Renowned urban thinker Anthony Downs wrote: “No jurisdiction is an island. Every suburb is linked to its central city and to other suburbs.” But intra-regional social and economic dynamics can sometimes make it appear as though there are actual oceans separating jurisdictional boundaries. The intra-regional social dynamic of homelessness is no exception.
Are there actually homelessness disparities within a region? If so, how large? I examine these questions in this article using the specific case of the national capital region.
But first, some background.
In The State of Homelessness in America 2012 (SOH12), we included an appendix with 2011 homelessness data for the 100 largest Metropolitan Statistical Areas (MSAs), as measured at a point-in-time. This includes data on nearly all of the metro areas in the country with populations over 500,000 people. Homeless point-in-time counts are reported to the U.S. Department of Housing and Urban Development at the geographic level known as the Continuum of Care (CoC), which is a local planning network designed to facilitate and encourage coordination of local efforts to address housing and homeless assistance. These CoC boundary lines are organized based on numerous local decisions of which the primary consideration should be to design a system that will most effectively meet the needs of the homeless population.
CoC boundaries may or may not reflect other demographic patterns or economic realities that shape how people interact in the physical environment. MSA boundaries, on the other hand, are determined by commuting to work/employment patterns and, therefore, are more likely to reflect the full human ecosystem. But the truth about MSA boundaries is that they are not representative of an incorporated jurisdiction, like a city or a town. Instead, MSAs are simply a statistical measurement instrument used by the U.S. Census and researchers alike to more effectively comprehend regions.
I’m going a long way to describe CoC and MSA boundaries here for two reasons. The first reason is so that I can say that coming up with estimates for metro areas required some spatial analysis work of matching CoC with MSA boundaries since point-in-time counts data are only reported at the geography of the CoC.
But the other reason I expounded on and on about the geography of the data was to show the value of work that derives estimates for metro area homelessness. And the value is that with such data we can have a more nuanced picture of the regional shape of homelessness in any particular metro area, especially when we take into account other factors, such as the general population. By taking population into account, we can look at rates of homelessness and make comparisons. More specifically, we can identify geographic disproportionality by comparing the rate of a single CoC to the rate of the whole MSA, or the rate between CoCs within an MSA. 
Data on the Washington metro area homeless population show that an estimated 13,205 people were homeless at a point-in-time in 2011, which ranks the area as the 8th highest total homeless population in the country. The area’s homelessness rate is 24 homeless people per 10,000 in the general population (~5.5 million people). Though ranking in nationally at number 8 in overall homeless population, D.C.’s homeless rate ranks as the 21st highest in the country. D.C.’s rate is lower than many major metropolitan areas, including Boston (ranked 20th), New York (13th), San Francisco/Oakland (12th), Los Angeles (6th), New Orleans (2nd), and Tampa/St. Petersburg (1st).
The more interesting data on Washington metro area homelessness, I believe, are found when you look at the geographic distribution of the population in the region (see the map above and table below). Nearly half of the metro area’s homeless population lives in the District of Columbia. Fairfax-Falls Church has 12 percent of the population, followed by Montgomery County (9 percent) and Charles, Calvert and St. Mary’s Counties (9 percent). Prince George’s County is the only other jurisdiction with at least 5 percent of the metro area’s total population. The remaining 15 percent of the population lives in the 6 other CoCs.
But when you take the general population data into account and look at rates of homelessness within the metro area, you can see the disproportionality among the jurisdictions. Four of the eleven CoCs in the Washington metro area have rates of homelessness that are higher than the national rate of 21 per 10,000. These four CoCs are: Arlington (22 per 10,000); Alexandria (30); Charles, Calvert, and St. Mary’s Counties (34); and the District of Columbia (109), which has a rate over 4 times that of the region as a whole and more than 5 times that of the nation as a whole.
The jurisdictions with the three lowest rates of homelessness in the region are: Prince George’s County (9 per 10,000); Fredericksburg/Spotsylvania and Stafford Counties (7); and Loudon County (5).
There are certainly a number of regional and CoC-level dynamics that account for the variation in the rates of homelessness in the Washington metro area—such as differences in housing affordability, jobs, etc.—but one thing that’s clear is that there is significant variation in the available data.
 One thing to note about the analysis of all the jurisdictional variation is that when making comparisons across CoCs, caution should be used as jurisdictions’ methodologies for estimating their population counts do vary. But, still, the rates of the point-in-time counts of people experiencing homelessness do provide us with a method for making reasonable comparisons.