More

Episode Summary: Does gentrification lead to increased displacement of vulnerable low-income households? To date, research findings have been surprisingly mixed. One explanation may be that most gentrification studies focus on individual cities, which vary substantially from place to place, or the entire U.S., which may overlook local or regional differences. Hyojung Lee joins us to discuss his new study with co-author Kristin Perkins which categorizes the country into eight unique geographies according to shared characteristics, searching for differences in how gentrification impacts displacement of low-income households. It persuasively finds that gentrification does lead to more household moves — and importantly, more downward moves — and can hopefully inform further research and more location-appropriate anti-displacement strategies.

  • “Building on earlier findings (Glass 1964; Marcuse 1986), recent studies sometimes assume that gentrification causes or is associated with physical displacement of original residents, particularly lower-income, vulnerable residents. Some studies of gentrification, especially those based on qualitative data, present evidence that gentrification has negative consequences for longtime residents of changing neighborhoods (e.g., forced mobility and/or increases in housing costs) and for neighborhood and city character (Brown-Saracino 2017). Other research suggests that the displacement of low-income residents from gentrification is far less dramatic and conclusive (Ding, Hwang, and Divringi 2016; Ellen and O’Regan 2011).”

 

  • “We argue that metropolitan heterogeneity in the link between gentrification and displacement may help reconcile these conflicting findings. Just as the process of gentrification and neighborhood change unfolds differently across different types of neighborhoods (Hwang and Sampson 2014; Owens 2012; Rucks-Ahidiana 2021), the effects of gentrification may also vary based on the broader demographic and socioeconomic context … Although some studies discuss metro-, city-, or state-level differences and neighborhood-level patterns and processes of change, we are unaware of any that examines metropolitan variation in the association between gentrification and residential mobility. With these analyses, we heed an increasingly urgent call for a more robust sociology of variation that documents heterogeneity in social processes (Brown-Saracino 2015; Spring and Charleston 2021).”

 

  • “We use a large, individual-level national sample—collected by the American Community Survey (ACS) in 366 US metropolitan areas after the Great Recession—to test our hypothesis that the association between gentrification and residential mobility differs by the type of metropolitan area in which gentrification occurs and the intensity of gentrification. We define neighborhoods at risk of gentrification in 2010, identify which of these neighborhoods gentrify between 2010 and 2019, and follow individuals living in both gentrifying and non-gentrifying neighborhoods to assess how gentrification is associated with residential mobility and, among movers, destination neighborhood quality. Our geocoded, individual-level data allow us to observe individual education level, among other characteristics, to focus on the individuals presumably most vulnerable to disruption associated with displacement.”

 

  • “Results from linear probability models suggest that, overall, gentrification is associated with a significant but modest increase in residential mobility compared with non-gentrifying neighborhoods. The magnitude of the association increases as gentrification intensifies; moderate and intense gentrification are both associated with downward mobility in terms of neighborhood quality. We disaggregate results by metropolitan cluster and find significant positive associations between moderate and intense gentrification and residential mobility in Large Coastal, Large Southern/Midwestern, and College Town metros. Moderate gentrification is associated with elevated residential mobility in Retirement Destination metros and intense gentrification is positively associated with residential mobility in Plains Mountain, Southern, and Small Midwestern metros. Gentrification is weakly or not associated with residential mobility in Inland Empire/Texas Border metros. Movers from gentrifying neighborhoods generally end up in neighborhoods with higher poverty rates and lower median incomes than their origin neighborhoods. These results suggest that properly assessing the impact of contemporary gentrification requires considering the broader ecological context in which it occurs. Moreover, policy interventions aimed at mitigating displacement costs may be more necessary in some metros than in others.”

 

  • “Given concerns about displacement from qualitative portraits of gentrifying neighborhoods, there have been many attempts to quantify the amount of displacement induced by gentrification. Two primary methods have been used to estimate the magnitude of displacement from gentrifying neighborhoods (Carlson 2020; Freeman 2005; Jackson 2015). The first is a compositional approach comparing aggregate characteristics of neighborhood residents after gentrification begins with characteristics of neighborhood residents prior to gentrification. Displacement has occurred when there is population loss of a particular subgroup over time. The second is an individual-level approach that tracks residential mobility across neighborhoods but seldom accounts for the reasons individuals move (Freeman (2005) and Martin and Beck (2018) are exceptions). Ideally, research estimating the effect of gentrification on displacement would employ detailed individual-level data including motivation for moves, but such data are not available for sufficiently sized nationally representative samples at small enough geographic levels to assess conditions in both origin and destination neighborhoods (Carlson 2020; Jackson 2015).”

 

  • “The large and growing body of gentrification research has not yielded consensus on whether or not gentrification causes residential displacement. One explanation for inconsistent findings across studies could be different definitions of gentrification used in the analyses—relative income ratio versus growth in college-educated residents and increases in housing prices versus difference in housing price percentile and income percentile (e.g., see Appendix A in online supplement). Most of these studies, however, report that their results are robust to different measures of gentrification. A more compelling explanation for the inconsistency between single-city and national-level studies is that the association between gentrification and residential mobility differs between different types, or clusters, of metropolitan areas. Metropolitan areas differ markedly across a number of dimensions. For example, the recent reversal in regional income convergence means that metropolitan areas are becoming more unequal in terms of economic status, with the richest people and places pulling away from less wealthy areas (Manduca 2019). Poor neighborhoods are also now much more heterogeneous across metropolitan areas than they were in 1990 (Small, Manduca, and Johnston 2018). Patterns of residential mobility are conditioned by characteristics of the metropolitan areas in which the movers live. Metropolitan-level racial composition, Black-white racial segregation, the distribution of neighborhood poverty, and supply of new housing are all associated with residential mobility and neighborhood attainment (the racial composition and poverty level of the mover’s destination neighborhood) net of individual-level characteristics that predict mobility (Crowder, Pais, and South 2012; South, Pais, and Crowder 2011).”

 

  • “Following the well-established approach of previous literature, we examine the association between gentrification and displacement by comparing mobility outcomes among residents from gentrifying neighborhoods with mobility outcomes among residents from gentrifiable yet non-gentrifying neighborhoods. If gentrification involves displacement, the residents in gentrifying areas should have higher mobility rates and may be more likely to move into lower-quality neighborhoods than those in similar non-gentrifying neighborhoods, controlling for other covariates.”

 

  • “One key issue is defining gentrifying and non-gentrifying neighborhoods. Building on prior research, we conceptualize gentrification as an influx of residents with higher socioeconomic status into low-income, central city neighborhoods. Thus, we define a gentrifiable neighborhood as a low-income and central urban census tract. A census tract is low-income if it has a median household income in the bottom 40% of the income distribution across census tracts within its Metropolitan Statistical Area (MSA) at the beginning of our observation period (Dragan, Ellen, and Glied 2020; Ellen and O’Regan 2011) … Among the neighborhoods at risk of gentrification, we classify a tract as gentrifying when the neighborhood experiences a substantial increase in its within-MSA percentile rank of college graduate share over the time period examined. A benefit of this measure is that it captures relative change in an MSA. We start by ranking the tracts in a metropolitan area from lowest to highest in terms of the share of population 25 and over that has a college degree at the start of our observation period. Suppose a gentrifiable tract starts with a low share of population with a college degree, placing it at the 25th percentile in the MSA. Over our study period, this tract experiences an increase in the share of its population that has a college degree that outpaces the overall increase in the college graduate share in the MSA such that when we re-rank the tracts from lowest share of college-educated to highest share of college-educated at the end of the observation period this tract is now at the 50th percentile, experiencing a gain of 25 percentiles over the observation period. We consider the intensity of gentrification rather than using a dichotomous measure, common in prior research: we classify a neighborhood as “moderately gentrifying” when its ranking improves by 5 to 14 percentiles and “intensely gentrifying” when its ranking improves by 15 percentiles or more. We classify the remaining gentrifiable neighborhoods as “non-gentrifying.””

 

  • “Our key variables are mobility status in the survey year and residence one year prior. Using these variables, we identify individuals who moved to a different neighborhood in the prior 12 months and the characteristics of their origin and destination neighborhoods. For example, using the 2019 ACS, we identify individuals who moved from gentrifying neighborhoods between 2018 and 2019, and their neighborhood characteristics in 2018 and 2019 (origin and destination neighborhoods for movers). We control for potential individual-level confounders between gentrification and mobility using age, gender, marital status, race and ethnicity, and educational attainment as proxies for life cycle factors (Horowitz and Entwisle 2021; Rossi 1955). We include neighborhood characteristics such as racial minority share, college graduate share, median income, median home value, and median gross rent, often found in relevant studies (Ding, Hwang, and Divringi 2016; Dragan, Ellen, and Glied 2020), as controls in the regression models.”

 

  • “These data are not longitudinal, so we analyze one-year mobility outcomes among residents from gentrifying and non-gentrifying neighborhoods cross-sectionally, pooling annual data from 2011 to 2019. Therefore, we focus on the relationship between gentrification and short-term residential mobility, similar to recent research using CCP data (Ding, Hwang, and Divringi 2016; Hwang and Ding 2020; Hwang and Shrimali 2021), whereas Brummet and Reed (2019) examine gentrification’s effects on long-term residential mobility outcomes.”

 

  • “Estimating the association between gentrification and residential mobility in all 366 MSAs is informative, yet average national estimates may disguise substantial spatial heterogeneity in the association across MSAs with different economic conditions and population characteristics … We provide a more comprehensive picture by exploring this spatial heterogeneity. Specifically, we identify groups of MSAs with distinct common characteristics using PCA and cluster analysis. This approach allows us to group MSAs more thoroughly and objectively based on a holistic set of characteristics rather than imposing arbitrary rules and thresholds based on single characteristics like population size or region … The resulting MSA clusters are shown in Table 3. Since we obtained eight principal components that had an eigenvalue over one, we grouped the 366 MSAs into eight groups: Large Coastal metros generally include East and West coast cities with large shares of college graduates, high-income earners, and high-status occupations and high housing costs (e.g., New York, Los Angeles, Washington, DC); Large Southern/Midwestern metros represent large metros that have similar characteristics to the first cluster, but to a lesser degree (e.g., Dallas, Atlanta, Minneapolis); Inland Empire/Texas Border metros consist of Central and Inland Empire cities in California and Mexican-Texas border cities with a majority Hispanic population (e.g., Riverside, Fresno, and El Paso); Southern metros are made up of cities in Southern states with large shares of Black residents and single-family homes (e.g., Memphis, New Orleans, Birmingham); Plains Mountain metros include low-density cities that have large shares of non-Hispanic whites, married couple households, and low levels of income inequality and residential segregation (e.g., Oklahoma City, Salt Lake City, and Des Moines); Small Midwestern metros include mid- and small-sized MSAs that have a large share of non-Hispanic whites, formerly had a strong industrial base, and offer affordable housing options (e.g., Dayton, Grand Rapids, Akron); Retirement Destinations metros include many rapidly growing Florida metros with large shares of seniors 65 and over, non-Hispanic whites, and homeowners (e.g., North Port and Cape Coral); and our final cluster, College Towns, includes small metros with high turnover rates, college graduate share, and high income inequality (e.g., Durham-Chapel Hill, Tallahassee, Ann Arbor).”

 

  • “Table 4 reports results of linear probability models using the aggregate national sample from the 2011–2019 ACS microdata. Results from the base model, controlling for MSA and period fixed effects, indicate that the probability of moving to a different neighborhood among residents of moderately gentrifying neighborhoods is 0.30 percentage points higher than among residents of non-gentrifying neighborhoods. Residents in intensely gentrifying neighborhoods are even more likely to move than those in non-gentrifying neighborhoods, by 0.95 percentage points … The estimated coefficient for gentrification does not change substantially after introducing neighborhood characteristics into the model: controlling for individual and neighborhood characteristics, residents in moderately and intensely gentrifying neighborhoods have 0.38- and 1.00-percentage point higher probabilities of moving to a different neighborhood compared with those in non-gentrifying neighborhoods. Considering that the mobility rate is about 9.2% in the sample (Table 2), the probability of moving is about 4.1% and 10.8% greater in moderately and intensely gentrifying areas than that in non-gentrifying areas, controlling for individual and origin neighborhood characteristics. The magnitude of the association between gentrification and residential mobility is similar to what Ding, Hwang, and Divringi (2016) found in Philadelphia.”

 

  • “Compared with movers from non-gentrifying neighborhoods, movers from moderately and intensely gentrifying neighborhoods end up in higher poverty neighborhoods (0.68 and 1.40 percentage points greater) and lower income neighborhoods (median household income $1,351 and $2,848 less). Summarizing results from the aggregate national sample, gentrification is positively associated with the residential mobility rate and negatively associated with destination neighborhood quality.”

 

  • “The results indicate that the association between gentrification and residential mobility varies across metropolitan clusters. In the Large Coastal metros (reference group), the probability of moving into a different neighborhood is 0.42 and 0.92 percentage points higher among residents in moderately and intensely gentrifying neighborhoods compared with non-gentrifying neighborhoods (plotted in Figure 2). The magnitude of the associations in Large Southern/Midwestern metros is similar, 0.49 and 0.93 percentage points as shown in Table 7 and Figure 2. In these clusters, intense gentrification is associated with higher mobility out of neighborhood and city and downward mobility in terms of poverty rate and median income.”

 

  • “In Retirement Destination and College Town metros, moderate gentrification is associated with moving to a different neighborhood, city, and MSA, and the magnitudes of coefficients are significantly greater than those in Large Coastal metros. Out-of-neighborhood mobility rates in moderately and intensely gentrifying neighborhoods in College Town metros are 2.5 and 2.1 percentage points higher than non-gentrifying neighborhoods. Unlike in Large Coastal and Large Southern/Midwestern metros, however, gentrification was not significantly associated with downward mobility in these MSA clusters.”

 

  • “In the Southern, Plains Mountain, and Small Midwestern metros, intense gentrification is associated with within-city moves (1.19, 1.75, and 0.90 percentage points, respectively, see Figure 2) and, in Small Midwestern metros, downward mobility in terms of median income, whereas gentrification’s association with longer-distance moves is not consistently significant. Only the association between intense gentrification and mobility out of the MSA is significant in the Inland Empire and Texas Border cities; gentrification is not associated with within-MSA moves in this cluster. This geographic heterogeneity urges caution before generalizing results from a limited number of cities or national averages and warrants further research.”

 

  • “We investigate the association between gentrification and residential mobility in the nation’s 366 metropolitan areas in the 2010s and find that moderate and intense gentrification are significantly associated with increases in the probability of moving among less-educated residents of gentrifying neighborhoods compared with similar residents of similar neighborhoods that did not gentrify. These findings are consistent with the findings of recent research that uses a national sample and concludes that gentrification is associated with higher rates of residential mobility (Brummet and Reed 2019), but are counter to the finding of no elevated rate of residential mobility in gentrifying neighborhoods from a national sample during the 1990s (Ellen and O’Regan 2011).”

 

  • “Gentrification and neighborhood change unfold differently across neighborhoods with different characteristics (Hwang and Sampson 2014; Owens 2012), and our results provide evidence that the effects of gentrification may also differ by the broader metropolitan context in which it occurs. Gentrification has a relatively large positive significant association with residential mobility in Large Coastal and Large Southern/Midwestern metros, large cities that have high levels of residential segregation. Prior research in Philadelphia, New York, and the San Francisco Bay Area found that low-resource residents of gentrifying neighborhoods were no more likely to move compared with similar residents of non-gentrifying neighborhoods (Ding, Hwang, and Divringi 2016; Dragan, Ellen, and Glied 2020; Hwang and Shrimali 2021). Our findings for the clusters containing Philadelphia, New York, and San Francisco show that gentrification is associated with higher mobility rates out of neighborhood, city, and MSA, in contrast to prior research.”

 

  • “We contribute to the literature on gentrification and neighborhood change by showing that the association between gentrification and residential mobility depends on the intensity of neighborhood socioeconomic upgrading and the MSA context in which it occurs. Significant interaction terms support the conclusion that moderate gentrification is associated with a significantly higher probability of residential mobility in Retirement Destinations and College Towns and a significantly lower probability of mobility out of the MSA in Southern and Small Midwestern metros compared with Large Coastal metros.”

 

  • “Characteristics of the movers into gentrifying neighborhoods may also explain variation in the association between gentrification and residential mobility across MSA clusters. Movers into gentrifying neighborhoods in Large Coastal metros fit the stereotypical image of gentrifiers: disproportionately young, white, unmarried, native-born, and renters (see Appendix G in the online supplement). The racial/ethnic composition and socio-economic status of newcomers to and original residents in gentrifying neighborhoods, compared to non-gentrifying counterparts, are noticeably different in other MSA clusters. In particular, the profile of movers in to gentrifying neighborhoods in Retirement Destinations and College Towns differs from Large Coastal metros: In-movers are similar to original residents in terms of race (Retirement Destinations) or disproportionately non-white and foreign-born (College Towns) and in both clusters a large share of movers come from different MSAs and abroad. Movers to gentrifying neighborhoods in some MSAs come from local suburbs whereas in others, new residents of gentrifying areas arrive from suburbs and central city neighborhoods in other MSAs (Smith, Pride, and Schmitt-Sands 2017).”

Shane Phillips 0:04
Hello, this is the UCLA Housing Voice podcast, and I'm your host, Shane Phillips. This week's episode is with Professor Hyojung Lee, and we're talking about how gentrification may have different displacement impacts depending on the kind of metro area where it's occurring, and how big those differences actually are. The relationship between gentrification and residential displacement has been surprisingly difficult to confirm with empirical research. That's partly because definitions of gentrification differ. But even more importantly, because measuring displacement is really hard especially when we try to isolate involuntary displacement from other kinds of moves. By grouping together different kinds of metro areas by their shared characteristics, Hyojung and his co author Kristin Perkins, arguably have made the most persuasive case yet that there really is a statistically significant association between gentrification and the more harmful kinds of residential mobility. At the same time, they've also given us a deeper understanding of how gentrification impacts can differ depending on whether someone lives in a large coastal metro area, a smaller Midwestern one, or retirement destination, or a college town or one of several other place categories. We talk with Hyojung about how they put together this major research effort, what they found, and what it means for how we think about gentrification, and what we still need to learn to better inform our efforts at preventing and mitigating displacement. The Housing Voice podcast is a production of the UCLA Lewis Center for Regional Policy Studies. With production support from Claudia Bustamante and Jason Sutedja, and Divine Mutoni helping us out on transcripts. You can email me at Shanephillips@ucla.edu, and you can give the show a five star rating at Apple or Spotify. Now let's get to our conversation with Professor Hyojung Lee.

Hyojung Lee is Assistant Professor of Housing and Property Management at Virginia Tech in what might be my favorite department name ever, the Department of Apparel, Housing, and Resource management, and he's here today to talk with us about the different forms gentrification can take including a classification system he's developed with one of our previous guests, Kristin Perkins. Gentrification is sometimes talked about like it's a monolithic concept or force, but hildren and Kristen's work is helping shed some light on how its characteristics and impacts differ depending on the kind of community you're looking at so that's what we're going to talk about here. Hyonjung, thanks for being here, and welcome to the Housing Voice podcast.

Hyojung Lee 1:52
Thanks for having me. Happy to be here today.

Shane Phillips 2:45
And my co-host today is Mike lens. Welcome, Mike.

Michael Lens 2:48
Thank you, Shane. Welcome to Hyojung and really looking forward to today's conversation.

Shane Phillips 2:53
So we're going to start off with a tour of a place our guest knows well, and he has picked Seoul. I will say he kind of wanted to do Los Angeles but we've heard quite a bit of that, and I don't know that we've had a tour of Seoul so we kind of pushed him in that direction. I've never been, I don't know about you, Mike..

Michael Lens 3:08
Nope

Shane Phillips 3:09
But Hyojung, where are you taking us?

Hyojung Lee 3:11
Sure, so I've lived in many places and loved most of them for different reasons. But if I must choose one, I would like to invite you guys to Seoul, which is the capital city of the South Korea, and a huge city that has about 10 million people, so quite similar to LA's pattern size right? The thing that I liked the most about the city is its wide variety of urban spaces. I mean, if you've ever tried Korean food, Bimbim Bap, have you ever tried it?

Michael Lens 3:39
Yes.

Shane Phillips 3:40
It's Los Angeles, cmon

Hyojung Lee 3:41
Like it's kind of mixed rice with a lot of meat and vegetables. Seoul is just like that, you know, it has so many places with unique local character, and a very organic way that those neighborhoods are mixed makes the city very diverse and dynamic. And in the city, you can find 100 years old palaces or traditional Korean houses right next to skyscrapers and modern architecture. And then you know, for those planners, maybe I would recommend visiting Cheonggyecheon, which is a restore string in downtown Seoul. So that was a site (or) a major project that basically remove a highway in the heart of the city like I-10 in LA, or I-9 in Chicago, and created open spaces all along those turned lines, restore the strip.

Shane Phillips 4:34
If you've been if you're like an urbanist, who's been online in the past 10 years, you've definitely seen a before and after of this former freeway now, a stream and kind of public open space yeah

Unknown Speaker 4:46
Sure. So you definitely saw that kind of pictures of Cheonggyecheon before Yeah, so it was so dramatic, and you know, that was the moment that people in Seoul, in Korea in general, realize that there are more important things like the environment, like quality of life, and sustainability, things just kind of efficiency and growth that we had pursued during the, you know, rapid economic growth period. So I think that was a great kind of example of great sites to visit for urban planners.

Michael Lens 5:17
So, this is pretty simplistic, but I gotta imagine that there's a lot more familiarity with Korean culture to an extent for Americans these days due to the kind of explosion of Korean cinema, and I mean, you know, I don't know if I would call Squid games and whatever, like, cinema, but you know, it's, it's speaking very anecdotally and personally, like, I feel like, you know, through Netflix or, you know, cinema or whatever, like, I'm actually observing something about a country that, you know, I knew mostly as like a growing economic powerhouse in Asia but didn't have any window into kind of like what life was like there, but I don't know if others feel the same.

Shane Phillips 6:05
Well, there's the cinema, there's the TV, there's also like K-Pop, K beauty. I mean, that's not necessarily giving you the sense of what it's like there but the cultural inroads have been massive in the past, like, decade or so for sure.

Michael Lens 6:18
Yeah, that's what I'm trying to say. Better said Shane

Unknown Speaker 6:23
So in terms of that kind of cultural theme, just one thing that I want to introduce more is the nightlife in Seoul. Like nightlife is something that makes Seoul special, and that's something that I miss a lot in here in the United States. So you know, suppose that you arrive in Korea, like 11pm, somehow can't afford sleep due to maybe jetlag, or something like that. But no worries in Korea, there are so many places to go like from 24 hour restaurants, to karoke to night market and street food, there's so many things to do. So if you saw kind of 24 hour kind of restaurants in Korea, Korea Town, that is our culture so basically, the city runs 24/7. And I like to kind of lively atmosphere within the city so much.

Michael Lens 7:15
Yeah, I definitely am down to travel to Seoul and like hop off the plane and go to karaoke, I got no problem with that.

Shane Phillips 7:22
A hundred percent. This is very much going on a tangent here but I heard something a while ago, some other podcasts. But they were talking about how that culture actually developed in South Korea, and how it was, at least in part, a result of the kind of repressive political system, you know, 30 years ago, whenever and how people, there was like a curfew, you couldn't be out after a certain time so people would go out to these clubs and bars and just stay until daylight so they wouldn't get in trouble. And then it just became part of the culture which is such an interesting story. I mean, correct me if that's entirely wrong, but that's my recollection.

Hyojung Lee 8:02
That's exactly correct. So, you know, I recently saw a movie 1986, what happened in Argentina during their authoritarian regime after the coup d'etat, and that was exactly seeing what happened to Korea. So in the 60s and 70s, people couldn't go out during the night and there was an oppressive, authoritarian kind of government in Korea so people couldn't have that kind of freedom. And then, in 1980s, and 90, we had some spring, and so a lot of the things we enjoy. So that is part of reason why those K-Pop things and then like K-drama movies are so thriving in these days, because we really wanted to express ourselves through movies, dramas and songs, and we couldn't do that for several decades, and now you're allowed to do that so you're kind of..

Shane Phillips 8:56
Yeah, well, let's just pretend like we've got a great transition from that to gentrification. I mean, hey, like clubs and bars coming in, that's kind of gentrification, who knows?

Michael Lens 9:05
We can talk about these forever but you should move on, right?

Shane Phillips 9:09
Sure, so the paper we're discussing was published last year in the Journal Social Forces, and it's titled 'The Geography of Gentrification and Residential Mobility'. In this study, I would say you're looking at gentrification through three lenses. First is the intensity of change, or how much a places is gentrified. Second is residential mobility, which is sort of a proxy for displacement but not a perfect one, because only a fraction of moves are forced or involuntary. And third is the characteristics of the place itself where that gentrification is occurring, and in whatever ways it's occurring. Bringing these all together, the idea is that if you were to control for the amount of gentrification that's happening, and we'll get into how you measure that momentarily, you might find that the amount of residential mobility or displacement specifically varies depending on whether you're in a highly productive coastal city, an area that's popular with retirees or college students, or a rapidly growing metro in the sunbelt for example. A lot of previous gentrification research only looks at the US as a whole, which could be hiding some local variation, or at a single city or metro area, which can be hard to generalize to other places. So your work represents, I think, a big step forward by looking at gentrification and residential mobility in clusters of places that share similar characteristics. You did indeed find that gentrification affects residential mobility differently depending on what kind of metro area a neighborhood is located in. So we're going to get into those findings in this conversation. But before we get there, I think we first want to make sure that we're on the same page about definitions. gentrification is, at this point, a notoriously slippery concept that for some people has a very specific meaning. Like, for example, a change in neighborhood demographics that's accompanied by a rise in forced or involuntary displacement of lower income renters. For other people, the meaning is much less precise, and could really refer to pretty much any neighborhood change, whether that's a bike lane coming in, or a new wave coffee shop, or the construction of a four story building. Residential mobility is something we discussed in our episode with Kristen Perkins, and it's sort of an umbrella term that includes involuntary displacement, but also many other kinds of moves that don't necessarily have negative connotations. To make sure that we don't have you using these words in one way, and our audience interpreting them in another, tell us how you define gentrification and residential mobility, When it comes to your research, what processes or changes or impacts do you associate with those terms, and we can come back to the ways that you actually measure gentrification and mobility a little bit later. But for now, let's just get on the same page about what these mean to you as concepts.

Sure, this was exactly the reason that I avoided using the term gentrification in any of my prior research to this one. Just like you know, other urban phenomena like urban sprawl, suburb, and neighborhood; we have a vague idea in our heads but that concept does not necessarily have a very clear-cut kind of definition. So when we talk about gentrification people often say, what does it mean? Or, you know, gentrification isn't what you may think it is, right. So as you mentioned, something gentrification involves the displacement of original, and often low income residents, and sometimes substantial changes in racial and ethnic character. But others don't agree with that. So yes, we need to first clarify what gentrification means when we talk about that. And in this paper, we consider changes in racial and ethnic composition, and displacement of original resident as potential outcomes of gentrification rather than its elements. So basically, we conceptualize gentrification, as you know, substantial socio-economic neighborhood upgrading of formerly low-income and urban neighborhoods, which is largely driven by an influx of people with higher socioeconomic status from some other neighborhoods. So that is basic idea that we had for gentrification, and how to measure that using some variables is kind of different story.

Michael Lens 13:39
So Hyojung, the way, you're conceptualizing gentrification. And we're, I think we're trying to start with like, the high-level concept of gentrification in your mind, and in Kristen Perkins' mind, your co-author, what you're capturing there to me is really focused on socio-economic change of the individuals that occupy a neighborhood as opposed to the physical changes that might be in a neighborhood that might accompany that socio-economic change, it might not, it happen before, but to have it after, etc. So you're really focusing on your first concept is socio the-economic change in a neighborhood., and then, you know, another thing that I think we want to specify there is like, that can be the result of lower-income people moving out at a faster rate than they normally do. It can be a result of higher-income people moving in faster than other places but we haven't even gotten to that part yet right? It's just changes in neighborhood composition, according to the socio-economic features of the individuals that live there, and then when I said, you know, higher income, lower income that can I can also be raised, it could be education, right?

Unknown Speaker 14:56
Sure. I mean, so there might be some other kind of past aspect like cultural gentrification people with different cultural background comes in. So there are some conflict between the two groups and then one group displaced, some displaced out something like that. Or there might be some other concept, but here we are using, you know, socio-economic terms, maybe part of the reason why we are doing this is we wanted to compare our research with previous research. And previous research usually adapted this kind of idea of social commitment to education. And maybe thepart of the reason why they this was they could find much easily find those socio-economic related variables, rather than some variables that captures cultural differences like number of maybe Korean restaurants in the area, number of Starbucks in the areas like that, right? Yeah,

Shane Phillips 15:46
Yeah, it's a much harder thing to measure cultural change, especially the data we have available. And I did want to just highlight here that what you're doing is, you're separating gentrification, from displacement. I think for some people those things go together. Often, that's how I've talked about gentrification is like, it's not really gentrification unless people are being displaced, you know, more than in other places. But that's just like kind of a personal preference, and I think there's value in what you're doing here in looking at places that have this socio-economic change happening, and looking at, you know, whether the displacement outcomes actually differ, depending on the kind of location. So I think there's a good reason to separate those two things but I just want to make very clear that gentrification can happen or not in a place, and that displacement can happen or not displacement, and can be an outcome of that, but it might not be.

Hyojung Lee 16:39
Exactly, so in regards to residential imobility, you know, unfortunately, our definition includes both good and bad moves - that is both forced and voluntary moves on all those people who are in gentrifying neighborhoods within a year. So that was largely due to the fact that the American Community Survey, which we use for our study, does not have any variable about the motivation for moving so thatw as one reason. And our other reason was that like, even, even if the ACS had that kind of reason for moving question, like other services, such as PSA SAT or case ID, there have been some kind of measurement error, meaning that, you know, people when they decide to move, they may have more than one reason to do so. And the reasons that people report may not capture that kind of involuntary displacement that you're interested in.

Shane Phillips 17:35
Or they might, you know, really emphasize that, because that's really most prominent to them, but they might have actually ended up moving anyway, and just kind of like attached to that explanation. It could go any direction, (it is) very hard to know when people have multiple reasons for moving.

Hyojung Lee 17:49
Right, I mean it could be as housing reasons but actually that was due to gentrification. And you don't know how they would answer that reason for moving question, because there might be some kind of systemic differences in their reporting patterns across different types of people maybe you know so you just wanted to go with just to use the conceptual version mobility that includes all moves. So basic assumption that we had was that if gentrification involves "displacement", it should also increase the resident mobility of origin residents in those areas. So if there are some relationship between gentrification and racial mobility, there could be a supportive evidence of arguing that, you know, there is some relationship between gentrification displacement.

Shane Phillips 18:39
Yeah, so let's talk about before we get into your research in more detail, let's talk about the state of research on gentrification and displacement more broadly, there's been a lot published on these subjects, and the findings, I think, are pretty mixed and hard to make sense of. It seems like your goal in this study, or one goal was to try to reconcile earlier findings by weaving together those three threads that I mentioned: the intensity of gentrification, the amount or type of displacement, and the metro area characteristics. So where did things stand when you embarked on this research, and what unresolved questions were you hoping to answer with it?

Unknown Speaker 19:18
Sure, we often simply think that, you know, gentrification causes displacement of original residence, especially when they are low-income people and racial minority. But recent studies shows that it is less clear that there is such a strong relationship between the two. So some of the previous researcher, especially those based on qualitative approach based on survey or ethnographic approaches, show evidence that you know, gentrification has very negative impact on longtime residents in terms of forced moves, or increased housing cost burdens, and dramatic kind of changes in the original character. But other studies are often based on like large scale datasets with reverse fences, econometric methods show somewhat kind of mixed results or in show much of the last adverse gentrification results that we produce before. So when Kirsten and I began our research, what we wanted to do was not only just empirically examine the relationship between gentrification and mobility, but also wanted to figure out why previous studies had reported somewhat kind of inconsistent findings. One possible explanation for the inconsistent finding was different definitions and measures of gentrification using this kind of different studies as we just have before. So to address this kind of issue, we carefully selected a measure of gentrification that we thought that was the most vulnerable to the concern. And then we tested findings based on various definitions used in other studies. So that was something that you would argue that our contribution is split, or is that the findings in terms of the relationship between gentrification and reality that we found were pretty consistent across different measures? So you know, Appendix, there's a table that includes all the regions based on that and solve different definitions, you can confirm that. And, you know, I know, it's kind of crazy to run all of them, and tabulate the readers in kind of one cup make shoes table. So we basically record it kind of plays table, and then a lot of candidate was geographic heterogeneity in the relationship between the two, maybe, you know, different areas might have somewhat different or distinctive local contacts, that makes them have better or worse outcomes from the same urban phenomenon. But look, there has been a little less discussion on this issue. So sometimes some empirical study is basically focused on what happened in a single city like New York, Philadelphia, like San Francisco, and it's pretty difficult to generalize their findings to other cities with very different characteristics like Denver or New Orleans. And in other cases, researchers have simply put everything you know, from half 50 metros, 100 metros, is shows us wrong, miss our estimate of the association between the two and say, Yeah, gentrification displaced people, or no, it does not. But what we suspected was, maybe it depends, right? I know, it's a medical word that works every time. But that's actually something important to know. Yeah, it's important because understanding whether there are some variations across different types of metros that is important and necessary to find different ways to address, you know, potentially negative effects of gentrification, according to different local contests that can, that we can talk later.

Michael Lens 23:01
Yeah, so I wanted to underline some of that, for sure. I think I want the listeners to maybe appreciate like how, I don't know controversial, the previous research on gentrification and displacement are particularly on the quantitative side. So, you know, the common understanding of gentrification is that it at the very least, is a process that leads to the displacement of low income people and racial minorities and how June, I think, articulated that very well. But as far as the quantitative research and evidence that we have on on those processes, more often than not, I would say the evidence says Not really. And so I'm thinking of probably an earliest paper on this is Freeman and Bocconi. 2004 I think, Ellen and O'Regan and 2011. And then if you're you're listening at home, and you want to look at like a summary of some of this really good research Lei Ding and Jackie Wong have a great paper, I think, from 2019, that really spells out, you know, what we know about the causes of, or the causes and consequences of gentrification, one of those consequences potentially be in displacement. And all that is is again to underline like we, you know, as far as quantitative researchers, we all kind of shrug our shoulders and say, we can't really find evidence that changing neighborhoods means like, there's some big acceleration and displacement on average. And I think the best explanation for that has been that maybe some people pay more money to stick around in neighborhoods as as they're changing, you know, maybe the neighborhoods are improving along some dimensions that they're willing to pay more for. That's not necessarily like a great outcome, because then people are spending more of their household budget on rent to stay in a gentrifying neighborhood. But, you know, it does, I think, give one explanation for why there's this He's kind of cloudy relationship between neighborhood gentrification and what we really think of as displacement.

Shane Phillips 25:07
And if that were at least part of the explanation for what's going on, I think it also helps sort of explain the difference between the qualitative research findings and the quantitative, right, because this is something that came up in our interview with Elizabeth Delmelle, which also kind of touched on displacement around transit. And she made the point that these quantitative analyses are really looking at averages, essentially, but but there is no average person, people don't live their lives as averages. And so if some people are being displaced, and other people are staying longer and willing to pay more for that, and that sort of washes out to no increase displacement, that doesn't really mean anything to the person who was displaced the fact that someone else stayed longer and averaged out their displacement, like, what do they care, right, we should still be concerned with the what happened to that person. And so this gets all back to Hillsong. And Kristen's work on whether this is even happening. And maybe some of the washing out are canceling out is actually happening at the metro area level where some metro areas are experiencing more displacement, others less and it's kind of when you put them all together, it doesn't look like much, but maybe, depending on the the conditions or circumstances in different kinds of metros, we see different results, right. So let me just jump ahead here. And say, at the heart of this analysis, what I think makes it novel and a step forward in this area of research is that grouping of metro areas into clusters based on characteristics that they share with each other. And that set them apart from others. You identified eight clusters or groupings in total. And I'll just list those here really quickly. They were large, coastal, large, Southern and Midwestern, Inland Empire slash Texas border, southern plains mountain, small Midwestern retirement destinations, and college towns. So these are again, like metro areas that all share some are many characteristics and distinguish them from other kinds of places. And again, the goal here was to assess whether there are different displacement outcomes in each of these geographies. If we hold the intensity of gentrification constant, aside from their names, which I think do a pretty good job of describing the kinds of places that we're talking about, or at least their, you know, geographic location. What else should we know about these different geography clusters you identified? And how you group them together?

Hyojung Lee 27:39
Yeah. After we decided to see some difference between different types of metros, we had to answer the question of how to do so right. So we originally did that by running regressions by maybe census regions, and by population size, by maybe housing prices, etc, etc. And then realize that there are some issues with that. The first one was that there might be some confounding factors. And we may cherry pick whatever research that we want to show, and the second one was metros that are in the same state quintiles AutoSum same group might have very different characteristics. So for the first issue, for example, suppose that we found relate to the strong relationship between the two among those a larger kind of matches that you yoga live in cities. But that doesn't necessarily means that we can conclude that either the city size that derives a stronger relationship was stronger gentrification, fattier, because it could be actually maybe public transit system or urban amenities that played an important role in creating such close association, strongly fast, but also that that kind of, you know, things could be found more often in those kind of figures sentence. So for that kind of approach, we cannot identify the real kind of causal factors. But we are just saying that these kind of things and conclude that this is something so we just don't want to do that. The second thing is that each Metro has unique local contacts that cannot be easily characterized by one or two variables. For example, if you can imagine like Boston, Detroit, Phoenix, Riverside and San Francisco, those metros have quite similar populations at around 400 500 1000s. But they're very different in many ways, right? So we just don't want to use the kind of one or two variables to kind of classify or like 366 MSA is, so we wanted to use more kind of, you know, more systemic or subjective way to categorize them, which is called principal component analysis and cluster analysis. That's kind of tech ecard stories. So long story short To

Shane Phillips 30:01
Create that,

Unknown Speaker 30:02
Yeah, or have the most like variables into the model. So, we did that we got eight clusters, but in doing so, we sometimes add in some variables or remove some of the variables based on your comments thoughts by just customers and viewers but groupings are pretty robust. So he said like, such kind of groupings or clusters in the United States some of those metros so to us. It was interesting to see in large coastal metros that included largest cities in the nation that yoga ladc that have lots of well educated high income people and have high housing costs. And then you know, large southern Midwestern metros have pretty similar characteristics to the largest coastal but to a lesser degree. So these two stands out. And the among the rest of them three Metro cluster showed very kind of distinctive characteristics such as Inland Empire, and Mexico, Texas border metros have lots of the Hispanic people and higher poverty rate. And retirement destinations are those Florida cities or cities in Arizona have senior population in college towns or something places some places like Blacksburg or was another harvest house. And three other categories was just kind of like Southern Plains mountain, and small Midwestern not only kind of similar characteristics, although they were kind of geographic, Carly kind of concentrated in a certain region. So that was kind of interesting to see.

Shane Phillips 31:37
And really quickly, just to clarify, you cluster these before looking at displacement outcomes, right? So the displacement outcomes, were not driving the selection or the grouping of these metro areas?

Hyojung Lee 31:50
Not at all. So yeah, so we wanted to exclude any kind of displacement or residual mobility related variables from our modal that these are kind of not giving a certain New Jersey outcome variable.

Shane Phillips 32:04
And then to take a step back here, I think it's important just to think about or discuss, like, why would we maybe expect different effects of gentrification on displacement in different kinds of places? Like what do we think is different about large coastal cities or planes, mountain cities or retirement destinations, that might lead to more or less displacement in those locations when gentrification occurs?

Hyojung Lee 32:31
So when we think about these kind of different metropolitan areas, we thought they may have very different and distinctive characteristics in terms of their economy in terms of their Metro amenities in terms of temperatures you see and you guys led Southern California I mean eastern part

Michael Lens 32:50
Raining a lot here though. Feel bad.

Hyojung Lee 32:56
Rainstorm or something like that.

Michael Lens 32:58
Yeah, there there are actual storms, I want my I want my money back.

Hyojung Lee 33:04
But on the other hand, you know, metros are the lover that constitutes local labor and housing markets. And that means that people can change their jobs, you don't change your place of residence, or they can find a new home we need a reasonable commuting distance without changing their jobs. And in this kind of situation, metros economy characteristic can influence certain mobility decisions, or their influence their neighbors, mobility decisions by changing housing prices, putting more burdens on housing, leaving a cause is interested trust. So we thought there could be very close relationship between gentrification and rigid mobility Metro.

Shane Phillips 33:46
Okay, yeah, that makes sense. Another thing I found interesting is that you didn't just categorize neighborhoods as either gentrifying or not gentrifying. Instead of that binary. You've got a spectrum of three possibilities. Not gentrifying moderately gentrifying, and intensely gentrifying. Tell us how you measured that gentrification. You know, in a little more detail. We talked about socio economic upgrading, but I think we're talking very specifically about education here. And then what is your dividing line between a place that is moderately gentrifying versus intensely gentrifying?

Unknown Speaker 34:21
Sure. So, as you sat Kristin, I conceptualize gentrification as influx of residents with higher socioeconomic status into low income and Central City labor. So there'll be kind of denominator or universe that you choose to to feel neighborhoods, and among them, there could be moderately and intentionally gentrifying neighborhoods, right. So the first thing that we did was to define a Chintu feasible neighborhood or a neighborhood that is at risk of gentrification as low income and central urban census tracts with There are two kinds of criteria. So one is low income neighborhood. Second one is central urban census tract. So for the first condition, we define census tract as low income when it has median household income in the bottom 40% of income distribution within a certain metropolitan area. And then to be considered like urban central neighborhood, that must be really largest priests per city in each MSA is an or second or third largest city, if that has any 100,000 residents. Okay. So for example, if it is new Metro, it'd be New York, you are in White Plains. It is Los Angeles Metro, it'd be in Los Angeles, Long Beach in Anaheim. So they are considered as urban.

Shane Phillips 35:47
Yeah. But I do think that's important to know, because that's leaving out tons of places. And you know, maybe we can assume that there's, this is sort of generalizable that the results in LA Anaheim and Long Beach can be, you know, generalized to Inglewood and Santa Monica and Glendale, but like, maybe not, but you had to draw the line somewhere.

Hyojung Lee 36:08
Exactly. So the first response after seeing these kind of lists was, what about Pasadena? What about Santa Monica in July, I thought they are a kind of an urban area. But somehow you wanted to use a process that is consistent with previous studies that they basically analyzed urban that way. And we wanted to apply that to all metros, so we couldn't make any exceptions or exclusions. And among those kinds of changes, variable neighborhoods or labels and risk stratification, you classify a trap as trying to find when that neighborhood experiences substantial increase in its within MSE percentile rank of college graders share, from time to time to see some of the listeners to see what I mean, it is not nearly increase in college graduate share, but on improvements in its illative status. We didn't MSA in terms of cars graduate share,

Shane Phillips 37:09
Right. So like if the entire metro area has a rising share of college graduates. And your neighborhood is sort of on track with that same rate as the Metro overall, you're probably not changing your rank. And that's what you're measuring. But if the college graduate share in your neighborhood in your census tract is increasing at a faster rate than the metro area as a whole, then you might qualify as gentrifying, if you're reaching one of these two thresholds that you said,

Unknown Speaker 37:37
Exactly. And if a neighborhood ranking improves by, say, five to 14 percentiles, it will be classified as moderately gentrifying. And if that improves by 15% cars or more, it will be an investigation defined, and that had a 15% higher threshold was selected, because that was the standard deviation. And we did this kind of analysis. So there was kind of threshold. So if a certain neighborhood status improved from a fourth percentile to 60th percentile, that'll be intensely centerfire. Neighborhoods, from 20 to 30, is tend to be moderately easy to find. So that's how we classify them.

Shane Phillips 38:17
And I think moving that percentile, you know, 10, or 20 percentiles in a 10 year period, just to be clear, is pretty significant.

Michael Lens 38:25
Yeah, that's a lot. And I think, you know, one thing that this gets at maybe a little bit indirectly is just how hard it is to time a gentrification process as an independent variable or even a variable, you know, because we don't, it's a very dynamic process. Neighborhoods change at different paces, at different times. And I think by trying to capture as you do the intensity of gentrification, you're hopefully doing a little bit better job than some previous analyses at catching a neighborhood when it is actually gentrifying. Because you're not just using this binary measure, you're just a little bit more nuanced that hopefully gets at the timing a little bit better to

Unknown Speaker 39:09
Exactly so one thing that we want to mention is to be tested the robustness of this definition of gentrification by changing low income thresholds in other ways, and changing that kind of BA plus share to house prices or some other measures

Shane Phillips 39:24
BA plus being bachelor's degree or higher education.

Hyojung Lee 39:27
Thank you. You can find it in our crazy table, the findings main findings, one pretty consistent definition. So maybe people think very seriously about how we define gentrification in this way in that way. But the key thing is whether we can find a consistent findings across different definitions or not, rather than this definition is best capture the idea of education versus

Shane Phillips 39:53
so we talked about how, you know, residential mobility is is an umbrella term really it captures things that we would consider displacement but also things we would just consider good, you know moving for a new job moving to move into a bigger home a nicer home, a neighborhood you prefer to live in whatever, I'd like to hear a little bit more about how you measured residential mobility, and at least tried to isolate the quote unquote, bad moves are the displacement in those results, you did a few things to try to identify those moves specifically, even if you can't do it precisely. And so tell us a little bit about the choices you made and the maybe filtering of certain moves where you didn't consider them versus the ones you did.

Hyojung Lee 40:35
So as I mentioned before, unfortunately, we could not distinguish good and bad moves. But what we could identify was whether a person who took a different neighborhood within a year, so for example, if we use the 2011, ACC, or we can identify these, those people who are intensifying it neighbors in 2010. And those people who are Tenterfield, who have non gentrifying neighborhoods in 2010, and what happened to them in 2011. So comparing these kind of two groups, if you find people in gentrifying neighborhoods, but more affluent than those people, not just to find neighbors, we can say that there's a passive relationship between occasion and reasonable pity that that was our kind of regular default outcome. And although we could not distinguish those bad moves from good moves, have, we also use some other mobility outcomes that could include potentially some bad moves. So for example, we could identify whether a person moved to a different city or different Metro, assuming that kind of that kind of long distance moves might have more disruptive effects, especially for those kind of vulnerable people. And then we also measure neighborhood Porter rate and median household income, assuming that if somebody moved to a neighborhood that has a quarter rate greater than his or her origin neighborhood, or moved to a neighborhood, lower household income, compared to their ordinary, then it may we indicate a downward mobility. So we examine the relationship between education and mobility, not only in terms of quantity, like mobility rate, but also in terms of its quality, whether they move to worse neighborhood.

Shane Phillips 42:22
I think that's a really important measure. And, you know, someone might be thinking, well, sometimes people move to poor neighborhoods, and it's not a bad thing. But I think when you're finding or if you were to find that those kinds of moves were more common in gentrifying neighborhoods, either monetarily or intensely than in non gentrifying neighborhoods. That might be some sign that actually these are moves made under duress or you know, involuntarily where they couldn't afford to live in the place they were in anymore, or got evicted, or whatever, and had to move to somewhere a little less expensive. I did want to call out also, I don't think you mentioned this, but my recollection is that you're really only looking at moves for people who don't have a bachelor's degree. And so you're looking again, at kind of a less educated and presumably lower income, a little more vulnerable population, and not so much worried about, you know, the moves of people who have bachelor's degrees and above who probably have a little more options and aren't so vulnerable to involuntary displacement.

Hyojung Lee 43:27
Exactly. So we restricted our sample to those people 25 And older with less than bachelor's degree, because we think, you know, gentrification is was having much more greater impact on those people with fewer resources on one hand, but on the other hand, we test his test at that his idea as well, were renters, low income people, or just over 25 and older population, including those people with bachelor's degree, but readers are pretty consistent. So it seems like gentrification do have more destructive impact on or vulnerable people with fewer resources. It's true, but the results are pretty consistent, even when we include those people as well.

Shane Phillips 44:07
Yeah, yeah. Something I didn't think about there was how, how, even if you are displaced involuntarily, but you have a bachelor's degree or an advanced degree, then you know, it's still bad. But you probably are able to recover from that a little more easily than someone with less resources just on average. Again, this is not going to be true for everyone. But I think it's important to note. So something that we've talked about on this podcast before is how a lot of attention in the media and I think in the research community is given to gentrifying neighborhoods, but there are many more neighborhoods across the country that are actually becoming more poor or more segregated over time or both. To put that another way, if you pick a random neighborhood on a map of the US, it's not very likely that it's gentrifying. And that's not to dismiss the problems occurring in gentrifying neighborhoods, especially Here's someone living in LA where this is really a problem. But I'm noting it here partly as corrective but also to introduce this question, to give us a sense of scale. How many neighborhoods in your study were classified as gentrify rubble? And how many did you not include in your analysis? at all because you didn't consider them gentrify Abul during the study period. And then among those identifiable neighborhoods, what share of them fell into each of the categories of non gentrifying moderately gentrifying, and intensely gentrifying? I just want to be sure that we have a sense for how much of the country is really being analyzed here? And how much isn't even really concerned with gentrification and this may be experiencing totally different problems?

Unknown Speaker 45:49
Sure. So I definitely totally agree with you on that. And our research actually shows that the majority of neighborhoods have not experienced gentrification. Actually, one of Kristin's papers shows that the most common trajectory for neighborhood is just stability, kind of no change. And yes, maybe there has been too much attention paid to those gentrifying neighborhoods or permanent scars. Given that these commodify neighborhood has long been experiencing lateral issues such as underinvestment instability, and some social inequality and segregation with not our analysis shows that it is not an insignificant number of laborers had experienced certification. So let me throw some numbers here. So in the 2010, census, geography, there's about 73,000 census tracts or neighborhoods, and about 60,000. Among them are within 366 MSA. So 660,000 neighborhoods are in metro area, like 80%, right? Sure, about 80%. Yes. Among that about 12,500, or 13,000. Census Tracts are about 20% of those neighborhoods, MSA are considered to be tend to feel meaning that they're located in urban areas near low income neighborhoods. And as you said, you know, those neighborhoods will be our geographical interest or denominator, universe, whatever we can say, right. And among those tend to feel the neighborhoods of us 36%, or 4500. Census Tracts were identified from 2010 to 2019. And another 64% were 8000 neighborhoods, or 20. feasible, but not too far. So about 36 versus 64. And then 36% can be divided into two groups. So 21%, that is moderate to 25. And 15% is in tessellation to five. So that's that's the basic idea that we can have. So I know there are a lot of the numbers but hope that just kind of give us how many neighborhoods we're trying to feel born in 2010. And how many more attributes are defined to independence?

Shane Phillips 48:02
Yeah, there were enough steps there that I think I lost, like the final rough calculation. But right, maybe you could tell me, so we started with they're just 72,000 census tracts period, right? Or 70? Something 1000? How many census tracts Did you find were moderately gentrifying? And how many were intensely gentrifying?

Unknown Speaker 48:24
So there's a lot of 73,000 in the United States as a nation. And then 60,000. In matches,

Shane Phillips 48:31
let's just let's just skip, skip from the 73,000 or 6000. I keep forgetting the number. Let's just skip from that to just the number of tracks that were gentrifying in both kinds if, if you have those figures, I'm just like, is it I'm struggling? Is it 5% In each of the intensive gentrifying categories, or and we're going to delete, like what I'm saying right now, just to be clear, or is it 10 or 15%? Like, what does it add up to?

Hyojung Lee 49:00
Okay, so as I mentioned, are about 73,000, nieghbors. Nation. And originally, there was about 4500 census tracts that was 25, from 2010 to 2019. That gives us 6.2% of nature's neighbors were identified in our definition. Okay.

Shane Phillips 49:18
Definition. And yeah, as you say, that's not an insignificant number. And as a reminder, we're only looking at those urban census tracts because the only ones you even considered identifiable. And in the case of Los Angeles that's excluding anything lower than the third largest city so Pasadena, Santa Monica, Glendale, Inglewood, all these places would not yet even be considered, but they might actually be gentrifying, we just don't, they weren't part of this analysis.

Unknown Speaker 49:47
Right. So you can take it as a very conservative kind of process to restrict our sample to rear or urban area. But actually what Kristin and I are doing is analyzing what happened a lot happened in suburban U turn. So for several gentrification, so stay tuned. Research. Great.

Michael Lens 50:05
So one thing that's interesting as a researcher to me about this paper is that you can't just grab public census data and get information on residential mobility like this. So can you talk about what sorts of data you used to identify movers and and what you can do with those data?

Hyojung Lee 50:30
Yeah, and the data set that we used was select kind of unique, it's in toner version of the American company survey can be only access from what it's called, like federal status corps research theater centers, and it's long buy by some authorized researchers. So, in that regard, I must say the research and conclusions mentioned here in this podcast is ours and to necessarily reflect the views of Census Bureau and those reserves discussed here, whether they are showing off paper appendix or not, or it should be reviewed by the Census Bureau's disclosure Review Board. And I wanted to say this, for the record. So for us access to data was necessary as internal ACS data provide kind of detailed geographic information regarding current residence or or individual hours and previous wisdoms or recent movers as census block level. And that kind of accuracy was the key for this study, as public version of the micro data set called ACS poems, or public use micro data sample has the same Bureau for but at primmer level, warm Public Use Microdata area level, it is certainly greater than zip code, or sometimes even greater a county for some rural areas. So it has been not possible to conduct this type of neighborhood lavar analysis using the ACS public data set, but it became possible with the internal and confidential version of the ACS data. Nice.

Shane Phillips 52:04
So I think we have now you know, pretty far into this at this stage enough that we can talk about results, and have the mean something to our listeners and not be misinterpreted as meaning something different. You did this analysis for the entire nation as a single sample, but then also for each geographic cluster. What did you find for the national sample? And what were some of the highlights for the different geographies?

Hyojung Lee 52:28
Sure, so let me first tell you what we found in our national sample. So when we run the regression using the aggregate national sample 366 metropolitan area, you found that the probability of moving to a different neighborhood out of is about point four percentage points higher among those people in moderate age and defined neighbors compared to non gentrifying neighbors, and about one percentage points higher on those people who are in intensive gentrifying neighborhoods, controlling for everything. And we also tested some other mobility outcomes that I mentioned before, right? So first one, we tested whether gentrification is associated with short distance versus long distance moves. It turned out that gentrification is especially intense gentrification is associated with moves across in the same boundary. So longer distance moves. Exactly. So somebody laying moved to San Francisco,

Shane Phillips 53:28
Probably not that move specifically. Involuntary they're not going to move to an even more expensive place. But Las Vegas I take your meaning, yeah. Let it ride,

Michael Lens 53:41
Go to Las Vegas.

Hyojung Lee 53:43
And additionally, we have seen whether gentrification is associated with destination already wanting our mobility. Our reserve basically shows that movers from moderately and intensely gentrifying neighborhoods ends up in higher poverty rate, higher poverty neighborhoods than their origin neighborhoods, saying about seven percentage points or 1.4 percentage points higher. And in terms of median household income movers from gentrifying neighborhoods tend to move to they would be lower. So both variables indicates that those people from gentrifying neighborhoods not only have higher probability of moving, but also, you know, higher product or downward moving. So, both quantity and quality of mobility, they're worse.

Shane Phillips 54:34
And how about for some of these individual geographies? Are these clusters of geographies? Like how did the results sort of align with the national results and how do they differ in some places?

Hyojung Lee 54:47
Right. So and then we examined whether that kind of association between the two varies across different types of metro areas, right. So in the large coastal metros, like LA New York, Chicago, Chicago, it's not cool. surplus to large. The prior to moving was somewhat similar with national numbers. It is understandable they had lots of the population that can drive most of the things that we found in national estimates. And then estimated clip sets were slightly higher in large Southern and Midwestern metros compared to those large coastal metros. So they are pretty similar, a slightly higher,

Shane Phillips 55:23
Slightly more displacement given you know, whatever level of gentrification

Hyojung Lee 55:27
I will say slightly more poor to moving and certainly more probably all our mobility and then in return destination and Hearthstone Metro, so somewhat, you know, distinctive patterns as motor gentrification is associated with moving to a different neighborhood to confi percentage points and more personal cars, that kind of large number. But gentrification was not necessarily associated downward mobility in terms of poverty and median income in those areas. So they move more often if epsilon is in 25 level, but that doesn't necessarily lead to tau and in southern plains money and small Midwestern metros, you know in test gentrification matter and was associated with the higher profile moving to a different neighborhood, but not necessarily so for moderate future plan. So that was somewhat differences across different type of metros.

Shane Phillips 56:30
Something I want to point out here that I think supports your findings. Is that because you did this spectrum of non gentrifying moderately gentrifying and intensely gentrifying? If the theory that gentrification leads to displacement or at least residential mobility of certain times, then we might expect that places that gentrify more would have higher rates of displacement, or the the rate would increase by a greater amount. And that's pretty much what you find across the board, both in the national sample and in most geographic clusters. And so I think that's just kind of further support, that there is really a connection between these things. If what you were seeing was, you know, the moderately gentrifying places had maybe the same displacement or mobility rates as intensely gentrifying, or especially if they had higher rates of mobility than these intensely gentrifying places, we might start to question like, Well, is it really the gentrification? Because why wouldn't you expect more mobility as gentrification increased, but that's what you found. And I just, I think that's a really important thing to call out here. So thinking about ways people might challenge this research, what do you think about the critique that metro areas are just too heterogenous, to variable for each to be classified into a single group, you got your PhD in Los Angeles, as you said, so you're very familiar with the variety of neighborhoods that you can find here. I'm thinking of neighborhoods like palms, Highland Park, North Hollywood, the community around USC in South LA, all of which are or recently were pretty low income. Those neighborhoods are all part of the Los Angeles metro area. And in that sense, they have a lot in common, but they also each have unique characteristics and demographics, and economic and social pressures that might result in different responses to gentrification. And even though I use la here, as an example, I'm sure the same could be said, for every metro area in the country. I realized that this isn't entirely a fair question or critique, because you've already taken a big step forward in spatial differentiation, let's call it compared to previous gentrification and displacement research. But I'm wondering if this kind of grouping actually creates maybe new analytical problems in some ways and what it might mean for future research?

Unknown Speaker 58:53
Sure. So I do you agree that there could be a wide in trauma to pretend variation, especially in larger metros like New York, LA, Chicago, San Francisco and San Jose. But, Kristin and , I first wanted to make sure that there's some level of you know, spacers, you need across areas, and if there are some spatial heterogeneity than we thought Metro might be the first geographical tasks given that there's a wide variety among metropolitan areas in terms of say size, demographic competition, city, urban form and other dates. And also, you know, demographic and socio economic characteristics of in movers can be very different across different types of metros as someone who for techie jobs, what others moving forward for homes for Roger families and others for their retirement is nice, right? So they and when they first determine which cities to move, and then they compare the neighborhoods within the city, meaning that there'll be more variation across the metropolitan area than within that Your question area in terms of economic characteristics in university, that was our our thinking? Yeah, yeah. But I think it is totally possible that gentrification happening, you say, around the USC area will be totally different from what's happening in Pasadena, or some other areas. So definitely further research will be needed for that kind of thing. And in part, we are doing that as well. So you are doing, you're applying the method that we use for these kind of natural classification to neighborhoods, and see what's happening in different types of the neighbors. So whether it is like expanding enclaves versus working class neighborhoods versus multifamily neighborhoods versus single family?

Michael Lens 1:00:44
So given given the types of metros that you see the biggest and smallest effects? I think there is some evidence there that high or rising rents is playing a strong role. Do you agree with that? And you know, you've touched on some mechanisms, but are there other mechanisms that you think are going on there?

Unknown Speaker 1:01:06
Sure. So about the factors that is associated with the relationship that we found, right? Yeah. So Kristin, and I thought, this papers main contribution would be just exploring the spatial heterogeneity, and the relationship between the two using the most maybe arguably the most comprehensive DRM approach. But on the other hand, we thought it'd be amazing, if you could find the specific factors, right, explaining these kind of differences, but long story short, we could have let me say how we ultimately got matures. So first, basically, what we did was we classify these kind of 366 MSA s into three groups like total, you have 33%, military 4% and 33%. By a start little protect characteristics that we used for PCA analysis, for example, it could be you know, Congresscritters, shear, it could be poetry, it could be, you know, resistance, integration in gas, etc, etc. And our assumption, our thought was that if we find if we can find some strong relationship among top three 3%, MSS versus 33%, MSA is based on these metropolitan characteristics. And it could be an indicator of disease. And it's related to relationships, these could be potential factors around the relationship. But when we run a regression on this kind of regression, we couldn't find any kind of conclusive relationships, but still to share of newly built housing units. And vacancy rates, explain some heterogeneity across areas. And population growth was also show some leadership, but long, and there was some statistical significance. But that was not strong enough to argue in our manuscripts. So we decided just focusing on describing this kind of magic, there's unity, rather than arguing that this is a factor associated.

Michael Lens 1:03:10
Right. And that's a common thing we have to do is to not speak past with the data can conclude right, that's good. Another thing is that we've talked certainly a couple of times about, you know, whether gentrification and displacement are kind of the most common or pressing problems in a lot of urban areas. But there's another piece of of that kind of line of thinking, which is, how strong the effects are here, you know, what is the magnitude of this connection that you find? So if I can interpret at least one of the key findings is that residential mobility rates are roughly one percentage point higher in the most intensely gentrifying neighborhoods. And, you know, I don't want to sound glib about that. But it doesn't seem like a big difference to me, particularly when you consider that a lot of those moves are going to be for perfectly fine reasons are not forced, etc. So after reading your paper, I'm definitely more convinced that gentrification affects displacement. But I'm not yet necessarily convinced that it's like this big driver of displacement. So you know, do you agree with that take or what might I be missing there?

Hyojung Lee 1:04:21
Sure. So as you mentioned, it put an impasse Leeson defined as one percentage points higher productive moving, and others. So it may sound not that kind of substantial, right? But here are two things to consider. So first one is the in our sample, the average will be read rate was about 9.2%. And that makes that one person's points pretty crucial, as that indicates that prior to moving will be 11% higher in gentrifying neighborhoods compared to those people into fear of over nonsense. So just 1% of coin sizes More 11% sounds almost bigger. Yeah. On the other hand, the second thing to consider is that our measure of version mobility is about migration within 12 months or a year. So for those people who answered 2011 ACS, they asked where they lived in 2010. So is a one year morbidity rate. That means that so 9.2% mortality rate, so about one tank of all people, or people with less than bachelor's degree in those low income neighborhoods are moving every year. And to feel better neighborhoods have one percentage points higher mortality rate every year. So that might give some I mean, there might be some frequent movers moving here to there, like churning labels over time. But if we add up that kind of difference over say, five years or 10 years, you'll be able to see a quite substantial differences in relation mobility, or retention between gentrifying and non to define neighborhoods. So that's something that I would argue, thank you.

Michael Lens 1:06:03
As I said in the previous question, I'm more convinced after reading your paper that gentrification is a factor in displacement. With this evidence, what should policymakers do about this problem?

Hyojung Lee 1:06:19
So the team that says that we originally wanted to deliver in this article was that gentrification can be found in many different types of metros. And when you talk about its relationship with racial mobility, work displacement, local context matter. So you should think about how you know substantial naval upgrades will lead to improvements can be translated into the everyday life of resonance in the specific local context. And then there cannot be a one size fits all type of policies that can fix ordinary outcomes on to defecation in all the different types of Naturals. I mean, it may sounds like me a little bit overly cautious, but I think it's better to be extremely careful, rather than be reckless or irresponsible in any kind of policymaking process that can determine some winners or losers. And that kind of local kind of specific policies can be thought of as something like, you know, for example, like recent paper by Tacey, Dragon and Sherry Clyde, basically shows that mobility rates are notably lower understand is living in subsidized housing or public housing, even when they are located in gentrifying neighborhoods, and children into defined neighborhoods expensive, larger improvements in their neighborhood environment, such as lower crime rates or higher school performances. So if we think about if there are some were performing local PTA, or public health agency, and marking mechanism cannot be able to provide like low income family houses and that kind of area, maybe public housing might be a good solution for that specific area. Especially gentrification is associated with downward mobility. For some other cases, if gentrification is not necessarily associated with our mobility, or higher mobility rate, than we might think gentrification might not necessarily be a bad thing. But we should do some more research to find out other than these kind of measures, if there are some other community pass certification in such neighborhoods, such as, like why we talked about right, like economic or cultural displacement, rather than purely social creativity space.

Shane Phillips 1:08:36
If I can push on this a little bit, I feel like definitely the takeaway, that context matters, and we need to take that into account makes all the sense in the world. But I'm just thinking about, like, in practical terms, what we do with that, because I think there's a lot of different ideas about what causes gentrification and what you do about it. And so, and also, there's the fact that gentrification is happening in many places, like, regardless of whether you're making investments, necessarily, a lot of this can be driven entirely by private individual decisions of people moving into or out of neighborhoods. And so, you know, is this about being careful about where we invest public resources? Is it about displacement protections and mitigations? And that kind of thing? I mean, I'm sure the answer is like, it's all these things and more, but like, Could you say a little bit more about how this all fits together? And maybe how we go from like, recognizing that the context matters, and you might need different solutions to like, what we actually do with that information or what kinds of policies or decisions that awareness should inform.

Hyojung Lee 1:09:46
Yeah, I mean, this is a part that I'm saying I'm maybe overly cautious because I cannot argue something for a certain type of neighborhoods because what we found was is clear differences across metropolitan areas, rather different factors. across different areas. So I will say there should be more and more research conducted in this matter. To find out that kind of factors protective factors are determined are mechanisms that drives the relationship between gentrification in regional mobility and in certain areas. So, the reason why I'm saying this is that, in general, researchers have anything more focused on what's happening in big cities like superstar cities, like New York, Los Angeles, San Francisco. But it turned out that some previously overlooked metropolitan areas, smaller metros, or cottage towns or retirement communities also experiences these types of gentrification and potentially displacement issues. So what we are arguing is that these kinds of neighborhoods deserve some attention. And researchers should recognize the difference between different areas, and should do some more research to find out the factors. So that answer your question on Mike's question about what should we do on based on what we found, rather than to saying that you should do this in this kind of neighborhood? Or do that kind of thing? Without knowing anything about to come here?

Shane Phillips 1:11:13
Yeah. And I think at the end of the day, what we're not saying here is we just need to wait for more research to decide what to do in different kinds of neighborhoods, but we're just going to have to keep trying different things, doing our best based on what we do know about neighborhoods, seeing how different kinds of interventions or protections or what have you actually work. And I mean, that's at some level that needs to be done so that people like us can study how effective or ineffective they are. But ideally, with the research we have now, that should inform at least a little bit of the kinds of decisions we're making, you know, knowing that the people moving into a neighborhood are kind of young, upwardly mobile, college educated, versus retirees versus immigrants from Mexico like that should have some bearing on the kinds of policies and programs we put in place to try to protect people and provide some housing stability for them.

Hyojung Lee 1:12:12
Thanks. I like your answer more than mine.

Shane Phillips 1:12:17
Well, with that wonderful compliment, I will. I'll close this up here. hildren Lee, thank you so much for coming on to housing voiceohpodcast.

Hyojung Lee 1:12:25
Thank you, Shane. Thank you, Mike.

Michael Lens 1:12:26
Thank you.

Shane Phillips 1:12:31
You can read more about children's research on our website. lewis.ucla.edu. Show Notes and a transcript of the interview are there to the UCLA Lewis Center is on Facebook and Twitter. I'm on Twitter at Shane D. Phillips and Mike is at MC_lens. Thanks for listening. We'll see you next time.

About the Guest Speaker(s)

Hyojung Lee

Hyojung Lee is an assistant professor of housing and property management in the Department of Apparel, Housing, and Resource Management at Virginia Tech and a research fellow at the Joint Center for Housing Studies of Harvard University. His research has focused on the impacts of demographic change on housing markets, the consequences of neighborhood change for urban policy, and the jointness of mobility, residential location, and housing tenure choice.