Policing Can Be Disrupted
It is difficult to identify a single innovation in modern policing that was counter-intuitive at the time of its introduction. Here’s one.
A central complaint lodged against any effort to reimagine policing is that doing so will undermine public safety. Over the years, a variety of cost-effective alternatives to traditional policing have been proposed but not implemented. Here, I propose a real innovation to police practice, one that dramatically increases the likelihood of solving a psychologically important crime, that dispatches fewer armed law enforcement officers to crime scenes, that has a false positive rate close to zero, and that is relatively cheap to operate. This is not a new idea I just conjured—it has been in the literature for a decade. But not implemented, because of deep inertia. Read on…
***
American policing is under siege, and this should be a very good thing. Despite exponential growth in technology, the American policing model is more or less identical to the 1970s version. Police conduct their business from patrol cars, look for trouble, and respond to radio calls for service. This is the policing model of the famed ADAM-12 drama of the early 1970s and it is overwhelmingly the model today in America’s 18,000 law enforcement agencies. I have no doubt readers will respond with examples of innovation from a department with which they are closely familiar, and there are doubtless many innovations to be found.
But the protests and the civil unrest reflects a broader reality—that little has changed, that the anger is broad and deep, that over-policed people of color remain perpetually over-surveilled, over-sanctioned and oppressed. At the margins, perhaps policing is a little different but only in the direction of more, not less. Militarized SWAT teams are executing more warrants more aggressively and data-led policing is targeting increasingly narrow slices of space and time, albeit within the same geography of the usual suspects that have dominated for decades. Changes in policing reveal no deep self-reflection, no self-awareness that any particular way of doing business should change toward less. Indeed, it is difficult to identify a single innovation in policing that was counter-intuitive at the time of its introduction.
Why has widespread policing innovation been limited only to expanded use of military-gear and tactics? This derives from a knee-jerk response to policing criticism, which is: we don’t know what the right answer is, but we know you are wrong. Any effort to propose a reimaging of policing is met with a fusillade of counter-attacks, based on around the premise that when the re-imaginer feels threatened, they will call the police and are thus hypocritical in their thinking. This is a haymaker of an argument, and it generally connects.
Progressive opponents of policing would do well to acknowledge the limits of their own approach. The protests have pulled back the curtain on policing and revealed the structurally racist core that lies within, and this should be cheered. And I do! But, the next step goes wanting. Any ideas around reform are met with a similar broadside of criticism. The haymaker of an argument here is that anything short of disbanding modern policing preserves structural racism. And this punch lands with equally devastating effect.
So, policymakers are left to consider two competing haymakers-as-arguments, and no real solutions. And so what do policymakers do? They wait. They wait for a consensus to emerge for a compromise. But as long as ‘I don’t know what the right answer is but I know you are wrong’ dominates the discourse, there is no compromise and no policy change. And eventually the policy window closes, the status quo wins again, and we wait for the next conflagration to see if maybe we can do better this time.
Solutions
I think the problem of structural racism in policing has four solution areas.
Prevention. America has the wealth to lift people out of desperate situations that led to crime and violence. We know what the risk factors are: discrimination, poverty, lack of supports in health, mental health and substance abuse, lack of education, inequality, declining economic, social mobility and stagnant social and emotional learning. And more. These are the risk factors and there are solutions. America simply needs to choose to invest.
Who answers the call? Perhaps the most progress on reforming policing has come from an acknowledgement that a call for help and a call for armed police are not the same call. When a house is on fire, a call to 911 is answered by the fire department, not the police. The same logic can be applied to many other types of service. CAHOOTS is the leading example of someone else answering the call, but it is not the only answer.
What do police do and what don’t they do? How do police spend their time, and how should police spend their time?
When is police violence ok? America has armed its police to the teeth so that they are capable of overwhelming violence. But no Violence Doctrine has accompanied the armament. There is nothing Geneva Convention-like on police use of force. There should be.
These are all very fruitful paths to reimagining policing with a lot of common ground. Everyone is for more prevention (but no one wants to pay for it). Patrol officers will be the first to tell you they are ill-equipped to answer many calls for service, especially when the call leads them into contact with someone in a crisis. Police violence is harder, because we are nation with an extraordinarily well-armed citizenry, and police have to have the means to protect themselves. But the status quo of excess death and disproportionate use of force against people of color is immoral. But perhaps the most fruitful area of inquiry in terms of immediacy is changing what the police do and how they do it.
Ending Residential Burglary as we know it
One idea that is obviously right that gets little attention is that Americans are despairingly afraid for their personal safety. Here’s a recent industry report that finds that Americans spend more than $50 billion each year on security. That’s everything from commercial security for businesses and offices parks, to do-it-yourself Internet of Things solutions. That’s $50B after we pay for police and prisons, the military and data security. And it is increasing: the security industry expects a 50% increase in spending on security in just the next five years, to $78.9B.
Aside from what they spend, there is also what they say. Almost every year, more than half of all Americans report to Gallup that they believe crime has increased from the previous year. As Maya Angelou said, “When someone tells you who they are, believe them the first time.” Americans are telling you they are afraid. You should believe them.
It turns out though, that we don’t believe them. Or, more accurately, perhaps we believe them but we think they are wrong. Talking heads and academics spend a lot of time dismissing and diminishing Americas fear of crime. Some typical headlines:
Many Americans Are Convinced Crime Is Rising In The U.S. They're Wrong. (FiveThirtyEight.com)
Americans don’t know crime has plummeted. In fact, they think it’s gone up. (Vox)
Voters’ perceptions of crime continue to conflict with reality. (Pew Research)
So, what are Americans afraid of? On theory kicking around is that this is an effect of the post-9/11 focus on security. Certainly physical security investments increased massively after 9/11. There was also a huge increase in federal spending on security and non-military federal policing. Color coded threat warnings dominated the airwaves, and the 2004 election pundits spoke earnestly about ‘Security Moms’ and their outsized role in shifting partisan politics toward safety measures. The data seems to support this story as well.
The red circle highlights 9/11/2001, and it is clear that there is a reversal in the long term trend toward less fear of crime in America—a decline in fear in the 1990s that closely tracked the actual decline in crime. Americans obviously have the ability to reasonably link actual declines in crime with real declines in crime. It is just that after 9/11 a moderating variable appeared that did not exist before—the threat of terrorism. It is an existential threat, and there is little or more likely nothing that the average person can do to prevent an attack or reduce their own risk. So it is reasonable that American’s mentally transfer this risk to something more tangible.
That’s one story. Another story that is equally clear in the same data is that Americans fear levels seem to track pretty closely with property crimes. The pre-9/11 declines in property crimes closely match the declines in fear levels, and both decline a little bit more than the decline in overall violence. This suggests it worth thinking about whether something is happening with property crimes that is affecting Americans fear levels. Criminologists tend to use homicide as the barometer for ‘crime’ in general and ‘violence’ in particular—mainly because it is the most accurately measured crime type. But we know that homicide tends to occur between people who know each other, and there are relatively few homicides (5.4 per 100,000 in 2016) compared to say burglary (469 per 100,000). So Americans have a lot more experience with property crimes then violence and perhaps are more sensitive to changes in those types of crime.
After 9/11 declines in both property and violent crimes slow way down, but continue. But fear levels are way up, almost to 1990 levels. So what is going on, what are Americans afraid of?
When you ask Americans about relatively more common types of victimization than homicide, it turns out they are afraid of having their home broken into or their car stolen. I wish Gallup asked this question prior to 2000, but the trend here is pretty clear. There is a hint of a substantial decline in perceived risk of having a home burglarized or a car stolen in 2000-2001 as part of the long-term crime decline. After that, perceived risk of victimization remain steady throughout the next two decades, perhaps falling again after 2014.
So, let’s look at the crime data. Two of the three crime types show big declines in this period that are not reflected in the fear data. Motor vehicle thefts are way down in this period, down 42% between 2000 and 2010 and another 41% by 2016. Muggings (robbery) are down 20% between 2000 and 2010 and another 17% by 2016.
What fits the data much better is the change in burglary. Among all major crime types, burglary declines the least in this period. Between 2000 and 2010, burglary declines by less than 4% which correlates to the unchanged sentiment in the same period. Burglary, however, declines a lot between 2010 and 2016 which is the one segment in the fear time series that shows a decline.
This is hardly definitive evidence of residential burglary as the principal mechanism to explain Americans fear of crime. We spend huge portions of our lives in our homes, far more than on the streets, and it makes good sense to believe that perceived threats to the home are an important driver of levels of general fear of victimization.
What Does Fear of Burglary Have to do with Reimagining Policing
As we reimagine policing, a critical place to start is with a review of how police spend their time (my take is here). Modern policing does a very poor job of investigating and solving burglaries. Clearance rates for burglary have been stubbornly stuck on about 13.8% for decades (meaning that only 1 in 7 or 8 burglaries result in either an arrest or a definitive conclusion by the police that they know who the identity of the perpetrator.
The combination of low clearance rates for a psychologically important crime has a devastating effect on policing legitimacy. A key reason why burglary clearance rates are so low is that police expend little effort to investigate burglary. Burglary is seen as a low priority call, because it is rarely solved (the obvious endogeneity problem is that burglary is rarely solved because it is a low priority investigation). In practice, officers involved in the burglary study I led on behalf of the National Institute of Justice reported that if a suspect was not found at the scene, or was identified by a witness, the case was essentially closed more or less immediately, and certainly within a couple of days.
There is an innovative solution to this problem that has been known for a decade. Among non-sexual property and violent crimes, burglary is uniquely suited to forensic investigation, specifically the collection and testing of DNA. DNA evidence is often left behind at crimes scenes (burglars touch things without gloves, but more importantly they eat, drink and smoke and leave other DNA samples about). That evidence is probative—unlike a cigarette butt found in a recovered stolen auto or a shell casing with DNA at a shooting, DNA evidence at a crime scene directly connects the suspect to not only the place of the crime but the time as well, and leaves no confusion about whether the evidence was left accidentally.
What happens when DNA is found at a burglary crime scene? Finding DNA at a crime scene doubles the likelihood that an otherwise unknown suspect is identified and doubles the likelihood of an arrest. In the study city (Denver) with the greatest fidelity to best practice, those numbers were even better, with a clearance rate three times business as usual investigation.
These results were even better than the clearance rate for investigations that relied solely on fingerprint evidence, Burglary cases investigated with recovered DNA evidence were five times as likely to identify an unknown suspect as were cases investigated with fingerprints alone, and, nine times more likely to yield an arrest. And going against stereotype, the suspects were serious, chronic offenders. Suspects identified with DNA-led investigations averaged 2.5 prior felony convictions for all types of crime—burglary, assault, sexual assault and homicide. The intuition of burglary as a ‘gateway’ crime, that is a stepping stone to real offending is not supported in the data.
And, finally, using DNA to investigate a crime scene is cost-effective. In Denver, again the city with the most fidelity to best practice, the additional cost per new arrest was $1,500. That is, $1,500 to arrest with strong evidence someone for burglary, who has more than two prior felony convictions. This is a true bargain in a bloated criminal justice system.
So here is how it would work. A call for a burglary comes in to 911. IF dispatch determines the burglary is in progress, police are dispatched. If it is not in progress, evidence technicians are dispatched instead (evidence technicians are generally just that, technicians and they are not sworn law enforcement, are generally not armed, and do not have the power of arrest). The evidence tech processes the scene, the evidence is submitted to the FBI for comparison against the national database of (generally) convicted offenders (CODIS) and the suspect is identified (with almost no margin of error). Work obviously needs to be done on the next step—serving an arrest warrant without harming innocent people. But the process greatly reduces risks of wrongful policing, and greatly increases the legitimacy of policing in the process of getting this case right, and getting more cases right more often.
Coda
One obvious criticism of this approach is that it relies on the use of the DNA database (CODIS) for a list of potential suspects to match. And that list of potential suspects is not, of course, generated at random. It is a list of people who have previously been arrested (in some states), have previously been convicted of a felony (many states) or have previously been convicted of a serious felony (all states). If you believe as I do that there are serious racial disparities in each of these steps (arrest and conviction) then the CODIS database is itself racially biased. This is not up for debate. One solution would be to include everyone in the database…
Book
I suppose everyone who makes book recommendations has recommended a Gentleman in Moscow. It is a study in grace in a perfect metaphor for the moment. I recommend it without qualification.
Musical Interlude
To all those who believe no reform of policing is possible, today or ever, this is for you.