May 9, 2019

Enterprise Releases New Interactive Tool for Comparing Definitions of Gentrification

Gentrification in communities

Enterprise today is launching an unparalleled, interactive tool that examines one of the most contentious topics in community development: gentrification. 

The Gentrification Comparison Tool (GCT) is the first online resource to present multiple definitions of gentrification side-by-side. It allows users to analyze three different definitions of gentrification in 93 U.S. cities over four decades, providing a hyper-local view of changing conditions in over 14,000 neighborhoods. The data are from the 1970-2010 Neighborhood Change Database, purchased under license from Geolytics, Inc., which normalizes tract-level characteristics to 2010 tract boundaries for all Decennial Census years.

The GCT complements to our 2018 report, Gentrification: Framing Our Perceptions, which looked at how gentrification is defined across dozens of recent studies. The report found that, while gentrification is generally understood as the change that occurs in low-income communities experiencing rapid residential turnover because of the in-migration of higher-income households, precise definitions are much murkier. Rather than coalesce around a standard approach, most researchers develop their own methods for identifying neighborhood change, with different indicators of income change, residential turnover, and housing conditions all deployed as proxies for gentrification.  

This lack of consensus around how to identify gentrification leads to inconsistent findings about where and how it occurs, as well as its consequences for low-income neighborhoods. These results in turn suggest different needs for policy interventions, either to encourage revitalization of formerly-distressed neighborhoods or to mitigate the negative consequences for existing residents. 

With the GCT, users can now visualize what our 2018 report described – how different definitions of gentrification often identify different sets of changing neighborhoods. The tool operationalizes measures derived from three well-known studies: Freeman (2005), Ellen & O’Regan (2008), and McKinnish et al. (2010). It then applies these definitions to a common set of data – Decennial Census tract-level reports from 1970 to 2010 – to document the effect of different approaches on observed results. 

All three definitions use a two-step process to measure gentrification. They first identify a set of neighborhoods deemed “eligible” to gentrify (i.e. sufficiently low-income and disinvested) based on their economic and housing conditions at the start of each decade. From this subset, they then distinguish those areas that exhibited enough improvement over the decade to suggest that gentrification occurred. Each measure thus creates three categories of neighborhoods: gentrified, eligible but not gentrified (i.e. remained low-income over the decade), and not eligible (i.e. higher-income at the start of the decade).

The GCT maps the results of this analysis side-by-side-by-side for all 93 U.S. cities that have at least one eligible tract under all three definitions in each of the four most recent decades (the 1970s through the 2000s). By selecting a city and decade, users can view the neighborhoods each definition classifies as gentrified, eligible but not gentrified, and not eligible. The tool also shows the inputs into each definition, so users can determine what drives differences in the tract-level classifications. Finally, it calculates the overlap across the three definitions in the number of places deemed gentrified in each city and decade.


An example of the tool’s output is visible in the screenshot above, which shows gentrification in Washington DC in the 2000s according to each definition. The maps use a distinct color – green for Freeman, red for Ellen & O’Regan, and purple for McKinnish et al. – with the darker shades signaling gentrified tracks, and the eligible-but-not-gentrified tracts in the lighter shades.  (Tracts shown in white represent the non-eligible tracts within each measure, and grey areas lie outside the city border.) 

Freeman’s definition (on the left, in green) identifies the largest number of gentrified tracts in the city, covering much of the northeast and southeast parts of the capital. Ellen & O’Regan’s definition (center, red), meanwhile, finds fewer gentrified tracts, which are mostly concentrated near the center of the city. Finally, McKinnish et al.’s definition (right, purple) has the fewest eligible and gentrified tracts, indicating that most of the city was too high-income (as of 2000) to be able to gentrify. 

It is important to note that our new tool is not intended to suggest that one measure is more accurate or relevant than the others. Each of these – indeed, every measure of gentrification – reveals something about how neighborhoods change over time. The problem lies in labeling all these different approaches “gentrification,” which obscures that word’s meaning to the point of irrelevance. Nor can any set of aggregated data provide a complete picture of what occurs, at a more granular level, to people, buildings, institutions and communities. Without that local context and perspective, no quantitative measure of gentrification will ever give the full story.

Enterprise invites users to explore these definitions of gentrification and apply their knowledge about different cities and neighborhoods at different points in time to the data provided. Key questions to consider include: 

  • How different are the results from each definition? 
  • How do the results of these definitions align (or not) with your own understanding about neighborhood-level changes in these places over time? 
  • What do these definitions lack that would create a more complete perspective? 

We hope this tool introduces some much-needed nuance to conversations around gentrification and how it is identified, especially in quantitative analyses based on aggregated data sets. While such analyses are valuable for revealing broad changes in neighborhood conditions, they say little about what these shifts mean for the low-income communities experiencing increased residential demand from higher-income households. Future analyses of gentrification should also be more forthcoming about what their measures do and do not capture, and more cautious with the conclusions they draw about where and why gentrification occurs.

Opp360 logo

For full access to our tools and resources, please provide the information below.

We use this data to better understand our users; we do not sell or share this data. By providing this information, you can expect to receive newsletters and other updates from Opportunity360.