Going Home (or Not): How Residential Change Might Help the Formerly Incarcerated Stay Out of Prison

John Laub: Good morning, I'd like to welcome you to today's seminar. This is the first in this year's NIJ's Research For The Real World seminar series. My name is John Laub, and I'm the Director of the National Institute of Justice, and I want to thank you for taking the time this morning to join us.

Today's presentation is entitled, as you could see, Going Home (or Not): How Residential Change Might Help Former Offenders Stay Out of Prison, and it's featuring David Kirk from University of Texas at Austin. The first systematic study of recidivism in the United States was conducted by Sheldon and Eleanor Glueck and presented in their book Five Hundred Criminal Careers, which was published in 1930. In this work they found, and I quote, "That out of the 510 men who left the Massachusetts reformatory during the years 1911 to 1922, 80 percent were not reformed five to 15 years later but went right on committing crimes after their discharge." What was especially troubling about this finding was at the time the Mass Reformatory, which is now known as the Concord Prison in Massachusetts, was considered to be one of the most enlightened institutions in the United States. Fast forward more than 60 years later, we have data from the Bureau of Justice Statistics showing that among the nearly 3,000 prisoners released in 15 states in 1994, 67 percent were rearrested within three years. Throughout the 20th century and now into the new millennium, it's painfully evident that reducing recidivism is one of the most challenging and complex problems in criminology and criminal justice today. That's why I'm delighted that David Kirk is here to share with us the results of his research on this topic.

In his presentation, Dr. Kirk will explore the link between residential change and reduced recidivism. He'll discuss how the forced residential migration caused by Hurricane Katrina affected the likelihood that ex-offenders from New Orleans would be reincarcerated. Dr. Kirk will also examine potential strategies for fostering residential change among ex-prisoners and focus on parole residency policies as well as the provision of public housing vouchers.

Now it is my pleasure to introduce my good friend and colleague, Dr. David Kirk. David is Assistant Professor in the Department of Sociology, and a faculty research associate of the Population Research Center at the University of Texas, Austin. He was formerly an assistant professor in the Department of Criminology and Criminal Justice at the University of Maryland from 2006 to 2009, and I'll never forgive him for leaving. Prior to earning his doctorate in sociology at the university of Chicago, Dr. Kirk worked at the Urban Institute, and his current research explores — in addition to the topic today — explores the social context in neighborhood change on behavior and a lot of his work used the project on human development in Chicago neighborhoods data which was funded in part by the National Institute of Justice. One ongoing project that he has now is examining the structural and cultural predictors of neighborhood violence. Dave's work has appeared in the leading journals in our business, Criminology, American Journal of Sociology, American Sociological Review, Demography, and so forth. And I'd like to join you now in welcoming Dr. David Kirk.

[Applause]

David Kirk: Thank you.

Thanks for the introduction, John. I'll just be forthcoming and say that certainly the work I'm going to talk about today has been inspired by some of John's work with Rob Sampson on life course criminology. One of the things that came out of the narratives presented in the 2003 book, Shared Beginnings, Divergent Lives, is this notion that a lot of the desisters in that study, those individuals that managed to move away from crime, that they, or some of them talked about the importance of getting out of the old neighborhood, of separating from the old contacts, old neighborhood contacts because it wasn't a good thing. And so that little nugget was in the back of my head when I started thinking about Hurricane Katrina as a natural experiment and how it affected where people lived and ultimately started to think about the implications that it might have for their behavior in the future.

Anyway, so what I want to do today is to build a case for considering residential change as one solution for recidivism. So let me just jump right into it.

And so we have roughly 730,000 folks coming out of U.S. prisons each year, and this number has been... it's flattened out recently but certainly from '94 has been growing quite a bit. I think in the mid 1990s, we're talking about 500,000 individuals coming out of U.S. prisons, now we're on the order of 730,000 individuals. And we all deal, a lot of us deal with numbers like this every day and so many it's not shocking to us anymore, but it's still really shocking to me. This is a very large number.

One other number to keep in mind — and John mentioned the recidivism rates a minute ago — but actually more than half of individuals coming out of prison each year can expect to be back in prison within three years. This comes out of some studies produced by the Bureau of Justice Statistics, and John mentioned that in terms of rearrest, we're talking about two-thirds of individuals are going to be rearrested within three years after release from prison.

So I think one thing to keep in mind is that there's a huge impact on public safety, but also budgets. We're talking about high recidivism rates like this. I sure wish that the state of Texas spent less on incarceration and spent more on public schools because I've got some school-age children that are going to be entering the public school system pretty soon, and I wish that a lot of teachers didn't have to lose their jobs last year because of budget cuts, but anyway, that's a side point.

Let me get to some motivating facts. A lot of excellent research out there, particularly by the Urban Institute, has demonstrated that we have a pattern of concentrated prisoner re-entry in this country where folks coming out of prison tend to go back to urban areas, tend to go back to select neighborhoods in urban areas, that are oftentimes resource-deprived. The other part about this geographic pattern is that oftentimes individuals return home, if not to their very same neighborhood where they resided prior to incarceration, at least in proximity to where they resided.

So why are these facts consequential? Well, there's quite a few reasons, I'm going to talk about a couple of different reasons. One is peer influence; this notion that the likelihood of criminal behavior increases with exposure to others that are engaged in criminal behavior. So the implication is that if we remove individuals from their criminogenic social network, then maybe they have a decreased likelihood of engaging in crime. Now I think there is a fair amount of researched evidence that supports this contention. And so I'm going to give you one example: My colleague at the University of Texas, Mark Warr, has investigated peer effects and has found that at least one reason for the decline in criminal offending with age is because people are less likely to hang out with criminal peers as they age.

He's also looked at the reasons why marriage tends — is predictive of desistance from crime, and has found that it's because with marriage, individuals are less likely to hang out with criminal peers.

Now it doesn't take a Ph.D. to understand this scenario. The cartoon is that The New Yorker talks about the very same thing. Let me just say that I stole this slide from a presentation that John has made in the past because I think it's very effective in demonstrating a point about marriage and the importance of marriage for reduction in criminal behavior. And so what we have here is a gentleman sitting in his chair, about four or five beers into a six-pack, perhaps getting ready to hit the town with his friends, and presumably his spouse, or his spouse has a different idea. And so she's got the cue cards right there that instructed him what to say to his friends on the phone, and the text, the fonts are a little small, so I'll tell you what it says. "Sorry, guys. I'm spending the evening with my lovely wife."

The message here is that there's a variety of mechanisms why marriage is predictive of desistance from crime. But one reason is because it means that individuals are less likely to hang out with criminal peers. What that means from my study is I've also tried to think about what other mechanisms, what other things may be in place that will lead to a fragment of criminogenics social networks. What other things can lead individuals to hang out with their criminal peers less often? And that's led me to residential change. If somebody moves to a new neighborhood, perhaps they're less likely to hang out with their criminal peers.

Another thing to keep in mind when thinking about the importance of residential change for behavior is opportunity. When individuals return home to the same environment where they reside prior to incarceration, hanging out with criminal peers, exposed to the same set of conditions and criminal opportunities that got them into trouble with the law in the first place.

There's actually a lot of fascinating research coming out of the addiction field that shows that drug addicts when confronted with familiar stimuli, familiar environmental stimuli, oftentimes have a physiological reaction to place, and so in some sense, sending...if a drug addict goes back to a place where they have bought drugs in the past, or used drugs in the past, it's a losing situation. They have a higher likelihood of engaging in continued drug use. And so, residential change, going to a different neighborhood may be one way to separate a drug addict from those environmental cues that contribute to further drug use.

And so moving forward, one of the things that I thought a lot about are turning points in the life course of crime and what are the various things that can produce a turning point. Marriage, military service, these are the types of things that certainly come out of the research by John and Rob Sampson, but I want to explore residential change as a turning point in the life course of crime that contributes to desistance from crime, desistance being a process that supports determination of criminal offending. And so if somebody moves to a new area, has a fresh start, does it mean that they're less likely to engage in criminal behavior in the future?

And so I mentioned all these points about opportunity, peer influence, desistance to suggest that. There's one other thing that we should be thinking about in terms of turning points in life course to crime, and that's residential change, and so the rest of my time today I'm going to explore the importance of that particular turning point.

And so how am I going to do that? Well, so I've been engaged in a study for about the past five years, and I'm going to talk to you guys today about some recent findings on what's kind of a larger project, but I've used Hurricane Katrina as a natural experiment for investigating the importance of residential change for criminal behavior, specifically recidivism. It's been six years since Katrina hit, so I'm going to put up some pictures and some maps just to remind us all in case we needed a little reminding about Hurricane Katrina.

This shows the progression of Katrina. August 23rd, 2005, Katrina developed around Cuba, and that's Point 1 on the map. Made its first landfall in the United States around the 25th of August, southern tip of Florida. I believe it was a Category 1 storm at the time. Then it moved westward into the Gulf of Mexico, rapidly intensified, becoming a Category 5 storm by the 28th of August, and then the morning of the 29th of August made a second landfall in the United States, the Southern tip of southern Louisiana and moved north towards New Orleans soon thereafter, and we all know the story. The levees in New Orleans breached the city, started flooding and property damage, certainly all throughout the New Orleans metropolitan area but certainly throughout the Gulf Coast region as well.

So this is a map of the flood depths taken in New Orleans a couple of days after the city started flooding. The scale goes from...red is least severe flooding down to green and blue which is the most severe flooding. Now I want to be clear, no demographic group was unaffected by this storm, but it is true that those areas hardest hit by property damage from Hurricane Katrina tended to be poor minority neighborhoods, and it's in these poor minority neighborhoods where ex-prisoners when they've come out of Louisiana prisons would typically reside. At least that's where they resided prior to Hurricane Katrina. And so what Katrina did was it devastated the housing stock all through New Orleans and in particular the neighborhoods where ex-prisoners would reside, and so I want to explore the implications of that for prisoner re-entry.

Just to make the point even stronger, flooding throughout the area, highways submerged, streets submerged.

Neighborhoods submerged.

And when the water finally receded, many neighborhoods looked like this. Housing structures absolutely decimated, rubble. If the structures were still intact, they were likely moved off their foundation down the block, and so again, extreme amount of housing destruction in the New Orleans metropolitan area.

Back to this notion of a natural experiment: So what this did was it induced some individuals to move to new areas who otherwise would not have moved had it not been for the hurricane. And I'm going to show some figures in just a minute about the extent of the change in residential patterns, but ultimately what this allows me to do is it allows me to compare the likelihood of recidivism for two different groups. The assumption here is that people that were coming out of prison in Louisiana, pre- versus post Katrina, roughly similar on characteristics such as family background, employment history, criminal history, so on and so forth, all the predictors, a lot of predictors of recidivism that we talk about, but a difference between the pre- and post-Katrina folks was that those individuals coming out of prison post-Katrina were much more likely to move, to move to a different parish. And so it sets up this comparison, and I basically just look at the recidivism rates across these different groups.

This slide I use to demonstrate some of the changing patterns of residence, pre- versus post-Katrina. And so what this demonstrates, this is based on sample individuals pre- versus post-Katrina that are originally from the New Orleans metropolitan area, prior to going to prison. So if you focus on the leftmost bars, Orleans parish, those individuals originally from the New Orleans area who came out of prison prior to Hurricane Katrina, roughly half of them would go back to Orleans Parish. Orleans Parish is the central part of the metropolitan area. Post-Katrina, the distribution changed, changed significantly, about 20 percent would go back to Orleans Parish. So where do these folks go if it wasn't to Orleans Parish? Well, some went to Baton Rouge, two percent before the storm versus 11 percent afterwards. If you focus on the rightmost bars, which is labeled other, proportionally more folks went to places like Lake Charles, Lafayette, Shreveport, and so the message here is that the large difference in the residential patterns pre- versus post-Katrina of where people who originally from New Orleans would go once they got out of prison.

Similar point: Pre-Katrina, individuals that came out of prison, about a quarter of them would go to a different parish relative to where they resided prior to incarceration. Louisiana is actually a little bit different than a lot of states. There's no parole condition or law that says that parolees have to go back to their old county. They are allowed some flexibility in Louisiana, and prior to Hurricane Katrina a quarter of them would go to a different parish. Post-Katrina, it's 50 percent, and so what we can assume then is what Katrina did was it induced an additional 25 percent of individuals to move to a new parish. And so the way I set up the analyses that I'm going to show you in just a second, I mean basically what I'm doing here is I'm taking these 25 percent of individuals that were induced to move and I'm comparing them to their pre-Katrina counterparts who went back home.

So let me quickly go through some of my data.

So I've been fortunate enough to get access to some data from the Louisiana Department of Public Safety and Corrections, provided me a lot of individual-level data on ex-offender characteristics, but have data on reincarceration through three years. Importantly, also have information on place of residence. In order to isolate the specific effect of residential change, moving to a new area relative to where somebody resided prior to incarceration in order to isolate that effect. I also use a lot of different information on the socioeconomic context, the social context of where people resided.

And so point 2 on that slide, I've data from the Census, HUD, Louisiana Department of Labor, and ESRI, to establish what the neighborhood environments are like, and what the county environments are like where people were residing. So information on average wages, fair market rent, unemployment, things like that.

I also control for in my analyses pre- versus post-Katrina, differences in the operation of the criminal justice system. This is important because Katrina might have affected how the Criminal Justice system operates. And so I have information that I include in my analyses on average contacts made by parole officials, pre- versus post-Katrina. It's actually the case that when people left New Orleans, the size of the parole staff in New Orleans actually stayed the same so they had less caseload and more opportunity to scrutinize offenders in New Orleans, and so I tried to control for this change in the Criminal Justice practices by including the variable — analyses in the average contacts by a parole officer.

Quickly about the sample. So I have two different cohorts to correspond with the two different time periods, pre-Katrina, individuals who were released from prison from September 2001 through February 2002, and then the post-Katrina period, September 2005 to February 2006. These individuals I use in my analyses are those individuals who are committed from the New Orleans metropolitan area, originally from the New Orleans metropolitan area, not a statewide release cohort, just those from New Orleans.

Let me just jump into questions and hypotheses.

One of the things that I was wondering is, after showing those figures, we know that there has been a change in the residential patterns, pre- versus post-Katrina, and so the question that came to mind for me was, well, are those individuals that move, are they moving to better neighborhoods, worse neighborhoods or is there no difference? And I hypothesize that it's likely they're moving to similar neighborhoods or perhaps even more disadvantaged neighborhoods, and I say more disadvantaged because the social context after Hurricane Katrina was extremely difficult for lots of individuals, including ex-prisoners. So if people moved, I would suspect it wasn't that they were moving to fantastic environments with lots of jobs and things like that, they may be going to slightly more disadvantaged contexts. And the reason this is important is if I'm trying to isolate the effect of residential change on recidivism, there may be a number of different mechanisms why that may occur. Somebody moves, they may have lower likelihood of recidivism; it may be because they are moving to an area with lots of jobs. Or it may be because they're moving to an area where the criminal peers no longer are present, and so by looking at social context and trying to dissect the reason why moving may be important.

 

So let me just quickly summarize what I found. Unemployment rate was actually significantly but not substantially higher where the movers went to. And the distinction here is movers versus stayers, the stayers are those individuals who returned to their older parish when they exited prison; movers are individuals that went to a different parish relative to where they used to reside.

Unemployment slightly higher, wages, a little bit lower, household income, a little bit lower for the movers. Fair market rents, quite a bit lower for the movers, and the reason for this is that if somebody is moving, they're more likely to move outside of New Orleans, and New Orleans has the highest fair market rents in the state. Segregation, roughly similar across the groups.

But the message here is that in terms of the social economic context, those individuals that moved to a new place were going back to a similar environment, maybe even slightly more disadvantaged than those individuals who moved back home. Moved back home to their home parish. And so again, if we were thinking, if these were the only predictors of recidivism, then we may conclude that those individuals who moved should actually have higher levels of recidivism. And just to give you the punch line that I'm going to explore in just a minute, I actually found the opposite, that those people that moved had a lower likelihood of recidivism. And my argument is that moving may still be consequential, may still be a good thing for individuals, even if they're not moving to a great neighborhood in terms of jobs and socioeconomic conditions, but moving may still be beneficial for desistance from crime if individuals are allowed to, if we give them an opportunity for a fresh start, give them an opportunity to separate from the criminal peer networks and from some of the criminogenic influences that influenced their behavior in the past.

That leads me to my second question. Does residential change serve as a catalyst for desistance from crime and long-term behavioral change? And my hypothesis is that it does; that the likelihood of recidivism is lower when individuals reside in the geographic area different from where they resided prior to incarceration.

The first set of results I published out of this project back in 2009, I found that there was a substantial significant marginal effect. Difference between those who move back to their old parish versus those who went to a different parish. Now let me just state that the reincarceration measure that I used in this slide is very inclusive. It includes reincarceration in a Louisiana prison facility, which makes up the bulk of recidivism represented by these bars, but it also, the measure includes reincarceration in local facilities, federal facilities, and other state facilities. So these numbers are high, shockingly high, even among the movers but part of that is because it's a very inclusive measure of recidivism. But anyway, so the point I want to draw your attention to is the difference in the height of these bars. So those individuals coming out of a Louisiana prison who went back to their home parish, almost 40 percent of them back in prison within... back in some form of detention within a year, versus those who moved to a new area, a new parish, 10 percent lower.

And just coming back to what I was saying before about mechanisms, I propose that the reason is because those folks who moved to a new area have benefited from a fresh start separating from the criminogenic influences in their past, criminal peers. Those people who have moved have had a greater likelihood of establishing a new set of daily routines that are conducive to desistance from crime.

After I published those initial results, the skeptics came out, and I'm really glad that they did because I've been engaged in some follow-up work for the last couple of years to see if the one-year findings that I published back in 2009 are still true if we extend the time period by a few years if we look at three-year reincarceration rates. And so there's a few reasons come to mind about why the effect, this treatment effect may have disappeared over time, and so I'm going to talk you through but basically I think there's three hypothetical scenarios which may have happened, which I ultimately investigate.

What do we mean by short-term pause in behavior? Those individuals who moved may have been temporarily delayed in their eventual recidivism, eventual criminal behavior, and so when somebody moves, if we think about criminal opportunities, maybe they're temporarily blocked from criminal opportunities, but eventually criminal opportunities are all over the place. If somebody wants to find an opportunity to engage in crime, they eventually will, even if they're temporarily delayed in doing so. If we think about criminal peers, if somebody moves, maybe they are fragmented from their old criminogenic social network, but it doesn't take long for somebody to establish new criminal ties, even when they move to a new area. Finally, scrutiny by law enforcement. When somebody moves to a new area, perhaps they fly under the radar for a little while, police don't recognize their behavior. But if somebody is engaging in criminal behavior with enough frequency, doing it long enough, eventually the police are going to become aware of their reputation, become aware of their transgressions, and so somebody maybe committing the same number of crimes may not be caught in the first six months or a year because the police don't really know about it, but eventually the police may end up doing something, end up catching them and arresting them, sending them through the court system.

So there's I think a lot of reasons why the one-year effect that I published in 2009 may have disappeared, may have diminished over time, may have reversed course, and so that when I've conducted some follow-up research, it's been to investigate this idea of whether there was temporary change and behavior from moving or whether there was true change.

And if you're a visual person like I am, let me just show up, put up some hypothetical graphs which demonstrate what I'm talking about. This is the hypothetical hazards of reincarceration within 12 calendar quarters, after somebody is released from prison. And so what I mean by short-term pause is there's initially a gap between movers and stayers in their likelihood of being reincarcerated assuming that they had not, if they had not been reincarcerated already. But around the 12-month mark, 15-month mark, ultimately those curves converge and so that there's no distinction between the patterns after a short period of time, so short-term pause.

Another scenario is what I call catch-up; that those individuals who have a high likelihood of recidivism that moved, maybe they delayed their eventual behavior for a year or so, but eventually they were only incarcerated, so the stayers were reincarcerated immediately, the movers, it just took a little while. And so what we have here is an initial gap and then they basically cross over time.

This third scenario is what I call real behavioral change. The idea that there's a consistent gap through 12 calendar quarters after prison release. And so even three years after somebody has been released from prison, when the likelihood of reincarceration is pretty low, there's still a gap between those two curves. And so these are the scenarios, the hypothetical scenarios that I sought to investigate when I did my follow-up work after I published those initial studies, initial findings.

I have a forth-coming article due in April or May or next year. I'm happy to distribute the copy to you guys if you'd like, just send me an e-mail or I have the University of Texas Sociology Department, I have a homepage where there's a link to a copy of this article.

So I'm going to show you the actual results, and basically what I found is scenario number three, what I characterize as real behavioral change that every point in the first three years after individuals had been released from prison, those who moved to a new parish are significantly and substantially less likely to be back in prison than those individuals who went back to their old parish.

And so a similar slide as I showed before. Before it was a 0.01 difference after one year and after three years it's expanded, it's a 0.15 marginal effect. So those individuals who stayed in their own parish, meaning when they got out of prison, went back to their old parish, 69 percent expected to be back in prison within three years. Again, this is a very inclusive measure of reincarceration, versus those people who moved to a different parish, 0.54.

This is what is known as a survival curve. Same story, just a different way to present the data. Of those individuals who moved to a new parish, 46 percent of them survived, meaning they did not end up back in prison within three years, 31 percent of those people who went back to their old parish survived. The rest went back to prison.

So what now? So I want to talk about some implications from my findings. The Urban Institute has done a lot of research on prisoner re-entry and one of the things that comes out of that research is that, in interviews and discussion with returning prisoners, oftentimes individuals exiting prison don't want to go back to their home neighborhoods, but end up doing so anyway, even though they realized that it may be bad for them, it may contribute to criminal behavior in the future. And so why do they do so? I mean honestly, there's lots of reasons why individuals go back home. Familiarity with place, attachment to place, loved ones, things like that—

—but there's also some institutional reasons that I wanted to talk about today, and one, parole residency policies. In most states, prisoners released to parole supervision are legally required as a condition of parole or mandatory supervision, to go back to the county where they were convicted, or the county where they resided prior to incarceration.

My home state Texas is certain an example of this. Texas Government Code 508.181, says that those individuals as a condition of parole or mandatory supervision are required to go back to the county where they committed their crime, or where they resided when they committed the crime, or if they lived, or if they're from out of state, they have to go back to the county where they did commit their crime.

And so my point is that I think about my findings from Louisiana, allowing ex-prisoners, parolees, to go to different counties may in fact be beneficial when we're talking about public safety. So may be fruitful for states to reconsider residency restrictions put upon ex-offenders.

Now of course, even if we as a society allow greater flexibility in our parole policies, that doesn't mean that individuals aren't going to necessarily just get up and move to a different country, because there's other reasons why people move home to their old neighborhoods. And so, for those of us that think about prisoner re-entry a lot, one key issue is housing. As a whole we're talking about individuals on average didn't have a lot of income before they went to prison, coming out of prison without much money in their pocket, and so therefore it's hard to find a place to rent in the private housing market, even if they had income, but there's also a potential discriminatory practice in the private housing market against individuals with felony convictions on their record. And in terms of public housing, there's also, difficult to gain access to public housing whether we're talking about housing developments or housing vouchers. But one of the things that I wanted to talk about today was, housing is so key, especially housing, providing housing in an area that's different from where somebody resided in the past then what can we do as policymakers to try to support or promote residential change and so incentive moving, but how would we do that. And one idea is to make housing vouchers available to individuals, greater access to housing vouchers to individuals with felony convictions but to do so in areas that may be some distance from where they resided in the past. And this is a realistic idea? Well, I wanted to come here today and hear from you guys whether this type of idea would really fly but say I was encouraged from some things I saw this spring and this summer that maybe providing greater access to public housing, housing vouchers for ex-offenders is something that we can do.

As you've probably seen, the Federal Agency Re-entry Council has released a lot of these myth busters about what's fact versus what's fiction with prisoner re-entry and one of them deals with public housing, and so the myth here is that individuals who have been convicted of a crime are banned from public housing. The realty is, is there's two instances, where under federal law somebody is deemed ineligible for public housing. Lifetime registered sex offenders and individuals convicted of making meth in federal public housing. Otherwise there's discretion amongst public housing authorities about how they use their funds.

I know you can't read this, but this is a letter from HUD, from Secretary Donovan and Assistant Secretary Henriquez, who talks to... addressed to Public Housing Authorities that addresses this issue.

And so I've cut out a blurb here that I think is quite interesting, and I'll read it. As President Obama recently made clear, this is an administration that believes in the importance of second chances – that people who have paid their debt to society deserve the opportunity to become productive citizens and caring parents, to set the past aside and embrace the future. Part of that support means helping ex-offenders gain access to one of the most fundamental building blocks of a stable life – a place to live. And so I've been encouraged by the signs from our federal government that this recognition that housing is a major issue when we're talking about ex-offenders, a lot of the reason why people end up back in prison in a pretty short amount of time is because they don't have housing, they don't have access to housing, and don't have a place to... don't know where they're going to put their head that first night, that first week out of prison. And so maybe there are some things that the policymakers can do to at least solve this one problem, this one issue that ex-offenders face. By my argument, if we're going to think about a greater access to public housing and resources for ex-offenders, maybe it makes sense to think about where those housing vouchers should be issues. So in New Orleans if they're going to receive, the ex-offenders, if they are going to receive housing vouchers, maybe it shouldn't be in New Orleans, maybe it should be in Baton Rouge is the point that I would make.

So let me switch gears and talk about money because I think when you're talking about policy, you have to come back to dollars. And so I want to tell you what the cost savings have been like in Louisiana from a reduction in recidivism. So just for parole revocations, Louisiana has about 600 fewer nowadays than they did prior to Hurricane Katrina. By Louisiana estimates, they spend a little more than $43 a day to re-incarcerate an offender. Parole violators spend, again from the data, spend about 18 months once they go back to prison.

 

So we're talking about $14 million per release cohort in savings. No I think you saw there's a figure published by NIJ or BJS that suggests that we spend something like $74 billion on incarceration in the United States. I can't recall if that's just state or if that's state and federal, so in the grand scheme of the total dollar spent, $14 billion, or $14 million is certainly less than $74 billion, but this adds up. I mentioned the Austin independent school district a while back, I sure wished that school district had 14 more million dollars per year. I'm sure John wished that NIJ had 14 more million dollars per year. So it adds up and so if, for example, loosening some parole residency restrictions lowers recidivism and represents cost savings to the states, then in my mind that's something we should think about. Lots of the programs that have been effective in lowering recidivism, they cost money, this may be unique, moving may be unique in that it could be done through lower recidivism through a cost savings, kind of the double whammy if you ask me.

So what now? So basically what I've talked about today are results from a natural experiment, a horrible catastrophe that just happened to have a bit of a silver lining for some ex-offenders. But what I've been thinking about for the past couple of years is how do we perhaps promote what I've learned, what I've learned from a natural experiment, how does it translate to a real world policy environment and so I've come over the...and this is before we do some kind of whole-scale rewriting of parole policies and start telling public housing authorities how to use their money.

I've thought about an idea, starting with a small demonstration project. It may look something along the lines of these type of details. Assigned 300 volunteers, people coming out of prison asking them if they want to volunteer for a demonstration project. Three hundred of them, individuals, set to be released to parole go onto a treatment group where they are required to move to a new city, or assigned relative to where they lived in the past, or assigned to parole, to a parole office in that location away from where they loved in the past, and they receive a housing voucher. Have a comparison group that goes back to the old city but still gets the housing voucher. A control group that goes back to the old city, but doesn't get the housing voucher and so setting up kind of this three-group design would allow a researcher to investigate the effect of moving to a new city, but would also allow them the possibility of investigating whether their greater access to housing vouchers would be predictive of a reduction in recidivism. And also the combination of the two, if a housing voucher is beneficial for an individual, is it better if that housing voucher is issued in an area some distance where somebody lived in the past. So the final piece is just for the evaluation team, the research team to just compare recidivism outcomes across groups.

So anyway, I was anxious to come to D.C. today to talk about my research but also to put an idea out there, and you guys are the experts. I sit in an office in the UT Tower and crunch numbers and occasionally try to impart some lessons to my students, but I wanted to learn from you guys about the ideas that I presented in terms of residential change as a way to produce a turning point in the life course of crime, how can we do that? How can we promote that as policymakers and practitioners, and here's my idea but I'd love your feedback on it. Thank you.

[Applause]

Laub: We're going to open up the floor for questions.

Bill Sadler: Bill Sadler, BJS. I have two questions. So when you started your presentation and talked about the national experiment, I was interpreting Katrina as forced mobility, kind of like Todd Clear's coercive mobility, but it turns out there's really choice. And so when I'm looking at the results, a quarter of your study went back to their old neighborhoods, if I've got the data right. So how much of that difference is really due to selection effects, those people who are given little incentive decided to stay, or really a difference between groups of offenders? So one, if you could address the selection effect.

Second, I was in Cleveland in the 2000s and a similar type of thing was going on with HOPE VI where they were demolishing public housing and using vouchers, and one of the things I was able to observe between '90 and 2000 at that time was the change in the neighborhoods people moved into as a result of the vouchers. Do you have any results that look at particularly if there is high concentrations of offenders moving into new neighborhoods, what's the impact on the new neighborhoods? Because in Cleveland there was massive social change going on at that time including increases in crime rates, in the moved-into neighborhoods. So if you could talk about those two things I'd appreciate it.

Kirk: Thanks, Bill. So start with the second point. The effect of changing residential patterns on receiving neighborhoods I think is a great question. With the folks coming out post-Katrina, there has been what I would characterize as a dispersion of the population. Some reconcentration, but not like we saw prior to Hurricane Katrina. And so you've got folks spread out all over the place, and it's certainly a question I need to investigate further, but if we've got three in Shreveport and five in Monroe, and ten in Baton Rouge, that there's not this concentration and I would suspect not a huge impact on the neighborhood environments from a reconcentration of parole, but it's certainly an important question and one that deserves a lot of attention.

In terms of...so we're talking about the fact that there was still some choice, and even if individuals post-Katrina couldn't move back to their old neighborhood, they still had I guess some choice to move somewhere, right, if it wasn't New Orleans, and that's a challenging thing, I'll admit, that's a challenging thing to account for. But the thing about the post-Katrina context was there was some choice, but very minimal, and part of it was because of competition with all the other people that were displaced from New Orleans. So there wasn't a lot of...probably less choice than in a normal situation. So I think that handles the selection issue to some extent, but I'll be honest, I mean the benefit of a natural experiment is it has allowed me to address selection to a certain extent but I haven't totally eliminated it. So I'll just leave it at that, that using a natural experiment, throw in a bunch of control variables, run a bunch of sensitivity analyses, but there's always a possibility for some selection, and so that's why, one reason why I've said okay, I think we've learnt a lot, or I know I've learnt a lot from Katrina, let's do a small demonstration project, randomized demonstration process. Let's push the science even further, to make sure that this works, and then if it works, then let's think about a broader scale implementation. Thanks, Bill.

Eddie Ellis: Hi, how are doing? My name is Eddie Ellis, Community Representing One by One. I have a question and statement. My statement is I've actually done 15 years myself in prison. What's the difference between someone's who done 15 years and 15 days, how would you all address that, first of all? And my question is, I mean my statement is what I want to say is, when dealing with this type of program, and when you move people, the support when people come home, a lot of these men and women who are dealing with mental health problems, how would you all address these mental health problems because now these mental health problems are really not being addressed now. So how would you all address these problems when they are really not being addressed now?

Kirk: Right. Great question. The first one on the notion of an individual that's been in prison for a while, 15 years versus 15 days. So the initial findings that I presented back in the 2009 publication, I broke it out by whether somebody had been in prison multiple times versus just on one occasion, and that the gap between those bars was consistent whether we're talking about repeat offenders or first time releases. And so what that means to me is that different types of offenders could all benefit from a fresh start, from getting out of the old neighborhood.

In terms of services, I think it's a great question. So when I am talking about this demonstration project of saying, find a way for folks to move to a new area, ideally this would be done where they would still receive the same services they would have received if they had gone back to their old city. The services are the same and that helps, well, with the randomization process and trying to figure out if moving or vouchers is what led to any kind of change and behavior. In terms of just across the board, it had to be provide better mental health resources and things like that. I'd open it up to the audience, the people that deal with budgets, part of it is finding money in a tight economy to better fund those types of things, but services for mental health and drug addiction are certainly vital. We're talking about successful prisoner re-entry and reducing recidivism, now that is absolutely important.

So when I talk about place of residence and changing of place of residence, it's not an all-purpose elixir, all-purpose solution to recidivism, it's one component of a process. But as I said towards the end of my slides, what I think is an interesting part about it, is that there may be some cost-effective ways to promote residential change. And then those cost savings maybe could be used towards as a form of justice reinvestment, used for more mental health resources, used for more police, and things like that, so mental health resources are certainly a key aspect of it, and need to find ways to fund it. Thank you for your question.

Ellis: Okay. Thank you.

Nancy La Vigne: Hi, David. Nancy La Vigne with the Justice, the Urban Institute, Justice Policy Center. A very good presentation, very provocative, really got me thinking and you did reference our own prisoner re-entry research, and there was one big component of our research and our findings that you were unable to include in your own research because it involved interviewing people behind bars, following them over a year's time in the community, and that is the role that family plays in prisoner re-entry. And I would venture to say of all our findings from our Returning Home study, which was a multistate study, a total of 1,500 men and women that we tracked over time, is the important role that family does play in prisoner re-entry and then successful outcomes. And we were actually pretty stunned and surprised to learn how important family is, the amount of support that they provide of returning prisoners regardless of the criminogenic nature of the families themselves. There typically is someone there to support nearly every person who leaves prison, and so I was thinking about your findings through the lens of our own findings about family support, and the degree to which family support, if you had, certainly if you had a policy where you encourage people to live elsewhere would they be removed from that support system. And then looking at your own research I guess a question whether that unmeasured variable could have had a lot of impact on your findings whereby people who moved elsewhere moved elsewhere to be with their families, and therefore, had more support and so then creating a policy that encourages people to move elsewhere based on a very interesting study that I just don't see as generalizable from a policy perspective. I have concerns about taking this and running with it and saying people should live elsewhere. I mean we even found that negative family influences had less of a negative impact on re-entry outcomes than did positive family influences is another thing we heard from people, because it's not just peers, it's negative family influences that you might be thinking you should remove these people from. But based on our extensive research, I would argue otherwise. So that's one point.

The other's just a question because I was formulating this question in my mind while you were talking about cost and cost savings. Did you include the savings of averted victimization costs, so the societal cost in your estimates?

Kirk: So on that second question, no. Those are just the cost savings from reduced incarceration. There's more savings, sure.

La Vigne: No, I would argue against including that although I think the real cost to victims is not real cash that you can reinvest, so I'm glad you excluded those.

Kirk: Right.

La Vigne: So I'd be very interested in hearing your thoughts on this unmeasured variable.

Kirk: Sure. And that's an excellent point. I'm glad you brought it up. Thank you for that. So yeah, when folks go back to their old neighborhoods, part of the reason they do so is to be with family, and families can be supportive. Not always, but they can be supportive. I love reading research that talks about when families are helpful for promoting desistance crime and under what circumstances they may not be all that great. In terms of pushing this idea and pushing the science on it, I'd say well, something like this, maybe moving with that supportive housing is the best idea. The families are supportive, they help individuals steer clear of future criminal behavior, well then, if we're going to provide vouchers for individuals, why not both for the individual and their immediate family. And test that idea, in a demonstration project for the very reasons that you're talking about. And so I agree, families can be very important but perhaps there's a way to include them into the demonstration in order to foster kind of the ideal circumstance. Separating individuals perhaps from their old buddies which may...may be the individuals that were selling drugs and whatnot, but move the ex-offenders with their family to help them stay out of prison. So I think a combination is perhaps the way to go.

Linda Rose: Hi, my name is Linda Rosen; I work with OJJDP. My question concerns youth who are coming out of public housing, I'm sorry, coming out of facilities and don't want to live in public housing because they don't want to be in a concentrated area with people with similar problems, they may have a felony on their records. Can these vouchers be used for other forms of housing other than "public housing?"

Kirk: I know there are some folks from HUD in the room that could speak to this issue probably better than I can. So we had a fundamental shift in the country and the way we handle public housing, so away from the dense model of high rise public housing, more towards use of housing vouchers in scatter site housing, for the reasons you're talking about, this idea that concentrated public housing is not necessarily a good thing. And so in that type of situation, I'm specific when I talk about housing vouchers that those could be used in a private housing market. I think the reality is that the private building owners and managers have discretion in whether they want to accept them so there's still some concentration even when we talk... of people with housing options in terms of where they can live, there's still some concentration, perhaps not as much concentration as with the dense high-rise facilities that characterize kind of post-World War II housing developments.

And so, back to your initial question, in your scenario, a youth coming out of detection doesn't want to go back to public housing complex, public housing development, maybe they could get a voucher and try their hand and try to see if they could find some place in the private housing market, outside of the old neighborhood, see if they can get a building order to take that voucher.

D.J. Ervin: My name is D. J. Ervin; I'm with the Milton Eisenhower Foundation. I was wondering, were you able to break down you results by type of crime, drug crime, non-drug, violent, non-violent, serious Part I versus Part II, anything like that?

Kirk: Not very well is the answer. Part of it is kind of tracing the steps back from parole revocation to the initial reason somebody was imprisoned; it's kind of difficult to follow the path. But in Louisiana prisons, I mean as you can imagine, we've got, there's a lot of individuals incarcerated for drug violations and violent offenses as well, so a lot of the releases are coming out, especially those with kind of shorter terms or individuals that were locked up for drug crimes.

Donald Zimmerman: Good morning. I'm Donald with the National Homecomers Academy. As I look at these findings, you have 300 homecomers that are coming home that are going to be going to a new city with a housing voucher and going to have treatment. Have you defined what kind of treatment they're going to have because I know by just giving people a housing voucher, that's an art, to come home and have to take care of a household. And also, with the other 300 that are going to go to the same city with a housing voucher, again, define the kind of treatment that they're going to get. And as far as with the other 300 homecomers that are going to be coming home with nothing, I don't think that should happen. I actually think that everybody should get something whether it's some type of treatment because we're dealing with people's lifestyles. We try and change the culture of what they are doing and not try to do stats. It truly has to come to the sense of taking these people into consideration that they do have families and we can't play around, we can't mess up.

Kirk: Thanks for that question. Yeah, let me clarify. So the language that I'm using in this slide is, the language of it, experimental design. So when I talk about this bottom group of what I call control, this is kind of the status quo group. This is how we're doing it now in large regard. Should we have greater services for individuals? Sure. So when I talk about a comparison group and a treatment group, the treatment that I'm talking about here is basically the distinction between the groups. The treatment is whether somebody moved and whether they received a housing voucher as opposed to treatment for drug addiction or mental health services. So even for all of these three groups, I would say that each should receive, if we're just thinking about the science behind it in order to try to isolate the effect of moving in the receipt of housing provision, each group should receive the same extent of services on average in terms of drug addiction, mental health treatment, job training, things like that. Do I think that we as a society should invest in those forms of services? Absolutely. But just to clarify, so the treatment I mean here is just the difference between the groups moving in the receipt of the housing voucher. Thank you.

Yolanda Kent: My name is Yolanda Kent, I'm parole and probation officer in the city, and I've been a PO in other jurisdictions, and I guess some of the issues that I had were as far as this voucher is concerned, is...how would they prioritize that because the waiting list for vouchers is already in here, in D.C. it's about 10 to 20 years. So if you have people coming out, are you going to say, well, the people that don't have criminal records, they have children and other issues, are no longer a priority and offenders coming out are going to be the priority? How would that work? That's one of my questions.

Kirk: Did you have a second question?

Kent: I did. When you were talking about the cost, I was wondering about the cost of the jurisdictions as the other gentleman had mentioned, the people, the jurisdictions that would be receiving the offenders. I'm probably certain, I remember when my best friend worked in Raleigh in housing, and when that happened a lot of the Katrina people came here, a lot of people came there, and I remember us having back and forth conversations about how, I remember here, they were housed on New York Avenue, a lot of them, I mean in the hotels. And of course that's not a regular neighborhood, but I remember Fifth District and their staff were saying that a lot of their crime had risen and a lot of it was attributed to the Katrina residents coming here. I'm not saying all the ones that came were criminals, but the crime did have an impact on our police districts, so that was a question that I had, had you done any studies on impact of the receiving areas.

And the last thing, as me being a PO and dealing with the thought processes of the people that we're speaking of, the resistance factor is great. I mean you have people coming out and there are programs when they can go directly from prison to transitional housing, or to drug treatment, and they would tell you that they don't want to go, that there's no mandate for them to go, they're not going to go. So I would be curious as to how you come to the conclusion or hypothesis that they will be willing to just willy-nilly go somewhere else.

Kirk: Great questions. So let me start with the third. So resistance to...I think ethically we're talking about a small demonstration project or even a larger policy, we were talking about where people live. I'm not in favor of forcing somebody to go some place, giving them some option, but incentivizing their choices, and so people having a resistance to joining a certain program, well, then they don't have to volunteer. But getting back to that Urban Institute research that I was talking about, what's some of the...and perhaps Nancy can talk more to this, but from some of those qualitative interviews with offenders, they often heard that people wanted to go someplace else. Not all, I don't imagine all but some did and those individuals that do want to move to a new environment to have a better chance at successful life out of prison then are there ways, my question is, are there ways that we can provide that opportunity as a society?

Your second question about the impact on receiving areas. So let me just be clear about, so in my particular study, the way the cohorts were set up, these are individuals that came out of prison, onto parole, in Louisiana, a vast majority, especially in the post-Katrina period of individuals that came out of prison onto parole stayed in Louisiana, post-Katrina kind of makes sense that because if people left the state, including some people that were at that time under the supervision of the criminal justice system may have gone to Houston or Raleigh or wherever and committed crimes, those states may have been reluctant to do any kind of interstate compact with individuals coming out of prison after Katrina. So my study is based on the vast majority of individuals who came out of prison, post-Katrina, onto parole, stayed in the state and then basically compared them to pre-Katrina counterparts.

I think there are some studies out there that have looked at the impact of receiving areas, from Katrina, those folks that were displaced by Hurricane Katrina, but I think you're bringing up an important point or this notion of the impact on receiving areas is an important point to think about when I'm proposing something like this. And so one way to do this is to not just say, okay, we're going to give everybody that lives in D.C. a housing voucher to go live in Richmond. I think the way to do it is what I would call a two-way swap, where you also provide individuals that are from Richmond, a housing voucher to use in D.C. so that there's no net gain or loss in the total number of ex-offenders in a given area. Both areas have the same, but I would hypothesize that the likelihood of reincarceration would be lower in both areas for those people that moved. But doing this two-way swap kind of accounts for this notion that there may be some impact on receiving areas.

And your first point about housing, public housing waiting list. That's an excellent point. And so the question is, if we're going to do something like this, should we...individuals with criminal record jumped the line? And how does that fare to individuals that have been on a waiting list for a long, long time? There's lots of public housing waiting lists that are very long. I think some closed and there is a serious need of housing, so one thing to think about is greater access to or greater money spent on housing vouchers in general, not just for offenders. But to justify, I go back to this statement from HUD, how can we justify spending money on housing for ex-offenders when there's lots of non-offenders that are on the waiting list? Well, it may turn out to be a good thing for all of us. If we can enhance public safety, that's a benefit that we all get, so that's perhaps one way to justify it. The other way to justify is again, there is some cost savings, whether from less money spent on incarceration or less money spent on courts or victimization, things like that, there's some cost savings from reduced reincarceration that could be used to justify the expense you used to pay for housing vouchers, for individuals. And so I think it's an important question to think about how we would pay for it, but again, I think there's some public safety benefits from this idea of housing vouchers, and maybe some cost savings.

Barry Steffen: Barry Steffen from HUD. I noted that in your presentation, or you noted that the movers moved to neighborhoods that had lower fair market of rents than the stayers and keying in on that idea that housing costs are a barrier. Could it be that there needed to be some control for housing costs in your model, some kind of instrumental variable or something? That's part one.

Kirk: So let me answer that one. Yeah, so all those, so I had the five different comparisons in that one table that I put there. All of those variables are included as controlled variables in my analyses along with some variables that represent the characteristics of the criminal justice system, so ultimately the statistical model accounts for the fact that the movers or sustainers maybe living in areas with different fair market rents.

Steffen: Okay. Thank you. Part two, the housing choice voucher program is funded by federal taxpayers, but the benefits that you noted in your cost benefit calculation I think accrued to states that states pay for prisons and...so there will be local benefits to preventing recidivism. And so I'm wondering if there's an implication here that states and localities could do more, if they actually have an incentive that they don't recognize to do more to promote affordable housing in their jurisdictions dispersed across the jurisdictions so there are not concentrations of poverty and negative effects that way. But if this is really an unrecognized social cost of the way we've structured out land use.

Kirk: No, that's an excellent point and I'm, as I said before, one of the reasons I was excited to come to D.C. to talk about this, and learn from folks that know a lot more about housing and the policy implications, yeah, I mean, so federal funds distributed to the states and look at public housing authorities to them distribute to recipients, and how that process takes place I think it's important to think about especially in talking about an idea such as this. I think that's a question perhaps better answered by somebody that deals with those decisions.

Amy Solomon: Hi, Amy Solomon at OJP. I'm going to throw you a couple of softballs here after—

Kirk: [interposing] All right, let's have them.

Solomon: —but first one quick comment. On the issue of there not being jumping the line, there not being enough housing availability, one idea that's been discussed and with HUD partly lobbying these discussions is having people be able to come onto family leases. Leases that already exist but legally, and that's one way to get people back into public housing but without taking an apartment from someone else. So my questions, they're not really easy questions but they're simple questions. Data questions. One is the pre-Katrina movers, I'm wondering if they have a similar lower reincarceration rate than the post-Katrina movers?

Kirk: Yes. So I've actually done my analyses where I've looked at only pre-Katrina, pre- versus post-Katrina and just post-Katrina, in fact all... well, to answer your question, if we look at just pre-Katrina, there's a gap. If we look at only just post-Katrina, there's a gap. Even larger gap. This is, well, this is a sensitivity analysis that's present in the paper where I just look at post-Katrina, and the reason I do this sensitivity analysis is if there is some kind of Katrina shock then I'm not accounting for my statistical model, but it's also possible to do this analysis with just the post-Katrina folks and still set it up as a natural experiment and then the gap is even further, so the answer is yes, moving across whatever time period seems to have some benefit.

Solomon: Okay. That's good because obviously for applicability, so it's not the post, the Katrina part of it. The other question is, did you look at arrest, rearrested and not just reincarceration? Because I'm trying to think about how to take parole violations out of the mix for having higher reincarceration rates for those who are still under local parole supervision.

Kirk: No, I wasn't able to access arrest data. I would love to do that, if not, it's not possible in Louisiana to kind of link people to future arrest records, certainly possibly maybe in other states. And if somebody wants to do that type of research with me, let me know, I'd love to do it. But it's an important question, so it's something I've thought about, I just was unable to pull off.

Barbara Solt: I'm Barbara Solt, and I work primarily as a social worker in mental health aftercare of the institutionalization and so there is a lot of data around setting and type of setting for the mentally ill post-residential care. My question for you is did you, other than looking at location, did you look at type of setting, or placement or housing? Were these single individuals living singly, or had their families perhaps relocated and they went to live with their family because we have had a lot of discussion around the role of family. Did you look at that?

Kirk: Yeah, excellent question. The quick answer is no, I didn't have data specific to the type of setting where folks are residing. The best I was able to do was I mapped out, I geocoded service providers, all throughout the state of Louisiana, and I looked at proximity to different types of service providers relative to where somebody was residing, and the idea that if they have greater access to all types of services, that it would be negatively predictive of recidivism. So that's what I've done to partially get at your point, but I didn't have specific information on the characteristics of the setting where people were living.

John Lopit: I'm John Lopit, National Institute of Justice. Two questions, Dave. First I'm curious as to what the reaction of the findings were from the Louisiana Department of Corrections, particularly the parole officers. And second, just thinking about Nancy's question about families, I mean there's moving and there's moving, and so were you able in the sensitivity analysis at all to look at distance or proximity to New Orleans relative to larger moves, further moves, and what effect might that have had on the findings because I think it's an important issue in terms of pushing the analyses before we go to the grand experiment. And I guess related to that is, is there anything that you could call from the incarceration data, DOC data that would tell you something about the different family structures in the parolees?

Kirk: Thanks, John. In terms of the family contexts, so I know when somebody is married, there's the data on parenthood is okay, if not great, has some missing data, missing data problems but anyway, those individuals that are married are less likely to be reincarcerated. Parenthood, I don't want to speak to it because it has missing data but certainly the family context is evolving when individuals get out, individual's trying to repair relationships if they've been harmed in some way. I don't have good information on that. I'd love to because it seems certainly important, so what I have is kind of limited to this notion of marriage basically.

Distance, a great question. So in the findings I published back in 2009, I did subanalyses of distance and the more the better but with diminishing returns. So if somebody doesn't need to move to... from New Orleans to Alaska to benefit from separation, from their prior context, do they need to move to Baton Rouge or Shreveport? Well, that's probably better than moving next door. But I think the implication of your question and maybe this is where you're going, if we're going to talk about the demonstration project that I was mentioning, perhaps needs some precise information on what's the cutoff point, 100 feet doesn't work, half a mile may not work, but how many miles is it? And so in suggesting this idea, I've used kind of just a generic number, 25 miles, 50 miles, assuming that for that type of thing is enough separation from someone's old context. Could it be shorter? Maybe. But there's some interesting findings out of the moving to opportunity experiment, Rob Sampson has a paper that was published in American Journal of Sociology in 2008 where he looked at the distance that some of the MTO families moved relative to their previous location, and so a fair number of them in the Chicago side at least may have moved to lower poverty but they didn't move all that far from their old neighborhood and so in that sense perhaps the social networks of the youths in that program did not change all that much. They're still hanging out with the same individuals, still involved in the same routine activities that they were doing before their move and consequently maybe that's why the results are fairly mixed when it comes to problem behavior of males. And so when I'm talking about an idea of residential change I'd say let's make a significant change, let's make it a good amount of distance, not a mile or two, but you know, 25 or more.

Your question about reaction, Louisiana folks have been very receptive to these findings, are extremely interested and lowering their recidivism rates, and I will say that not necessarily in Louisiana, but as I've gone across the country talking about these results, I have received a little bit of skepticism from parole officers when I've talked about this idea of letting people moved and I'm certainly not saying that the services that a parole office provides to their parolees is useless, I mean the more services they could provide, the better. I mean that's an important aspect of successful prisoner re-entry. I'm just talking about maybe an additional thing that our criminal justice system can do in terms of maybe promoting this idea of residential change to a new area and a fresh start, and then provide some supervision, provide job training and so on and so forth. So when I've had some skepticism of this idea and in talking through it I've been able to clarify that, I think parole supervision is crucial, I'm just talking about tweaking it a bit to allow greater flexibility in residential patterns .

[Applause]

Kirk: Thank you.