Does Gentrification Contribute to Segregation? A Study of Urban Displacement.*

Does Gentrification Contribute to Segregation? A Study of Urban Displacement.* Abstract Gentrification is a concept with a broad set of implications, but at its base, represents the economic revitalization of neighborhoods and a reinvestment in local markets and infrastructure. Gentrification is often preceded by the arrival of more highly educated artists and professionals and is followed by the arrival of tech elites and other monied interests seeking to capitalize on the revitalization of these neighborhoods. On the flip side, gentrification has also been associated with displacement of poorer neighborhood residents as well as the disruption of indigenous social communities. The recent focus on gentrification of American cities has led to increased interest on the effects of this process, both good and bad. While it has been shown that, in the short run, gentrification may lead to more diverse communities largely through an influx of non-hispanic whites, the long-term effects of potential displacement on resident diversity are less well known. As a result, we seek to explore the effects of gentrification on racial/ethnic and income segregation in the United States by utilizing census data over the past four decades (1970-2010). Although later analyses will examine additional decades, our results from 2000 to 2010 indicate that tract-level reductions in both the non-white population and those in poverty are higher in tracts that gentrified over the decade compared to tracts that did not; declines in the black population were especially notable in gentrifying tracts. Despite tract-level evidence of racial and economic compositional changes associated with gentrification, metropolitan-level analyses reveal a slight buffering or non-existent relationship between gentrification and economic and racial segregation. *Draft manuscript. Please do not cite without permission from authors.

Introduction The definition of gentrification tends to center around its perception as a unifying or dividing force for neighborhoods and communities (Atkinson 2003). At its most basic, Smith defines gentrification as the process by which central urban neighborhoods that have undergone disinvestments and economic decline experience a reversal, reinvestment, and the in-migration of a relatively well-off, middle- and upper middle- class population. (1998, p. 198). However, alternative definitions that are consistent with colloquial and journalistic representations argue that gentrification represents the replacement of low-income, inner-city working class residents by middle- or upper- class households, either through the increase in market value of existing housing or demolition to make way for new upscale higher-cost housing construction (Hammel and Wyly 1996, p. 250). In other words, many view gentrification through the lens of displacement of the urban poor, regardless of its role in the economic revitalization of urban spaces; this dovetails with the origins of the term (Glass 1964). Even further, some view gentrification as a process diluting vibrant ethnic neighborhoods and destroying indigenous social communities (Atkinson 2000). The debate over the definition of gentrification tends to vary with one s interpretation of its consequences. This debate may continue to unfold for the indefinite future. While some predicted that gentrification would not have a significant impact on the American urban landscape (Bourne 1993), more recent empirical estimates suggest that gentrification is here to stay and may overtake America s cities (Brueckner and Rosenthal 2009). A key question is which population sectors gain vs. lose from gentrification. Thus, research on understanding the effects of gentrification on existing residents of communities with significant previous disinvestment that undergo gentrification is critical and needed to inform equitable community development strategies.

Gentrification and Its Effects. An evidence-based review (Atkinson 2002) of the effects of gentrification on urban areas describes a host of positive characteristics of gentrification, including: stabilization of declining areas, increased property values, reduced vacancy rates, increased local fiscal revenues, encouragement and increased viability of further development, reduction of suburban sprawl, increased social diversity, decreased crime, and property rehabilitation. Noted negative effects of gentrification include: displacement through rent/price increases, secondary psychological costs of displacement, community resentment and conflict, loss of affordable housing and unsustainable speculative property increases, commercial/industrial displacement which may in turn be associated with fewer employment opportunities for lower income and working class residents, increased costs and changes to local services, housing demand pressures on surrounding areas, loss of social diversity, increased crime, and under-occupancy and population loss to gentrified areas (Atkinson 2002). This review further notes that the negative effects of gentrification may be small compared to larger issues of urban decline and the general abandonment of some central city areas; similarly, one study argues that gentrification effects pale in comparison to larger trends in the increase in high-poverty neighborhoods which have tripled over the past four decades (Cortright and Mahmoudi 2014). Several studies find that residential turnover is similar in gentrifying and non-gentrifying neighborhoods (Freeman 2008; Freemand and Braconi 2004; Freeman 2005; Vigdor 2002; McKinnish, Walsh, and White 2008). However, similar rates of turnover can still result in displacement if movers are concentrated among more poor households. However, Freeman (2005) finds that demographic changes in gentrifying neighborhoods are due largely to the relative affluence of in-movers and not due to high rates of displacement among out-movers.

Nonetheless, Freeman finds a small rate of displacement among poor movers, a finding which was not supported in previous research (Vigdor 2002; Freeman and Braconi 2004). Subsequent research finds not only null evidence for displacement, but finds that gentrification of predominantly black neighborhoods creates favorable conditions for current residents (McKinnish, Walsh, and White 2010). These finding are tempered by a study which finds that gentrification is attenuated in neighborhoods that exceed 40% black residents (Hwang and Sampson 2014) indicating a persistent racial divide that influences gentrification decisions by White movers. It is important to note that at any given time, nearly 1 in 10 (9.6%) American households report a desire to move (Mateyka 2015), while over over 1 in 3 (35.4%) households has actually moved in the past 5 years (Ihrke and Faber 2012). Thus, neighborhood turnover is a fact of modern life and even moderate rates of turnover in a neighborhood could result in changes in neighborhood composition. As former suburbanites return to the inner city, major compositional changes are possible. One study (Freeman 2008) compares gentrifying neighborhoods to nongentrifying neighborhoods within the same census areas and finds gentrifying neighborhoods to be more diverse with respect to both race/ethnicity and socio-economic status. Given high rates of moving in U.S. neighborhoods, this would certainly make sense in the short term as more educated, higher income white residents move into lower income neighborhoods. It may also be the case that racially/ethnically homogenous neighborhoods may not be deemed by whites to be ideal neighborhoods for gentrification, conflating cause and effect in this study. In addition, gentrification may result in increasing rents, higher property taxes, and increased foreclosures/evictions factors that could take years to play out and result in displacement over time. In the context of significant income inequality and income segregation in the United States,

gentrification becomes a significant factor in further shaping options for low income and minority residents as they are priced out of gentrifying neighborhoods. As such, investigating longer term trends in gentrifying neighborhoods is crucial in order to draw conclusions and inform development strategies. Hypotheses Our key hypothesis is that gentrification, although marked by increased diversity in the short term, will ultimately displace poorer households leading to higher levels of income and racial segregation. It is important that any analysis of compositional changes within neighborhoods accounts for larger compositional changes at the regional level. As such, we explore the contribution of gentrification to segregation measures at the metropolitan level. Data and Analysis We use data from 40 years of the US census from the Longitudinal Tract Database (LTDB) (Logan, Xu, Stults, 2014), including 2010 data from the American Community Survey (ACS, five-year estimates). Additional variables not available through this database were extracted from SF 3 in the 2000 Census and from the American Community Survey (through American Factfinder). The LTDB provides linked files for a broader array of variables than the Neighborhood Change Database (NCDB) (Tatian 2003) 1 and allows for merging on non-census sources of data. In addition to the LTDB, we also anticipate using the NCDB for linking the 1990 tracts, as the NCDB has more reliable estimates for 1990 based on its interpolation method based on block populations and proxies for population density; both databases use similar areabased interpolation methods in 1970 and 1980 (Logan, Xu, Stults, 2014). From this data, we 1 Both databases suffer from sources of error, including allocating individuals into newly split blocks in proportion to the area of the block fragments and assuming that when tracts split that the parts of the tract the form new tracts or merge with other tracts are homogenous in terms of population composition (www.s4.brown.edu/us2010/researcher/bridging.htm).

construct measures for both the tract-level as well as metropolitan areas. Following Logan et al. (2010), we restrict our sample to metropolitan areas with populations of 500,000 or greater (116), and additionally restrict our sample to metropolitan areas with black and Hispanic populations of 10,000 or greater for the racial segregation analyses (112). Operationalizing Gentrification. Based on an operationalization developed by Freeman (2005), a neighborhood must meet the following criteria to be classified as gentrifying: 1. Be located in the central city at the beginning of the intercensal period. 2. Have a median income less than the median (40 th percentile) for that metropolitan area at the beginning of the intercensal period. 3. Have a proportion of housing built within the past 20 years lower than the proportion found at the median (40 th percentile) for the respective metropolitan area. 4. Have a percentage increase in educational attainment greater than the median increase in educational attainment for that metropolitan area. 5. Have an increase in real housing prices during the intercensal period. Neighborhoods are then classified as a) non-gentrifying, b) potentially gentrifying (first three criteria), and c) gentrifying (all five criteria). This definition effectively captures the construct of gentrification without relying on an income driven definition and without making reference to changes in racial/ethnic composition which could presuppose displacement. Statistical Models. Our regression models are adapted from Reardon and Bischoff (2011) and are as follows: = + + + +... Equation 1 where and index metropolitan areas and census years, respectively, and and are variables indicating changes in segregation and the percent of metro-area tracts gentrified between censuses, in metropolitan area in year. The models include both metropolitan area ( ) and year ( ) fixed-effects, as well as a vector of metro-year covariates ( ) as control variables. The coefficient on gentrification represents the average within-metro area association (over time) between gentrification and segregation, net of any metropolitan area

specific secular trend. We also account for the clustering of observations by estimating robust standard errors. Operationalizing Segregation. We utilize four measures of economic and racial segregation for our outcomes. The first, the Information Theory Index (H) compares variation in family income at the tract level with that at the metropolitan level to create an income segregation measure, but is not confounded with changes in income inequality (Reardon 2011; Reardon and Bischoff 2011). Higher values, from 0 to 1, indicate greater correlation between a family s income and that of their near neighbors income. We also capture racial segregation through three oft-utilized measures: the Dissimilarity Index, the Isolation Index, and the Exposure Index. The Dissimilarity Index is a measure of evenness, or the percentage of a group that would have to change residence to have the same distribution within a tract as occurs within a metropolitan area (cite). The Isolation Index captures the extent to which minority members are exposed only to one another, (Massey and Denton, 1988: 288). Finally, the Exposure Index measures possible contact between groups. Following Logan and colleagues, we also included a host of controls (Table 2) which we will explicate in more detail in the next draft, including population and socioeconomic composition measures. Results We first conducted exploratory tract-level analyses to investigate the extent of change in economic and racial composition from 2000 to 2010. Using a series of t-tests, we tested the mean difference in change between gentrified and non-gentrified tracts (Table 1). The percent nonwhite grew by 5.11 percent in non-gentrified tracts, on average, whereas it grew by only 1.46 percent in gentrified tracts, on average. Similarly, the percent Hispanic grew disproportionately in non-gentrified tracts (3.41 percent) relative to gentrified ones (1.98 percent). Tracts that

gentrified lost their black population rapidly, a reduction of 2.16 percent, on average, relative to a 0.05 percent change in non-gentrified tracts. Additionally, we also see larger growth in individuals and families in poverty in non-gentrified tracts relative to gentrified ones; in fact, the percent of families in poverty shrunk by 0.12%, on average, in gentrified tracts. All of these differences are statistically significant at p<0.01. These results suggest that, at least at the tractlevel, gentrification may result in racial and economic segregation. However, it is unclear whether these tract-level changes will be associated with racial and economic segregation at the metropolitan level, including and beyond the central city. We examined the association of gentrification with four metropolitan-level measures of economic and racial segregation (Table 3) for 2000 to 2010; additional decades will be included in the final draft. For the Information Theory Segregation Index (H), we find that the percent gentrified within the CBSA s central city is positively related to H, but contrary to expectations, this relationship is not significant. In the covariate adjusted model, the average change in per capita income and increases in economic inequality as measured by the Gini coefficient are positively and significantly related the H. Likewise, we also find that the percentage gentrified is positively, but not significantly, related to the Dissimilarity Index for racial segregation. Dissimilarity is, however, negatively related to increases in the foreign born population, and positively related to increases in the percent black and Hispanic. Like H, racial segregation also increases with growing economic inequality. Gentrification is positively, but not significantly related to racial isolation before controlling for other metropolitan-level change; however, after controlling for our host of covariates, gentrification is associated with small but significant declines in racial isolation. As with the Dissimilarity Index, the percentage change in those with less than high school predicts

increases in racial isolation. In contrast, the percentage black decreases racial isolation and the change in the percentage of Hispanics is unrelated to racial isolation. After controlling for covariates, gentrification is associated with increases in racial exposure between blacks and whites. A percentage change in foreign born individuals is associated with greater racial exposure between blacks and non-hispanic whites. For each racial segregation outcome, a percent change in female-headed households is associated with decreasing racial evenness and isolation and increasing exposure. Income inequality is not significantly related to racial isolation or racial exposure. In a series of robustness checks, we also examined whether the percent change in tracts with the potential to gentrify independently predicted the segregation outcomes. A one percent increase in the tracts with the potential to gentrify over the period predicted small declines in the Dissimilarity Index and Isolation Index, and small increases in exposure (Appendix 1); However, these estimates either lost significance after adjusting for covariates or were only moderately significant (p<0.10). Additional analyses (Appendix 2) tested an alternate measure of gentrification: the percent gentrified within the entire CBSA rather than within the central city. The results were extremely similar to those presented in the body of this paper. Future Directions In addition to adding additional decades of data, we will also investigate other functional forms for gentrification and other measures of economic and racial segregation, including multiethic measures of racial segregation. We also plan to highlight particular communities where gentrification led to high rates of displacement and/or community disruption.

References Atkinson, Rowland. 2000. Measuring Gentrification and Displacement in Greater London. Urban Studies 37(1): 149-165. Atkinson, Rowland. 2002. Does Gentrification Help or Harm Urban Neighborhoods? An Assessment of the Evidence-Base in the Context of the New Urban Agenda. CNR Paper 5: June 2002. Glasgow: Center for Neighbourhood Research. Atkinson, Rowland. 2003. Introduction: Misunderstood Saviour or Vengeful Wrecker? The Many Meanings and Problems of Gentrification. Urban Studies 40(12): 2343-2350. Bischoff, Kendra and Sean F. Reardon. 2014. Residential Segregation by Income, 1970-2009. In Diversity and Disparities: America Enters a New Century, Ed. John Logan. The Russell Sage Foundation: New York. Bostic, Raphael W. and Richard W. Martin. 2003. Black Home-owners as a Gentrifying Force? Neighbourhood Dynamics in the Context of Minority Home-ownership. Urban Studies 40(12): 2427-2449. Bourne, L.S. 1993. The Demise of Gentrification? A Commentary and Prospective View. Urban Grography14(1): 95-107. Boustan, Leah Platt. 2011. Racial Residential Segregation in American Cities, in Oxford Handbook of Urban Economics and Planning, ed. Nancy Brooks and Gerrit-Jan Knaap, Oxford University Press, 318 39. Brueckner, Jan K. and Stuart S. Rosenthal. 2009. Gentrification and Neighborhood Housing Cycles: Will America s Future Downtowns be Rich? The Review of Economics and Statistics 91(4): 725-743.

Chapple, Karen. 2009. Mapping Susceptibility to Gentrification: The Early Warning Toolkit. Center for Community Innovation August 2009. Cortright, Joe and Dillon Majmoudi. 2014. Lost in Place: Why the Persistence and Spread of Concentrated Poverty not Gentrification is Our Biggest Urban Challenge. City Report December 2014. Coulton, Caludia, Brett Theodos, and Margery A. Turner. Residential Mobility and Neighborhood Change: Real Neighborhoods Under the Microscope. Cityscape: A Journal of Policy Development and Research 14(3): 55-90. Freeman, Lance and Frank Braconi. 2004. Gentrification and Displacement. Journal of the American Planning Association 70(1): 39-52. Freeman, Lance. 2005. Displacement or Succession? Residential Mobility in Gentrifying Neighborhoods. Urban Affairs Review 40(4): 463-491. Freeman, Lance. 2008. Testimony of Lance Freeman to the National Commission on Fair Housing and Equal Opportunity.. Fry, Richard and Paul Taylor. 2012. The Rise of Residential Segregation by Income. Pew Research Center: Social and Demographic Trends August 1, 2012. Glass, Ruth. 1964. Introduction: Aspects of Urban Change. Centre for Urban Studies, ed. p. xiii-xlii. London: MacGibbon and Kee. Hammel, D.J. and E.K. Wyly. 1996. A Model for Identifying Gentrified Areas with Census Data. Urban Geography 17(3): 248-68. Hwang, Jackelyn and Robert J. Sampson. 2014. Divergent Pathways of Gentrification: Racial Inequality and the Social Order of Removal in Chicago Neighborhoods. American Sociological Review 79(4): 726-751.

Iceland, John, Daniel H. Weinberg, and Erik Steinmetz. 2002. Racial and Ethnic Residential Segregation in the United States: 1980-2000. U.S. Census Bureau, Series CENSR-3. U.S. Government Printing Office: Washington, D.C. Immergluck, Dan and Geoff Smith. 2003. Measuring Neighborhood Diversity and Stability in Home-Buying: Examining Patterns by Race and Income in a Robust Housing Market. Journal of Urban Affairs 25(4): 473-491. Ihrke, David K. and Carol S. Faber. 2012. Geographical Mobility: 2005-2010. U.S. Census Bureau P20-S67. Keating, Larry. 2000. Redeveloping Public Housing: Relearning Urban Renewal s Immutable Lessons. Journal of the American Planning Association 66(4): 384-397. Kennedy, Maureen and Paul Leonard. 2001. Dealing with Neighborhood Change: A Primer on Gentrification and Policy Changes. Brookings Institution Center on Urban and Metropolitan Policy Discussion Paper April 2001. Logan, John R. and Brian J. Stults. The Persistence of Segregation in the Metropolis: New Findings from the 2010 Census. US2010 Project March 24, 2011. Logan, John R., Zengwang Xu, and Brian Stults. 2014. Interpolating US Decennial Census Tract Data from as Early as 1970 to 2010: A Longitudinal Tract Database The Professional Geographer 66(3): 412-420. Mateyka, Peter J. 2015. Desire to Move and Residential Mobility. U.S. Census Bureau P70-140. McKinnish, Terra, Randall Walsh, and Kirk White. 2008. Who Gentrifies Low Income Neighborhoods? NBER Working Paper Series 14036.

McKinnish, Terra, Randall Walsh, and Kirk White. 2010. Who Gentrifies Low Income Neighborhoods? Journal of Urban Economics 67(2): 180-193. Newman, Kathe and Elvin K. Wyly. 2006. The Right to Stay Put, Revisited: Gentrification and Resistance to Displacement in New York City. Urban Studies 43(1): 23-57. Reardon, Sean F. and Kendra Bischoff. Income Inequality and Income Segregation. American Journal of Sociology 116(4): 1092-1153. Redfern, P.A. 2003. What Makes Gentrification Gentrification. Urban Studies 40(12): 2351-2366. Sharkey, Patrick. 2013. Stuck in Place: Urban Neighborhoods and the End of Progress Toward Racial Equality. University of Chicago Press: Sheppard, Stephen. 2012. Why is Gentrification a Problem? Center for Creative Community Development. Smith, Neil. 1998. Gentrification. Pp. 198-199 in The Encyclopedia of Housing, edited by W.v. Vliet. London: Taylor and Francis. Tatian, P.A. 2003. Neighborhood Change Database (NCDB) 1970-2000 Tract Data: Data Users Guide. Washington, DC: Urban Institute. Taylor, Paul, Rich Morin, D Vera Cohn, and Wendy Wang. 2008. American Mobility: Who Moves? Who Stays Put? Where s Home? Pew Research Center: A Social and Demographic Trends Report. Vandergrift, Janelle. 2006. Gentrification and Displacement. Urban Altruism Spring 2006. Vigdor, Jacob L. 2002. Does Gentrification Harm the Poor? Brookings-Wharton Papers on Urban Affairs.

Table 1. 2000-2010 Tract-level Change, by Gentrification Status Change Non-Gentrified Gentrified Central Difference Central City Tracts City Tracts % Non-White 5.11 1.46 3.64** % Black 0.05-2.16 2.22** % Hispanic 3.41 1.98 1.43** % in Poverty 2.74 0.67 2.07** % Families in Poverty 1.71-0.12 1.83** N 16,351 4,638 Note: Includes only central city tracts eligible for gentrification. N varies slightly across variables; listed N is for % Non-White. * p<0.05; ** p<0.01

Table 2. Descriptive Statistics for Regression Model Variables, 2000-2010 Mean Std. Dev. Minimum Max Outcomes, Change in: Information Theory 0.032 0.006 Segregation Index, Income Dissimilarity Index 0.006 0.063 Isolation Index -0.003 0.013 Exposure Index 0.003 0.013 Covariates % Gentrified 20.81 10.180 % with Potential for 6.32 8.987 Gentrification Change in: Population (ln) 0.39 0.254 % Less than HS -4.27 1.385 % College Degree 3.64 0.984 % Population over 65-3.65 3.843 % Population under 18-1.11 1.022 % Unemployment 2.38 1.528 % Manufacturing Employees -2.83 1.500 % Foreign Born 1.97 1.193 % Female-Headed Households -0.004 1.033 Per Capita Income 5566.68 1816.311 % Black -1.20 2.120 % Hispanic 3.02 2.164 % Non-Family Households 18.74 1.974 Gini Coefficient 0.072 0.013 N 116

Table 3. Regression Results for Income and Racial Segregation, 2000-2010 H H D D I I E E % Gentrified a 0.00000-0.00003 0.00003-0.00031 0.00006-0.00025-0.00006 0.00025 (0.06) (1.21) (0.07) (0.79) (0.46) (2.67)** (0.46) (2.67)** Change in: Population (ln) -0.00019 0.01608-0.01424 0.01424 (0.07) (0.45) (1.65) (1.65) % Less than HS 0.00015 0.01251 0.00256-0.00256 (0.51) (3.11)** (2.62)* (2.62)* % College Degree 0.00034 0.00109-0.00139 0.00139 (0.98) (0.24) (1.27) (1.27) % Population over 65 0.00026 0.00328-0.00076 0.00076 (1.55) (1.52) (1.45) (1.45) % Population under 18 0.00045-0.00190 0.00181-0.00181 (1.27) (0.40) (1.59) (1.59) % Unemployment 0.00034-0.00700-0.00115 0.00115 (1.34) (2.13)* (1.45) (1.45) % Manufacturing Employees 0.00047-0.00120-0.00022 0.00022 (2.07)* (0.40) (0.30) (0.30) % Foreign Born 0.00009-0.01838-0.00502 0.00502 (0.25) (3.99)** (4.49)** (4.49)** % Female-Headed Households -0.00068-0.01731-0.00346 0.00346 (1.63) (2.69)** (2.22)* (2.22)* Per Capita Income 0.00000 0.00001 0.00000-0.00000 (2.53)* (2.48)* (3.59)** (3.59)** % Black -0.00003 0.01143-0.00292 0.00292 (0.12) (3.99)** (4.21)** (4.21)** % Hispanic -0.00007 0.00967-0.00053 0.00053 (0.42) (4.30)** (0.96) (0.96) % Non-Family Households 0.00010 0.00066 0.00007-0.00007 (0.63) (0.30) (0.12) (0.12) Gini Coefficient 0.43084 0.76670 0.06236-0.06236 (16.85)** (2.25)* (0.75) (0.75) Constant 0.03204-0.00153-0.00151-0.01428-0.00446 0.01311 0.00446-0.01311 (25.21)** (0.35) (0.14) (0.24) (1.51) (0.90) (1.51) (0.90) R 2 0.00 0.82 0.00 0.54 0.00 0.63 0.00 0.63 N 116 116 112 112 112 112 112 112 ^ p<0.10; * p<0.05; ** p<0.01 H=Information Theory Segregation Index, Income, D=Dissimilarity Index, I=Isolation Index, E=Exposure Index a Within central city

Appendix 1. Regression Results for Income and Racial Segregation, 2000-2010 H H D D I I E E % Gentrified a 0.00001-0.00002-0.00049-0.00044 0.00002-0.00030-0.00002 0.00030 (0.14) (0.74) (0.98) (1.08) (0.14) (3.13)** (0.14) (3.13)** % with Potential 0.00001 0.00005-0.00150-0.00063-0.00011-0.00025^ 0.00011 0.00025^ for Gentrification a (0.22) (1.34) (2.84)** (1.16) (0.77) (1.90) (0.77) (1.90) Population (ln) 0.00041 0.01077-0.01633 0.01633 (0.15) (0.30) (1.90) (1.90) % Less than HS 0.00013 0.01250 0.00255-0.00255 (0.45) (3.11)** (2.65)** (2.65)** % College Degree 0.00030 0.00139-0.00127 0.00127 (0.86) (0.31) (1.18) (1.18) % Population over 65 0.00027 0.00315-0.00081 0.00081 % Population under 18 (1.63) (1.46) (1.57) (1.57) 0.00042-0.00159 0.00192-0.00192 (1.18) (0.34) (1.71) (1.71) % Unemployment 0.00034-0.00705-0.00117 0.00117 (1.36) (2.15)* (1.49) (1.49) % Manufacturing Employees 0.00048-0.00139-0.00029 0.00029 (2.13)* (0.46) (0.41) (0.41) % Foreign Born 0.00011-0.01856-0.00509 0.00509 (0.32) (4.03)** (4.61)** (4.61)** % Female-Headed Households -0.00067-0.01699-0.00333 0.00333 (1.63) (2.64)** (2.16)* (2.16)* Per Capita Income 0.00000 0.00001 0.00000-0.00000 (2.87)** (1.76) (2.49)* (2.49)* % Black 0.00000 0.01108-0.00306 0.00306 (0.00) (3.85)** (4.44)** (4.44)** % Hispanic -0.00002 0.00905-0.00077 0.00077 (0.12) (3.92)** (1.39) (1.39) % Non-Family Households 0.00008 0.00095 0.00018-0.00018 (0.47) (0.42) (0.33) (0.33) Gini Coefficient 0.43130 0.75672 0.05844-0.05844 (16.93)** (2.22)* (0.72) (0.72) Constant 0.03185-0.00276 0.01886-0.00159-0.00291 0.01809 0.00291-0.01809 (20.64)** (0.62) (1.47) (0.03) (0.81) (1.24) (0.81) (1.24)

R 2 0.00 0.82 0.07 0.54 0.01 0.64 0.01 0.64 N 116 116 112 112 112 112 112 112 ^ p<0.10; * p<0.05; ** p<0.01 H=Information Theory Segregation Index, Income, D=Dissimilarity Index, I=Isolation Index, E=Exposure Index a Within central city Appendix 2. Regression Results for Income and Racial Segregation, 2000-2010 H H D D I I E E % Gentrified b 0.00021 0.00002 0.00059-0.00117 0.00064-0.00020-0.00064 0.00020 (1.61) (0.35) (0.56) (1.30) (2.29)* (0.87) (2.29)* (0.87) Population (ln) -0.00026 0.01366-0.01508 0.01508 (0.10) (0.38) (1.69) (1.69) % Less than HS 0.00013 0.01293 0.00263-0.00263 (0.43) (3.22)** (2.60)* (2.60)* % College Degree 0.00029 0.00075-0.00169 0.00169 (0.84) (0.17) (1.51) (1.51) % Population over 65 0.00024 0.00332-0.00082 0.00082 (1.44) (1.55) (1.52) (1.52) % Population under 18 0.00043-0.00155 0.00168-0.00168 (1.20) (0.33) (1.42) (1.42) % Unemployment 0.00031-0.00758-0.00147 0.00147 (1.22) (2.34)* (1.81) (1.81) % Manufacturing Employees 0.00042-0.00117-0.00051 0.00051 (1.84) (0.40) (0.70) (0.70) % Foreign Born 0.00018-0.01920-0.00477 0.00477 (0.49) (4.13)** (4.08)** (4.08)** % Female-Headed Households -0.00056-0.01770-0.00305 0.00305 (1.33) (2.77)** (1.90) (1.90) Per Capita Income 0.00000 0.00001 0.00000-0.00000 (2.36)* (2.48)* (3.20)** (3.20)** % Black -0.00003 0.01128-0.00293 0.00293 (0.12) (3.95)** (4.08)** (4.08)** % Hispanic -0.00008 0.00945-0.00067 0.00067 (0.53) (4.24)** (1.20) (1.20) % Non-Family Households 0.00011-0.00013-0.00002 0.00002 (0.69) (0.06) (0.03) (0.03) Gini Coefficient 0.42925 0.82931 0.07830-0.07830 (16.42)** (2.43)* (0.91) (0.91) Constant 0.03029-0.00254-0.00612 0.00892-0.00896 0.01252 0.00896-0.01252

(24.05)** (0.55) (0.58) (0.14) (3.23)** (0.79) (3.23)** (0.79) R 2 0.02 0.82 0.00 0.54 0.05 0.60 0.05 0.60 N 116 116 112 112 112 112 112 112 ^ p<0.10; * p<0.05; ** p<0.01 H=Information Theory Segregation Index, Income, D=Dissimilarity Index, I=Isolation Index, E=Exposure Index b Within CBSA