The Evidence on Policing and Crime

Subscribe here: Apple Podcasts | Spotify | YouTube | Overcast | Pocket Casts

The police murder of George Floyd in 2020 brought radical ideas—such as defunding police or abolishing police departments altogether—into mainstream policy conversations. Activists highlighted instances of police brutality and the unjust treatment of unarmed citizens, right as violent crime began to skyrocket during the pandemic.

In a New York Times op-ed from that year famously titled “Yes, We Mean Literally Abolish the Police,” the prominent organizer Mariame Kaba argued that police were not set up to go after “the worst of the worst” criminals, and that they spend most of their time “responding to noise complaints, issuing parking and traffic citations, and dealing with other noncriminal issues.” Kaba also argued that the cost of this “safe” world was that racial minorities were kept “in check through threats of arrest, incarceration, violence and death.” Kaba had lost faith in reform efforts that had failed to prevent deaths like Floyd’s, representing a growing perspective that policing and the criminal-justice system were making people less safe.

But although a majority of Americans remain critical of the criminal-justice system, voters have recently backed so-called tough-on-crime ballot measures, such as one in California calling for harsher penalties for theft and drug crimes, and another in Colorado involving parole eligibility for violent offenders.

So what do the data tell us about the connection between policing and safety?

On today’s episode of Good on Paper, I asked that question to Jennifer Doleac, an economist and expert on criminal-justice policy who is the vice president of criminal justice at Arnold Ventures, a foundation focused on evidence-based policy making.

---

The following is a transcript of the episode:

[Music]

Jerusalem Demsas: A little over a year ago, Gallup asked Americans to choose between two approaches to lowering the crime rate: first, direct more money and effort to go into addressing social and economic problems like drug addiction, homelessness, and mental health. Or option No. 2: more money and effort for strengthening law enforcement. Almost twice as many respondents chose “address social problems.”

Now this is a bit of a false dichotomy. People didn’t have an option to answer “both,” and it is possible to invest in both of these. But there are monetary constraints, particularly for the local governments in charge of most policing. What this poll indicates is a widespread belief that the preferable crime-fighting strategy isn’t focused on policing.

This is a complicated conversation. The question of whether police keep us safe raises many more questions before it can be answered: How are they keeping us safe? Safe from what? And, of course, who is “us”?

[Music]

My name is Jerusalem Demsas. I’m a staff writer at The Atlantic, and this is Good on Paper, a policy show that questions what we really know about popular narratives.

Today’s guest is Jennifer Doleac, an economist and expert on criminal-justice policy who is currently the vice president of criminal justice at Arnold Ventures, a foundation focused on evidence-based policymaking.

Jennifer has been working on these issues for years. And I asked her to come on the show to unpack what we know about the impact of policing on safety: Is it true that investments in social-welfare programs can make police obsolete, as many activists have called for in recent years? Does increasing the number of police officers actually make us safer?

Here to answer those questions, and more, is Jennifer. Welcome to the show.

Jennifer Doleac: Thank you so much for having me.

Demsas: So before the pandemic, it seemed that crime rates were on this long-term secular decline, right? Since the ’90s, murder and violent crime were down significantly nationwide. Property crimes were down nationwide. It felt like we were maybe approaching utopia.

We had a bunch of great conversations about, Oh, all the criminal-justice reform we can have now, given that crime rates are really low. More people were amenable to that. What explained that long-term decline, before we get into what happened during COVID?

Doleac: We don’t really know, which I think then made the uptick during COVID more concerning. There are a bunch of different theories. One is: We did a bunch of things during the late ’80s and early ’90s in response to rising crime during that period. And so we hired a lot more police. Prison sentences got longer—all kinds of things. And so probably some combination of that worked, and crime started falling.

But there are also other theories about, you know, there were dividends to the war on poverty that had happened during the ’60s, and the next generation was much better off. We took lead out of gasoline, and so lead exposure for young kids was much lower. And so when those kids grew up, they were less likely to commit crime. There’s just a whole bunch of stuff that probably combined to pay big dividends over time and contributed to that long-term decline that we saw in crime.

Demsas: And can you draw out—this is one of my favorite things—the connection between lead and crime. How does lead impact criminality?

Doleac: Yeah. I should be better at explaining this because I talk about it all the time, but I will caution that I’m not a scientist. The reason lead is so damaging is that it mimics calcium in the body. All of our receptors in the brain that look for calcium—basically lead latches onto those and crowds out the calcium. And so it affects brain development in that way.

And then it winds up leading to more impulsiveness and aggression and learning disabilities. And so it winds up reducing educational attainment, and that itself could increase criminal behavior later if you have worse labor-market options. But because it increases impulsiveness and things of that nature, it increases violent behavior during the teenage years. And so we see an uptick in arrests for, especially, violent crime beginning for teens and into early adulthood.

Demsas: If you were to make a bet, what sort of theories are most attractive to you? Because from my perspective, given that policing is super local in this country—we do have state troopers and things like that, but you just have all these different police departments all over the country—there’s tons of variation in the way that they’ll hire people and the policies that they have.

Given that that’s true, and you still see this long-term secular decline, I’m kind of pulled towards these more non-police policy explanations for things than the ones that are referencing specific changes that departments may have made.

Doleac: Yeah. I think that’s my hunch, too. I still think it’s probably a combination of stuff. We’re not going to find the one thing we did that led to this big decline.

But I think overall, I’m really convinced by the evidence on lead exposure and the impact that lead exposure has on later criminal behavior—and so not necessarily research around the crime decline in the ’90s, but subsequent research and just exposure, in general. And so it’s compelling to me that that could have had a really big effect, as well as these other anti-poverty measures that we took.

But I think you’re right that things like policing are very local, but they do respond and are funded heavily by the federal government. And so through COPS grants, for instance, the federal government put a lot of money into local communities to hire more police at that time. So there was this sort of grand experiment where we increased the amount of policing during that period. It wasn’t just one-off local decisions.

So yeah, I think my hunch is the same as yours, that it’s probably the other stuff that was going on then that contributed the most. But I don’t think we can totally rule out that some of the crime-reduction measures we intentionally put in place were effective.

Demsas: So I want to switch tacks here a bit because the thing that pushed me to invite you to come on the show, even though you’re someone I’ve been thinking about for a while to bring on, is you tweeted something, I think a couple weeks ago. And you wrote: “One of the most consistent research findings in the [criminal-justice] space is that hiring more police reduces crime. We need more and better policing, not less—and that will require more funding.”

And this is, I think—depending on your media diet—either the most anodyne statement that could possibly be made, or it’s extremely controversial. I want you to walk us through, how do we know—when you say something is one of the most consistent research findings in this space, what are you drawing from there? What are some seminal studies that you point to or meta-analyses in this space that have pushed you to believe that hiring more police definitely is causally reducing crime?

Doleac: Yeah. I love this research area both because I think the punchline is surprising to most people, and I always like telling people research that can change their mind, but also because the really strong causal research really changed the field’s perception of this area.

So it used to be that we—just based on strictly correlational evidence—you regress crime on the number of police officers, and you look at the correlation between those two things, and there’s no negative correlation. So it looks like, maybe, police cause more crime. But it turns out that that’s not telling you the causal effect of crime, because places with more crime will tend to hire more police officers. And so you’ll get this positive correlation, but it doesn’t mean police are causing the crime. It’s because the crime is causing the police, basically.

And so then the credibility revolution happened in economics, and people got really serious about looking for natural experiments that could distinguish correlation from causation and measure these causal effects. And suddenly, every study that is serious about measuring that causal effect using these natural experiments finds, consistently across a whole bunch of different contexts, that hiring more police, putting more police on the streets in various ways, reduces crime.

And so the way that they’re doing this—there are, I would say, two broad buckets of these studies: One is looking at actual hiring of police officers. The natural experiment is that the federal government gives out COPS grants to enable local departments to hire more police. And so departments apply for this funding, and then some get it; some don’t. There’s some sort of ranking of applications, and the ones just over the threshold get the funding, and the ones just below don’t.

And so using that kind of setup, you can look at the departments that just barely got the funding, compare them with the departments that just barely didn’t. And the departments that got the funding hire additional officers, so the money matters and is used for its intended purpose. And you see crime fall, especially violent crime. So that, I think, is very compelling evidence that hiring that additional officer reduced crime rates. That was really the only difference between those departments.

The other bucket of studies—

Demsas: I’m sorry—is that across all crime rates, or is it violent crime? What’s it looking at?

Doleac: People look at all different crime rates. I think, overall across these types of studies, there are effects on both types of crime rates, but it’s most consistent and largest for violent crimes, especially homicide.

I think, you know, one additional caveat there is that putting more police around could increase the reporting of crime. So you might wind up seeing, you know—there could be different effects that are canceling each other out for the reporting of theft, for instance. Maybe theft is falling, but it’s reported more often when it does happen, and so you get this more muted effect on net. Homicide is very consistently reported, and so it’s generally a really good measure of what’s happening with violent crime.

Demsas: And what’s sort of the magnitude we’re talking about here, like, per police officer or whatever? What kinds of declines in violent crime are we seeing?

Doleac: Okay. So the best estimate is that you prevent one murder a year for every 10 to 17 police officers that are hired.

Demsas: Okay. So that’s pretty expensive.

Doleac: It’s expensive, but also, I mean, murder is expensive, right? Murder is really costly. So I think most economists think of that as a pretty big effect.

Demsas: Okay. And then the second bucket of studies?

Doleac: The second bucket of studies is on how you allocate police officers, or thinking about police presence. And so for instance, there are a lot of randomized controlled trials of hotspot policing. So that’s where you would pick a bunch of higher-crime areas of a city, and you put police officers on this street corner, and you don’t put police officers on that street corner. And then you see what happens across the two street corners. And crime goes way down when the police officer is standing there.

And so that, you know, is very consistent with an overall idea that what’s happening here is that police are deterring crime. Just the presence of having more police out and about is having a big deterrent effect on crime, because you’re increasing the probability that people will get caught, and people respond in a very consistently large way to that.

Demsas: So I think that if we hadn’t seen really prominent instances of police brutality and political movements kind of form to respond to that, this would be probably a pretty widely accepted finding. But I think that there’s a sense that these sorts of findings miss the types of violence that police themselves might perpetrate. Is there research that looks into the question of how much police brutality increases when you have increased police officers, or is it the case that more police officers reduce the incidence of police brutality? Like, how do we see that effect kind of play out?

Doleac: Yeah, we don’t have great evidence on this, because police use of force is not consistently measured. Even if it’s collected internally, it’s not consistently shared with researchers. And so I think there’s a lot of effort right now going into measuring that sort of outcome, but it’s been harder to get our arms around it because the data just isn’t as easily available as reported crime.

But there have been a bunch of efforts to measure things like the effect of a police killing, for instance—especially what seems like an unjustified police killing if the victim didn’t have a weapon—on people who live in the local area, especially young people and young Black and brown people who live in the local area. And you see that, you know, educational attainment goes down. Interestingly, voting goes up, which is perhaps good news—people respond to that perceived injustice.

But yeah, I think one big development over the last several years is that researchers have become much more serious about trying to quantify what the costs of different types of police behavior are so that we can compare them with the benefits. And I would say that, right now, the real policy and research frontier is figuring out ways to maintain the benefits that we get from policing while we mitigate the costs.

Demsas: Can you help decompose the effect here a bit? Because I’m interested in the mechanism by which more police are reducing crimes. What are the pathways that this is happening? Is it the case that, you know, with the hotspots, is it just the visibility? They’re there, so people are like, All right, I’m just not going to do this right now, because a cop’s there, but I’m going to wait for the cop to be gone, or I’ll go somewhere else and commit a crime. Or are they really good at solving crimes, or what’s actually happening here?

Doleac: Yeah. So there could be these two different channels, right? One is pure deterrence that you see—that there’s an increased probability of getting caught, and that is what you’re responding to, or you have this sense that you’re more likely to face consequences for your actions, and so you don’t commit the crime.

The other one is incapacitation. And so that’s where, you know, the police are just arresting everyone who’s committing crime, and so there’s no crime anymore. And what studies consistently show is, using the kinds of natural experiments, that you hire—you get a COPS grant, and you hire more police officers—and so crime goes down, but arrests don’t go up. And so that is pretty consistent across all of these studies.

And so that leads to a general conclusion that what really seems to be happening here is that police presence—I sometimes globally refer to it as almost like a scarecrow effect. Like, you just put a cop on a corner, right? And you do see sometimes in a lot of cities. Generally, cops hate doing this—but sometimes their job is just to sit in their car in a high-risk area of town just to kind of be there. But you know, they’re assigned to do that because it’s really effective.

Demsas: My friends in New York sometimes will complain that, you know, NYPD in the subway is just sitting there scrolling on their phones, but is your sense then that those people are actually really helping to deter crime?

Doleac: Probably. Yeah. I mean, maybe they would deter crime more if they looked like they were looking around instead of at their phone, but yeah, I think that’s the theory.

Demsas: So there’s this well-established idea in the criminal-justice literature that certainty of punishment is effective at deterring crime. So it’s not about making someone go away for 20 years, but if they know they’re going to get caught for committing a crime, it reduces the likelihood that they’ll commit further offenses.

How does that interact with something you once wrote, which is that the best available data from the U.S. suggests that two-thirds of people released from prisons will be caught again within three years? So if people who are caught, arrested, and punished are likely to reoffend, how does that interact with arguments that certainty of capture and punishment is what reduces crimes?

Doleac: I’ve never thought about it in those terms, but I think my general reaction to that is that people are probably committing a lot of crime before they get caught again. And so the general problem is that right now, for most crimes, the probability that you’re going to get caught is actually very low.

If you just look at clearance rates—which we could have a whole separate conversation about what clearance rates really measure, etcetera, but it’s basically the probability that the police are going to make an arrest for a particular type of crime—the highest category is homicide, which is about 50 percent nationally. It’s a coin flip whether or not you’re going to get caught for committing homicide. When you look at stuff like motor vehicle theft, it’s, like, 10 percent.

And so there is lots of evidence that people are deterred much more by the probability of getting caught than the punishment. That is because most people who are on the margin of committing crime are just not that forward looking. And so it doesn’t matter what the consequences are in 10 years. You know, adding five years to an already-long sentence, people are thinking about tomorrow, maybe, right? And so increasing the swiftness and the certainty of consequences, even if the consequences are relatively limited, will have a much bigger effect on behavior than continuing to have it be, like, a 10 percent chance that you get caught, but if you do win that very unlucky lottery, then you go to prison for 10 years. It just isn’t affecting behavior.

Demsas: I guess this theory kind of implies that people will commit crimes if they have the opportunity to commit crimes or that there is at least some subset of the population. Do you have sort of a generalized theory of why people commit crimes, and do we have a sense about that at the population level?

Doleac: As an economist, I think people respond to incentives, and that’s kind of the core, right? But incentives aren’t entirely about punishment or the probability of going to prison in the same way that they’re not entirely about money, right? It can be about just disapproval, social disapproval, social norms. And so if everyone around you is following the law, then you’re more likely to follow the law because you want to be liked, and you want to be an accepted part of society.

If, you know—we see this often in countries where paying your taxes is relatively uncommon; you feel like a sucker if you pay your taxes, and so no one pays their taxes. And so there is this element to which there are social norms that we want to maintain, because that winds up governing behavior often much better than having to actually go out and enforcing the law through force can.

Demsas: And I think I asked you this question because I think that there is sort of an alternative theory of mind here, which is that crimes are essentially always committed because of need—that people are committing crimes because there is some sort of deprivation, either in their literal physical need, like hunger, or they need to steal clothes for themselves, or that’s what’s going on, or they don’t have an opportunity to make money in a legal way, and they found a better way of making money through illegal channels. How much purchase do you put into that sort of theory of criminal behavior?

Doleac: I think it’s certainly a factor but, clearly, for some types of crimes more than others, right? So for revenue-generating crimes—robbery, theft, selling drugs—that totally is going to be part of the motivation. And so giving people other better options can be part of the solution. For crimes like homicide, assault—just this whole other category of crimes that are more crimes of passion, where there’s no money coming from it—there’s not a financial incentive. It’s not just because you’re poor. So it just becomes more complicated.

Demsas: So I want to return to the original kind of framework of this conversation, which is around this question of the question of policing’s impact on crime. And I think one of the things that complicates this for me is: Back in 2020, I wrote this article when I was at Vox—well, actually, I think it’s 2021—but it was looking at a preprint by this Ph.D. student, Travis Campbell, who was an economist. Or he was a Ph.D. student in economics, and I think now he’s an economist. And he was at the University of Massachusetts Amherst, and he found that the effect of BLM protests was a 15 to 20 percent reduction in lethal use of force by police officers. So that was roughly 300 fewer police homicides in census places that saw BLM protests.

But there’s a secondary effect that he also observes, which is essentially that there’s an increase in crime. And the way that I think that he would explain this, and the way that made the most sense to me, is that you see this increase in crime as a result of police officers—this is called the Ferguson effect—reducing their desire to kind of do their job. So either they’re somehow on strike, or they feel—maybe the most sympathetic version for them is that they feel—you know, worried about engaging in policing activities because of the backlash, or they feel kind of embarrassed or ashamed about their jobs now. They don’t really want to do them. And then the worst is it’s a kind of like an intentional F you to the public for being mad at them.

And so that’s kind of like—I mean, you see kind of the wide range of that that exists. And it’s interesting because, I mean, the vast majority of BLM protests are nonviolent protests. We now know this from a lot of research. And so how do you think about this problem of largely nonviolent protests leading police forces to reduce their willingness to do the kinds of policing that reduce crime? Doesn’t that kind of feel like you can’t criticize the police, otherwise they’ll just stop reducing crime rates, you know?

Doleac: Yeah, I mean, that’s definitely the taking their ball and going home kind of interpretation of that whole period. It does not make police look good; I agree. I don’t know that paper that you just mentioned, so I’m now eager to go read it. But there have been a bunch of papers looking at whether there does seem to be a Ferguson effect—that police backed off, reduced their effort, and basically what the impact on crime seems to be, and my general read is, similarly, that crime went up.

So the reduction in effort was especially apparent with regard to low-level arrests, which actually doesn’t seem to matter very much. Like, that’s the kind of enforcement that maybe we do want police doing less of, right? Just arresting everybody for trespassing or disorderly conduct.

But yeah, I mean, I think I agree that it is problematic if police essentially seem to be, like, holding communities hostage, saying, like, It’s our way or the highway. You know, You can’t criticize us, or we’re just not gonna enforce the law anymore.

At the same time, the more generous interpretation here is that, you know, everyone wants to feel valued in their profession. And this is a hard and dangerous job. And suddenly, if the community that you are serving and protecting and putting your life on the line for every day is insulting you every time you turn the TV on or every time you go out in public, you maybe aren’t that excited to put your life on the line for them anymore.

And that, you know, it’s still—most police departments across the country have a really difficult time hiring now. They’re all extremely understaffed. And that is forcing a lot of creativity about, like, Okay, well, what do we really need someone with a badge and a gun for, and where can we hire civilians to do some work? But given that we know that putting more police on the streets reduces crime, and given that I think we all would like for police chiefs to have the ability to hold their officers accountable and say, you know, If you’re behaving badly, I want to fire you, right, then we need a line out the door of other people who want to be cops.

[Music]

Demsas: After the break: the wild connection between air pollution and violent crime.

[Break]

Demsas: So you mentioned clearance rates earlier, and you said, you know, it’s about 50–50. That’s definitely true overall, but Pew Research from April, looking at FBI data, found that clearance rates for both violent and property crimes are at their lowest levels in at least 31 years. Police cleared roughly 37 percent of violent crimes they came across in 2022. Is your sense that this decline has to do with the Ferguson effect, or is something else going on here?

Doleac: I think this is bigger than that. This has been a longer-term trend. People I talk to—I think it’s a combination of real changes in policing and perhaps less of a focus on solving crime and more of a focus on just being present, which maybe is by design. I think there’s also a potential positive spin on this, which is: If clearance rates are measured by just the probability that you’re making an arrest in a particular case, well, police officers can just go out and make an arrest. They can arrest somebody, right? It might not be the right person.

What we want is them to arrest the right person. And as, especially, technology has improved over the past several decades, it has become much more difficult for police to get away with just arresting someone, right? Or a more generous way to put that is: It becomes much easier for police to get evidence that tells them whether this is the right person or not and prevents wrongful arrests and wrongful convictions.

And so I would expect clearance rates to fall just because of that, but I don’t think we have any really rigorous studies of this—of how much should be attributed to that, and how much should be attributed to changes in types of investigations and policing. But this is something that, again, my team is talking to people all the time about. It turns out, you know, detectives in this country don’t get any real training. You get a bunch of training when you sign up as a new cop.

Demsas: That’s just absurd. I can’t believe that. (Laughs.)

Doleac: But there’s no detective training, right? So I’m having lots of conversations on the—

Demsas: They just put you out on the street, and they’re like, Go solve crime?

Doleac: They’re like, Figure it out. You’ve been a police officer for a while, and so now you’re usually promoted to be a detective. And so I’m having conversations with people now about, like, Okay, what if we started a detective school? What are some big, crazy ideas we could think of here to help people solve more crimes?

Demsas: So one thing I want to get into here is how different the American experience is from the rest of the world and whether we can really even know if we have comparable analysis of this. Here in the U.S., of course, as we all know very well, crime increased. Was that an American phenomenon? And if so, why would that be?

Doleac: I admit, I don’t pay as much attention as I used to to what’s going on outside of the U.S., because my job is now so focused on U.S. policy. My impression had been that crime went up everywhere during COVID. But even within the U.S., it was a very specific type of crime, right?

In general, crime went down. It was really homicides and shootings that went up. So in most countries, there are not as many guns, and so you’re not going to have shootings go up. And guns are very effective at killing people, as you know. You maybe don’t see the big spike in homicides, but during that COVID period, people were just not out and about as much. And so if you’re at home, your house is not gonna be broken into. But for other types of crimes, you’ll see an increase. And so there was an increase in motor vehicle theft because maybe you aren’t using your car as much, and so someone can steal it more easily.

What also seems to have happened—to the best of our knowledge, our best guess of what happened here—is that because people weren’t out on the street, you don’t have witnesses. You don’t have bystanders. You don’t have the Jane Jacobs–style eyes on the street to deter people from bad behavior.

Basically, the only people who were out on the street were people who wanted to cause trouble. And so you wind up with more shootings, more homicides. And then that just became a vicious cycle where there’s retaliatory violence. And so even after the lockdowns ended and people were moving around again, you didn’t see homicides start to fall until a couple years later. So that seems to be what happened in the U.S.

I think, again, similar to what was happening in the ’90s and afterwards, it’s really difficult to pin down the exact reason for any of these broad trends because there’s just so much else going on. I think of it as the crime equivalent of trying to guess what happens to the stock market. There could be lots of really good advice about how to make your business profitable, but that doesn’t mean you can tell me what’s going to happen in the stock market tomorrow. And it seems similar to me. Like, I can talk all day about what we should do to reduce crime. I cannot tell you what’s going to happen to crime rates next month.

Demsas: So I want to steel man a bit the position of “defund the police” advocates, because many of the people who were arguing for cutting police budgets during 2020 were advocating redirecting those resources towards investments in social-welfare programs. And you have, in your research and work, pointed to many such programs that have welcome impacts on crime rates.

But what I want to ask is: How do we weigh these things against one another, given that, of course, municipalities have limited budgets, right? You have this pretty robust finding that 10 to 17 more officers is reducing one murder a year. Are there similarly cost-effective social-welfare programs that can prevent violent crimes at the same rates as police? Or is this just complimentary, but you cannot actually achieve the same effect?

Doleac: Yeah. I think the biggest challenge here with substituting completely is just the time horizon we’re talking about, right? So if the problem you’re trying to solve is gang violence that’s on your street right now, investing in K-12 education is not going to help, right? That’s a long-term solution. Or investing in lead abatement, right? That’s going to maybe reduce homicides 15 years from now, but it’s not going to be compelling to the community that is struggling right now.

That, to me, was the biggest problem with the argument that we should just be shifting money from police to these other things. I think there’s lots of evidence that we should also be investing in these other things. So summer jobs for teens is one big example—lots of evidence that that reduces violent crime, not just during the summer, when people are working, but after the fact. Lead abatement, as I said, we’re looking at stuff like—there’s evidence that real-time exposure to air pollution can affect violence—

Demsas: I think you tweeted out a study that was just so absurd. It was like, Short-term air pollution increases fatal accidents in real time. Like, air pollution—

Doleac: Yeah. It’s crazy. It really makes you wonder or think about all the ways that air-pollution exposure is probably affecting our thinking and reasoning all the time that are not pushing us to commit a crime, right? But that’s a really extreme outcome. But yeah, there’s a study that I love that looks at a highway, and they say, If the wind is blowing the exhaust from the highway to the west one day, violent crime goes up over there. And if the next day it blows the exhaust to the east, violent crime goes up over there. And so it’s just like, Okay, let’s think about putting air filters in all the schools or something.

Demsas: Do we have free will, Jennifer Doleac? (Laughs.)

Doleac: Right. (Laughs.) Yeah. There’s lots of stuff that we can do that is outside the scope of the traditional criminal-justice system that would dramatically improve public safety. And I am 100 percent in favor of investing in all of those things. I think where it gets tricky, and thinking about the trade-off with policing—or to the extent to which these things are compliments or substitutes for policing—is going to depend on, in large part, what the time horizon is you’re thinking about.

And so long term, yeah, it probably will help dramatically, and we probably could shift a lot of our criminal-justice spending from policing, but also especially prisons. I actually think a lot more about prison spending as being the target here, like shifting those resources to these other types of areas.

Demsas: You mean shifting away from prison spending?

Doleac: Away from prison spending. Right? I think there’s a really good argument here that if we invested in lead abatement, if we invested in summer jobs, if we invested in police, even—investing in, in general, those social investments, but also increasing the probability of getting caught—that’s our ticket out of mass incarceration, is my view.

But so we could dramatically, long term, shift that spending to these other things. If we invest in these other things, we would see dividends in terms of reduced prison spending and perhaps also a reduced necessity for police spending. But in any given moment—I mean, even Norway has police, right? I don’t see a world where, because we’ve invested so fully in mental health and education and all the rest, there’s just no crime anymore. I think there’s always going to be crime, and I think there’s always going to be a need for police, but I think there are a lot of things we could do to reduce people’s propensity to commit crime.

Demsas: So what are the ways that, when you’re reviewing literature here, you can actually reduce the negative externalities of policing, particularly when we’re talking about not just the most, you know, I think, well-known instances of police brutality that end in murder, but also ones where, I mean, often, when you have this pretty prevalent policing—foot policing, foot traffic, hotspot policing—it often means that there’s low-level harassment of people going on in that community.

And you know, whether you call that harassment or you call that, you know, police are just checking in, like, What are you up to? What are you doing? I mean, you can characterize it how you want. But anyone who’s interacted with a police officer in a way where you’re being questioned about your daily life and goings-on—it can be kind of stressful.

It can also make you feel like you’re being watched and make you feel like you’re doing something wrong, like you’re not supposed to be there. Often, they will just tell you to keep it moving or go somewhere else or do something different, even if you’re not breaking the law, and you need to just comply in order to not escalate the situation.

So these sorts of things can be costly upon a population of people, even if it’s reducing crime rates overall. So are there ways of reducing that cost on society, or is that just a necessary part of what policing looks like?

Doleac: I am very hopeful that it’s not a necessary part of what policing looks like. So I think this is a place that police departments are experimenting. Researchers are eager to test what different places are doing.

There is an ongoing randomized trial of—basically a follow up to a previous RCT—procedural-justice training for police, so basically trying to train officers to think of the end goal of any interaction, that everyone involved to feel like they respect the outcome, even if they disagree with it. And so the intervention is largely around changing how police officers interact with their bosses in the department. So the boss sort of models this procedural-justice interaction for them, and then the idea is that they’ll go on and do that out in communities.

And so there was an RCT several years ago in Seattle of this training, and it was on the small side, which is why they’re following up and doing another larger-scale RCT. But it did seem to reduce—I believe it looked at—use of force and some other bad things, and so generally seemed positive and seemed to have beneficial impacts. And so we’ll see in this larger multicity RCT if it continues to work.

Another type of training that seems to be effective and that I’m eager to see replications of elsewhere is a situational decision-making training that was developed by some researchers in Chicago in partnership with the police department that’s based on cognitive behavioral therapy. So basically helping police officers—like anyone else in a high-risk, high-speed, high-pressure environment—avoid what they call “thinking traps,” and avoid jumping to conclusions, and just think about all the different options. And that seemed to reduce use of force, reduce arrests, and also reduce injuries of officers, so win-win all around. And so there are a bunch of other trainings like that that I think are being tried and just to just see what happens.

I also think, you know, there are probably a lot of people that just should not be police officers, right? If you’re not cool under pressure, like, this is probably not the job for you. So my hunch is that a lot of the police killings that we see, especially the ones that make the news, I think a lot of them really are that, you know, the cop was genuinely afraid in the moment—not all of them, certainly, but, like, a large share. But if you’re that afraid in the moment, you probably shouldn’t have a gun.

So there’s some combination here of training that could be effective, and we’ll figure that out, and then there’s probably an element of screening in terms of, like, Who should we be hiring? Who should we retain? How do we make sure that departments have the ability to discipline people when they do something wrong? But also just have the tough conversation with someone and say, You’re not a fit for this job anymore. You’re fired. So I think it’ll be a combination of those things, but I’m hopeful to see that go forward.

Another bucket of potential interventions here that we haven’t talked about is technology.

Demsas: Speeding cameras. Right.

Doleac: And so, you know, I’ve thought a lot about, you know—it’s like putting surveillance cameras everywhere. Like, we could be like London, right, and have cameras everywhere. I’ve done research on DNA databases where people who are convicted of crimes, sometimes arrested for certain crimes, are added to a law-enforcement DNA database, which increases the probability that they’ll get caught if they commit another crime. And for all these kinds of interventions, you see a huge deterrent effect afterwards. People respond to that increased probability of getting caught.

Now, people worry, then, about privacy costs, but I think there’s a real conversation to be had about whether the privacy cost of having a camera on every street corner is really worse than the social cost of having a police officer on every corner. I think there are a lot of communities that would probably prefer the camera, and it could have the same sort of effect.

And so anyway, I think there’s a lot of opportunity here to just experiment with some different ways forward. And I think this is where all of the research and policy attention is in this space right now, is trying new things on these dimensions.

Demsas: Yeah, I’ve been surprised at the conversation regarding traffic violations and cameras because, I mean, when you look at the research, you can basically eliminate racial disparities in traffic violations. If you turn it over to just automated cameras, Were you going over the speed limit? But they’re so unpopular, obviously. People do not like to use cameras. Someone once said to me, like, No, I want the cop to catch me.

Doleac: (Laughs.) That’s funny. I mean, my general very cynical interpretation of that is that, you know, the main way that disparities come about is that police use their discretion in favor of, or to let the favored group off the hook. And so lots of white people are used to talking their way out of tickets, and you can’t talk your way out of a ticket from a speed camera, and so they don’t like it. But it is fair.

Demsas: Yeah. There’s also a bizarre argument that some people use, which is just that there’s a racially disproportionate effect, because in places where there are highways, people are often going into minority neighborhoods because they’re placed near highways in less-desirable locations. And as a result, you might have more people who are Black and brown being caught by these cameras. And I’m like, Yeah, but that means those people are speeding through Black and brown communities. That’s who’s getting impacted by the speeding and getting pedestrian deaths and things like that. Like, that’s the whole argument.

Doleac: Right. You’re trying to protect people in that community. And also, I mean, yeah. Exactly. It’s like, if you’re really worried about that, then put more cameras in the white communities. Just, I mean—just the answer to me seems to be more, not fewer.

Demsas: And so, you know, there’s been a renewed interest in broken-windows policing, and I’m interested in your thoughts on the relationship between order-maintenance policing and, you know, violent crime. I think many people kind of think, like, I want cops to reduce murders. We don’t want more murders. We don’t want assaults. We don’t want sexual assault. Like, those sorts of things are really, really harmful.

And I went back to the 1982 Atlantic article that I think is the genesis of the broken-windows policing in the public discourse, and there’s a passage that describes what this sort of thing looks like that I’m just going to read. And I want you to tell me whether you think that this sort of order-maintenance policing is really important for reducing crime. So it reads,

The officer—call him Kelly—knew who the regulars were, and they knew him. As he saw his job, he was to keep an eye on strangers, and make certain that the disreputable regulars observed some informal but widely understood rules. Drunks and addicts could sit on the stoops, but could not lie down. People could drink on side streets, but not at the main intersection. Bottles had to be in paper bags. Talking to, bothering, or begging from people waiting at the bus stop was strictly forbidden. If a dispute erupted between a businessman and a customer, the businessman was assumed to be right, especially if the customer was a stranger. If a stranger loitered, Kelly would ask him if he had any means of support and what his business was; if he gave unsatisfactory answers, he was sent on his way. Persons who broke the informal rules, especially those who bothered people waiting at bus stops, were arrested for vagrancy. Noisy teenagers were told to keep quiet.

So almost none of these are laws. You know what I mean? Or some of these might be laws, but they’re not even being really enforced. Like, you’re not allowed to drink in public. It doesn’t matter if it’s in a paper bag or not. The idea that, like, lying down is somehow different for loitering than sitting up—I don’t know if that’s written in any codes or anything like that.

But there is a sense in which I think a lot of people feel like there’s a general social disorder that they’re frustrated by and mad about, particularly in many cities. And I wonder: How much of policing is about this type of order maintenance, and what relationship does that order-maintenance policing have to do with the violent-crime rates that everyone’s concerned about?

Doleac: This is a really tough question. I agree—it is one that is on everyone’s mind right now. My general sense is that this type of order maintenance is not directly related to crimes like homicide. I haven’t seen great evidence on that, though, so I could definitely be wrong about that. We do have growing evidence that the kind of low-level arrests—like arresting more people for disorderly conduct or using drugs in public or whatever else—that does not reduce index crimes, the more serious crimes. And so, you know, people are finding various natural experiments that affect the number of those low-level arrests that police make and that they don’t seem to be helpful for reducing more serious crime.

That said, this broader type of disorder matters and definitely makes people feel unsafe and can contribute to a broader kind of vicious cycle where, you know, people just stay home, and then you have fewer eyes on the street, as Jane Jacobs said. And then that can increase crime. And so there is a way in which it can matter in the long term. But it also matters in the short term because it just makes people feel unsafe, even if they themselves are at relatively low risk of being the victim of a crime.

And so, I mean, my team at AV has an advocacy component. And they all came back from this spring’s legislative session saying, you know, We had all these great ideas about how to reduce crime, and the only thing anyone wants to talk about is disorder. What do we do about disorder? And I think, in general, the research consensus is: We don’t want to go back to broken-windows policing.

The conventional wisdom had been that that’s why crime fell in New York City. But that was during the early to mid to late ’90s, and crime fell everywhere. So it wasn’t just New York City. It wasn’t just broken windows. And so the consensus, I think, in the research literature is that broken-windows policing is not really helpful, in part because communities need to trust the police in order to cooperate with them.

But all of that said, I think we’re in a period right now where we are reconsidering the ways we’ve handled disorder. I think the pendulum had swung a little too far toward leniency, you know, especially pre-COVID, due to that big crime decline and really ambitious criminal-justice reform. I think a lot of people were saying, Look—we don’t need any of this. Everything’s great. And as the pendulum swung too far left, people responded, and we got more of this disorderly behavior that makes people really uncomfortable.

And so my hope is that we don’t swing all the way back to super “tough on crime.” But I think we just—we’d gone too far. And so we need to kind of find some middle ground, but we are definitely in a period of experimentation right now to figure out what some new answers will be.

Demsas: Well, Jennifer, that’s a great place for our final question, which is: What is something you once thought was a good idea but ended up just being good on paper?

Doleac: I love this question. So many things—I’ve tried so many things that didn’t work. But the first example that comes to mind is totally different from anything we’ve talked about, which is: Several years ago—I grew up with dogs, always wanted to have a dog, was finally at a place in my adult life where I was like, I’m ready to get a dog. And then I fostered a very sweet pit bull mix.

Demsas: Okay.

Doleac: And suddenly realized that it completely changed my life in ways that I was unprepared for.

Demsas: Oh, interesting.

Doleac: I was an academic. I got up every morning, was just in my head, like, to wander the coffee shop, spend most of the day there. And suddenly, I had this creature in my house that needed to be fed and watered and walked, and needed me around.

Demsas: Yeah.

Doleac: And it made me really sad how poorly this was going. And eventually, I decided this is just not going to work. And I had to give the dog back. And it was heartbreaking. But fast-forward to during COVID, and I was like, I need something, anything to think about other than work and Netflix. I know! I’ll get a dog. I remember how much work that is. And so then I adopted my adorable rescue pup, Chula, who is now an integral part of my life. So the revised version of the plan has been going very well. But yeah, the first version was definitely good on paper only.

Demsas: Well, Jennifer, thank you so much for coming on the show. This was fantastic.

Doleac: Oh, thank you so much. This has been super fun.

[Music]

Demsas: Good on Paper is produced by Jinae West. It was edited by Dave Shaw, fact-checked by Ena Alvarado, and engineered by Erica Huang. Our theme music is composed by Rob Smierciak. Claudine Ebeid is the executive producer of Atlantic audio. Andrea Valdez is our managing editor.

And hey, if you like what you’re hearing, please leave us a rating and review on Apple Podcasts.

I’m Jerusalem Demsas, and we’ll see you next week.