Abstract
Neighborhood change is a complex phenomenon that may result in a range of physical, demographic, and economic changes in a locality. Using four case studies of transit neighborhoods in Los Angeles, this study utilizes a mixed-methods approach to examine a particular aspect of neighborhood change—gentrification. The article also compares and contrasts the type of data gathered by different methods to help us understand each method’s potential and limitations in capturing gentrification trends in neighborhoods.
Introduction
Since the term gentrification was first used by sociologist Ruth Glass (1964) in the mid-1960s, a rich literature has emerged of studies that seek to identify the magnitude of change and document its impact on gentrified neighborhoods. While these studies discuss mostly the processes and impacts of gentrification, we are not aware of studies that focus on the methodologies of studying gentrification. In general, a methodological dichotomy characterizes much of the existing gentrification literature, as studies are either quantitative, “macro” analyses or qualitative, “micro” inquiries of neighborhoods in the form of case studies (Hammel and Wyly 1996). But there is often a “disconnect” between quantitative and qualitative approaches that are designed to gather and analyze different types of data, and only few gentrification studies adopt a mixed-methods approach.
The presence or absence of gentrification is often hotly debated by residents, who experience impacts such as increased rents or store closures in their neighborhoods, and policy makers and academics, who use aggregate census data to document neighborhood change and may not see these impacts. This article seeks to show the importance of incorporating knowledge from both quantitative and qualitative methods to study gentrification. But although utilization of both quantitative and qualitative methodologies can paint a richer picture of gentrification, planners and policy makers often do not have the time to employ multiple methodologies. Therefore, we are also interested to understand what each method contributes; under which circumstances planners may be able to identify a “methodological short-cut” in studying gentrification; and which cases require more elaborate and multiple methodological approaches.
The study examines the contribution of three methods to our understanding of gentrification processes in four Los Angeles transit neighborhoods: (1) measures based on secondary data; (2) systematic street- and parcel-level observations; and (3) interviews with representatives from community-based organizations (CBOs) and public agencies active in the four neighborhoods. The goals are to examine the degree to which gentrification exists in the four neighborhoods; compare and contrast the type of data gathered by each method; and understand each method’s potential and limitations in capturing neighborhood change.
The study draws from a larger project that examined the impact of transit investment on gentrification around stations in LA County and the Bay Area, finding that the presence of a transit station was a significant independent variable, but whose effects varied across time periods and neighborhoods. Transit-oriented development advocates often describe the potential of transit stations as catalysts of neighborhood change (Dittmar and Ohland 2004). And indeed we found that, on average, station neighborhoods have experienced more development than nonstation neighborhoods. But they have also witnessed greater increases in white, college-educated, and higher income households and are changing more in the direction of gentrification than neighborhoods without transit stations (Chapple et al. 2017). This systemwide evaluation of station impact offered us an opportunity to test a mixed-methods approach and examine the presence and extent of gentrification in some diverse low-income neighborhoods. Thus, in the present article, we focus on four station neighborhoods from the larger study—Chinatown, Hollywood/Western, 103rd St/Watts, and Mariachi Plaza—to highlight what we have learned from the methods employed.
In what follows, we present a literature review of the methodologies used to measure neighborhood change and gentrification, discuss the study context and methods, and detail findings gleaned from each method. We conclude by discussing each method’s potential and limitations.
Literature Review: Methods for Capturing Neighborhood Change
Neighborhood change is driven by at least three dynamic processes: movement of people, public policies and investments, and flows of private capital (Zuk et al. 2015). The nature and intensity of change may vary across neighborhoods in a metropolitan area because of spatial variations in these factors. An example is the geographic distribution of infrastructure investment such as the building of a transit station.
Neighborhood change may have positive or negative effects for residents. Following suburbanization, deindustrialization, and white flight after World War II, many inner-city neighborhoods in the United States witnessed negative change—sharp disinvestment and decline. Early studies of neighborhood change focused on such decline or “descent,” as well as disinvestment, demographic shifts, and discriminatory practices (Massey and Denton 1993). More recently, following investments to regenerate inner cities, some studies have focused on the consequences of the upward trajectories of neighborhoods, or neighborhood “ascent” (Zuk et al. 2015).
Gentrification is a commonly studied outcome of neighborhood ascent (Zuk et al. 2015). The influx of higher-income new residents may lead to displacement of existing residents. Finding data that allow for the simultaneous measurement of physical, cultural, economic, and demographic shifts ensuing from gentrification, however, can be an arduous task (Benton 2014). Thus, much research has only measured the magnitude of particular aspects or impacts of gentrification (often housing and demographic shifts) and has relied heavily on the US census (Schwirian 1983), described as the “most comprehensive and comparable source of data” on neighborhood change (Hammel and Wyly 1996, 248).
Most quantitative studies of gentrification have taken a “macro,” census-based approach (Hammel and Wyly 1996). Their focus has been the measurement of demographic shifts over a number of years that indicate gentrification, such as changes in the racial/ethnic composition, income, and educational attainment of residents (Barton 2016). More complex quantitative approaches link noncensus data from large-scale surveys to census measures and geographies. For instance, Freeman (2005) links geocoded data from the Panel Study on Income Dynamics (PSID) to decennial census data and analyzes the data at the census tract level to compare displacement in gentrifying tracts to low-income tracts that did not gentrify using characteristics such as location, income, and educational attainment. Bostic and Martin (2003) use a similar approach, aggregating 1970–1990 data from the Home Mortgage Disclosure Act to the census level to study the role of minority groups in the gentrification of neighborhoods. Their research builds on previous work by Wyly and Hammel (1999), who examined gentrification in eight American cities during the 1990s.
Studies utilizing census-based quantitative methodologies to measure neighborhood change have given insights about segregation, population dynamics and “tipping points,” neighborhood life cycles, and neighborhood revitalization and gentrification (Schwirian 1983). However, one shortfall of the quantitative approach is that it does not compare census data with what is actually on the ground, thus not verifying census-based findings (Hammel and Wyly 1996, 248). Additionally, the geography of the census tract is not always ideal to understand neighborhood processes, and aggregated quantitative data at the census tract level may miss subtler changes occurring in neighborhoods.
Qualitative work has usually taken a case study or ethnographic approach to provide an in-depth look of neighborhood change. Such studies typically use a combination of built environment analyses and observations and stakeholder interviews. For instance, in his case study of West Town in Chicago, Betancur (2002) uses interviews, media coverage, and field observations to examine the role of local dialectics of power, class, and race/ethnicity in the process of gentrification. Brown-Saracino (2009) draws from interviews and ethnographic research in Chicago to identify three types of gentrifiers that have different goals and motivations: urban pioneers, social homesteaders, and social preservationists.
Other qualitative studies focus on the opinions of proponents or opponents of gentrification. This includes ethnographic work on the political discourse of diversity in Rogers Park in Chicago (Berrey 2005), and on grassroots resistance to gentrification in San Francisco’s Tenderloin (Robinson 1995). While most studies examine the motivation of middle- and upper-class, white in-movers into communities of color, some studies have also examined the relocation of black, middle-class in-movers to low-income black neighborhoods (Taylor 2002; Boyd 2005; Freeman 2005; Hyra 2008; Moore 2009; Pattillo 2008).
Qualitative studies are often richly detailed ethnographic accounts of neighborhood change. They usually focus on a single neighborhood or small group of neighborhoods experiencing gentrification due to demographic shifts (Barton 2016). This is the case in several studies of gentrification in New York City neighborhoods (Mele 2000; Freeman 2005; Maurrasse 2006). The very small number of neighborhoods examined typically prevents qualitative research from using control measures for comparison with other neighborhoods. Another shortcoming of qualitative approaches is that they generally do not integrate analyses of census data to verify findings from neighborhood-based fieldwork (Hammel and Wyly 1996).
More recently, the volume of neighborhood-level data has increased dramatically. Parcel-level data (from the County Assessors’ office) and business data (e.g., from Dun and Bradstreet) are now easily accessible annually. Researchers can also draw information about neighborhood urban form and its changes from the Street View’s archives of Google Maps, various crowdsourcing data, and the American Community Survey (ACS) annual data. While such information may be time-consuming to compile, and at times lacks consistency and accuracy, or in the case of ACS, uses limited sample sizes, it nevertheless gives researchers the capability of developing annual neighborhood profiles to assess neighborhood trajectories. Thus, a third line of research links census and non-census quantitative data with qualitative data to triangulate information about neighborhood change. An example is the work by Hammel and Wyly (1996), who groundtruth census reports of neighborhood upscaling using field observations of visible housing reinvestment, and the work by Sampson (2012), who collected extensive street-level data from observations of neighborhood social and physical disorder in Chicago. Chapple (2009) also uses a mixed-methods approach to map the Bay Area’s susceptibility to gentrification, employing first quantitative analysis to explore the link of gentrification to the presence of different factors relating to neighborhood location, sociodemographics, housing characteristics, and neighborhood amenities. Based on this work, she later develops an “early warning toolkit” for gentrification, and uses qualitative research to test it in the Lake Merritt neighborhood. Similarly, Hwang and Sampson (2014) examine gentrification in Chicago using a diverse array of quantitative and qualitative data, including census-based indicators, police records, community surveys, city budget data on capital investments, and built environment observations from Google Street View.
The methodology presented in this study draws from this third line of research. It seeks to understand neighborhood change, using a mixed-methods approach that combines secondary data from the census and administrative records, qualitative observations of residential and commercial ascent, and perceptions of gentrification by stakeholders, often contrasting these data to test their accuracy.
The Context: Four Los Angeles Station-Neighborhoods
The study uses different methods and data sources to evaluate gentrification in four Los Angeles transit neighborhoods, focusing on the area within a half-mile radius from their station. Gentrification is a hot-button issue in Los Angeles, at a time when housing affordability challenges in the city and in California are “as bad as they’ve ever been” (Dillon 2017). The large majority of new construction takes places around transit stations, where the city has decided to strategically concentrate its new and higher-density development (Loukaitou-Sideris 2007). There is, thus, considerable concern that gentrification is lurking at many transit neighborhoods in Los Angeles.
Los Angeles and its transit neighborhoods represent then a natural selection for studying gentrification pressures. For this study, we chose four station-neighborhoods, which we considered vulnerable to gentrification, because they are occupied by high numbers of low-income/low-education and minority populations. Diversity of station-area conditions also influenced our selection of these neighborhoods, since each of them represents a different land use and demographic mix, and stations (belonging to different lines) were introduced into the neighborhoods at different times (see Table 1 for summary statistics of each neighborhood).
Summary Statistics of the Four Neighborhoods.
Source: 2009-2013 5-year American Community Survey, aggregated to the block group, area weighted for a half-mile radius; prevailing rent and home values are from Zillow.com for the zip code station falls in; 2013 dollars.
Chinatown is a mixed-use, medium-density, ethnic neighborhood north of downtown Los Angeles. Although predominantly an Asian neighborhood, Chinatown also has Latino, black, and non-Hispanic white residents. Confined into an ethnic enclave by legislation and racial backlash, early Chinese merchants developed family-owned “mom-and-pop” stores. Today, many of these small businesses continue to cater to the shopping needs of residents and visitors, but shopping centers and mini-malls have also popped up over the years. Community groups believe that the area is currently experiencing gentrification as they see transformations, including the loss of traditional businesses and the development of new housing options, public services, and activities that are inconsistent with its historic identity (Figure 1) (Mai and Chen 2013; CCED 2015).
New Transit-Oriented Development (TOD) housing in Chinatown, 2016.
Hollywood/Western is a mixed-use, regional destination in East Hollywood. The neighborhood surrounding the station is one of the most densely populated in Los Angeles. Beginning in the 1960s, immigrants from East Asia, Latin America, the former Soviet Union, and the Middle East started settling in. Today, the neighborhood is home to a diverse mix of residents, with non-Hispanic whites composing the largest racial group (51 percent), but also hosting the ethnic enclaves of Little Armenia and Thai Town. The neighborhood has a substantial stock of multifamily housing and has witnessed significant development in the last decade (Figure 2), which raises concerns about gentrification.
New development at Hollywood/Western, 2008.
103rd St/Watts is a low-density neighborhood in South Los Angeles (Figure 3). The area in the half-mile radius around the station is a residential commuter district, about 13 miles south of downtown and away from other large employment centers. For years a disinvested and poor African American neighborhood, largely affected by the 1965 Watts riots and their aftermath, Watts has experienced significant demographic transition in the last decades. Presently, the neighborhood has a Latino majority (74 percent) and an African American minority (25 percent). The desire to promote local economic development by the public and private sectors in the wider South Los Angeles area brings to the fore the prospect of gentrification (Watts Community Studio 2013).
103rd/Watts Towers: View from the station, 2016.
Mariachi Plaza is a mixed-use, relatively low-density, ethnic neighborhood in Boyle Heights, East Los Angeles. Its station is located directly adjacent to Mariachi Plaza, the commercial center of the area, which since the 1930s has served as a gathering place for mariachi musicians looking for work (Figure 4). The neighborhood has historically been home to different ethnic groups, but today, it is predominantly Latino (87 percent) and working-class. The station is surrounded by restaurants, stores, and the Boyle Hotel, one of the oldest commercial structures in Los Angeles (Los Angeles Conservancy 2016).
Mariachi Plaza: View from the station.
Methods and Data Sources
In our larger study, we used census tract–level information to create a multivariate model to examine residential gentrification in Los Angeles County (Chapple et al. 2017). We classified all Los Angeles County tracts as either vulnerable or not vulnerable to gentrification based on certain socioeconomic characteristics. Indicators of vulnerability included household income, educational attainment, percentage of renters, and race characteristics. Table 2 shows the criteria used to classify a census tract as vulnerable to gentrification in 2000, and those used to identify which of these vulnerable tracts were considered as gentrified or gentrifying in 2013. Of course, not all low-income/low education and high-minority neighborhoods are equally vulnerable, as other indicators (e.g., significant or historic architecture, proximity to natural features, etc.) may make a neighborhood appealing to gentrifiers. However, one of the weaknesses of census tract–level data is that it does not give much information about urban form characteristics or the residents’ lived experiences, which are better captured by qualitative research.
Criteria for Tract Classification as “Vulnerable” or as “Gentrified or Gentrifying.”
Source: Chapple et al. (2017).
Table 3 and Figure 5 show the results of the gentrification model for the four neighborhoods. The model did not show gentrified or gentrifying tracts in 103rd/Watts and Mariachi Plaza, even though it depicted both areas as vulnerable to gentrification. The model indicated that Chinatown and Hollywood/Western have undergone some change in the past decade. Most of the change in Chinatown can be seen along the outskirts of the half-mile radius, while in Hollywood/Western change has occurred near the transit station.
Number of Census Tracts Partially or Completely Within Half Mile of Station, 2000–2013.
Gentrified Census Tracts, Los Angeles County, 2000–2013.
To complement our model and get a better idea if and to what extent gentrification is happening, we compiled secondary data for the four neighborhoods, conducted field surveys collecting visual information from each neighborhood, and interviewed representatives of local CBOs and public agencies active in the four neighborhoods.
Secondary Data
We acquired and analyzed various secondary data sets to track neighborhood change along the following three dimensions associated with gentrification: sociodemographic shifts, changes in the building stock, and changes in housing affordability. Table 4 lists the type of secondary data collected.
Type of Secondary Data Collected.
Note: SF = single-family; MF = multifamily; HCV = Housing Choice Voucher Program; LIHTC = Low-Income Housing Tax Credit; HUD = US Department of Housing and Urban Development; ACS = American Community Survey.
To track changes in the neighborhood building stock, we acquired parcel-level Assessor’s data. To determine if a single-family property had a major renovation, we examined changes in its improvement value from 2007 to 2013. California caps property taxes at 1 percent of the assessed value of a home at the time of purchase and prevents taxes from increasing more than 2 percent annually, or more than the rate of inflation, whichever is less, unless there is a sale or major renovation. Anything above this indicates an improvement or renovation. We considered a residential property as experiencing major renovation if the percentage change in its improvement value was greater than the rate of inflation from 2007 to 2012 (10.7 percent), and the amount in real dollars of improvement was greater or equal to $5,000.
We examined change in affordable rental units from 2000 to 2013 using the decennial census and ACS for block groups partially or completely within the half-mile radius. We defined affordable rental units as those with median gross rent of less than 80 percent of the 2000 County median. For 2013, all units below the 2000 baseline were considered affordable. The difference between 2000 and 2013 was normalized as a fraction of the housing stock (divided by total housing units) in each year.
We calculated the change in Section 8 housing voucher recipients from HUD’s Picture of Subsidized Housing database for 2000 and 2013, as well as the number of Ellis Act Evictions from 2007 to 2014 as documented by the City of Los Angeles Housing and Community Investment Department. The Ellis Act allows landlords to evict tenants, if they change the use of their building (e.g., from rental units to condos). Lastly, we used the Low-Income Housing Tax Credit (LIHTC) database to calculate the change in low-income units between 2000 and 2013.
Visual Surveys
We conducted systematic visual surveys of the four neighborhoods, making an inventory of visual indicators, and using them to document change. Two trained researchers per neighborhood collected these data on predesigned recording sheets (see Appendix B). We conducted a series of trial observations to check the appropriateness and validity of the forms and interobserver reliability (Landis and Koch 1977). Researchers walked around the neighborhoods, documenting observations at the street-segment and parcel levels, and photographing each block and parcel aiming to capture a variety of indicators of possible presence or absence of gentrification. The methodology for observing built environment changes was, in part, adapted from Hwang and Sampson (2014); however, our observations included both commercial and residential parcels, and also sought to document urban form elements explicitly linked to demographic shifts, upscaling, and presence or absence of social disorder. Appendices A to C (online supporting information) show the visual survey instruments.
We observed a total of 116 residential and commercial parcels and 84 street segments in the four neighborhoods. We surveyed all blocks immediately adjacent to the station, within a quarter-mile radius and in each cardinal direction, and a smaller sample of blocks within the half-mile radius (but outside of the quarter-mile). For parcel-level analysis, we mapped and visited all parcels with new construction, renovation, condo conversion, or sales to single-family homes, multifamily buildings, and commercial properties between 2008 and 2013. Table 5 summarizes the type of data collected through visual surveys.
Visual Observation Data Collected.
The data listed on Table 5, and in particular the type of land uses, the appearance and condition of buildings, the newness of building stock, and signs of construction, renovation, and property flipping (e.g., “for sale” signs) offered visual clues regarding the presence and extent of gentrification. We coded these particular visual indicators and mapped where and to what extent they occurred in the transit neighborhood. These visual data helped to compare gentrification signs among the neighborhoods, understand where gentrification occurs (within the quarter mile or further away from the station), and also groundtruth the secondary data (see below).
Interviews
The people who live or work in a neighborhood are the first to notice neighborhood changes. We, therefore, complemented the secondary data and field observations with interviews with representatives from local CBOs and public agencies active in the four neighborhoods. The part of the interviews relevant for this article centered on the respondents’ perceptions regarding neighborhood change and gentrification, and how the station has impacted the transit neighborhood.
We conducted a total of seventeen semistructured interviews in the four neighborhoods, with representatives from different public agencies (Los Angeles Department of City Planning, neighborhood councils, and city council offices) and ten CBOs active in the neighborhoods (Strategic Alliance for a Just Economy [SAJE], Southeast Asian Community Alliance [SEACA], Chinatown Community for Equitable Development [CCED], Thai Community Development Center [Thai CDC], East Los Angeles Community Corporation [ELACC], Watts Community Studio [WCS], Trust for Public Land [TPL], Union de Vecinos, Los Angeles Alliance for a New Economy [LAANE], and LA Voice). Public agency interviewees were staff from agencies who have worked on projects in the four neighborhoods. We located interviewees through Internet searches, referrals from colleagues, and recommendations from other interviewees. Interviews lasted 30–45 minutes, were completed in person or by phone, were recorded, and followed a pre-prepared questionnaire (Appendices D and E, located in online supporting information). We coded the complete interview transcripts and clustered codes expressing similar themes to form categories of responses. Codes indicating gentrification included “rising rents,” “changing neighborhood character,” “new people,” “hipsters,” “new stores,” “new development,” “condos,” “property flipping,” “tenant buyouts,” “business turnover,” “evictions,” and “displacement.”
Findings: What Did We Learn from the Three Types of Data?
Secondary Data
Table 6 summarizes the collected secondary data; it shows changes in the sociodemographic characteristics of the station-neighborhoods, and how they compare with LA County and All-TOD averages. Data was aggregated to the block groups that fall partially or completely within a half mile of the transit station. All dollar values were adjusted to 2013 dollars.
Sociodemographic (Percentage Point) Changes in the Four Neighborhoods: 2000–2013.
Source: 2000 Census, 2009–2013 ACS 5-year ACS; aggregated to the block group level.
Note: TOD = Transit-Oriented Development; LTHS = less than high school; ACS = American Community Survey.
As shown in Table 6, from 2000 to 2013, LA County experienced a decrease of 4 percentage points in white population, while the four neighborhoods either witnessed a slighter decrease (Mariachi Plaza), no decrease (103rd/Watts), or an increase in non-Hispanic white residents (Hollywood/Western, Chinatown). The percentage of low-income households in all four neighborhoods dropped significantly more than the LA County average and consistent with all-TOD averages, while all four neighborhoods saw a modest rise of high-income households, in contrast to the County trend. The mean gross rent increased in the four neighborhoods, but by less than that in the County and all-TOD areas. Consistent with the County and all-TOD areas, the percentage of rent-burdened households increased in all four neighborhoods.
Table 7 summarizes changes in the neighborhood housing stock. Conversion of apartments to condominiums may signify a gentrification trend, but as shown in Table 7, this only happened in significant numbers in the Hollywood/Western neighborhood. There was very little new construction of single-family residences within the quarter-mile area of all neighborhoods; however, 103rd/Watts witnessed the construction of seventy-two single-family houses within the half-mile area. There was a more significant addition of multifamily units in three neighborhoods, mostly outside the quarter-mile but within the half-mile radius from the station.
Changes in Neighborhood Housing Stock.
Source: Assessor’s data.
Note: MF = multifamily; SF = single-family; SFH = single-family home.
Table 8 summarizes changes in neighborhood affordability. The four neighborhoods experienced significant drops in the percentage of their affordable housing stock from 2000 to 2013, quite higher than the 13 percent reduction in affordable housing units experienced by the County as a whole. During the same time, the increase in their Section 8 housing was very small (0.5–3.6 percent). With the exception of 103rd/Watts, the other three neighborhoods saw increases in LIHTC units but these were not large enough to offset the total loss of affordable rental units. Lastly, the number of Ellis Act evictions (for condo conversion) was miniscule (0–6 units).
Changes in Housing Affordability.
Source: Decennial census, ACS, HUD, LIHTC database, City of LA database.
Note: TOD = Transit-Oriented Development; HCV = Housing Choice Voucher Program; LIHTC = Low-Income Housing Tax Credit; HUD = US Department of Housing and Urban Development; ACS = American Community Survey.
Field surveys and visual observations
The gentrification model shown in Figure 5 indicated that Chinatown has experienced some change along the outskirts of the half-mile radius around the station, but not close to the station, where most of the commercial parcels exist. Street-level observations also did not reveal signs of commercial gentrification in Chinatown, finding primarily older, established businesses and no signs of trendy new stores (boutiques, yoga studios, high-end grocery stores, or galleries). Some commercial parcels had minor cosmetic renovations (e.g., fresh paint). Observations captured, however, some signs of residential gentrification that were not evident from the secondary data. Chinatown had the highest prevalence of new construction on residential parcels among the four neighborhoods, which may indicate a quickly growing residential market. About one fourth of residential parcels had upscale landscaping, and one fourth were newly renovated.
For Hollywood/Western, the gentrification model showed that only the area southwest of the station has gentrified in the last decade, while the area to the southeast has undergone little development (Figure 5). The tracts north of the station were not considered vulnerable to gentrification. Observations, captured more signs of commercial and residential gentrification than those found by the model. This area had the highest percentage (15 percent) of new construction in the commercial blocks surveyed, while about 15 percent of the blocks had minor or moderate renovations. Two new commercial multistory buildings were out of context in size from the surrounding parcels. Compared to the other three neighborhoods, Hollywood/Western had the highest concentration of “hipster” establishments, with a yoga studio, a specialty food shop, a Starbucks, a Crossfit specialty gym, and other brand-named retail stores. Additionally, Hollywood/Western showed multiple signs of residential gentrification. About 20 percent of the blocks surveyed had new constructions, which was the highest among the four neighborhoods, and about 40 percent showed signs of renovation. Half of the blocks observed had upscale landscaping, the most among the four neighborhoods. Of the residential buildings, 9 percent were new, 27 percent renovated, and 36 percent had ongoing renovations.
For 103rd St/Watts, the gentrification model did not find gentrification around the station. Observations were consistent with this finding, showing no evidence of commercial gentrification and only limited commercial land uses within the quarter-mile area. Only about 6 percent of the surveyed block segments had signs of new commercial construction and mostly minor renovations. The few new commercial properties housed mostly small mom-and-pop stores. Only one block had predominantly commercial/retail uses catering to a lower-income demographic. Similarly, we found minimal signs of residential gentrification. While Assessor’s records show a high amount of transactional activity in residential parcels, a change in ownership has only occasionally resulted in the improvement of a parcel’s appearance and minor cosmetic renovations. Only four blocks had new residential construction, while the great majority of new residential structures were outside the quarter-mile radius, but within the half mile from the station, and were similar in quality/appearance to their surrounding residential structures.
For Mariachi Plaza, the gentrification model did not show gentrification. However, observations pointed to signs of early residential and commercial gentrification. The commercial establishments retained a strong ethnic character, and there was evidence of public investment (pedestrian lighting, colorful murals, and an ornate gazebo) near the station. Two blocks had higher-end Mexican cafes. Indicators of residential gentrification were present, but less significant than at the Hollywood/Western station. There were six residential and two commercial new constructions. Most buildings (84 percent) were older, but one third of the surveyed blocks had buildings with recent renovations.
Interviews
Our third method—interviews with representatives from CBOs and public agencies—revealed some facets of gentrification that had not been uncovered by the analysis of secondary data or the street-level observations. For example, our observations in Chinatown that failed to find any major evidence of commercial gentrification differed from the perceptions of CBO representatives in the area, who expressed concerns that a growing number of new neighborhood businesses are not catering to the needs of long-term Chinatown residents, but instead target a new clientele. As stated, “New development and incoming retailers are catering to new residents or more affluent commuters” (SEACA, interview, February 4, 2015). CBOs reported some flipping of commercial properties (CCED, interview, April 15, 2015), and noted that business turnover and displacement have led some long-term residents to leave because they no longer feel a cultural and economic connection to Chinatown (SEACA, interview, February 4, 2015).
Chinatown CBOs also expressed concerns over the threat to affordable housing units. In the words of one interviewee: “Chinatown has had affordable senior housing since the 1980s, but many of the affordable units have expired or are set to expire, and some affordable senior units are converting into market rate units” (CCED, interview, April 15, 2015). Between 2007 and 2014, at least 14 Ellis Act residential evictions have occurred in the census tracts within half mile of the transit station. One CBO representative reported that “tenants are often offered buyouts to move out of their units” (CCED, interview, April 15, 2015). CBOs believe that developers see an opportunity to attract higher returns on their developments, which may have negative effects for a neighborhood like Chinatown that has many low-income renters (CCED, interview, April 15, 2015). Conversely, a city planner argued that the new “infrastructure investments are attracting developers to the area interested in turning existing commercial and industrial properties into housing” (LACPD, interview, April 15, 2015).
Representatives of CBOs interviewed in the Hollywood/Western neighborhood underlined the residential gentrification that the area is experiencing. According to one interviewee: “Real estate speculation is forcing long-term, low-income renters out of their neighborhood.” Indeed, the mean rent has increased in this neighborhood by more than 40 percent since 1980. An organizer from LA Voice (an interfaith local community group) estimated that 30 percent of a Hollywood church congregation has moved to San Fernando Valley because of rising rents in Hollywood (LA Voice, interview, April 10, 2015). On the other hand, city planners noted that the area has experienced less gentrification than the core of Hollywood, thanks to the concerted efforts by CBOs, business owners, and the council district to fund programs that benefit the existing population.
Both the gentrification model and visual survey in the 103rd/Watts Towers neighborhood did not find visible signs of gentrification. CBO and public agency representatives confirmed the lack of noticeable changes near the station. CBOs, however, noted instances of illegal evictions and slum conditions in the wider area of South Los Angeles (SAJE, interview, February 4, 2015). The CBO representatives interviewed believe that this station-neighborhood remains underserved, and economic and community development efforts in the Watts area—a community heavily impacted by the 1965 Watts riots and the 1992 Los Angeles riots—have been largely unsuccessful (LAANE, interview, February 13, 2015), which may explain the reluctance of wealthier residents to move in and the absence of gentrification. According to a planner interviewed: “In Watts, there has been little development around the Blue Line. Unlike in many areas where [rail] transit has come in, transit has not been a draw. The area is lacking in services, business, [mix of] uses, jobs, and housing” (LACPD, interview, April 15, 2015).
At Mariachi Plaza, CBOs expressed concerns that a growing number of new neighborhood businesses are not catering to the needs of long-term residents. As noted “since the opening of the station, Mariachi Plaza and its weekly market are increasingly being marketed for cultural tourism. The plaza no longer serves the community but is mostly portrayed and promoted as a place for tourists” (Union de Vecinos, interview, February 14, 2015). On the other hand, city planners disputed that the new transit station has displaced existing small businesses, arguing that “many local mom-and-pop stores are deep in residential neighborhoods, surrounded by duplexes/triplexes, multifamily and not near the transit station. The zoning where these stores are located does not lead to a financially attractive use of land for developers” (LACPD, interview, April 15, 2015).
Data Triangulation
In addition to getting a deeper understanding of gentrification processes in the four neighborhoods, we also sought to compare and contrast the data gathered by different methods and test the accuracy of secondary data. Thus, we used data triangulation to compare the secondary data to the primary data collected through field observations. Specifically, we examined the consistency between Assessor’s data on land uses, residential improvements, and new construction with data collected from field observations. Table 9 shows the level of consistency between the Assessor’s data and observed data about residential land uses. We found a significant inconsistency in Hollywood/Western and some inconsistency in Chinatown between the assessed and observed single-family homes.
Percentage of Residential Land Use Matched.
Source: Tabulated by authors from County Assessor’s data; and observations collected in March and June 2015.
We defined a “major residential improvement” as one where extensive renovation was visible, which would have likely required a building permit. Table 10 shows a discrepancy at Hollywood/Western, where observations found only 2 percent of properties with major improvements while the Assessor’s data indicate 9 percent.
Percentage of Major Improvements for Observed and Assessor Single-Family Parcels.
Source: Tabulated by authors from County Assessor’s data; and observations collected in March and June 2015.
Table 11 shows the match between reported and observed construction for single-family parcels. We found consistency in the two data sets for the Hollywood/Western station, where there was no reported or observed new construction for single-family homes. A large inconsistency was noted in Chinatown (32 percent observed new construction compared to 4 percent reported), and a more modest inconsistency at 103rd/Watts (13 percent compared to 5 percent) and Mariachi Plaza (7 percent compared to 1 percent). For the most part, we found more consistency between the secondary data, field observations, and model results in areas with little development, but less consistency in areas undergoing more significant development.
Percent of Constructions for Observed and Assessor Single Family (SF) Parcels.
Source: Tabulated by authors from County Assessor’s data; and observations collected in March and June 2015.
What Did We Learn about Gentrification?
So what did we learn about each neighborhood in regard to gentrification? We found that gentrification is a complex and context-specific phenomenon that cannot be simply captured by one indicator. Take for example rent burden. Simple statistics cannot easily capture change because the initial wave of higher-income residents who benefit from existing rent levels could keep the rent-to-income ratio low, or even lower it. At the same time, existing census information is not sufficient because it does not depict the fine-grain changes, such as the number of individuals displaced. We can only indirectly get a sense of that, and the better data (such as evictions) are not readily available across jurisdictions. Given this complexity and disparate trajectories, it is important to have multiple indicators of gentrification.
Table 12 is a visual representation of how different gentrification indicators in the four neighborhoods (gathered from the three methods) compare to County trends. Taking into account these indicators, we can tell that the 103rd /Watts neighborhood has not gentrified, the Hollywood/Western neighborhood is at a late stage of gentrification, while Chinatown and Mariachi Plaza exhibit early stages of gentrification.
Gentrification Indicators in the Four Neighborhoods as Compared to LA County.
Note: HH = household.
Significant change towards gentrification compared to County averages/trends;
Moderate change towards gentrification compared to County averages/trends;
Small or no change compared to County averages/trends;
Absence of this gentrification indicator.
In addition to using multiple indicators, it is also important to employ different methodologies to measure these indicators. This is shown in Table 13, which summarizes the assessment from each method. We can see that in two neighborhoods, where gentrification was either clearly visible or clearly absent, the assessment of all three methods converged. Indeed 103rd/Watts has not experienced gentrification, while Hollywood/Western has experienced significant commercial and residential gentrification near its station. On the other hand, the assessment of the presence or absence of gentrification was somewhat different depending on the method in two other neighborhoods. In Chinatown, the secondary data showed only signs of residential gentrification at the southwest edges of the neighborhood; visual surveys showed also some additional residential gentrification closer to the station; while the interviews confirmed the observations but also revealed the flipping of some commercial properties and early signs of commercial gentrification. Lastly, the secondary data did not reveal residential or commercial gentrification at Mariachi Plaza. However, observations and interviews with CBOs pointed to early signs of both residential and commercial gentrification. It should be noted, however, that some planners interviewed disputed that commercial gentrification is taking place in this neighborhood.
Assessment of Presence or Absence of Gentrification per Method per Neighborhood.
Conclusion
This study used a mixed-methods approach to “drill in” the topic of gentrification in four transit neighborhoods. It showed that gentrification is a fluid process that cannot be captured by one indicator or one source of data. The employment of a mixed-methods approach offered a richer picture of neighborhood change and revealed the strengths and limitations of the different methods and data sources. It also showed how gentrification can be manifested from the compilation of different empirical indicators that planners could use in contingent ways.
Secondary data with sociodemographic and housing indicators can give planners a longitudinal view, showing how these indicators have changed over the years, as well as a cross-sectional view, showing how they compare to those of other (control) neighborhoods or the County averages. But secondary data cannot always tell us what is on the ground, such as new or unpermitted renovations that community residents can see but are not reflected in the Assessor’s data, or different merchandise or increased prices that some businesses may enact in response to a new clientele. As our triangulation exercise showed, in areas of significant change, secondary data may not always give a fully accurate picture of changes in the built environment and current land uses. Planners should not assume that secondary data are always precise, and should carefully evaluate such data for anomalies before using them in models.
Systematic field observations can give a wealth of information about a neighborhood’s urban form and social activities at the parcel and street level, and can help groundtruth secondary data. In this study, observations revealed some visual signs of gentrification such as new “hipster” establishments, and new buildings that were more upscale than their surroundings. Field observations, however, cannot capture numerical changes in neighborhood demographics or real estate transactions, which are better identified by secondary data. Unless field observations are compared with data sources that give information about the past context of a neighborhood (e.g., street-level photographs, aerial photographs of previous years), they can only give a static view of what exists in a neighborhood. Additionally, case study research that focuses on a specific neighborhood at one point in time does not offer opportunities for comparison and control with other neighborhoods cross-sectionally and longitudinally, and thus cannot explain what may have triggered neighborhood change. Collection of primary data is also tedious and time-consuming, and for this reason, cannot easily cover large geographic areas. Lastly, visual observations may at times be inadequate to distinguish between housing types, such as condos versus apartment complexes, and may require further verification. Field observations require significant training of the individuals undertaking them, as well as the preparation of detailed guidelines and survey templates (see Appendices A to C in online supporting information) for different assessments of the urban form to avoid inconsistencies among the different field surveyors.
Interviews with knowledgeable local stakeholders can help uncover information about neighborhood change that is not available in secondary data or readily visible by field observations. Those on the ground may be knowledgeable of patterns of change that are not captured by other data sources. Thus, our interviewees gave anecdotal evidence of, for example, what is happening at particular Chinatown or Mariachi Plaza stores that seem to target new incoming residents or tourists, and parishioners in an interviewee’s congregation, who have left Hollywood/Western because of rising rents. Data from interviews, however, are subjective and may reflect the biases, priorities, advocacy, and broader concerns of the observer, interviewer, and interviewees. Therefore, such data also need to be triangulated and solicited by different groups that may have differentiated or even opposing views.
As other researchers have argued, gentrification is a dynamic process that has different stages (Kerstein 1990). While each method contributes to a richer understanding of the extent of gentrification in a neighborhood, planners may not always have the resources and time to pursue multiple research methodologies. This study showed that an examination of neighborhood change based only on secondary data is likely adequate for neighborhoods that either do not have signs of gentrification or experience a late stage of gentrification. However, the employment of only secondary data may lead to inaccurate assessments, if a neighborhood is at an early stage of gentrification. In such cases, planners will do well to complement secondary data with qualitative methodologies, visit the neighborhood, groundtruth the census data, and even hear from local stakeholders.
Neighborhoods evolve and change over time in complex and different ways. We often encounter discrepancies in indicators and beliefs about the nature and extent of neighborhood change. This is due to the complexity of identifying, measuring, and characterizing change, but also the existence of different information sets, and even the inaccuracy of some data sources. While small inaccuracies and discrepancies may not matter much when data show a clear presence or absence of gentrification, they may lead to wrong assumptions if gentrification is in early stages. In such cases, the utilization of multiple indicators and data sources that involve both secondary data as well as empirical work such as field observations and stakeholder interviews complement each other and can give a more complete picture of neighborhood change brought about by gentrification.
Footnotes
Acknowledgements
We would like to acknowledge the support of the California Air Resources Board as well as Karen Chapple, Dan Chatman, Paul Waddell, and Miriam Zuk, who were members of the research team for the larger project.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research for this article was drawn from a larger research study which was funded by the California Air Resources Board.
Supplemental Material
Supplemental material for this article is available with the manuscript on the JPER website.
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
