Focused Deterrence: A Strategy for Reducing Homicides in American Cities

- Baltimore Maryland has had high homicide rates for decades, and it got worse after the death of Freddie Gray In 2015. In 2022, Baltimore launched a new strategy of focused deterrence that concentrates on persons involved in gang related gun violence. Over the next three years, Baltimore's homicide rate fell 60% shootings and carjackings also fell. How did Baltimore achieve these gains? And what can we learn from Baltimore's experience to help us reduce gun violence in other American cities? Welcome to Economic Supply. I'm Steven Davis. Senior fellow at the Hoover Institution with me are two researchers who study crime and how to prevent it. Aaron Chafin is a criminology professor at the University of Pennsylvania. He's a senior fellow at the Skain Center and a research affiliate at the University of Chicago Crime Lab. Max Capin is an economics professor in the Brooks School of Public Policy at Cornell. He's also an affiliate at the University of Chicago Crime Lab. Welcome both of you, and thanks for joining the show. Thank you very much. Thanks so much. So let me, let me ask you just to, well, it's obvious, and people know that homicide and violent crime are huge problems in many American cities, but it, it may be a little hard for most people to appreciate the scale of the problem. So can you give us some numbers or some other information to help people who are fortunate enough not to live in a high crime area, what it's like in, in, in some parts of America? Yeah. So nationally in two, in 2021, this is the year before the crime decline begins in a, in Baltimore, around 21,000 Americans are, are, are killed in a homicide. And, you know, that, that number probably, you know, sounds bad, but it's, it's hard to think about how to contextualize a number like that. I think. So here's a way to, here's one way to think about it. When you, when you consider how concentrated homicide is in a small number of neighborhoods, in a small number of US cities, and you think about the risk falling primarily on young men, there are young men right now, even with the crime decline that we've seen over the last couple years, who probably face something like a 3% risk of dying in a homicide over the next year. So, you know, that's, that's a really big risk, right? If you, if you think about living your life, it's probably the risk that someone who is maybe 80 years old of faces with respect to, to mortality. And it's, this is single source, right? This is only only talking about homicides, right? It's a risk that's than what US soldiers faced at the peak of the Iraq war. So it really, I think that really can underscore how, how high risk this is, Right? And so the Iraq war comparison is useful 'cause we often think of soldiers who, who come through combat zone experiences is having some kind of traumatic stress associated with that. And it's certainly a ti a time of it's certain ti stressful while you're in it. But these, these young men in some par some American cities or some neighborhoods in American cities, they're living through this for years. And the fear, so it's not just that you might randomly be shot, you're, you're thinking about it, you're trying, you're worrying about it. Your brother, your mother, your, your other family members, they're thinking about it. They're worried about it. So it's really a, it's, it's an omnipresent feature of life, I think for somebody that faces that, that such a high mortality risk. Yeah, No, you, you put it very well. Right? So for the people who are most exposed, you know, this is something that really can dominate, dominate one's life and, and change one's thinking about their life possibilities for other people who maybe are, are on the periphery of this, they experience cost too, right? So people are living in places where there is less investment in, in, in opening new businesses, less economic growth. People are willing to take, you know, very long car trips or, or train trips to commute into the city for work. And I, I think in an age where people can easily work from home, the the entire project of cities, of, of cities is at risk if violence isn't controlled. Yeah. Appropriately. Yeah. Well, that's a point I've made in other settings. I've got a, a research program on work from home. But one thing it means, as you've just alluded to is if cities don't get the, don't bring safety and security to the streets, people have options. They have more options now, especially professionals than they did before the pandemic because they're no longer necessarily so tied to jobs in the city center. Okay. So I, it's just one more point's worth making. It's, it's, you said something like 21,000 Americans were victims of homicide in 2021, I think you said, if I heard you correctly. That's, that's right. Yeah. And of course, many more people were gunshot victims. They just didn't die from it or they were shot at, or they, they were next to somebody who got shot. So if you put all that together, we're talking about probably somewhere on the order of a hundred thousand Americans a year with direct experience of death by gunshot, injury by gunshot shot at and missed, or next standing next to somebody who was killed by gunshot or shot. Right. Have I got that about right? Yeah. I, I, I, I think, I think that's about right. It's a, it's somewhere on the order of maybe a hundred thousand or so who've, who've either directly been, you know, either homicide victims or, or who've been struck by gunfire, but, but survived. I think if you start factoring in, you know, people who've, who've been direct witnesses to it or who've been victims of an armed robbery or, or an assault that didn't involve them necessarily being shot, yeah. You start to get into the hundreds of thousands of, of people. Yeah, that's, yeah. And, and, and as Aaron said, we're talking about, you know, this is all happening in a, in a, in a relatively concentrated, heavily in, in a, in a handful of communities where, you know, if, if you're, if you're directly in the center of that activity, it's it, as Aaron said, it kinda shapes your life. It is every it is, right? It governs your day-to-day activities in, in a dramatic way. If you're not in the center of it, but you're just, just on the periphery, you know, you might still end up having, you know, knowing friends and family who, who've been victims suffering, PTSD, et cetera. It's, it's, it's enormously consequential special thing. Okay. Alright. So I think, I think we've made the point now about what gun violence means in some parts of America and for how people live their lives and the fear that they live with. So how does Baltimore fit into this overall picture? You know, I mentioned in my opening remarks that Baltimore has really had for I think more than half a century, at least among the highest homicide rates of any large American cities. It got worse after 2015. So to set up the intervention, the focused deterrence intervention that we're going to talk about that is really what you look at in your study. Tell us what happened in 2015 and bring us up to the period right before the launch of this focus deterrent strategy so we have some context within which to evaluate it. Yeah, sure. So yeah, so if you think about the US as a whole, homicide falls quite a bit in the 1990s, then flattens out, starting around 2015, you start to see a gentle increase in homicide in a number of US cities. But in Baltimore it's not so gentle. So homicides really start to go up after 2015. This is when Freddie Gray is killed in police custody. There's some unrest around that, you know, and, and, and, and homicide really goes up quite a bit. Now, Baltimore did not have the kind of spike in 2020 that other cities had, which is interesting. Hmm. So it's not like homicides doubled in 2020 in Baltimore, like they did in, I see. In Philadelphia. But, you know, they remained very high. And then really after 2022 is where you start to get this really steep decline in, in homicide in Baltimore. And, and, and, and note that I said homicide not crime. The, the decline in crime that we see in Baltimore really is concentrated on things like homicide shootings and, and other crimes that have a, a tight nexus to, to gun violence. Right. Okay. And so is it correct then to think about at the launch of this focused deterrent strategy, there was a certain level of exasperation in Baltimore with what had been decades long high homicide rates, mostly from guns, mostly with perpetrated with guns, and then even worse and even worse experience that pushed Baltimore further up into the homicide rate distribution after 2015. Is that a, is that how I should think about this onset of this focus deterrent strategy? I, I, I, I, I, I think, I think that's a fair characterization. I I would say, you know, from, from that point on in 2015 that you mentioned that that sort of turning point, Baltimore went from being a city that was, you know, let's say at the 95th percentile of large cities in terms of homicide rates to, to trading off with, with St. Louis as the sort of number two, number one in a typical year. So, so definitely a certain amount of exasperation. And I would say, you know, it's also in the public consciousness, Baltimore and in particular the Western district, where, where, where a lot of our study is focused has really come to occupy a sense of sort of a place of almost being synonymous with gun violence. I mean, I'm sure a bunch of your listeners and viewers know the Wire and other sort of David Simon work that that's set, set there. It's just become this, this almost, it's inextricably linked to, to the gun violence problem, which I, I, I've, I've seen, I've heard podcast interviews by, by by Brandon Scott, the mayor of, of Baltimore, kind of ruining this, this, this association that they would rather not be known for this, but, but unfortunately tragically that, that this has come to be kind of something they're linked with. Yeah. Maybe you can help them become known as the city that successfully cut gun related homicides more than half. So, alright, let's, let's talk about this focus deterrent strategy. And this is a, this is somewhat complex concept. I think there are multiple elements to it. So why don't you describe what it is in general terms, but also with an, with attention to how it was implemented in Baltimore and first in the Western district. If I understand one of the highest homicide rate parts of maybe the highest, I'm not sure, but certainly one of the most violent crime prone neighborhoods in Baltimore. That's where it was rolled out first, as I understand it. That's Right. So, so what is focused disturb deterrence? What's the thinking behind it? Yeah, sure. I'll, I'll I'll start out and then, you know, max may have more to add. So first, just, just thinking a little bit about the Western district, not only is it the most violent police district in Baltimore, it's just incredibly impoverished. So poverty rate is among the highest of any urban areas in the United States. 36% of the housing is vacant, which is just astounding. And, and in, in a very visible way, right? Buildings are boarded up. It's, it's not something that you can see in most US cities, right? So just to set the stage a little bit, this is a really, really difficult place to live and, and place to work in for law enforcement and for social service providers. So, so folks deterrence, it, it's, it's funny, like in economics we refer to it as the intervention, but it's really more of a strategy. And then the intervention is the way that the strategy is actually implemented in a particular setting. So focus deterrence is really leveraging this empirical irregularity and crime research that a very, very small number of people are responsible for a large share of the crime, right? It's really the Pareto principle and steroids. And these are, these are people who are generally known to law enforcement. They're people with long criminal histories. They're people who are, who are generally in what you might call a gang or a group. So there's, there's group dynamics that govern this, right? If Max and I get into some kind of spat and, and Steve, you and I are friends, like, you know, it expands, it expands the conflict. And so folks, deterrence is a strategy that recognizes these R dynamics, recognizes the concentration and says, look, you know, policing doesn't have to be one size fits all, right? You can imagine just sort of saturating an area with a lot of police officers say, you know, go ahead, make a lot of stops, make a lot of arrests, be visible, be proactive, be out there. Or you can say, look, you know, within a community like the Western district, it's, it's 2% of the population, maybe at most that are members of gangs and groups. Let's focus our, our attention. Really, really, really, I just wanna un underscore this point. This is not, let's identify violent crime, violence prone neighborhoods and just put a lot more police on the, on the ground there. Its not that, okay, or, or more aggressively enforce small scale infractions. That's not what this is about. That is not what this is. Which, which I think is quite a, because that's been, that's been tried and advocated in other settings, but that's not, that's not what the focus deterrent strategy is about. Yeah. No Focus, deterrence is much more surgical, right? So, you know, when you, when you concentrate police in a high crime area, there, there is a payoff. Crime generally does go down. It's a modest payoff. When you have the police make a lot of stops and a lot of low level arrests, the, the public safety benefits just aren't that high because most people who are arrested for a low level crime are just not very high risk people. They're, they're probably not gonna shoot someone, right. You know? So instead what this is about is saying, let's focus proactively on the gangs and groups that are driving the violence. We know what these spats are about. You know, we have some intelligence about this for people who are trigger pullers. You know, let's do proactive investigations. Let's do gang takedowns, let's arrest these, these particular individuals and try to incarcerate them, right? Because they've committed serious violence. But for the other people on the periphery of all this, people who no one thinks of has pulled the trigger yet, but are, you know, sort of involved in these criminal networks, let's try to see if we can affect behavioral change differently. Yeah. I just, I wanna make this very concrete. Johnny has previously been convicted of a violent crime with a gun. Okay? So he's in the first category of people you mentioned. He might be in the first category, he Might be, well, I'm just, I'm just trying to, but you, you do a distinction between the trigger pulls and people who associate with the trigger pulls. That's the, that's the, which I think as I understand this focused deterrent strategy will then take different approaches to those Different people. That's, yeah. If someone, if someone suspected of, of committing a shooting that they haven't been prosecuted for, and there's evidence, right? This is not, oh, let's just give this guy a pass. This is no, let's see if you can make that arrest. But also if there are others who participated in that crime, you know, bring conspiracy charges to bear, right? There's all these tools that law enforcement can use for the people in the periphery who are, you know, maybe committing some crimes, but haven't really done something quite like that yet. Let's make the poll of law enforcement very, very salient. Yes. Okay. We are watching you. How do you do that? How do you do that? Yeah, so, so essentially what this strategy has done to operationalize this and to make deterrents really, really salient is teams of people go into the community and knock on doors. So if, if Max is someone who is thought to be at risk for committing a gun crime or being the victim of a gun crime, there's a team of people who might knock on his door. One is probably gonna be a police detective or police officer. And that person is there to relay the message that you are on our radar. We know what you and your friends are up to. We're watching you, we're gonna be on you. There's someone also from the community who's there to present the community's moral voice. Listen, you know, I live in this community, I've lived here my whole life. My parents lived here, right? Gun violence is unacceptable. Gun violence is destroying the fabric of this community. And then someone else is gonna be there at present for that conversation as well, which is gonna be a service, a social service provider, someone to say, look, there's an alternative here. You know, we can help you get a job, we can help you get skills. You don't have to live like this. Right? And, and the idea is that it's a carrot and a stick. You know, law enforcement wants to really maximize the salience of their capability, but you also want to give people an alternative because these are people who have grown up in pretty tough circumstances and really could use better opportunities. Yeah, there's, there are carrot and stick elements to it, but as I listen to you, it also occurs to me that this focus deterrent strategy is, it recognizes the, the, the complexity of human beings and their multiple motivations. So one motivation, part of the stick is I don't wanna become incarcerated, like, you know, somebody else did. So the police are on, they're gonna be watching me. Second is appealing to their, their moral reasoning really. You say, look, I, I, somebody comes in and says, I grew up in this neighborhood, my kids live in this neighborhood, my parents are this day. Your, your parents kids are, your parents are living in this neighborhood. Don't you want it to be a good, safe neighborhood that's appealing to their conscience? And then there's, there's the part where, okay, how can you do better for your own life? You're kind of trying to give them a sense of opportunity and aspiration that you can, here, here's a pathway to meet those aspiration. So it's, it's more complex than just carrot and stick. It's trying to recognize that even people who might under some circumstance commit violent crimes are in fact complex human beings. It's A very pragmatic and realistic strategy, right? That, you know, recognizes that when people pull a trigger and take a shot at somebody, doesn't necessarily mean that they're a monster, right? But it means that they have assailant disputes, they don't see another way at the moment to settle those disputes. And you really want to deter that behavior. You wanna give people a chance to see a different path forward. So yeah, it's very, very pragmatic. I, i I think it's meant to, you know, really it's, it's inspired by empirical, empirical evidence and, and I think a, a grounded conception of human, human nature. Okay. So, but I, I wanna know one other aspect of what you just said, because it will feed into our discussion of how scalable this strategy is across cities. So you described the strategy, but then the execution of this strategy where you had the police detective, you had the credible voice within the community, you had the social worker, you might have the jobs trained, the skill development person as well. So this, this requires a concerted, coordinated effort among different parts of the municipal authority, the go, the city government, the county, whatever, as well as local civic leaders. So I think the underlying logic behind this is appealing, but it does take considerable competence and coordination to effectively implement this strategy. I think, I think, I think that's right. Not only competence and coordination, but also having all those pieces even in place, right? Yeah. So you need, you need the infrastructure in place to actually have the, the social service organizations, a and, and, and, and, you know, to be able to plug people into, into those kinds of things. So you're, you're absolutely right. I would say that though, it's not, you know, there are a lot of challenges that cities face that that can be characterized this way as, in other words, as needing a lot of coordination among different types of roles to solve. Seldom is it the case that one single agency is gonna be able to solve the problem entirely on its own. Certainly in the case of public safety. I think if you talk to, to folks in policing who are kinda at the forefront of thinking about how to approach the challenges the police face today, they'll say, this is not a job that police can do entirely on their own. It's gonna involve some degree of a, kinda a whole of government or, or, or even beyond government approach. So yeah. Okay. I take, I take that point, but the reason I'm raising this is because, let's, I'll just put it this way. There are recent years have thrown up many reasons to have concern about the competency of various governmental units in the United States. I'll just put it mildly like that. And this, this does require an effective local government in coordination with some parties that are not part of the government, per se. So I can see how this can be done. I can also see why in a community where the civic leaders aren't on good terms with the police, or the police and the social workers don't really coordinate very well because they have very different outlooks and very different emissions. It might be hard to implement this strategy in an effective way. And This is the third time Baltimore's tried this, right? Okay. Just, I think this underscores your point. Okay. Right. This was, this was attempted nineties. It was attempted actually just before Freddie Gray's death in custody, which, you know, helped to drill that attempt. I see. So, yeah, this isn't, this isn't easy. I, I, I think there are reasons to think that this could scale very well, but there's also reasons to think that some jurisdictions will struggle to do something like this. You know, I, I, I think more than intervention though, I, I think this is an, an aspirational approach to reduce violence, right? This is how cities should be organizing themselves. This is how investments should be made to try to focus law enforcement effort and social service effort on the people who are most info marginal. Right. You know, there will be hiccups along the way, right? And we don't know what will happen in the future in Baltimore, but I, I, I think more than anything really aspirational. Well, and I like the way you put it at the outset. You do it, you drew a distinction between the strategy and it's actual implementation. Yeah. So there's a clear strategy here. The way it gets implemented, it's gonna have to be tailored to local circumstances, local capacities, and so on. So the details might differ somewhat in the implementation of this strategy from one locality to the next, if we try to scale it up and roll it out. I presume there are some efforts underway to do that, given the success, the parent success in Baltimore. Well, so yeah, let's go to Baltimore. So you brought us up to 2022. Describe the way this rolled out in Bal Baltimore. As I, as I understand your paper, it was rolled out first in the West district that you mentioned earlier. And then because it was successful there, rolled out more broadly across the city. Tell me if I've got that right, but more generally, just tell me about the rollout and the implementation of this strategy, this focused deterrent strategy in Baltimore. Yeah, I, I, I, I think, I think you've got the, the, the basic picture, right? So it launched in, in 2022, the beginning of the year in 2022 in the Western district. Prior to that, there was some, some preparatory work, let's call it, to try to identify some of the individuals who were actually involved in this group based violence that, that Aaron talked about, such that when, when it launched in 2022, they could really hit the ground running. And it didn't, it didn't end up scaling to another district until 2023. So a full year later, which is itself a little bit unusual because I think the way that this often works in, in local government is, you know, once something appears to be working well in one place, there's a really strong incentive to go and quickly push it out to as many places as possible. And then in doing so, potentially dilute it, its impact and, and undermine its efficacy. And I, I think Baltimore deserves some credit here for, for withstanding that impulse and instead being very thoughtful, deliberate about how they rolled it out. But in that, in that, so in the 2022 to, to into 2023, that's the really the period that we focus on a lot in the paper. So you can think of this, the way this this works on a, on a sort of regular basis is that, you know, there, there's this process by which, you know, individuals are identified on a recurring basis. So law enforcement gathers together, they, they've examined recent shootings that they think have some nexus to group violence. They identify the individuals involved as, as Aaron was mentioning, either people who they think might have been the, the trigger pullers, people who've committed a crime versus people who might, might have been the victims of the crime, might be future victims, might be potentially future trigger pulls, in other words, who might retaliate for that incident, right? And so those two groups are identified, and then correspondingly are, you know, there, there's an effort to go out and notify them, you know, that they're on the radar of law enforcement. I'll say sort of on, on on that point too, you know, 'cause the name focus deterrent, especially for, for someone sort of approaching this as, as an economist, you know, you might think of this as, as, as, as sort of in indicating that the way this principle works is through deterrence. And, and I think that is a certainly a, a major channel through which a strategy like this can work. But, you know, the way that deterrence, you know, with respect to law enforcement normally works is, you know, somebody who's sort of living in a, in an area like the Western district, they might hear about people that they know coming into contact with law enforcement. They might see an increase in officers sort of patrolling their neighborhood, but they're getting these kind of noisy signals about the degree to which enforcement attention is either rising or falling. This is really a way to try to cut through that noise and get to people directly by literally having people show up on your doorstep and say, you are, you are a focus of this particular strategy. And so that, that is really a way to try to ratchet up that, that, that, that deterrence by cha changing the perception that you are being, you're being sort of watched in some way by this intervention. And so, you know, the folks who are, who are suspected of either retaliating in the future, being victims in the future, they're sort of routed toward, towards services. They get connected to a service provider, to jobs and education and, and, and life coaching and therapy and all those other sort of basket of interventions. And then those who are targeted for arrest are non surprisingly targeted for arrest. So there's, there's gang takedowns that Aaron talked about earlier where entire group is, is arrested over the course of a couple of days, and there's also some individual people who are targeted for arrest that way. So this, this sort of process repeats itself on a, on a regular basis as sort of new folks identify new incidents happen and so on. And will, will these arrestees typically be prosecuted for a crime that, where there's a reasonable prospect of conviction? So even if you can't, you think somebody might have been involved in a homicide, you can't pro, you don't have enough evidence to convict of that, but you can convict them of car theft or for example, is that, is that the strategy or that the implementation? So certainly with, with, with the group takedowns, I think there's, there's a, there's an effort to try to build strong cases against individuals in that group and, and gather as much evidence as possible to, to try to, you know, make those charges stick. We, we currently, with the data that we have, we, we can't see quite that many steps far down the criminal justice pipeline. So we can't see the, the exact outcomes of, of convictions and so forth. We're working on that as we speak, but we do see at least that initial arrest does happen, and, and, and those individuals were, are targeted for arrests are in fact very often arrested. Okay. Alright. So I mentioned at the outset that Baltimore had a 60% reduction in its homicide rate from 2022 to 2025. And so how much of that do you attribute to this intervention? And, and how do you know? And, and here you, you, you've got a very dense long paper here. So here you're gonna have to give me your best, most understandable evidence as how we know that this intervention played a material role in the reduction of Boston's Baltimore's homicide rate. Yeah. So I, I think the, the short answer is we will never be able to disambiguate, you know, between how much of this is due to this folks' deterrence intervention and how much is due to sort of other factors, reversion of the mean, there's a change in, in the prosecutor's office in Baltimore, there's no way to formally disambiguate. But I think there's some hints that this is, that this is Really, that's too, that's too, like you can say that for the referees to get published, right? But there's a weight of evidence in your paper. Yeah. There's a, there's Weight of evidence that informs our discussion. So yes, what give us, give us the most powerful evidence you've got. Yes, Yes. That makes, With that caveat behind me. Yeah. Okay. So your referees will now let you go ahead. I, we hope so, yes. Yeah. So, so first, you know, we can think about the western district compared to other parts of Baltimore in the first 18 months, af after, after implementation of the Western, no other district in Baltimore is receiving this. Okay? So we, we studied this for 18 months and we find that in the Western district, there's large improvements in public safety relative to other parts of Baltimore that are subject to the same general crime environment and, and trends that, that, you know, Baltimore is subject to. So in the Western we see about a one third reduction in homicides and shootings, a 40% reduction in, in carjackings. There's really very little evidence that there, that this spilling over to other areas. We don't see crime increasing in the areas that don't see, that don't have focus deterrents. And the way we're doing this is we're using synthetic differences in differences. So we're finding communities in Baltimore that have very, very similar pre-intervention trends. And then starting in January, 2022, when the intervention begins, you see them diverge fairly dramatically. Let, let me just say this in simpler terms of what I think got this Western district and compare what's hap where the, where this treatment of focused deterrent is being applied. So that's something different is happening there than in other parts of Baltimore. And you're comparing the Western district to other parts of Baltimore that are basically similar in many observable respects. Yes. You see the big reduction in, in homicides and other, some other violent shootings in the Western district within a year, but not in these other similar neighborhoods. Right? That's the basic nature of your evidence. And so Correct something could be happening, you know, something you weren't able to detect, weren't able to measure, could be happening differently or, you know, some other causal force could be operating in the Western district. But the natural interpretation of this pattern is the focus deterrent is having a major effect, positive effect. Yeah. That, that's, that's correct. Max, go ahead. I was gonna say, I, I go even a little step further. Not only are we seeing, as Aaron mentioned, we saw this, this de decrease in homicides and shootings and carjackings two kinds of crimes that are very, very directly tied to to, to gun violence. We actually don't see any clear evidence of drops and other kinds of crimes that are, are, are, are further removed from the gun violence problem. Right? Right. Which again, is sort of a signature that this is, this is picking up the impact of focus deterrence rather than something else that might have coincidentally been happening in the western district at the same time, right? Like we're seeing it exactly in like Property theft, home break-ins, that kind. Okay. So, right. We're not seeing anything in those, we're seeing it just specifically in homicides and shootings, you know, and we're not seeing that that sort of displaced to other, to other parts of other parts of Baltimore. And, and, and in addition to the, to that, to that sharp decline in, in those specific categories of gun violence, we're also not seeing any detectable change in overall enforcement. Right. So one concern, I mean, you raised this point at, at the beginning of our conversation, right? Is this just a strategy where we're flooding the western district with officers who are going out making arrests and stops. We're not seeing an increase in overall arrest. We're not even seeing an increase in the more discretionary kind of arrests like drug arrests that you might sort of associate with that type of a policing strategy. What we are seeing, however, is a pretty large increase specifically in arrests for serious violent crimes, right? Which, which again looks a lot like from where we're sitting a redirection of police effort and focus towards serious violent crime and away from the kind of high volume policing activity that, you know, you might otherwise think might be happening here. Okay. So I guess there would be a concern if you think the police have limited resources that now they're not really gonna worry so much about property crime 'cause they're focused, they're more focused on these gun related violent crimes and therefore you, you might have eventually the, the people who engage in these nonviolent crimes figure that out and feel like they have a freer hand. But it sounds like you did not find much evidence or any evidence of that. Am I correct? Yeah. That, That's, That's a theoretical concern, right? Absolutely a theoretical concern. We don't really see evidence that those crimes are going up in the Western district relative to other parts of Baltimore. I I also wanted to connect with the other part of your question, which is how much of Baltimore's 60% decline, right? Is due to this focus deter deterrence intervention. And you know, here we, we also think that there's, you know, a signature that, that is really consistent with this focus deterrence intervention playing an important role. So when we compare Baltimore to other US cities that have very similar crime trends over the last decade, we see that Baltimore and many other cities aside from Baltimore, have seen crime declines in the last couple of years. So Baltimore's not unique, Right? Even when you compare Baltimore to other cities that saw crime reduction, you see that Baltimore's homicide reduction is about 25% larger than similar cities, cities with similar homicide trends up until 2022. And, and to Max's point that he made about other crimes within Baltimore, you see the same thing when you look across cities that the crime reduction in Baltimore, that, that really is remarkable is the homicide reduction. When you look at other assaults, other forms of violence, Baltimore's crime reduction looks a lot like other cities that were experiencing similar trends. So this is really, really concentrated among gun violence, lethal violence. It is absolutely the signature of something like folks' deterrence, whereas something like a change in prosecution, like maybe more effective prosecution prosecutorial practices, you might think that that would affect a lot of different violent crimes, not only homicide. Right, right. Okay. So the way I would, my, my read of all this, you can tell me if you disagree, is while, while we can't confidently you can't confidently pin a precise number on the percentage of the decline in Baltimore's homicide rate due to the focus deterrent strategy. There are several pieces of evidence, I won't repeat them here, but several pieces of evidence that suggest a major part of the story is the effect of focused deterrence operating as intended to reduce homicides, especially gun related homicides. Is that a, is that a fair way to summarize your evidence? Yeah, I think so. Yeah. Okay. Agreed. Alright, great. Great. So let's go on. You, you know, you, you go beyond just trying to persuade the reader that this program has some success in reducing homicide towards trying to understand how and why it works. And this goes back to our earlier discussion about the complexity of human motivations. There's the fear of incarceration, there's the aspiration for a better life. There's appeals to conscience and you actually try to get at some, I dunno whether you wanna call 'em mechanisms or channels through which you're trying to understand, okay, this seems to have worked, but why and how, what was it that led the targets of the focused intervention to alter their behavior? 'cause obviously some people were just incapacitated. You take, you take the guy, the trigger pulls off the street, they can't pull the trigger while they're off the street. Okay. But there's more to it than that. So tell us, tell us what you learned in this respect and how you went about it. 'cause I think you, you, you turned to some different approaches here. Yeah. So, so we, we, we have a couple of different approaches that we, that we do to try to get at this. And maybe it just, just to take a step back, it might be worth naming some of these different mechanisms or channels. I think either one is, is, is is a fair term, right? There's the one that's sort of right in the name deterrence itself, right? So getting people to change their behavior, you know, through, you know, making them afraid of potential consequences if they were to engage in, in, in, in violent crime. There's incapacitation, arresting people, incarcerating them and so forth. But there's also sort of related mechanisms sort of floating around in the periphery, right? There's, there's kind of, you know, for example, providing people services, right? That that, that maybe help them help them change their behavior in ways that don't have to do necessarily with putting the fear of, of further contact with law enforcement in them or, or, or things like that. So we have a couple of different approaches that we have for trying to, trying to get at this. One of the things that that makes this project particularly exciting for me is that previous studies of focus de terms have also tried to estimate similar numbers to the ones we just talked about a moment ago, right? What was the impact of on violent crime in a given area where, where focus terms was tried. What we have here that's a bit unusual is we actually have individual level administrative data, so we know who the specific individuals are, who were identified by the strategy, whether they were identified specifically to receive a communication and services or whether they were identified for, for arrest. And we can actually track them in the administrative data to see what ended up happening with them, right? Were they subsequently arrested for a violent crime? Were they subsequently a victim of a crime and so forth. And we also have one of our, one of our colleagues on this project, Brian Wade, who's an excellent qualitative researcher, actually went out and conducted a number of focus groups and interviews with individuals who participated. So we have sort of both ends of, of, of the spectrum in terms of evidence here and, and looking across them, you know, if you think about something like deterrence, right? Which is so the, the hypothesized to be the key channel through which this works, you know, the, the administrative data actually doesn't have a whole lot unfortunately for us to say on this, right? We, we looked at the individuals who identified, who received these notifications on their doorstep, right? This message of deterrence, unfortunately, the, the, the outcomes are still too noisy for us to really be able to rule out that, that it actually had much of an impact on deterrence. We don't know for sure whether they were less likely to be arrested in the future on account of having received this message. But in some of Brian's work, in doing these interviews, you know, folks did report meaningfully taking that, taking that message to heart, changing their behavior, avoiding certain individuals, certain locations that they thought would put them at greater risk for potentially being apprehended by law enforcement. So there's some evidence of deterrence, but it's largely coming from the qualitative evidence. It's really not, we're not seeing it so much in the, in the data. However, on the incapacitation side, you're not Seeing it. You can't infer it from the administrative data. Exactly. You seeing it in these conversations, they're basically interview conversations with the targets of the focused interventions. And I, I found that that's not economist style research for better or worse, but I found that part of the paper to be quite interesting because it tries to get inside the heads of the people whose behavior you were trying to modify. And they will tell me exactly things like, you know what, I don't even go down to the corner grocery store anymore because I might run into some of the people I'm trying to stay away from who get me into trouble. Yeah. Ex exactly. And that the kind of thing, they explain their reasoning And they, they tell you why they are modifying their behavior. So I Exactly, I don't, I don't think economists do enough of that kind of research. I think it's quite powerful in this context. Yeah, I I I completely agree and I'm very lucky we have, we have Brian as a, as a, as a co-author on this project for that reason. The other sort of main mechanism that people think of when they think about focus deterrence is, is incapacitation, right? So some of these individuals are targeted for rest and they, and they do indeed go out and get arrested. We see that in our data. We see very clear signature, an order of magnitude increase in the likelihood of being arrested in that very first quarter when a persons identified for arrest by the strategy. So they are, law enforcement does go out and apprehend those people. However, you know, it doesn't look like from at least what we're seeing in the, in the arrest data, it doesn't look like they're necessarily being detained for a very long time. The reason we, we can say that is because if somebody were arrested and held for, for extended period of time, we would expect that in subsequent periods, second quarter, third quarter afterwards, we would see decline in arrests because they're no longer able to be arrested and we don't actually notice that. So the, the, these look to us to be relatively short stints of incapacitation. That being said, I don't wanna undersell the potential importance of that because even a short period of time spent incapacitated, just given the dynamics of some of this group violence could actually have a very meaningful impact. Right? A lot of these conflicts are, are, are, are active and ongoing. If somebody is not in the picture for even a short period of time, that could potentially be a victimization averted, right? So that that, that, that is again, Also, I presume word gets around so and so just got arrested last week. The, the cop said they're gonna be honest and they are honest. That's there, there it is. There's the clear signal. Yeah. Yeah. It's, it's, it's a go following through on the threat of enforcement is sort of also key to making the deterrence piece stick, right? So if you don't do that, then it's gonna be hard to sort of make that operational. Exactly. Okay. So you didn't, in your, in your discussion of mechanisms, you didn't really mention, and I wanted to prompt you on this to see whether it's important I, my coded in my head is messenger legitimacy or credibility. So you have these civic leaders or just people in the neighborhood who are respected speaking, knocking on the door of these targets of the focused interventions. How important is that piece of the overall effort and or do you know, can you say anything about how important that piece is? Yeah, I, IIII think it's, it's a, it's challenging to get sort of an empirical like magnitude of, of how important it is. Certainly from some of the qualitative work that that, that Brian did. I, i it comes through in some of those conversations that, that folks really do respond to the fact that those credible messengers, those those community moral voices rather are, are part of this conversation because it's entirely plausible that without their presence given, given the lack of generalized lack of trust in law enforcement, the, the, the, the sort of the message might not have gotten its foot in the door, so to speak without their presence. It's hard to know for sure as a, as a counterfactual would've happened if they weren't there. But it seems at least plausible that, that they are an important element complimentary to the rest of the strategy, to making sure that that message is, is received. I I should also note one other thing channel that I neglected in my, my last answer to sort of talk about, which is, which is services, right? A big piece of this of course is we're connecting individuals to, to jobs and to life coaching and other sort of services. It, again, it's, it's hard to get an empirical magnitude on this. We don't have a clean control group that we can compare these individuals to from some other work that, that myself and others have been involved in in Chicago looking at similar kinds of programs that provide, you know, things like cognitive behavioral therapy and jobs, you know, using our best sense of what the, the, the magnitudes of those estimates apply to this context. You know, we think that those make, I think a modest contribution to the overall decline in gun violence that we saw in the Western district, but it's very unlikely that those services alone could have explained the size of the decline that we're seeing in the Western district. It's, it's gonna have to be a combination of some of these other components like, like deterrence that it's harder to put a magnitude on. Okay? So better, better prospects, better job training opportunities, those things by themselves aren't gonna solve the problem. That's, that's what I just heard you say. I, I, I think the evidence that we have is that in, in, for, for the specific individuals who, who are offered those programs in other contexts, we do see quite large reductions in involvement in violence. It just seems like here, based on what we're seeing, the magnitude of the decline that we're seeing in the Western district, it seems unlikely that on its own accounts for the totality of the decline that we're seeing. I think, I think the thing to really, I think one of the things that really distinguishes this folks deterrence approach in Baltimore from a lot of the research in, in economics and criminology on, on violence reduction is that this is something that was trying to have an at scale, at scale impact, right? So there have been RCTs of really, really interesting and promising approaches. You know, 300 kids get a particular set of services, 300 kids don't, there's appreciable violence reduction, but it would be very hard to scale something like that up, right? And it's never really intended to have at scale impact, like, you know, you're not gonna treat 300 kids and, and, and probably see a citywide reduction. So when you think about the number of people who are treated here, and then the fact that you see these massive reductions at the, at the level of the entire police district and even a city, it seems unlikely that like you, you'd have to make sort of heroic assumptions about the crime reduction that you would see from treating a relatively small number of people, about 90 people with services to think that that could explain, you know, the, the magnitude of the decline that we're seeing. So I I I think that, you know, you, you don't want to compare like this 30% reduction in violence, you know, for like an RCT, a particular sample of people to 30% decline over an entire community. Right? Okay. But there, this a little complicated here, but it's an important point. It, it does distinguish your study from many others as I understand it and as as you just pointed out. But I, let me see if I understand what you're saying. It would be possible to get large reductions, or at least it's plausible that you might get large reductions in homicide rates through a strictly service oriented approach, but you'd have to carry it out at a very large scale involving essentially all the people who might potentially be at risk in this neighborhood or in this city. Whereas this strategy is very highly focused on people, a small set of 2% of the neighborhood, or if I understood correctly, is that the, is that the distinction you're drawing? Yes. You put that more artfully than I did. Yes. And also that this is, this is intended to generate deterrence and to create this, you know, a general equilibrium effect as opposed to just a partial equilibrium effect, right? You're, you're trying to like treat a small number of people directly, but indirectly treats, you know, so many others. Okay. Okay. Okay, great. Okay, so let, let's turn to the last, last thing I want to talk. So c, can this be scaled across cities and if so, how challenging is it? We, we got into this little, a little bit earlier, what's happened so far in this respect? I just wanna, this sounds, I take this to be as a consumer of your work and some related work, I take this to be promising not the easiest thing to implement effectively as we talked about earlier, but, but certainly within the realm of the feasible and reasonably cost effective compared to other strategies one might think about pursuing. And so give us your sense of whether we can scale it up across many cities and have a big effect, not just in Baltimore, but, but on homicide rates more generally in American cities. Yeah, so I'll take a stab at this hard question. So first lemme just sort of set the stage by saying, this is what we wanna see, right? What, you know, what we really wanna see is a large reduction in serious violence without widening the net of the justice system, right? And making many, many more arrests. Is this scalable? So I think that there's some reasons to be optimistic and some reasons to maybe be concerned about scalability, right? On the one hand, it's very cheap. This intervention in terms of new resources that ha that required financial outlays is maybe like, I don't know, somewhere between three and $5 million depending on how you count. It's, it's actually not a lot of money and relative to the benefits, it's a tiny, tiny amount of money. So, you know, often scalability comes down to like, yeah, this thing has to be cheap enough to, you know, to to scale this thing is cheap enough to scale, it doesn't require creating a new government agency or, or something like that, right? You're working within the existing frameworks, police, social service, providers, courts, prosecutors, right? That's another reason why, you know, this is something that could, that that can scale and it really is consistent with everything that we know about crime and offenders, right? That there's a small number of people who drive the problem. And we have more data than ever now, right? We have way more data than we had in the 1980s that can be brought to bear to, to better target people and communities and gangs and groups. There's new investigative tools that involve, you know, forensic investigation like cell phones and, and things like this. There's, there's camera evidence. So I, I, I think there's many reasons to think that this is something that more cities can be taking greater advantage of at the same time, right? It's not some, it's, if you think about something like street lighting like that can scale because you know, you, if you know how to install one streetlight, you know how to install a thousand streetlights right here. This is really something that involves, you know, human capital, right? You know, prosecutors and police need to need to work well together, police and, you know, need to understand when do we go to the prosecutor and ask, you know, like, when should we try to get a search warrant, right? What do we need to get that search warrant? How do we actually go about this? You need prosecutors who have real experience doing these long-term investigations. You need social service providers that have credibility that deliver, you know, a, a good service to people and actually do meaningfully change their lives. I, I think that is the hardest thing probably in all this to scale, right? Because, you know, if it takes a visionary leader to create a great social service provider, those people are probably not an, you know, an infinite supply. So I, you know, there are gonna be challenges, you know, I don't think every city that does this is gonna see, have the same experience that Baltimore's had in the last couple years, but I very much believe that everything we know about crime, about the justice system, about people who commit crime, about victims suggests that this is the sort of thing that cities should be investing in and, and creating a, a, a, a long-term sustained infrastructure to, to nurse along, right? You don't wanna just sort of say, well here's some, you know, we have $10 million do this, and then there's no budget in the future to ever continue it. Right? You, you want to create something sustainable. Okay. Max, you you want to add anything to that? Yeah, I just wanna add one, one other point on on on scale. So Aaron mentioned earlier that, you know, we have RCT evidence of small scale interventions, tutoring programs, jobs programs, et cetera, working with a small group of people and, and limited evidence on what it's like to take those interventions to a citywide scale. On the one hand, the what we're studying here is already being done. In some sense it's that neighborhood and citywide scale. I think that the, the, the, the difference between those small interventions in this one with the scale question is with those small interventions, the challenges often as we think about treating more and more people, how do we get this kind of scarce inputs necessary? The skilled tutors, the great job coaches, et cetera, you know, those individuals aren't short supply, et cetera. And so it's, they hit an inherent sort of wall in trying to scale to a large population size here. I think that the dimension that's, that's relevant is less the, you know, hiring a certain scarce number of inputs and more sustaining this over a period of time. So all the, all the pieces that Aaron mentioned, right, the, you know, the cooperation of the police and the prosecutors and social service agencies and so forth, right? Each of these organizations is under, is going to undergo change as time goes on, new leadership comes in, et cetera. It's being able to sustain that working relationship and really making that sort of a bedrock foundation of how these agencies and organizations work. That's the piece that I think we're, you know, I think the jury's still out on whether or not cities can really do this because the intervention itself is not new. I mean, probably more than 50 cities in the US and, and I'm sure elsewhere in the world have tried, implement a version focused deterrence that that's not new. It's really can you sort of implement it well and sustain that fidelity over a long period of time that that's the real challenge. Okay, thanks. So there's a strategy here that can work, has worked in some settings. There are challenges in rolling it out and sustaining it, but those challenges can be met. Hopefully they will be met with enough political leadership and goodwill by community members. Look, this has been, you guys are doing really important work. I, I wanna applaud you for it and for your efforts to, I'm sure you engaged in efforts to try to scale this beyond Baltimore. Yeah. Our, our colleagues at the University of Pennsylvania Crime and Justice Lab are responsible for helping Baltimore with some of this work. They do work in other cities and it's my great hope that we can see some funding to do some cross city research in the future and really get some better evidence on how well this, this can scale. Alright, well thank and, and I hope, I hope you get the funding and to, to carry this out. So thank, thanks a lot for joining us and thanks again for doing your work and telling us about it. Keep keep it up. Take care. Thanks. It was a pleasure. Thank you very much. Thanks So much.