Washington Post investigative journalist Radley Balko just published a story about a Little Rock, Arkansas police unit conducting no-knock raids with “explosive entry”—literally blowing the doors off of homes—with little or no justification. It turns out many of these drug raids were based on the word of a confidential informant with a lengthy criminal record. As Balko lays out in his story, the informant and the officers who relied on him are unreliable, at best, and very possibly criminal. While the Little Rock story is extreme in some ways, the use of confidential informants to make drug busts invites injustice and police abuse.

The rules for using confidential (also called “criminal”) informants in criminal investigations vary from jurisdiction to jurisdiction, but generally speaking, employing CIs introduces three systemic flaws into the criminal justice system. First, the use of confidential informants definitionally requires secrecy and opacity, which shields CIs and officers alike from sufficient oversight and accountability. Second, the informant system relies on bad inputs—namely, drug-addicted individuals and other people immersed in criminal activity to act as agents of the government—and thus effectively becomes a subsidy for criminal behavior. Third, the use of confidential informants creates some bad incentives for law enforcement actors and the CIs themselves, which skew toward case production and away from public safety and security. Taken together, and in the context of our everyday justice system, these flaws produce an array of bad individual and public policy outcomes while providing only superficial benefits for law enforcement.

Secrecy and Unaccountability

From the outset, confidential informants stand in contrast to the original purpose and structure of our criminal justice system. The U.S. Constitution, and particularly the Sixth Amendment, contemplate a criminal system based on fairness and openness. All criminal trials are public record, not to shame the accused, but for the government to present its case to the people at large as fair and just. Moreover, the accused has the rights to counsel, a speedy trial before an impartial jury, and to confront (i.e., challenge) any witnesses who testify against him.

The term “confidential informant,” on the other hand, is just another way to say “secret source,” also colloquially known as a “snitch,” inside and outside of law enforcement circles. They provide criminally implicating information about others to police, most often maintaining their anonymity and escaping cross-examination. As law professor Alexandra Natapoff explained in her book Snitching: Criminal Informants and the Erosion of American Justice, “[S]nitching flows from two dominant characteristics of our criminal justice system: plea bargaining, and a tolerance for a high level of law enforcement discretion.” While there is case law that protects the anonymity of confidential informants at a criminal trial, roughly 95 percent of all criminal convictions nationwide are the result of plea bargains, so most potential cross-examinations of witnesses are precluded because trials are so rare anyway.

Plea bargaining exists independently from CIs, but it helps insulate law enforcement sources from public examination because the prosecutor has already enticed (or coerced) the defendant to waive their right to a jury trial, the right to appeal, and many other rights to take a reduced sentence. In a perfect system, procedural shortcuts would be a positive hallmark of efficiency, assuming that only the guilty are punished and fairly so, but in reality, plea bargains invite injustice and wrongdoing because there are almost no checks on police or prosecutors’ actions. Without trials, confirmed liars can help law enforcement build case after case, because the defendant never gets an opportunity to challenge the liar’s story.

Bad Inputs and Bad Actors

While TV shows like “The Sopranos” portray how informants can be used to infiltrate organized criminal networks like the Mafia, in real life, most CIs are used at the local level to make street-level drug and firearms deals, to build cases against suppliers. They are typically recruited when they are caught selling or using drugs and, again, either convinced or coerced into working with the police to catch the proverbial “bigger fish.” This method of recruiting snitches will then tend to draw young, low-level dealers and problem drug users, which each pose distinct problems when they become informants.

Sarah Stillman wrote a piece for The New Yorker in 2012 that told the stories of several individuals who had been enlisted by police into informing on other dealers. One informant, Rachel Hoffman, was a 23-year-old Florida State graduate who was tasked with buying a large amount of cocaine, some Ecstasy pills, and a gun from a drug dealer after her second arrest for marijuana distribution in Tallahassee, Florida. Three weeks after becoming an informant, Hoffman’s body was found murdered, alongside the gun the police had sent her to buy. The police took an unnecessary risk with her because, frankly, she and those in her situation are relatively expendable, particularly when compared with undercover narcotics officers in whom police departments have invested significant personal and institutional capital.

On the other hand, individuals like Ventura “Benny” Martinez, who became a star informant for the Philadelphia Police Department, are duplicitous con artists that always seem to be running some grift to score pity, drugs, money, or another benefit. Martinez was a key figure in the Philadelphia Daily News’s Pulitzer Prize-winning series “Tainted Justice,” and the subsequent book Busted by journalists Barbara Laker and Wendy Ruderman. Laker and Ruderman chronicled how Martinez and the narcotics squad he worked with took down both legitimate and illegitimate scores. The officers got their arrest numbers up, while Martinez got paid for the high volume of targets he ratted out for police. Martinez’s remarkable number of scores, among other things, provided cover for the narcotics squad’s own illicit activities—including shaking down bodegas and other legitimate merchants for cash—because high-producing units earn favorable reputations and are less likely to be second guessed. While Martinez hardly got rich from his efforts, his ongoing substance abuse and petty schemes were subsidized by the PPD while bad cops tarnished the badge by enriching themselves.

Policing Must Be More than Arrests

Of course, most informants aren’t murdered or providing cover for dirty cops. Yet, the use of CIs for drug cases is itself a type of systemic corruption that turns police units into arrest producers rather than public guardians. Part of this problem may stem from basic managerial decision-making.

Like many jobs, police supervisors and leadership can judge officers by certain metrics, such as tickets issued for traffic cops or arrests made in open homicide cases. Police routinely deny that arrest quotas exist, but there is ample anecdotal evidence that vague benchmarks are common, and in specialized units, such as narcotics squads, they would be expected to arrest individuals somewhat frequently. It only makes sense, then, that officers who are incentivized to make arrests will try to meet expectations. Providing opportunities for offenders to stay out of jail and then paying informants per arrest is an efficient way to ensure those expectations are met.

According to the Centers for Disease Control, in 2016, an estimated 10.6 percent of residents 12 and older reportedly used illicit drugs in the previous 30 days. That means millions of Americans continue to use illicit drugs, providing police officers with a virtually endless supply of targets for arrest. Yet CIs are more often used in economically depressed black and Hispanic neighborhoods, arguably because they suffer from higher violent crime rates, but such targeted drug enforcement further adds to the problems affecting those communities. Typically, these communities already suffer strained relationships with the police. Traps to arrest residents—or, worse, to recruit more informants—are more likely to damage police-community relations than strengthen them. As a result of this, and of other aggressive policing methods, officers become very adept at policing high-crime communities for minor offenses but lack the community trust needed to solve major crimes like homicide when they occur. Put another way, aggressive, arrest-driven policing may lead to less safe neighborhoods rather than safer ones.

Deadly Drug Raids

It is in this environment of deception and arrest production that a recent wrong-door drug raid happened in Prince George’s County, Maryland. While that incident ended with two officers being wounded by the innocent resident who fired at them in self-defense, other informant-driven raids have had even more tragic results. Kathryn Johnston, a 92-year-old woman in Atlanta, was killed by police who were acting on a bad informant’s tip. Ryan Frederick was sentenced to 10 years in prison for killing Detective Jarrod Shivers, who was part of a team conducting a drug raid on Frederick’s residence in Chesapeake, Virginia. Frederick claimed he started sleeping with a gun by his bed because his home had recently been burglarized and he mistook the police for unlawful intruders. The raid was precipitated by a 20-year-old informant who had recently broken into Frederick’s place to find evidence of drug distribution. And just a few years ago, an infant in Georgia was maimed by a police flash-bang grenade that was thrown into her crib during a raid based on bogus informant claims.

These examples are only a few of the innumerable accounts of police raids, triggered by CI information, that undermine the security of residents and endangers the police and community. What’s more, CIs obscure the intended transparency of our justice apparatus, exploit and subsidize criminally involved individuals, and perpetuate a system that values arrests over public safety and well-being. The mistrust their use sows in communities erodes police effectiveness in other realms, like homicide clearance. Using confidential informants to enforce drug prohibition is an endless Sisyphean chore that imposes high social costs, particularly in the United States’s most vulnerable communities, without ridding them of the underlying danger—drugs—that brought the police there in the first place. Indeed, the police often indirectly support an informant’s drug habit to make more drug arrests.

Given the violence posed by organized criminal syndicates, such as MS-13 and the Aryan Brotherhood, recruiting confidential informants already involved with those organizations may be one of several imperfect and more dangerous options that should be available to law enforcement. But those are extraordinary circumstances—and very often, interstate federal operations—to disrupt particularly dangerous organizations whose criminality goes beyond drug distribution. In contrast, CIs should not be cultivated by officers and narcotics squads simply to “make their numbers” and prove to their bosses in the brass that they’re doing something.

The common sense answer to these problems is drug legalization or decriminalization, thus removing the criminal law from the street-level drug trade. The most serious advocates for legalization recognize that drug abuse is nevertheless a legitimate policy problem, but one to be handled with drug treatment and other programs to alleviate the circumstances that drive the substance abuse and crime in the most at-risk communities. Such programs may work in concert with law enforcement, such as diverting drug addicted individuals to treatment after committing a crime, but using CIs to seek-and-arrest is helping no one outside of the station house.