Public Opinion Watch

Ruy Teixeira
Ruy Teixeira

(covering polls and related articles from the week of January 3–9, 2005)

In this edition of Public Opinion Watch:

• More on Exurbia
• Is It All about White People with Kids?
• Mr. Popularity
• Show Them the Money: How the Democrats Can Take Back the House

More on Exurbia

I have argued that the term exurban has been used in a scandalously confused way, categorizing so much of suburbia as exurban that the term loses any meaning. In particular, I took a close look at the definition of exurbia used by the National Committee for an Effective Congress (NCEC), a leading Democratic targeting firm. Here is some of what I reported:

“[NCEC’s original criteria] indicate that pretty much any suburban county that does not contain a large city can be designated as exurban, if it is relatively downscale in terms of occupation, income, and education or if it falls below a certain density criterion. [Note: these original criteria also appear to categorize any non-rural county with a minority population over 50 percent as an urban county.] . . .

1. Some entire MSAs (metropolitan statistical areas) are designated exurban, like the Canton MSA in Ohio and the Pensacola and Sarasota MSAs in Florida.

2. In other MSAs, only the county containing the MSA’s main city is designated “urban-surburban” while every other county is designated exurban or even rural. For example, in the Columbus, Ohio, MSA, only Franklin county is termed urban-suburban, while five other counties are designated exurban and two are considered rural. Similarly, all Ohio counties in the Cincinnati MSA are designated exurban except Hamilton county.

3. Medium-sized metro areas wind up being classified almost entirely as exurban or rural. In Florida, for example, there are 16 counties in medium-sized metro areas. Of these, just three are classified as either urban-suburban (1) or suburban (2), while 13 are classified as either exurban (11) or rural (2).

4. Almost no counties are simply designated “suburban.” In Ohio, there are only three (compared to 30 exurban counties); in Florida, just five (compared to 21 exurban counties). . . .

Collapsing all but the most urbanized parts of big metro areas, almost the entirety of medium-sized metro areas and outer suburbs everywhere into exurbia does considerable violence to the concept and clarifies little.”

Arguably, what the NCEC criteria are really doing is categorizing all suburban counties where downscale white voters predominate as “exurban.” Not surprisingly, given recent voting trends among white, working-class voters, these suburban counties tend to vote Republican. But calling these counties exurban simply confuses the issue: why not call them Republican-leaning suburban counties (RLSCs) instead? That would be clearer and more analytically justified (though less trendy).

I have now obtained a categorization of every county in the United States on the basis of the NCEC’s original criteria and have conducted some analysis using their categories. (Note: they appear to have modified their criteria slightly since their original criteria were elaborated, but I do not have access to these modified criteria.) This analysis produces some interesting results which further underscore, I think, the need for much more careful and selective use of the term exurban.

1. By NCEC’s definition, 581 counties in the United States are exurban and just 131 are suburban.

2. By NCEC’s definition, 29 percent of the U.S. population live in exurbia and just 19 percent in suburbia (!). (If you’ve got a geographer friend, tell that one to him/her to get a good laugh.)

3. NCEC’s exurban counties provided 31 percent of the vote in 2004, two points over their population share of 29 percent. Note that these counties provided 30 percent of the vote in both 2000 and 1996, so the exurban share of the vote, even under NCEC’s peculiar definition, is increasing very slowly, not rapidly.

4. Once you adjust the increase in votes in these counties for population increase (see my earlier analysis on this subject), their adjusted rise in turnout in 2004 was actually less than in rural, suburban, and urban counties, as defined by NCEc=

5. Republican domination of these counties is, as I argued previously, nothing new. Even under NCEC’s definition, Reagan carried exurban counties by twenty-seven points in 1984, compared to just fifteen points for Bush in 2004. In fact, Bush’s papa in 1988 actually did better than his son in these counties, carrying them by seventeen points. The only area where Bush bested his papa was in rural counties (by NCEC’s definition), carrying them by nineteen points, compared to eleven points for his father (though Reagan carried them by twenty-four points). And Bush did way worse than his father in NCEC’s suburban counties, losing them by five points, while his father carried them by eleven in 1988 (and Reagan carried them by twenty-one in 1984); he also did much worse in NCEC’s urban counties, losing them by nineteen points, while his father lost them by only five points in 1988 (and Reagan actually carried them by four points in 1984). All this underscores the “Reagan lite” nature of Bush’s coalition.

Bottom line: we’re still looking for a definition of exurbia that clarifies more than confuses and adds real analytical value. I am in touch with some geographers who are trying to come up with a clear, tight definition rooted in standard practices in their field. I’ll report back when their efforts have (hopefully) borne some fruit.

Is It All about White People with Kids?

An interesting variant on the exurban argument (critiqued above) is the white families with kids argument. This argument contends that exurbs are busting at the seams with white married households with children who have moved there for more land, bigger houses and a safer, more traditional and (let’s face it) less racially diverse environment for their kids. Since these kind of voters naturally tend to favor the Republicans, and since the exurbs are fast-growing, this must give the Republicans a big edge over the demographically stagnant Democrats. This argument popped up recently on the New Republic website in an article “Parent Trap” by Joel Kotkin and William Frey. The same basic argument is developed in more detail by political analyst Steve Sailer in his article “Baby Gap” in the American Conservative.

I have already had much to say about the problems with the fast exurban growth part of this argument. But the white families with children argument has another, deeper problem: white married households with children are not only declining relatively, as a percentage of households, they are also declining absolutely—that is, the number of these households is actually falling over time. Between 1990 and 2000, for example, the number of white married households with children declined by almost 7 percent. This is true even in the NCEC-designated “exurban” counties (or “Republican-leaning suburban counties” [RLSCs] as I prefer to call them} I discussed on Monday: white families with kids declined by 1 percent in RLSCs between 1990 and 2000. And in NCEC-designated rural counties, they decreased by 9 percent.

That makes the whole white families with children argument sound pretty weak. Aren’t there any areas where these households are at least growing in absolute terms? Sure there are, but to find them you have to adopt a fairly strict definition of exurbs, like the one I’ve used in the past (fringe counties of large metro areas). If you do that, it turns out that these exurban counties had 11 percent growth in white married households with children in the 1990s. But these exurban counties are also just 4 percent of the population and contain only 6 percent of the nation’s white families with children. In other words, the only category of counties that remotely fits the white families with children argument is too small to have the big political impact the argument alleges.

It’s also interesting to note how slowly the distribution of these households is changing. Using the NCEC categories, in 1990, 34 percent of white families with children were in “exurban” (or RLSC) counties compared to 35 percent in 2000. In rural counties there was no change (24 percent in both years). In NCEC’s “suburban” counties (many of which are not really suburban and are selected so that they tend to lean Democratic), there was a slight increase, from 20 to 21 percent of white families with children over the decade. And in NCEC-designated urban counties, which are typically in only the largest urban areas with the heaviest minority populations, there was only a slight decline over the decade, from 21 to 20 percent of these housholds.

And even using my strict–and more accurate–definition of exurbia, the proportion of white families with children in this category of counties only rose from 5 percent in 1990 to the 6 percent mentioned above.

Thus, not only are the absolute numbers of these families declining almost everywhere, but the distribution of these families across different types of counties is actually changing very slowly–in fact hardly changing at all.

In short, the attempt to construct a dynamic, demographic argument around white people with kids just doesn’t hold water. Of course, the basic observation that white people with kids do tend to vote Republican remains true, but the attempt to gussy up this fact with “parent traps” and “baby gaps” should be taken with an entire cellarful of salt.


Mr. Popularity

Ipsos poll of 1,001 adults for AP, released January 7, 2005 (conducted January 3–5, 2005)

Just how popular is Bush these days? Check out this excerpt from a recent Associated Press (AP) story, discussing the most recent Ipsos-AP poll:

“Bush’s approval rating is at 49 percent in the AP poll, with 49 percent disapproving. His job approval is in the high 40s in several other recent polls—as low as any job approval rating for a re-elected president at the start of the second term in more than 50 years.

Presidents Reagan and Clinton had job approval ratings of around 60 percent just before their inauguration for a second term, according to Gallup polls.

President Nixon’s approval was in the 60s right after his 1972 re-election, slid to about 50 percent right before his inauguration and then moved back over 60 percent. President Eisenhower’s job approval was in the low 70s just before his second inauguration in 1957.”

That 49 percent approval rating in the AP poll is no fluke; it is exactly Bush’s average approval rating in the last month across nine polls, as shown by Chris Bowers over at MyDD.

It’s worth highlighting the rest of Bush’s underwhelming approval ratings as well. Even his rating in his best area, “foreign policy issues and the war on terrorism,” is a mere 50 percent approval/48 percent disapproval. His rating on the economy is 47 percent/51 percent; his rating on Iraq is 44 percent/54 percent; and his rating on “domestic issues like health care, education, the environment and energy” is 43 percent/56 percent.

Rather than celebrating his impending inauguration, many Americans seem to be developing a case of buyer’s remorse.

Show Them the Money: How the Democrats Can Take Back the House

The Democrats can’t take back the House because of redistricting, right? In fact, redistricting has reduced the number of competitive seats so drastically that the Democrats must wait until after the next round of redistricting (that is, 2012) to even have a chance of taking back the House (and that’s only if this future redistricting is done far differently than the previous round of redistricting).

But there’s a huge problem with this argument, accepted as gospel by so many political observers: the idea that redistricting has destroyed competition for House seats is dead wrong. That’s what Alan Abramowitz, Brad Alexander, and Matthew Gunning of Emory University’s Political Science Department show in their new paper, “Incumbency, Redistricting, and the Decline of Competition in U.S. House Elections,” presented at the recent meeting of the Southern Political Science Association. They present evidence that suggests very strongly that redistricting has almost nothing to do with the decline of competitiveness in House elections. And if that’s true, then redistricting can’t be the root of Democrats’ difficulties in taking back the House and the whole argument about having to wait until 2012 falls apart.

But if redistricting isn’t the cause of decreased competitiveness in House elections, what is? And what, if anything, can Democrats about it?

Abramowitz, Alexander, and Gunning provide a clear answer to the first question:

“The evidence presented in this paper indicates that declining competition in U.S. House elections is explained by two major factors: a shift in the partisan composition of House districts and a decline in the ability of challengers to compete financially with incumbents. Since the 1970s, and especially since 1992, there has been a substantial increase in partisan polarization among House districts. The number of marginal districts has been declining while the number of districts that are safe for one party has been increasing. Redistricting appears to have little or nothing to do with this trend: almost all of the change in district partisanship has occurred between redistricting cycles. . . .

The effects of increasing partisan polarization have been reinforced by the second trend uncovered by our study—the decreasing financial competitiveness of House challengers. Not only are there fewer incumbents in high-risk districts, but even in these districts, incumbents running for reelection are less likely to face financially competitive challengers. Fewer and fewer challengers are able to raise the amount of money that is now required to wage a competitive campaign against a well-funded incumbent. As a result, competition is now confined to open seats and a handful of races involving exceptionally vulnerable incumbents and/or exceptionally well-financed challengers.”

That couldn’t be clearer. It’s not redistricting, it’s partisan polarization and (lack of) money that are at the root of Democrats’ difficulties taking back the House. Since Democrats can’t do much about the first problem (at least, in a purposive way), I suggest they concentrate on the second. Find high-quality challengers and show them the money. They need not wait for 2012.

But where exactly should the Democrats pursue such a strategy? Abramowitz provides some useful detail on where Democrats might want to concentrate their resources:

“In 2006 Democrats would be wise to target Republicans representing high-risk districts: districts that lean Democratic in presidential elections. Such districts account for a disproportionate share of incumbent defeats and party turnover in House elections. For example, in 1994, 32 percent of Democratic incumbents in high-risk districts were defeated compared with only 7 percent of Democratic incumbents in all other districts. Although only 34 percent of all Democratic seats in 1994 were in high-risk districts, 70 percent of Democratic seat losses occurred in these districts.

So where are these high-risk Republican districts? There are currently 25 such GOP districts: Colorado 7; Connecticut 2, 4, and 5; Delaware AL; Florida 10 and 22; Illinois 10; Iowa 1 and 2; Kentucky 3; Nevada 3; New Hampshire 2; New Jersey 2, 3, and 4; New Mexico 1; New York 3, 13, and 25; Pennsylvania 6, 7, 8, and 15; and Washington 8.

There are certainly vulnerable Republicans in other districts, but the GOP representatives in these high-risk districts deserve special attention. In order to maximize their gains in the 2006 midterm election, Democrats need to recruit strong challengers in these high-risk GOP districts and make sure that these challengers have the funds needed to wage competitive campaigns. That will take a lot of money, but Democrats showed in 2004 that they can compete financially with Republicans. We only need to gain 15 seats to regain control of the House. With a major effort and a little help from a Bush Administration that seems determined to cut social security benefits for future retirees, it should be possible.”

Ruy Teixeira is a joint fellow at the Center for American Progress and The Century Foundation.