Episode 20: J. J. Prescott

 
Ep20_Prescott_photo.jpg

J. J. Prescott

J.J. Prescott is the Henry King Ransom Professor of Law at the University of Michigan.

Date: January 7, 2020

Bonus segment on Professor Prescott’s career path and life as a researcher.

A transcript of this episode is available here.

Episode Details:

In this episode, we discuss Professor Prescott's work on sex offender registries:

"Do Sex Offender Registration and Notification Laws Affect Criminal Behavior?" by J.J. Prescott and Jonah E. Rockoff.

OTHER RESEARCH WE DISCUSS IN THIS EPISODE:

 

Transcript of this episode:

 

Jennifer [00:00:00] Hello, Probable Causation listeners. Before we get started today, I want to encourage you to support the show on Patreon. For just five or ten dollars per month, you get access to exclusive bonus content, such as interviews with book authors hosted by David Eil and bonus segments with the scholars I interview on the show, talking about their background and life as a researcher. Plus, you'll know that your contributions help keep the show running, something for which the entire team is grateful. To subscribe, go to patreon.com/probablecausation. There's also a link on our website. Thank you in advance for your support. Now, on to today's show.

 

Jennifer [00:00:42] Hello and welcome to Probable Causation, a show about law, economics, and crime. I'm your host, Jennifer Doleac of Texas A&M University, where I'm an Economics Professor and the Director of the Justice Tech Lab.

 

Jennifer [00:00:53] My guest this week is J.J. Prescott. J.J. is an economist and law professor at the University of Michigan. J.J., welcome to the show.

 

J.J. [00:01:01] Thanks for having me.

 

Jennifer [00:01:03] So we are going to talk today about your research on sex offender registration and notification laws, but first, could you tell us about your research expertise and how you became interested in this topic?

 

J.J. [00:01:15] Sure. So I have a PhD in economics from MIT and a law degree from Harvard. And I think when I was in law school, I thought I might be really interested in becoming a Torts professor, which would be a professor who is interested in the causes and policy surrounding accidents. But I wound up clerking for Merrick Garland after law school, and while I was clerking, I spent a lot of time doing sentencing related cases. And and that wound up leading me to be very interested in criminal justice and criminal law. With respect to sex offender registries, I think- it's a little bit of a it's a little bit of a story that that goes back to an advisor, a mentor of mine, who who—in in talking about things that people who are eventually going on the market ought to be thinking about—said, you know, just just do something that's exciting and for want of a better phrase, sexy.

 

J.J. [00:02:15] And and and, you know, the thing about sex offender registries in the area of criminal law and criminal justice, there isn't a lot that's new under the sun. I mean, we have kind of restricted ourselves from using certain kinds of punishments. And we've been using mostly the same types of punishments for a long time. We've mostly restricted behavior in the same kind of ways. But sex offender notification and registration are actually pretty innovative and new, and they don't have a very long history. I mean, they- there is a an earlier generation of them that happened in the 30s, and you can make analogies, but really they started in the late 80s and early 90s. And the idea behind them was let's let's leverage information, let's try to use information, either information going to law enforcement or information going to the public as a way to control behavior. And that was, first of all, really innovative in the context of criminal law and just really interesting to me.

 

Jennifer [00:03:11] So we'll focus, for the first part of the show, on a paper you wrote with Jonah Rockoff titled "Do Sex Offender Registries and Notification Laws Affect Criminal Behavior?" This paper was published in the Journal of Law and Economics way back in 2011. But you followed this topic closely since then. So we'll talk about how this literature has developed a little later on. But why don't you kick things off for us by telling us a little bit more about what sex offender registries are? What do these policies look like and why do they exist?

 

J.J. [00:03:41] Sure. So, I mean, I got a little bit ahead of the game in my answer to your last question about information. But essentially what registration notification laws are designed to do is I think first and foremost, they're they're targeting recidivism. So the the inspiration behind them, which goes back to some of the earlier- the earliest cases or, you know, surprising, terrifying, aggravating newspaper articles about recidivist sex offenders and how we might have done something different. So imagine somebody who has committed a sex crime, is released from prison, they've served their sentence, and and there's some possibility that they will reoffend and commit another sex crime. What can we do other than committing them as if they are mentally ill? What what can we do other than that to try to reduce the likelihood of their returning to crime? And one one possibility is that we could let people know either the police, just the police, who could then monitor released sex offenders, that that usually goes under the name private registration or registration. And and the thinking here is that, you know, if you're a sex offender and you know, the police know where you are, they can first monitor you. And second, if there is a crime that happens near where you are or seems consistent with your behavior in the past, the police will have a leg up in potentially identifying the likely offender. So that's registration.

 

J.J. [00:05:23] And a second strategy, which is sometimes called Public Registration—or goes under the moniker, Community Notification—involves making the identity and criminal history of a convicted sex offender public. For those of you who are familiar with it, you know, the most common way this is done today is through a web registry or an Internet registry where you can go on and type in your address and find out who has committed a crime near you. And this was designed really to have two effects. One, again, is to improve the likely quick apprehension of somebody who has reoffended. But also brings in the public to both monitor released sex offenders for signs of bad behavior and also to allow the public to to engage in precaution-taking, precautionary behavior. So if you know that there's somebody who has been convicted of a crime in the past near you, you can change the way you behave, maybe interact with them differently so that you are less at risk. And that's pretty much what registration and notification are designed to do. And really, you know, it's it's a pretty innovative strategy, one that we might find a little bit of a cop-out because the police and law enforcement are essentially putting the burden on people to protect themselves. But, of course, there's a natural, I think, psychological tendency to want to know this sort of information. So they've met with great success. Once they once the innovations happen, we you know, they they pretty quickly spread throughout the the U.S.

 

Jennifer [00:06:59] I'm realizing we should probably also clarify what a sex offense is in this context. Is this just really serious offenses like rape or does it- do these registries tend to include more minor offenders, too?

 

J.J. [00:07:09] Great. So there is some variation across law, especially in the beginning. They mostly focused on very serious crimes. But over time, what counts as a sex offense has expanded. So usually it- a sex offense would include both activity that is sexual in nature—so it might include, for example, a child pornography crime, which isn't maybe like a rape, but both seem to be related to sex—or it can involve body parts. So, for example, even indecent exposure or urinating in public, these are often also considered sex offenses for purposes of winding up- causing you to be included on one of these registries. So there are there are a few other crimes that are maybe less sexual in nature that wind up being included here. For example, like kidnapping, the kidnaping of a child, kidnapping of a child in some places puts you on the sex offender registry, even if the kidnaping didn't result in any kind of sexual activity whatsoever. So there are a few, a few kind of transitional crimes that that can generate inclusion. But but most of them fall into those two categories.

 

Jennifer [00:08:23] And how high are recidivism rates for convicted sex offenders? Is this a major policy problem?

 

J.J. [00:08:29] Well, I think probably recidivism in general is a major problem. I think we've spent a lot of time policy-wise trying to figure out how, you know, once somebody has committed a crime, how we can reintegrate them into society, make them a contributing member of of society. So in general, it's a problem. With respect to sex offenders, it's tough to know. So there is a a belief out there, a mistaken belief, that that sex offenders, once they commit a crime, they're sort of revealed as as if they are impulsive and incorrigible and will continue to offend. In fact, sex offenders recidivate at considerably lower rates than most other crimes. So somebody who's convicted property offenses or someone who's committed theft or something like that is much more likely to recidivate than somebody who has committed a sex crime.

 

J.J. [00:09:26] And so- and there is there is actually a fair amount because there have been these laws targeting sex offenders in particular. We've had a Supreme Court, for example, that has described recidivism rate, recidivism rates among sex offenders as frightening and high, citing to something that suggested, as you know, rates as high as 70 or 80 percent, that that claim has now been debunked. Rates are actually much lower, you know, somewhere between five and and considerably, you know, 20 or 30 percent, depending on the way you count recidivism. How many years you how many years you measure recidivism over. That's not- it's not trivial, though. The recidivism rate is not trivial. And it's, in my view, I mean there have been a lot of people who've talked about whether or not we ought to try to innovate in this area to do something to reduce recidivism. I think the answer is yes, so long as we actually think that what we're doing is improving on matters. And, you know, as I'll describe some of the work that's come out- a lot of the work that has come out has suggested that using SORN policies, Sex Offender Registration and Notification policies, doesn't really help. So but we'll get to that down the road, I assume.

 

Jennifer [00:10:44] Yeah. So people who haven't thought much about these policies might think that any effects they could have would be very straightforward. They help us catch repeat sex offenders. Right. But as you detail in this paper we're going to talk about and elsewhere, there are actually a variety of ways these laws could potentially affect behavior and not just the behavior of registered sex offenders, but behavior of the police and potential victims and potential first time offenders as well. So talk us through all of the potential mechanisms we should have in mind for how sex offender registries might affect crime.

 

J.J. [00:11:16] Yeah, so this is this is complicated because there are a bunch of different ways to to cut this. But let me start just with with something simple and straightforward: private registration, where essentially people are required under these laws to provide identifying information in their address information and and and maybe where they work and how to contact them to the police. So how would we expect this to affect the various players? Well, first of all, at least in theory, you know you know, once a law is passed and you realize, hey, if I am convicted of a sex offense, I'll have to provide a lot of information to the police and I'll have to do so regularly, that might cause you to rethink whether or not you want to commit a sex crime. This would be a form of deterrence. And so that's one possible mechanism that we'd have, potential first time offenders deciding not to actually offend when they otherwise might.

 

J.J. [00:12:17] With respect to people who've already been convicted of sex crimes, and a lot of these laws were retroactive, so they they wound up applying to people who had been convicted in the past. Now, essentially, you're required under the law to provide your information. So you start doing that and you realize, again, the police might be able to to monitor you. And and so you may decide, again, not to commit another sex offense. I refer to this usually as recidivism reduction—or people who would otherwise have recidivated—might not now. That's, I think, some of the sort of the two obvious ways that this might matter. What we'll find, though, is when we we think about recidivism reduction, that these laws, if you've already been convicted of a crime, wind up burdening you quite significantly, less so with private registration, to be sure. But even with private registration, you regularly have to deliver information to law enforcement. Sometimes you might need permission on whether you can travel or how and when you're able to move. These also interact with residency restrictions and so on.

 

J.J. [00:13:28] Now let's move to community notification, where you provide this information to the public. Again for first time potential offenders, there's a much I think- you know, this is this is not really based on data, but at least at least in thinking about hypothesizing, in effect, the prospect of being labeled publicly a sex offender, it seems quite salient. People are aware of these lists. And so we, again, might expect there to be quite a large deterrent effect for people who are not yet on the registry. They may decide for whatever reason to to engage in a different crime or maybe a worse outcome here would be to to engage in a crime that was harder to detect because of because of the possibility of winding up on a public list.

 

J.J. [00:14:17] For those who are already on the list, however, we now have a couple of different effects moving in opposite directions, at least theoretically. So we have police and and victims who are aware of who these people are, at least in theory, or at least some subset of them. And and so the prospect of committing a crime successfully, meaning like completing the crime and and not being arrested or convicted for it, goes way down. So there might be, again, a recidivism reducing effect because the laws essentially work as they are supposed to.

 

J.J. [00:14:59] But but there's another effect. And this is a really important one and one that policymakers almost always forget, which is, you know, for people who have been convicted of a crime before and go on these lists, they may they may be, you know, somebody who never likely to recidivate in the first place under the previous regime. But now they are essentially a pariah. They are publicly listed, makes it difficult for them to find housing, makes it very difficult for them to find a job, makes it difficult for them to form new relationships. And and it turns out that we know from other reentry work that all of these things are real are real significant risk factors for returning to crime. So on the one hand, we have public notification or community notification leading to a higher likelihood of getting caught if you're reconvicted, at least if you think that these laws work the way they should. On the other hand, everybody who is who is listed on the registry now finds it much harder to reintegrate into society. And and so that that would suggest the possibility that these laws could actually be counterproductive, at least for people who are listed on the registries. They might might wind up becoming more likely to commit sex crimes than they would have in a world where there was just private registration or nothing at all.

 

J.J. [00:16:27] And you also mentioned other types of players here. So, of course, in order for notification and registration to work, you know, you need to have the public and law enforcement using this information. So with private registration and with public notification, we have law enforcement—who who are aware of of where sex offenders are—can keep a closer eye on them. Presumably they do some of that in some systems where there's more active involvement of police. They do sort of checks. They you know, they show up and and find out whether or not anything looks unusual at someone's house.

 

J.J. [00:17:07] When it when it comes to notification and involving the community, letting them know, there are a couple of assumptions that go into into the claim that these things make a difference. The first is that people know about and use these registries, these public registries, and that seems to be a true assumption. So a lot of people know about them. A fair number of people check them. I remember when I first started working on this, I had some undergrads who were collecting some information for me, and more than one of them said that, as you know, in high school, they were directed to to look up on the registry, whether there were any sex offenders nearby. So so so that that part of this the claim is not a hard one. One one area where I haven't done research, but other people have is to find out how people respond to knowing that there is a sex offender nearby. And and and it turns out that the evidence here suggests that people don't do very much with it, so they often talk about people who are on registries. But when you ask them what they do to lower their risk, there there isn't there there isn't a lot of evidence that they do anything too helpful. And there is some evidence that they use this information to to, you know, to harass or to to force people out of neighborhoods and so on. All of which goes back to my earlier comments about potentially increasing recidivism rates through notification.

 

J.J. [00:18:42] And I and I, you know, when it comes to notification, we're going to have some neighborhoods that have listed offenders in those neighborhoods. And one thing you might worry might happen is that if you live in a world where in your neighborhood, you know that there are two or three convicted sex offenders, those might be the those those might be the people you focus your attention on when you're worried about your kids or your family members. And that raises the possibility that, well, that listing somebody on a registry might distract from other from other potential risks. So one of the things we know from the sex offender context is that a lot of people believe that sex offenses are really a stranger danger. So, you know, somebody will jump out of the bush bushes or there will be somebody who tries- who pulls up in a van and tries to take your child. Well, you know, those things, they do happen. But it's much likelier that a friend or neighbor or uncle or coach or something of that sort will be the source of the real threat to your family. And and one concern is that by giving people a list of of of potential recidivists that live in their neighborhood, it distracts them from from generally being wary of other people. You know, presumably people who have never been convicted of a crime and therefore don't appear on the registry. Did that cover most of it?

 

Jennifer [00:20:17] That does it. Yeah, that was great. Yeah, so there's a lot that you might want to measure here. So let's talk about the empirical challenges to studying what the effects of the policies are in the real world. So as you and Jonah started thinking about this, what were the primary hurdles you had to overcome as you were thinking about how you would measure what impact sex offender registries had on criminal behavior from these these various actors you just talked about?

 

J.J. [00:20:42] Yes, sure. So, I mean, in in the criminal justice work that is done by economists, a lot of time is very sort of reduced form. I mean, there's lots of actions I've described, many of which we can't measure. But we started from the premise that, you know, the first thing we need to do is really come up with a way to to identify features of these laws and then come up with a way to measure those laws both in time and strength. And one of the pretty standard ways of thinking about understanding the effects of of of law generally, policy generally, but especially criminal justice, is to take advantage of the fact that states enact laws that are in the same category but look different from each other and and cover different populations and are put into place at different times.

 

J.J. [00:21:38] And so from our perspective, there was beginning in, roughly the late 80s, a movement across the country, starting with a few leading states, but then essentially an avalanche over time, of these laws being passed. They were mostly similar in design. They also progressed through a similar sort of phase. So they began with private registration and moved on to public registration. And and so the first task was to get a handle on the differences over time and across states and in terms of how in terms of how these laws were both enacted and then put into effect. And and that's a big difference. That's something that was important to us because these laws were often enacted at a- at Time A but didn't really go into effect until sometime later. And if you just spent time looking at statute books, you might be mistaken in terms of when you thought these laws really ought to start changing behavior.

 

J.J. [00:22:47] So that's one side that's like let's let's figure out what we're measuring on the law side. And then the question is like, what are the outcomes that matter? And here we used federal data, the NIBRS data, National Incident Based Reporting program System data, and this is the data that winds up feeding up and creating the the uniform crime reports data that comes out from the FBI and we did that for a reason. So there there's been some other work, including my collaborator, Amanda Agan, that has looked at the effects of sex offender registries and community notification using the UCR data. But one of the nice things about the NIBRS data, even though it didn't cover as many states and didn't start as early as we would have liked, is that it actually included some information on the relationship between the offender and the victim. And that was important to us both from a theoretical perspective in terms of, you know, how these laws were supposed to work by revealing information to people who were actually in a position to use it, but also because it gave us another potential identification strategy that we were going to be able to to kind of at least assume that certain types of relationships between offenders and victims, certain types of offenses, would not have been affected by registration and notification. So to give you an example there, we might think that a note of- a crime that occurs between parents and children was, you know, it was unlikely to be affected by SORN, at least- the registration notification, at least when it comes to recidivism reduction, because it seems pretty unlikely that within the nuclear family, a history of of sex crime would not have already been known to the children in that family or or other members of the family.

 

J.J. [00:24:52] So so that that's that's how we approach this. And so two big categories. One are the outcome measures which we used incidence of reported crime. You can move on from there and and look at other other measures of of crime, depending on where they are in the process. So an incident of crime that is reported may actually be nothing. So maybe we should only look at an incident that leads to an arrest or an arrest that leads to a prosecution or a prosecution that leads to a conviction. We focused first and foremost on reported incidents, in part because we you know, we wanted the broadest possible measure. And there are lots of reasons why crimes don't progress to future stages that have very little to do with whether or not the crime actually occurred. And then we use some of those other measures to to check our inferences about that.

 

Jennifer [00:25:54] So before this study, what had we known about the effects of sex offender registries, what other research had existed?

 

J.J. [00:26:01] Great, so primarily there had been some studies that had looked at small groups of of offenders and looked to see whether or not if they were released before registration notification laws were put in place, did they- were, you know- let's let's come up with a baseline. What was the likelihood that they would recidivate and and did that change or did that rate change for individuals released after after SORN laws went into effect? These were almost entirely studies that focused on a single state. So, you know, to give you a classic example, we might take one of the early adopters of this kind of information, Minnesota or or Washington state or New Jersey. And we'd say, hey, let's let's let's take a set of sex offenders who were released, let's say, five years before SORN happened and and see how likely they were to recidivate. And then let's take a set of offenders who were released after SORN went into effect and see how likely they were to recidivate. Either, you know- unfortunately, if you have a set of identified offenders, you have to look at least to people who have been arrested, otherwise you wouldn't know whether or not it was an offense that came from a potential recidivist.

 

J.J. [00:27:31] And so there are some problems with this. For instance, there are a lot of other things that could be going on in a particular state. There could be other criminal justice policies. And on top of that, you are only looking at the effects on one dimension. So I mentioned before that these laws are designed to reduce recidivism, but they might have important potentially offsetting effects on first time offenders and including, you know, first time offenders who take advantage of the fact that there's public notification to to fly under the radar, for example. And and and and most of these studies did not look at sort of overall levels of of offenses going on in a state. Presumably the goal of SORN was not just to- I mean, or at least we would care whether or not total numbers of offenses seem to be going down, not just the effects of the crime on- of the of the of this kind of law on on recidivism.

 

J.J. [00:28:37] So earlier studies mostly look like that: within a single state, looking at just that recidivist crime. And almost all of them showed no effect, no effect in part and usually the estimates were very close to zero, sometimes even positive. But there wasn't even a a a typical sign associated with SORN laws. There was a study that came out just before hours that showed in the context of a very narrow, targeted public public registry scenario where really the worst of the worst were put onto a public registry, some indication of reduced recidivism. But by and large, there was a message that came out that these things just don't make a difference, even to recidivism. You know, leaving aside all their other potential effects, they don't seem to work the way they were designed or people have assumed that they were going to work.

 

Jennifer [00:29:48] OK, so in this paper, you're going to use this gradual rollout of these policies and a difference-in-differences framework, comparing trends in sex offenses in places where a sex offender registration policy goes into effect, with trends in places where those policies change at different times. And so if we think of sex offender registries as a shock to the probability that repeat offenders will get caught or to victims ability to protect themselves, then we should see crime rates change at the same time the policy changes. But what you're doing is actually a little bit more complicated than that—standard diff-in-diff—because you're also interested in differentiating between the effects of private registration laws and notification laws, as well as whether changes are due to changes in recidivism or first time offending. So tell us a little bit more about your empirical strategy and what exactly you're testing for.

 

J.J. [00:30:38] Sure. So, you know, you got it right there. So one of the first things we did was recognize—and I should have made this clear in my earlier answer—the existing work at the time Jonah and I started working on this really conflated registration and notification. They were treated the same. But as I, you know, as I tried to argue before, they're really quite different. One is really focused on letting the police know where you are so that they can keep an eye on you and potentially apprehend you if a crime happens nearby. But other than the police knowing, the public would not know, I mean, unless they did some sort of background check. So you wouldn't you wouldn't necessarily have trouble getting a job or finding housing or reintegrating into society, whereas notification really works by putting the public on notice and making sure that you're keeping your your nose clean. And that has, as I mentioned, the possibility of of- it has the potential consequence of making reentry much more difficult. So we took into account that most states progressed from private registration to public registration and sometimes separated by, you know, only a few months, but in other places, separated by many years. And so we we took advantage of that.

 

J.J. [00:32:07] And we also were interested, as you mentioned, in the deterrent effect on potential first time offenders and the recidivism reducing effect on people who had been convicted of the crime, convicted of a sex crime in the past. And the way we tried to tease that out was taking advantage of the retroactivity of these laws and the size of the registry. So let me give you an example of how that would work. So some states mostly, you know, mostly—I mean, we argue essentially randomly—based on previous views of what their constitution required or what was an appropriate way to proceed with designing these laws. And I could talk a little bit about the law later if if you're interested. But some states made these public registries apply and public and private registration apply prospectively. In other words, only people who were released starting the day after the effective date would go on the registry. If you had been convicted and released prior to the effective date of the law, you would not go on the public registry. So that means on the first day, we had a registry that was effectively size zero. And so if you're thinking about the difference of deterring first time offenders who don't want to wind up on the registry versus this strategy that tells us that that people who are on the public registry will will will- the registry will alert their neighbors and acquaintances and potential victims to watch out for them. We know that that theory of behavior change isn't operating in those states at that time. And essentially over time, the registry grows.

 

J.J. [00:34:00] And so the identifying assumption we make is on day one in a state that has a very, very small registry or no registry at all, any behavioral changes that is is the result of deterring potential offenders who who are not yet on the registry and do not want to wind up on the registry and that that that effect begins to shrink relative to the recidivism reducing effect as the registry grows. And so we kind of we assume the deterrence effect is is fixed over time. And so if the effect changes over time, we we argue that that change over time is the result of the registry growing over time and giving us a chance to explore the recidivism reducing possibility of these laws.

 

Jennifer [00:34:53] OK, so let's talk about the results. What do you find is the effect of private registration laws on sex offending?

 

J.J. [00:35:01] Great. So remember, there are two dimensions we're interested in. One is deterrence of of first time offenders, or at least those offenders who who even if they've been convicted of a crime before, were not going to be subject not eligible for this. And so they might want to not commit another crime so they don't wind up on the registry. For private registration, we find no real deterrence effect whatsoever. And that that's not that surprising, in part because it is a an administrative burden, but doesn't come with the pariah status or registration doesn't doesn't bring the pariah status that that that comes later with a public notification.

 

J.J. [00:35:39] However, we do find that recidivism is reduced among people who who who are on private registry. I should be clear that we don't actually have in our data any connection between people who are on the registry and whether or not they committed crimes. But instead, we find is that the the amount of sex crime goes down as the registry size goes up and we attribute this to recidivism reduction. So that's private registration. So doesn't really deter anybody. But the monitoring that law enforcement can engage in seems to have a positive effect, meaning that there are fewer sex crimes as the as the registry expands.

 

J.J. [00:36:31] With public notification, we almost find the opposite. So, first of all, again, not surprising, I don't think, if you think about it a little bit. We find significant deterrence effects. So, people who are not on a public registry don't want to be on the public registry and are either committing fewer sex crimes or are committing different kinds of crimes that are less likely to be reported. In any event, we see reported sex offenses going down. But when it comes to recidivism reduction, we actually see either no effect or actually a subtle increase in sex offenses. So to be clear, the larger and more expansive the registry is, the the higher is the level of crime. OK, and we explore in the paper various possibilities for this. But from our perspective, we think the the most likely story is that that essentially for people who are who are potential recidivists, the environment in which they can, you know, reintegrate into society has become much more hostile. And so, again, I described before that there's a basic tradeoff, like maybe these laws work to cause victims to be more cautious, allow allow the public to monitor more effectively and so on. There's there's maybe some effect there. On the other hand, we are dramatically increasing the risk factors for recidivism among those who are on the public registry. And our evidence suggests that the the latter effect outweighs the former. In other words, people become less likely to to stay on the straight and narrow and become more likely to commit crimes as a result of public notification.

 

Jennifer [00:38:23] And then I think you also are able to look at the effects of these laws on who is victimized, am I recalling that correctly? So-.

 

J.J. [00:38:29] That's right.

 

Jennifer [00:38:30] -What the relationship was to the offender, what do you find there?

 

J.J. [00:38:33] Yeah, so so, again, you know, at least if you believe that notification is supposed to work the way, you know, proponents argued it was, we wouldn't expect it to make much of a difference with respect to family members because they should already know things. And and strangers—people that you don't you don't know and come from other neighborhoods and that sort of thing—this kind of information wouldn't be very helpful to you either, because they're too far afield. In an ideal world, what you would do is learn about the people that are acquaintances or maybe somebody you're thinking about going out on a date with or somebody who lives in your neighborhood or maybe is is not quite a friend, but is somebody that you're sort of in that middle range, not a complete stranger that you've never seen before, but somebody who is in your environment. And with with respect to public notification, the deterrence effect shouldn't shouldn't differ by these relationships, because if you get convicted of any crime, it doesn't matter whether it's against a stranger or not, you're going to wind up on the registry. So, and that's what we find, the deterrent effect is the same in terms of the reduced incidence of offenses, you know, across across different types of relationships.

 

J.J. [00:39:56] With recidivism reduction, we also find that it does not vary across relationships. And remember, we actually find that recidivism increases. And and this suggests in our view, and this is what we argue in the paper, is that if what is really making a difference here and what really matters is not that acquaintances and neighbors learn about somebody's status and are able to protect themselves, but instead that offenders lives now are just much worse. They have housing difficulties and employment difficulties and and isolation isolation and loneliness and other psychological problems. We have no good reason to think that that would have the effect of increasing their likelihood of recidivating against any particular group. And and that's what we see, sort of an equal amount an equal increase against against all categories. So we have three categories. So really close in the strike zone and then people who are far afield strangers and we see roughly similar results across.

 

J.J. [00:41:10] Now, that we don't find with private registration. And private registration, you'll remember that we didn't find any deterrent effect, but we did find some evidence of recidivism reduction. And we argued there that this is actually a situation which there shouldn't be too many issues related to recidivism enhancing risk factors because the burdens were much lower and the stuff was supposedly kept confidential. And yet you're going to have police who are better informed and can keep an eye on people, at least in their neighborhoods when they are where police know of them. And that's actually what we find. We find that there is a reduction in crimes with with private registration, where the the incidence of crime goes down as the registry increases in size for for close close member, and this could include members of the family. And remember that the police would also be more likely to to investigate those sorts of crimes. Right. And and also acquaintances and neighbors.

 

J.J. [00:42:20] But we don't find any reduction in crime with respect to strangers. And that makes sense because the police having this information would probably be unable to to to really interfere with or monitor people when they're away from their neighborhoods or around victims who would constitute strangers. And so we we interpret that as all consistent with our earlier view that registration, private registration, works to reduce recidivism. But public notification, although it does as a threat, serve to deter first time offenders, it winds up exacerbating recidivism doesn't work the way we we think it will. And and this comes out of the fact that we don't see it operating to improve even the relative situation of acquaintances and neighbors versus other other potential victims.

 

Jennifer [00:43:13] And just to clarify what all this means for kind of current policy, so am I correct that basically everyone, every state, now has a notification law. These private registration laws don't really exist anymore? Or are they in place?

 

J.J. [00:43:27] That's right. That's right. So so what, you know, now we don't have anything like that. It matters outside of the U.S. because, you know, because all of the social science here has really consistently indicated that this doesn't really work to reduce reduce recidivism. Other countries have not, I mean, there may be one or two now that has gone the way the U.S. has. But last time I looked, Canada and Australia and and a number of other countries have private registration systems, but not public registration systems. And so, you know, this evidence would suggest they ought to stop there or at least go on from there. We know it has a ton of costs both to offenders and their families, but also to law enforcement to kind of to kind of keep people constantly registered despite the burdens on them of doing this. So it would suggest stopping there.

 

J.J. [00:44:26] And for for the US, where we might be thinking about this, you know, you can you can imagine a world in which you have a two registry, two registries going on. And there are certainly states in the U.S. that do this where you have a broad registry, anybody who's convicted of a sex crime, tier one, tier two, tier three, in terms of I mean, it's kind of a common categorization of people posing a particular risk. Everybody is on the the list that law enforcement knows about, but only a subset of those are put on the public registry. And if we take into account our work and also the the the the the paper I mentioned before, Duwe and Donnay, that find some evidence of recidivism reduction when you really target public notification on the worst of the worst. So, you know, they in their situation, it's only 10 percent of the highest risk offenders. So it's a pretty narrow sliver. It suggests a more targeted approach to public notification. So don't apply this willy nilly to everybody. Instead, there are likely to be some very high risk individuals who we think are highly likely to either impulsively commit new crimes or whatever the drivers of their behavior are are more predictable predictable. You know, reforming public registration systems, notification systems, to kind of focus on that group might be the way to go.

 

J.J. [00:46:02] I think there are a lot of people in this field who think that even that, although it makes sense kind of in the abstract, it's not the best approach that focusing on recidivism risk factors is the way to go. So instead of doing the opposite of what we do in so many other contexts where we work on trying to help people reenter society, reintegrate, we ought to be doing a lot more of that in the sex offender context: helping people find stable housing, helping them find employment, helping them succeed in treatment. That sort of thing is is likely to be I think almost all experts would agree, much more effective at reducing recidivism than trying to deputize the public to watch people, because whenever you deputized the public to watch people, they inevitably decide not to hire them, decide not to rent houses to them, decide not to form relationships with them. And and that has real costs to society.

 

Jennifer [00:47:04] It is interesting that like there's this broader movement in the U.S. right now to think about people with criminal records as human beings, right, who deserve our attention and some investment. And it does seem like sex offenders have been left out of that movement to a large extent. But as you say, it's a very similar problem. So I do want to make sure we talk about other work that has that has happened in this space. I know you've continued to work on this topic. You're coauthoring a book chapter with Amanda Agan, that summarizes the literature. So so let's talk about what else we've learned since you wrote the study you just talked about. As you do in your chapter, I'm going to divide the literature into four broad categories so we have the effects of sex offender registries on the recidivism of convicted sex offenders. Second, the effects of these policies on non registered individuals' criminal behavior. Third, the effects on the geography of where sex offenses occur and the victims who are targeted. And four, the effects of failure to register offenses. So these are offenses that are basically created by the existence of these registries. So whether whether the fact that a sex offender does not register with authorities actually tells us anything about their public safety risk. So let's go through these categories one by one. First, what else have we learned about the effects of sex offender registries on the recidivism of convicted sex offenders?

 

J.J. [00:48:28] I would say, you know, the the work continues in in much the same way that the work before ours continued. So there's a couple of of nice papers that focus on using criminal history records. So Kelly Socia and some coauthors have and there have been some other papers like this that have, again, just looking at a single state have taken- have been able to get criminal history information so they can study actually—OK, here is somebody who has been convicted of a sex offense. And let's look sort of pre and post the application of SORN to them and see how that changes the distribution and consequences of of sex offenses. And one of the big things that we learn from this, which I don't I don't think, you know, it came out- this paper came out roughly contemporaneously with my work with Jonah Rockoff is that is that most sex offenses, as many as 95 percent of sex offenses are committed by first time offenders. And so if you're spending a lot of time trying to reduce the number of sex offenses by focusing on focusing on recidivists, you're you know, you're you're probably focused on the wrong thing. And I think it's interesting because in lots of spaces, we think of crime being primarily committed by a small group of repeat criminals. And and that doesn't appear to be the case in the context of sex offenses. Now, part of that may be because if people are convicted of sex offenses, they're going to go to prison for significantly longer than for a lot of crimes. But if we if we think that criminal records are the way we're going to be able to kind of differentiate people who are going to be on something like a registry or not, it's it's important to learn that like 95 percent of the crimes that will be convicted next year, for example, in New York, will be by people who do not have a record of having committed a sex offense in the past. So that's one big lesson I think that's come out more recently.

 

J.J. [00:50:44] I'd say there have been a number of studies, some that have kind of replicated across state stuff like Amanda and I did in our papers, you know, looking at a set of four or five states. Some have done some work on juveniles and found similar types of stories coming out of them that, you know, that recidivism reduction doesn't really happen here. There there has been some additional work on deterrence and whether or not did the enactment of these laws have has reduced overall sex offenses? I'd say a lot of that work is either, you know, mostly finds that, you know, goes with the null hypothesis there, doesn't reject the null. So I'm not sure how much we've learned from that. Jillian Carr, there's a recent paper that uses changes in the length of registration as a way to identify the effects of registration on on crime. And and she also finds that there's not much going on here. So I'd say that even though our paper is eight or eight years old now—and since the first draft came out even older—we know that the story has not changed. I mean, other than this one paper that I mentioned, nobody has found any reliable, positive consequences on recidivism for public notification. And when it comes to private registration, there really hasn't been a way to test it because it's mostly no longer a policy option in the U.S. I mean, it'd be interesting to try to do it in Canada or Australia or one of these places where private registration existed a lot longer. But but that hasn't really been done.

 

J.J. [00:52:32] So let me move on to the effects on non registered individuals' criminal behavior. There, I think, as I mentioned there, there has been a little more work. And I'd say that the jury is still out on that. We found some evidence that—Jonah Rockoff and I found some evidence—that that it matters. But I think the the you know, we just don't have enough, I think, work since then to really know whether that that that that conclusion will stand the test of time.

 

J.J. [00:53:08] And when it comes to geography of where sex offenses occur and the types of victims who are targeted, there has been some more work. Amanda Agan and I worked- we wrote a paper in 2014 where we we actually had precise addresses over time of registered offenders in Maryland. And we laid that residential information over a geographic distribution of where sex crimes were happening, and part of this was just a descriptive question. We wanted to know whether or not this assumption that underlies a lot of this work that, listen, if you're living near a sex offender, you're at a higher risk of being victimized. We wanted to know whether that was true. And part of the reason we wanted to know that is because there is- there was some pre-existing work on how sex offenders commit their crimes. And it turns out that like many criminals, they don't commit their crimes right where they live because committing crimes right where you live tends to not be a great strategy in terms of avoiding detection and so on. And and and so, you know, there's a hypothesis that had come out about this. You know, there's a donut, right? Like, people don't travel five hours, but they do travel a half an hour or an hour away to commit their crimes. And so if that were true, you'd have this really weird situation where you would actually prefer to live near sex offenders because at least those sex offenders would not be committing crimes around you where you live. And and so we wanted to test that.

 

J.J. [00:54:43] And indeed, what we found in Maryland is that victimization risk was actually lower the more registered sex offenders lived in a particular census block group. Now, there are potentially a lot of reasons why that might be true. Let me just give you one. You know, sex offenders, as I mentioned, if they're if they're on a public registry, are socioeconomically going to be struggling. They may live in more disorganized neighborhoods that have fewer parks, fewer schools, fewer children generally. And so maybe the reason why they don't offend there is there are many fewer victims there. And so actually on a per victim basis, maybe even this lower offense rate is actually a sign that if you're in that category, there's higher risk. And so in the paper, we spent a lot of time checking those kinds of theories, and we really don't find any evidence to suggest that that explains our findings. So we wind up concluding that one of these strange you know, one of these assumptions that comes with this whole set of laws is that if you live near a sex offender, you are you are at more risk than somebody who doesn't live- who lives in a neighborhood without a sex offender. And that turns out not to be true.

 

J.J. [00:55:50] We then check to see whether or not that changed with the with the implementation of of sex offender notification. And what we find is that, if anything, the reported number of sex offense incidents in neighborhoods with a sex offender went up following a movement to public registration. So once people found out that there was a sex offender living in their neighborhood, the number of reported crimes went up. Now, this you know, there are a couple of potential explanations here. One is that, you know, now things that maybe, you know, they didn't have any idea who who who might the you know, who the who the offender was. Maybe, you know, now it made sense to actually report a crime. So, you know, reporting could have changed as a result of notification. I mean, even the existence of notification could have caused people to think that the government would be better able to help them if they reported crimes. And maybe that effect was stronger in neighborhoods with sex offenders. Who knows? But at least it seems that the reported number of incidents went up in neighborhoods with a sex offender. However, and that's by the way, I should add, that that's consistent with with public registration, increasing recidivism. Right. On the other hand, it doesn't go up so much that those neighborhoods are actually- doesn't go up so much that those neighborhoods are actually riskier than neighborhoods without sex offenders. So it increases- closes the difference a little bit. But neighborhoods with sex offenders are still relatively safer than neighbors without sex offenders.

 

J.J. [00:57:24] And you know that difference, by the way, the increase or decrease in the number of offenses varies a lot by the nature of the offense. So so that's that's another thing. And you mentioned types of victims. There's also variation across types of crimes. And we found that, you know, for example, the the the reported number of child molestations, you know, it's not very precise. The estimate seems to have gone down a little bit in in neighborhoods after notification neighborhoods with sex offenders. So so there's a lot of variation. I think there's still a lot of work to be done here to really understand that. And I think that's one of the main messages of this research, is that, you know, sex offenders include a lot of people, a lot of different types of offenders, and there are a lot of different types of victims. And and and so believing that these laws just have a single kind of average effect, maybe useful for deciding whether to have it or not. But we really have to get into the heterogeneous effects across different groups and different crimes to know whether or not there are some types of offenders and types of victims for for whom SORN laws might actually be really good or especially bad.

 

J.J. [00:58:34] And then you also asked me, I guess, about failure to register offenses. And there I'd say, you know, we really don't know that much here. So some people thought that failure to register may help us accurately predict recidivism like sex sex offense recidivism. And the results seem to be pretty mixed on this score. So there is some evidence that people who fail to register are more likely to commit crimes generally, but not any more likely to commit sex offenses. So they're you know, they're just generally less organized and more likely to to get into trouble generally. And one indicator of that is their failure to register. But it's not an indication, for example, that somebody is likely to recidivate specifically with with respect to sex crime. So anyways, I think hopefully that covered most of what you were interested in.

 

Jennifer [00:59:24] Yeah, that was great. So putting all this together, the results of your own, your own work and these other studies that you just talked about, what are the policy implications here. When policymakers ask you whether sex offender registration and notification laws are worthwhile? What do you tell them?

 

J.J. [00:59:42] I tell them that that even though it seems like common sense that finding out that there's a sex offender in your neighborhood, it seems like it's common sense that that would be, you know, either helpful or at least not not a bad thing. That that common sense inference is is actually not only wrong as a matter of logic, but also likely to lead them in the wrong direction. So, you know, these policies are essentially anti-reentry policies. And unless you believe that the information we're providing people makes them so much better able to avoid being victimized, that it compensates for the recidivism enhancing risk factors, that these things actually, again, counterintuitively make the world less safe. Now, I mean, there there there are people who thought a little bit about this. Why do you know, given that we it seems common sense, but now that we've seen the social science on it, why why don't we take it, take it, take the idea of of moving away from public registries more seriously? And I think there are psychological explanations for for wanting to know things even when we can't affect them. Or even if we knew that they might, they might actually make us less safe. We still have just a strong desire to not be ignorant about potential risks. And so there's, I think, a really strong demand.

 

J.J. [01:01:18] But you'll you'll hear people saying, policymakers, first and foremost what they need to do is move away from those who say, listen, if SORN- if registration and notification can save the life of one child or can save one child from being victimized, I don't care about the burden it places on offenders. And I think what this work shows us is actually your starting assumption that that at a minimum, these laws might say one child is wrong. Actually, they may put many more children at risk. And so I think policymakers need to go back to the to the drawing board and spend time thinking about how to how to combine these laws either with reentry oriented policies that can kind of counteract the negative consequences of publicity, or they need to think about restricting public registries to the set of people where we think the empirical tradeoff really makes sense, i.e., they are so likely to recidivate regardless of their situation, that giving people information is likely to be a good tradeoff.

 

Jennifer [01:02:24] And so what is the research frontier here. What are the big open questions in this space that you and others will be thinking about in the years ahead?

 

J.J. [01:02:30] Oh, that's a great question. So, I mean, I would love I would love to know more about about private registration and how effective it is. Like I said, it's a little bit hard to do research on that now that we're living in a world of just public registration. So I'd say, you know, people who are interested in in this possibility, you know, would be would probably do well to look at other countries and and see where where progress can be made there. I'd say, you know, there's lots of design details here. For example, what what coverage in terms of types of offenses should we be focused on? You know, if you really think that sex offenses of a certain type are committed by acquaintances or neighbors, maybe that's where we ought to be targeting registries. And if there is somebody who has been a rapist, who has gone out and and violently raped people who are strangers in other states, that's their M.O. Then this is a person for whom putting them on the registry is just it's not going to make a difference. It's you know, it's not it's not going to be a solution to the style of offending that this person has adopted. And I think the sooner states start moving in that direction and, you know, the world will you know, these things will will become either well, they'll become better in the sense that they will produce, I think, fewer negative consequences. And the more that we can get into the nitty gritty, gritty, on heterogeneous effects of these laws, the better. So I'd say that's mostly where the frontier is.

 

Jennifer [01:04:16] My guest today has been J.J. Prescott from the University of Michigan. J.J., thanks so much for doing this.

 

J.J. [01:04:21] My pleasure. Thanks so much for having me.

 

Jennifer [01:04:28] You can find links to all the research we discussed today on our website, probablecausation.com. You can also subscribe to the show there or wherever you get your podcasts to make sure you don't miss a single episode. Big thanks to Emergent Ventures for supporting the show. And thanks also to our Patreon subscribers. This show is listener supported, so if you enjoy the podcast, then please consider contributing via Patreon. You can find a link on our website. Our sound engineer is Caroline Hockenbury with production assistance from Elizabeth Pancotti. Our music is by Werner, and our logo is designed by Carrie Throckmorton. Thanks for listening and I'll talk to you in two weeks.

Jennifer DoleacDifference-in-Differences, Re-Entry