Long before George Floyd’s death this past May, police reform debates have centered around de-escalation. The interest has stemmed, mainly, from high profile police shootings of mentally ill individuals. Reformers hope to prevent such incidents by giving cops special training to de-escalate crisis situations.
De-escalation’s not universally popular. Defund the police advocates would prefer diverting resources to social services. Academic researchers have cast doubt on whether de-escalation reduces shootings. City budget deficits pose the greatest threat to de-escalation. Getting de-escalation right costs money, and the COVID-19 pandemic has pounded governments’ finances. With signs emerging of diminished commitment to de-escalation, now’s a good time to rethink the case for it.
The goal of de-escalation programs is best described as reducing the number of fatal police shootings of mentally ill individuals that could be prevented with de-escalation. We don’t know how many shootings fit that bill. The Washington Post has tracked all fatal police shootings since 2015 and coded them based on "signs of mental illness." It also codes them based on whether the victim was armed and/or attacking the officer. If one defines shootings that de-escalation could have prevented as those in which the victim was mentally ill, unarmed, and not attacking the officer, the count appears to be around 5-10 incidents per year.
De-escalation trained officers have shot to death mentally ill individuals. De-escalation-trained officers have themselves been shot to death by mentally ill individuals. Some departments have seen the number of police shootings of mentally ill individuals increase after adopting de-escalation. On September 13th in Lancaster, Penn., a mentally disturbed man was shot to death after charging at a cop with a knife. We should discuss how to prevent tragedies such as the incident in Lancaster. But de-escalation is the wrong answer to that question.
Now, surely some police shootings of victims who were technically “armed,” but containable, could have been prevented. Examples include the June 2014 death of Jason Harrison in Dallas and the October 2016 death of Deborah Danner in the Bronx. One 2018 analysis found that half of all police shootings of mentally ill people happen at home, where the containment potential should be the greatest. “No one wants Rambo answering a mental health call” says Pete Earley, the author and prominent advocate. But if there’s only a small handful of true “Rambo”-style incidents each year, de-escalation’s benefit will be hard to quantify. Further complications, from a program evaluation standpoint, include that de-escalation is often brought online amidst other mental health reforms and the extremely decentralized nature of American policing. There are around 18,000 public safety agencies nationwide. De-escalation programs can vary widely in quality, based not only on the nature of the training itself but who gets it. Certain officers are going to be better at de-escalation than others. A de-escalation training program that’s given to every cop in a department is not likely to be a very high-quality program. And yet, half of all departments in the nation employ ten or fewer cops; over 80 percent employ fewer than 50. To maximize the availability of de-escalation trained cops, small departments may have no choice but to train everyone.
If we don’t know whether de-escalation works, why do it? Progressive critics offer two alternatives to reducing police shootings of mentally ill individuals. Invest more in “upstream” social services or respond to mental illness-related calls for service with emergency personnel who are not armed cops.
An effective mental healthcare system would, in theory, eliminate all mental illness-related police shootings. Mental health reform is the right answer to how to prevent cops from shooting to death mentally ill individuals who are attacking them with knives. But before investing more money, we should ask why the current mental healthcare system’s not more effective. The police reform debate is supposed to be about accountability. The 200-250 police shootings of people with mental illness every year, plus the hundreds of thousands mentally ill people in jails and prisons stand as abundant evidence of system failure. Yet who’s ever forced to resign their office when a city’s jail population sees an increase in the number of schizophrenic inmates? Which nonprofit loses its contract when a psychotic man is shot to death for attacking a cop with a knife? Many mental health service providers have little interest in working the type of person who would attack a cop with a knife. A system that’s geared mainly towards reducing incarceration and homelessness among the mentally ill, and would accept the consequences of failure, would merit increased investment at a very large scale. But that is not the mental healthcare system we have.
Now let’s consider de-escalation critics’ second recommendation, about replacing police as first responders with other emergency personnel. This recommendation is more sensible. Erratic yet innocuous behavior can be sometimes mistaken for dangerous behavior, and with fatal consequences. Perhaps responding to mental health crisis calls with unarmed social services personnel or, simply cops with community knowledge, would reduce the risk of thinking harmless mentally ill people are threats. Governments could direct 911 to dispatchers decide whether to send out cops or social workers, or a mix of both, or set up a separate hotline for mental health crisis calls.
We don’t know how effective any of that would be, though. It may well be true that the majority of mental health crisis calls don’t need police to arrive as first responders. But it’s likely also true that the minority that do are those that hold the greatest potential for fatal outcomes. Earlier this month, a Manhattan, Kansas man was shot dead by police after mental health responders failed to de-escalate a crisis situation after hours of effort. Not all mental health crisis calls are coded as such by the dispatcher. Sometimes, all that’s known by the first cops on the scene is there’s been a report about a potential firearm or someone armed with a knife. If governments set up alternative dispatch systems, will everyone call the mental health crisis number instead of 911? (In the recent Lancaster Penn. incident, one of the victim’s relatives did call a local non-emergency help line while another relative called 911.)
We should not accept mental illness-related shootings as inevitable and should debate solutions. We should also acknowledge that all solutions will have evidentiary problems. Good programs shouldn’t be conflated with weak programs. To do de-escalation right, oversight and funding are necessary. State or regional authorities need to exercise some oversight over programs’ quality. Many de-escalation proponents emphasize the importance of a community-orientation, which points in the direction of local control. But localism can’t get in the way of program integrity. A community should not be allowed to decide on a superficial de-escalation program. Sending a handful of cops to a daylong mental health first aid course shouldn’t count as de-escalation. Ideally, the training should consist of a full 40 hour week course for cops with at least a few years of patrol experience. Special teams should be selected for dispatch purposes based on officers’ interest and dedication.
The training itself doesn’t necessarily need to cost very much. Some highly-regarded de-escalation training programs use mainly volunteer instructors. But a de-escalation program with integrity allows patrol officers to spend two hours resolving a single call if that’s how long it takes. In the 2019 documentary Ernie & Joe: Crisis Cops, de-escalation officers in San Antonio do check-in visits with people months after assisting them on a 911 crisis call. Communities should reconcile themselves to longer response times to non-mental health crisis calls if that allows for more resources to dedicate to de-escalation. Slashing a police budget to make a political point, even if it does not directly affect de-escalation training proper, could do deep harm, indirectly, by preventing officers from dedicating the time they need to use the training.
De-escalation’s about owning the idea that police work is social work. We must reject the idea promulgated by an unholy alliance of progressives and pro-cop traditionalists that police should only focus on “serious” crime. The mental health reform debate needs the involvement of people from criminal justice. Because that’s who deals with the hardest cases: beat cops in midtown Manhattan, staff at jails and prisons, judges and district attorneys. Instead of taking the attitude of “step aside, cops, real social workers will take it from here,” de-escalation engages the police in the mental health reform debate. Mental health reform should be left to the experts, but cops are the experts. Recognizing that helps to define the challenge and point towards solutions.
Stephen Eide is a senior fellow at the Manhattan Institute and contributing editor of City Journal.
