In Search of the Elusive Swing Voter

It almost doesn't matter who the Democratic candidate is. In terms of strategy, the road map for the coming presidential campaign was set long before the primaries—and it runs straight through the handful of states with the largest numbers of independent voters. Any candidate needs to hunt them down

During the primary season we are accustomed to focusing on the drama and intrigue in a few important states such as New Hampshire and Iowa. A candidate who can win these, it is believed, is all but certain to carry the nomination. Suddenly that's less true than in years past. Because so many candidates are running, and because the contests are not winner-take-all, formerly inconsequential states such as Oklahoma, New Mexico, and even Delaware could be critical in determining who finally emerges as the next Democratic nominee. In contrast, we tend to view the general election as wide open, when in fact fewer and fewer states have determined the outcome in recent presidential elections. The reason for this is the growing polarization of the American electorate.

In fact, there is empirical evidence to suggest that this year's contest may be the most partisan in history. In the 2000 election less than 10 percent of Democrats voted for George W. Bush, and a similarly small percentage of Republicans voted for Al Gore—the lowest voter crossover ever documented. The Supreme Court's decision in Bush v. Gore, the Bush Administration's hard-nosed tactics, and the war in Iraq have only widened this divide.

As the American electorate becomes ever more polarized, the number of undecided voters and the number of states in which the two parties will truly compete have diminished considerably. Two decades ago as much as a third of the electorate was deemed to be in play, and there were grand debates, particularly in the Democratic Party, about whether the best way to win "swing" voters was to pursue a southern strategy or to target the Rocky Mountains. "A basic postulate of American politics today," says the political demographer Mark Gersh, a Democratic strategist, "is that the swing vote is much, much smaller than it used to be." Strategists in both parties have narrowed their focus to no more than 10 percent of the electorate (some have narrowed it even further), and both parties plan to seriously contest only about fifteen states in November. This shrunken playing field, along with hardening lines in the electorate, all but guarantees a close race. That in turn limits the strategic possibilities for both parties to the point where it is possible to predict in considerable detail what the next campaign will look like—even without knowing the identity of the Democratic nominee.

There is a widespread misperception that the course of a presidential campaign flows directly from the candidate's persona. Naturally, a Howard Dean campaign would differ in style and atmospherics from one featuring Wesley Clark or John Kerry or Richard Gephardt. But with so little room to maneuver, the Democratic formula for victory will depend less than ever on the identity of the nominee. Instead it will be dictated by geographic and demographic necessity—how best to cobble together the necessary 270 electoral votes. The candidate must carry a sufficient number of swing states, and success in each one will depend on highly specific combinations of constituencies and issues—many of which can already be identified. In other words, just as the genetic blueprint for human beings and chimpanzees is 95 percent identical, the campaign blueprint for the Democratic candidates will be nearly the same, regardless of which becomes the party's nominee.

The unprecedented closeness of the 2000 presidential election has had dramatic effects on the political world, from the news media's hesitancy to call election results to the parties' renewed emphasis on voter turnout to the newfound superstition among speechwriters that they must prepare three versions of a candidate's election-night remarks: the dignified victory speech, the gracious concession, and a third in case the election is too close to call. To political demographers, who digest ungodly amounts of data in an effort to understand and predict the behavior of the American electorate, the 2000 contest provided a sort of Rosetta stone: a demographic snapshot of a nation in perfect balance, which has become the starting point for strategy in 2004.

All told, twelve states in the previous presidential election were decided by fewer than five percentage points. Along with two or three other states where demographic changes portend a similar closeness, they make up the battleground this year. The most significant states are scattered across the Pacific Northwest (Oregon, Washington), the Southwest (Arizona, Nevada, New Mexico), and the Rust Belt (Ohio, Pennsylvania, West Virginia), with outliers on the East Coast (Florida and New Hampshire) and others along a lengthy stretch of the Mississippi River, from Minnesota and Wisconsin down to Arkansas and Missouri. The next Democratic campaign will closely follow this map.

The Northwest

In numerical terms the most striking aspect of the 2000 election remains the number of votes Al Gore lost to Ralph Nader. "Democrats created an opening for Nader in 2000 by not taking the Green Party seriously enough until it was too late," says Doug Sosnik, a White House political director under Bill Clinton. To head off a similar catastrophe the Democratic nominee will probably begin his campaign with an early pilgrimage to the Pacific Northwest, where Green Party support is strongest, to quell a potential challenge. Such a move would not only strengthen the candidate's standing in Oregon and Washington, two states Gore won narrowly, but also provide a platform for talking about the environment—one of the few "wedge issues" available to Democrats, and an issue pollsters believe is the primary motivator for six percent of voters.

The South

The other great political truth revealed in 2000—and reinforced in 2002—is the Republicans' consolidation of the South. The long-standing axiom that the Democrats must carry southern states to win the presidency still holds sway among many political consultants, and at least partially accounts for the premium placed on southern candidates such as Wesley Clark and John Edwards. But Democratic strategists are increasingly aware that that goal has become nearly unattainable. With the exception of Florida, the South has trended away from the party. Bill Clinton's success in the 1990s was not indicative of a southern Democratic resurgence—rather, it masked this erosion. Georgia, which Clinton carried in 1992, went Republican in 1996. Arkansas, Kentucky, Louisiana, and Tennessee, which Clinton carried in both his elections, all followed suit in 2000. Thomas F. Schaller, a professor of political science at the University of Maryland, Baltimore County, warns, "Pursuing a southern strategy in 2004, instead of looking ahead to other areas, could relegate the party to minority status for years to come." Indeed, the futility of a southern strategy is tacitly acknowledged in the list of swing states that Democratic groups are planning to contest this year. Of the seventeen states targeted by America Coming Together, a coalition of liberal interest groups aimed at mobilizing Democratic voters, only Florida and Arkansas are in the South.

The Southwest

The Democrats' new area of opportunity is a swath of formerly Republican territory where an influx of Latinos and transplanted white Democrats is changing the demographic profile. Gore's lone win in this region was New Mexico, where his margin of victory was even narrower than his margin of defeat in Florida. But newly elected Democratic governors in Arizona and New Mexico and booming Hispanic populations there (25 percent and 42 percent, respectively) should persuade this year's nominee to spend considerable time and effort in the region. Nevada, too, has become a case study for Democratic optimism: although Bush carried the state in 2000, the Latino population surged by 15 percent in just the next two years; Clark County, which leans Democratic, is among the nation's fastest-growing counties; and Las Vegas, in that county, is rapidly unionizing. Furthermore, Nevada presents an enticing opportunity to raise the issue of Yucca Mountain, where President Bush recently decided to dump nuclear waste after vowing during his campaign not to do so.


Latinos, who have historically been identified with the Democratic Party, now represent an important swing vote. As it became clear in the closing weeks of the 2000 campaign that several battleground states would go down to the wire, the Republicans spent an unprecedented amount of money on Spanish-language television advertising; overall, they ended up spending more than twice as much as the Democrats. According to Adam J. Segal, the director of the Hispanic Voter Project, at Johns Hopkins University, the Bush campaign poured money into Florida media markets in particular, stoking Hispanic anger over the Clinton Administration's handling of the Elián González affair—and ultimately helping to deprive Gore of the state and the election. (In contrast, Gore outspent Bush nearly three to one in New Mexico, and won.) And with some strategists believing that the 2004 election, too, could hinge on Florida—and that Florida could hinge on the heavily Hispanic I-4 corridor between Tampa and Orlando—the Democrats are sure to avoid making the same mistake again.

The Upper Midwest

The culturally cautious Rust Belt states that were a key to Bush's win have been particularly hard hit by the net loss of three million jobs since Bush took office, 2.4 million of them in the manufacturing sector. As the Democratic contenders delight in pointing out, Bush stands to become the first President since Herbert Hoover to see the country lose more jobs than it gained on his watch. Even if the economy improves, a critical component of the Democrats' regional rhetoric will be reminding voters exactly how many manufacturing jobs have been lost in states such as Michigan (127,000), Pennsylvania (132,500), and Ohio (151,800).

The Mississippi Basin

Finally, for all the ribbing it drew, Gore's four-day riverboat tour along the Mississippi after the Democratic convention is likely to be repeated in some fashion by the next Democratic nominee. Though Bush's campaign manager, Karl Rove, dismissed it at the time as a corny gimmick, he later changed his mind. By floating down the river Gore hit small, difficult-to-reach media markets in such key midwestern swing states as Minnesota, Wisconsin, Illinois, Iowa, and Missouri. (He won all but Missouri.) When Rove later sought to target some of these same areas, he discovered that no airport nearby was large enough to land the Boeing 757 that served as Bush's campaign plane. "We never got Bush there," Rove lamented afterward, and he laid the blame for Gore's narrow win in Iowa on that fact.

If the 2000 election supplied the road map for the next campaign, the 2002 midterm elections gave both parties an urgent mandate to reach swing voters. Democratic campaigns mostly outperformed their Republican counterparts in the elections of 1996, 1998, and 2000. (Gore did, after all, win more votes than any U.S. President except Reagan.) This was thanks largely to Election Day voter-turnout efforts, which got Democrats to the polls and often proved decisive. But in 2002 the Republicans shocked the Democrats by besting them on this front, nullifying an important edge. This has set off a pitched battle to capture the narrowing sliver of what pollsters call "persuadables"—the undecided voters who will make the difference in any close election.

In fact, it has sparked a kind of demographic arms race. For the first time, both parties are embracing sophisticated and costly demographic technology that until recently was the province of consumer market-research companies. The Democratic National Committee has acquired a database of 158 million voters it has dubbed the "DataMart." Appended to every name are as many as 306 "lifestyle variables" gleaned from voter files, consumer databases, and other sources. From these, candidates can find out a citizen's voting record, number of children, kind of car, favorite television shows and magazines, and even number of pets. Not to be outdone, the Republican National Committee has its own Orwellian construct, called the "Voter Vault," which contains records on 165 million people.

By drawing samples from the DataMart, the thinking goes, Democratic pollsters and interest groups can create intricate predictive models of where the most sought-after voters will be found. "In a crowded marketplace," the pollster Geoff Garin explains, "it's about being able to know the architecture of the people most likely to be supportive of you and seeking them out."

It is no longer enough to posit that a broad notional category such as "soccer moms" will decide an election. Advances in computer and database technology now offer infinitely more detail, promising campaign staffs the capacity to learn not only which issues matter most to a particular soccer mom but also her home address, the phrasing likeliest to persuade her, and when, how, and by whom she might prefer to be approached. Karen White, the political director for the pro-choice women's group Emily's List, which is working with DataMart information, says, "In the past we've always tried to bring voters to us on our issues. This time we're getting so much insight into their personal lives that we can actually bring what they need to hear to them, on their terms."

The New Democrat Network, a centrist political organization, was among the first in this election cycle to use polling to sketch out a profile of the latest generation of swing voters. Data shared with each of the Democratic candidates (and provided to The Atlantic) describes them as mainly white and also younger, less likely to vote, and more likely than self-identified Democrats or Republicans to characterize themselves as "workaholics." They are most heavily concentrated in suburbs and small cities, and though they disapprove of many Bush Administration policies, they tend to be more religious and to admire military service more than most Democrats do. "On many issues their attitudes correspond strongly with the Democratic Party even though demographically they are closer to Republican voters," says Peter Brodnitz, of the firm Penn, Schoen and Berland, which conducted the poll. The New Democrat Network identified civil liberties and the environment as the two issues on which independents and Republicans most strongly disagree—and, indeed, many of the Democratic candidates have sounded precisely these themes. (Buried in the report's "tactical recommendations" is information that both sides in the next campaign may find useful: independents listen to a disproportionate amount of country radio, and they watch SportsCenter more often than other Americans—a taste, the poll reveals, that corresponds more closely with Democrats' than Republicans'.)

Other organizations, including Emily's List, have conducted broader studies to sort independents into smaller "lifestyle clusters," the better to target them in the fall. Emily's List has identified four basic groups: disengaged "Bystanders," who when motivated to vote lean Democratic; "Senior Health Care" voters, whose gender (predominantly female) suggests an inclination to support Democrats; "Education First" voters, 64 percent female and 66 percent pro-choice but currently more supportive of Bush and the Iraq War than the typical Democrat; and the "Young Economically Pressured," many of whom work more than forty hours a week and may care for an elderly parent. Though this last group tends to support the Democratic position on funding public schools and other issues, its members live predominantly in small towns or rural areas and are culturally conservative.

The challenge for the next Democratic candidate will be reaching all these independents, many of whom live in small cities and suburbs that are gradually abandoning the Democratic Party. The suburban vote, which Bush won narrowly in 2000, continues to grow. Suburban women already tend to vote Democratic, so the nominee must make a special effort to appeal to men, whose vote fluctuates more than women's in presidential elections and who have lately deserted the party in large numbers: men now prefer Republicans over Democrats by 19 percentage points. Efforts to do this are under way on gun control and other issues. Gore was widely thought to have lost blue-collar swing voters in West Virginia and Ohio because of his position on guns and—pollsters argue—how he spoke about it: gun owners believed that Gore would take away their weapons, and voted accordingly. But pollsters discovered that if the discussion had simply been reframed to acknowledge the Second Amendment right to bear arms (as in the phrase "with rights come responsibilities"), 20 percent of gun owners—seven percent of the electorate—would have been inclined to vote Democratic. It's probably no accident that none of the leading Democratic candidates have echoed Gore.

From the archives:

"Blue Movie" (January 2003)
The "morality gap" is becoming the key variable in American politics. By Thomas Edsall

One early lesson that ought to figure in the Democratic campaign is the importance of values, which have replaced income as the best indicator of voting behavior. In the past income corresponded strongly to party preference: voters supported the Democratic Party when they were poor and grew increasingly Republican as they moved up the income scale. That's no longer true. The exodus of white working-class voters from the Democratic Party has been well documented ever since the Republican revolution of 1994. A similar migration away from the Republican Party by affluent suburbanites alienated by the party's social conservatism has received less notice. Just as white working-class voters swung Georgia and other once Democratic states to the Republican Party, affluent suburbanites turned formerly Republican states such as Illinois and New Jersey into Democratic strongholds. The Democratic nominee will have to hold on to these upscale voters while winning back working-class voters.

"We're a party that prefers to talk about issues, not values," says Bruce Reed, who was Bill Clinton's domestic-policy chief. "Clinton demonstrated that if we want to expand our reach, we have to talk in terms of values." Clinton successfully rewrote the language of values to fit his own policies—family and medical leave, the V-chip, school uniforms—rather than those of the social conservatives who popularized it. In fact, both Clinton and Bush set an example for the next Democratic nominee by rising above party stereotype to attract independents: Clinton by endorsing fiscal conservatism, fighting crime, and opposing welfare dependency; Bush by using enough compassionate rhetoric to persuade independents that he wasn't another meanspirited Republican like Newt Gingrich.

It is a rare point of agreement among Democrats that the party owes its win in the 2000 popular vote to an unprecedented mobilization of minority and union voters. Any serious discussion of another close election must be premised on a similar performance. The strong Republican turnout efforts in the 2002 and 2003 elections have only increased the pressure on the next Democratic campaign to keep pace.

One way Democrats hope to do so is by improving their methods of reaching voters. Party bosses once relied on local precinct captains to impart a measure of personalization to even the largest campaigns. In the 1970s and 1980s, however, television developed into a more efficient medium; direct human contact waned, eventually giving rise to the maddening wall-to-wall carpet-bombing of TV attack ads that are the hallmark of modern campaigns. Today, in light of the proven benefits of voter-turnout efforts, strategists in both parties are hoping to combine demographic information with political research in order to repersonalize campaigns and lure back dropout voters who are disillusioned with politics. Emily's List has even asked voters in certain elections to keep a diary of every political contact they received, recording each instance in which a television ad, phone call, or direct-mail brochure caught their attention.

From the archives:

"'The Democratic Party Suicide Bill'" (July/August 2003)
The fate of McCain-Feingold ultimately rests with the Supreme Court. But principle has already cost the Democrats plenty. By Seth Gitell

Perhaps most striking about the Democratic effort to shape the next campaign is its urgency. The McCain-Feingold campaign-finance-reform laws have had the effect of directing much of the available campaign money away from the Democratic National Committee and toward a broad range of liberal interest groups. Within these groups there is a palpable sense of desperation about the party's predicament and a corresponding willingness to experiment that the hidebound DNC rarely displayed. There is also the stinging example of Karl Rove, who brilliantly understood that rigorously pursuing a patchwork of distinct constituencies could add up to an unlikely electoral victory. But above all there is the pervasive fear that the next Democratic candidate will once again be a hairsbreadth from victory—and once again manage to lose.