Association for Public Policy Analysis & Management

During the past decade and a half, college-educated and high-income individuals have increasingly located in urban neighborhoods, many of them originally low-income. The effects of this gentrification process on original residents are much debated but not well understood. On the one hand, in-movers could bid up the price of housing, increasing rents for original residents or causing them to leave the neighborhood. On the other hand, increasing income and education levels in the neighborhood could contribute to improved labor market opportunities or increased neighborhood quality, such as lower crime or better schools (Diamond 2016). The overall effect of gentrification on original residents is therefore ambiguous.

We use novel longitudinal data on individuals' locations, characteristics, and outcomes to describe the gentrification process and estimate causal effects of gentrification on the location and well-being of original residents. We construct a national panel of individuals and their locations, characteristics, and outcomes over time using unique Personal Identification Keys (PIKs) developed by the Census Bureau’s Center for Administrative Records Research and Applications (CARRA). We use PIKs to match individuals responding to both the 2000 Decennial Census long form and any year of the 2005-2014 American Community Survey. In both years, we observe in the Census data each individual's block of residence and block of work (if working); demographic characteristics; employment, wages, and income; homeownership status; and rent paid or home value. We assign each individual in each period to a geographically consistent neighborhood (tract) of residence, neighborhood of work, and metropolitan area. We construct measures of neighborhood quality from a number of sources (share poverty, school quality, crime, transit, and retail options) and match them to individuals by neighborhood of residence. The final sample includes millions of individuals in all urban neighborhoods in all metropolitan areas in the U.S.

These data allow us to establish a rich descriptive picture of gentrification in the U.S. We first explore how the process of gentrification works by disaggregating observed neighborhood education and income changes into combinations of in-moves, out-moves, and changes for stayers by individual education and income levels. We then explore out-moves of original residents and the kinds of neighborhoods they move to. Finally, we turn our focus from moves to the wages, rents, and neighborhood quality of all original residents, whether they move or stay. We describe these changes over time for original residents of gentrifying and non-gentrifying neighborhoods.

We then estimate the extent to which gentrification has a causal effect on the location, wages, rents, and neighborhood quality of original residents. We do this using instrumental variables and difference-in-differences methods to isolate plausibly exogenous increases in neighborhood housing demand. We perform various robustness checks to assess the validity of our results.

Overall, our results will provide new insight into how current trends transforming cities are affecting original residents. They will inform how policymakers can direct housing policies to avoid potential harms and harness potential benefits to help foster more inclusive, equitable, and higher opportunity neighborhoods and cities.