Powering Smarter Market Decisions™     Toll Free 877.944.4AGS

Why I Really, Really Dislike The Census Metropolitan Areas

Posted by on Nov 1, 2015 in Blog | No Comments
Why I Really, Really Dislike The Census Metropolitan Areas

On site-based demographic reports, it is often helpful to have a benchmark column so that one can readily assess how a particular location is changing over time relative to a larger area.   A benchmark area should be both immediately recognizable and geographically relevant.   The notion of a metropolitan area seems like a reasonable choice.  For example, we could compare a number of sites in different suburban cities of Nashville, TN, against the concept of the metropolitan area as a whole.

The question has always been how to define a metropolitan area?  Cities really do change in extent over time so we require an ability to change the definitions over time, but we would also like to have a consistent and well known base geography from which to construct them.

For several decades now, the Census Bureau has been using counties (or in New England county subdivisions or towns) as the ‘building blocks’ for its metropolitan areas.   The rules for inclusion are relatively straight forward, using cross-county labor force movements to determine which counties become attached to the metropolitan area in question.   Even with changes over time to the definitions, it is relatively easy to reconstruct the history of areas by rebuilding with the current county definitions.   Overall, these have been very stable geographies over time, and thus very often used as benchmarks.

In most of the states east of the Mississippi, counties are relatively small and generally have a compact shape.   To the west, however, the counties become very large, in some cases the size of small northeastern states.metropolitan areas

The first map above shows the coverage of the metropolitan and micropolitan areas (the latter in light red).   From all appearances, California is pretty much completely urban and certainly one unfamiliar with the state would surely think that it was overwhelmingly urban.  And while the Los Angeles area goes on seemingly forever (especially when stuck on the dreaded 405 freeway), the metropolitan area includes substantial areas which are effectively unpopulated and several areas which are, despite being in the included counties, almost unrelated to the metropolitan area itself.

A classic case of metropolitan area strangeness is Santa Maria-Santa Barbara, CA, which is simply the Santa Barbara county.  Rather than being one central city and suburbs, the county really includes three very different and geographically separated areas of urbanization — coastal (Goleta, Santa Barbara, and Carpinteria), the Santa Ynez valley (Lompoc, Buellton, Solvang, Santa Ynez), and the Santa Maria valley (Santa Maria, Guadalupe).

As a benchmark area, nothing could be more misleading than comparing a site in Santa Maria to the metropolitan area as a whole.   Santa Maria and Santa Barbara are roughly the same size cities, the former being an agricultural service center and the latter being the tourist center of the ‘American Riviera’.    Comparisons are effectively meaningless.

The standard metropolitan area maps give one a sense that cities are the dominant feature of the American landscape.   A look at the extent of the Census Bureau’s “Urbanized Areas” tells a completely different picture of where we live.   This program has a set of simple definitions of what constitutes an urbanized area including minimum population size and population density.   They very much mirror the concept of a ‘built-up’ area for each center.

Urbanized Areas

Urbanized Areas

This is a vastly different map than the county-based urbanization map first shown.   The vast urban areas of the west simply vaporize.  It is here where we realize just how vast and generally open this country really is.

The issue of what constitutes a good benchmark for a metropolitan area remains.   I have personally tried a number of different approaches, none of which proved to be worth preserving.   Until then, the question of what constitutes a good benchmark is probably best determined by asking ‘what do you want to know?’ — not satisfying, but probably the best answer for now.