« The "Values Voters" Debate Continues | Main | WSJ Poll: Bush 'Mandate' Limp At Best »

The Exurban Myth

Exurbs have been getting a lot of attention in analyses of the 2004 election. According to some observers, Republican domination of these areas was the key to Bush’s re-election victory and, because of the phenomenal level of mobilization in these fast-growing areas, Republicans should continue to out-point the demographically stagnant Democrats in the future.

A careful look at the data suggests to me that there is a great deal less than meets the eye to this thesis. I have already pointed out how exurban counties made only a modest contribution to Bush’s net gain in votes in 2004 and how Republican domination of exurban counties is nothing new and was, in fact, more pronounced under Reagan than it is now.

An interesting angle I haven’t covered yet is how mobilization in exurban counties stacked up to mobilization in other types of counties. In post-election analyses, mobilization and turnout in counties has generally been measured simply by comparing the vote in 2004 to the vote in 2000. The higher the percent increase in votes cast, the more mobilization has taken place, is the general assumption.

But this assumption is not warranted--it leaves out an important variable that affects the number of votes cast: population growth. The more population grows, the more votes should be cast, even if there is no change at all in the level of mobilization; more possible voters = more votes, all else equal. Therefore, if we are interested in the extent to which mobilization changed between the 2000 and 2004 elections, we need to measure the change in votes cast relative to the growth of the population between 2000 and 2004.

Once we measure mobilization in this way, exurbs do not appear to have been more mobilized than most other types of counties. Most of the 22 percent increase in votes cast in exurban counties, it turns out, is attributable not to extraordinary mobilization, but rather to population growth with ordinary mobilization.

In fact, my analysis shows that exurban couties actually increased their vote less relative to population growth (8 percent) than did the central counties of large metropolitan areas (9 percent), counties in medium-sized metro areas.(12 percent), counties in small metro areas (11 percent) or even most types of rural counties (9-11 percent), except for the most extremely rural, where votes cast grew by only 5 percent relative to population growth.

These findings are yet another reason to examine claims about the political potency of exurban counties with considerable skepticism. Mark Gersh, head of the NCEC and leading Democratic number-cruncher, heads in the other direction, however, in his interesting article, “Battlefield Erosion” in the latest issue of Blueprint, the DLC’s magazine.

In this article he ascribes substantially more importance to exurban counties that I have done. The article is based around analyses of three key states--Ohio, Florida and Pennsylvania–and he presents data that seem to implicate exurban counties heavily in Bush’s wins in Ohio and Florida and Kerry’s narrowed margin in Pennsylvania.

What explains the difference between his analysis and mine? Several things actually, but the most important one is this: he defines exurban counties much more broadly than I do.

This difference in definitions raises an interesting question: what exactly do we mean by an “exurban” county anyway? How do we know an ”exurb” when we see one? Surprisingly, despite the loose way the term is now thrown around, there is very little rigorous–or even semi-rigorous--discussion anywhere of criteria for defining an exurb.

But here are a couple of definitions I found on the web which fairly reflect the general view of the exurb:

The expression "Exurbs" was coined in the 1950s to describe the ring of prosperous rural communities beyond the suburbs that, due to availability via the new high-speed limited-access highways, were becoming dormitory communities for an urban area (Wikipedia)

Exurb: A region or district that lies outside a city and usually beyond its suburbs (Outline of American Geography)

Note that both definitions allude to exurbs being beyond the conventional suburbs–on the very fringes of metro areas and not suburbs in the conventional sense. That is the idea that led to the identification of such counties as Douglas county, CO, Scott county, MN, Loudon county, VA, Frederick county, MD, Pinal county, AZ, Forsyth county, GA and so on as exurban counties. These are the kinds of counties cited by David Brooks in his influential New York Times article on the exurban voter and that are included as exurban in my analysis, based on their status as fringe counties of large metro areas. I include 133 counties nationwide in my exurban category. They are also the kinds of counties included in a category of New Metropolis/Suburbs of Suburbs counties developed by geographer Robert Lang. In his paper on these counties, Lang enumerates 47 counties that fit this category.

In contrast, Gersh/NCEC designate 30 counties (!) in Ohio alone as exurban. How did they arrive at a definition of exurban so broad that it would generate so many exurban counties in a single state?

More on exurbia tomorrow!


Ohio has only 88 counties. If Gersh is counting 30 counties as being "exurban", what counties wind up being included in the urban and suburban category? Does any county in Ohio really count as "urban", since Ohio doesn't allow for separate cities in its municipal structure. Even Cuyahoga county must count as suburban since Cleveland is dominated by the ring of old suburban communities around it, which you couldn't distinguish from the city itself as you drive through the respective polities. Gersh's analysis must be just plain screwed up. Wouldn't that enable someone to arrive at whatever conclusion one was trying to push that day?

I don't think there's even such a thing as an "exurban county." Maybe an "exurban precinct." It's really just a convenient label for places having lots of wealthy people who vote for Republicans. During the Reagan years the south Denver "exurbs" included areas of Arapahoe and Jefferson Counties, but those same counties also included inner suburbs. As infill development has occurred, demographics and voting patterns have moved toward a more suburban and urban profile in those counties. The same is happening in Douglas County - as the population grows, it starts to look more like suburbia and voting patterns will moderate. Most of the land area of Douglas County is rural and always will be, but the rural population will be a smaller fraction of the county total.

I happen to have lived in one of Gersh's "exurb counties" in Ohio and can state that Lorain County is no exurb. Lorain County consists of a belt of Cleveland suburbs in the east, two mid-sized industrial cities of around 50,000 population (Lorain and Elyria), and rural areas to the south and west. There may be Clevelanders moving to small towns there, but they are by no means representative of the rest of the county.

My understanding of an exurb is Dutchess or Columbia Counties in New York's Hudson Valley. Both were predominantly farm country until recently and remain mostly rural in character. There are no major cities of any size and no suburbs at all. Their growth is fueled by city and suburbanites from New York and, to some extent, Albany seeking a country lifestyle.

Lorain County was solid Democratic for years (and was once again in 2004), the heavy union vote in Lorain and Elyria dominating the rural voters. Dutchess and Columbia are rock-ribbed Republican (FDR, born and raised in Dutchess County, lost it every time while winning the Presidency) and have, if anything, moved a bit towards the Democrats lately, though some of this is a local reaction against one-party rule. My conclusion is that the exurb process is simply the shift of some of the Republican vote from an urban area to a rural one, without an increase. In some places, this might produce an electoral vote shift (e.g., voters moving from Chicago suburbs to Indiana and Wisconsin rural areas, but the idea that such voters are new Republican strength has no basis in reality that I can see.

Actually Frederick is in a historically red area of the state (MD) that is becoming more blue as folks get squeezed out of Montgomery County by the high housing prices and move to Frederick and beyond to get more house for their money. It'll be interesting to see if that area will turn blue in 2006 for the Governor's race.

How's this for a Dem proposal: state initiatives to eliminate counties. We don't need them. In fact, they help Repubs. From my experience in Lane County, Oregon, the city I live in, Eugene, and the State I live in Oregon are Bluer than blue. My county, on the otherhand is RED because of exurban Repubs. Those functions that counties perform like roads and elections would become state functions and cities would have more influence. So all those people living in their blissful exurbia would be required to get involved when the power shifted, even slightly, to cities and inner ring suburbs.