Wendell Cox and his population density daffiness - 2

As stated in a previous post, se find it difficult to understand why Wendell Cox expends time and effort comparing the gross population densities of various cities and metropolitan areas - without reference to land area and other important factors. His approach might make sense if there was a point to make. However, the relationship between gross population density - nothing else considered - and transportation needs, traffic levels and so forth is less strong that Cox implies (albeit tacitly). Professionals and advocates know this well, and we think that Cox does, too.

It is difficult to understand how the "number of people per km2" (or sq mi) is meaningful in any context other than place-specific descriptions that include other important statistics. Capital Cities in Comparison (Statistik Berlin, 2002 pdf format) compares Berlin, Budapest, Paris, Prague and Warsaw, and includes small sketches of each city - to the same scale.

(A "same-scale" resource of broader interest: Subway systems of the world, presented on the same scale, Neil Freeman, 2004.)

We posed a few questions for Cox in a previous post. We would like to know what reasonable conclusions, related to transportation planning or analysis, might one draw from comparisons of two (or more) places with vastly different land areas based on gross "population per km2" or sq mi - with land area adjusted on an ad hoc basis. We would also like Cox to explain his apparent belief that adjustments for differences in land area are not to be made - e.g. by comparing a Paris-sized area of New York City with Paris proper. (We don't expect to receive a answer anytime soon.)

We admit: if all metropolitan areas were "unicities," having a single municipal administration for the central city and all suburbs, then Cox's comparisons would have some degree of utility. "Some:" it would still be necessary to take factors other than gross population density into account when considering transportation needs. However, most metropolitan areas worldwide are not "unicities" - and most conurbations described as "unicities" (e.g. Calgary, Indianapolis, Winnipeg) do have some population housed outside of the "unicity" boundary.

Administrative boundaries within urbanized areas reflect historic and political factors that are specific to countries, regions and metropolitan areas. Thus, differences among cities with respect to the location of administrative boundaries are, in effect, arbitrary. In one U.S. state, it is possible to argue that the location of administrative boundaries is indefinite.

Texas cities have "extraterritorial jurisdictions" (ETJs).

The ETJ of a city is the contiguous unincorporated land adjacent to its corporate limits that is not within another city's ETJ. The size of a city's ETJ varies according to its population, ranging from one-half mile for communities with less than 5,000 persons, to five miles for cities greater than 100,000 in population (Annexation Policies, Planning Department, City of San Marcos).

ETJs have no municipal government of their own. The adjacent city exercises limited authority within its ETJ. All subdivision and road development within the ETJ must be approved by the city (Ryan, Molly. 2010. MUDs and Houston profit from annexation system. Commuity Impact Newspaper, Friday, 20 August 2010.

Texas provides no financial assistance to municipal governments. However, it permits municipal governments to exercise powers to annex land.

Annexation is the process by which a city extends its municipal services, regulations, voting privileges and taxing authority to new territory. Cities annex territory to provide urbanizing areas with municipal services and to exercise regulatory authority necessary to protect public health, safety and welfare. Annexation is also a means of ensuring that residents and businesses outside a city's corporate limits who benefit from access to the city's facilities and services share the tax burden associated with constructing and maintaining those facilities and services. Annexation may also be used as a technique to manage growth (Annexation Policies, Planning Department, City of San Marcos).

Two types of annexations exist under Texas law: "general-purpose" and "limited purpose" (Ryan 2010.)

Because Texas cities exercise significant control over unincorporated land located outside of their formal corporate boundaries, it is possible to argue that the location of administrative boundaries within urbanized areas is indefinite. This, in turn, implies that significant imprecision is associated with crude population density statistics.

In Texas, cities may annex land located across a county line. Online maps facilitate comparisons between Dallas County (Dallas)and Harris County (Houston, pdf format) in terms of the relative shares of land within the county that is also located within the urban core.




City population



City land area

887.2 km2 (342.5 sq mi)

1,501 km2 (634 sq mi)

County population



County land area

3,649 km2 (1,407 sq mi)

4,478 km2 (1,729 sq mi)


Notes: Population statistics for the city of Dallas and Dallas County are 2010 estimates. Statistics for the city of Houston, and Dallas and Houston metropolitan areas, are 2009 estimates. The statistic for Harris County is a 2008 estimate.

Dallas accounts for nearly 39 percent of the land area of Dallas County. Houston accounts for nearly 37 percent of the land within Harris County - but Harris County has more than twice the land area of Dallas County, and Houston has nearly twice the land area of the city of Dallas. In addition, Houston accounts for a much larger share of its metropolitan area (Houston - Sugar Land - Baytown) than Dallas (Dallas - Fort Worth - Arlington).

Some countries and sub-national jurisdictions facilitate the growth of cities through annexation of adjacent others. Others restrict or prevent this. A very good example is France - with specific reference to Paris. For historic reasons, the French state has confined the area controlled by the municipal government of the French capital to a small area - about four percent of the total urbanized area.

If the "underlying issue" of interest is "residential density" - the number of people per square unit of "residential land" - then one will need to do more than merely divide gross population by gross land area. In the U.S., residential density is expressed typically as people per acre - that's "per acre" of "residential land," used for some form of dwelling (plus garage, back yard and so forth). One wanting to obtain such information will probably need to contact the metropolitan planning organization (MPO) for the area(s) of interest. This statistic cannot reasonably be estimated by dividing crude regional population by crude land area.

What does all of this population-density stuff "prove?"

A few years back, we, um, "discussed" this with a Cox supporter. We won't use the density statistics he quoted (they proved to be incorrect), but we will use the cities (with the single addition of Paris). These we have listed in ascending order of crude population density (links are to aerial images):

Sankt-Peterburg, Russia: 3,239 / km2 (8,389 / sq mi, 2002).

Moskva, Russia: 9,682 / km2 (25,080 / sq mi, 2002).

New York City (NY), US: 10,630 / km2 (27,532 / sq mi, 2009 est.).

Manhattan, NYC, US: 27,394 / km2 (70,951 / sq mi, 2009 est.).

Paris, France: 20,807 / km2 (53,980 / sq mi, 2007)

Hong Kong SAR, China: (31,825 / km2 (82,791 / sq mi, 2006).

Hong Kong statistic excludes Southern District of Hong Kong Island, and all parts of the New Territories except for the Kai Tsing District.

We have listed the same cities above in (ascending) order of annual passenger traffic density carried by metro systems (links are to Urbanrail.net pages):

Paris, France: 32 million pass-km per km of system length (2004).

New York City (NY), US: 36 million pass-km per km of system length (2005).

Sankt-Peterburg, Russia: 78 million pass-km per km of system length (2006).

Hong Kong SAR, China: 80 million pass-km per km of system length (2005).

Moskva, Russia: 113 million pass-km per km of system length (2005).

Based on crude population density as the "predictor" factor, the worlds "busiest" metro system (in terms of annual passenger traffic density) "should" be Hong Kong's MTR.

However, this is not true. The metro system that carried (at 2005) the world's highest annual passenger traffic density was the Moskva Metro.

(The full formal title of this undertaking translates as something like "'State Unitary Enterprise for the City of Moskva, [titled] "The V. I. Lenin 'Order of Lenin' and 'Order of the Red Banner of Labor' Metropolitan Railway of Moskva."'")

Although the area served by the Hong Kong MTR had a gross population density more than three times greater than Moskva, the Moskva Metro carried an annual passenger traffic density more than 40 percent greater.

The annual Moskva Metro passenger traffic density statistic for 2005 was down a bit from the "all-time" high, which was 119 million pass-km per km of system length, at 1984 (on a significantly smaller network and with significantly fewer private autos in the city). At 2005, the busiest line carried 135 million pass-km per km of system length. In other words, 135 million people, on average, traveled, over each km of line during the year. The world traffic density record holder, the Western Railway corridor in Mumbai, does not carry much more: about 145 million pass-km per km of system length.

Traffic densities at this level do not occur in New York - and never have. NYCT carried 36 million passenger-km per km of system length at 2005. The IRT subway network - which was quite small compared to today's system - carried 52 million pass-km per km of system length at 1914. The combined IRT subway and elevated systems carried 43 million million pass-km per km of system length at 1929. The good citizens of Moskva put up with even higher levels of crowding than New Yorkers do.

We assert that, if crude "population density" were truly the key determinant of transit ridership that Cox evidently wants people to believe, then the results above would be much different.

Gross population density statistics are significant when studying urban areas and urban transportation systems - but are less significant that one might infer from various publications by Wendell Cox.


For Further Reading:

Extraterritorial Jurisdiction Issues and Request for Consent to Creation of a Municipal Utility District. 2007. Dallas: Economic Development Committee, Dallas City Council.

Houston, Scott. 2008. Municipal Annexation in Texas: Cheat Sheet for City Officials. Austin: Texas Municipal League.

__________. 2004. Municipal Annexation in Texas. Austin: Texas Municipal League.