An Emerging Democratic Majority

By Paul Starr
The American Prospect, November-December 1997

The 1994 election devastated the self-confidence of the Democratic The 1994 election devastated the self-confidence of the Democratic Party, and 1996 only partially restored it. After narrowly escaping the "Republican revolution," many Democrats have lowered their expectations and become resigned to the prospect of center-right government. And now President Clinton's budget and tax deal with the Republicans in Congress has left his own party without a clear long-term agenda or any resources for new initiatives. Especially on the party's liberal side, Democrats are thoroughly demoralized, gloomy about the prospects for recovering control of Congress in 1998 and reviving momentum on what at least used to be the party's distinctive progressive concerns.

Skepticism about progressive possibilities does not simply reflect the latest voting returns, opinion polls, or signals from the White House. Even sympathetic observers don't see why the underlying trends in American society and politics should return the Democrats, much less liberals, to a majority position. The conventional wisdom is that the Republican Party has become the "sun" and the Democratic Party merely the reflecting "moon" of American politics—to use a metaphor first suggested by Samuel Lubell in 1954, when the parties seemed to occupy the opposite roles. Democrats themselves do not have a believable narrative of the future that explains how and why they can become a majority party again. But their long-term prospects may not be as dire as they look. Although my purpose here is not to predict a new majority, I want to suggest why certain social and economic trends over the next 30 years could help Democrats to achieve it—if they can develop the ideas, strategies, and organization to capitalize on the opportunities that these trends represent.

Of course, new majorities are rare, while dreams and theories of new majorities are more common—hence mostly illusory. In recent decades, two theories of new political majorities have proved, if not exactly correct, at least substantially valid. Both were based not merely on a hope, a prayer, or a debatable historical lesson, but on long-term changes in American society that could be the rational basis of new political strategies.

The first was the theory famously proposed in 1969 by Kevin Phillips in what remains the single most brilliant recent work of political forecasting, The Emerging Republican Majority. Published when Republicans were far outnumbered in Congress and had just barely won the presidency after losing seven of the nine previous races, the book should ironically be an inspiration as well as a benchmark for Democrats today. Much of the analysis still stands up a quarter of a century later, even though the author's own views have evolved.

Phillips's original new majority formula was one part political realignment, one part geodemographic transition: The Democratic Party's embrace of black interests had opened the South to the Republicans, while rapid economic and population growth in the Sunbelt presaged a continuing shift of power toward the most reliably conservative region of the country. The analytical force of the book came from Phillips's command of patterns of ethnic settlement and county-level voting since the Civil War. Putting those data together with the growth of the Sunbelt, he correctly anticipated the sources of the Republican ascendancy that would make Ronald Reagan President and Newt Gingrich Speaker of the House. That ascendancy did not happen automatically; the Republican Party drew new leadership from the South and West and altered its policies to take strategic advantage of the opening that Democrats had provided.

The second theory of a new majority, also originating in the late 1960s, was the conception of the New Politics or new liberalism that emphasized such issues as civil rights, consumer protection, broader political participation, openness in government, feminism, and the environment instead of traditional lunch-bucket concerns. The immediate impetus for this vision of a new majority, including the young, minorities, and women, was of course the wave of political energy set in motion by the Vietnam War. But this strategy also built on a long-run trend: Surveys from the 1950s to the 1970s show that Americans did become more liberal on such issues as race, the role of women, sexual behavior, and the environment (though not on economics, taxes, or crime) in a historic shift of opinion that has not been reversed. The new liberalism also took advantage of the opening that the Republicans' Sunbelt strategy was giving Democrats in other regions. And while many analysts now hold this version of liberalism responsible for the decline of the Democratic Party, it provided new vitality (particularly in the form of hard-working, highly committed candidates) and helped Democrats keep control of Congress and state legislatures for another quarter century after the 1968 election and Phillips's forecast, for a total run of 62 years, about twice the duration of typical party regimes.

But how can these two theories, with opposite implications, have both been right? As a result of the trends that they identified and strategies that they suggested, the parties have reached a position of rough parity in electoral strength, each with the capacity to form a new majority—that is, a majority different from the one it previously assembled. Republicans can now usually count on majorities among men, Democrats on majorities among women. Republicans win majorities among whites; Democrats can sometimes assemble majorities from whites and other groups combined. The parties have exchanged regional bases with the South trending toward Republicans, New England toward Democrats. The rough parity between the parties has produced a divided federal government in 22 of the past 28 years. In 1996 the total vote for the House of Representatives was split almost evenly—49 percent for the Republicans, 48.7 percent for Democrats. The Republicans maintained control primarily because of the way in which the votes were distributed; they won the overwhelming majority of close races, while Democratic votes were clustered in districts where they won by lopsided margins. Even so, the Republican House majority in 1997 is the smallest in four decades.

Rough parity in electoral strength does not, however, mean parity in all respects. Rising to parity creates a different sense of direction from falling to the same point. Some years ago, after Harvard scored two touchdowns in the final minutes of the Harvard-Yale game, the Harvard Crimson ran a headline: "Harvard Beats Yale, 24-24." Like Yale, the Democrats seem to have been losing tie games. While many observers have talked of party decline and "de-alignment" as if they afflicted both parties equally, the changes have been asymmetrical, as my colleague Robert Kuttner persuasively argued a decade ago in his book The Life of the Party. It's the Democrats whose machinery has deteriorated most (the party as organization) and who have lost most in popular self-identification (the party in the electorate). Since 1994, Democrats have also surrendered much of their own agenda to stay politically competitive. They have had fewer resources and run into more trouble (and scandal) in scrambling to obtain them.

Financial scandals have decimated the leading parties of Italy and Japan in recent years, and they could similarly do severe damage to the Democrats in the wake of the 1996 campaign (if only by chilling donors in a system still dependent on private money). Yet if we look to the long term, there are signs more favorable to the Democrats: demographic growth among groups of voters with Democratic affinities; economic trends likely to emphasize the importance of issues identified with the Democratic Party; historical shifts as Democrats finally shed some of the burdens they have carried since the 1960s. These developments pose two related strategic and intellectual challenges: Are the Democrats capable of capitalizing on these emerging tendencies? And in the face of scandals and cynicism, can they revive themselves not just as a party but as a cause?


The 1996 presidential election diverged in several ways from the patterns of political support that Phillips had predicted in 1969. Clinton did better, for example, among Catholics and in the Midwest. But, most remarkably, he won a series of states across the southern rim of the United States—Florida, Louisiana, Arizona, New Mexico, and California—that were supposed to be anchors of the new Republican majority. What makes these results especially significant is that, except in Louisiana, Clinton and other Democrats received critical support from two groups whose numbers will increase dramatically in coming years—Hispanics and the elderly. Continued Democratic support from these groups certainly isn't guaranteed, but their growing numbers provide a historic opportunity for a flip of the lower, "Latinized" Sunbelt back to the Democrats.

Although 1996 was not generally a realigning election, it may have had something of that character for Hispanic voters. Realigning elections characteristically see both an increase in turnout and a swing in party support, and among Hispanics both took place in 1996. Nationally, the Hispanic vote rose an estimated 22 percent over 1992, and Hispanics cast 72 percent of their votes for Clinton, up from 55 percent four years earlier. (These and all other exit poll data for 1996 that I cite come from the Voter News Service exit poll; some of the figures were generated from the data on the CNN/All Politics site on the World Wide Web.) In what may be a signal of future bloc voting, 78 percent of Hispanics under age 30 voted for Clinton. In Arizona, which no Democrat had won since 1948, Hispanics put Clinton over the top with 81 percent of their votes, as they did in New Mexico, where Clinton "merely" won 66 percent of Hispanics. Perhaps the single most electrifying results were in California, where Loretta Sanchez upset Robert Dornan in a congressional race in what used to be the conservative bastion of Orange County, and where the Democrats retook control of the state assembly and chose a Hispanic, Cruz Bustamente, as the new Speaker. Clinton won 75 percent of the California Hispanic vote; he even won half of the Hispanic vote in Florida despite long-time Republican strength among Cubans.

According to Census Bureau projections, Hispanics will represent an astounding 44 percent of net population growth in the United States through 2025. The source of this growth is not only continuing immigration, but also Hispanics' relative youth and high fertility rate. The median age of Hispanics is 26, compared to 35 for the overall U.S. population; thus even if Hispanic women had children at the same rate as non-Hispanics, the Hispanic population would grow more rapidly. Census projections for 2025 show Hispanics growing to 18 percent of the population in the United States as a whole, but to 32 percent in Arizona, 38 percent in Texas, and at least 43 percent in California.

Moreover, among Hispanics, the slowest growing group is the most Republican, the Cubans, with a median age of 41, while the most rapidly growing groups are those from Mexico and Central America, who tend to be more Democratic. Thus the internal dynamics of the Hispanic population augur stronger Democratic leanings.

To be sure, several things could upset these projections. The Hispanic population will be smaller if immigration is sharply reduced or if Hispanic fertility rates converge more rapidly with the general population than the Census assumes. Some critics, such as the columnist and population watcher Ben Wattenberg, argue that Census forecasts of fertility are generally too high. But even if Wattenberg is right, non-Hispanic fertility rates might fall in parallel with the Hispanic fertility rate, leaving as large a differential. And tighter immigration laws might not halt the growth of the Hispanic population if, as Douglas Massey, a sociologist at the University of Pennsylvania, argues, greater economic integration between the United States and Mexico (and other Latin American countries) increases the flow of people along with goods regardless of immigration laws.

Hispanics also might not vote in numbers that reflect their share of the population. Today Hispanics represent a much smaller percentage of the electorate than of total population because of their low median age, the high proportion of noncitizens, and low voter turnout. Nationally, Hispanics made up 10.5 percent of the population in 1996 but only 4 percent of the electorate; as they rise to 18 percent of the population, they have the potential to double or triple their share of the vote. Whether they will close the gap in turnout with other groups is impossible to say; the spurt in 1996 could turn out to be a special case. But as their median age increases and a larger proportion become citizens because they have naturalized or were born here, the Hispanics' share of the electorate should grow faster than their share of the population.

The Hispanic turn toward the Democrats in 1996 could also prove ephemeral. Republicans might increasingly appeal to Hispanics on the basis of conservative cultural values or by running more Hispanic candidates, and Hispanics themselves might become more conservative as they advance socioeconomically. The history of other immigrant groups suggests, however, that early political identifications tend to be highly persistent; Irish Americans, for example, have maintained their identification with the Democratic Party long after its original basis disappeared. Some writers have properly cautioned that Hispanics are not as reliably Democratic or liberal as African Americans. But African-American voting patterns (roughly 90 percent Democratic) aren't a reasonable standard. Not even Christian fundamentalists vote Republican at that rate.

Of course, the Hispanic preference for Democrats in 1996 was well above prior levels because of the alarm created among Hispanics by Proposition 187 in California, the congressional cutoff of welfare benefits and other services to legal immigrants, and Republican support for making English the exclusive language of public business. Yet even if Republicans soften their stands, there is no mistaking which party will remain the home of both nativist sentiment and opposition to social programs that benefit groups with large numbers of poor working families. Family incomes among Hispanics, again except for the Cubans, continue to lag far behind those of non-Hispanic whites. Given recent trends toward growing income inequality and relatively slim gains among low-wage workers, Hispanics seem likely to remain predominantly working-class in orientation and more favorable to the party that supports increases in the minimum wage and earned income tax credit and is more closely identified with unions, expanded educational opportunities, and broader access to health care.


Nationally, voters over age 65 favored Clinton over Dole by 51 percent to 42 percent in 1996. Although this nine-point margin was just above the average for all voters, it was significantly higher than among voters between the ages of 50 and 64, who split for Clinton by only 46 percent to 44 percent. Except for the elderly, age was positively correlated with voting Republican; the deviation from this pattern among the over-65 voters suggests some distinctive influence affecting those in retirement. The preferences of the elderly particularly mattered in Florida, where they favored Clinton by 56 percent to 40 percent and tipped the state to him, giving Democrats their first win in a presidential race in Florida since 1976.

In 1996, the elderly made up about 13 percent of the national population and 16 percent of voters; in 2025, they will make up one out of five Americans and perhaps about one-fourth of the electorate. As with Hispanics, the growing elderly population in coming years will be regionally concentrated; Census projections for 2025 show the elderly rising from 19 percent to 26 percent of Floridians (and probably close to a third of voters). The regional concentration of Hispanic and elderly voters has particular relevance to presidential elections. During the 1980s, some observers spoke of a Republican lock on the Electoral College in large part because the party's base in presidential elections seemed to include California, Texas, and Florida. By 2025, these states will be the nation's three most populous, and if the concentration of Hispanic and elderly voters gives Democrats an edge in those states as well as in traditionally Democratic New York (the fourth most populous state in 2025), Democratic candidates may begin presidential races with a big electoral college advantage.

Compared to the Democratic leanings of Hispanics, however, those of the elderly are much weaker to begin with and therefore more uncertain in the future. One key question here is whether their voting patterns mainly reflect formative political experiences earlier in life, their current economic interests (such as Social Security), or demographic factors, such as differences in mortality rates. Today's elderly came of age during the middle decades of the century when there were high levels of unionization and Democratic partisan identification. The elderly of 2025 will be drawn mainly from today's middle aged—the most Republican cohorts in 1996—who formed their views when unions and Democratic identification were declining. If such generational effects predominate, we might expect a shift among the elderly toward more conservative voting.

Some evidence does suggest generational differences between today's elderly and those just behind them, but the data from the 1996 presidential race are ambiguous. The generational effects should apply no less to men than to women, but men 65 and older gave Clinton about the same proportion of their votes (44 percent) as did men between the ages of 50 and 64. Clinton's wider margin among the elderly than among the 50- to 64-year-olds was due entirely to a four-point-wider edge among elderly women and to the larger proportion of women among the elderly population because of their lower mortality rates. These patterns suggest that, at least in 1996, the Democratic margin among the elderly was related to the gender gap.

Voting patterns among women under age 65, particularly differences by marital status, may offer a clue to future trends. Among the married middle aged, there was no gender gap in presidential voting; married 50- to 64-year-old women voted for Dole by 51 percent to only 42 percent for Clinton, much as their husbands did. In contrast, unmarried 50- to 64-year-old women favored Clinton by 63 to 31 percent, displaying the same voting preferences as younger unmarried women, more than 60 percent of whom also voted for Clinton. Single women might be more partial to Democrats for a variety of reasons: more experience in the workforce, higher probability of depending on government programs, and—not least of all—less influence by more conservative men.

As the 50- to 64-year-old cohort ages, the proportion of women will increase, and more of these women will become single through divorce or widowhood (though the latter may have less impact on political attitudes). On the basis of these demographic factors alone, the elderly of 2025 will probably become more Democratic than they were in middle age.

And as the 50- to 64-year-old cohort retires, Social Security and Medicare should also become more salient issues for them. But how they construe their interests as beneficiaries may depend on whether those programs continue to exist in their current form. Extensive means testing, for example, could remove the more affluent elderly from the program and turn them into opponents of more generous benefits. Similarly, privatization of Social Security could expand the number of the elderly who see themselves as investors and reduce the number who see themselves as beneficiaries. This is precisely the objective of many who favor means testing and privatization. And some version of these changes may well result from the bipartisan reform of Social Security and Medicare that Clinton is now calling for. Even with some means testing and partial privatization, however, the most likely outcome is that the elderly will remain the age group most dependent on public social protection—policies historically identified with the Democratic Party.


In 1996, the age group that supported Clinton and the Democrats most strongly was actually not the elderly, but the youngest voters. Those between the ages of 18 and 29 favored Clinton by 53 percent to 34 percent; first-time voters gave him an even higher margin, 58 percent to 40 percent; and, according to the pollster Stanley Greenberg, surveys of high school students showed still stronger support. This is a reversal from the pattern in the 1980s, when the young were more Republican; as Reagan tutored new voters then, so Clinton and Gore may be doing in the 1990s. No doubt Dole's age cost the Republican ticket support among the young, a factor unlikely to be repeated. Clinton also did well among the young because of demographic characteristics, such as low income and unmarried status, that will become less pronounced as these young voters age.

But the Democratic leanings of the young may also herald a historical shift. Beginning in the late 1960s, Republicans were able to paint Democrats as being weak on crime, morality, and national defense and to win over much of their traditional white working- and middle-class base. Clinton's ability to reclaim these voters may stem not only from his personal success in reframing the social issues, but also from the diminishing resonance of appeals rooted in the experiences of the 1960s and 1970s. The fading power of the past may be showing up first among younger voters, who have no memory of those years. And as time lifts that onus from the Democrats, the Christian right is creating new burdens of the opposite kind for Republicans.

The swing among young voters may also be connected to economic issues that work in favor of Democrats. Stagnant earnings and cutbacks in fringe benefits have acutely affected workers in their twenties. New jobs, particularly in small firms and the service sector, often do not carry the health insurance and pensions, much less job security, that were long part of the standard employment package. If younger workers and their families are going to receive health coverage and other benefits, they are almost certainly going to need government's help, either directly in public programs or indirectly in employer mandates. The Democratic Party is the only political vehicle available for such demands.


Democrats certainly cannot take Hispanics, the elderly, the young, or any other group for granted. The trends only open up possibilities. Some of the trends even threaten to produce cleavages among the very groups that Democrats seek to unite. The aging of the population brings higher costs for Social Security and Medicare, but because total spending will likely be constrained, the politics of the budget could turn even uglier than in the past—in the nightmare scenario, into a civil war of the welfare state with older whites on one side and younger Hispanics and blacks on the other. Support for public education has already eroded because of the disparity in racial and ethnic background between urban school children and taxpayers; given the rising share of Hispanics in the schools, white support for public education may erode even more. The growth of the Hispanic population may also further arouse among whites anxieties already evident in the vote for Proposition 187 and the English-only movement. Thus, the same demographic trends that might benefit Democrats could also divide them.

To maintain support among these and other groups, however, Democrats do not need to be single-minded advocates for interests narrowly conceived; they have to be the responsible guardians for legitimate interests, anchored in broadly shared values. Democrats need to make clear their fundamental concern for immigrants by strongly defending their civil rights and opposing the English-only movement, but they should be wary of supporting high volumes of legal immigration and thereby undercutting the economic position of low-wage workers. Democrats ought to be clear about protecting the integrity of social insurance programs and, for that very reason, be willing to compromise on such measures as raising the age of eligibility for Social Security; the current and soon-to-be elderly will likely accept a marginal reduction in benefits in exchange for the assured longevity and solvency of the programs. Democrats should similarly support expanded educational opportunities, a living wage, and other policies that benefit young workers and their families, but they do not need to develop separate programs that exacerbate racial and generational cleavages.

In 1996, the Republicans drove the elderly and Hispanics toward the Democratic Party by supporting measures inimical to their interests, and they alienated the young with a candidate who seemed to belong to another era; Republicans are unlikely to keep repeating the same mistake. But what happened in 1996 does reflect more than a casual Republican impulse. The conservative antagonism to government is likely to keep threatening those who need it. While the elderly depend on social insurance programs, Hispanics and African Americans depend on public spending for education and other social services because they are disproportionately young and poor. The other demographic groups that supported Clinton—unmarried women of all ages and young men and women—tend to face more economic insecurity and have more need of government than older men and middle-aged married women do. The core of the Democrats' emerging majority consists, as it has since the New Deal, of the groups that are struggling hardest to take care of themselves and their families. Helping them realize that aspiration ought to be central to the purposes of the Democratic Party. Democrats should appeal to these groups not merely because they make up a new majority, but because their aspirations are a just and necessary cause.


Whether the Democratic Party itself can become a cause for the movements historically allied with it is an open question. Democrats are divided about what kind of cause the party represents, and each of the competing factions has its own theory of a new majority. On the left, populist-progressives see a "sleeping majority" that requires stirring nonvoters from their political slumber, and on the right, New Democrats see a new information-age centrist majority that includes independents and moderate Republicans allied with moderate Democrats.

The difficulty with the populist strategy is arousing enough nonvoters to win elections; people who tune out politics are inherently hard to reach. A hard-edged populism may also inadvertently mobilize opponents as well as supporters and thus have a negligible or even counterproductive impact. As a short-term proposition, the New Democrats' approach is more likely to succeed. Just as it is easier to sell a new brand to those who have bought another brand of the same product than to people who haven't bought any, so it is easier to sell a candidate or reformed image of a party to independents and moderates who vote than to nonvoters.

But while attracting middle-class independents and Republicans requires narrowing and blurring the differences between the parties, activating low-income nonvoters could create an electorate more friendly to progressive ideas. In the long run, Democrats would be better off with an expanded electorate in which the median voter was closer to their position than with a smaller electorate in which they moved closer to Republicans—better off because even if they chose to make tactical moves toward the center, the electorate would be weighted further to their side. The populist approach would also be more likely to maximize the effect of the demographic trends favorable to Democrats. The growing Hispanic population turned out to vote in larger numbers in 1996, but it still lagged far behind the rest of the country. A politics addressing the needs of low-income workers may bring more of them into the electorate. Similarly, an inclusive, progressive approach to education and living standards is more likely to engage young people. The long-term interest of Democrats is to invest in a broader electorate and to develop ideas and networks of organization that connect with the currently disengaged.

The New Democrats, however, have not articulated a program that addresses, much less stirs, the politically disengaged and economically insecure. The vision of America favored by the Democratic Leadership Council and its Progressive Policy Institute highlights the benefits of the information revolution and global economy but downplays the losses to those least capable of taking advantage of them. Like Gingrich, some New Democrats have accepted the view derived from Alvin Toffler that the United States is entering a new technological era that dictates "demassification" of large institutions, including "big public systems." In line with that view, they have supported partial privatization of Social Security, Medicare, and public education. But why the information revolution should favor privatizing these services is obscure. What is clear is that privatization would aggravate inequalities in these spheres and undermine the already depleted sense of common social obligation in America. These policies threaten to alienate groups vital to a new majority, drive a wedge through the Democratic Party, and give conservatives the necessary margin (and cover) to enact their agenda.

To their credit, the New Democrats and President Clinton have helped to reconstitute the moral authority of the Democratic Party by redefining the political middle ground on the social issues, such as crime and "family values," that hurt the party badly in recent decades. The New Democrats' "tolerant traditionalism," as Bill Galston calls it, has more popular support than either the conservatives' intolerant traditionalism and what is perceived to be (and unfortunately sometimes is) the indiscriminate postmodernism of the left. Although often presented as a repudiation of liberalism, the New Democrats' views are not especially conservative on the issues championed by the "new liberalism" of the 1960s—civil rights, the role of women, environmentalism, openness in government. The divisions inflamed by the Vietnam War have now faded. However much they may vex each other, the right and left of the Democratic Party are much closer than they were during the long period when southern Democrats were bitterly opposed to the national party.

Each side brings valuable assets to the task of building a new majority. Liberals and progressives are vital to party renewal because the progressive project has the greater capacity to inspire commitment to the party as a cause and expand its reach across the electorate and among nonvoters. Liberals are unlikely to make up a majority in the general population, but like conservatives among the Republicans, they can realistically aspire to be a majority within America's majority party. Managing this role requires a sense of both strengths and limitations. Liberals and populist-progressives have a right to insist on their core role in setting the party's agenda. But their influence will often be less than their share of party activists might appear to warrant because moderate voters (and, alas, donors) will continue to provide the additional votes and resources needed to win general elections. If the Democratic Party is to build a new majority, it will need both its liberal and New Democrat wings. The party won't fly without them both.

This article is adapted from an essay in the book, The New Majority, edited by Stanley B. Greenberg and Theda Skocpol.

Copyright 21997 by Paul Starr Preferred Citation: Paul Starr, "An Emerging Democratic Majority," The American Prospect, November-December 1997. This article may not be resold, reprinted, or redistributed without prior written permission from the author. Direct questions about permissions to