2016.
 Criminal Justice Policy Program, Harvard Law School, Moving Beyond Money: A Primer on Bail Reform, October 2016, p. 22 (“This can lead to a vicious cycle: because pretrial detention has been shown to lead to worse criminal justice outcomes, the characteristics of the individuals detained pursuant to risk assessment will gain an even stronger association with pretrial failure over time, thus strengthening the seeming predictive power of those features. Indeed, because APRAIs are based on empirically-derived factors, it is possible that risk assessment tools will not only entrench but exacerbate existing racial and socioeconomic disparities by appearing to give a scientific imprimatur to unequal outcomes.”). See also Open Justice, “How do arrest rates differ by race over time?,” 2015, http://ift.tt/2oovDFj (accessed March 29, 2017).
 Bernard E. Harcourt, “Risk as a Proxy for Race: The Dangers of Risk Assessment,” Federal Sentencing Reporter, vol. 27 (2015), pp. 237-243.
 Ibid., p. 239.
In Baltimore, according to city arrest data, between 2013 and 2016, black people made up 81.5 percent of all people arrested and 84.6 percent of all arrested on drug charges, but only 63.7 percent of the population. Joseph George, “Justice by Algorithm,” CityLab, December 8, 2016, http://ift.tt/2oyaudQ (accessed March 29, 2017).
 Italics added. Santa Cruz County Probation Department, Alternatives to Custody Report 2015, April 2016, p. 11.
 Zach Dal Pra, “LJAF Public Safety Assessment – PSA,” Laura and John Arnold Foundation, slides 6, 50-54.
 Criminal Justice Policy Program, Harvard Law School, Moving Beyond Money: A Primer on Bail Reform, p.21.
 For example, in Arnold’s presentation to San Francisco, they calculated the chance of committing a new violent crime, in their highest risk category, as 11.1 percent―almost a 90 percent probability of not committing a new violent crime. For the lowest risk category, the chance is 1.3 percent. Zach Dal Pra, “LJAF Public Safety Assessment – PSA,” Laura and John Arnold Foundation, slide 33. The presentation was made by Justice System Partners (JSP) on a contract with Arnold.
 Jeff Adachi, “SF bail algorithm falters on promise: Bail algorithm may increase incarceration,” Los Angeles Daily Journal, 2016.
 Sarah Lawrence, “Managing Jail Populations to Enhance Public Safety: Assessing and Managing Risk in the Post-Realignment Era,” Stanford Criminal Justice Center, June 10, 2013, pp. 14-15.
 Ibid. p. 15.
 Marie VanNostrand and Gena Keebler, Pretrial Risk Assessment in the Federal Court, US Department of Justice, Office of the Federal Detention Trustee, April 14, 2009, http://ift.tt/2oodrf5 (accessed March 29, 2017), pp. 29-33; Christopher T. Lowenkamp and Marie VanNostrand, Exploring the Impact of Supervision on Pretrial Outcomes, Laura and John Arnold Foundation, November 2013, http://ift.tt/2oygofd (accessed March 29, 2017).
 Marie VanNostrand and Gena Keebler, Pretrial Risk Assessment in the Federal Court, pp. 31-32. Drug treatment and testing requirements also lowered success rates.
 Ibid., pp. 4, 6.
 Ibid., p. 4.
 Ibid., pp. 5-6.
 Ibid, p. 4.
 Ibid., p. 8.
 Crime in California, 2015, California Department of Justice, California Justice Information Services Division Bureau of Criminal Information and Analysis, Criminal Justice Statistics Center, p. 6.
 Senate Budget and Fiscal Review Committee, Office of Senate Floor Analyses, Senate Bill 863 (June 18, 2014); Senate Bill 863, Filed with Secretary of State June 20, 2014; Senate Bill 844, Filed with Secretary of State, June 27, 2016.
 Board of State & Community Corrections: SB1022 Adult Local Criminal Justice Facilities Construction Funding Awards, March 13, 2014; Public Policy Institute of California Data Set—Key Factors in CA Jail Construction: AB900 Jail Construction Projects; Senate Bill 863: Adult Local Criminal Justice Facilities Construction—Summary of Awarded Projects, November 12, 2015.
 Sarah Lawrence, “Managing Jail Populations to Enhance Public Safety: Assessing and Managing Risk in the Post-Realignment Era,” Stanford Criminal Justice Center, June 10, 2013, p. 21.
 California Constitution, art. 1, section 7; see preceding section on International Human Rights Law.
 Individualized risk assessment contrasts with the statistical predictions of profile-based risk assessment by taking into account the circumstances of the defendant’s individual situation.
 In the case, US v. Salerno (1987) 481 U.S. 739, the Supreme Court outlined fundamental procedural protections required to justify the limited use of “preventive” or no-bail detention, pursuant to provisions of the Bail Reform Act of 1984.
 US v. Salerno, p. 750.
 Human Rights Watch telephone interviews with Armando Miranda, San Francisco County deputy public defender, April 19, 2016; and Joshua Norkin, New York Legal Aid Society, New York, October 2016.
 California Welfare and Institutions Code sec. 637; 18 USC sec. 3142.
 US v. Salerno, p. 751.
 Salerno, p. 751. See also California Constitution art.1, sec.12, which requires “clear and convincing evidence that there is a substantial likelihood the person’s release would result in great bodily harm to others; or … there is a substantial likelihood that the person would carry out the threat if released.” This requirement currently applies to preventive detention. Judges avoid it by setting high bail.
 International human rights standards regarding pretrial detention are predicated not only on fundamental rights; they also are grounded in recognition of the stress and suffering detainees may endure from being confined in jail. In addition to the emotional impact of confinement, detainees are separated from family, friends, and community, are uncertain as to their future, worry about their legal position, and are anxious about their economic future and that of their family. See Centre for Human Rights, Crime Prevention and Criminal Justice Branch, Human Rights and Pre-trial Detention: A Handbook of International Standards relating to Pre-trial Detention (New York and Geneva: United Nations, 1994), p. 8.
 International Covenant on Civil and Political Rights (ICCPR), adopted December 16, 1966, G.A. Res. 2200A (XXI), 21 U.N. GAOR Supp. (No. 16) at 52, U.N. Doc. A/6316 (1966), 999 U.N.T.S. 171, entered into force March 23, 1976, ratified by the United States June 8, 1992, art. 9(1).
 Ibid.; Manfred Nowak, UN Covenant on Civil and Political Rights: CCPR Commentary, (Arlington: N P Engel Publisher, 1993), p. 172.
 Ibid., p. 173.
 Van Alphen v. the Netherlands, Human Rights Committee, Communication No. 305/1988, U.N. Doc. CCPR/C/39/D/305/1988 (1990), para. 5.8.
 ICCPR, art. 14(2).
 See Section IV: “Bail Devastates Poor and Middle-Income Defendants and Households.”
 United Nations General Assembly, Standard Minimum Rules for Non-Custodial Measures (The Tokyo Rules), Resolution 45/110, U.N. Doc. A/45/49 (1990), rule 6.1. See generally, Centre for Human Rights, Human Rights and Pre-trial Detention, p. 14 (Substitutes for Confinement).
 See, for example, UN Human Rights Committee, General Comment No. 8, Right to liberty and security of persons (Sixteenth session, 1982), Compilation of General Comments and General Recommendations Adopted by Human Rights Treaty Bodies, U.N. Doc. HRI/GEN/1/Rev.9 (vol. I)(2008), p. 179, para. 3 ("Pre-trial detention should be an exception and as short as possible."); United Nations Standard Minimum rules for Non-Custodial Measures (The Tokyo Rules), G.A. Res. 45/110 ("Pre-trial detention shall be used as a means of last resort in criminal proceedings…alternatives to pre-trial detention shall be employed at as early a state as possible. Pretrial detention shall last no longer than necessary.").
 Centre for Human Rights, Human Rights and Pre-trial Detention, p. 18
 Ibid., p. 16.
 Ibid., p. 15.
 Nowak, UN Covenant on Civil and Political Rights: CCPR Commentary, p. 177.
 The IACHR is the principle oversight branch that monitors OAS member states’ compliance with the American Declaration of the Rights and Duties of Man and the American Convention on Human Rights, both adopted by the US See generally, OAS, “Basic Documents in the Inter-American System,” http://ift.tt/2oow2Yu (accessed June 20, 2016), footnote 4.
 IACHR, Report on the Use of Pretrial Detention in the Americas, OEA/Ser.L/V/VII, Doc. 46/13 (2013) [hereinafter OEA Report], para. 185. Additionally, the evidence presented “must be based on facts, expressed in words; that is, not on mere conjectures or intuitions.” Suárez Rosero v. Ecuador, IACHR Series C no. 35  IHRL 1418, Nov. 12, 1997, para. 77. European courts have interpreted article 5.3 of the European Convention on Human Rights to reject pretrial detention decisions based on formula and not on careful consideration of all relevant facts. See Gabor Nagy v. Hungary, No. 33529/11 Feb. 11, 2014: failure to consider all relevant circumstances made the pretrial detention unlawful.
 OEA Report, para. 298.
 Nowak, UN Covenant on Civil and Political Rights, p. 458.
 Equality and nondiscrimination operate as positive and negative statements of the same concept. Anne F. Bayefsky, "The Principle of Equality or Non-Discrimination in International Law," Human Rights Law Journal, vol. 11 (1990).
 ICCPR art. 2(1) states: "Each State Party to the present Covenant undertakes to respect and to ensure to all individuals within its territory and subject to its jurisdiction the rights recognized in the present Covenant, without distinction of any kind, such as race, colour, sex, language, religion, political or other opinion, national or social origin, property, birth or other status." See also, International Covenant on Economic, Social and Cultural Rights (ICESCR), adopted December 16, 1966, G.A. Res. 2200A (XXI), 21 U.N. GAOR Supp. (No. 16) at 49, U.N. Doc. A/6316 (1966), 993 U.N.T.S. 3, entered into force January 3, 1976, art. 2(2); General Conference of the United Nations Educational, Scientific and Cultural Organization, Convention against Discrimination in Education, 429 U.N.T.S. 93, art. 1 ("For the purpose of this convention, the term ‘discrimination’ includes any distinction, exclusion, limitation or preference which, being based on race, colour, sex, language, religion, political or other opinion, national or social origin, economic condition or birth, has the purpose or effect of nullifying or impairing equality of treatment in education…"); Sarah Joseph, Jenny Schultz, and Melissa Castan, eds., The International Covenant on Civil and Political Rights (New York: Oxford University Press, 2004), p. 532.
 Committee on the Elimination of Racial Discrimination, General Recommendation XXXI, para. 26.
 See, for example, Human Rights Watch, Every 25 Seconds: The Human Toll of Criminalizing Drug Use in the United States, October 12, 2016, http://ift.tt/2e4zmXv; Jamie Fellner (Human Rights Watch), “Race, Drugs, and Law Enforcement in the United States,” commentary, Standard Law and Policy Review, June 19, 2009, http://ift.tt/1Q3NECC.
 UN Human Rights Committee, General Comment No. 18, Non-discrimination (Thirty-seventh session, 1989), Compilation of General Comments and General Recommendations Adopted by Human Rights Treaty Bodies, UN Doc. HRI/GEN/1/Rev.9 (vol. I)(2008), p. 197, para. 7.
 Ibid., para. 13.
http://ift.tt/2oZWlXK Source: https://www.hrw.org