Where did the Democrats lose in 2002? A lot rides on this question; wrong answers will produce poor targeting and ineffective politics, and the Democrats can afford precious little of either. But right answers can set the stage for future gains in the 2004 election and beyond.
Perhaps the most common answer -- certainly among Democrats -- has been base mobilization. In this view, the Democrats' 2002 campaign failed to excite their political base, which consequently turned out at low levels and didn't offer the consistently high support the party has enjoyed in previous elections.
Let's take the issue of support levels first. We are handicapped here and elsewhere by the absence of exit polls in this election, with the exception of a Los Angeles Times exit poll in California. However, a combination of telephone surveys taken right before, during and after the election, plus analyses of county and precinct voting results, does allow for some provisional assessments.
These data suggest that the most resolute element of the Democrats' base -- the minority vote -- did not skimp in its level of support for Democrats in the 2002 election. This is particularly true of blacks, the most pro-Democratic minority. In the 2000 election, Al Gore got 90 percent of the black vote; in 2002, Democratic senatorial and gubernatorial candidates received a similar -- if not slightly higher -- percentage of the black vote.
Similarly, Hispanics appear to have voted 2-to-1 -- their usual ratio -- or a bit better for Democratic senatorial and gubernatorial candidates. The one clear exception to this rule was in Florida, where Gov. Jeb Bush received a majority -- 56 percent -- of the Hispanic vote, according to one poll. But then George W. Bush received 49 percent of the Hispanic vote in the 2000 presidential election, reflecting the influence of Florida's conservative Cuban Hispanics. Another relatively bright spot for Republicans was New York, where extremely liberal Republican Gov. George Pataki may have received as much as 38 percent of the Hispanic vote, according to an analysis of precinct returns by political scientist John Mollenkopf. But overall, there were few signs that Hispanics' strong support for Democrats was slipping or that such slippage played much of a role in this year's election results.
Finally, data on Asians are hard to come by in this election, but the California exit poll, at least, indicates that Asians voted for Gov. Gray Davis over Bill Simon 54 percent to 37 percent, similar to their preference for Gore over Bush in 2000. So, one can fairly conclude, overall minority support for the Democrats looked healthy in this election.
Even if minority voters supported Democrats at high rates, however, their low turnout could have hurt the Democrats in a number of key races. Close examination of polling and county-level returns does turn up some evidence of a turnout problem for Democrats. For example, while polls of voters or likely voters are imprecise instruments for assessing turnout trends, they do suggest that the share of minority voters -- the percentage of the overall electorate that they constituted -- in many state electorates was down compared with 2000. Even taking these data at face value, a relatively low share of minority voters didn't matter much in a state such as California, where the Democrats did well anyway, or in a state such as Florida, where the Democrats were so far behind that any reasonable increment of minority voters could not conceivably have saved the party's candidate. But in close races such as Missouri's, it could have been a factor.
More broadly, county-level voting returns suggest that in Democratic-leaning large cities and inner suburbs, the turnout -- even where it did not go down -- did not keep pace with Republican-leaning outer suburbs and rural areas. Indeed, Republican areas seemed highly mobilized this year. For example: In Missouri, the increase in votes cast this year from 1998 was less in strongly Democratic St. Louis city or Democratic-leaning St. Louis County than in the strongly Republican suburb of St. Charles County or, especially, in rural and fervently Republican Cape Girardeau County. The same pattern can be seen in Minnesota, where many of the more rural counties cast almost as many votes in 2002 as in 2000, while the more urban counties lagged behind.
The Rise of the Exurbs
Republicans are delighted to take credit for their voter-mobilization efforts in the 2002 elections. But some analysts want to go further and suggest that Republican gains in 2002 reflect the rise of the "exurbs" -- fast-growing, mostly Republican-leaning counties on the fringes of big metropolitan areas. David Brooks made this case in a Nov. 10 New York Times article.
While Brooks is right that exurbs made important contributions to Republican victories in 2002, his assertion that they were central to these victories is on much shakier ground. Let's look at his two main examples, Colorado and Maryland. It is true that Colorado's quintessential exurb, Douglas County, with its stunning 191 percent population growth rate during the 1990s, voted overwhelmingly for Republican Wayne Allard over Democrat Tom Strickland, 66 percent to 32 percent; that was about the same margin Douglas County gave Bush over Gore in 2000. But it's also true that the Denver-Boulder area, which includes Douglas County, as a whole gave Strickland a 6-point edge, which is more than the 3-point victory Gore received in 2000. Strickland's problem was not so much conservative exurbs turning the vote in metro Denver-Boulder than it was his poor showing in the rest of the state.
The example of Maryland fits even worse, and that's leaving aside the Democrats' pickup of two House seats in the 2002 election and Gore's 17-point victory over Bush in 2000. It's true that Republican gubernatorial candidate Robert Ehrlich did very well in such genuinely exurban counties as Frederick, northwest of Washington, and Harford, north of Baltimore -- both of which Brooks specifically mentions and both of which tend to vote Republican to begin with. But Ehrlich's real coup was carrying counties Brooks doesn't mention -- such as close-in Baltimore County, the third-largest in the state, and Howard, the fastest-growing county in Maryland that has a population greater than 100,000. Both counties tend to vote Democratic (Gore easily carried them in 2000) and have become more so over time. In short, Democrat Kathleen Kennedy Townsend, the incumbent lieutenant governor, was a lousy candidate who lost many counties she should have won and lost counties badly where she should have at least come close. Ehrlich's victory hardly suggests an impending era of Republican exurban dominance in the oh-so-blue state of Maryland.
Indeed, exurban counties are frequently part of much larger metropolitan areas that are trending Democratic. And as these exurban counties get bigger, denser and more diverse, they generally become less -- not more -- Republican. In short, Republicans should be glad to have many of these exurban counties in their camp for the moment, but if they're counting on long-run political dominance from these same counties, they may be sorely disappointed.
The Democratic-Leaning Suburbs
Both of these views of Democratic defeat in 2002 -- the dwindling Democratic turnout and support and the growing Republican base in the exurbs -- leave out a very important factor: the Democrats' inability to carry Democratic-leaning suburban or mixed suburban-urban counties by the margins previously enjoyed. These counties, generally close to cities and frequently at the heart of the technologically advanced postindustrial areas John B. Judis and I have termed "ideopolises," have been steadily giving a larger and larger share of their vote to Democratic candidates. But in 2002, the Democrats generally did not do so well in these counties because Republicans picked off swing voters.
For example, in St. Louis County (which doesn't include St. Louis city) in Missouri, Jean Carnahan's margin over Jim Talent was only 3 points, down from the 8 points by which her late husband carried the county in 2000. That alone cost her 16,000 votes and -- combined with her reduced margins of victory in nearby Jefferson County, in the Jackson County suburbs outside Kansas City and in Boone County around the university town of Columbia -- was easily enough to cost her the election.
Similarly, in Hennepin and Ramsay counties in Minnesota around Minneapolis-St. Paul, Walter Mondale's margins over Jim Coleman were substantially less than Mark Dayton's in 2000 or Paul Wellstone's in 1996. If Mondale had simply run as well as either of these Democratic candidates in those two counties, he would have defeated Norm Coleman by about 10,000 votes.
That's two Democratic Senate seats down right there. It's clear that relatively poor performances in Democratic-leaning suburbs were a key reason -- perhaps the main reason -- the Democrats lost the Senate in 2002 and had such a dismal showing overall. This suggests that while the Democrats need to improve base mobilization, they cannot afford to neglect the task of reaching swing voters in the suburbs, particularly those in the close-in Democratic suburbs. Of course, there's no reason to stop there: Cutting into Republican strength in the exurbs is a fine idea as well, if not quite as significant as David Brooks argues. These tasks urgently suggest the need for a new Democratic approach. It is now clear that the party's 2002 program of working for a Medicare prescription-drug plan, defending Social Security and having little to say on other issues was a loser all around. It didn't excite the Democratic base, it didn't effectively reach swing voters in the Democratic-leaning suburbs and, lord knows, it didn't make much of an impression on swing voters in the exurbs (and they do have them out there!). None of this reflects a dearth of good Democratic ideas, but it does reflect a dearth of Democrats actually expressing good ideas. It's time for that to change.