Policy Studies Alternatives to Arrest

The conservative case for jail reform

The attached policy study originally appeared in the Fall 2017 edition of National Affairs.

Vishnu Phagoo was released from jail in May 2014 after spending three years awaiting trial for the murder of his sister. Few would pity an accused killer, but Phagoo was not guilty, a fact a jury would determine after only four hours of deliberation. Why was Phagoo in custody for so long? Because his family was unable to pay his bail, set at $750,000. Following his acquittal, Phagoo lamented that the authori­ties “need to pay for what they did to me … I didn’t see or talk to my daughter in those years. I lost out on her life, my family’s life.”

But just 11 months later, in April 2015, Phagoo was back in jail, this time on charges of biting, kicking, and punching a relative at a family gathering. Why would he risk wasting more years of his life in a jail cell? One possibility is that a long stint in a jailhouse makes a person — in­nocent or otherwise — more violent. Before his biting incident, Phagoo told reporters, “I got into fights [in jail]. . . .You are with criminals, rob­bers, whoever. [If they] don’t have what you have . . . they steal and if you don’t stand up for yourself you are just food [to them].” In addition to the emotional scars, Phagoo had also left jail with scars around his neck, from where another inmate had attempted to strangle him.

Of course, it may be that Phagoo is just incorrigibly violent, and that it is better to incapacitate him through incarceration. But his story is representative of the correlation many experts have found between jail stints and reoffending. It is also indicative of a system at odds with some core American principles.

The Republican Party has long declared itself to be the “party of law and order.” And in recent decades, its commitment to being tough on crime has yielded lower crime rates. Both property crime and vio­lence are far less common today than they were in the late 1980s, when Michael Dukakis lost the 1988 presidential election to tough-on-crime George H. W. Bush, and the rate of violent crime has declined by almost half since peaking in the mid-1990s.

But these improvements in public safety have come at the cost of bloated jail and prison populations. Jail is the first stop for many in the criminal-justice system. Over 11 million people go to jail each year in this country, and there are currently 2.3 million people under some type of confinement. Local jails hold defendants — the ones who cannot afford or are denied bail — before a trial, and also detain persons convicted of lower-level crimes carrying sentences of usually less than a year. Over time, jails have increasingly become detention centers for citizens who are legally in­nocent; more than 60 percent of people sitting in jail cells on a given day have not been convicted of a crime and are awaiting trial. Given the presumption of innocence the founders embedded in our Constitution, a tension now ex­ists between reality and our values. The Republican Party in particular touts “limited government” and “the rights of the people” as bedrock principles; it is hard to reconcile those principles with a lock-them-up approach to criminal justice.

There is a host of reasons why conservatives should be concerned about the effects of U.S. corrections policies and the front end of our criminal-justice system. We are now dealing with the cultural fallout of decades of these policies. Nearly one-third of working-age Americans have a criminal record, making it much harder for them to get jobs. Incarceration separates offenders from their families, which increases rates of homelessness and single parenthood. Approximately 17 million children are currently being raised without a father, a growing social problem that only perpetuates cycles of violence and crime.

While the sheer number of people incarcerated alone should shock the consciences of conservatives, our pretrial-jail system in particular challenges core Western principles of justice and liberty. The Magna Carta in 1215 declared that “[n]o freeman is to be taken or imprisoned … or in any way ruined … save by lawful judgement of his peers or by the law of the land.” There are currently 630,000 individuals in local jails throughout the United States, almost 450,000 of whom are awaiting trial. Pretrial detention retards a defendant’s case. And, desperate to be released, many accused will accept unfavorable plea deals; even the innocent will plead guilty to escape the misery of jail. What’s more, the people awaiting trial behind bars are often the wrong people — they tend to be not the most dangerous, but the poorest.

Public safety should be of paramount concern in any kind of criminal-justice reform. As such, we must ensure that pretrial detention is used for those who pose a threat, rather than those who cannot afford bail. Fiscal responsibility should also be part of any policy discussion; taxpayers, who pay to fill and run the bloated jails, are then squeezed again to pay for the societal effects of an overly aggressive system.

Most important, we should pursue a more just and moral criminal-justice system. It’s hard to advocate the moral importance of following the law when the system in place to enforce that law is less than moral itself.

These competing ideals and priorities do not point to a simple set of answers. But they do add up to a need for jail reform, and it is time to take that need seriously.

Principles of Government

“You are under arrest.” There are few phrases that could better capture the power of government. With these words, a government agent can physically restrain a citizen, commit acts of violence if the citizen re­sists, and even kill, if such resistance intensifies. Government agents are granted latitude for error in such judgment calls. In light of the power given to law enforcement, it is critical to ensure that such power remains limited within the bounds of a fair system.

The notion of a limited government should be dear to all Americans. The founding fathers purposefully and specifically rejected the model of government that forces citizens to trust in the benevolence of a mag­nanimous state, and suspicion of the sovereign is at the heart of how we tend to think about politics. American history is full of examples demonstrating that such skepticism is warranted: from the internment of Japanese-American citizens during World War II, to several of the military campaigns against Native Americans, to the testing of the ef­fects of radiation on unknowing U.S. military personnel.

In addition to being limited, the government the founders created was meant to be consistent and fair in how it dealt with its citizens. They were inspired by John Locke, who was concerned about a govern­ment that acted arbitrarily when the freedom and fortunes of its people were at stake. Locke wrote in his Second Treatise of Government, “The legislative, or supreme authority, cannot assume to itself a power to rule by extemporary arbitrary decrees, but is bound to dispense justice.”

Unfortunately, there is ample evidence that the jail system has several arbitrary and capricious facets. In many jurisdictions, an individual’s ability to be released before trial is linked directly to his wealth, of­ten with little or no consideration of the dangers he might pose to the community or the risk of flight, which are supposed to be the main rationales for pretrial detention.

Languishing in jail cells places incredible pressure on defendants to accept plea bargains, regardless of their actual guilt. It is rare for a de­fendant, even one who professes his innocence, to reject a plea deal that allows him to escape jail. Indeed, one of the injustices of the current jail system is that it pushes poor defendants to end their pretrial deten­tion with a plea, as a means to prosecutorial efficiency and informal punishment. Looking back to Phagoo’s case, he also had been arrested for attempted arson before the murder charge was leveled against him. In that case, he claimed he pleaded guilty to attempted arson, despite maintaining his innocence, because he could not coordinate bail and was offered release in exchange for a guilty plea. At this point, it is im­possible to determine the veracity of Phagoo’s claim, but his story is one that has been told countless times by countless defendants. If only a small portion of such stories are true, it is too many.

Moreover, individuals locked up as part of a pretrial-detention ap­paratus receive significantly longer sentences upon a conviction or plea deal than similarly situated defendants who are released pretrial. Recent research commissioned by the Laura and John Arnold Foundation found that even detained defendants deemed low-risk (based on factors that include ties to community, past criminal behavior, and seriousness of the pending charge) receive harsher sentences compared to those who are released, when all other factors are considered.

Reasons for these disparities include the increased leverage the gov­ernment has in securing “favorable” plea deals from those who are locked up, as well as the fact that incarcerated defendants face increased difficulty building a defense with their legal counsel. By comparison, de­fendants who are free during pretrial can prove themselves trustworthy by not committing any more crimes and attending all court appoint­ments. They thus can more often be deemed suitable for probation, instead of prison, if their case ends in a conviction.

None of this is to say the system as a whole is unjust. But it would be unbefitting our democracy to turn a blind eye to those injustices that do occur. Conservatives should be able to concede that certain economic biases are inevitable, which can lead to uneven enforcement of the law. And an arbitrary and capricious government can’t be a limited one.


In addition to such straightforward injustice, the current jail system also undermines the very purposes for which it exists — especially public safety and the reduction of crime.

Research has found that, after only two or three days in jail, low-risk individuals are almost 40 percent more likely to commit new offenses than individuals who are held for 24 hours or less. The effect is also magnified the longer an individual is detained. Low-risk inmates who were held more than 31 days are 74 percent more likely to reoffend than those released within 24 hours. It isn’t entirely clear why this is the case; loss of em­ployment, housing, and problems created within the family all may be contributing factors. By contrast, for high-risk individuals, studies show no correlation between time in jail and the likelihood of new crimes.

Looking at the other goal of pretrial detention — ensuring defen­dants appear in court — we see the same pattern. Low-risk individuals held two to three days are 22 percent likelier to fail to appear at court appoint­ments than similarly situated defendants held for 24 hours or less. Those held for 15 to 30 days were 41 percent likelier to fail to appear than those held 24 hours or less. For high-risk defendants, the length of time in jail had no significant impact on their likelihood of missing court.

These statistics could be interpreted to reflect that the poorest among us also lack the ability to follow the rules, perhaps due to a lack of com­munity, family support or education. Under this theory, those with a propensity to commit crimes after release also will lack the sophis­tication to make bail within 24 hours — hence the strong correlation between time in pretrial detention and recidivism. But were this strictly the case, there would not be such a drastic difference between low-risk and high-risk defendants. Instead, this difference supports the evidence suggesting that time spent in jail is linked to a deterioration of public safety, particularly in the pretrial stage of criminal proceedings.

The length of an individual’s stay in jail has effects that reach be­yond short-term recidivism rates. The Arnold Foundation study found after controlling for other variables, the more time an individual spent in jail pending trial, the more likely it was that he would commit crimes for as long as two years after the case had completed. Individuals held approximately three days were 17 percent likelier to engage in criminal activity within two years than those held for less than 24 hours. For those held four to seven days, there was a 35 percent increase in likelihood of reoffending. If a defendant was held eight to 14 days, he was 51 percent likelier to commit new crimes within two years.

Those individuals who go to prison almost always serve lengthy sentences and return to society older and with a lower propensity to commit crimes. Indeed, individuals typically cease committing prop­erty crimes like burglary or grand larceny in their 20s. Most violent offenders cease their offending behavior in their early 30s. Individuals age out of their criminal behavior for numerous reasons. Studies in neu­roscience tell us that the brain’s functions that govern risk and reward are still developing for individuals into their mid-20s. Moreover, poverty is closely linked to criminal activity (though this does not hold true for many white-collar crimes), and young people are more likely to be poor. Age also obviously limits some offenders’ physical capacity to commit certain crimes, particularly violent ones.

Because jail inherently is designed for short-term custody, virtually all jail detainees — except for those who will plead guilty to or be con­victed of serious crimes before being moved to the state or federal prison system — will be released back into the community while they are still young and healthy. They thus still have all the attributes that helped land them in jail in the first place. This amplifies the importance of the link between time detained in jail and recidivism.

Compounding this problem is the fact that jails typically provide much less in rehabilitative programming than prisons do. Thus the public-safety benefits of pretrial criminal incapacitation appear to be offset by the increase in crime that stems from turning petty offenders into hardened criminals.

Moreover, even brief periods in jail are associated with disturbing rates of suicide. This is due not only to the shock caused by the sudden loss of freedom, but also the fact that defendants often lose their jobs and their basic sense of normalcy. Many of these suicides occur before the cases are over. Indeed, the rate of suicide is seven times higher for pretrial defendants than for those who are convicted.

Ensuring that only those defendants who truly pose a public-safety risk are locked up ought to be a key reform embraced by the conservative movement. Marc Levin of the Right on Crime coalition summarized the point best:

Thus, the stakes are high, as is the potential for government to do more harm than good. Yet, traditionally, it has been those on the left who were primarily concerned with overincarceration, including in the pretrial context. Yet, among the most ascendant principles in the modern conservative, and especially Tea Party, movement are limited government and individual liberty. It is increasingly difficult to square current pretrial policies and proce­dures with these principles. Therefore, it is perhaps not surprising that a growing consensus is emerging that, even as a relatively small number of violent and dangerous individuals must be kept behind bars pending the resolution of their case, we must not allow pretrial incarceration to grow in scope like so many other government programs, which often begin with a core function and then metastasize over time.

Such growth in pretrial incarceration is extremely costly to both those incarcerated and the public at large. In financial terms alone, the mass incarceration of low-risk pretrial defendants is inexcusable. This is es­pecially true given that how we currently run our jail system does not actually keep us safer — and that release rates can be increased with little risk of new crimes. Indeed, holding low-risk defendants longer than 24 hours appears to increase the likelihood of new crimes, not only before the disposition of the current case but also years into the future. Conversely, the same observation is not true of high-risk defendants.

Judges and pretrial services make noble attempts to separate low- and high-risk individuals, but they often lack empirically driven risk-assessment tools. Indeed, throughout the United States, only about 10 percent of courts employ data-driven risk-assessment tools to make pretrial-detention decisions. Judges are thus left to their own devices. This, too, is problematic from a fiscal perspective, because it costs taxpayers approximately $9 billion annually to house, feed and protect pretrial defendants. Ensuring that judges have proper information to separate low- and high-risk individuals is a necessity.

On any given day, 2.3 million people are incarcerated across the United States. Of that staggering number, 27 percent are sitting in local jails. The majority are in the pretrial stage, and many of them are simply too poor to post bail. The pretrial-defendant population has driven a 99 percent growth in jail populations over the past 15 years. Throughout the course of an average year, 11 million people churn through our jail system.

This churn costs states dearly. Texas is seventh highest for incarcera­tion rates but highest overall in total number of incarcerated individuals at more than 223,000. Of that 223,000, the jail population is 66,000, second highest in the country. Approximately 24,000 of those are low-risk individuals awaiting disposition of their cases. On average, it costs $59 per day to house one inmate in Texas; thus, the state spends slightly less than $1.5 million a day, or $520 million a year, to house their 24,000 low-risk offenders.

In California, which is No. 1 for total jail population, it costs about $190 a day to house inmates. With more than 82,000 total inmates in jail (both low- and high-risk), California spends more than $15 million a day to house its jail population, which amounts to over $5.5 billion per year — and that does not include the cost of housing the nearly 130,000 held in California’s state prisons. This is a staggering sum. Consider, for instance, that the state funds the entire University of California system for $3.3 billion a year.


Republicans have long struggled with negative public perceptions. Many on the left view conservatives, and the Republican Party that claims to represent them, as racists who disdain the poor and cater to the rich. Conservatives in particular have had their political identity shaken since the election of Donald Trump. Jail reform offers a perfect opportunity to regain a sense of purpose and moral clarity.

The Republican Party was born from the 19th-century anti-slavery movement. Abolitionists looked to both the Democratic and Whig Parties for support and, finding little, turned to forming grassroots organizations. By midcentury, the slavery issue began to splinter the Democratic Party and contributed to the collapse of the Whigs. Out of these political ashes, the Republican Party — the party of modern conservatives — was born. That party should strive to be particularly sensitive to the injustice of capricious incarceration.

The connection of the Republican Party to the abolition movement is also relevant because African-Americans (and Hispanics) are hurt dispro­portionately by the current system. Together, they represent nearly 60 percent of the total incarcerated population, though they account for only 29 percent of the whole U.S. population. Conservatives should recognize the damage this does to minority families and communities — and the moral harm it does to the nation. Republicans have had a bad reputation for caring more about big business and the wealthy, and for having little compassion for the struggles of average Americans. If they would like to shake that reputation, they should start by showing that it is wrong.

A political case for taking such concerns seriously can reinforce the moral case. Among black voters, President Trump captured only 8 percent of the vote. This is comparable to both President Ronald Reagan in 1984 and President George W. Bush in 2000, who each received 9 percent of the black vote. Among Hispanic voters, President Trump did comparatively poorly, winning over only 28 percent, a significant fall from Reagan and Bush, who captured 34 percent and 35 percent, respectively. The na­tion’s long-term demographics underscore the need to arrest this trend, as the Hispanic population continues to grow at a steady pace. When Reagan was elected president in 1980, there were 14.8 million Hispanics, amounting to 6.5 percent of the total population. Today, there are more than 55 million Hispanics, representing more than 17 percent of the population; by 2050, they will comprise nearly one-third. If they want to continue having a voice in the political arena, Republicans must make an effort to serve all their constituents.

But more important are the moral arguments for reform. While conservatives certainly do not hold a monopoly on Christian values, Christian interest groups do hold special power within the conservative movement — a fact sometimes decried on the left by those who fear con­servatives attempt to use the political system to impose religious values.

Jail reform is a place where both sides can come together. It ap­peals to the social-justice warriors who are vocal on the left, and is fully consistent with the Christian message that is such a large part of the conservative identity.

A limited government is essential to the conservative political ideol­ogy and is the bedrock of liberty. Yet, many on the right appear to pick and choose when outrage is appropriate. When it comes to the Second Amendment, conservatives are unwilling to allow any movement, for fear of the slippery slope. However, when it comes to the power of the state to rob citizens of their freedom, the same passion for a limited gov­ernment and the instinct to be suspicious of power is too often absent from the conversation.


Local leaders have recognized the need for jail reform, and there is a wide range of reform efforts currently playing out across the states. Among efforts to reduce jail populations and increase public safety, some of the most promising reforms include behavioral-health diver­sion, police diversion and objective risk assessments.

Behavioral health among inmates has become a particularly press­ing issue. Local jails are increasingly becoming health-care facilities; data from the Bureau of Justice Statistics show that 64 percent of jail inmates have a mental-health problem. According to a former county sheriff, Los Angeles County jails are “the largest mental health provider in the country.” The problem, of course, is that jails are woefully unprepared and ill-suited to handle the administration of mental-health services.

Jail is a uniquely terrible place for those with behavioral-health prob­lems. Confined, chaotic, and violent jail conditions predispose detainees with mental-health problems to alarmingly high recidivism and miscon­duct rates. Moreover, housing mentally ill offenders in the criminal-justice system is costly, and we spend far more on imprisonment of the mentally ill than we would spend on treatment. A 2011 report by California’s Task Force for Criminal Justice Collaboration on Mental Health Issues found that, in 2008-09, the average prison inmate cost California about $51,000 per year (and those with mental illnesses often cost far more), whereas the yearly cost per individual for a supportive housing program in Los Angeles was $20,412. Such program costs vary widely, of course, but more cost-effective support options should be explored further as an alternative to incarceration for the mentally ill.

More often than not, detainees with mental-health problems are jailed for minor offenses, such as trespassing, disorderly conduct, public drinking, disturbing the peace or illicit drug use. It makes little sense to throw the mentally ill in jail for petty crimes, where their symptoms are exacerbated and they incur significant costs to taxpayers.

A growing body of research suggests that mental-health court-diversion programs can reduce both crime and incarceration rates among the mentally ill, and many counties across the country have taken steps to adopt such reforms. In Pima County, Arizona, where mental illness and substance abuse affect 60 percent of the jail population, of­ficials decided to implement behavioral-health screenings before pretrial detainees’ initial court appearances. This simple step gives judges safe alternatives to incarceration for nonviolent offenders with significant behavioral-health afflictions. Post-booking, judges divert these select individuals into rehabilitative programs that seek to address the root causes of delinquency. The county also monitors diversion participants after the program with health experts who can ensure a smooth transi­tion back into society.

Through its plans, Pima County aims to save taxpayers more than $2 million annually and reduce its jail population by 18 percent over the next three years. Washington, D.C., implemented a mental-health court in which participating defendants were 36 percent less likely than the comparison group to be arrested again in the year after completing the program. These types of programs are typically reserved for nonviolent offenders who still face charges and will go to trial; the difference is these programs attempt to rehabilitate those with illnesses instead of putting them in a jail cell.

Similar to these mental-health court-diversion programs, police-deflection programs can offer robust benefits to taxpayers and low-level offenders. These programs rely on police officers’ knowledge of their neighborhoods. As the first line in the criminal-justice system, they are best suited to evaluate who poses the greatest risk on the street.

In 2011, Seattle launched Law Enforcement Assisted Diversion (LEAD) in the downtown neighborhood of Belltown. Based on success­ful arrest-referral programs in the United Kingdom, the LEAD program allows police officers on the street — who know the neighborhood’s drug dealers, users and prostitutes — to decide who receives immediate access to treatment and other services.

The program targets “frequent flyers” in the criminal-justice sys­tem — the chronically homeless, low-level dealers, drug users and prostitutes. A survey of the LEAD program found that the 54 individu­als most frequently contacted by police in Belltown had been arrested a collective total of 2,704 times. Those repeat, low-level offenders often have particular needs that require an alternative to jail; these individuals are hand-selected by a group of police officers and assigned casework­ers who can provide a hot meal, a warm coat, and a safe place to sleep. The program provides longer-term services for drug and mental-health treatment, stable housing and job training.

Ten cities across the country have now implemented LEAD or similar programs, and dozens more are in the development process. Like court-diversion programs, police deflection can alleviate jail overcrowding and address the sources of misbehavior, which also reduces recidivism.

Finally, objective risk assessments have also proven promising. In deter­mining whether to release or detain a defendant, pretrial judges have the unenviable job of making predictions regarding human behavior. Will this person show up for trial, and will he abstain from committing new, violent crimes while he is out on release? Getting this wrong means that judges will either detain an innocent individual, at high cost to that per­son’s freedom and livelihood (and to taxpayers), or the judge will release a dangerous defendant who may commit a heinous crime before his trial.

The judicial profession is full of erudite, well-meaning individuals who have a uniquely tough job. That they get things wrong from time to time is no mark against them. It is, however, the inevitable product of their human fallibility, cognitive myopia and various subconscious biases. But technology can compensate for human idiosyncrasies and produce more just outcomes. One simulation that tested a risk-assessment algorithm based on machine learning found that such programs could cut crime up to 24.8 percent with no change in jailing rates, or reduce jail populations by up to 42 percent with no increase in crime rates. These gains applied across the board, including for minority populations.

Today, there are approximately 60 risk-assessment algorithms used across the United States. Assessments differ in form, length and con­tent. The simplest tools use only criminal records, while others conduct a defendant interview and integrate the results into the algorithm to pro­vide a risk score. The score produced by the algorithm correlates to a risk level — usually broken up into high-, medium-, and low-risk tiers. Those tiers are associated with a recommendation for release or detention. New Jersey has had success with this strategy; the state recently reformed its bail system and replaced it with risk-assessment software. The result has been a 19 percent decline in the jail population in the first five months alone.

These tools can save localities and states millions of dollars annu­ally while at the same time lowering detention and crime rates. The innovation represents more than simple reform — objective risk-assessment tools, as they transition from traditional statistical models to more sophisticated machine-learning programs, could revolutionize the pretrial process.

Data-driven risk tools are not without controversy, however. Judges themselves may resist an increased reliance on risk algorithms, but virtu­ally every researcher and developer of these tools stresses that they ought to be used to supplement, not replace, judicial opinions.

A more pressing concern is that these tools could be biased. One well-publicized study released by ProPublica suggested that a particular algorithm used in Florida was racially biased, though there have been several other studies that have shown potential for algorithms to re­duce bias in pretrial decisions. Reforms must always be compared to the status quo, so the most pertinent policy question is whether actuarial programs are more biased than human predictions. Proper safeguards can ensure that this remains unlikely.

Lastly, there is the issue of transparency. If private companies pro­vide assessment tools, they may resist disclosing their proprietary algorithms. Defendants, then, may be unable to discover how a tool ar­rived at its recommendation. This is a valid constitutional concern that merits attention. Possible remedies exist, however, such as mandating open-source codes and regularly auditing tools for bias. If a particular algorithm has a racial bias, it should be fixed. An algorithm is only as fair as the code that runs it. But the beauty of such codes is they can be rewritten and updated as new information comes to light.

Despite these concerns, it’s clear that data-driven risk tools have tremendous potential to improve the fairness and efficiency of the criminal-justice system. The question should not be if we should imple­ment technological advances into pretrial decisions, but how.

Given the tremendous financial and moral costs of our current ap­proach to criminal justice, the moment for reform is here. Millions of defendants — many of them innocent — languish in jail awaiting trial simply because they cannot afford to pay their bail, while more dan­gerous offenders with greater financial means walk free. This is both perverse and tremendously expensive. There are proven paths to im­provement, and conservatives should recognize that such reforms can create a system more in line with their values.

Image by sakhorn


Featured Publications