In the News: Gentrification Is More Widespread Than You Think

By: Josh Ishimatsu, Director of Research and Capacity Building at National CAPACD
From: Rooflines
Posted: November 17, 2015


In Miriam Axel-Lute’s recent post here, “Place Matters But Place Changes,” she references “a study done by Governing magazine that found a 20 percent gentrification rate for census tracts in the past decade in the largest 50 cities in the country, a greatly accelerated rate from the previous decade.” She goes on to note that, while an increase over past rates of gentrification, a 20 percent gentrification rate still means that 4 of 5 low-income neighborhoods are not gentrifying.

These are basic, straightforward conclusions to draw from the Governing study.  However, there are a few huge, inter-related problems with the underlying study in being able to adequately describe our current round of gentrification.

The Current Round of Gentrification Is Driven by Rental Housing

The Governing study, and most quantitative studies of gentrification, use rising home prices as the primary indicator of gentrification. This is fairly standard practice and therefore not a problem unique to the Governing study. But the current round of gentrification is driven by the rental housing market, not the homeownership market.

The Joint Center for Housing Studies of Harvard University’s “The State of the Nation’s Housing 2015” report talks about the current decade as being “on track to be the strongest decade for renter growth in history.” The foreclosure crisis turned many homeowners into renters and made buying a home more difficult, and the share of people looking to rent housing is at a 20-year high. Rental housing construction, while strong, is not keeping up with demand, and new units are primarily being constructed for the high end of the market. Vacancy rates are down for all segments of the rental housing market, but especially for units renting at $800 or less per month. And, per the JCHS, “falling vacancy rates have lifted rents, improving the financial performance of rental properties, but straining the budgets of millions of households unable to find units they can afford.”

So, to be relevant, metrics of gentrification (at least for the current round of gentrification) need to reflect changes in the rental market.

Gentrification Is Happening Faster Than our Ability to Track It With Census Data

Not that I want to call out the study authors, per se, as this is not so much a problem of their analysis but with the limitations of the data (and about how we understand/talk about recently released studies which are based on old data). The study shows 1 in 5 low-income census tracts gentrifying in 2013 (actually it’s worse than that–the Governing study uses fiver-year ACS data, which means the numbers skew slightly more toward 2009 through 2013–an appropriate data set to use when conducting census tract-level analysis but not the best for measuring for trends that happen quickly. My ears-to-the-ground, anecdotal sense is that gentrification and displacement have intensified over the past year, even over the past several months, such that 2013 is a more accurate starting point to measure the current round of gentrification, not the ending point. The authors used the most recent data available from a reputable, public source–if I had designed the Governingstudy, I don’t know if I would have done it much differently. But the study doesn’t represent what is happening in low-income neighborhoods right now. And we shouldn’t mistake it for such.

And the past year has been a doozy in terms of rising rents. Looking at rental data from proprietary websites, which scour online rental listings to calculate regional rental rates, has its own methodological questions, but at least provides us with more up-to-the-month data. And this more recent data is dramatic. For example, over the time period October 2014 to October 2015, shows double digit percentage increases in median rents for 18 of the 50 cities that it tracks, with several more cities showing annual increases on the order of 9 percent. In contrast, from 2005 to 2014, the average annual increase in U.S. median rent was on the order of 3 percent per year. Per the table below, recent and dramatic run-ups in rents look to be widespread–in far many more places than as indicated by the Governing study.

480 Sources: [1] =; [2] = 2005 to 2014 1-year American Community Survey


The above chart is not straight apples to apples, but it at least shows that something more is happening than what can be seen within the Governing study time period–a real reason why right now, gentrification is, as Miriam says, “on everyone’s tongue… from Sacramento to Nashville to Lexington…”

Gentrification Isn’t Only a Problem for Low-Income Neighborhoods

This point probably deserves a blog post in and of itself. In looking for gentrification, theGoverning study authors define a baseline universe of census tracts that were eligible for being gentrified. Per their definition, all of these census tracts were low-income in 2000. That is, in seeking to define gentrification rates, the study tracks only low-income neighborhoods (or more precisely, neighborhoods that were low-income in 2000). As an example, the city of San Francisco has 196 census tracts of which only 16 were “eligible to gentrify” per the Governing study, and thus only 3 of these 16 showed evidence of gentrification. (A side observation here is that San Francisco was already so gentrified in 2000 that less than 10 percent of its neighborhoods were low-income in the context of the five-county San Francisco MSA!) The point here is that gentrification doesn’t only happen to low-income neighborhoods–it is also a problem in its impact on middle-income and mixed-income neighborhoods.

If we are focused on the impact of gentrification on low-income people, we should be concerned about gentrification in all neighborhoods. A typical census tract may have anywhere from 1,000 to 3,000 housing units. In most places, but especially in dense urban settings, a neighborhood of 1,000 to 3,000 housing units will contain many units that are above and below whatever the median home price or monthly rent is; this is the very definition of “median.” In gentrifying places, there is not only upward price pressure on all housing units, but particular pressure on the segments of the market that were "informally" affordable housing (i.e., cheaper housing but not formally covenanted as affordable), as whole buildings, for example, are flipped (emptied, fixed-up then re-marketed) to target a higher market segment. The median moves up as all prices rise–but also as the cheaper options disappear. Part of gentrification is that mixed-income places–regardless of whether the neighborhood started as a low-income neighborhood–become less mixed-income as low-income people are priced out.

Gentrification, therefore, is not just about what happens to low-income neighborhoods. It is the story of the loss of mixed-income neighborhoods, a phenomenon not completely registered within the Governing study. If we listen to Raj Chetty about mixed-income places producing greater economic mobility for low-income people, we should all care about the loss of these mixed-income neighborhoods too.

We need better measures of gentrification. Metrics that track rental data, that are more current, and that recognize that gentrification affects more than low-income neighborhoods. If we had such metrics, I think a much more widespread and insidious problem would be much more visible.


(Photo credit: courtesy of Jason Tester Guerilla Futures, via flickr, CC BY-ND 2.0)