Ep 75: Segregating the Built Environment with Ann Owens

We often talk about residential segregation by race or income, but we rarely explore it in the literal sense — as in segregation of residences: of one kind of housing from another. Ann Owens joins to discuss her research on how segregation manifests itself in our built environment in cities and neighborhoods across the U.S.

Transcript

Shane Phillips  0:05  
Hello! This is the UCLA Housing Voice podcast, and I'm your host, Shane Phillips. Our guest this week is Ann Owens, and our topic is one that is surprisingly overlooked, but very, very important. We all spend a lot of time talking about residential segregation, usually by race or income. But we rarely talk about residential segregation in the literal sense, as in segregation of residences — of one kind of housing from another. This is a little strange, given what we know about the relative costs of say, detached single-family housing versus small multifamily housing, or about the barriers to entry for homeownership compared to renting. Those costs and barriers play a big role in determining who can live in a neighborhood, with different effects on people of different races and incomes, among other things. And yet these things are often taken for granted, despite the fact that the mix and distribution of single-family versus multifamily and owner-occupied versus renter-occupied housing varies so much from city to city, and from metro area to metro area. Ann's study tackles this important subject, and it's the topic of our conversation with her here. So enjoy! 

The Housing Voice podcast is a production of the UCLA Lewis Center for Regional Policy Studies, with production support from Claudia Bustamante, Jason Sutedja, and Gavin Carlson. My email is shanephillips@ucla.edu, and you should tell your friend, colleague, or mother about our show. But with that, let's get to our conversation with Ann Owens.

Ann Owens is Professor of Sociology and Public Policy at the University of Southern California, and she's joining us today to talk about how segregation of building types like single-family and multifamily and tenures, which is to say owners and renters, contributes to segregation by income. Ann, thanks for joining us, and welcome to the Housing Voice podcast.

Ann Owens  2:17  
Thank you so much for the invitation. I'm really excited to be here. 

Shane Phillips  2:20  
And Paavo is my co-host today. Hi Paavo.

Paavo Monkkonen  2:23  
Hey Shane and hi Anne, it's great to see you. I'm very excited you're here to talk about your research, which I've followed for a long time. And I want to make sure that we don't miss mentioning your website, the Segregation Index. I don't know if you're planning on mentioning it later, but it's super cool.

Ann Owens  2:43  
I am and we've rebranded. We're now the Segregation Explorer.

Paavo Monkkonen  2:46  
Awesome, excellent. Okay. Well, we'll talk about creating online resources to look at segregation later.

Shane Phillips  2:54  
Yeah, we might have something we can learn from you on that with our own projects here. All right, so we always start the interview by asking our guests for a tour of a city they know and want to share a few highlights from with our audience. So Ann, where did you pick?

Ann Owens  3:08  
Well, I grew up in Chicago, but I've lived in downtown Los Angeles for 11 years, and I've lived there without a car for 11 years and so I thought I would take the listeners for my morning walk. We'll start at Grand Central Market, which is a food hall in downtown LA. If you're hungry before the walk, we'll stop for a donut at Donut Man. We're going to walk north to Grand Park, which opened about a dozen years ago in downtown LA. There's a beautiful new courthouse that I like to walk through. If you would like a view of the city you can go up to the observation deck at City Hall for free. And then you have a choice: We can walk north to Chinatown and the State Historic Park, east to Little Tokyo and the Arts District, and further north and west to Echo Park or Silver Lake. But I'm going to keep it in downtown, so let's meander through Grand Park to the music center. We'll walk around the Department of Water and Public Works or whatever DWP stands for — a building which has finally restored their water feature —

Shane Phillips  4:05  
Water and Power.

Ann Owens  4:06  
Water and Power, thank you. They always have beautiful buildings. Walk past the Disney Concert Hall, go upstairs to check out the garden in the fountain. Go to the Broad or MOCA for free art, and then cut through California Plaza, think about how downtown used to be just beautiful Victorian homes, and now it's California Plaza, which is also cool. But we can honor the history by taking Angel's Flight, the world's shortest railroad, back down the hill to Grand Central Market in time for lunch. So that would be my my downtown tour. 

Shane Phillips  4:36  
We got rid of the most rail tracks in the world, the longest rail essentially, but we held on to the smallest rail at least.

Ann Owens  4:44  
We moved it a few blocks, but we kept it, and actually now the A Line is the longest light rail in the world.

Shane Phillips  4:49  
Oh, yeah. That actually sounds about right from Azusa to Long Beach. That's a it's a long trip. I didn't know you could go through the federal courthouse like just kind of wandering around. I'm assuming that's the one You're referring to you

Ann Owens  5:01  
can go through the grounds I'm buddies with all the security guards so they might let me but you could you can go through the grant that sort of little maze and walk through the plaza if you want.

Shane Phillips  5:09  
Yeah, there's there's multiple courthouses downtown, but that one's the Justice Cube, as I call it. 

Paavo Monkkonen  5:14  
Yeah, I was waiting till we were gonna wave to the people parking for jury duty.

Ann Owens  5:18  
I do a lot of directing to jury duty in the morning. 

Paavo Monkkonen  5:21  
People lost at 6am are probably looking for parking for jury duty.

Shane Phillips  5:25  
So this week, we are covering a study authored by an that was published in Sociological Science in 2019. It's titled "Building inequality: Housing segregation and income segregation." Most studies of residential segregation, at least that I'm aware of focus on racial and ethnic segregation, or income segregation, or both, but these types of segregation don't occur in a vacuum. Quoting from the article, "The type and affordability of housing available in a given neighborhood is a primary factor shaping households' desire and ability to live there." And so she explores four additional measures of segregation based on housing type and cost. These are segregation between single family and multifamily housing between owner and renter occupied housing between communities with higher and lower rents and those with higher and lower home values, and finds that people are more segregated by income in places with greater degrees of housing type and cost segregation, and that this is true at the neighborhood level, as well as between cities within a metro area. Quoting again from the article, as long as housing opportunities remain segregated, income segregation will change, little unquote. So efforts to desegregate our communities will require changes that include creating more rental options in predominantly owner occupied neighborhoods, and more multifamily options in predominantly single family neighborhoods. Among other things. If you're familiar with the Lewis Center's work on fair housing, and increasing access to higher resource more advantaged historically and often presently exclusionary neighborhoods, then you'll understand why we wanted to bring Anne onto the show for this conversation, we will start as we usually do by establishing a bit of context. And please add anything you would like to that summary I just gave, but tell us what motivated you to study this topic and help our listeners understand why we should be concerned about income based segregation. And just to be clear, these are patterns of living where people with similar incomes tend to be grouped together in the same communities. So if I'm rich, then I'm less likely to have many poor or even middle class neighbors. And if I'm poor, then I'm unlikely to have many middle class or rich neighbors. What kinds of outcomes do we observe for people who live in income segregated communities? 

Ann Owens  7:43  
Yeah, well, I've been studying racial and economic segregation for a long time and research on residential segregation tends to have identified a few big explanations. So our neighborhoods are segregated because of economic inequality, because of discrimination because of people's preferences and information networks. And these are all important. But all of this plays out against the backdrop of housing and the housing conditions and the housing markets vary from place to place. And so I felt like, we need evidence on kind of that other piece of the puzzle that housing options that people bring their their resources and their preferences and their actions to and in this literature, people often will say, housing segregation to refer to the segregation of people by race or by income, and it made me realize that like, we actually don't measure housing segregation, per se, and so I wanted to do that, you know, obviously, studies have recognized the importance of housing before often using zoning data, which is, you know, increasingly improving, but it historically difficult to get zoning data at a granular level. So I just wanted to say, well, what is the actual segregation of housing look like, which is related to zoning and other policies, but but lots of other things as well, that I'm sure we'll get into, as far as you know, why segregation at all? Why, you know, why are we concerned about this? I think we have pretty good evidence at this point that both racial and economic segregation are associated with with unequal opportunities and outcomes, right. There's a long line of research in neighborhood effects studies and and macro level segregation studies, that shows that where you grow up matters and growing up in socially and economically disadvantaged places can have negative impacts on people's economic educational, health, criminal justice, you know, other outcomes. Neighborhoods, like families, and like other social contexts provide political, institutional and social resources for, you know, research by Raj Chetty, which has been you know, widely cited and viewed. For example, He has data on nearly every American and where they've grown up and his work shows that neighborhood segregation is one of the strongest predictors of unequal outcomes in adulthood. So I think there's a pretty good evidence base that illustrates that name. provides structure opportunity. And that's why I'm interested in segregation. 

Paavo Monkkonen  10:04  
And how many people have said to you, wow, why didn't I do this paper before? Because it's I mean, it's such a great, important piece of research. But it's like fairly one of these things that's fairly obvious once somebody else does it. So I wonder if you if you get that a lot. And then I just wanted to ask you if maybe we could talk a little bit about your work on school segregation, which we might not get into that much. But you've done a lot of this. And I think it's probably one of the most important parts. Well, one of the several core parts of why residential segregation is such a problem and completely under studied by urban planners especially. 

Ann Owens  10:37  
Yeah, well, when I was working on the paper, at first, I was like, well, obviously, someone has done this. And I, you know, looked around and and as I said, you know, there's a lot on zoning, there's other literature on subsidized housing and segregation that I've contributed to, but it seemed like just housing writ large, you know, it's mentioned in a few places, but just kind of measuring it empirically in the way I do, I think, had not been done. And so it's actually a, you know, it's a paper that I it's one of my one of my favorites, you don't always love your work. But I do love this paper. And I think it is because, you know, I'm hopefully introducing a new set of tools that people can continue to use and track and really just get down into the nitty gritty of there's so many more questions to ask on this, you know, I look at the relationship between housing segregation and income segregation, but there's so much more to do. And so for interested scholars, I would encourage them to pick up some of these measures. As far as the relationship between neighborhood segregation and school segregation. Yes, that is, I think, extremely important. I always think about neighborhood and school segregation as relating to each other in a cyclical way. So I think often the conventional wisdom is, well, schools are segregated because neighborhoods are segregated. And, you know, we have often student assignment policies that kids go to school based on where they live. And that's certainly true. But I think what my work tries to do is show well, but these things are all a feedback loop. Because when schools are segregated, that's going to shape where parents want to live. And they'll, you know, I think that schools and the structure of schooling and school options is a really important factor in residential patterns and should be a bigger point of interest in urban policy. I do think increasingly, urban housing and schooling policy are coming together in ways that I think are critical for actually addressing these problems. I will also add that I have a recent annual review of sociology article with my collaborator, Peter rich that goes through empirically and theoretically, the connections between Neighborhoods and Schools. So anyone interested c  an check that out.

Shane Phillips  12:32  
Yeah we can put that in our show notes as well. 

Ann Owens  12:35  
I'm just gonna plug, so you can cut any plugs you want? 

Paavo Monkkonen  12:38  
No, that's great. 

Shane Phillips  12:39  
Well, kind of to Paavo's point about this research not having been done. And it being fairly straightforward, but obviously really important, something I've seen folks drawing attention to in the last few years, which is well after you published this study, and it's just looking at the average household incomes of people based on the housing that they live in. And I just pulled these numbers for Los Angeles using American Community Survey data for 2022. And what it shows is that the average household income of people living in single family detached houses in the city is $140,000 a year in single family attached housing, it's $100,000 a year in duplexes, it's 83,000. And then for everything else is basically somewhere between 69,070 $3,000 per year. And so the incomes of people living in single family detached are twice as high, on average. So there's a really, really big difference here. And you see that duplicated all over the country. And I should also say that if you look at just people who have moved into their homes in the past 234 years, these numbers are even larger, the average income for people in detached homes is more like $170,000. If you moved in, if you threw will there be bigger? Well, there's you tend to see, you know, incomes for homeowners are roughly double renters, but the wealth is 1020 times greater because home equity is a huge source of many, if not most people's wealth. Another kind of clarifying context setting question here that I think is important, is what you say to people who think that income segregation is just the natural way of things, or maybe even the way it should be. I'm certain you've heard people say things like, I worked hard to get where I'm at. And usually that place is higher income, mostly single family, often disproportionately white community. And it is not fair to now open up my neighborhood by adding multifamily or rental housing. And they might also say that if you allow those changes, then the neighborhood won't be as nice, whether because of increased traffic or crime or whatever else they imagine. And so by opening up the neighborhood, you will destroy what made it so appealing and maybe what made it what gave it its positive associations with economic opportunity with education and so forth. How do you respond to that concern or to that claim?

Ann Owens  14:56  
Well, I'm a sociologist, so nothing is natural, right? Everything is a risk. All kinds of choices and the people and the institutions that that get to make those choices. And the neighborhoods that people are proud to have been able to access were often created via exclusionary policies that just made it impossible for some people to reach them. And I think, you know, if we think about intergenerational processes that often enable a lot of wealth building homeownership, even access to education, you know, it wasn't that many generations ago that entire swaths of our American population, were just not able to access these things. So I'm sympathetic to this concern. I'm someone I grew up in, my parents owned our home, I grew up in a predominantly white suburb of Chicago, you know, I understand these concerns, especially for homeowners, because we have tie, you know, we've made homeownership, our main wealth building tool in this country. And so when you have that system, the stakes feel super high. And it's just sort of expected that your your home value is going to continue to appreciate but obviously, nothing is guaranteed to stay the same. So I mean, I guess I would say, well, most people are working hard, right? I think that the National Low Income Housing Coalition often has a stat where I believe there isn't a county in this country where minimum wage would enable you to be able to afford housing, right. And so most people are working hard, right. But if you're born on third base, you didn't hit a triple, it's a lot easier to get home. And I think most people don't see the structural advantages they've had access to whether that's in low levels of pollution, safety, high quality schools, etc. So I guess I would say, ultimately, I think equalizing neighborhoods is beneficial for everyone, even from a purely instrumental point of view, you know, allowing people access to higher quality places can put lower strains on the social safety nets. As a country, it's better for educated workforce. So even if you care less about the kind of, it's the right thing to do, you can you can make sort of an instrumental argument for it as well.

Shane Phillips  16:51  
Yeah, and I do think people who make these kinds of arguments, and you know, I share the sympathy that you mentioned, but I think people often imagine that someone who did what they did, could earn their way into their neighborhood. But that's often not the case, someone who bought into a neighborhood 20 years ago, that has these characteristics, that is exclusionary, that is not adding any more housing, often, if that person had been more than 20 years later and had the same trajectory in life, they would not be able to move into that neighborhood. And so it's actually in some meaningful way, reducing opportunity over time, which is, we should be increasing it, but at the very least, not making it worse. And it seems to be what's happening in many neighborhoods. And that's

Ann Owens  17:33  
certainly I think, the conversation in California where many homeowners who bought their homes in the 70s 80s 90s now have children who are interested in buying their own homes, and they've done everything they can to give their children you know, a good education, whatever it might be even assistance with a downpayment. And sometimes their children can't even buy a home. So maybe that reality will help people realize that it's a systemic issue we need to address. 

Paavo Monkkonen  17:56  
All right, so let's switch a little bit to some research methods, the most exciting segment of the housing Boise podcast, I want to talk about measuring segregation, and maybe can you talk a little bit how you do it housing type cost, and then in terms of people by income and race, so I guess, to give our listeners a sense of, you know, what is a low score? What is a high score? And how different are they in the real world?

Ann Owens  18:19  
Yes, I would love to talk about measuring aggregation. So it is much more common right to measure the segregation of people by race or by income. And so you know, perhaps most commonly, there's a lot of research on the segregation between, for example, white and black people. And the idea is we want to determine the degree to which these two groups share neighborhoods or live in separate neighborhoods. And so I sort of brought that analytical approach to housing and I wanted to understand to what degree housing units of different types so I look at renter versus owner occupied units, and single family homes versus multifamily buildings, share neighborhoods or locate in separate neighborhoods. And so in segregation world, we can think about two different types of measures. One class is typically called exposure measures. The other are called evenness measures. And so I use an evenness measure here. And the basic idea is to say, well, how evenly are housing units with different characteristics distributed across neighborhoods in a metro area? And so essentially, what the measure is doing is saying, Okay, what does the composition let's say for renter and owner occupied look like in the metro area? And then how well do the census tracks which is how I measure neighborhoods, how well does the housing composition of the housing stock, and each track kind of match that overall? So in other words, if all tracks looked exactly like what the Metro composition was, we would say that that would be you know, no segregation, it would score a zero on this index. If some tracks are 100% owner occupied and others are 100% rental occupied units. That would be complete segregation. There would be no neighborhoods that would have a mix of housing stock and so that it'd be a one on a zero to one scale to concretize that a little bit, we can use those exposure measures I talked about and basically compare, well, what does a typical rental units neighborhood look like versus what the overall Metro looks like. So I looked at the 100 largest metropolitan areas in my study, and in 2014, the average large metropolitan area was 37% rental units. So if there were no segregation, every neighborhood should be 37% rental units. But rental units clustered together. And so the average rental unit was located in a neighborhood that was about 48% rentals, they're sort of an 11 point higher rate than what we'd expect with no segregation. So those are the sorts of tools we can use to measure segregation and the aggregate index, you know, the zero to one score is kind of accounting for that departure in neighborhoods from the metro rate. When we go into the segregation measures by housing cost, which I also do by rents and by home values, and then also residential segregation by income, we use a slightly different measure that just accounts for the fact that now we have a bunch of ranked categories rather than just two. And if all of this sounds confusing, you don't have to estimate these measures yourself. Because in May, I launched the segregation explorer with my collaborator, Sean Riordan, at Stanford. Right now, our we have a interactive map and data tool, and we also have data downloads for researchers and data enthusiasts. Right now the data are available for school segregation, but neighborhood segregation are coming online shortly. So our dataset Currently has five different types of segregation measures at geographies as smallest the school district up to the nation. So you can check out sec explorer.org. And keep an eye on what we're doing. But the data are coming for neighborhood segregation soon.

Paavo Monkkonen  21:53  
That's super exciting. Yeah, I've followed the work of you and your co author Shawn Rhoden for a while, and I guess maybe I'll just plug another segregation website, what you want to check out an international effort to compare segregation internationally, it's called Urban structure.sites.luskin.ucla.edu.

Ann Owens  22:16  
You gotta buy you got to buy a shorter domain.

Paavo Monkkonen  22:20  
So this is an effort that my PhD student, Andre Comandon, and mostly did all the work on but we worked on for a while. And it was using some of these same measures to compare segregation in cities, I think a dozen countries, it's not as fully fleshed out as yours is. So maybe check out yours first. And then maybe if you want to see Sydney, Australia, check out that one. Cool, but I did want to stay on methodologies for just one second. So one of the innovations that Sean Riordan brought to the world of sociological segregation studies, I'll just note that it's interesting in this academic world, like the geographers really like their techniques, and when I tried to publish a paper using what Reardon had developed in a geography journal, they got like, there was some reviewer got extremely angry and said, No, the geography approaches to segregation measurement. But one of the innovations that readin brought that I really like is the multi scalar approach. So maybe you can just explain, like maybe the checkerboard analogy is, I think, quite useful in explaining, you know, measuring segregation at a neighborhood or a larger geographic scale. Yeah,

Ann Owens  23:23  
so there's a few different ways that Shawn has addressed and others to have addressed scale in the work because most of our measures of segregation actually end up being pretty a spatial, you don't necessarily know, you know, if segregation is at a certain level between neighborhoods, you don't know if that's because, you know, if you can envision, let's say, a box with six smaller boxes in it representing a city with six neighborhoods, you could have complete segregation, and all of the boxes on one side of town or on one side of the the larger figure could be, you know, 100%, black versus 100%. white on the other side, for example, or it could be, as Pablo mentioned, more of a checkerboard where they're sort of alternating neighborhoods as 100%, black or white, and the aggregate measure would not distinguish between those two configurations. And so, you know, there's a few different ways to think about this, we tend to rely on census tract data. Shawn and others have done work saying, Well, what if you define neighborhoods based on more of a radius, sort of a geographic distance instead of tracks, um, that actually hasn't picked up traction as much as I would have thought? I'm surprised because it's extremely useful. The technological advance, Shawn was doing Sean and his collaborators at Penn State and at Cornell and other places were doing this work 1015 years ago, and we've sort of just reverted back in mind. I'm totally guilty. We've reverted back to census tracts. So that's one consideration of like, what is the right neighborhood unit? And it also might vary depending on you know, if you're interested in the segregation of housing or people by race or by income, so that's, I think, a consideration. And then the other thing I think that has started to get more attention is to think about of segregation at various scales and to be able to compare those scales to one another. So one thing I did in this study is to look within metropolitan areas, both at segregation between census tracts between neighborhoods, but also between what the census calls places which are like municipalities, towns, cities, you know, whatever is sort of a jurisdiction. And so for example, you could think about, okay, just segregation, whether it's by housing units, or race or income, is it happening between neighborhoods in the Los Angeles metropolitan area? Or is it because there's big differences between Los Angeles and Culver City, Beverly Hills, etc. And, you know, in most in most situations, it's both. But I think it's important to track those different scales. So we can think about policy solutions. It's something that's very common in school segregation literature as well, thinking about in metropolitan areas, we can think about segregation between schools, both within school districts, but then also between them. And you know, a very enduring fact in the school segregation research, because of Supreme Court cases in the 70s is that most segregation is happening between districts because a Los Angeles Unified looks very different from Culver City of Pasadena, Beverly Hills, rather than the schools within those districts looking different from one another. So the measure that I use and are several other evenness measures are what we call decomposable, which we can say, Okay, what's the total segregation among the smallest units we're interested in, whether that be schools or neighborhoods, and then we can say and how much of that is due to segregation kind of between these larger units or within them? So this is, I think, a really useful tool that Dan Lichter has a great paper on racial residential segregation, and other people are starting to pay attention to this macro segregation scale,

Shane Phillips  26:45  
really quick, clarifying question on the scores themselves, maybe just thinking about neighborhoods here, for each city? Are you essentially just having one score for neighborhood segregation in aggregate? Or are we talking about in some way, comparing each neighborhood to each other neighborhood?

Ann Owens  27:03  
Yeah, so the segregation index indices are at an aggregate level. So for every metropolitan area, we have a zero to one score, the building blocks of that measure is sort of the diversity of each tract and how it compares to the metro. So in building that measure, we do kind of assess how each tracks composition or diversity of housing stock compares to the metro, but I don't use those analytically. I just use the Metro level score. Okay,

Shane Phillips  27:29  
so the whole metro, you know, there are there are hundreds 1000s of neighborhoods, census tracts within each metro but the neighborhood score for each Metro, it's just a single score that is created from how far each neighborhood differs from the average Metro white. Exactly.

Ann Owens  27:46  
So on a zero to one scale, you know, for example, we would have an average level of housing segregation by renter versus owner occupied something like point two on a zero to one scale, and that would that would be the average among the 100 metros, but if you pulled out any individual Metro, you know, it would be point 1.4, or whatever the case may be.

Shane Phillips  28:04  
Got it. Alright, I think we can get into some results here. Your study looks at how these measures of segregation have changed over time from 1990 to 2014. And I think that's a good place to start. Big picture. How much segregation do we see at the different levels between neighborhoods, cities and towns? And then central cities and their suburbs? Which are your three geographic areas of analysis, or comparison? And how has segregation by housing type and cost changed over time?

Ann Owens  28:33  
So I look at segregation in a few different ways, as you mentioned, and again, I'm looking at the 100 largest metro areas that results aren't too different if I look at all, so the first thing I did was measure housing segregation by type and type here is like tenure. So renter versus owner occupied or single versus multifamily units. And so segregation between single family units and units in multifamily buildings increased by about 9% On average, and in 72, of the 100. Largest metros from 1990 to 2014 segregation between renter and owner occupied units was you know, it moved around a little bit during this time. But comparing 2014 to 1990, things didn't look that much different, it increased in about 60 metros declined in 40. So that's been a little bit more stable for both of these measures of housing segregation by type segregation is around, you know, point two on a zero to one scale, which is sort of a moderate level, I guess we would say. And then the second way I looked at housing segregation was by cost so either by rents or by home values, and I will say that my data source for all of this is census and American Community Survey data. And so rents and home values are reported by the residents and we can talk a little more about what that might mean. But just to note the data source their segregation by home values is higher than by rent both are in the kind of the point two to point three range. And we see some you know, ups and downs during In this period, probably because I'm looking at 1990 to 2014. Well, we had the great recession, one of my data points spans 2007 2008. So we sort of have that disruption in housing costs in the middle. So we sort of see a decline, especially from 2000 to 2007, in housing cost segregation, and then an increase after 2007. Finally, I consider segregation at different geographic scales. As I was mentioning before, you know, we can think about segregation either within places between neighborhoods, so within Los Angeles, among neighborhoods, or between places. And so I can say, again, using the measure that I use, how much total segregation between neighborhoods is happening between places. I also pay particular attention to segregation between the central or the major city of a metro and this is just a census to find, such as the first named one in a metro area. So in the Los Angeles, Long Beach, Glendale metropolitan division, Los Angeles is the main city and everywhere else is a suburb. This is not perfect. The Census does not define suburbs, which is very unhelpful. But this is, you know, one way to kind of capture that, looking across these four ways to measure housing segregation, rent or owner single multifamily, by rent or by housing costs. Across the four measures, about a third to 45% of housing segregation is occurring on this macro scale between places. So place or jurisdiction distinctions in you know, zoning policy composition, physical features, do stratify housing options, and that's important. The city suburban quote unquote, distinction is non trivial. It accounts for like just looking at segregation between that one place and every other place in the metros, on average accounts for depending on the type of housing segregation, we're looking at about seven to 15%. So I think place distinctions are important and paying attention to the macroscale is important. But on the other hand, the majority of housing segregation by type and by cost is occurring between neighborhoods within places. And so you know, municipal level measures of zoning or other variables aren't really can we know wouldn't really capture this variation or contribute to it. That's sort of what I find. It's certainly a both and story. But the segregation of housing units within cities seems to account for the majority of segregation. Yeah,

Shane Phillips  32:29  
one thing that stood out to me was in figure one, in this article showing the segregation scores at the at the metro area level, in 1990, and 2014 2014, was one axis 9090 versus the other, and you just see kind of a straight 45 degree line for each of these metro areas where if it was, you know, a point to in 1990, it's probably around 2.2, in 2014, if it was a point five in 1990, it's probably around 2.5, in 2014. And so is there anything we can say about just how persistent segregation seems to be within each of these places? Or metro areas?

Ann Owens  33:06  
Yeah, that's certainly true. I mean, I think residential segregation patterns are sticky. But housing segregation is maybe even stickier because it is about the built environment. And so for a metropolitan areas, segregation score, to really move on that, you know, something would have had to drastically change the location of their rental versus owner occupied or multifamily versus single family units, and the housing market would have had to change in such a way to make rents or home values less spatially separated.

Shane Phillips  33:38  
Yeah, in the case of the multifamily single family in particular, you'd have to literally either tear down a lot or build a whole lot in a pretty short period of time for it to change significantly. Right,

Ann Owens  33:49  
exactly. And yeah, the other things are more maybe market based, and those could have changed more, but we'd have to think about it, you know, it wouldn't just be the case that a metropolitan area was somehow experiencing like a decline in housing costs, it would also have to have this spatial component where things became sort of more or less even across space. So yeah, it's interesting. I think that, you know, one future paper that I or someone listening to this podcast should, right is really trying to isolate in places that have seen zoning reforms or other policies that we would expect to affect the distribution of housing. Have we actually seen that? And I know we'll talk more about policies later.

Paavo Monkkonen  34:27  
Yeah, it's almost like if cities had to plan for a more fair housing, in their land use they would have to change these things.

Ann Owens  34:34  
It's wild and then if we could you know, enforce it and have carrots and sticks imagine that world.

Paavo Monkkonen  34:40  
I guess I just wanted to mention your the on the central city suburb dichotomy, I think, was in Chicago, that was the most dramatic in that setup, I felt those numbers of you have at hand are useful in thinking about kind of what the central city for suburb means. Yes,

Ann Owens  34:56  
I do think Chicago is a good example. It's my hometown. I love Chicago, but Chicago needs to work on it segregation. So as you said, Chicago is among the 10 metropolitan areas with the most segregated single family versus multifamily. And this is a city in a suburb story. And you know, anecdotally, I know this from growing up in a suburb of Chicago, where, you know, I like all good housing nerds Do I look at the census data for my suburb a lot. And you know, there's like, literally maybe one or two rental units in the whole suburb. And so one statistic that I often use when I talk about this work is in Chicago, 60% of the multifamily units in the metro area are in the city, and only 19% of single family homes are in the city of Chicago. And so whether it's your preference or what you can afford, if you cannot afford a single family home, or you choose that you don't want to live in a single family home, most of your housing market is going to be in the city, right, as opposed to a place like Las Vegas, where about 30% of both single family and multifamily homes are in the actual city. And then 70% of both types are in the outlying areas. This

Shane Phillips  36:07  
is also making me think about how and this isn't in your article. And but a place like Chicago, if renters are over represented and renters are poorer, and are not paying as much in taxes and so forth, they might also require more services, because they're poor, it's placing a disproportionate burden on the city of Chicago to provide for these residents with less of a budget to do so. And so beyond just sort of the immediate individual impacts and costs and everything, there's just this larger scale or municipal burden that's being created. And I mean, this is a common thing with suburbs, where they sort of it might even be the responsibility to, to do larger cities or more urban places. But I think that just it just occurred to me. Yeah,

Ann Owens  36:52  
and I, you know, a couple of points on that. One is it is this idea of the concentration of poverty, right, you know, classic concept in sociology and other literature's William Julius Wilson, among others, but just this idea of any individual poor family is maybe gonna have a hard time making it work. When you cluster many, many poor families together, those challenges multiply, and we see this in the education literature as well, you know, it's challenging to address the needs of a single poor student, it's a lot harder when 75% of the class is poor. Right. And so thinking about, again, you know, an instrumental pitch for more integration is that overall, it you know, our society can function better when we're able to kind of take everyone take on the even share of the social safety net. Of course, you know, the second point on, I think, Pablo said, it's almost like suburbs do this on purpose. You know, many a suburb has been created specifically to kind of piggyback on to the social services and municipal services of large cities. So, again, nothing is natural. There's we have chosen this, we have chosen the levels of segregation that we have.

Shane Phillips  37:57  
Well, I think people might be surprised to hear that the Bethesda and New York metro areas are more segregated between single family and multifamily housing, and by owners and renters than metros like McAllen, Texas, and Little Rock, Arkansas. In fact, the first two have among the highest levels of housing type segregation and the ladder to have among the lowest in scores in the most segregated metros, we're about three to five times higher than those in the least segregated Metro. So there's a very big gap there. How should we think about that? Yeah, there

Ann Owens  38:32  
is a sizable range in the housing segregation landscapes of metropolitan areas. And I think just that fact alone was something that I was trying to emphasize with this article. And it is a super obvious point. But it is so odd to me that in almost all studies of residential segregation, the housing landscape is just like not part of the empirical work. And you know, all of the things that we know matter for residential segregation, income inequality, preferences, those are going to operate in very different ways, depending on what the housing landscape looked like. So just the first point of there is variation in the levels of housing segregation is something that I was trying to emphasize. And then as you mentioned, there's we see patterns of different types of places that have higher or lower segregation. So higher cost metropolitan areas, places where the residents are higher SES, that is positively associated with higher single versus multifamily segregation. I think one interpretation of that is, you know, economic inequality creates maybe a market for some people looking for more exclusive enclaves and the building industry is responding to that and policymakers are responding to that in terms of zoning to sort of create these more single family enclaves. The policy context certainly matters you know, I haven't done specific studies of each of these places, but you know, in terms of zoning, and also the size of cities and how they are carved up so East Coast metropolitan areas, may be more likely to have lots of Little jurisdictions, that's certainly the case in school segregation. There's many more school districts in the East Coast than there are in the south or the West. And anytime you sort of carve up large places into smaller units, then people are kind of trying to match their preferences more specifically. So that can increase overall segregation. So, yeah, I think there's a lot more research to be done on the causes, or the factors that contribute to housing segregation. But those are just some of the things that I was thinking about when I wrote the paper.

Paavo Monkkonen  40:28  
So mathematically, larger cities are more segregated. What's the zips law of size and segregation. So

Ann Owens  40:36  
the segregation measures that we're using shouldn't be too affected by size, because they are sort of just taking the variance basically, of any characteristic within a metro and then comparing it to tracks. So it's not necessarily a math story. Some of it can be an on the ground story of you know, when it's a physically smaller place, like there could just be fewer neighborhoods. And that might mean that, you know, people sort less or more actually, but there's fewer neighborhoods to sort of sort between. So I think that the measures themselves aren't too sensitive to size. But it could be just a real, you know, larger cities also tend to be more diverse people may, you know, urban economy story. Yeah. So I think it's, it's less of a technical point.

Paavo Monkkonen  41:20  
And like you said, I mean, the point about as cities grow and get more expensive, there may be greater benefits to exclusion. And I think that that showed up in the research I did with Mike lens on on zoning and housing regulations and economic segregation. And we find that actually, you know, it's it's a fact that you, I'm sure know, very well, having studied this for a long time, but most people don't think about that it's the higher income households that are creating most of the Metropolitan segregation, right. So they're much more isolated from middle income households and low income households than low income households, in fact, are from Middle and High income households, right? Yes.

Ann Owens  41:55  
And your paper with Mike was definitely an inspiration for this paper. And I think that's a really important point. I don't attend to it in this article. But when we think about housing segregation, by cost by rent or by income, that can be because of the concentration of poverty, like I was talking about earlier, but it can also be because of the concentration of affluence. And I didn't look at it carefully in this paper, but it is certainly something that we should pay attention to, because segregation creates winners and losers.

Shane Phillips  42:23  
So next up, you looked at the relationships between these four segregation measures, and a whole bunch of other metrics, including income segregation, and you found some pretty strong associations, neighborhoods with more segregation by housing type by tenure, and by both measures of cost also had much higher levels of income segregation. I also saw that the share of non Hispanic white residents was lower in neighborhoods with more single family to multifamily segregation, and more cost segregation. Can you break some of this down for us and also share any other associations you think are particularly interesting or important?

Ann Owens  43:02  
So I think that I spend more time in the paper the most time in the paper looking to see how housing segregation is associated with income segregation, but I think there's a lot more to do on predictors of housing segregation. So as you mentioned, I find that in metros with fewer white residents, single versus multifamily segregation is higher. And this This is a correlation and it's not a causal relationship necessarily, but it could reflect that in more racially ethnically diverse places. Single Family and multifamily homes are located in different neighborhoods due to a long history of exclusionary policies, right, you could think about, you know, back to suburbanization and white flight in the 50s 60s, etc. Those policies then have long standing impacts on the on the built environment. Some of that is due to white families trying to avoid neighbors that are not white. So in less diverse places, and more white places, you may not have had these same policies and patterns that affect the built environment so starkly. 

Shane Phillips  44:01  
So you also looked at how predicted levels of income segregation change as the levels of these four measures of housing segregation increase. So for example, on a scale of zero to one if neighborhoods at the 10th percentile of owner renter segregation have an income segregation score of point three, what score does your model predict for neighborhoods at the 90th percentile? Does that income segregation score increase 2.33 Just 10% higher or point six which would be 100% increase? I think this is just another way of thinking about how strongly these housing and Cost Segregation measures are associated with income segregation, seeing a large increase in the former produces a large or small increase in the ladder. Tell us what you found there for each of these measures, segregation by single family, multifamily owner renter rent and home value.

Ann Owens  44:51  
So comparing the values of renter owner segregation and if we look at that metropolitan areas at the 10% I love renter owner segregation. So a pretty integrated place versus the 90th. So much more segregated place on the renter owner segregation metric, income segregation is about 30%, higher at the 90th. And the 10th percentile. Similarly, income segregation is about 24%, higher at the 90th, and the 10th percentile of single versus multifamily unit segregation. And if we go to the housing cost metrics, it's a slightly smaller relationship, income segregation sort of jumps by, you know, 10 to 16%. If we compare the 10th, and 90th percentiles of housing cost segregation. So that might be hard to get your head around. I think one thing that I did was in a lot of income segregation research, the key predictor that people look at is income inequality, right? Because the idea is, well, in places where we just have big gaps in the income that folks at the top and the bottom of the income distribution earn, there's been a big gap in the sort of housing that they can afford. And so there's great studies showing that income inequality is a strong predictor of income segregation. And so what I find I just sort of replicate that finding here using similar models. And I find that the housing segregation metrics actually come in at you know, sort of a similar magnitude in terms of predictor. And so to me, that's telling me that, yes, income inequality is an important predictor of income segregation. But housing segregation is also robust. And I think kind of the complementary piece of the puzzle that has been previously overlooked here. 

Shane Phillips  46:30  
Something that stood out to me in these results was just how how income segregation increases more between neighborhoods that are the most and least segregated by single versus multifamily and owner versus renter compared to segregated by rent or by home value. That seems kind of surprising. It seems that rent and home values are direct measures of cost. So you might expect them to be more correlated with income segregation, since income determines how much costs you can afford. So what do you think was going on there? Or am I interpreting that correctly?

Ann Owens  47:03  
I think that is an interesting point. And a couple things. So the first is, you know, I in the paper to try to make the results more interpretable, I sort of talked about things that kind of the 10th, or the 90th, or 50th percentile of the distribution among the metro areas that I study. But those four measures of housing segregation all have different distributions. So if you instead kind of look at them all on a zero to one scale, rent, or owner occupied actually is sort of the like, the biggest coefficient there. But it is true that the cost segregation has kind of a smaller coefficient, which is potentially curious. So one hypothesis that I haven't tested is, there could be a measurement error in these things. So we're asking people to report the rent that they're paying right now, which is not necessarily the rent that something would have if it was back on the market. And same thing with home values. We're asking people what their estimate of the home value is, you know, there's been some research kind of assessing like, How accurate is this? And it's actually not too bad. But there's certainly some --

Paavo Monkkonen  48:07  
There's a good old John Quigley paper on that. 

Ann Owens  48:09  
Yeah, yeah, there's there is essentially, you know, there's there's potentially measurement error there. 

Paavo Monkkonen  48:13  
There's bias. The longer you've lived there, the less you assess it at.

Ann Owens  48:17  
So that's maybe part you know, I didn't look at kind of market data. So that's maybe one source of discrepancy. And it's also I don't really measure kind of like affordability, right. So I'm looking at housing cost segregation, it could be that high and low cost housing is located in like very different neighborhoods, but it's like not a huge, it's like I didn't really look at the correspondence between income and costs, right. So it could be the case that yes, the lower income folks end up in the lower cost areas, but that sort of correspondence could vary across places. So you know, for example, in Los Angeles, the highest cost neighborhoods are going to be just completely inaccessible to all but the highest income folks were in a more moderate cost metropolitan area that just sort of correspondence between cost and income. I feel like I'm not quite there's like a little bit of slippage there that I don't quite capture, but it's a great, it's a great point. 

Paavo Monkkonen  49:15  
Plus, I think it's probably a bit more dynamic. I mean, housing, the variation in housing costs within a metropolitan area changes over time, faster than the variation in housing typologies, for sure. And so there's probably some of that. Yeah. So and I wonder on that, I mean, do you think about the dynamic relationships at all? I mean, not explicitly that that much in this study, but having done this, do you think about how neighborhoods rents going up happens because higher income people are moving in and how that might affect its role in the overall metropolitan level of segregation? I mean, gentrification matters. 

Ann Owens  49:51  
Yeah, no, it's absolutely a cyclical relationship. You know, I set up kind of the theory of this paper to think about how housing segregation predicts income segregation. empirically, I don't know totally nailed the directionality of that relationship. And I think a more complex model would think about the feedback loops between the two measures. So certainly something to look into.

Shane Phillips  50:12  
Is this something where we might expect cities, you know, that are not meeting the demand for housing and have a bunch of gentrifying neighborhoods to actually appear better, at least on income segregation, not on the built environment, necessarily, but just having poor people leaving having richer people moving into poorer neighborhoods, it's all kind of homogenizing. Right. And maybe those bigger, richer cities are pulling people away from other metro areas that are in some ways, maybe becoming a little more segregated. But does that seem like a plausible thing that would be happening?

Ann Owens  50:45  
Yeah, I think the capturing the empirical relationship between gentrification and segregation is sort of tough, because it's tough to get the timing right. So I think like, there can be moments where a gentrifying neighborhood looks like an income diverse neighborhood. But then five years later, it's homogenously. higher income. And you know, one thing I'm interested in that I haven't had a chance to look at empirically to is thinking about, going back to this macro scale of segregation is thinking about when cities have stronger gentrification pressures, do you start to see higher macro level segregation, because lower income folks are pushed entirely out of the city, you know, thinking about a place like like LA, and we sort of end up with these curbs, exurbs, low income suburbs,

Shane Phillips  51:31  
Where neighborhood segregation might actually go down, but between place, between city right, segregation might go up.

Ann Owens  51:37  
And we have these like very wealthy cities, and then kind of a spatial distribution of income around them that is much less equal. And you know, we know that you know, suburbs are increasingly diverse, but what that's hiding is like suburbs as a whole are more diverse. But what tends to happen is you have sort of very high income and very low income are very white and very non white suburbs. So that diversity as a whole in the class of places that we define a suburbs is masking between suburb segregation.

Paavo Monkkonen  52:03  
And on that I'll just note, I mean, if if listeners are interested in further reading, the dissertation of aforementioned student, Andre comando on I think is fantastic. I don't think he's published papers out of it yet, but it's free on the UCLA dissertations website. And you know, he had folk has this focus on neighborhoods with stable diversity, because he looks over at, you know, five decades of time and kind of this gentrification, you when you realize that gentrifying neighborhoods segregation patterns are changing, then actually, what becomes important is stable diversity, rather than diversity at one given point in time. It's a complicated thing to study, but I think, extremely important for urban scholars.

Shane Phillips  52:43  
I did have a question on timing here, I guess maybe two questions. The last two time points in this study were 2007 and 2014, which coincide with roughly the peak of the housing market before the global financial crisis. And then at the tail end of Lowell after the global financial crisis after the housing crash that lasted for quite a few years. So I'm curious if you have any thoughts on how that timing may have affected any of these results. And beyond that, it's now been about 10 years since that last time point. And the housing market looks very different, much less affordable. I think we can agree than it did at that time. So I'm also just wondering if you've done any analysis on how these measures of segregation have evolved since 2014. I guess this might be at...

Ann Owens  53:36  
SegExplorer.org

Shane Phillips  53:37  
So curious to hear where things stand today. If we have data on that?

Ann Owens  53:42  
Yes. So I think you can definitely see the impacts of the 2007 crash. If you look at the trends and housing cost segregation from that 2000 to 2007 time point, you know, you see declines and those things, presumably, because the high end is just sort of coming down costs are declining there. And so you don't see this high level of segregation by cost. So I took a quick look at the data, looking at the 2018 to 22, American Community Survey five year aggregation. So you know, if we think about the midpoint of that is 2020, which has its own set of, of issues, I guess, we can't go seven years without a crisis, but or one year, or one month. But you know, back of the envelope calculations that I would not stake my career on show that all four metrics of housing segregation are actually up from 2014 to this kind of ACS year. There's some data issues and changes with the way the data are published that I would caution hanging your hat on that, but it does seem like the housing segregation metrics are up and so as I said before, I would be curious to think about a study that could try to compare trends in places that have or haven't implemented like drastic zoning changes to try to understand have those things moved the needle and also what's kind of the time lag on that right if in place abolish the single family zoning, while the neighborhoods obviously aren't going to look different the next day or the year later or the five years later. So I would be interested in thinking about designing a study where we could think about different policies that maybe should have moved the needle on this. And if we're starting to see results yet, but it might be too soon to tell.

Shane Phillips  55:18  
Last question from me thinking about policy solutions to this. What is your, what are your prescriptions for reducing these various forms of segregation over time? And, you talked about this in the article a bit, but what do we need to be thinking about so we don't end up worsening other problems as we try to resolve these ones?

Ann Owens  55:38  
Yes. So we need to get housing of a range of types and sizes and costs into all neighborhoods, right? We need sort of an everything, everywhere, all at once approach here. Maybe not all at once. But you know, ASAP, especially in California, and in so many metropolitan areas in cities where we just haven't built, or we haven't built in certain places for so long. I think zoning reform is necessary, but not sufficient, right. We need zoning that permits a range of housing types, which maps on to tenure and costs in, you know, every neighborhood in a metro area. But it's not sufficient. We also need — and this is one of the reasons that I just took it to the actual "where are these types of housing units actually located" rather than using zoning as a as a proxy — we also need incentives for private and public developers, and city and municipal leaders to welcome housing and to plan for housing in their neighborhoods. And we need enforcement. We need priorities and commitment from the state, and enforcement when they do not do so. 

I've also done research on the Low Income Housing Tax Credit Program, which is our federal supply side program. So on the one side, it's an important part of both our affordable housing stock, and also our multifamily housing stock overall, but it does not build enough housing for everyone that needs it. But in the Low Income Housing Tax Credit program, states include priorities for where affordable housing should be located. And almost every state — or every state actually, now — includes incentives for building in higher cost, high opportunity, in scare quotes, areas, which they measure in lots of different ways that make more and less sense. So I published a recent paper in the Journal of Urban Affairs with Rebecca Smith, analyzing how the introduction of California's Opportunity Map, which categorizes neighborhoods by their economic, educational, and environmental resources for children's development. That map is now used in the Low Income Housing Tax Credit program, and developers have competitive advantages if they propose to build in those areas. And so we do find an increase in the probability that developers proposed to build in those higher resource areas, and ultimately an increase in the number of affordable housing units located in those places. So that's good news. But, I think, you know, there's a lot of additional thinking that needs to be done there. How are we measuring opportunity? What's the right balance of LIHTC across neighborhoods of all types, right. We need to continue to invest in historically underserved communities, and especially in gentrifying areas to help folks stay there. And I do think some of this is a municipal level issue as well. I did interviews with affordable housing developers, and one thing they told me was, like, we... we sort of go back to the cities where we have a good relationship, where we know that they're kind of housing friendly. And like, we're not even going to try in some of these neighborhoods that declare themselves, you know, mountain lion sanctuaries, just as a random example. So I think that, right now, the LIHTC program is like, is not large enough, where... you know, developers can still find sites and they can still... they have plenty more sites and proposals than they have funding. And so they don't have to seek out land in every possible jurisdiction. And so I think some of this is sort of at a — again, thinking about this multi scalar level, at the jurisdiction level, and then at the within-jurisdiction, neighborhood-level, just trying to build, build, build and get housing of a range of types and a range of costs so that people have choice, and can actually bring those preferences, resources, constraints to bear in a housing market that gives them a lot more opportunity to live wherever they'd like. So, those are some of the things that I have been thinking about.

Paavo Monkkonen  59:08  
I'll just say, well, Ann, I'm so excited we got to chat with you about your paper, and I really recommend people check out your research, especially the school segregation stuff as well. I guess we don't have a school choice podcast to have you on for that, but I hope we can point people to that research because it's extremely important. 

Ann Owens  59:26  
Great, thanks. It's been a pleasure to chat with you.

Shane Phillips  59:28  
Maybe next year we'll start a school one. We'll just pretend like we're experts in it. Yeah.

Ann Owens  59:33  
I will host a limited housing and schools spin-off of the podcast. There we go.

Paavo Monkkonen  59:41  
That sounds cool. We should do a crossover. Yeah.

Shane Phillips  59:48  
You can read more about Ann's work on our website, lewis.ucla.edu. Show notes and a transcript of the interview are there too. The UCLA Lewis Center is on social media. I'm on Twitter at @shanedphillips and Paavo is at @elpaavo. Thanks for listening, we'll see you next time.