Transit-Induced Gentrification in U.S. Metropolitan Areas

Abstract

Problem: Gentrification is a term used to describe the process and changes that commonly occur in lower-income and/or minority neighborhoods with the influx of more affluent residents who are more likely to be white, increases in property values, the renovation of housing, the upscaling of local commercial and retail properties, and potentially, the displacement of current residents. Some studies in the past two decades have found that the sustained growth of the high-quality public transit systems—both rails and buses—may have triggered or accelerated gentrification in some U.S. metropolitan areas (MSAs). Some studies define that phenomenon as the “Transit-Induced Gentrification” (TIG). Until now, nineteen empirical studies have examined TIG in one or several cities, with twelve of them focused on American cities. Some fundamental studies have been conducted to explain TIG based on previous theories and hypotheses of the traditional gentrification. Through the summary of literature, three unsolved issues on TIG have been identified. First, current studies have not reached consensus on the pervasiveness of TIG, partly because of their different operational definitions and measures of gentrification and their different research designs. Second, current studies have not found sufficient empirical evidence to support the hypotheses explaining TIG, and factors associated with the probability of TIG are not clear. Third, scholars are still not clear about whether displacement always happens during the TIG. Research strategy and findings: This dissertation is designed to address the first two unresolved issues. With a quasi-experimental design, this dissertation examines the hypothesis of TIG in all neighborhoods newly served by rapid transit stations that opened from 2000 through 2009 across the U.S. This dissertation confirms that TIG is likely but not inevitable by comparing the pretest-posttest results between all new rapid-transit-served neighborhoods and a control group selected by nonparametric propensity score matching that controls for neighborhood characteristics and the impact of Great Recession. This dissertation provides the first comparison of the likelihood of gentrification associated with both rail and bus rapid transit (BRT) and shows that rail stations are more likely to induce gentrification than BRT stops. This dissertation also shows that TIG is more evident over long-term than over short-term for rail-served neighborhoods. Methodologically, although some previous studies have used Census block groups (CBGs) as the areal unit of analysis, most have used Census Tracts (CTs), and none has compared results from simultaneous analysis using both CBGs and CTs. This dissertation makes a contribution by comparing results from using both CBGs and CTs as the areal unit of analysis. The comparisons show that CBG-based analyses better approximate the areas served by transit stations, are more consistent with theory, and therefore provide more valid results. This dissertation also applies the multi-level (hierarchical) logistic regressions to identify and examine factors that are likely to be associated with the probability of TIG, including both MSA and neighborhood characteristics. The results show that MSA characteristics are less stable and provide less and evidence of the probability of TIG than neighborhood characteristics. Some socioeconomic characteristics of neighborhoods, mainly measures of poverty, show consistent significance in the examinations for their impact on the likelihood of TIG. Take Away for Practice: The findings of this study have some policy implications. The BRT is less likely to induce gentrification compared with rail transit, and thus could help sustain the transit service to the most vulnerable without the same likelihood of gentrification. The identification of neighborhoods with higher probability of TIG, such as the neighborhoods with higher poverty rates and people of color and lower proportions of college-educated residents, enables policy-makers and urban planners to target policies such as affordable housing and rent ceilings to assist the most vulnerable areas and residents. In addition, the use of different definitions and measurements of TIG results in substantial differences in the percentages of neighborhoods classified as experiencing gentrification and in identification of different factors that affect the probability of TIG. These findings can be interpreted as evidence that policy makers and planners need to involve stakeholders, especially the low-income and people of color who are more vulnerable to gentrification, in their deliberations over definitions of TIG and when establishing anti-gentrification policies.