"The Myth of the Return to Cities"

[retired711]

https://www.nytimes.com/2017/05/22/...n-region&region=top-news&WT.nav=top-news&_r=0
https://www.nytimes.com/2017/05/22/...n-region&region=top-news&WT.nav=top-news&_r=0
The headline is too strong; actually the article shows some increased density in N.Y., Philadelphia and Chicago. But many areas are still expanding outward because many people can afford the suburbs more easily than cities. We've had discussions here before about urban density, and so I thought this New York Times piece might be interesting.

 

[Upstream]

NIRH is not going to be happy with you. He has 5 friends who prefer urban areas, and that is all the proof he needs to determine that millennials will always prefer cities, even as they grow into their 80s.

I actually interpret the data in that article slightly differently. With the exception of Seattle, the urban areas that became more dense only became slightly more dense (and some urban areas like LA and SF became slightly less dense). These slight changes are most likely within the range that you would expect from just random changes and normal movement of people. The slight differences show urban areas that are mostly stable, not indicative of increasing or decreasing density.

Seattle is the exception, and that is indicative of an urban area that is growing.

For the urban areas where you've seen significant declines in density, the article notes that the are urban areas that weren't particularly dense to begin with. So you are just seeing growth in those urban areas, but the growth is in the suburban sections because those were the larger sections to begin with. Again, it is not a trend toward high or low density. It is just a trend to continued growth of what previously existed.

 

[retired711]

NIRH is not going to be happy with you. He has 5 friends who prefer urban areas, and that is all the proof he needs to determine that millennials will always prefer cities, even as they grow into their 80s.

I actually interpret the data in that article slightly differently. With the exception of Seattle, the urban areas that became more dense only became slightly more dense (and some urban areas like LA and SF became slightly less dense). These slight changes are most likely within the range that you would expect from just random changes and normal movement of people. The slight differences show urban areas that are mostly stable, not indicative of increasing or decreasing density.

Seattle is the exception, and that is indicative of an urban area that is growing.

For the urban areas where you've seen significant declines in density, the article notes that the are urban areas that weren't particularly dense to begin with. So you are just seeing growth in those urban areas, but the growth is in the suburban sections because those were the larger sections to begin with. Again, it is not a trend toward high or low density. It is just a trend to continued growth of what previously existed.

Good analysis; you read the article more carefully than I did. As for NIRH, there is a thread on the Current Events board, speculating humorously (or not so humorously) on what he might be doing instead of posting here.

 

[MoobyCow]

I think they have some poor measures, or at least imprecise measures.

When you say a metropolitan area became less dense, that's actually impossible with an increasing population, unless you've now expanded the metropolitan area. Things that were formerly outside of the metro area, are now inside. That would seem to indicate those areas became more dense than they were previously, otherwise they'd still be outside of the metro area.

So yeah, Austin metro is less dense than it was, but the actual individual blocks they are counting have to be more dense, almost by definition. It's just the new area added to the metro is less dense than the existing area, which is completely unsurprising.

IMO, the data does not show what they say it shows, not by a long shot.

 

[retired711]

I think they have some poor measures, or at least imprecise measures.

When you say a metropolitan area became less dense, that's actually impossible with an increasing population, unless you've now expanded the metropolitan area. Things that were formerly outside of the metro area, are now inside. That would seem to indicate those areas became more dense than they were previously, otherwise they'd still be outside of the metro area.

So yeah, Austin metro is less dense than it was, but the actual individual blocks they are counting have to be more dense, almost by definition. It's just the new area added to the metro is less dense than the existing area, which is completely unsurprising.

IMO, the data does not show what they say it shows, not by a long shot.

Imagine ten people living in an apartment house. There are two babies born. But rather than stay in the apartment house, they split into two groups, living at a substantial distance from each other but still in the metropolitan area. It seems to me that in that instance, density goes down while the population goes up. In other words, the measure of density is not simply a matter of dividing the area by the population. It also has to do with the distribution of the population within the metropolitan area.

 

[MoobyCow]

Imagine ten people living in an apartment house. There are two babies born. But rather than stay in the apartment house, they split into two groups, living at a substantial distance from each other but still in the metropolitan area. It seems to me that in that instance, density goes down while the population goes up. In other words, the measure of density is not simply a matter of dividing the area by the population. It also has to do with the distribution of the population within the metropolitan area.

Nope. The metro area is the same size and there are two new people. The density has gone up for the entire area, there is no way to add to the count of people (or anything else) in a given fixed area and reduce the density. That is simple math and the very definition of density

 

[retired711]

Nope. The metro area is the same size and there are two new people. The density has gone up for the entire area, there is no way to add to the count of people (or anything else) in a given fixed area and reduce the density. That is simple math and the very definition of density

Still don't agree. Density is calculated per square mile or kilometer, not by the whole area. So when people spread out within a metropolitan area, density can go down even with the total area's population going up.

http://www.statisticshowto.com/population-density-definition/

 

[MoobyCow]

Still don't agree. Density is calculated per square mile or kilometer, not by the whole area. So when people spread out within a metropolitan area, density can go down even with the total area's population going up.

http://www.statisticshowto.com/population-density-definition/

Let's try this again. Density is calculated by the square mile, but we're not looking at an individual square mile we are looking at a metropolitan area, which is a set number of square miles. The population density of a city is a simple calc population/SQ miles. If you don't change the number in the denominator, but you make the numerator larger, the answer is a larger number.


Even your link says SQ miles. You calculate the density by population over the total number of SQ miles of the whole area. Otherwise you could never say what the density of an area larger than a mile is, and that's pretty obviously not true.

Now, maybe you're saying that you calculate all the individual SQ miles and average them, but that will get you the same answer.

 

[retired711]

What is your source? Suppose we have ten people in one square mile disperse to being six each in another square mile. Instead of a density of 10, we have a density of 6 (12/2). Suppose the same happens in every pair of square miles. Density has decreased even though population has risen. You seem determined that the measure of density is calculated per metropolitan area; instead, it is the average of individual square miles within the area.

 

[MoobyCow]

You think they strike an overall average; I think they aggregate the individual square miles without averaging them. I guess we'll have to agree to disagree. Still, I think my measure is a superior way to measure density.

That's nonsensical. What then would be your calculation for an area of more than 1 mile?

If you aggregate them, what is the denominator?

And, while I want this to be a polite discussion, it is not an agree to disagree type of question.

Just to give you a little background, this is directly related to my area of study in college (though I no longer work in the field), and my wife is on the national board of city planners, was the head on the NJ planners association and we discuss these matters often.

I am not just picking a fight, I know the correct answer and I am trying to explain it to you.

 

[retired711]

The denominator is obviously the number of square miles. I respect you and your wife's experience, but I cannot believe that a situation in which people are concentrated in one part of a metropolitan area is equivalent to one in which they are dispersed well over that area. I apologize if you think I'm not respecting you or your wife.

 

[retired711]

I'm not sure you're right, but I yield to your experience in the field. Note, however, that even northeastern metropolitan areas expand in size over time.

 

[MoobyCow]

The denominator is obviously the number of square miles. I respect you and your wife's experience, but I cannot believe that a situation in which people are concentrated in one part of a metropolitan area is equivalent to one in which they are dispersed well over that area. I apologize if you think I'm not respecting you or your wife.

It is not equivalent, it is an average, with all the qualifications that come when talking averages. The same as average salary or life expectancy etc. Outliers can have an impact.

That said, it is also generally a good measure. If everyone was clustered together in one spot, and no one in another, the second area, with few people, would not be part of the metropolitan area (assuming they were in a different county, see below).

As to your second post, yes NE met areas grow over time as well, but much more slowly as they are more fully developed.

 

[MoobyCow]

Also of note... Metropolitan areas are generally composed of counties, and when you have larger counties, as they do out West, you can add land mass quickly, even though the people may be in one spot in that large county. The problem with averages..
http://www.referenceforbusiness.com/small/Mail-Op/Metropolitan-Statistical-Area-MSA.html

 

[T2Kplus10]

NIRH is not going to be happy with you. He has 5 friends who prefer urban areas, and that is all the proof he needs to determine that millennials will always prefer cities, even as they grow into their 80s.

I actually interpret the data in that article slightly differently. With the exception of Seattle, the urban areas that became more dense only became slightly more dense (and some urban areas like LA and SF became slightly less dense). These slight changes are most likely within the range that you would expect from just random changes and normal movement of people. The slight differences show urban areas that are mostly stable, not indicative of increasing or decreasing density.

Seattle is the exception, and that is indicative of an urban area that is growing.

For the urban areas where you've seen significant declines in density, the article notes that the are urban areas that weren't particularly dense to begin with. So you are just seeing growth in those urban areas, but the growth is in the suburban sections because those were the larger sections to begin with. Again, it is not a trend toward high or low density. It is just a trend to continued growth of what previously existed.

[roll]

 

[T2Kplus10]

Why suburbs? Because sooner or later most people start families:

http://www.njfamily.com/Best-Towns-for-Families-The-List/

And cities suck for families.

 

[NotInRHouse]

Headline is misleading- several cities have become more dense, including this area.

The suburbs in NJ doing well are the train and/or excellent school ones.

Places like Sussex and Warren Counties shrink while Hudson gets bigger. Let Austin do Austin. Ain't happening here.

 

[Upstream]

When you say a metropolitan area became less dense, that's actually impossible with an increasing population, unless you've now expanded the metropolitan area.

Mooby -- your arguments here are correct when looking at total density for a metro area. But they aren't looking at total density for a metro area. They're looking at the average density of neighborhoods within a metro area.

The average density of neighborhoods within a metro area can drop with population increase (or population stability) if the neighborhoods are not all the same size. And that is probably a fair assumption where higher density neighborhoods in urban cores are typically smaller geographically than less dense suburban neighborhoods in the same metro area. So as people move from the urban core to suburban areas, or if population growth in suburban areas is faster, then average neighborhood density will fall.

This actually a pretty good proxy to see if the growth is in the urban core or suburban areas of metro areas. However, as I indicated earlier in this thread, I don't agree with the conclusion.

 

[MoobyCow]

Mooby -- your arguments here are correct when looking at total density for a metro area. But they aren't looking at total density for a metro area. They're looking at the average density of neighborhoods within a metro area.

The average density of neighborhoods within a metro area can drop with population increase (or population stability) if the neighborhoods are not all the same size. And that is probably a fair assumption where higher density neighborhoods in urban cores are typically smaller geographically than less dense suburban neighborhoods in the same metro area. So as people move from the urban core to suburban areas, or if population growth in suburban areas is faster, then average neighborhood density will fall.

This actually a pretty good proxy to see if the growth is in the urban core or suburban areas of metro areas. However, as I indicated earlier in this thread, I don't agree with the conclusion.

That would generally be referred to as weighted density (and when it is used it is often explicitly stated the methodology) which isn't what the appeared to be saying to me as the source was county population estimates (accourding to the chart) and this calc would generally be done at the tract level and I would be surprised to find that the tract level information was accurate enough on an estimated year like 2016, to make that a meaningful calc.

Though, if so I owe Camden an apology and a beer.

 

[Upstream]

It's kind of hard to figure out the methodology based on the article ... and I'm not sure if there is an underlying study we can reference.

 

[BoroKnight]

Four of the five fastest-growing "cities" in the past year, according to an AP story I had to edit less than four hours ago, are "suburbs" of Houston, Austin and Dallas/Fort Worth.

Woodbridge, Edison and Lakewood are among the most populated "cities" in New Jersey.

Yeah, it gets tricky to measure these things. "Metropolitan area" is a very bad way to measure what this thread wants to talk about.

 

[retired711]

It is not equivalent, it is an average, with all the qualifications that come when talking averages. The same as average salary or life expectancy etc. Outliers can have an impact.

That said, it is also generally a good measure. If everyone was clustered together in one spot, and no one in another, the second area, with few people, would not be part of the metropolitan area (assuming they were in a different county, see below).

As to your second post, yes NE met areas grow over time as well, but much more slowly as they are more fully developed.

That would generally be referred to as weighted density (and when it is used it is often explicitly stated the methodology) which isn't what the appeared to be saying to me as the source was county population estimates (accourding to the chart) and this calc would generally be done at the tract level and I would be surprised to find that the tract level information was accurate enough on an estimated year like 2016, to make that a meaningful calc.

Though, if so I owe Camden an apology and a beer.

You do *not* owe me an apology or a beer. Even if you were wrong (which we don't know), you were much closer to being right than I was.

 

[srru86]

Slate
Something is Wrong with Connecticut
And they don't even talk about putting your football stadium and hour away from campus on a abandoned air strip.
But seriously folks...we in NJ face some of the same challenges covered in the article.

 

[e5fdny]

You do *not* owe me an apology or a beer. Even if you were wrong (which we don't know), you were much closer to being right than I was.

And this type of response is why I would want to have a beer with YOU.

Cheers, mate! :beer: