The simple-minded comparison on city and suburb population growth rates is misleading and incomplete
Every year, in the late Spring, the Census Bureau releases its latest population estimates for the nation’s municipalities. That produces a raft of quick knock-on statistical analyses that flag which places are gaining and losing population. Inevitably, these estimates get aggregated to the national level, and pronouncements are made as to whether “cities” or “suburbs” are winning the race for more population.
It’s suburbs, by a neck – or is it?
The Brookings Institution’s Bill Frey published his nearly annual take on the Census Bureau population numbers late last month. His call on the City v. Suburb horserace: “the trend seems to be shifting toward a renewed suburban advantage.” That claim is likely to fuel arguments that America’s love affair with cities is over, that Americans (and especially Millennials) really want to live in the suburbs, and general harrumphing that urbanism is somehow past its peak.
As always, Frey’s math is impeccable. But we have to take issue with the analysis and the conclusions that people are likely to draw from the raw numbers. It’s too early, and this data is too ambiguous and incomplete to make any strong statements about the American desire for urban living. It turns out that municipal boundaries are problematic for making comparisons across space and over time; that when we take a longer-term perspective, the up-tick in city population is still quite noticeable (despite the data flaws), and finally, we ought to be paying much more attention to relative prices than relative growth rates, if we want to understand the value Americans place on urban living.
The limits of city limits
City boundaries are lousy for making these cross metro comparisons. Municipal boundaries aren’t standardized or consistent in any way. Some cities have copious amounts of low density sprawl inside their city limits (Jacksonville, San Antonio), while others are just a fraction of the urban core).
If we want to know the demand for urbanity and dense, walkable, transit-served city neighborhoods, you can’t just rely on aggregating census data for an entire city. What we and others have done to understand the varying spatial character of growth is to use much finer and more consistent measures, such as the 3-mile radius around the central business district. These data show a consistently strong performance for city centers in population and job growth.
This kind of aggregation also hides urban growth booms within cities. Chicago’s downtown and near North and South neighborhoods are booming, even as the city’s total population decreases (mostly due to declines in the lowest density parts of the city). In addition, population growth is down sharply in the region’s outlying suburbs.
What’s the baseline?
Whether cities are performing well or poorly relative to suburbs also depends on the time period one evaluates. Just looking at one or two year’s data isn’t necessarily very illuminating of long term trends. So it’s important to be explicit about the baseline for comparison. As we’ve pointed out before, while growth has slowed in the densest counties it remains above its long term trends there, and while growth has rebounded in less dense counties, growth remains below longer term trends.
Let’s take a closer look at what the Census data show, and how that’s changed over time. Frey (and others) make a lot of the fact that over the past year or two years, suburbs are growing faster than central cities. Here’s Bill Frey:
Now suburban growth again exceeds city growth, though at more modest levels than in the early 2000s, and, for many areas, due more to a city growth slowdown than a suburban growth pickup. It is still the case that city growth exceeds suburban growth in 17 of these 53 metropolitan areas, including Boston, Atlanta, Washington, D.C., San Francisco, and Seattle, though no longer New York (download table 2 here). Yet between 2010 and 2011, this was the case for 25 of these metros, and the trend seems to be shifting toward a renewed suburban advantage.
So the bad news is that only 17 central cities are outperforming their suburbs, down from 25 earlier this decade.
But consider where we were in the 1990s and the first decade of the 2000s. Helpfully, Bill Frey has been tracking this carefully in his Brookings Institution reports, so we went back to his 2012 retrospective on city growth trends to look at these two decades.
In the 1990s, only 7 central cities (in 53 top metros) grew faster than their suburbs: Charlotte, New York, Oklahoma City, Portland, Providence, San Antonio and San Jose.
From 2000 to 2010, only 8 central cities grew faster than their suburbs: Boston, Charlotte, Miami, Oklahoma City, Providence, Raleigh, Sacramento and San Jose).
So the way to look at it is this: How many cities outperformed their suburbs?
1990s: 7 cities
2000s: 8 cities
2010-11: 25 cities
2016-17 : 17 cities
So recent performance is a huge improvement over the previous two full decades, even if it’s not quite at the breakneck level of 2010-11. (And again, keep in mind that these municipal population figures are a weak and uneven proxy for whether urban living is becoming more or less popular). Also: This list has some what we might call “false positives” where the central city’s growth is really due to the fact that the main city encompasses a lot of low density sprawl (i.e. Charlotte, Oklahoma City, San Antonio), or which grew by annexation (i.e. Portland in the 1990s).
What does this say about the demand for city living?
The deeper question is whether city growth is slowing relative to suburbs because Americans don’t want to live in cities. There’s been a surge of suburban triumphalism, based on a quick reading of the city population data.
The bigger issue is we’re bumping up against the limits of city growth imposed by zoning. While cities could add population quickly for a few years by reducing vacancy rates, as housing fills up, you can only expand population by building more housing, which takes time, and in some cities (I’m looking at you San Francisco) is extremely expensive. A couple of articles make this point indirectly, arguing that Americans are moving to places where housing is cheaper.
As we’ve repeatedly shown at City Observatory, the premium that American’s are willing to pay for central locations has been steadily increasing. This constitutes what we’ve called the Dow of Cities. What really shows the value American’s attach to cities is the price they play for close-in urban housing, which is steadily rising relative to suburbs, virtually everywhere.
If city population growth isn’t keeping pace with suburbs, its not because Americans are desirous of suburban living, but rather because we have a shortage of cities.
The horse race metaphor makes for easy journalism, but in this case, doesn’t tell us much about the relative demand for city living. We really ought to be paying much more attention to prices.
Handicapping the City-Suburb Horserace
The simple-minded comparison on city and suburb population growth rates is misleading and incomplete
Every year, in the late Spring, the Census Bureau releases its latest population estimates for the nation’s municipalities. That produces a raft of quick knock-on statistical analyses that flag which places are gaining and losing population. Inevitably, these estimates get aggregated to the national level, and pronouncements are made as to whether “cities” or “suburbs” are winning the race for more population.
The Brookings Institution’s Bill Frey published his nearly annual take on the Census Bureau population numbers late last month. His call on the City v. Suburb horserace: “the trend seems to be shifting toward a renewed suburban advantage.” That claim is likely to fuel arguments that America’s love affair with cities is over, that Americans (and especially Millennials) really want to live in the suburbs, and general harrumphing that urbanism is somehow past its peak.
As always, Frey’s math is impeccable. But we have to take issue with the analysis and the conclusions that people are likely to draw from the raw numbers. It’s too early, and this data is too ambiguous and incomplete to make any strong statements about the American desire for urban living. It turns out that municipal boundaries are problematic for making comparisons across space and over time; that when we take a longer-term perspective, the up-tick in city population is still quite noticeable (despite the data flaws), and finally, we ought to be paying much more attention to relative prices than relative growth rates, if we want to understand the value Americans place on urban living.
The limits of city limits
City boundaries are lousy for making these cross metro comparisons. Municipal boundaries aren’t standardized or consistent in any way. Some cities have copious amounts of low density sprawl inside their city limits (Jacksonville, San Antonio), while others are just a fraction of the urban core).
If we want to know the demand for urbanity and dense, walkable, transit-served city neighborhoods, you can’t just rely on aggregating census data for an entire city. What we and others have done to understand the varying spatial character of growth is to use much finer and more consistent measures, such as the 3-mile radius around the central business district. These data show a consistently strong performance for city centers in population and job growth.
This kind of aggregation also hides urban growth booms within cities. Chicago’s downtown and near North and South neighborhoods are booming, even as the city’s total population decreases (mostly due to declines in the lowest density parts of the city). In addition, population growth is down sharply in the region’s outlying suburbs.
What’s the baseline?
Whether cities are performing well or poorly relative to suburbs also depends on the time period one evaluates. Just looking at one or two year’s data isn’t necessarily very illuminating of long term trends. So it’s important to be explicit about the baseline for comparison. As we’ve pointed out before, while growth has slowed in the densest counties it remains above its long term trends there, and while growth has rebounded in less dense counties, growth remains below longer term trends.
Let’s take a closer look at what the Census data show, and how that’s changed over time. Frey (and others) make a lot of the fact that over the past year or two years, suburbs are growing faster than central cities. Here’s Bill Frey:
So the bad news is that only 17 central cities are outperforming their suburbs, down from 25 earlier this decade.
But consider where we were in the 1990s and the first decade of the 2000s. Helpfully, Bill Frey has been tracking this carefully in his Brookings Institution reports, so we went back to his 2012 retrospective on city growth trends to look at these two decades.
In the 1990s, only 7 central cities (in 53 top metros) grew faster than their suburbs: Charlotte, New York, Oklahoma City, Portland, Providence, San Antonio and San Jose.
From 2000 to 2010, only 8 central cities grew faster than their suburbs: Boston, Charlotte, Miami, Oklahoma City, Providence, Raleigh, Sacramento and San Jose).
So the way to look at it is this: How many cities outperformed their suburbs?
So recent performance is a huge improvement over the previous two full decades, even if it’s not quite at the breakneck level of 2010-11. (And again, keep in mind that these municipal population figures are a weak and uneven proxy for whether urban living is becoming more or less popular). Also: This list has some what we might call “false positives” where the central city’s growth is really due to the fact that the main city encompasses a lot of low density sprawl (i.e. Charlotte, Oklahoma City, San Antonio), or which grew by annexation (i.e. Portland in the 1990s).
What does this say about the demand for city living?
The deeper question is whether city growth is slowing relative to suburbs because Americans don’t want to live in cities. There’s been a surge of suburban triumphalism, based on a quick reading of the city population data.
The bigger issue is we’re bumping up against the limits of city growth imposed by zoning. While cities could add population quickly for a few years by reducing vacancy rates, as housing fills up, you can only expand population by building more housing, which takes time, and in some cities (I’m looking at you San Francisco) is extremely expensive. A couple of articles make this point indirectly, arguing that Americans are moving to places where housing is cheaper.
As we’ve repeatedly shown at City Observatory, the premium that American’s are willing to pay for central locations has been steadily increasing. This constitutes what we’ve called the Dow of Cities. What really shows the value American’s attach to cities is the price they play for close-in urban housing, which is steadily rising relative to suburbs, virtually everywhere.
If city population growth isn’t keeping pace with suburbs, its not because Americans are desirous of suburban living, but rather because we have a shortage of cities.
The horse race metaphor makes for easy journalism, but in this case, doesn’t tell us much about the relative demand for city living. We really ought to be paying much more attention to prices.
Related Commentary