New York

The Buffalo Massacre Illustrates the Inherent Limitations of 'Red Flag' Laws

Predicting violence is a lot harder than people claim in retrospect, and a wider net inevitably ensnares more innocent people.


In retrospect, it may seem obvious that New York's "red flag" law, which is designed to prevent dangerous people from owning firearms, should have been invoked against the perpetrator of the racist mass shooting that killed 10 people at a Buffalo supermarket on Saturday. But the details of the case show why such laws are not necessarily effective at preventing such crimes, even as a lack of adequate due process protections creates the danger that innocent people will lose their Second Amendment rights.

New York is one of 19 states that authorize court orders prohibiting gun possession by people deemed a threat to themselves or others. New York's law, which took effect in 2019, allows a long list of people to seek such orders. Potential petitioners include police officers, prosecutors, blood relatives, in-laws, current and former spouses, current and former housemates, current and former girlfriends or boyfriends, people who have produced a child with the respondent, and school administrators or their designees, such as teachers, coaches, and guidance counselors. The "school personnel" covered by the law can report a former student if he graduated within the previous six months.

Although the law calls the restrictions it authorizes "extreme risk protection orders" (ERPOs), that is a misnomer. An initial, ex parte order lasting up to six business days can be obtained based on "probable cause to believe the respondent is likely to engage in conduct that would result in serious harm to himself, herself or others." The purported threat need not be "extreme" or imminent. At this stage, the respondent has no opportunity to challenge the claims against him, and the experience of other states indicates that judges routinely rubber-stamp initial orders.

After a hearing, a final ERPO can be issued based on "clear and convincing evidence" that "the respondent is likely to engage in conduct that would result in serious harm to himself, herself or others." A final order lasts up to a year and can be renewed. Again, there is no requirement that the threat be imminent. And while "clear and convincing evidence" is a more demanding standard of proof than "a preponderance of the evidence" (which is enough in five states and the District of Columbia), "likely" is a slippery concept in this context.

The red flag law refers to the definition used in New York's standard for "emergency" psychiatric commitment, lasting up to 15 days, which requires a "substantial risk of physical harm." That is better than the standard prescribed by many other red flag laws, which typically require a "significant" risk and in some cases merely a "risk," "danger," or "risk of danger." But contrary to the connotations of extreme and likely, people can lose their Second Amendment rights even when it is quite unlikely that they would use a gun to harm themselves or others.

Judges may consider "any evidence," and respondents have no right to legal representation if they cannot afford it. Nor do they have a civil cause of action against petitioners who lie, a potentially significant problem in light of all the people who are allowed to file a petition. The lack of such a deterrent increases the risk that an in-law, cousin, ex-spouse, ex-girlfriend, or former housemate with a grudge will abuse this process.

So far, The New York Times reports, state judges have issued 589 final ERPOs, which amounts to about 18 orders per month. The Times says "the court system does not track the number of applications that were denied," which makes it hard to assess whether judges are making a serious effort to distinguish between well-grounded and weakly justified petitions.

State Sen. Brian Kavanagh, a Democrat who represents parts of Manhattan and Brooklyn, was a lead sponsor of the red flag law. He thinks the Buffalo massacre raises questions about whether the legislation has been properly implemented.

"We passed the law specifically to ensure that people who exhibit signs of being dangerous to themselves or others can be denied access to guns," Kavanagh told the Times. "There's a real question about whether that law was effectively applied when this person was apparently detained after making threats….It's important that we understand what happened there and that every county in the state is implementing the red flag law that we passed."

When he mentions "threats," Kavanaugh is referring to a June incident in which the Buffalo shooter, then a 17-year-old senior at Susquehanna Valley High School near Binghamton, responded to a question about post-graduation plans. He wrote that he "wanted to commit a murder-suicide," the Times reports, citing "a law enforcement official briefed on the matter." While he passed that off as a sick joke, his subsequent posts on the messaging platform Discord showed it was actually a slip that revealed his true intentions.

"I got out of it because I stuck with the story that I was getting out of class and I just stupidly wrote that down," the future Buffalo shooter wrote in an invitation-only chat room. "That is the reason I believe I am still able to purchase guns. It was not a joke, I wrote that down because that's what I was planning to do."

The school incident was reported to the New York State Police, which ultimately decided not to seek an ERPO. The Times notes that "hundreds of school threats are called in annually." In response to such reports, it says, "authorities interview students and parents to determine whether students have access to guns" and "then try to make a reasoned call on what action to take."

In this case, the Buffalo shooter underwent a mental health evaluation. "I had to spend ~20 hours in that ER waiting for somebody to give me 15 minutes to talk to me," he wrote on Discord. "This proved to me that the US healthcare system is a joke." The Times says he was "released within a couple of days."

Under New York's Mental Hygiene Law, such psychiatric evaluations are supposed to assess whether a patient has "a mental illness for which immediate observation, care, and treatment in a hospital is appropriate and which is likely to result in serious harm to himself or others." The law defines "likelihood to result in serious harm" as a "substantial risk of physical harm to himself as manifested by threats of or attempts at suicide or serious bodily harm or other conduct demonstrating that he is dangerous to himself" or "a substantial risk of physical harm to other persons as manifested by homicidal or other violent behavior by which others are placed in reasonable fear of serious physical harm."

In this case, the examination may have been cursory, but its conclusions did not preclude police from seeking an ERPO, which does not require a psychiatric diagnosis. Even when someone does not meet the standard for involuntary treatment, he might still be "likely to engage in conduct that would result in serious harm to himself, herself or others."

Judging from the information that has been reported so far, however, the only basis for thinking this high school student posed a danger was the written comment that he claimed was a joke. That explanation evidently seemed plausible to the police, especially since the teenager had not mentioned any specific target or plans and at that point had yet to purchase the rifle he used in the attack.

That incident, the Times says, followed "a pattern of increasingly bizarre behavior." Specifically, "two former classmates said he showed up to class in hazmat gear after pandemic restrictions were lifted in 2020." Even if police were aware of that stunt, which presumably was intended as a commentary on COVID-19 precautions, it was consistent with the impression that they were dealing with a joker rather than a would-be murderer.

Citing Cassaundra Williams, a former student at the high school, the Times also says the future mass shooter "favored online coursework even as his classmates returned to campus." Again, that is hardly a sign of homicidal intent.

"He was always very quiet and never much said anything," Williams said. She added that "he was 'book smart' but had grown more reclusive over the years since she met him in elementary school." That description might apply to many introverted, awkward adolescents.

"We were just so shocked," Williams said. "We can't even wrap our heads around it still." That comment does not suggest that the shooter's fellow students viewed him as a threat.

"Even the few people close to [him] appeared to have little inkling about what was in store," the Times reports. It quotes Matthew Casado, a friend who had known the shooter since second grade. "Until Saturday, I always knew him as a good person," said Casado, who is Latino. "He was never racist towards me, or around me. I didn't know he was racist."

It may yet turn out that a more thorough investigation could have turned up troubling evidence that was not apparent to friends or fellow students. But as always in such cases, there is a risk that hindsight bias will obscure the limited information that police had at the time, which did not include the teenager's recently revealed Discord posts or the manifesto in which he described his plans.

While "the Buffalo shooter didn't have a previous criminal record," Vox reporter Nicole Narea writes, "he had made serious threats of violence that were brought to the attention of police." But those "threats," which apparently were limited to the written description of his post-graduation plans that he passed off as a joke, were serious only in retrospect.

From our current vantage point, it is easy to argue that the massacre might have been prevented if the state police had been more aggressive. But there is a downside to automatically erring on the side of seeking a red flag order, as illustrated by a Florida case I covered a few years ago.

Chris Velasquez, a college student who did not own any guns, had no history of violence, and had never threatened anyone nevertheless was an early target of Florida's red flag law because he said some stupid, deliberately provocative things about mass shootings on Reddit. Because of that trolling, university police searched Velasquez's phones, laptop, backpack, car, and bedroom, all with consent and all without discovering any evidence of homicidal intent. Velasquez's father surrendered his own revolver to the police.

Velasquez, like the Buffalo shooter, was required to undergo a psychiatric examination, which likewise concluded that he did not meet the criteria for involuntary treatment. He was interrogated by an Orlando police officer, who repeatedly pressed him to admit nonexistent homicidal plans. A sergeant later wrote an affidavit that falsely portrayed Velasquez as a ticking time bomb who had threatened to attack schools.

That was enough to obtain a temporary "risk protection order" against Velasquez. But after taking a closer look at the evidence, the same judge declined to issue a final order, saying the city "did not meet its burden to prove by clear and convincing evidence that Respondent poses a 'significant danger.'"

With a less skeptical judge, the case easily could have turned out differently, meaning that an innocent person would have lost his Second Amendment rights for a year or more. Kendra Parris, Velasquez's attorney, also has represented a man who was targeted for talking about guns online and a man whose pistol was confiscated based on a relative's disputed account of an argument. Another Florida case I covered in 2020 involved outlandish accusations by an estranged wife that fell apart at a hearing where the same sheriff's office that had obtained a temporary order conceded that it could not meet the test for making it permanent.

This is what aggressive enforcement of red flag laws looks like. People who favor that approach rarely consider the consequences for individuals who can be deprived of their Second Amendment rights based on little more than unvalidated allegations. Given the "hundreds" of reports that police in New York receive each year just in the school context, they certainly could be seeking and obtaining more ERPOs. But a wider net is bound to ensnare many people who do not actually pose a threat, including high school students who really are guilty of nothing more than making a stupid joke.

There is no easy answer to striking the right balance in this area. But the costs that red flag laws impose on innocent people are unavoidable, while their effectiveness at preventing violent crime remains speculative.

"Because almost all [red flag laws] have been enacted within the past 10 years," the Times says, "there is limited research on their effectiveness." But the oldest red flag law, Connecticut's, has been around since 1999, and Indiana enacted one in 2005. So far there is no solid evidence that such laws reduce violent crime, and research on their potential benefits in preventing mass shootings relies on counterfactual assumptions that could easily be wrong.

The very concept of "red flags" assumes that experts can reliably distinguish between harmless oddballs and future murderers. But there is little basis for that assumption.

"The notion that we can identify mass killers before they act is, as yet, an epidemiologic fiction," psychiatrist Richard Friedman noted in a 2019 New York Times essay. "These individuals typically avoid contact with the mental health care system. Even if they didn't, experienced psychiatrists fare no better than a roll of the dice at predicting violence."

That view jibes with the conclusions that University of Georgia law professor Alexander Scherr reached in a 2003 Hastings Law Journal article. "Over thirty years of commentary, judicial opinion, and scientific review argue that predictions of danger lack scientific rigor," Scherr wrote. "Scientific studies indicate that some predictions do little better than chance or lay speculation, and even the best predictions leave substantial room for error about individual cases. The sharpest critique finds that mental health professionals perform no better than chance at predicting violence, and perhaps perform even worse."

2012 study that the Department of Defense commissioned after the 2009 mass shooting at Fort Hood in Texas includes an appendix titled "Prediction: Why It Won't Work." The appendix observes that "low-base-rate events with high consequence pose a management challenge." In the case of "targeted violence," for example, "there may be pre-existing behavior markers that are specifiable." But "while such markers may be sensitive, they are of low specificity and thus carry the baggage of an unavoidable false alarm rate, which limits feasibility of prediction-intervention strategies." In other words, even if certain "red flags" are common among mass shooters, almost none of the people who display those signs are bent on murderous violence.

Supporters of red flag laws prefer to ignore this problem. After a mass shooting in a state that has such a law, they argue, as in this case, that it would have worked if only it had been used properly. But the problem goes deeper than that. However you weigh the risk of preventable violence against the risk of taking away innocent people's rights, this policy has inherent limitations that mean it is bound to fail in one way or, more likely, both.