In recent weeks, Daniel Pantaleo and Darren Wilson have both become high-profile examples of white police officers who killed unarmed Black men and got away with it. Tragically, however, they are far from the only ones to have done so. In fact, the history of white cops escaping accountability for their use of lethal force goes back as far as the grand jury process designed to foster that very accountability. The concept originally immigrated to the United States along with British colonists as a way to protect them from abuses by the Crown. But thanks to a combination of built-in structural biases that favor protecting, rather than exposing, violent police culture, what was conceived as a way to hold leaders responsible for abuses of power has instead become a tool to enable them.

Eric Garner’s case is the perfect illustration of this. Despite videotaped evidence showing Staten Island police officer Pantaleo using an NYPD-banned choke hold on Garner while the latter pleaded for his life, a New York grand jury failed to indict Pantaleo on criminal charges in Garner’s death this Wednesday.

How can this be possible? The answers, sadly, seem both easy to explain and impossible to remedy. It’s possible Pantaleo wasn’t indicted in Garner’s killing, just like it’s possible that Darren Wilson wasn’t indicted in Michael Brown’s, because prosecutors never wanted to see them charged to begin with.

When looking to bring criminal charges against anyone, whether police officer or civilian, state-employed prosecutors have two choices. They can bring the charges themselves directly, based on evidence gathered by their office and police officers. Or, they can call a grand jury, which is a group of civilians that hears evidence from the prosecutor concerning the possible crime. The grand jury process varies from state to state, but there are some basic similarities: Grand juries meet in secret and their proceedings are usually driven by the prosecutor, meaning he or she has sole discretion in which witnesses to call, what evidence to put forward, and, often, what criminal charge he or she wants the grand jury to decide on.

As in the Brown and Garner cases, juries can decide whether a violation of criminal law at the state level has occurred. Federal grand juries also exist, but because the U.S. Constitution is generally more protective of individual rights than the states are, that process is a little different.

And this can’t be said enough: grand juries are not trials. Let’s repeat that again. A grand jury is not a trial. That means grand juries are not tasked with determining the guilt or innocence of an accused person. Instead, their only job is to determine whether there’s enough evidence that a crime was committed, and if so, what the appropriate criminal charge should be. If someone is indicted, that doesn’t mean he or she is guilty of a crime; once that happens, the usual criminal justice proceedings move forward, with the indicted person entering a plea and the matter proceeding to trial. It’s there, at trial, where prosecutors face the high burden of “proof beyond a reasonable doubt” of someone’s guilt.

In order to indict someone, however, prosecutors only need to convince the jury that there is “probable cause” to believe that crime was committed at all. This is part of the reason why many people are so enraged about Pantaleo’s non-indictment, for example. In theory, footage of an officer using overwhelming force on a man who was unarmed and who had not acted with violence toward officers should count as “probable cause” that a crime was taking place.

The process of choosing jurors for a grand jury mirrors the usual juror selection process in most instances. For instance, in the case of Ferguson, Missouri, the St. Louis County Circuit Court selects the grand jury from a randomly chosen master jury list, comprised of citizens based on voter registration information, property tax rolls, and driver’s license records. Once a grand jury has been chosen, the prosecutor then presents the state’s case, usually by calling a series of witnesses. The grand jurors are, in general, allowed to ask the witnesses questions, but there are no rules that require any actual eyewitness to an alleged crime, nor the alleged victim of that crime, to appear. Normally, the accused person is not present during this process unless he or she is called as a witness. In fact, the accused often doesn’t even have counsel to represent them during these proceedings, even if he or she is called as a witness.

When it comes to police use of deadly force, though, this whole process often gets turned on its head. First, the state’s job is supposed to be to present a case for prosecuting officers like Pantaleo and Wilson. But, as both those cases show, the opposite is frequently true. Instead of presenting the case for charging the cops, prosecutors instead put forward cases that look more like “justifiable homicide” claims, which present the dead suspect as the aggressor and give the officers involved wide berth to testify to grand jurors. Wilson, for instance, was present as a witness during the proceedings; Pantaleo was as well.

Second, grand juries are also not required to follow the rules of evidence, which means prosecutors are free to present, and rely on, hearsay when seeking a grand jury indictment. Hearsay evidence is kind of like gossip, in that the person testifying can offer up as fact their understanding of what another person did or said, and there is no burden on the prosecutor to actually prove the truth of that witness’s statement. That means a police officer like Darren Wilson can—and did—simply testify as to Brown’s actions during the altercation that led to Brown’s death.

At the close of the state’s case, a grand jury either issues a bill of indictment if it finds probable cause to believe both that a crime has been committed and that the accused person is responsible, or a “no-bill” if it does not find probable cause.

With this is process in mind, it’s easy to see how prosecutors and police officers can work the grand jury system to almost guarantee a non-indictment in cases of excessive policing. To begin with, police officers are the front-lines investigators for the prosecution. Like we saw in Ferguson, where local police did not retrieve Michael Brown’s body for four hours, file a substantive incident report, or even photograph much of the scene, there’s evidently a very strong impulse to not carry out a thorough investigation in these cases. And those investigations would, presumably, be the building blocks of any grand jury charge.

Prosecutors also have institutional pressure to drag their feet and avoid prosecuting cops for excessive force claims. Their prosecution rates for civilians often depend on having good relationships with the police—again, on whose investigations they rely for evidence. And in turn, since prosecutors drive, almost entirely, the grand jury process, including suggesting possible charges, relying on prosecutors and a grand jury to indict a cop can be like relying on Wall Street to police its own.

Institutional, procedural hurdles are not the only explanation for a failure to indict officers in excessive force cases. It is also possible that in Pantaleo and Wilson’s cases, neither were indicted because citizens siting on grand juries—especially white citizens—are willing to support white officers’ excessive uses of force, thanks to latent or not-so-latent racism. Police departments are whiter than the communities they police, while Black men remain the predominate target of state police power. At BuzzFeed, Adam Serwer has this great piece on how the lingering stereotypes left over from slavery contribute to the hyper-policing of Black communities as well as the unwillingness to indict white officers when they police members of the Black community to death.

There is also a disconnect between police training on the use of force and the constitutional rights of citizens. The Constitution protects against the use of unreasonable force, and it seems like a given that killing an unarmed man for selling loose cigarettes like Eric Garner’s case is, by its definition, unreasonable. Even so, according to the Supreme Court, police can use lethal force if they have a “reasonable belief” they are facing danger or if a fleeing suspect poses a danger to others. But some states, like Missouri, allow an officer to use deadly force if the officer “reasonably believes” that force is necessary to arrest a “non dangerous” fleeing felon, which is broader than the U.S. Supreme Court’s ruling on the matter.

In addition to potentially confusing legal standards on when federal and state law permit the use of deadly force, most police officers are trained that a “continuum of force” exists, and that, in theory, an officer’s use of force should match the threat posed by the citizen at that moment. But that training doesn’t actually require officers to start with the least amount of force when confronting suspects. Instead, it defers to the officer’s sense of danger to determine the “reasonableness” of the force he or she uses in a particular situation. Again, this comes down to an arbitrary, split-second decision—and when the laws are murky to begin with, this training can mean unjust death for people who appear “dangerous.”

To make matters worse, law enforcement has also taken a page from the gun lobby and done whatever they could to make data on police killings difficult, if not impossible, to accurately track. That lack of data makes the political case for wide-scale policing reform harder to push in white populations—the very group refusing to indict officers to begin with.

So where does that leave us? If we know police killing of civilians, especially Black civilians, is on the rise, and we know the criminal justice system is especially ill-equipped to deal with police violence, is there any answer—short of burning it all down and starting over—that can provide some actual justice?

For one thing, there’s the possibility of a federal prosecution of both Pantaleo and Wilson for professional misconduct. Federal civil rights prosecutions in police misconduct cases are centered on 18 U.S.C. §242, which outlaws anyone who “under color of any law … willfully subjects any person in any State, Territory, Commonwealth, Possession, or District to the deprivation of any rights.” But in order to do this, federal prosecutors must show that police deliberately deprived victims of their civil rights (including the right to be alive) through excessive force. In addition to conducting their own investigations, federal prosecutors rely on evidence gathered by local officers and investigations taken on by local prosecutors to help build their federal case. Because of this, all the flaws inherent in prosecuting cops at the state level emerge at the federal level as well—meaning federal prosecution of civil rights claims against police are incredibly rare.

Other reform ideas include naming independent prosecutors or citizen panels to handle investigations in cases where citizens are killed by police officers.

Ultimately, however, none of those reforms or possible remedies will make any difference if the underlying structural racial bias in our criminal justice system is not addressed. And to address that structural bias, a majority of white Americans will first have to acknowledge that bias exists. As the stories unfolding from New York to Missouri tell us, this is still a long way from happening.