America Fifty/Fifty

G. K. Chesterton’s famous description of the United States as “a nation with the soul of a church” was never more apt than during the 2000 presidential campaign. Religion was in the limelight from the moment that George W. Bush named Jesus as his favorite political philosopher to the day when Joseph Lieberman joined the Democratic ticket, quoting the Book of Chronicles. And when it was all over, Bush entered office amidst a flurry of worship services, clerical blessings, and religious consultations.

The pervasive religiosity of the campaign helped produce one of the closest and most controversial elections in history. America seemed to have two souls”or one torn in half. Experienced observers from journalist Thomas Edsall to social thinker Francis Fukuyama saw in the results a massive cultural divide. Newspapers produced brightly colored maps limning the political boundaries between the rustic heartland and cosmopolitan urban centers. The sharpest differences between the official party platforms were on cultural matters, from abortion to school vouchers, and religious groups campaigned vigorously for their respective champions. In the end, cultural disputes and widespread dismay over the country’s moral state overshadowed the economic optimism that was expected to put Al Gore in the White House, allowing Bush to eke out a victory.

Despite the precarious electoral margin, the public was not as bitterly divided as the pundits thought. Even some deeply religious voters were only vaguely conscious of the stark cultural differences evoked by the campaign. And only about one-half of the eligible electorate even bothered to vote. Alan Wolfe might easily have seen “one nation after all,” albeit one strangely aloof from the struggle for its soul. Indeed, many citizens were thoroughly disengaged from public affairs”or, to paraphrase Robert Putnam, they were bowling alone and abstaining together. The candidates seemed to sense this civic malaise and drew up alternative blueprints for rebuilding civil society: Gore proposed a revitalized public sector, and Bush a reenergized private one.

All these tendencies converged in a “50 percent solution”: half the people voted, and half of those backed each major candidate. (See table, p. 21.) More fundamentally, the outcome revealed the consolidation of a new religious order in American politics, an altered relationship between faith and public affairs.
Religious Groups and the 2000 Election Presidential Vote (in percent) Nonvoters (in percent) Bush Gore Latter-day Saints (“Mormons”) 88 12 29 White Evangelical Protestants 76 24 51 Traditionalists 86 14 37 Centrists 75 25 58 Modernists 46 54 63 White Mainline Protestants 61 39 47 Traditionalists 75 25 43 Centrists 60 40 50 Modernists 48 52 46 White Roman Catholics 53 47 47 Traditionalists 75 25 41 Centrists 53 47 46 Modernists 30 70 55 Secular 35 65 52 Hispanic Catholics and Protestants 28 72 70 Jews 23 77 18 Black Protestants 4 96 47 Totals 49.8 50.2 50 Throughout American history, voting has been shaped by ethnoreligious loyalties in which distinct religious traditions, often wedded to ethnic and racial identities, exhibited characteristic partisan ties. Such connections reflected the special culture of each tradition, formed in local neighborhoods and churches, engendering both ideological commonalities within the tradition and persistent social conflict with other traditions. In the mid-twentieth century, for example, mainline Protestants supplied most of the Republican Party’s leaders and its most faithful voters, while Catholics, Jews, and other religious minorities”including regional out-groups such as southern evangelicals”constituted the bedrock of the Democratic Party. These alignments survived even the class politics of the New Deal.

Since the 1960s, however, these attachments have been transformed by three critical developments. First, some religious traditions have shifted their partisan loyalties. White evangelicals, for example, abandoned the Democrats for the GOP in droves, while African-American Protestants finally deserted the party of Lincoln, becoming monolithically Democratic. And many Catholics strayed from their ancient Democratic home, with some joining the Republicans but others becoming swing voters available to either party under the right circumstances. Evangelicals and black Protestants are now prominent among the elites of their respective parties, and Catholics are leaders in both. Such shifts have been a staple of electoral politics, reflecting the political values of the special cultures of religious traditions. This revised version of the old ethnoreligious politics is still taken for granted, although it is not often seen as particularly religious. Even polls sponsored by the religiously tone-deaf New York Times routinely distinguish the political behavior of “white Protestants,” “Catholics,” and “Jews.”

More recently, a new kind of partisan divide has opened, based on conflict within religious communities over belief, practice, and the role of religion in society. By the 1980s disputes pitting “traditionalists” of many stripes against a variety of “modernists” had partially restructured the larger American religious traditions, modifying their distinctive cultures. Traditionalists in all communities began to find more in common with each other than with modernists in their own traditions. Modernists had a comparable experience. The famous “culture wars” theory both describes and then overstates these trends by ignoring a crucial fact: many of the faithful have not been drawn into these disputes, but remain loyal to the “center” of their historic tradition.

Despite the continued presence of such “centrists,” the incipient lines of cleavage within traditions and the budding alliances across old boundaries have had substantial political ramifications. Traditionalists have gravitated toward the GOP, while modernists have joined the Democrats, fracturing”but not destroying”the old ethnoreligious alignments. These trends are most visible among religious and political activists; the real question is how far the partisan polarization will extend to people in the pews. Whatever the answer to that question, this new politics of religious believing and behaving has complicated the old politics of religious belonging.

The new religious order has a third feature: the persistent (and, some would say, growing) detachment of Americans from civic life. Turnout in presidential contests has steadily declined since 1960, with only modest and temporary upturns. Although political factors have played a role in depressing turnout, the deeper cause lies in the decay of the nation’s “social capital”: the everyday social and organizational connections that nurture citizens’ values, define their interests, and connect them to public affairs. As Robert Putnam and his colleagues have amply documented, voluntary activity in organizations has declined significantly since the 1960s, affecting groups from the American Bowling Congress to PTAs to the League of Women Voters. Not even religious life, a most resilient form of social interaction, has been exempt. The atrophy of mainline Protestant churches and dramatic reduction in Catholic mass attendance have been accompanied by the growth of the unaffiliated or secular population. Indeed, most religious institutions have suffered, even the conservative churches that have remained vital by dint of extraordinary effort. Simply put, good “organization citizens” are harder and harder to find. Thus, the complex politics of religious belonging, believing, and behaving occurs in the wider context of social disengagement.

These three changes in religious politics are closely related, but in complex ways. The partisan realignment of evangelicals and black Protestants and the partisan dispersion of white Catholics were strongly encouraged by the culture wars and fostered by an uneven pattern of decay in social capital. And there was a strong link between the culture wars and the decay of social life itself. In many respects, the cultural warriors of right and left have been responding to the decline of social institutions. On one side, traditionalists lamented the erosion of family life, religious institutions, and traditional values caused not only by impersonal social developments but also by the conscious assaults of modernists. For their part, modernists attacked traditional institutions in the hope of building broader and more inclusive communities. Indeed, modernists”and their secular allies”were often as disturbed by decaying civic life as the traditionalists, but preferred to envision new institutions, not resuscitate old ones.

In this sense, the collapse of civic life undermined the special cultural expression of many social groups, threatening to create Richard John Neuhaus’ “naked public square.” Ironically, this denuded public life gave an opening to the warriors of right and left, each supported by their small, yet cohesive, cultures. Using the leverage of hot-button moral issues, these rivals have been able to shift the dominant partisan allegiances of the major religious traditions. Unfortunately, the culture wars may have accelerated the decline of civic life by damaging institutions and further alienating the disengaged. Indeed, for all the tumult, the generals in the culture wars still lack troops: many Americans refuse to volunteer, some are adamant draft resisters, and many have lost their taste for service, even in areas far removed from the front.

To describe the electoral outlines of the new religious order, we use the third quadrennial Survey of Religion and Politics, conducted by the University of Akron for the Ethics and Public Policy Center, and sponsored by the Pew Charitable Trusts. This survey has a larger sample and more religious questions than most polls, allowing us to analyze key elements of the new order in considerable detail. But before we undertake that task, we need to explain our system of religious classification.

Religious traditions represent the “belonging” aspect of religion and are key to understanding its role in politics. These traditions comprise groups of denominations and congregations that share common beliefs, rituals, and historical experiences. Assigning a poll respondent to the correct religious tradition requires in-depth questions, a task that most survey organizations still avoid. The largest religious traditions in the electorate are the evangelical Protestant (26 percent in the Akron survey), mainline Protestant (16 percent), white Roman Catholic (17 percent), and black Protestant (9 percent). Hispanic Christians (both Catholic and Protestant) constitute distinctive ethnic sub-traditions (7 percent). The United States is also home, of course, to many smaller religious traditions, including Judaism, Islam, Latter-day Saints, and others. In addition, a growing number of citizens are essentially “secular” (16 percent), although this group obviously lacks some traits of a tradition properly defined.

Our classification also incorporates the culture wars divisions, based on religious beliefs, behaviors, and ties to religious movements. For purposes of analysis and illustration, we have used this information to divide evangelicals, mainline Protestants, and white Catholics into “traditionalists,” “centrists,” and “modernists.” Although the precise allocation procedure is too complex to present here, we can easily summarize the results. Within each community, “traditionalists” profess orthodox beliefs, exhibit high levels of public and private religious behavior, and identify with “sectarian” movements (e.g., for Protestants, “fundamentalist,” “Pentecostal,” or “charismatic”). “Modernists” hold more heterodox beliefs, exhibit modest levels of religious practice, and identify with church-like movements (ecumenism, liberalism). Naturally, “centrists” fall between the other two camps on all these measures.

We have not divided smaller religious traditions for two reasons, one practical, and the other theoretical. First, even in a large survey these traditions are too small to subdivide, although evidence indicates such elaboration might sometimes be useful (as in the case of Judaism). Second, most minority traditions are best thought of as persisting examples of the old ethnoreligious politics, in which common theological, racial, or ethnic traits overwhelm any internal theological differences, at least for electoral purposes. Black Protestants, Mormons, and Hispanic Christians fit this older model (as presumably do Muslims, Hindus, Buddhists, and other minorities not examined here).

The resulting religious classification (listed down the side of the accompanying table) provides us with a snapshot of the new religious order, as manifest in the 2000 election. We have listed religious categories from most Republican to most Democratic, as measured by presidential vote (congressional voting and partisan attachments show much the same pattern). The “Bush” and “Gore” columns report the proportion of the two-party vote received by each from the various religious groups. The final column shows the proportion of a group that did not vote. (The few minor-party voters are omitted.)

A quick glance shows that religious groups made vastly different electoral choices. The Latter-day Saints gave Bush 88 percent of their vote, with Gore receiving only 12 percent. Obviously, there was little real political polarization within this solidly Republican group. All the major traditions, however, reveal at least some culture war divisions. Evangelical Protestants were the most Republican, giving Bush 76 percent of the vote, but evangelical traditionalists gave the GOP candidate 86 percent, while centrists gave him 75 percent, and modernists only 46 percent. Mainline Protestants, the GOP’s traditional ethnoreligious bastion, were less Republican (61 percent overall), but exhibited the same internal pattern, with the Bush vote ranging from 75 percent among traditionalists to only 48 percent among modernists.

These results confirm the strong bonds”some ancient, some more recent”of white Protestants to the GOP. One of George Bush’s political strengths was that he neatly straddled the evangelical/mainline split, being both “born-again” and a United Methodist. Contrary to some early expectations, he did very well among traditionalists in both groups, perhaps his reward for assiduous cultivation of conservative religious leaders. He attracted many centrists as well, depriving Southern Baptist Al Gore of the more moderate Protestant voters who might well have made him President.

Catholic voters were the object of much political, academic, and journalistic speculation in 2000. In recent years they have become a swing electoral constituency, often closely divided and usually tipping toward the winner. The Bush campaign targeted “church-attending” traditionalist Catholics, an effort hampered initially during the GOP primaries by the flap over his speech at fundamentalist Bob Jones University, an anti-Catholic stronghold. The campaign recovered quickly, however, and maintained the focus on Catholics throughout the fall. And for the first time we see incontrovertible evidence that the culture wars have polarized Catholic laity. Although Bush carried only a narrow majority among white Catholics (53 percent), he defeated Gore by a three-to-one margin among traditionalists, a big improvement over Bob Dole’s 53 percent in 1996. Centrists were closely divided, with a slight edge to Bush, but Catholic modernists voted Democratic in overwhelming numbers. These internal divisions among Catholics are the largest ever found by survey research, reaching all the way back to the 1930s.

The rest of the table demonstrates that secular voters and most religious minorities fell squarely in the Democratic camp. Secular voters favored Gore over Bush by an almost two-to-one margin. Hispanics also voted predominantly Democratic, although Protestants were somewhat more supportive of the GOP than were Catholics. Jews also remained in their traditional party location and black Protestants were almost unanimously Democratic. Other minority traditions (Buddhists, Hindus) also favored the Democrats, although some outside evidence suggests that Muslim voters were predominantly for Bush, following his endorsement by leading Islamic organizations.

Clearly, there is a new religious order in American electoral politics, one characterized not only by the distinctive partisanship of religious traditions, but also by theological polarization within the nation’s three largest traditions. Perhaps the culture wars will ultimately obliterate these old ethnoreligious attachments, but in 2000 religious traditions still mattered and they will for some time to come. The large centrist group in each tradition epitomizes its central partisan tendency, remaining somewhat deaf to the battle cries of the culture warriors. And ethnoreligious minorities are a vital part of the new religious order, mostly supporting the Democratic Party, the ancient home of cultural minorities of all sorts.

The practical political implications of the new order are more obvious if one considers the composition of each party’s vote. In terms of religious tradition, white evangelicals, mainliners, and Catholics provided the great bulk of the Bush vote (40, 21, and 20 percent, respectively), while black Protestants (20 percent), seculars (19 percent), and Jews and other religious minorities (18 percent) provided a solid majority of Gore votes. Looking at the coalitions from a culture wars perspective, evangelical, mainline, and Catholic traditionalists, combined with Latter-day Saints, provided 36 percent of the Bush vote, a figure that rises to 74 percent if centrists from those traditions are included. Even more striking, such traditionalists provided only 8 percent of Gore’s total vote, a figure that rises only to 31 percent if centrists are added. The two political parties have very distinctive religious constituencies to appease.

Of course, some observers will deny that religious differences are real, contending that they reflect the disguised operation of social class, gender, marital status, age, or regional cultures. And there is no doubt that such factors are also related to the vote. For example, voters earning less than $25,000 gave Gore 62 percent, but those over $75,000 only 43 percent”a difference comparable to that between evangelicals and Catholics. Voters with a high school diploma or less gave Gore 53 percent, while college graduates gave him 47 percent. (His best educational category was actually postgraduates at 54 percent.) This gap is slightly smaller than that between evangelical traditionalists and centrists. Similar differences appear elsewhere: Bush drew more men than women, more from the young than the elderly, and more married than single people.

But the modest size of these demographic gaps suggests that personal status factors will not eclipse the influence of religion. And, in fact, when we control for the impact of education, income, age, gender, marital status, and region, the resulting vote distribution looks very much like that in the table. Indeed, education and income”prime indicators of social class”virtually drop out of a multivariate analysis when these religious measures are included. All this goes to confirm that, to borrow a title from an earlier First Things article, “It’s the Culture, Stupid” (April 1994).

What accounts for these configurations? A simple question, but the answer is complex, and we can suggest only the beginning of one. Certainly these patterns express the distinctive party attachments of religious groups (what political scientists call “party identification”) and, for that matter, their evaluations of the candidates. Evangelical traditionalists are heavily Republican, and liked Bush; Catholic modernists are largely Democratic, and preferred Gore. But these factors are to some degree the product of the groups’ characteristic political attitudes, not only on the great cultural conflicts of the past three decades, but also”to a lesser extent”on economic and “scope of government” issues dating all the way back to the New Deal. The religious constituencies at the heart of the Republican and Democratic coalitions are distinctive on both sets of core issues.

To assess the role of moral issues, we used voters’ opinions on abortion, women’s rights, gay rights, and government-sponsored lotteries. Similarly, to measure New Deal issues and their progeny, we looked at questions on increasing government services, enacting a national health care program, government spending on hunger, aiding minorities, and protecting the environment. For each policy dimension we calculated an overall conservatism/liberalism score. If the new religious order is produced by the political values distinguishing religious groups, we should find the same basic pattern on these two issue complexes that we find in the vote. And we do.

Sharp cultural issue divisions certainly underlie the voting patterns. Not surprisingly, Latter-day Saints hold the most conservative moral values, but evangelical traditionalists finish a close second, followed at some distance by Catholic, and then mainline, traditionalists. In each major tradition, centrists hug the national averages, while modernists fall on the liberal (and Democratic) side, by only a small margin among evangelicals, but more decisively among Catholics and, especially, mainline Protestants. Indeed, modernist mainliners’ strong moral liberalism matches that of secular voters, and is exceeded only by Jewish voters. Black Protestants, Hispanic Christians, and non-Christian religious minorities fall solidly on the liberal side, close to Catholic modernists, but falling well short of mainline modernists. On the whole, though, the position of religious groups on moral and cultural issues closely mimics their voting pattern.

Although religious groups have distinctive views on social welfare issues as well, the pattern is a little different from that on moral questions. Once again, Latter-day Saints are most conservative, followed closely by evangelical traditionalists, and more distantly by mainline traditionalists. But at this point the story gets more complicated. First, both centrist and modernist evangelicals tilt to the conservative side, as do mainline centrists, perhaps reflecting classic Protestant economic individualism. Second, there is little difference among the three Catholic subgroups, as both traditionalists and modernists fall barely on the conservative side, while centrists sit right on top of the national mean. We can only speculate whether Catholic traditionalists’ relative liberalism (compared to their Protestant counterparts) reflects loyalty to the Church’s historic social teachings or simply a collective memory of the day when Catholics benefited from the New Deal’s social welfare policies. Finally, the liberal side on social welfare issues consists, in order of increasing liberalism, of secular voters, mainline modernists, Hispanic Christians, black Protestants, and Jews.

Thus, the new religious order clearly embodies the distinctive policy values of religious groups. And although the rank of most religious groups differs slightly from one dimension to the other, there is a fairly strong correlation between the dimensions, producing similar group patterns. If one looks for those defining the religious ends of the partisan continuum, one finds Mormons and evangelical traditionalists on the one side, and Jewish voters on the other, anchoring the conservative and liberal ends of both policy dimensions. We might also note that the distance between Republican and Democratic voters on both issue dimensions has become wider since 1992, paralleling the growing differences between (and within) religious traditions.

The distinctive political values of each religious group are also inculcated, reinforced, and activated by mobilizing forces. And, as we might expect, most groups are subject to at least some blandishments from both party and interest organizations. In the Akron survey we asked citizens whether they had been contacted by a number of religious and secular sources. In the electorate as a whole, we found secular mobilization efforts only slightly more frequent than those deriving from religious sources, a powerful indicator of the extent of religious mobilization. We also found a complex, but intuitively plausible, pattern reflecting the dynamics of the new religious order. In the three largest religious traditions, traditionalists usually reported more religious contacts than secular ones, while modernists reported a surplus of secular contacts over religious. Interestingly, in each religious group Bush voters were more often the targets of religious contacts than Gore voters, whatever the overall balance of religious and secular contacts. The total number of such contacts is still modest, however, suggesting that parties and candidates could significantly improve their vote totals with more intense and more targeted mobilization efforts.

If electoral choices strongly reflect the new religious order, it is still important to consider those not incorporated in that order. Indeed, nonvoters constituted almost half the potential electorate. The final column in the table reports the percentage of each group that did not vote in 2000, a crude but effective measure of disengagement. The results are intriguing. Among the “Republican-leaning” constituencies, traditionalists would seem to be most firmly attached to the new religious order, with smaller proportions not participating: Latter-day Saints (29 percent), traditionalist evangelicals (37 percent), traditionalist Catholics (41 percent), and traditionalist mainliners (43 percent). On the Democratic side, Jews are not only firmly partisan but also fully engaged, with only 18 percent not voting. Mainline Protestant modernists almost match their traditionalist brothers and sisters at 46 percent, but evangelical and Catholic modernists abstain more often (63 and 55 percent, respectively), and secular citizens show a fairly high rate of nonvoting (52 percent). Black Protestant turnout increased in 2000, certainly bolstering the virtually unanimous choice for the Democrats in that group. Only Hispanic Christians have little attachment to the new religious order, with 70 percent not voting.

What explains a citizen’s decision to stay home on election day? Political scientists have traditionally attributed nonvoting to a lack of political interest and knowledge, educational and income disadvantages, absence of organizational ties, and the failure of parties, candidates, and interest groups to reach citizens. And the Akron survey reveals that all were important factors in 2000. Clearly, psychological engagement was the best predictor of voting: those who knew most and cared most about politics voted at the highest rate. In a closely related vein, those who perceived major differences between Bush and Gore were also more likely to participate, while those who saw modest differences (perhaps because they knew little about either) were least likely to go to the polls. Educational attainment certainly influenced knowledge about and interest in politics, but encouraged turnout in other ways as well. Income by itself, on the other hand, was a much less important stimulus to voting.

These results are hardly unexpected, fitting well the models developed by academics for explaining turnout. More interesting is the way that social networks, organizational involvement, and group mobilization affected voting. Our findings confirm Putnam’s insight: Americans who live rich organizational lives, both religious and secular, were more likely to vote (and much more likely to participate beyond voting). And because of these connections, they were also more likely to find themselves in religious and secular communication networks that further encouraged participation.

Why, then, do some religious groups participate at a high rate, while others do not? Although our analysis is preliminary, we have already discovered that the impact of the factors discussed above varied substantially among religious traditions. Psychological engagement was important everywhere, especially for Catholic and secular voters, but much less so for evangelicals and black Protestants. Education was also an important predictor of turnout in all traditions, especially among Jews, while income is a significant but minor predictor across all religious groups. For evangelicals and Catholics, the married were more likely to vote than single people when everything else is taken into account, but this was not true for other religious groups. Among black Protestants and secular voters, women were slightly more likely to vote than men.

The more significant findings concern the impact of social capital, both religious and secular. Scholars have long observed that church attendance is consistently associated with higher turnout, but they have not been sure why. Do the faithful vote more frequently because they are reminded of elections at church? Or because the clergy provide political direction? Or because information is available in voter guides and other material? We asked many questions about politics in the congregation and discovered that church attendance is often a proxy for several different kinds of activation.

For evangelical and mainline Protestants, church attendance encouraged voting primarily by providing an opportunity for friends to exchange talk about the election. Such activity partially accounts for the strikingly high turnout among evangelical traditionalists, who still have some educational and other political disadvantages that might normally limit their electoral participation. These people were not only in church more often than most other religious people, but they were also especially likely to engage in political discussions there. For Catholics, political talk among friends in church was much less frequent; rather it was strong parish connections that appeared to stimulate political activities. Even among Jews, with their exceptionally high turnout, attending religious services stimulated voting. Among black Protestants church-based mobilization of voters achieved the status of an art form, confirming the legendary political prowess of the African-American church. Pastoral discussion of politics, conversations with friends, presence of voting material, and holding church office all added to the likelihood of voting.

Although religious institutions are one of the most significant sites of voluntary association activity, other kinds of social capital also affected voting. At first glance, that impact might appear limited. In our initial analysis, we found that voluntary activities outside the church were direct predictors of turnout only for mainline Protestants”the quintessential organizational generalists. As it turns out, though, these associations were equally vital for evangelicals, Catholics, black Protestants, and secular voters, but in an indirect way. For members of these groups, social and organizational involvement produced greater psychological engagement in politics, which in turn encouraged voting. Indeed, for secular citizens, largely detached from religious institutions, such social engagement was especially potent in connecting the citizen to politics. Similarly, direct contacts from political organizations”candidates, party organizations, interest groups”were also vital in encouraging turnout among secular voters, and to a lesser extent for mainline Protestants, Catholics, and even evangelicals.

How would greater turnout have affected the election results in 2000? The conventional wisdom is that greater turnout helps Democratic candidates, but this chestnut is not necessarily true. Although nonvoters tend to be centrist on issues, perceive fewer differences between the candidates, and care less about the outcome than voters do, these generalizations apply within the context of particular religious groups. For example, evangelical nonvoters fall between their brethren who voted for Bush or Gore on such measures, but they were closer to Bush on moral issues, and closer to Gore on social welfare questions. Thus, the impact of additional turnout is often unclear, in this or other religious traditions. Whether increased voting would benefit the GOP or the Democrats might well depend on which issues were salient and which groups increased their turnout. With such large contingents of nonparticipants in almost every religious tradition, mobilization differentials are politically crucial in the closely divided new religious order.

The election of 2000 left half the electorate pleased, and the other half dissatisfied. The former welcomed a Republican President who would restore integrity to the White House, advance conservative moral causes, and resist growth of the federal government. The latter wanted a Democrat who would support abortion rights, protect the environment and gay rights, and use the federal government’s power to promote national health care and education. Beyond the ranks of voters, half the population failed to participate, completing the “50 percent solution.”

Religion played a key role in determining both the partisan polarization and the disengagement that characterized the public in 2000. Proponents of the culture wars argument can certainly find some confirmation in the results. Religious traditionalists, including their Latter-day Saint allies, provided the core constituency for George Bush, while Al Gore depended heavily on religious modernists and secular voters. Religious traditions still matter, however, as Bush drew upon evangelical and mainline Protestants of all stripes and Gore gathered strength from black Protestants, Jews, and other religious minorities. But even in the religious groups most torn by the culture wars and in every religious tradition, many citizens were still disengaged, especially those outside the web of institutional ties.

As political scientist Walter Dean Burnham has suggested, such complex party coalitions are best described by geological metaphors. Today the ancient ethnoreligious bedrock of vote choice has been eroded by rising tides of disengagement, while simultaneously being fractured by the upheavals of cultural politics. Indeed, the religious formations we saw in 2000 have been developing for some time and have now solidified. This fact has vital ramifications for governance. In the future Republicans will remain solicitous of traditionalists, and evangelical traditionalists in particular, while Democrats will privilege the concerns of religious minorities, secularists, and modernists. Regardless of well-meaning admonitions to both parties to “move to the center,” ignoring such large core constituencies would be political suicide.

Of course, a recurring “50 percent solution” satisfies no one. Both parties will seek electoral dominance by mobilizing the disengaged in their core constituencies, by attracting more swing voters, or by making inroads into previously hostile groups. Republicans will find that bolstering their core vote will be easiest among evangelical traditionalists and centrists because they constitute the bulk of the large evangelical population, are quite pro-life, and are accessible through church-based organizing techniques. Reaching more mainline and Catholic traditionalists is not as easy, given their dispersion in politically pluralistic religious institutions, but it is equally necessary. In the same way, Democrats face not only the relatively simple organizational task of mobilizing black Protestants, but also the more difficult one of reaching other religious minorities. Activating the growing coterie of secular voters is even more problematic, as many are cut off not only from religious institutions, but from other organizations as well.

Building winning coalitions requires not only the care and activation of core constituencies, but also the cultivation of new supporters. Both parties have some clear options here. Republicans must try to expand their coalition among those on the conservative side of the center, especially Catholics, but the more promising route is to cut into the Democratic stranglehold among religious minorities. The centrality of the church, the conservative theologies, and the traditionalist moral leanings of many black and Hispanic religious communities provide a place to start. If the GOP can produce policies that address the daily concerns of minorities, economic as well as cultural, and find a role for community religious institutions to play in implementation, the party’s new religious tone might be an attraction, rather than a liability, to religious voters in other traditions. For the Republicans, the question is, can “compassionate conservatism” go beyond pious rhetoric to become a governing principle?

For Democrats, the best route is to expand their beachhead among modernist evangelicals and centrist mainline Protestants, and to retake the Catholic center. They can make such inroads by emphasizing moderation, rather than all-out liberalism, on moral issues, and by defending a vital, but restrained, role for the federal government in education, health care, and the environment. For the Democrats, the question is, can Bill Clinton’s “New Covenant” be reintroduced and transformed from a campaign slogan into a vision for moral governance?

The success of either party in transforming its characteristic “religious” appeal into a public theology may depend upon effective exploitation of social capital. Throughout this essay we have discussed the role that social capital plays”or no longer playsbringing Americans into the public square. We have found that those religious groups that maintain a strong institutional life, inside or outside the church, have become dominant actors in their respective parties. And yet there is another facet of this issue. Many types of organizational connection may enhance political mobilization: the tight religious cohesion of evangelical traditionalists or black Protestants serves just as well as the institutional activism of some secular liberals. But both these forms of involvement represent Putnam’s “bonding” social capital, in which social involvement takes place primarily within the group and does not extend to the larger community. Such social capital may provide electoral advantages, but it does not provide sturdy religious underpinnings for either “compassionate conservatism” or a “new covenant.”

What is needed is “bridging” social capital: activity that reaches beyond the religious group itself to work with others on causes that involve “loving thy neighbor,” but are not purely sectarian in nature. Traditionalist Protestants and Roman Catholics are remarkably generous in donating their time and energy to worthy causes: we find (as Putnam did) that they are much more engaged on the whole than religious liberals or secular people. But they are also more likely to volunteer in ways that bond them with one another, serving the needs of people within the community of faith, rather than connect to the needs of others beyond the fold. In this respect, the declining number of mainline Protestants is particularly disturbing, for this group has led the way historically in both “inside” and “outside” volunteering, as well as other efforts at civic revitalization. We do not mean to deprecate the value of “brightening the corner where you are,” but people of faith need to extend their light outside the perimeter of their own communities. Only then can citizens truly express compassionate conservatism or join in a new covenant, reinvigorating America’s civic culture. Such renewal, if it is to occur, must take place within the new religious order, for these matters are fundamental to our political soul.

James L. Guth is the William R. Kenan, Jr. Professor of Political Science at Furman University; Lyman A. Kellstedt is Professor of Political Science at Wheaton College; John C. Green is Director of the Ray C. Bliss Institute of Applied Politics at the University of Akron; and Corwin E. Smidt is Director of the Paul B. Henry Institute for the Study of Christianity and Politics at Calvin College. They write frequently on religion and politics for both scholarly and popular publications.