Episode 07: Residential Mobilization with Kristin Perkins

Perkins, "Reconsidering residential mobility: Differential effects on child wellbeing by race and ethnicity"

More

Episode Summary: Past research has shown that moving to a better neighborhood can improve life outcomes for children and adults, at least under certain conditions. However, these studies do not examine how impacts differ by race and ethnicity, and they tend to focus only on a narrow slice of the population, such as public housing residents. How does moving impact different households in the real world, outside of an experimental setting? We welcome Kristin Perkins of Georgetown University to the podcast to talk about her work, and the difficult (but perhaps unsurprising) finding that moving is more harmful to the well-being of Black and Latino children than white children.

  • Abstract
    • Residential mobility is a common experience among Americans, especially children. Most previous research finds residential mobility has negative effects on children’s educational attainment, delinquency, substance abuse, and physical and mental health. Previous research, however, does not fully explore whether the effect of mobility differs by child race/ethnicity, in part because many of the samples used for these studies were majority white or exclusively non-white or disadvantaged. In addition, previous research rarely fully accounts for factors that predict selection into mobility and that may also be related to the outcome of interest. This study simultaneously addresses both of these limitations by estimating the effect of moving homes on children’s emotional and behavioral wellbeing using first difference models and a diverse longitudinal sample from the Project on Human Development in Chicago Neighborhoods. I find that, after controlling for a wide range of individual, caregiver, household and neighborhood characteristics, the effects of moving among African American and Latino children are significantly worse than among white children.

 

  • “Previous research indicates that there is a strong association between the number of moves a child makes and various health outcomes, including increased behavior problems, depression, and increased use of professional psychological help (Gilman et al., 2003, Jelleyman and Spencer, 2008) … Researchers find both positive (Sanbonmatsu et al., 2011) and negative (Jelleyman and Spencer, 2008) effects of residential mobility on adolescent mental health. This inconsistency is likely a result of the fact that the population and types of moves examined varies across studies. Experimental data from a sample of low-income children and adolescents who are moving from extremely disadvantaged neighborhoods to slightly more advantaged neighborhoods show that moving can be beneficial (Sanbonmatsu et al., 2011).”

 

  • “The association between residential mobility and mental health outcomes that previous studies find may be confounded with other characteristics of the child’s environment. And yet, previous research has largely overlooked the potential contributions of these characteristics to the association between residential mobility and mental health. Much of the previous research fails to account for underlying differences between mobile and non-mobile families, such as parent characteristics, family structure, and neighborhood context, which could account for some of the association between residential mobility and child outcomes.”

 

  • “[Previous] studies suggest that research on residential mobility and child mental health outcomes would benefit from an attempt to distinguish the effect of residential mobility from other sources of instability, like family structure changes and school mobility, that may occur at the same time or near to a residential move. One way to conceptualize the effect of residential mobility on child outcomes is that residential mobility serves as a confounder in research that attempts to isolate the effect of family structure changes (Fowler et al., 2014, McLanahan and Sandefur, 1994). Another conceptualization is that residential mobility moderates the impact of family structure changes. Here, residential mobility and family structure changes matter more or less depending on whether they occur jointly or separately (Verropoulou et al., 2002). Isolating the potentially disruptive effect of mobility matters practically because it speaks to the wisdom of using residential mobility as a policy tool to improve the neighborhood environments of children who live in disadvantaged neighborhoods.”

 

  • Previous studies “are unable to determine whether the negative effects of residential mobility are significantly different for children belonging to other racial/ethnic groups or living in more advantaged families and neighborhoods. The analyses I present accomplish this by using data from a representative population and considering specific subgroup differences.”

 

  • “This study uses three waves of data from the Longitudinal Cohort Study of the Project on Human Development in Chicago Neighborhoods (PHDCN). The Longitudinal Cohort Study began in 1995 and follows a sample of over 6000 youth in seven age cohorts – birth (age 0), 3, 6, 9, 12, 15 and 18. The second wave of data collection began in 1997, and the third wave began in 2000 … I draw my dependent variables from all three waves of the Longitudinal Cohort Study. I define child wellbeing in terms of behavioral and emotional problems, operationalized here as youths’ scores on two subscales of the CBCL (Achenbach, 1991), internalizing behavior and aggression.” (Internalizing behavior refers to things like depression and anxiety, while aggression is also referred to as “externalizing behavior.”)

 

  • “The coefficient for moving in Model 1 is significant and positive, suggesting that moving is associated with approximately a 0.11 standard deviation increase in internalizing behavior, controlling for other characteristics. Other variables are also significantly related to internalizing behavior: having a parent with more education compared to one with less than a high school diploma and being a homeowner compared to a renter are both associated with lower internalizing scores while having a parent who receives public assistance and neighborhood disadvantage are associated with higher internalizing scores holding all else constant.”

 

  • “Model 2 in Table 2 reports the results from a first difference model … The residential mobility coefficient, at 0.11, is similar in magnitude to the coefficient obtained in the pooled OLS model (Model 1). This result suggests that moving still has an independent effect on internalizing behavior once I account for changes in time-varying characteristics and all of the fixed characteristics of the child and his or her parents, household, and neighborhood environment.”

 

  • “In Model 3, the move coefficient for white children is not significant and is negative (-0.127), suggesting that among white children moving is not associated with higher internalizing behavior and may even reduce internalizing behavior. Among blacks, moving appears to increase internalizing behavior, as a move is associated with a significant [0.157] standard deviation increase in internalizing score holding constant other changes and fixed characteristics.” The coefficient is also significant and positive for Latino children who move, a 0.134 standard deviation increase.

 

  • “Model 1 in Table 3 predicts aggression score controlling for child, caregiver, household, and neighborhood characteristics in a pooled OLS model. Net of the control variables, moving is associated with a significant 0.18 standard deviation increase in aggressive behavior. This aligns with prior research that finds a detrimental effect of moving on child wellbeing.”

 

  • “Model 2 in Table 3 shows coefficients from a first difference model predicting changes in aggression with moving between waves and the same covariates from the first difference model predicting changes in internalizing behavior … The coefficient for moving is 0.09 [but significant], suggesting that moving is not strongly associated with aggression among children in the aggregate controlling for changes in other characteristics and fixed attributes of children, families and neighborhoods.”

 

  • “Among white children, in Model 3, moving does not appear to affect aggression scores, as the coefficient of −0.16 is not significant. Further, the negative coefficient suggests that if anything, moving may reduce aggression among white children, a different conclusion than in the aggregate model where moving predicts an increase in aggression. Model 4 shows that moving does not significantly predict aggression among black children: the coefficient is positive, but insignificant. Among Latino children, moving predicts a 0.19 standard deviation increase in aggression holding constant other time-varying and fixed characteristics. This positive effect of moving on aggression appears to be understated in the aggregate model, with its smaller coefficient, and the significant interaction term of 0.35 (p < 0.01) from the fully interacted model run on the full sample further demonstrates the significant difference in effects between whites and Latinos.”

 

  • “It is reasonable to think that other characteristics potentially correlated with race and ethnicity may be responsible for the significant effect of moving on internalizing behavior among African Americans and Latinos compared to whites. For example, non-white children who are disadvantaged in another dimension – being exposed to violence, having parents who lack social support, being renters instead of homeowners – may be driving the negative effect of moving. In a series of supplemental models (results not shown) I added these characteristics to the set of covariates included in the first differences models; the effect of moving on internalizing behavior among blacks is robust to the inclusion of caregiver social support, baseline homeownership, and exposure to violence, and changes in exposure to violence. I added the same characteristics, plus immigrant generation, to the models predicting aggression among Latinos and again failed to explain away the effect of moving on aggression among Latino children.”

 

  • “Overall, one-third of the movers in the sample moved to a similar neighborhood while 42 percent moved to a better neighborhood and 25 percent to a worse neighborhood. These percentages differed by race, with a higher share of whites moving to similar neighborhoods (54 percent) and a lower share of blacks moving to similar neighborhoods (18 percent). Blacks had the highest share of moves to better neighborhoods (52 percent) but also the highest share of moves to worse neighborhoods (30 percent). That blacks experienced more neighborhood change, on average, than whites and Latinos may help to explain why black children have significantly higher internalizing scores after a move compared to whites. Indeed, once I interact the category of neighborhood change experienced by black children with residential mobility, the coefficient for a move to a better neighborhood is negative while the coefficients for moves to similar or worse neighborhoods are positive, though all of the coefficients are imprecisely estimated and not significant (results available upon request).”

 

  • “My first difference models control for changes in caregiver relationship status by including differenced dummy variables representing whether the caregiver is married or single at each wave. These variables are not statistically significant, though that is not to say that caregiver relationship changes are inconsequential for child emotional and behavioral wellbeing. It may be the case that an indicator for caregiver relationship status is not a fine enough measure to capture instability at the household level that could affect children.”

Shane Phillips 0:06
Hi, there this is the UCLA Housing Voice podcast. Our goal at housing voice is to bring important housing research to a broader audience, and to talk through how that research can be applied to making our cities more affordable and more equitable. As always, I'm Shane Phillips and I manage the Housing Initiative for the UCLA Lewis Center for Regional Policy Studies. But this time Mike Lens is on vacation, and I have a special guest co-host who will be introducing in just a moment. If you enjoyed the podcast, we'd love to have you give it a rating or review, or share it with a friend. And be sure to subscribe on whatever podcast app you use. All right, let's get to our guest.

Our guest today is Dr. Kristin Perkins, Assistant Professor in the Department of Sociology at Georgetown University and a Graduate of Cornell, Berkeley and Harvard in that order, Kristin studies inequality and social stratification with a focus on children, families and neighborhoods. And her research examines the impacts of residential mobility on children's educational and health outcomes. Kristin is joining us to talk about her 2017 paper in the Journal Social Science research titled 'Reconsidering Residential Mobility: Differential Effects on Child Well-being by Race and Ethnicity', and to tease the results here at the beginning, I'm just going to read the last sentence of your abstract here, which just says, "I find that after controlling for a wide range of individual caregiver, household and neighborhood characteristics, the effects of moving among African American and Latino children are significantly worse than among white children", which I think fits into the gentrification and displacement literature in a really interesting and illuminating way. And I think will help inform sort of a broader discussion on the way that the US approaches housing policy, and that we've been talking about a lot. So welcome to the show, Kristin

Kristin Perkins 2:01
Great, thank you, Shane, thank you for inviting me to be part of this set of conversations. I'm really excited to be here today and talk about this research in this paper.

Shane Phillips 2:11
We are very glad to have you. And we are actually joined today also by a special guest co host. And that is Paavo Monkkonen, Associate Professor of Urban Planning at UCLA and our first guest. Hey,Paavo.

Paavo Monkkonen 2:23
Hey Shane. How's it going? Good to see you. Hey, Kristin, it's so great to see you. It's really cool to welcome you to the UCLA Housing Voice podcast. I guess we should do a full disclosure that we've known each other for over a decade since we...

Kristin Perkins 2:37
Yeah, for 15 years now. I did the maths, and we were at Berkeley together 15 years ago, and it's great that we've been able to like connect on our work for all this time

Paavo Monkkonen 2:49
And I went and so we walked the hallowed halls of Werster and we edited the Berkeley Planning Journal together. And I remember, I vaguely remembered Rexford Tugwell was the main thing. I remembered you had a good pun, in essay about Rexford Tugwell, and so I went back and I saw that you were very much ahead of the times, you were already doing stuff on limited equity cooperatives and the New Deal and these things that have become like very kind of prominent in the housing conversation these days. So it's cool to see how your research has developed. And are you still thinking about Rexford Tugwell I guess?

Kristin Perkins 3:22
I haven't thought about Rexford in quite a while. But um, as you do say I've been interested in housing and housing policy since I was an undergraduate and definitely our time together at Berkeley shaped my interest in housing policy and my time in California, and have kept up that interest in housing throughout my studies and work since then. So I'm excited to talk today about how that fits into sort of my broader research agenda too.

Shane Phillips 3:52
And I I'm revealing my ignorance here. I'm sure this is like an academic thing. But I've never heard the name Rexford Tugwell, but it sounds amazing. I have nothing else to say.

We'll put it in the... and this goes in the show notes.

Yes, yes. We're gonna have to dig that up. So back to the research, Kristin said your paper is about residential mobility, aka moving. And there's a pretty big literature on how moving to a new neighborhood affects the people who move both the adults and the children, and how the effect depends on whether they're moving to a quote unquote better or quote unquote, worse neighborhood. There's also some really interesting work on how outcomes differ for people who move from gentrifying neighborhoods, and those who stay in those neighborhoods. Before we get into the design of your study and what you found, maybe we can just start with you summarizing some of that previous work that your paper is building upon.

Kristin Perkins 4:46
Sure, yeah. So I set out to investigate this question about the association between residential mobility and children's outcomes, because of some of the consistencies that I found in prior research. So as I was starting this project, I was reading research on residential mobility from a housing and housing policies perspective with my background in urban planning and my interest in housing policy. But I was also starting a Sociology Ph.D program so I was reading research from family sociologists who study family processes and child outcomes. And so I see myself as an urban-sociologist, and a family demographer. And I started like straddling these two sub-fields with the project that we're talking about today on residential mobility. So I was reading research on the 'Moving to Opportunity' experiment (MTO), and I think, you know, some of your listeners will know about that. But just the quick one sentence is that this experiment gave residents of public housing projects, housing vouchers to move away from their distressed neighborhoods. And this big social experiment about moving families and kids to better neighborhoods hadn't realize the dramatic improvements in well-being even among children who moved to neighborhoods with more opportunity than the neighborhoods where their public housing was located. And at the same time that these early results from MTO, were being researched, families sociologists were writing about residential mobility as this disruption to children's family environments, and found that kids who moved had lower test scores and lower educational attainment, they tended to have worse physical and mental health and higher rates of delinquency and substance abuse than kids who didn't move. So I was putting these two literatures together, and I wondered if part of the reason we hadn't seen these big positive results that we are expecting from MTO was because the move itself was disruptive, even if the new neighborhood was better than the old one. And I'll just say, as an aside that, like, I started working on this project before Raj Chetty, reanalyzed The MTO data and found the children who moved at a younger age did realize big gains from living in a neighborhood with more opportunity. So I don't want your listeners to be like, "no, that's wrong, it's very different now". And this, you know, putting you back into time when I was starting reading for this project, right, but I write in the article, there were these studies of residential mobility using samples that were predominantly white and other studies based on samples that were majority or entirely non white. But I hadn't seen anything that was explicitly assessing whether this association was different for kids who were away compared with black and Latin x children. And so that's where I sort of left off with my reading before getting into it.

Shane Phillips 7:25
Right, and yeah, I mean, it was it was news to me that, you know, moving generally is is known to have these kind of negative consequences, I think that was just not something I was really aware of, beyond the the impact, you know, how the neighborhood you're moving to or from affects things, just the fact that moving itself has believed to have these negative consequences.

Paavo Monkkonen 7:49
And if I can just jump in there, like, I think it's such an interesting focus, because there is kind of the housing policy implications, but as well, kind of the family demography implications. And it's just such a difficult moment to focus on, because you have to worry about the quality of neighborhoods, which is something that the planners would worry about for a long time. But then, like you said, the instability of the move itself, but then the kind of things that prompt the move, and kind of disentangling kind of, you know, households composition changes, or job loss or some traumatic event that sometimes prompts a move from, you know, like moving to a better job that prompts a move. And then not only that, I was thinking about like the correlates of households that are going to be having these unstable situations in the family, because of kind of whatever structural societal forces pushing that. So it seems like those four, those are the kind of four issues that you have to disentangle. Are there more issues there, Kristen, that I'm missing?

Kristin Perkins 8:44
Those are a lot of the main issues. And definitely I tried to address some of those things and do the best I can in this article. And at some point, we may talk about some of the other work that I'm doing that's trying to get at this question from different angles. So yeah, there's only so much you can do in one article. And I can explain and your readers, if they're interested in the article can see, what I tried to do. But there's a lot that you have to disentangle and isolate to get to this question.

Shane Phillips 9:11
And this one article does quite a bit. And so I just want to get more of a picture of what residential mobility looks like, generally, what do people not know about residential mobility other than, you know, the fact that people move a lot. For example, what do moving or mobility rates look like for owners versus renters, different household compositions, you know, single parents versus married families, people of different ages, things like that, just as like a baseline.

Kristin Perkins 9:40
I'll say the start, like my research focuses on the United States. So what I know about residential mobility is limited to United States and there are likely very different trends and statistics elsewhere. But to set the stage for this particular article, I could say that Americans are moving a lot less now than they did in the second half of the 20th. century, but it's still approximately 13% of Americans move each year. And most of the moves that people make are local, they're within the same county, and a much smaller proportion of moves involved like moving to a different state or abroad. So you asked about like different groups of people, renters and low-income households move a lot more frequently than homeowners and high-income households move. And moving among renters and among low income households can often be triggered by housing costs increases that these households cannot afford, right, their landlords raised the rents and they just can't afford so they have to find a new place to move. On the flip side, though, mobility is lower among homeowners, because moving involves more transaction costs for them to sell their house, they have to find a new place to live, rent or buy. So owners on average have less flexibility than renters do to move for new job opportunities or family changes, some level of mobility is good, right? But moving a lot can be disruptive. So moving rates are highest among young adults in their 20s and 30s. And this makes sense because this group is finishing their education. They're moving for jobs, they're forming households, and families. And all of these things, as Paavo mentioned, can prompt a move. But the housing affordability crisis, and changes in the labor market can make it a lot harder to move. And we see that dual earner households, so household with two people working, have lower mobility rates than single earner or single adult households, because it's just harder to find two jobs in a new location than it is to find one new job. And so I'll say that the Joint Center for Housing Studies at Harvard where I was a postdoctoral fellow before I joined the faculty at Georgetown, recently published a research brief and blog posts on this. And so I'm drawing some of my statistics from that work. And I'd encourage your listeners to check it out if they're interested in these trends.

Shane Phillips 11:55
Great. So getting to your study, it looks at Children's emotional and behavioral well-being after moving. And I think there are two ways or at least two ways that your paper is building on previous work. The first is distinguishing between these two different aspects of moving that could affect children's outcomes, as we talked about, researchers have tended to focus on the effect of moving to a different neighborhood, often a better or worse one in terms of poverty rates, educational attainment, racial and ethnic segregation, things like that. What you're looking at is the impact of moving itself. And you're trying to tease out the impact of the move from whatever motivated, as we said, which might be something like a divorce of parents or some other family structure change. Why is that distinction important?

Kristin Perkins 12:43
Yeah, you're right, that I'm trying to isolate this independent, unique effect of residential mobility. And in my view, the most important reason for trying to isolate the effect of residential mobility from change in neighborhood context, from other changes in the household is in terms of policy implications, that knowing that moves may be harmful to one set or some set of children could potentially inform a cost-benefit analysis that policymakers conduct if they're designing programs meant to help families move to opportunity neighborhoods greatly, which family should be targeted, and when ideally, should they move so that the potential benefits of the move outweigh the potential costs.

Shane Phillips 13:27
Right.

Kristin Perkins 13:27
And secondly, like, I hope that the results from my article and other scholars work on residential mobility can help parents, teachers, social workers, any other adults, identify children who may need more support after a disruptive event like a move. So I'm arguing that moves matter, net of parent divorce and re-partnering and net of changes in parent employment, and by arguing that I'm hoping to call attention to an event that might be distracting or upsetting to a child on top of or separate from other changes that they may be experiencing,

Paavo Monkkonen 14:01
And not to delay the getting into the paper, I mean, we're already in it but I think I mean, I think this point about the policy implications of this is really important, because I do think there has been a shift from I mean, I don't I wasn't like working on housing policy professionally in the 80s, or whatever. But I think back then, there was much more of a consensus that mobility was the strategy, right, and we need to do moving to opportunities and all this stuff. And now there's a much more emphasis on stability as a policy goal and housing. And kind of this pendulum shift has has come back the other way. I don't know if you use see the same thing in the literature and the policy conversations.

Kristin Perkins 14:37
Yeah, I mean, I think it's an important point. It's also a really long standing debate. I mean, sure, people making decisions about mobility programs in the 80s and the 90s and designing MTO. And then other people coming back and saying, "Oh, no, we need like durable urban investment".

Paavo Monkkonen 14:52
Right.

Kristin Perkins 14:53
And I like I don't think it's really one or the other.

Shane Phillips 14:59
We talked about this with Kathy O'Reagan as well, I feel like anytime this comes up, it's always... there's this debate like "well, which should we do?" And at the end everyone's like, "well, obviously we should do both, why are we even fighting over this?" yeah.

Kristin Perkins 15:12
Yeah, I mean, like to jump ahead, you could say like, "okay, I find that moving is bad for kids so that means we should do place investment rather than mobility programs". But I think we still need to provide opportunities for mobility to high-resource neighborhoods. I mean, we see from Raj Chetty and his colleagues, it can be really great for kids, you know, depending on when they move and where they're moving to. But we can't move everyone and we shouldn't move everyone. So we also need at the same time to like increase resources available to families living in these low opportunity neighborhoods.

Shane Phillips 15:43
Yeah. So I said, there were two ways that that paper builds on previous work. And the other one is by looking at how the effects of residential mobility differ based on a child's race or ethnicity. Other studies have found that moving to a better neighborhood is more beneficial to younger kids than older kids. I think that's Chetty, and probably others, and also that it's more beneficial for boys than girls in general. But prior to your research, we didn't really know if it was better for white kids than black kids, for example. And that was partly about the data sets we were using for those earlier studies right?

Kristin Perkins 16:18
Yeah, so before my article, we did have evidence from different studies that moving was detrimental for white children, and separately for black children. But because the evidence was from different studies using different data, we couldn't tell whether the effects of moving were more or less negative among white children compared with Black, and Latinx children. So in addition, some of the prior studies use data from samples of disadvantaged children, so those who had come in contact with child protective services, for example, or were who were participating in a social program that serve disadvantaged youth. So theirs, results were pretty limited in terms of generalizability.

Shane Phillips 16:56
Okay, got it. So what did you find? I gave the abstract one-sentence answer, but in a little more detail, how did the impacts vary by race and ethnicity? And I guess we haven't actually talked about what you were measuring exactly, what we mean by child wellbeing. And so maybe just tell us what you were measuring and what you found in those results.

Kristin Perkins 17:17
Right, so well being is sort of a general umbrella term that can have a lot of things underneath it. And it can be used to encompass children's physical and mental health, whether they're growing and developing according to expectations, their progress in schools, and their social skills, right? But in this article, I'm specifically focused on emotional and behavioral well-being. And I'm measuring that using two scales of questions that the child's caregiver answers about the child's internalizing behavior and their aggression. So it's asking questions about the child's emotions, their feelings, their mental health, their interactions with others and their aggressive behavior. So that's the the measure that I have for my outcome. And in terms of what I find, starting with internalizing behavior, which means sort of anxiety and depression, I find that among black and Latinx children, moving is associated with more internalizing behavior. So black and Latinx kids demonstrate more anxious and depressive behavior after they move compared with before their move. And this controls for changes in household size, it controls for changes in parents' employment and marital status, homeownership, and neighborhood characteristics. So in my regression models, I'm trying to account for some of these things that Paavo mentioned that could be happening at the same time, right? I don't find any significant difference in internalizing behavior among white children before and after a move. And in fact, the association goes in the opposite direction. So it's possible that white children demonstrate less anxious and depressive behavior after they move. But I can't claim that definitively based on my estimates.

Shane Phillips 19:00
Okay.

Paavo Monkkonen 19:00
Can I just kind of ask you to talk a little bit more about the... so you're looking at kind of otherwise similar households, in terms of their composition, their incomes, the neighborhood they're moving from, and to but are you able to measure anything about an event that would precipitate the move? Or is that something that is potentially explaining these differences?

Kristin Perkins 19:26
So I'm actually using each kid as his or her own control, right? I'm using a first differences model or a fixed effects model. So I'm not comparing kids who moved to kids who didn't move. That's one of my critiques about the prior literature that those groups of kids can be pretty different on a lot of things. And so I chose this different modeling strategy, where I'm following kids over time, and I look at their scores on the scales after they've moved compared to before they've moved and look at the difference in their scores. And I do also control for whether their parents employment status changed, you know, before and after the move, whether their marital status changed, how big their household is, like whether there are more newer people or more people living with them. And I have a measure of neighborhood context that I look at both in before they're moving after their move. So I do try to get at some of these potential triggers for a move and account for them in the model.

Paavo Monkkonen 20:28
Yeah, and in your other work, you look at changes in household composition, beyond just kind of parents getting divorced or not. And so you've shown that's, that matters a lot. So it's good that you can control for that, how confident do you feel that these controls are capturing the most important, I mean you know, in terms of kind of noise or other omitted variables, do you have a sense of, based on your knowledge of this field, like, how much of the major things are you accounting for?

Kristin Perkins 20:54
Yeah, so certainly, a limitation of even the models I chose - fixed effects models is that they're only working if I am controlling for all of the time varying things that relate to both the predictor of moving and the outcome of these scales. So to the extent that, you know, I'm talking about this with causal language, you need to keep in mind the assumption like, "okay, did she control for everything that she could have to account for this?" You know, I think I've controlled for what I can, to account for this, the paper has a reports series of like supplemental models, where I add different variables in to see if that changes my results. And for the most part, it doesn't. And so I feel confident that my estimates are a little bit closer, at least to a causal estimate, than what we saw in the literature before. But of course, you know, we can't all do a 'Moving to Opportunity' experiment where we randomly assign moving to some groups and not moving to others, where we can have a lot of confidence that the groups have the same on all the other characteristics. I think it's really important, and something that I like about the field of sociology is that you can approach observational data with rigorous methods and answer questions that you can't answer using experiments.

Paavo Monkkonen 22:10
Yeah, no, I mean, I think that's right. And the question about other precipitating events, were there other ones that aren't in there that you think I mean, I guess the question then is, if you're accounting for all the most common precipitating events before move, why else would people be moving?

Kristin Perkins 22:28
Right, so one of the things I add in the supplemental model is exposure to violence. That's not in my main set of models. But it's something that the data that I use, the project on human development in Chicago neighborhoods, captures, and those data are great because they capture a whole lot of different things going on in the context in the neighborhoods that other surveys don't. I added that as you could think that like maybe a family wants to get away from a neighborhood or a place because they're witnessing a lot of violence. And that didn't change my results. And so certainly there are other things, maybe about change in health status, that I don't include that could be related. I also think, you know, we may talk at some point about how this work inspired by subsequent work. And that was, as Paavo mentioned, thinking more about the other types of changes in household composition that kids experience. And so that's, I get it at that a little bit by controlling for household size in this paper. But in other work, I moved more specifically into that as a focus. And so it's certainly the fact that I couldn't account for it in this paper made me think, "oh, let me go in that direction and see what I can learn".

Paavo Monkkonen 23:37
Right, because you can imagine, like, you know, an uncle that is supporting the family moves out, but at the same time, you have another kid, so the household size stays the same, but actually, the nature of it is changing a lot.

Shane Phillips 23:49
I'm going to appease Paavo's methodological and statistical interests here, I'm not going to ask about coefficients specifically here, but I do...

Paavo Monkkonen 24:03
You can allude to them

Shane Phillips 24:03
Yes, I absolutely can, and will

Paavo Monkkonen 24:06
This podcast will acknowledge the existence of coefficients, we might not talk about them in more detail

Kristin Perkins 24:11
And I might slip up and say the words

Shane Phillips 24:11
It's fine, it's fine. What we're interested in, though, is like the size of the effect. I just want to make sure all of our listeners are aware that something can be significant, meaning, you know, there's there's some effect happening, but it could be very small, or it could be very large. And so where where do things fall in this case?

Kristin Perkins 24:34
Yeah, so, I was talking about the results from the internalizing scale. And among black and Latinx children, moving increases their internalizing score by 0.16 , and 0.13 of a standard deviation right? And standard deviations of a scale are not a particularly intuitive measure, right? But these effect sizes are similar to the difference in internalizing behavior, between a child whose parent receives public assistance and assistance and one who doesn't. And it's bigger than the difference in internalizing behavior between a child whose parent owns their home versus renting their home, controlling for other characteristics. So in that sense, I think the effect of moving is meaningful, right? It's statistically significant. But I think it's also meaningful in terms of what it might relate to in terms of changes in behaviour.

Shane Phillips 25:27
Yeah, especially for something that it's a one-time thing. It's just these other things, household structure, and public assistance and owner versus renter, those are all-encompassing, in a way for just a single move to have this effect is really impressive in a bad way I guess.

Paavo Monkkonen 25:43
Yeah, on that note, I wanted to... I had a question later on about, you know, how long the effects last, or how we can think about that issue

Shane Phillips 25:50
Wcan get to that now.

Kristin Perkins 25:53
Yeah, so that's not something I look at explicitly in this analysis, because I had this strategy of needing to measure the scores at these different waves and measuring whether there's a whether there's a move between the wave or not. So I cite some other research in the paper about a compounding effect of more mobility, right that some researchers find a negative effect of moving, but it gets bigger, if there are more moves that a kid makes. And I, you know, would hypothesize that there might be some sort of decay in the effect of moving right? If you have this disruption, you get to a new place, but then you're stable, maybe you can sort of bounce back to where you were prior. But a lot of other things may go into that in terms of other stuff going on at home, and the neighborhood that you're in and the school that you may have moved to etc. So just like this paper, figuring out the answer to that question about how long the effect of a move last would require a lot of different decisions and measures and assumptions about what's going on.

Paavo Monkkonen 26:56
You need some Swedish administrative data for that. I don't think we mentioned, what's the time period between the waves of the data?

Kristin Perkins 27:05
A couple years, so yeah, the first wave is in the mid 90s, and the second wave is in the, you know, 99 ish, and the third wave is maybe 2002. So it's not that long. And that sounds like maybe a long time ago in 2021. But the kids who were in the survey are young adults now. So I think still very relevant to, you know, current policy to me.

Paavo Monkkonen 27:31
But that would suggest that you're picking up the more immediate effects of moving rather than seven to eight years after kind of thing.

Kristin Perkins 27:39
That's right, that's right. Okay. And then I'm looking at a pretty immediate outcome, right? It's the score on this internalizing scale. Like, one of the things I can say about the type of regression model I chose to run is that I have to have a measure of the outcome multiple times. And so I can't like look at high school graduation, or having a kid as a teenager, because those things only happen once, right? And so you're right, that I'm then sort of forced into looking at this more immediate result, whereas you could pick a different strategy and a different method to look at a longer-term outcome.

Shane Phillips 28:15
So getting into implications and policy here a little bit. You know, reducing involuntary displacement, especially in black and Latino neighborhoods seems like an obvious implication of this research. But do these findings also imply that even voluntary or semi-voluntary moves are harmful to children of color? I think that's what it's saying. And if so, you know, what do we do with that information, given that moving is just a part of life, and also, often and maybe most of the time, a positive change for the household making the move?

Kristin Perkins 28:51
Right, so my findings do show that the average effect of moving among black and Latinx children, controlling for all these other things going on in the household in the neighborhood at the same time, is negative, right? But that's an average effect. And some moves for some Black and Latinx kids are good moves that could lead to better outcomes, right? Like, as you say, if parents get a better job, or they're moving away from a dangerous situation, we might expect that to be positive for the kids. So you're right that moving is simply part of life. And I'm certainly not advocating for zero mobility as an implication of the project. But instead just understanding moves as a potentially disruptive event and adults who are caring for and teaching children shouldn't just be prepared to offer additional support to kids who move

Shane Phillips 29:38
Out of curiosity, what does that support look like? I'm assuming there's some research on that as well.

Kristin Perkins 29:44
Um, I don't know what the day to day would look like. You know, I can think that, you know, if you see a kid who's withdrawing from activities at school or being more aggressive and bullying, then you can think like, "oh, this could be because they are having all this stuff go on" right? And maybe there's a way to talk to them about that or talk to their parents about what other continuity is they could offer to help address what's going on at home and then help them out more in school.

Paavo Monkkonen 30:14
Yeah, it would be really interesting to do a similar study, I mean I don't know the quality of schooling in Chicago area, but you know, in a place known for good schools, because you could imagine that would have ameliorative effect. I guess I wanted to go back to just this question of so you've controlled for many features of the neighborhoods and many features of the households. And you found this disparate impact in Black and Latinx children, why do you think that that's happening?

Kristin Perkins 30:42
Right? So this, like, finding of a difference by race and ethnicity might be pretty unsatisfying, like, okay, so like, "what's going on here, like, how do you explain this?" We talked a little bit already about the supplemental regression models, I ran about exposure to violence and immigrant generation to see if any of these things which are different for white, black and Latin x children could explain. And I didn't see much there in terms of explaining the difference in internalizing behavior and aggression. But there's other potential explanation that I explored was how much the child's neighborhood changed as a result of the move. So you know, I controlled for that in my models, but then I also looked at if, you know, isolate the kids who made upward moves versus moved to similar neighborhoods versus moves that were downward in terms of neighborhood characteristic, what do I find? And I found some suggestive evidence that the change in neighborhood context might help explain some of these differences by race, that black children who moved to a better neighborhood didn't demonstrate more internalizing behavior. But black children who moved to a similar or worse neighborhood did. And so my sample size wasn't big enough to say definitively that like neighborhood context alone is what's explaining these differences. But I think it's a question that's worth exploring with different data to see if that is is potentially the explanation.

Shane Phillips 32:11
Yeah, this part of the paper about where different racial and ethnic groups households are moving to was really interesting. And there were pretty big differences in the types of moves; for white households, 54% moved to a similar neighborhood, whereas only 18% of black households move to a similar neighborhood. 52% of black households move to better neighborhoods, which was more than any other group, but also 30% move to worst neighborhoods, which was also more than any other groups so there's just a lot more volatility, I guess. And this, it does feel really important. You know, it indicates to me that the 'Moving to Opportunity' Program is sort of a vision for how things could be like these positive outcomes when people are able to move to better neighborhoods. But you know, I think for one thing, not everyone can move to a better neighborhood, it's physically impossible for everyone to move up. So you know, moving can do a lot of good under the right conditions. But in practice, I think those conditions aren't always present. And when they're not these moves can do more harm than good. Does that framing of this ring true to you?

Kristin Perkins 33:20
Yeah, certainly. And I should say that the statistics I report about the quality of moves, I'm not the first person to talk about that I cite a couple of other papers in my paper that show that there are really big differences in the type of moves made by race. But I definitely agree with your assessment that we know now a lot more about the results of the MTO experiment than we did when we started this project. And especially the benefits to children who moved to opportunity neighborhoods at younger ages. But as I write in the article, and as you say, this sample that I'm using is representative of children in Chicago in the 1990s and 2000s. And in this group, the average move didn't result in a substantial change in neighborhood context, and that most moves are not like MTO moves. And this work is based on a sample that's representative of a much broader population, than the housing mobility experiments will ever reach. So I think in terms of policy design, we do need both types of research. We need these experiments that can isolate causal effects specific to the group of people that are going to be served by the policies. But we also need observational studies that are executed carefully to shed light on how these social processes work in a broader population.

Shane Phillips 34:41
Yeah, yeah. I'm not sure that we actually said what we mean by good neighborhoods, better neighborhoods, worse neighborhoods. How are those defined? What does it mean for for a household to move from one home to a home in a similar neighborhood versus a better or worse neighborhood?

Kristin Perkins 34:57
Yeah. So some neighborhoods are measured just by their rate of poverty, right? And you could say, at least 20%, or 30% of the residents of this neighborhood have incomes below the poverty line, and so that neighborhood is characterized by concentrated poverty. What I did in this article is make an index, including a few other characteristics in terms of the education level of the residents of the neighborhood, and the number who are single female headed households with children and tried to draw the unemployment rate tried to draw on a few more things and make it a sort of a more multi dimensional measure of neighborhood context than simply the poverty rate. So, you know, this is necessarily condensing some of the richness of neighborhood context by putting it into a scale and then applying these boundaries that if you move this many points on a scale, you're better or worse, but I'm hopeful that it does sort of map on to the lived experience in a neighborhood in terms of your you know, if you're moving to a better neighborhood, then you're moving to a place where fewer of your neighbors have low income, more of them are employed and have higher education. And you know, maybe along with that comes better elementary school, or better opportunities for education, better opportunities for employment, etc. So it necessarily have to have something that stands in to represent neighborhood quality. But I've tried to collect a few different characteristics that represent different dimensions of it.

Shane Phillips 36:36
Ultimately, we are researchers, and we demand categories to input into our models,

Kristin Perkins 36:42
Right.

Paavo Monkkonen 36:43
So you mentioned the neighborhood change factors. And then we had alluded to it earlier, the household composition, and there's some unmeasured things happening there right? That could be different. Since this paper, you've researched the topic of household composition, I don't know if you want to give our listeners a few kind of the most interesting findings from that body of work as well.

Kristin Perkins 37:04
Sure. So after I finished this article on residential mobility, I wanted to take this more comprehensive look at the changes and disruptions that kids face at home. And so in my dissertation work and some of my subsequent work, I examined children's exposure to changes in household composition involving extended family and non relatives. So we know from some of the research that I'm citing in this residential mobility article that moves often happen as a result of changes in parents' relationship status. But what I found in the subsequent work is that parental divorce and re-partnering are only one type of the change in household composition that kids experience. So over the course of childhood, so 18 years, nearly 40% of kids experience an extended family member or a non-relative joining or leaving their household. That was a surprising statistic to me, it seems pretty big.

Shane Phillips 37:59
Yeah, that's really big

Kristin Perkins 38:00
And in an article that I published in demography in 2019, I show that experiencing these non-parent changes, makes an adolescent less likely to graduate from high school and less likely to enroll in college. And in work, that's currently under review, I show that experiencing these changes makes adolescents more likely to have a child themselves as a teenager. So these projects together, the residential mobility and the household composition, are pushing me to think about households as a context, separate from families and separate from neighborhoods that are important in their own right. But also important in combination with family and neighborhood characteristics and dynamics. So my research agenda includes exploring residential mobility, and changes in household composition, as experienced independently and together. super interesting. And so just to be clear, it's the change that has the impact rather than having extended family in your household right?

Yeah, that's a really important distinction and a question...

Paavo Monkkonen 39:00
I don't want to discriminate against extended family households.

Kristin Perkins 39:04
That's right. I'm certainly not saying that, like living with your grandparent is bad, right? Because there's evidence that it can be really good for kids. And not...

Paavo Monkkonen 39:12
Maybe not so good for the parents, though, am I right?

Kristin Perkins 39:16
That's right. That's right. And I'm not saying that living with like cousins and aunts and uncles is bad. It's the disruption, right? That you have people join your household, and you have to like figure out new roles and responsibilities. And maybe there's conflict over child rearing because the grandparents did it one way and the parents want to do it another way. And then that changes again, if those people leave, so you could get adjusted to something and then you have to readjust or maybe your house is crowded because you have more people living with you. And you know, something that I want to explore is of course moving is attached to this too. So is it that it's bad because you have to move to your aunt and uncle's house and leave where you were. So a lot still to figure out.

Shane Phillips 39:58
Yeah, a lot of this is, I don't know, it's unintuitive to me in a strange way where it seems like the ongoing change would be the real impactful thing. And to just have this one time disruption, have such a big impact in these different waves is really, it's surprising to me.

Kristin Perkins 40:16
Yeah, I mean, some of where I was going, certainly during my postdoc at the Joint Center is to think about whether subsidized housing or housing assistance can help with this stability question. Certainly, in terms of moving but also in terms of the people coming in and out of your house, right? That if you, if you know, you're a kid and your parents get a housing voucher, you might make an initial move, but then hopefully, you have a more stable place to live for a while, and you're not moving again, when your landlord raises your rent, or something like that. And so that, okay, you know, I get a lot of questions about like, what are the policy implications of your grandparents or your aunts and uncles coming to live with you. I'm not gonna say like, that can't happen, that's bad. But maybe the people are coming to live with you, or you're coming to live with them, because nobody has affordable housing, and you need to double up to survive. And so if there was more stock and more affordable housing, then maybe we could reduce the changes that kids are experiencing?

Paavo Monkkonen 41:17
And what would you say that I mean, there's some research in Europe, where there's a lot more stable housing for many people, about the unemployment impacts of the inability to move and kind of being trapped in subsidized housing that you can't leave, because market rate housing is so expensive, or there aren't other housing opportunities in subsidized housing.

Kristin Perkins 41:40
Yeah, I guess I would say, make it all portable, and have enough supply everywhere that you could just, you know, hop to a different unit if you've got a better opportunity. I mean, Shane, you've written about this a little bit right with your thinking about shared equity or limited equity, that if there were these developments that were available in enough places, or enough of them, then you wouldn't be locked in?

Paavo Monkkonen 42:02
Yeah, I know, my reaction has always been "well, we're not there yet" right? So we can worry about that when we get to 30% of our housing as social housing?

Shane Phillips 42:12
Yeah, that's a problem for another day.

Paavo Monkkonen 42:16
Yeah, I was. I mean, if Kristin, are there other researchers that have kind of followed a similar approach since this paper was published in other cities? I mean, I think I was thinking about the differences of housing markets regionally and kind of how important they are for, you know, the ability to stay in your home or the need to move or the ability to move. And so I wonder if you've seen other people doing similar studies elsewhere?

Kristin Perkins 42:39
Yeah, that's a good point. And like I think, one you made in the episode I listened to this morning, about your work on LA, and whether you could apply that to other cities. And certainly, you know, I'm making an argument that the sample I use in this paper is more representative than subsidized housing population, but it's still one city, Chicago, and it's in the 1990s, in the 2000s. And so, I do think it'd be worthwhile to look in other specific cities and housing markets, but also if there is a way to, you know, look at a national sample and see if some of these same patterns are showing up. I don't know, of particular other studies in other cities, but I'm sure they're out there, in terms of people who are thinking about this in different markets, because certainly in San Francisco, or Seattle or Boston, it might be a different story than in Chicago, which has a slightly, you know, less crazy housing market. And then like, you know, the Midwest, other cities, smaller cities like Dayton and Columbus that have much more stock.

Paavo Monkkonen 43:45
And yeah, I'm also thinking about kind of the segregation context and the kind of the localized version of white supremacy and racism in a certain place and how that manifests, you know, because I mean, like, my impression is that a lot of that is percolating down into these coefficients, you know, on the disparate impacts, but you know, you could imagine, Houston has a very different context. And, you know, both in terms of the housing market, but just in terms of the social life of the city, and so you could imagine quite different impacts there.

Kristin Perkins 44:15
Exactly. I just heard on the radio this morning, there's a new report from some Center at Berkeley, whose name I didn't recognize, it's about segregation.

Yeah, Othering and Belonging

Exactly, about segregation in different metro areas. And, you know, here in DC, the radio was talking about DC is the 15th most segregated metro... something in the city. But I definitely think, you know, we're talking about these neighborhood moves, that it could be one implication of, you know, a systemic racism hypothesis, right, to the extent that historical and contemporary discrimination and racism are limiting the residential options for BIPOC families, like having financial and geographic constraints on where they can live certainly means they can make fewer upward moves in terms of neighborhood quality. And that could vary a lot by metro area.

Paavo Monkkonen 45:08
Yeah. Yeah, I don't I should know this. But I mean, it'd be interesting to just those same statistics on just regular moves. And are you going to a better or worse neighborhood or a similar that would be interesting to look at in different cities? And I was gonna ask you, did you move a lot as a kid?

Kristin Perkins 45:24
No, I didn't. Like, maybe this is me search, or maybe it isn't but I moved once when I was two and a half, and I have no recollection of, you know, the first place so...

Shane Phillips 45:37
But there is a little I mean, did you move a lot Shane?

I moved, I don't know, probably five times, seven times as a kid... quite a few times. I also didn't graduate high school so Anec-data right there.

Paavo Monkkonen 45:54
That's interesting

Shane Phillips 45:55
There was some disruption. Okay, on that cheery note, I'm gonna close this out. Thank you so much, Dr. Perkins for joining us. It was really great to talk to you.

Kristin Perkins 46:03
Thank you for having me. It was fun to talk about this work with both of you.

Paavo Monkkonen 46:06
Thanks. Bye. Later.

Shane Phillips 46:14
That's it for this one. Thanks, once again to Dr. Perkins for being so generous with your time. We got papers and articles mentioned during the interview in the show notes. And we always include our own summary of the key takeaways from the main paper and a very low quality computer generated transcript if you're into that kind of thing. You can find all that on our website at lewis@ucla.edu. But UCLA Lewis center is on Facebook and Twitter and I'm on Twitter @ShaneDPhillips, and Paavo is there @elpaavo that's E-L-P-A-A-V-O in the words of Ariana Grande: Thank you, Next episode in two weeks.

Transcribed by https://otter.ai

About the Guest Speaker(s)

Kristin Perkins

Kristin L. Perkins is an assistant professor in the Department of Sociology at Georgetown University. Her research examines the consequences of residential mobility and changes in household composition for children's educational outcomes, the neighborhood contexts of child and adolescent development, and the consequences of housing policies for individuals and neighborhoods.
Learn More

Suggested Episodes