Gentrification is a concern in communities across the country. As previously low-income neighborhoods see an influx of more affluent households, they are accompanied by rising rents, changing demographics, and a reshaping of the economic and cultural landscape. The preexisting residents see these changes in their community and worry that they and their neighbors may be forced to move as housing costs increase. Yet when people raise these concerns, they are pointed to a variety of quantitative studies arguing that their fears are unfounded—that gentrification does not displace people (Brummet and Reed 2019; Dragan, Ellen and Glied 2019; Ellen and O’Regan 2011; Freeman 2005; McKinnish, Walsh and White 2010). These studies are puzzling, given that they are both counterintuitive to what residents perceive on the ground, and that they contrast with other qualitative studies showing that displacement indeed occurs in gentrifying neighborhoods. Nonetheless, the finding is consistent across multiple studies drawing on diverse, high-quality datasets.
In an article in City & Community, I argue that when scholars say they are studying “displacement,” they are actually measuring quite different things. Drawing on data from the New York City Housing and Vacancy Survey, I compare commonly used measures of displacement and find that the discrepancies in what they are measuring are not trivial, and can lead to different conclusions from their research. I also find that the correlation between gentrification and displacement exists with some measures of displacement but not others. In doing so, I also document a variety of mechanisms that are missed by the other approaches to measuring displacement.
Measuring displacement
First of all, how do we define displacement? There are a variety of things that people mean when they use the word, including housing, cultural, and political displacement, among others. For my purposes, I define displacement as “involuntary residential mobility”—people moving from their homes when they did not want to move. This is important if we want to distinguish people moving because they were pushed out, versus people moving of their own volition to buy a home, take a new job, or find a larger residence. Neighborhood change could happen without displacement if people move and are then replaced by someone who is more affluent, higher educated, white, or whichever attribute we are concerned with. Indeed, Lance Freeman’s seminal study, “Displacement or Succession,” argues that gentrification does not displace people, but rather replaces them (Freeman 2005).
If we accept “involuntary mobility” as the underlying idea of displacement, then that implies two components: that someone moved, and that their move was in some way involuntary. I investigate whether prior researchers use one or both of those criteria in identifying displacement.
In a review of the literature, I find three general ways that people measure displacement—what I call a population approach, an individual approach, and a motivational approach. The population approach looks at the changing composition of a neighborhood between two points in time. Often using census data, researchers take their population of interest (low-income people, people of color, etc.) and measure them at two points in time. If, for example, there are 1,000 fewer low-income residents at the later period in time, then they say that 1,000 people were “displaced.” While this is one of the most common methods of identifying displacement, it does not directly measure people moving or whether those moves were involuntary. Another common method is the individual approach, which uses data that does identify individuals to see if they leave their neighborhood, but without ascertaining the reasons why people move. Lastly, the motivational approach both measures individual mobility and asks the reasons why someone moved, in order to distinguish whether a residential move was to buy a house or take a new job, versus moving in response to a rent increase or an eviction. [1]
There are multiple and complex reasons why someone moves, so all of these measures should be considered “proxies” for displacement. While I argue that the motivational approach is the closest direct measure of displacement, since it measures both mobility and whether it was involuntary, the others may be defensible if the data is more readily available and they are close enough to the motivational approach. I find that this is not the case.
To test this, I used each of the three approaches to separately rank all the neighborhoods across New York City by the volume of displacement occurring there. For each approach, the neighborhood with the most amount of displacement was ranked 1 and the least amount was ranked 55. Figure 1 compares both the population and individual approaches against the motivational approach by plotting their rankings and a best-fit line. A 45-degree line would indicate perfect correlation—that two approaches identically ranked a neighborhood with the highest quantity of displacement, followed by the second-most displacement, and so on. While allowing for some degree of discrepancy between approaches, a good proxy should have a very high degree of correlation, suggesting that one measure could be substituted for another. The graphs below show correlations of individual versus motivational (statistically significant) and population versus motivational (not statistically significant). While the individual approach is a better approximation of the motivational approach, there is still a high level of divergence, suggesting that they are measuring two distinct things.
Source: author’s own work.
Unmasking processes of neighborhood mobility
Why do these approaches to measuring displacement get such different results? I argue that it is because they miss some underlying processes. First, since the population approach only relies on snapshots of communities to see if their compositions change, it cannot identify people who might be displaced from their home but not their neighborhood. Indeed, I find that about a third of New York City residents find some way to stay in their neighborhood after they have been displaced. Second, the population approach cannot capture the in‑displacement of people into neighborhoods. If one low-income family is displaced from a neighborhood but another low-income family is able to find an affordable apartment in that neighborhood, then there would be a net zero change in the composition of the neighborhood. I find that every neighborhood in New York City both displaces people and receives displaced people. Third, by only counting the changes to low-income population, this approach misses displacement that may occur higher up on the income distribution. Thus, from these first three issues, the population approach risks under-counting the amount of displacement in a neighborhood. Lastly, the individual approach assumes that people moving from a neighborhood for any reason is proportional to the number of people displaced from it. I find that this assumption does not hold.
This study holds a variety of lessons for researchers, policymakers, and community leaders for how we understand gentrification and displacement. It means that the gentrification–displacement debates are not settled, since the dependent variable—displacement—is subject to wide variability. This is on top of the unsettled and highly differentiated ways that researchers measure gentrification (Preis et al. 2020). Displacement is still an issue across urban neighborhoods, though gentrification is not the only cause for it. This study also raises questions for future research, such as examining under what conditions people might be able to stay in their neighborhood after being displaced, and when displaced people from other neighborhoods might still be able to enter into gentrifying neighborhoods. People are displaced from non-gentrifying communities as well. The pressures leading to displacement also may lead to people being “squeezed” in their neighborhoods as housing costs rise, but not to the point of pushing a household to move—at least for the moment. By distinguishing between people being pushed from their homes versus communities becoming more affluent and whiter in their composition, it suggests that policymakers need to focus on both sides of the coin—on establishing renter protections for people to stay in their home as well as increasing affordable housing options in gentrifying communities.
Bibliography
- Brummet, Quentin and Reed, Davin. 2019. The Effects of Gentrification on the Well-Being and Opportunity of Original Resident Adults and Children, working paper, July, Philadelphia: Federal Reserve Bank of Philadelphia.
- Dragan, Kacie; Ellen, Ingrid; and Glied, Sherry. 2019. Does Gentrification Displace Poor Children? New Evidence from New York City Medicaid Data, working paper w25809, Cambridge (Massachusetts): National Bureau of Economic Research.
- Ellen, Ingrid Gould and O’Regan, Katherine M. 2011. “How Low Income Neighborhoods Change: Entry, Exit, and Enhancement”, Regional Science and Urban Economics, vol. 41, no. 2, pp. 89–97.
- Freeman, Lance. 2005. “Displacement or Succession? Residential Mobility in Gentrifying Neighborhoods”, Urban Affairs Review, vol. 40, no. 4, pp. 463–491.
- McKinnish, Terra; Walsh, Randall; and White, T. Kirk. 2010. “Who Gentrifies Low-Income Neighborhoods?”, Journal of Urban Economics, vol. 67, no. 2, pp. 180–193.
- Preis, Benjamin; Janakiraman, Aarthi; Bob, Alex; and Steil, Justin. 2020. “Mapping Gentrification and Displacement Pressure: An Exploration of Four Distinct Methodologies”, Urban Studies, 20 February, 20 pp.
