A Dynamic Approach to Understanding the New American Electorate

Red America and Blue America. Political polarization. Debt ceiling standoffs and government shutdowns. Increasingly, it seems that our nation is divided into two camps whose beliefs and goals are irreconcilable. And while the right is engaged in an on-going internal battle over how to move forward, the long-term solution for many on the left has been to assume that these two camps will remain locked in and diametrically opposed for decades. Democrats take solace in the fact that their coalition is growing and believe that it will eventually form a working majority to deliver consistent political victory. This report challenges the conception of political identity which underlies that prediction and instead posits that, rather than functioning as an unchanging tribal affiliation, political identity is dynamic and evolving. This new understanding has deep repercussions for American politics going forward: both promising the possibility of change and warning of the danger of Democratic electoral apathy.

In the fall of 2012, Republican strategists began claiming that the public polling showing President Obama leading the election was skewed. They argued that the polls failed to include enough Republicans, based on numbers from the past, and that if the polls accurately reflected the number of self-identified Republicans, Democrats, and Independents, a more representative result would emerge showing the race tied or with Romney in the lead. Indeed, a now defunct website, unskewedpolls.com, purported to show the real results of surveys once they were re-weighted to reflect the partisan distribution.¹

The “poll truthers” demonstrate a core problem in political analysis: viewing partisanship as a demographic characteristic. When pollsters weight data, they do so to match the demographic breakdown of the population. Data is weighted to match state and national demographic targets from the Census Bureau. But partisanship isn’t—and shouldn’t be—considered a demographic category. Gender, race, age, and education are justifiably the categories often used to match the data. Party identification, by contrast, is attitudinal, can evolve, and is not locked in.

Viewing partisanship as a demographic category rather than an attitude is neither new, nor unique to Republican strategists.* It is often an underlying assumption in analyses of voting behavior and explanations about party coalitions. In this report, we consider existing models for understanding partisanship and voting behavior. Examining recent changes in the electorate—particularly the increase in nonwhite and younger voters—we posit a new approach, one that considers voting behavior to be attitudinal and historically situated rather than a permanent characteristic. We then apply our dynamic framework to Hispanic, Asian, and Millennial voters, historically situating their recent voting behavior in a broader frame of identity construction, and explaining what that evolution could hold for the future of both political parties.

A similar argument about the partisan make-up of public opinion surveys was made by some on the left in 2004.

Understanding Political Identity: Traditional Views

What drives political identity? Why do some identities become politicized and others don’t? Is political behavior stable over time? Does calling yourself a Democrat or a Republican create a real and enduring connection, woven into your political DNA? Or do party labels simply indicate a voter’s strategic cost-benefit analysis of who is on their side economically, or which party will give them the most tangible benefits?

In academic analyses of ethnic identity, three broad approaches exist, each with a corollary in discussions about political identity. The first we can call primordial: the idea that some identities are pre-existing features with which you are born, inextricably linked and woven into our political DNA—like fur, it cannot be changed. This kind of identity is perceived to be real and stable (barring a major exogenous shock)—think tribal affiliations, or Israelis and Palestinians. The second view is instrumental: the vision of identity as a strategic alignment one chooses in order to get something one wants. In this view, identity may be activated if there is sufficient reward for doing so—it is not especially meaningful in the first place—similar to one’s wardrobe.

In this paper, we employ a constructivist approach: the view that identity is historically constituted and reflects an ongoing process of making and remaking group identity within concrete political contexts. A constructivist orientation allows us to avoid the ahistoricism and interest focus of the instrumentalist approach and the determinism of primordialism by accounting for how group identity is formed—the conditions under which it exists, the mechanisms leading to the adoption of an identity, and the processes through which identity is reproduced and evolves. In this understanding, analogous to choosing your clothes for a certain day based on the weather and personal style, identity is malleable, but it is also meaningful, because it reflects real pieces of a person’s affinities, characteristics, and values. It simply recognizes that how a person assembles their overlapping identities in a single moment will differ based on what is happening around them.

The conception of political identity underlying the primordial and instrumental perspectives provides a lense through which we can understand dominant perspectives about political identity. But they are insufficient to explain the modern electorate, and in fact, if true they would have some serious negative consequences. The following section will explain how the primordial and instrumental perspectives have taken hold in political discourse, what their limitations are, and the implications of their analyses. The next section will then offer a constructivist understanding of political identity, illustrating how it better reflects the new American electorate, and explain what that means going forward.

Primordialism: Partisanship as Tribalism

I. Primordial approaches to identity and partisanship

Amongst scholars of politics, a dominant approach is to view political identity as relatively stable, a product of political socialization occurring at a young age. In this vein, political identity can be considered quasi-primordial, where it is relatively unchanging and woven into one’s political DNA. This makes party alignment largely external to politics and governing. Here, the bonds that form between partisans are viewed as real and persistent group attachments—Democrats bond with Democrats and own that label within themselves, and Republicans do the same, identifying with others who share that label first and foremost. Since party affiliation is immoveable, then change in party alignment can be driven only by external changes—in underlying structures or preconditions. Within this broad conceptualization of partisanship, there are several variations, emphasizing forms of connections such as psychological attachment, social characteristics, and ideological alignment.

Among the earliest systematic inquiries into public opinion and voting in the U.S. were a series of studies focused on the effects of the mass media in the 1940s at Columbia University. These surveys concluded that primary social groups, such as the family unit or religious affiliation, impacted individual opinion more than other factors in making partisan voting decisions. After conducting panel research in New York during the 1940s, Paul Lazarsfeld and his colleagues, working at the Bureau for Social Research at Columbia, concluded that:

For many voters political preferences may better be considered analogous to cultural tastes—in music, literature, recreational activities, dress, ethics, speech, social behavior. … Both have their origin in ethnic, sectional, class, and family traditions. Both exhibit stability and resistance to change for individuals but flexibility and adjustment over generations for the society as a whole. Both seem to be matters of sentiment and disposition rather than ‘reasoned preferences.²

In its initial formulation in the 1940s and 1950s, the primary social determinants of partisanship were considered to be socioeconomic status, religious affiliation, and residence.³Specifically, the working class, those in urban areas, and Catholics tended to identify with the Democratic Party, while higher social classes, those in rural areas, and Protestants tended to identify with the Republican Party. Later scholars added to the categories of social group membership that they believed could impact voting behavior, examining the influence of race, gender, and other demographic factors.

A rival to the Columbia Studies, Angus Campbell and his team in Michigan emphasized the role of deeply rooted psychological attachment to parties in their seminal work The American Voter. They argued that party identification functions as its own form of enduring psychological attachment—that is, people feel a bond of kinship with other partisans and view themselves as aligned with similar partisans. They observed that most voters tend not to pay close attention to news and many follow politics in a limited manner. Political values and issue positions are less important then, they posited, since they simply reflect existing party linkages. Thus, “identification with a party raises a perpetual screen through which the individual tends to see what is favorable to his partisan orientation. The stronger the party bond, the more exaggerated the process of selection and perceptual distortion will be.”⁴

A recent effort by Donald Green and his co-authors inPartisan Hearts and Mindslinks these psychological and sociological perspectives, arguing that partisanship functions as other social identities and that people form deep psychological attachments to those parties. They argue that party identification based on social groups answers two questions: “What kinds of social groups come to mind as I think about Democrats, Republicans, and Independents? Which assemblage of groups (if any) best describe me?”⁵

In a break with these traditions rooted in psychological and social identification, more recent analyses argue that an individual’s ideological beliefs drive his or her partisanship and voting behavior. Thus partisanship is rooted in deep political values about the role of government in the economy and morality. But here, as in the analyses above, ideological alignment is conceived of as stable and driving partisan decisions. For example, if you are an ideological conservative, you support Republicans and align with the GOP, because they align with your own ideology. If you are a liberal, you support Democrats and align with that party. This ideological sorting by party results in tribal stability, under this view, where conservative Republicans and liberal Democrats are forever entwined in a zero-sum battle over the proper role of the state in economic, socio-cultural, and foreign policies. Hence the dominant view of Red v. Blue America.

II. Limitations of these approaches

While these approaches illuminate different aspects of partisanship and voting behavior in the U.S., they do have limitations. For example, emphasizing social group membership as the primary determinants of partisanship assumes durability in party coalitions. Black and Jewish voters have consistently aligned themselves with Democrats. Yet they appear to be the exception, not the rule. Even when looking just since 1980, the demographic make-up of the party coalitions have fluctuated rather than remaining stable.

But a potentially bigger problem is the assumption of internal group homogeneity. A social group approach presumes common experiences and thus common political values, opinions, and behaviors among group members. There is little room for individual agency in this structural analysis, and it lacks attention to the ability of people to recast their own identity. Social group analyses also ignores the fact that different aspects of one’s identity can become meaningful at different times or when different values or issues are invoked.

We explore these issues in greater detail in subsequent sections. But to give you an example of the types of differences evident within social groups, consider both Hispanics and younger voters. Differences on issue positions within the Hispanic community are evident on such issues as immigration. Later generation Hispanics—those who have been here for three or more generations—tend to be restrictionist (favoring a more restrictive approach rather than expansive approach) in immigration policy debates. Conversely, Hispanics who have arrived in the U.S. more recently—first or second generation—have a less restrictionist approach.⁶

And among younger voters, racial and ethnic differences have emerged recently in both voting and party identification. For example, in 2008 President Obama won white 18–29 year olds by ten points, but in 2012 he lost white 18–29 year olds by seven points, and he even lost white female 18–29 year olds by one point.⁷Among 18–29 year olds in 2012, 23% identified as a Republican, 36% a Democrat, and 36% an Independent. But if we look just at white 18–29 year olds, the share identifying with Republicans rises to 30%, the number identifying with Democrats falls to 27%, and the share calling themselves Independent increases slightly to 39%.⁸

This is important since even small variations in how a particular group votes can impact who wins an election, as evident in national exit polls. When we examine presidential voting by demographic groups over time, we find considerable variation, suggesting that demographic groups are not reliable partisans, and thus social group membership does not reflexively determine one’s political identity.

For example, the women’s vote has fluctuated both in which party garners a bigger share and in the size of the margin. Democrats have tended to win the female vote, with the exception of the 1980s. But that doesn’t always translate into presidential success. Kerry’s three point margin in 2004 wasn’t enough to put him over the top against former President George W. Bush, who was credited that year with appealing to “Security Moms.” Not only is the women’s vote winnable for Republicans, if they can shave just a few points off of the Democratic margin, they might be able to win at the presidential level.

This fluctuation is visible among other social groups as well. Democratic presidential candidates have won the youth vote since 1992. However, they have lost the presidential election three times while winning younger voters. As with the women’s vote, there has been a significant amount of variation in the percentage of the youth vote garnered by Democrats, and winning younger voters has been no guarantee of electoral success. Many analysts point specifically to the large Millennial Generation as a Democratic bulwark against older, more conservative generations. Millennials have been aging into the electorate, and in both 2008 and 2012, the 18–29 year old vote was comprised exclusively of Millennials. But in just four years, this group’s support for the President dropped eleven points. This does not suggest a permanent partisan alignment, but rather continually shifting ground.

Hispanics have consistently voted for Democrats, although that margin rises and falls election to election. President George W. Bush was successful in 2004 partially because he peeled off enough Hispanic voters from John Kerry. In fact, Bush lost the group by only a nine-point margin, compared to Romney’s forty-four-point margin loss in 2012. Perhaps Republicans are unlikely to win Hispanics outright in the near future—but they may not have to do so in order to achieve success at the presidential level. Narrowing the margin could put them over the top—even in an electorate that is slightly more Hispanic than the one President Bush faced.

One of the biggest surprises for many pundits in 2012 was the Asian vote. Asian voters supported Republicans in the 1990s, but they have become increasing aligned with the Democratic Party since then. As two prominent Asian American political scientists noted, the 1990s were “a period when the Democratic Party developed a new pro-business image, economic growth was strong, Asian Americans naturalized in unprecedented numbers and Clinton made public efforts to woo them, including nominating the first Asian American to the Cabinet.”⁹Asian support tipped towards Democrats in 2000 and 2004, but it wasn’t until 2008 that we began to see large differences in the share of the parties’ votes. By 2012, Asians were supporting Democrats by potentially wider margins than even Hispanic voters (Democrats won them by forty-seven points, compared to forty-four points for Hispanics).

Homogenizing social groups and then attributing any position, value, or perspective that has majority support among the group to its individual members obscures important distinctions and oversimplifies the political identity and behavior of voters in these groups. Asian voters, in particular, have flipped their partisan support by seventy-one points in twenty years. Why then would we assume that they or any group would stay completely static into the future?

Further, the psychological attachment between voters and their party that Campbell and others found may be neither durable nor deep. When we examine generational cohorts—thereby controlling for temporal events impacting individuals during formative periods of political socialization—we find considerable fluctuation over time. For example, according to Pew, the Silent Generation was “once one of the most Democratic generations, today they are the most Republican.”¹⁰In 1992, 54% identified as a Democrat or an Independent leaning towards the Democrat Party; by 2011, only 46% identified similarly. The Baby Boomer Generation, in Pew’s analysis, had “very little allegiance to the GOP during the 1960s and 70s, but were increasingly drawn to the Republican Party starting in the 1980s. Since then, they have tilted to the Democratic Party.”¹¹

In addition, some analysts have asserted that the party in office at the time one comes of age plays a major role in determining lifelong party identification and voting patterns—thus sparking a lifetime of reliable party allegiance. But the data suggests otherwise. Of the nine House elections between 1994 and 2010, the results show a decidedly mixed bag. For example, people who came of age during Democratic President Truman’s term voted more Republican than the national average in six elections and more Democratic in two. Rather than being overwhelmingly Democratic, these voters trended toward Republicans.Similar reverse trends appear for those coming of age during the Kennedy/Johnson years (they had six Republican-leaning House votes between 1994 and 2010 and only three Democratic-leaning House votes) and the Nixon years (seven Democratic-leaning House votes as opposed to two Republican-leaning House votes).¹²These formative periods appear to have resulted in relatively weak partisan bonds.

And increasingly, partisan attachment appears to be even weaker than it was in the past. Now, fewer voters identify strongly with either party and many are prone to vote switching. In 1970, only 18.5% of the Democratic Party coalition was composed of Democratic-leaning Independents (Dem Leaners). By 2008, Dem Leaners composed a full third of the coalition. Republican-leaning Independents (Rep Leaners) composed 25% of the Republican coalition in 1970, which rose to 31.5% in 2008.¹³The shift away from strong partisans and towards Independents who lean towards the parties challenges the view of tribal partisans with deep and enduring partisan attachment.

These changes in the party coalitions also have implications for winning elections, since leaners and weak partisans (people who say they are Democrats or Republicans, but rather than being strong partisans, respond that they have a weak affiliation with the Parties) do not vote the same as reliable partisans. For example, panel data from the 2000, 2002, and 2004 U.S. federal elections indicates that Independents leaning toward one party and weak partisans are prone to party switching. By 2002, 31.4% of Dem Leaners no longer identified with the Democratic Party. And a similar trend is found with Rep Leaners, 27.2% of whom no longer identified with the GOP after just two years.

The trend was less pronounced for self-identified partisans. But even here we see some individuals changing their party identification over time—a trend more pronounced among Democrats than Republicans. By 2002, 12.3% of weak Democrats had bolted the party, and by 2004 that figure rose to 15.8%. Similarly, 8.3% of weak Republicans had left the party by 2002, and by 2004 it was 11%.With so much movement just over a four year span, it seems evident that party identification at the individual level is not as durable as we might think.

Further, congressional voting patterns demonstrate that many leaners and weak partisans were apt to voteagainsttheir espoused partisan leanings, once again a trend more pronounced among Democrats than Republicans. In 2002, 45.9% of Dem Leaners and 33.9% of weak Democrats voted Republican. And by 2004, 38.2% of Dem Leaners and 23.5% of weak Democrats supported a Republican House candidate.¹⁴

Finally, we must consider the depth of attachment between voters and their party. In a 2013 survey, 36% of Republicans felt their party was out of touch with the American people, while 23% of Democrats felt the same about their party.¹⁵Looking at political party affiliation over time, increasingly Americans are identifying as Independents in record numbers. Pew put that figure at 38%, with 32% identifying as a Democrat and 24% as Republican, in 2012.¹⁶Gallup found that 40% of Americans identified as an Independent in both 2011 and 2012; the latter, they noted, was particularly high for a presidential election year.¹⁷ In our own analysis of partisan voter registration data in 8 presidential battleground states between 2008 and 2012, we found that the number of registered Independents soared by 14.4%, while Democratic registration fell 2.5% and Republican registration inched up by 1.3%.¹⁸

As the data above demonstrates, party loyalty is not a given over time. Even within the same generational cohort, party defection and vote switching is common. Further, the trend has been away from strong partisanship and towards weak partisanship and a large percentage of Independents. While a small and devoted group of activist partisans clearly exists on both sides—often referred to as the “base”—the plurality of Americans now eschews party labels. Partisan attachment for the vast majority appears shallow and in flux, rather than deep and lasting.

As we noted above, durable and strong attachments between individuals and political parties may also result from ideological agreement rather than social identification. But here, too, we find limitations with this approach. First, ideological principles associated with liberalism and conservatism are often not well understood nor consistently applied by voters and politicians. Second, people often opt-out of ideological labels, select moderate over liberal and conservative, or adopt issue positions not in accordance with their own self-professed ideological principles. Indeed, in national exit polls, moderates have outnumbered liberals and conservatives, forming a plurality of the electorate, in every presidential election since 1976 (at least). These factors limit the usefulness of relying solely on ideology in analyzing political identity and behavior.

Early work on the role of ideology demonstrated at best a weak attachment between ideological concepts and political positions. Philip Converse laid the foundation in 1964, demonstrating that most people do not have coherent and consistent responses to open-ended questions about political issues and values. Further, he found that people do not necessarily view politics through an ideological lens. While there was a strong correlation among ideological beliefs and the structure of political opinions among elites, Converse found that such linkages were much weaker at the level of the average voter.¹⁹

Subsequent research demonstrated that elites send cues—signaling which political values are at stake in a debate and thus how the public should vote—thereby clarifying the ideological stakes for voters. Some argue this process has come to play an even bigger role in political identity in recent years. According to ideological realignment theory, due to “the growing ideological polarization of the two major parties since the 1980s, Americans have increasingly been choosing a party identification on the basis of their ideological preferences, leading to a gradual realignment of party loyalties along ideological lines.”²⁰Proponents argue that realignment—the mapping of ideology onto party, where the two increasingly overlap and there now are fewer conservative Democrats and fewer liberal Republicans—has both increased polarization and provided easier cues for the voting public to know whom to support. Others argue that this amounts to simply sorting people by ideology into the two parties, where most liberal voters call themselves Democrats and most conservative voters call themselves Republicans—noting there is little evidence for increased ideological polarization in the general population.

Despite the dominance of the liberal and conservative ideological divisions in Washington, the plurality of voters, if not the outright majority, considers themselves moderate. The American National Election Studies has included a question about ideology in their biennial elections surveys since 1972. Respondents rank their ideology on a 7-point scale from extremely liberal to extremely conservative, with “moderate or middle of the road” as the midpoint. They are also offered the option of “haven’t you thought much about this.” In general, half of all voters gave the responses of “moderate” or “haven’t you thought much about this,” with minor exceptions in 1996 (49%), 2002 (44%), and 2008 (47%). The percentage of voters identifying as a moderate ranged from a low of 20% in 1980 to a high of 28% in 1986 and 1998. The percentage of voters who selected “haven’t you thought much about this” ranged from a low of 22% in 2002 to a high of 36% in 1980 and 1982.

These moderates may be policy centrists or ideologically cross-pressured (think “economic conservative and social liberal”), but they are hardly staunch ideologues defending a uniformly progressive or conservative worldview. They are not genetically liberal or conservative. It thus seems unlikely that ideology drives political identity and behavior for at least the half of the country who labels themselves as a moderate or eschews ideological labels altogether. Further, in analyzing the ideological distribution of opinions on substantive policy issues, such as health insurance and military spending, scholars have often found greater centrism than polarization.²¹

In 1964, Philip Converse also noted that peoples’ belief systems may not always be logical. He expected that since the average voter didn’t “think about” politics and beliefs as much as elites did, public views regarding politics would have less coherence. As Converse put it, “One cannot believe that government expenditures should be increased, that government revenues should be decreased, and that a more favorable balance of the budgets should be achieved at the same time.”²²And yet, in 1964 he noted that many voters, in fact, simultaneously support all 3 of those positions regarding government budgets when asked in a survey.

Along similar lines, a number of polls during the 2012 election demonstrated that Americans wanted the federal government to cut its budget and spend less—just not on any particular program or service. Yet in another survey, Pew asked respondents to choose between decreasing, increasing, or keeping funding levels the same for 18 priorities, ranging from education to Social Security to the military. In not one area did a majority of respondents say we should decrease funding.²³Maybe Converse was correct. The more things change, the more they stay the same.

Instrumentalism: Partisanship as Rational Calculation

I. Instrumental approaches to identity and partisanship

Instrumental approaches to political identity emphasize the rational behavior of individuals. Here, vote choice is driven by people weighing which candidate or party will best provide for them. People will evaluate candidates and parties, selecting the one they perceive will best maximize their utility. Thus, partisanship results from rational choices by individuals.

This can function in a few different ways. Partisanship may be driven by prospective voting—assessing who or which party will benefit an individual or group in the future. Or it could be driven by which candidate or party is closer the individual voter’s policy preferences on specific issues. In this vein, partisanship serves as a proxy or an informational shortcut for voters to adjudicate between candidates.²⁴Vote choice might also be retrospective, reflecting a “running tally” or accumulation of empirical assessments of the parties over time.²⁵For example, during the 2012 election, some Republicans cynically argued that 47% of voters were lost to the Party and aligned with Democrats due to the “gifts” and other benefits Democrats distributed to those communities.

II. Limitations of these approaches

The biggest problem with this perspective is that it assumes that people are able to make rational decisions about politics and how policies will affect their interests, and that they can then make the link to parties and candidates with the policy outcomes they desire. That is, it assumes that voters can link abstract policy debates to kitchen table issues. It also doesn’t fully explore how policies can be framed in different ways to draw support from specific groups—even those whom these policies would not directly benefit.

There is mounting evidence that people do not objectively assess political data presented to them, but rather they evaluate information based on their preconceived notions. For example, assessment of economic data (e.g., whether you think unemployment figures are good or bad) is impacted by political context—who is in the White House and whether you approve of that person. Thus, rather than consumer sentiment explaining economic perception and then driving political choices, the effect of partisanship is evident in both individual and aggregate economic assessments of presidents.²⁶For example, the average partisan difference in approval ratings of the economy during President George W. Bush’s terms was 12.8 points between Democrats and Independents and 21.5 points between Republicans and Independents.²⁷All else being equal, people think the economy is worse when they didn’t vote for the person in charge.

Further, we have evidence that people will accept policies and issue positions they normally would oppose if they are presented by an individual’s preferred candidate or party.²⁸For example, there is evidence that people with higher incomes vote in substantial numbers for Democrats—even when they campaign on raising taxes on the wealthy, as President Obama did in 2012. Then, he won 44% of those making $100,000 or more, and in 2008 he won 52% of those making $200,000 or more.²⁹While President Obama’s tax policies may not have been in their direct interest, nonetheless they identified with him and pulled the lever for him not once, but many twice.

Of course, people can define their policy preferences in different ways. For example, a married gay couple with a high income might support Democrats because Democrats support gay equality, including marriage recognition for that couple. The freedom to marry for gay couples could be viewed as a pocketbook issue due to tangible marital benefits like taxes, Social Security benefits, and range of material interests grounded in legal recognition—in addition to the deep emotional resonance it bears. Still, most people do not follow policymaking closely, and on most issues rival candidates haven’t adopted such oppositional positions, which reduces the ability for voters to clearly adjudicate which ones will most directly benefit them.

Why it Matters: The Stakes in the Debate over Identity

The stakes in the debate over identity and partisanship are daunting. If we are a primordial society, divided by tribal partisan allegiances, then room for compromise and negotiation in our political system is permanently stunted barring a massive shock to the system (think 9-11). Thus, winning can result only by annihilating the other Party. Politics becomes a zero-sum game. The space for bipartisan problem-solving shrinks. And the tiniest of infractions to the Party line become flashpoints for modern-day witch hunts. If, on the other hand, we are an instrumental society, fighting primarily for individual material gains, then we ignore history and deny the existence of anything more than a thin veneer of genuine political beliefs and values. Politics would be nothing more than handing out gifts to coopt different groups.

But there is another perspective, one in which identity is socially constructed—neither fixed and pre-ordained nor conjured by political entrepreneurs for personal gain. Rather than ahistorical instrumentalism or deterministic primordialism, we can view political identity through the process of interest and identity construction, emphasizing the process through which campaigns engage different groups and tap into pieces of their identity to create coalitions.

At this moment in American political history, it is a good time to reconsider our models of political identity and behavior. We need a new approach to take account of the changing nature of the electorate. The population increases among Hispanics, Asians, and Millennials coincides—and is perhaps a cause of—changes in the process of political socialization. For example, the number of Independents has soared to a record high 46% in recent years. As new voters enter the system in a different environment than previous eras, party labels may not carry the same meaning and weight as in the past.

Political socialization today differs from past eras in a few ways. How individuals engage with the political system has shifted over the past decade, particularly as historic forms of voter organization have given way to new patterns, and both individuals and groups have political experiences that diverge from the recent past. The cultural resources and backgrounds of the electorate have also shifted, particularly with the influx of immigrants. This has impacted individual and group values and beliefs, and equipped voters with the unique political, economic, and socio-cultural traditions of communities from which those immigrants came. Finally, the explosion of communications technology and availability of information has altered the once dominant role of a few information gatekeepers, resulting in heterogeneous sources and instant information sharing. The next section offers a new understanding of political identity for this new age.

A Dynamic Framework to Understanding Partisanship

In this section, we offer a new approach which we hope can shed some light on the dynamics of political identity formation, and offer a future that does not sentence our country to fierce tribal political wars ad infinitum. To begin, we lay out our framework for approaching the study of political identity and behavior. Then we sketch the key forces impacting individual identity before examining how campaigns interact with these forces. After sketching out our approach to political identity, we apply this model to Hispanic, Asian, and Millennial voters in the 2012 election.

Dynamic Framework

Political identity is constructed, neither ascribed by social identity nor the result of objective calculations. Identity is rooted in pre-existing social group membership, values and beliefs, and life experiences. Identity is not frozen at a certain moment in life but rather is fluid and can be deconstructed into its constituent parts and reconstructed in different times and contexts throughout a person’s life. Political behavior results from meaningful and affective ties forged through political contestation by candidates and political parties.

In contrast to the orthodox explanations, our framework posits that political identity is constructed rather than determined by social identity or instrumental calculation. Individuals do associate themselves with those with whom they find an affinity. But rather than being stable for all time, we conceive of political identity as being fluid. That is, partisanship may change over time due to shifts in perceptions about the stakes of a debate, real world experiences, or recasting shared interests and commonality.

<>When determining which party to align with, people draw upon existing resources, including patterns of social group formation (e.g., race/ethnicity, class, religion, gender, sexual orientation), an established— although at times fluctuating (e.g., during formative periods of personal identity, or during upheaval or institutional change)—set of beliefs and values, and real world (individual/familial/group) experiences. Pre-conditions thus structure and shape, but do not determine, outcomes. In our view, identity is pragmatic and leaves room for individual agency.

Political identity, as with other forms of identity, answers questions about the self and the other: Who is like me? Who am I similar to? Who is not like me? Who am I different from? It provides a common language and frame of reference for group members. But identity is also multiple, fluid, and overlapping (e.g. a voter is female and Hispanic and wealthy and religious). Which elements of an individual’s identity are invoked in a political struggle depends upon many factors and can be affected by the way a policymaker indicates the values and principles that are at stake in the course of framing a policy issue, experiences with similar issues or policies, or the types of people who a voter believes shares her or his interests.

Policymakers, candidates, and party leaders are also constrained by the past, which creates preconceived views of political parties (party brands). Past performance also impacts the public’s evaluation of the parties. However, this is not a purely “rational” decision-making process, but one that is contested in the public sphere, often resulting in stereotypes about the parties that may not hold up to scrutiny (e.g. that Republicans are the party of balanced budgets, even during the W. Bush era).

In appealing to voters and constructing a coalition, parties can also draw upon common elements of American political culture, including national symbols. Such symbols may include invoking the Founding Fathers or weaving a narrative of American identity rooted in a revisionist reading of Horatio Alger, Jr.’s “pick yourself up by the bootstraps” mentality. In this vein, we view symbols as structures of significance which draw upon historical conceptions to help frame and give meaning to contemporary debates.

What were voters’ pre-existing social group memberships, political values and beliefs, and experiences in the lead up to the 2012 election? What resources could the Obama campaign draw upon in forging a winning coalition from each of the major demographic groups from whom he won support?

In the next section we consider these questions in detail for three of the key demographic groups in the Obama Coalition. Within each group, there are intragroup differences. Rather than treating demographic categories as representing internally homogenous communities—objectifying the subjective—we specifically consider internal divisions in political identity and behavior, how they develop, and their potential consequences. Further, we assume that the variables influencing political identity—group membership, experiences, and political values and beliefs—do not each exist in a vacuum. Indeed, in many situations, experiences influence group membership, for example, or political values and beliefs serve as a lens through which people interpret experiences.

But first, we define the scope of our variables and lay out a series of questions we used to test our framework.

Social Group Membership:Consists of real or perceived (on the part of the group and/or by significant outsiders) commonalities between members based on shared characteristics. Social group membership is a form of cultural identity, a type of individual and group identification, which carries a sense of belonging, history, and affectively meaningful, orienting connection to other people (the living, the dead, and the not yet born), to particular lifestyles and values, to locales/ecosystems, and to a sense of destiny (historical and future).³⁰When analyzing the key demographic groups in the 2012 electorate, we asked:

• What elements are unique to this group?
• What intragroup divisions exist?
• Why was this group constituted? What caused these people to be “linked” together?

Political Values and Beliefs: Encompasses one’s worldview or perspective on how the world is organized and operates, and how itshouldbe organized and operate. It also includes principles that form the basis of how decisions are made about new or existing policy issues and goals to pursue through political processes. When analyzing the key demographic groups in the 2012 electorate, we asked:

• Which policy and ideological attitudes are held in common by a social group that are distinct?
• What are their policy priorities and what does that signal about their values?
• What principles or values do they hold most dear in the political sphere?
• How do they view the society/political system and what would they keep, eliminate, or fix?

Experiences: Includes encounters and events, or combinations of events, that can be common (e.g., September 11th) or unique to a specific group (e.g., Stop and Frisk, sexual harassment). It may include family structure, socio-economic status, political institutions, cultural norms, and technological changes. When analyzing the key demographic groups in the 2012 electorate, we asked:

• What experiences are unique to this group?
• What historical, cultural resources are drawn upon by the group to frame, understand, discuss, and explain their perspectives?
• What are the dominant political, economic, and socio-cultural institutions for different groups?

Political Contestation and Campaign Effects

Politics is fundamentally about power—who has it, how it is exercised, and to what end. Elections are competitions for who will control the levers of power. They are visible and measurable examples of the outcome of political contestation, or the push and pull of political debate. After examining specific elements of three key demographic groups in the Obama Coalition, we assessed the “push” and “pull” effects of the 2012 presidential campaigns. That is, how did the Obama campaign successfully pull key demographic groups (e.g., Millennials) into their coalition? Conversely, how did the Romney campaign push key demographic groups (e.g., Hispanics) away from the party?

After examining the cultural identity groups that were central to the 2012 Obama victory, we consider what influenced their choice. While group membership, experiences, and peoples’ political values and beliefs are important in structuring individuals’ partisanship and voting decisions, elections are fundamentally battlegrounds over immediate choices. This is political contestation, a process by which each side vies for support by demonstrating shared affinities on important political values and issues of the day. And without it, political identity cannot translate into political power.

We know the outcome of the 2012 election. Thus we can analyze the specific ways in which the campaigns pushed away or pulled in key groups of voters. In effect, modern campaigns are about identity construction—weaving a thread of commonality through disparate communities (often linking history, the present, and a shared vision of the future with their political values and issue positions in ways that resonate within particular communities). In every election, one side is more successful than the other at achieving this goal. Understanding why one side prevails can illuminate what happens when political strategy encounters society. Specifically, in order to lay out a path forward we must know the answers to the following questions:

• How did the Obama campaign pull different groups in—or successfully link their policies and message with the values, attitudes, and beliefs of different groups?
• What meaningful, within-group resources did the Obama campaign invoke when appealing to different groups?
• How did the Romney campaign push different groups away—or fail to link their policies and message with the values, attitudes, and beliefs of different groups?
• Was the Romney campaign’s failure to successfully appeal to different groups a result of the candidate, issues positions, or Republican branding?

By analyzing the political identity of the key groups in the Obama Coalition and better understanding how it was assembled in 2012, we hope to cast light on what it portends for the future for both political parties and why Democrats should not take these groups for granted as permanent party-line voters.

A Dynamic Approach to the New Electorate

In this section we apply our dynamic model to Hispanic, Asian, and Millennial voters in the context of the 2012 presidential election. In an earlier report, we argued that, contrary to conventional wisdom, Hispanic, Asian, and Millennial voters appear neither overwhelmingly liberal nor Democratic in survey data, and should not be viewed as reliable Democratic voters.³¹Here, we examine three groups in the Obama Coalition, thier similarities and differences, and elements that are unique to thier members, broadly defined. We consider their political values, attitudes, and beliefs about key issues and the experiences they have encountered, be they shared by group members or distinct from other communities. A central element is individuals’ own sense of place within the political system: where do voters place themselves in standard left-right metrics of American politics? We then explore how President Obama and Democrats wooed these voters—the issues and themes they drew upon—and where Romney and the Republicans failed to understand and integrate key values of these communities in the political struggle. Finally, we consider the implication of 2012 for the future.

Hispanic Voters and the Dynamic Model

Social Group Membership among Hispanics

People defined as “Hispanic” or “Latino” in the U.S. are considered part of a common ethno-linguistic group. That is, membership is predicated on the group’s real or perceived shared characteristics of ethnicity (sense of common descent or relatedness) and language (assumed to be Spanish-speaking). However, the Hispanic label is externally ascribed (by non-group members), and people categorized as Hispanic are more likely to label themselves based on their family’s country of origin (51%) or call themselves simply Americans (21%).³² In recent surveys, only 24% primarily use the label “Hispanic” or “Latino.”³³ And strikingly, 69% of Hispanics believe that they do not share a common culture with one another, compared to only 29% who believe they do.³⁴

These patterns of self-identification differ based on length of residency in the U.S. While 62% of first generation and 43% of second generation Hispanics primarily identify with their family’s country of origin, only 28% of third generation and later agree.³⁵Nearly half (48%) of third generation and later Hispanics describe themselves simply as American, with 21% claiming Hispanic or Latino.³⁶Differences between first and second generation Hispanics and those third generation and later are evident not only in labels but also in policy positions, as reviewed below, and they portend interesting consequences for future political coalitions relying on increasing numbers of voters who others would call “Hispanic.”

The Hispanic story in America should not be read as necessarily parallel to earlier waves of immigration and assimilation. Compared to other ethnic groups coming to the U.S. (e.g., Irish), Hispanic arrival and cultural assimilation has important differences. Notably, Hispanic immigration is continuing, rather than being confined to a particular time period. As a result, there are constant new within-group arrivals, invoking the process of acceptance and inclusion, and keeping those issues fresh in the minds of even more established communities. And Hispanics tend to maintain ongoing cultural, political, and economic ties to their country of origin—through family, friends, and colleagues. Finally, Hispanics have maintained their ethnic identity in many ways, such as speaking Spanish at home. Thus, Hispanic assimilation is likely different today, and in the future, than it has been for other groups historically.³⁷

Additionally, there are intragroup differences based on, among other things, ancestral country. For example, Cuban Americans, centered in Miami, have tended to be more conservative and more closely aligned with the Republican Party. They also have the lowest poverty rates of any Hispanic group and are more likely to own their own home.³⁸ However, subsequent generations of Cuban Americans have gradually moderated their positions. This is evident in the fact that Republican dominance is eroding within the community. For example, in 2012, President Obama received 48% of the Cuban vote in Florida. By comparison, Al Gore won only 25% in 2000.³⁹

Other sending countries and waves of Hispanic immigrants have had different historical experiences—and thus different processes of socialization. Time in the U.S. and method of arrival have impacted their legal and socioeconomic status. Salvadorans, Dominicans, and Mexicans tended to come to the U.S. at different time periods and for different reasons. Salvadorans fleeing their war-torn country in the 1980s were given legal status. In turn, they could seek different types of work—typically higher paying—than an undocumented immigrant.⁴⁰Puerto Rico is an American protectorate; thus, Puerto Ricans are counted as American citizens. These differences may undermine any sense of a common, pan-ethnic Hispanic identity. Indeed, the term Hispanic was widely put into use only after the 1970s when the U.S. Census Bureau adopted the category, and, as noted above, it is typical ascribed by those outside the group, not adopted by those within.⁴¹

There have also been regional variations in patterns of settlement, which impact how communities are received and integrated not only into local economies, school systems, and neighborhoods, but also into the political system. Existing networks and institutions in a neighborhood, for example, can impact both cultural identity and participation in politics.⁴²Despite these internal variations, most American political commentators and political actors on the national level continue to homogenize the Hispanic experience and treat individuals perceived to be Hispanic as a monolithic group.

Experiences: The Influence of Immigration Reform

The battle over immigration reform in 2006—a response to the passage ofThe Border Protection, Antiterrorism, and Illegal Immigration Control Actof 2005 (H.R. 4437)—may have had the unintended effect of unifying the diverse Hispanic population in the political arena, at least temporarily. In the wake of the House bill, which increased penalties on both undocumented immigrants and anyone assisting them, unprecedented organization within the Hispanic community began to take root. As researchers have documented, a political threat was turned into an opportunity “that activated multiple Latino constituencies, including the Latino citizenry and organizational elite, to come together insolidaridad, or group solidarity, for immigrant rights.”⁴³Through the Catholic Church, Spanish-language media, and local schools, the community organized approximately 160 rallies and marches throughout the country. Evidence suggests that between 3.5 and 5 million people participated in these events.⁴⁴

Prior to this unprecedented confluence of political organizing, attitudes on immigration were divided within the Hispanic community. Polls taken in 2004 indicated a gulf between first and second generation Hispanics and third generation Hispanics. Specifically, third generation Hispanics were less likely to want toincreaselegal immigration from Latin America and more likely to say illegal immigration hurts the economy than more recent immigrants.⁴⁵This is not surprising. Historic patterns suggest that as immigrants structurally integrate into American society and assimilate to the dominant cultural values, they will support more restrictionist immigration policy.

But by 2006, as the table above demonstrates, the generational dichotomy on immigration was beginning to collapse. Further, there was broad agreement that the immigration marches signaled the start of a long-term social movement (64.1% for first and second generation and 64.2% for third generation) and a shared belief that the movement would result in higher voter turnout come November (77% for first and second generation and 75.1% for third generation).⁴⁶And when looking at trends in party identification, 2006 does appear to be a turning point. Hispanics have tended to identify as Democrats more than Republicans or lean in that direction, and after 2006 that gulf widened. In Pew’s analysis of Hispanic party identification, which includes leaners—Independents who say they lean towards one or the other party—the Democratic margin rose from a low of D+22 in 2006 to D+34 in 2007, D+39 in 2008, and D+48 in 2012.⁴⁷

This data suggests that the immigration debate had a constitutive effect by bridging the gulfs within the Hispanic community. In many ways, it may have turned what was once merely a label ascribed by outsiders onto people who claimed no common group identification into a self-described and unifying political identity. As positions on immigration issues aligned within the community, the marches themselves functioned as collective vehicles for building a social movement around immigration. Hispanics felt this coalition would be durable—the opening act, not the climax—and would translate into political behavior (voting). As the party identification data indicates, the 2006 immigration marches and legislative battles altered what had been a relatively stable trend of partisan identification by Hispanics. Taken together, the construction of cultural commonality, the perception of unfairness by political actors (Republicans in the wake of H.R. 4437), and the belief in their own political efficacy demonstrates how political mobilization was intertwined with and created ethnic solidarity or group consciousness.⁴⁸*

“Group consciousness is defined as instances when a group maintains a sense of affinity and group identification with other members of the group, which leads to a collective orientation to become more politically active” Sanchez, p. 428. “Ethnic solidarity is rooted in social conditions that trigger heightened ethnic awareness across other identities…Solidarity is not predictive of lifelong behavior” Barreto, Manzano, Ramírez, and Rim, p. 738. Determining which of these best captures the Hispanic experience and resulting affective orientations will impact the stability and durability of their partisan identity in the future.

Political Values and Beliefs among Hispanics

Interestingly, while the issue of immigration played a central role in the construction of a recent national Hispanic political movement, it has not been the dominant priority of this community.* Since 2008 (at least), Hispanic voters have cited jobs and the economy, education, and healthcare as their top three most important priorities. In 2012, 55% said education was extremely important, followed by 54% for jobs and the economy and 50% for healthcare. Immigration came in fifth, at 34%.⁴⁹And unlike non-Hispanic white voters, Hispanics have a favorable view of thePatient Protection and Affordable Care Act, 48% to 19% (32% say they don’t know). By contrast, whites view the law unfavorably, 48% to 30%.⁵⁰

It is possible that Hispanics consider immigration to be an issue that is unique to them and believe that if they do not hold that issue deeply, no one else will. Thus, it may not be their most important issue, but their passion on immigration may be strong.

To be sure, there remain differences in political values and beliefs within the Hispanic community. For example, religion, nativity, and time in the U.S. are interrelated and shape political positions and attitudes on issues. Overwhelmingly, Hispanics are Catholic (62%), with only 19% identifying as Protestant and 13% as Evangelical. Among the U.S. public generally, Catholics represent 23% of the population, Protestants 50%, and (white) Evangelicals 18%. But we see greater identification with Protestantism among U.S.-born and third generation or later Hispanics. While 69% of foreign-born U.S. Hispanics identify as Catholic, only 51% of native- (U.S.) born and 40% of third generation or later call themselves Catholic. We see a reverse trend with Protestants, where only 16% of foreign-born and 22% of native-born Hispanics identify as Protestant, but 30% of third generation and later do.⁵¹

These differences have implications on a range of political issues. For example, a majority of Hispanics believe abortion should be illegal in all or most cases (51% to 43% for Hispanics, as opposed to the reverse for the general population, with 54% saying it should be legal in all or most and 41% illegal in all or most cases), with agreement among Catholics (52% to 42%), Protestants (65% to 32%), and Evangelicals (70% to 28%). But opposition to abortion fades as length in the U.S. increases. While first generation Hispanics agree that abortion should be illegal in all or most cases (58% to 35%), majorities of second (56%) and third generation and later (54%) Hispanics believe that abortion should be;legalin all or most cases.⁵²While time in the U.S. appears to moderate abortion opinions, the rise of Protestantism among later generation Hispanics may also act as a crosscurrent.

On the size and role of the government, Hispanics overwhelmingly support a bigger government providing more services (75%) compared to a smaller government providing fewer services (19%), a value usually associated with Democrats. The U.S. population generally is more evenly divided, with 41% supporting a bigger government with more services and 48% a smaller government with fewer services in recent surveys. But once again, we see differences within the Hispanic community based on time in the U.S. While 81% of recent immigrants and 72% of second generation Hispanics support a bigger government with more services, only 58% of third generation and later agree.⁵³

Finally, Hispanics in the U.S. display more optimism about their own lives and the economic system than the general public. For example, in their 2011 National Survey of Latinos, Pew found that 75% of Hispanics agreed with the statement “most people can get ahead with hard work” and only 21% ascribed to the view that “hard work and determination are no guarantee of success,” tracking more closely with Republican attitudes than Democratic ones on this question. For the general public, those figures were 58% and 40%, respectively.⁵⁴And there are virtually no differences among Hispanics based on birth place or time in the U.S. in adherence to the belief.

For some, this optimism is likely a reflection of comparisons (implicitly or explicitly) to their native homeland, especially for more recent arrivals. For example, on a range of issues, Hispanics rate the U.S. more favorably than their ancestors’ country of origin. Consider some things Hispanics say are better in the U.S.:

• The opportunity to get ahead (87% U.S. to 2% ancestors’ country);
• The conditions for raising children (72% to 10%);
• Treatment of the poor (69% to 7%); and,
• The moral values of society (44% to 21% with 32% answering same).⁵⁵

When we look just at recent immigrants (not second and third generation), their rationale for coming to the U.S. reflects these views. Economic opportunity ranks first, with 55% of Hispanic immigrants saying that is their primary reason for coming to the U.S. The second most cited is family reasons (24%), followed by educational opportunities (9%).⁵⁶This paints a picture of eager Hispanic immigrants coming to the U.S. for an opportunity to work hard and achieve, provide a better life for their family, and invest in their kids’ futures.

Political Identity among Hispanics

In terms of political identity and ideological values, Hispanics buck the conventional wisdom. In a 2006 survey, 31% of Hispanic Democrats responded that they didn’t think of themselves in an ideological way and another 7% said they didn’t know. Only 19% of Hispanic Democrats identified as liberal.⁵⁷In the same survey, Hispanics overwhelmingly rejected ideological norms, with nearly two-thirds eschewing both the liberal and conservative labels.⁵⁸

Of course, Hispanic support for the Democratic Party has been overwhelming in the past two presidential elections. But that hasn’t translated into overwhelming Hispanic identification as Democraticpartisans. Gallup analyzed 26,264 survey responses from Hispanics covering January 2nd through December 30th of 2012. Half of Hispanic voters surveyed identified as an Independent and about one-third as a Democrat. It is only when you include leaners—Independents who say they lean toward one party or the other—that Democratic identification rises to 51%.⁵⁹As such, Democrats cannot assume the Hispanic vote is stable, especially given how prone weak Democrats and Democratic-leaning Independents are to party-­ and vote-switching. Further, the high numbers of Hispanics who prefer the label Independent suggests that rather than a permanent and tribal alignment with the Democratic Party, their attachment is shallow rather than deep.

Hispanics have been pulled towards (or wooed by) Democrats and pushed away from Republicans both over time and in 2012 specifically. President Obama and Democrats appealed to Hispanics by making them feel welcome and included in the broader American identity (e.g., references to common background, a nation of immigrants). But Republicans largely neglected to engage the Hispanic community, and Republican “otherizing” of Hispanics over time has made group members less willing to identify with and vote for that party, at least on a national level, in recent elections. This is largely the story of 2012. And while immigration is important, as mentioned above, it is not the only piece in the puzzle.

The Obama campaign began appealing to Hispanic voters a year and a half before the election, engaging thought leaders and appealing to the community in the Spanish-language media on such issues as Race to the Top and theAffordable Care Act.⁶⁰And Democrats made extensive Spanish-language advertising buys—spending $12.4 million on over 15,000 advertisements compared to Romney’s $9.7 million on 8,500 advertisements.⁶¹Conversely, Romney backer and Kansas Secretary of State Kris Kobach helped draft the Arizona law allowing for the arrest of anyone who couldn’t prove their legal status at any time (the “show me your papers” law).⁶²Linking Republicans, Romney, and the unpopular legislation, which treated all Hispanic-looking people without legal documentation on their person as criminals, merely added to aura of Republican hostility toward this group of voters.

Republicans appeared aware of these problems. They attempted to field a diverse line-up of marquee speakers at the Republican National Convention, including New Mexico Governor Susana Martinez, in order to try and change the perception. Yet despite their attempts to appear welcoming, the whiteness of attendees overall reinforced a sense that the GOP is not inclusive of Hispanics. As one woman noted, “When I looked at the GOP crowd, I didn’t see any diversity. Just one type of person. When I look at Democratic Convention, I saw myself in so many faces. This is the party of the future.”⁶³

And it isn’t just optics that matter. The role of policies in the 2012 election and the “push” and “pull” effects are evident in the issues of theDREAM Actand the Obama Administration’s decision to grant temporary status to young people brought here illegally by their parents. Before the Iowa Caucuses, Romney said he would repeal theDREAM Actif it reached his desk. Romney, in effect, viewed these young people, many of whom are Hispanic, as outside of the big tent coalition. The Obama campaign capitalized on this statement, making sure it was the top story in key Spanish media outlets.⁶⁴

By contrast, not only did President Obama adopt Deferred Action for Childhood Arrivals—protecting undocumented young people brought here as children from deportation—but he urged support for theDREAM Actand comprehensive immigration reform with a path to citizenship. In reference to the “DREAMers,” President Obama said, “Now, these are young people who study in our schools, they play in our neighborhoods, they’re friends with our kids, they pledge allegiance to our flag. They are Americans in their heart, in their minds, in every single way but one: on paper.”⁶⁵Compared to “self-deportation,” President Obama and Democrats embraced Hispanics with open-arms, including them in the big tent Democratic Party coalition.

Implications for the Future

Hispanics overwhelmingly supported President Obama in 2008 and 2012, for all of the reasons laid out above. But the dynamic model demonstrates that they may not remain stable and durable Democratic voters. Hispanics largely reject the traditional left-right dichotomy, preferring to identify as Independents and eschewing the orthodox liberal-conservative dichotomies. Hispanics are not a homogenous group, and they hold complex views on government, morality, and foreign policy issues.

Republicans have lost Hispanics in every presidential election for which we have data. However, if the GOP worked to include them and appeal to their interests, they could shave enough votes off of Democratic margins to win the presidency—just as Bush did in 2004. This is the crux of the issue: will immigration leave a searing anti-Republican legacy within the Hispanic population? As Karl Rove, Grover Norquist, Matthew Dowd and other GOP luminaries believe, probably not, if immigration reform is passed in a bipartisan way.

And there is some evidence that Hispanics, especially these third generation and later, prefer Republicans on economic and deficit issues. But as Senator Marco Rubio (R-FL) noted in 2012, “it’s very hard to make the economic argument to people who think you want to deport their grandmother.”⁶⁶In a recent poll, 44% of Hispanics said they would consider voting Republican if Republicans took a lead on passing a comprehensive immigration reform bill. And over half said they had voted for a Republican for office in the past.⁶⁷

At the outset, we considered whether Hispanics were like black voters—a now stable and permanent bloc of Democratic voters. Will immigration leave a legacy in the Hispanic community that permanently drives a wedge between them and the Republican Party? The early evidence suggests no. For example, there are many prominent Hispanic Republican politicians who are seen as rising stars in the Party (e.g., Governor Susanne Martinez, Senator Marco Rubio), but very few prominent black Republican politicians. Further, as Hispanic voters assimilate over time, they are likely to resemble the median voter—as evidenced by the differences between first and second generation versus third and later generation Hispanics. And Republicans are actively recruiting and appealing to Hispanics, whereas with black voters they have generally ignored them or used them as examples of social problems (e.g., Republican Presidential Primary candidate Newt Gingrich and his comments about young black children working as janitors to learn about work ethic). Finally, if comprehensive immigration reform becomes law, there may be a legislative solution to a major wedge, which would eliminate a top Hispanic grievance toward the Republican Party and effectively remove the issue from politics.

Former President George W. Bush is credited with saying, “Family values don’t stop at the Rio Grande ... and a hungry mother is going to try to feed her child.”⁶⁸In 2004, he won 44% of the Hispanic vote—and reelection. Given the dynamic political identity of Hispanic voters, there is no reason to think future Republican nominees cannot match that level of support if they try.

Asian Voters and the Dynamic Model

While most pundits were unsurprised by the margins President Obama ran up among black and Hispanic voters, his performance among Asian voters had many scratching their heads. Asians are believed to be wealthier, more conservative, and less likely to benefit from many of the social welfare programs Democrats support (e.g., affirmative action) than other nonwhite voters. And historically, Asian voters were Republicans (Clinton lost them by double-digits). But in 2012, Romney lost Asian voters by 47 points—20 points worse than McCain in 2008. So why have Asian voters so dramatically realigned with Democrats, and will they stay there permanently?

Several explanations have been put forth. The first is that Asians are geographically concentrated in coastal states near urban centers—liberal bastions and strongholds for Democrats.⁶⁹Since fewer Asians live in rural, Southern areas, some say, they are culturally different than Republicans. But other analysts assume that Asians should naturally be Republicans, due to commonalities in their values, family traditions, and professional proclivities. These commentators argue that Asian voters rejected the GOP because Republicans have become known as “the party of Bible-thumping, anti-gay, anti-abortion creationists.”⁷⁰Some analysts highlight Republican hostility to science and education (e.g., Louisiana Governor Bobby Jindal calling the GOP the “dumb party”)—key values for the highly-educated Asian community.⁷¹Another variant on this cultural theme argues that Asians are less likely to see government as an impediment to hard work and success, due to their different cultural backgrounds, than those bred in the Protestant, American tradition centered on individualism and skepticism for authority.⁷²Yet none of these can fully explain the changefrom a Republican-leaning constituency to a Democratic-leaning one in recent years, or whether Asians will vote Democratic at the same astronomical levels in future elections.

Social Group Membership among Asians

Americans of Asian heritage are a diverse group, with over 17 million people in the U.S. representing 19 different countries. As with Hispanics, the Asian identity is largely ascribed by outsiders. However, the Asian American label encapsulates some shared experiences (see below for immigration and the treatment of “Asians”) and general commonalities, such as boasting higher average levels of education and the highest percentage of foreign-born group members of any major racial or ethnic group in the country.⁷³Yet these common elements should not mask important distinctions within the community.

The largest subgroups, based on ancestors’ homes, are Chinese (excluding Taiwanese) at 22% of the Asian population, Filipinos at 20%, Asian Indians at 18%, Vietnamese and Koreans with 10% a piece, and Japanese at 8%. Other Asian heritages comprise 2% of the Asian population or less.⁷⁴It’s not surprising, then, that only 19% identify as “Asian” or “Asian American,” with the bulk (62%) identifying based on their ancestor’s country of birth, and another 14% simply as American. As with Hispanics, there are variations based on nativity, whereby 69% of those who are foreign-born identify primarily with their birth country and only 43% of native- (U.S.) born Asians identify with their familial country of origin.⁷⁵

Asians tend to be stereotyped as highly educated, and data indicates there is some truth to that assumption. Nearly half (49%) hold a Bachelor’s degree or higher—far outstripping any other racial or ethnic group in our country, as well as the national average (28%).⁷⁶When comparing solely recent immigrants, 61% of recent Asian immigrants ages 25–64 have a high school degree, compared with only 30% of recent non-Asian immigrants.⁷⁷And 65% of recent Asian immigrants 18 and older have some college experience, compared to 58% for white, 38% for black, and 16% for Hispanic recent immigrants.⁷⁸

But educational attainment varies widely by ancestral country. The groups with the highest proportion of Bachelor’s degrees (or a professional degree) are the Taiwanese (73%) and Indians (68%). Conversely, the lowest levels are found in the Hmong (14%), Cambodian (14%), and Laotian (12%) communities—who, along with sizeable numbers of Vietnamese, came to the U.S. primarily as refugees.⁷⁹

Experiences among Asians: Racial Categories & Immigration

Immigration reform is an important issue for the Asian American community, and the history of discriminatory practices looms large. Immigration in the 19th century was driven by American needs for railroad workers. Women were explicitly excluded. In 1882, Chinese immigration was explicitly banned (in subsequent decades, this ban was extended to all Asian countries). As scholars of Asian history in the U.S. have noted, prior to the 1950s when Asian immigrants were legally permitted to naturalize, Asians “occupied a racial hinterland outside the traditional formation of Caucasian-white, Negro-black, and American Indian.”⁸⁰It wasn’t until 1965 with passage of the Immigration and Nationality Act that Asians (along with some other groups) were finally permitted to legally immigrate to the U.S. This law removed the national origin quota system and replaced it with one emphasizing professional skills and family reunification in applications for permanent residency—a change that allowed for a huge influx in Asian immigrants into the U.S.⁸¹

According to the 2007–2009 American Community Survey by the U.S. Census Bureau, about 60% of Asian Americans are foreign-born—far outstripping Hispanics (38%) or any other group.⁸²Of those who are foreign-born, 53% are American citizens. But there are wide variations in status based on home country. While 73% of those born in Vietnam have naturalized, only 47% from India are citizens.⁸³Japanese and Hmong Americans are the only Asians where a majority are native- (U.S.) born.⁸⁴As we demonstrate later, the high proportion of foreign-born Asians has likely resulted in less alignment with the political system. In short, these voters are still up for grabs.

Of the approximately 11.2 million undocumented immigrants in the U.S., nearly 1.3 million (or 11%) are Asians.⁸⁵And approximately 1.8 million (out of the 4.3 million) of the backlog in family visas are among Asians.⁸⁶ These facts on the ground impact the attitudes of Asian Americans in the political sphere. For example, 58% of Asians support a path to citizenship for undocumented immigrants, with only 26% opposed, and 14% having no opinion. This is a reversal from 2008, when only 32% supported a path to citizenship and 46% opposed.⁸⁷And 22% of Asians say that family visa problems are a very serious problem for them or their family, with another 16% saying it is a fairly serious problem. Less than half (46%) responded that family visa backlogs were not a serious problem for them or their family.⁸⁸

Political Values and Beliefs among Asians

As with Hispanics, Asians aren’t single-issue (immigration) voters. Their top concerns in 2012 were jobs and the economy (86% said it was “very important” to their vote), followed by education (81%), and health care (80%).⁸⁹On the deficit, they preferred the President’s approach, with 67% supporting tax increases on the wealthy and only 35% a cuts-only approach.⁹⁰And 51% of Asian Americans have a favorable view of thePatient Protection and Affordable CareAct, compared to only 18% who have an unfavorable view (the general public, by contrast, is much more evenly divided).⁹¹Finally, by 55% to 36%, Asians prefer a bigger government providing more services over a smaller government providing fewer services—a near reversal of views among the general U.S. public.⁹²

These current political values and issue positions of Asians seem to align with Democrats and President Obama. Indeed, Asians flipped from supporting Republicans to Democrats in the 1990s and early 2000s. The shift in attitudes is partially due to Democrats intentionally wooing Asian voters. The booming economy of the 1990s, coupled with the Democratic Party’s adoption of a pro-business identity, has also been credited with drawing Asian support.⁹³

The anti-immigrant fervor associated with Republicans in the 1990s particularly impacted the Asian community, as their blanket banishment from legal immigration loomed large in the group’s collective memory.⁹⁴Republican opposition to the nomination of Bill Lann Lee for Assistant Attorney General for Civil Rights at the Department of Justice—the first Asian cabinet member—also didn’t help their cause among Asians. And similarly, Republican policies in the 2000s, especially on opposition to healthcare reform and ending the Iraq War, pushed Asian voters away from the party.⁹⁵Finally, the emphasis on Christianity and Christian values, which has come to guide social conservatism within the Republican Party over the past few decades, has less appeal to Asians than the general public. Only 42% of Asians identify as Christian and 39% say religion is very important to their life. By contrast, 75% of the U.S. general public is Christian and 58% say religion is very important in their life.⁹⁶The Republican Party likely appears as an exclusive party to many Asians, with Christianity as a key unifying element, rather than an inclusive party welcoming others.

Political Identity among Asians

But even given the values and positions outlined above, we should not assume that Asians are ideologically liberal or permanent members of the Democratic Party. For example, in the 2008 National Asian American Survey, 21% identified as liberal, 15% as conservative, and 33% as moderate, while 31% did not identify with any of these choices.⁹⁷Nearly two-thirds of Asians essentially opted out of the liberal-conservative dichotomy.

While Asians overwhelmingly voted for President Obama in 2012, the exact margin is open to some debate. The official exit polls had President Obama winning 73% of Asians and Romney 26%, giving the President a larger margin among Asians than he had even among Hispanics. However, a post-election, national survey of 6,609 Asian American and Pacific islanders, conducted in 9 languages, found that figure may be closer to 68% to 31%.⁹⁸

In comparing the language (English vs. Asian language) used during the interview, the results demonstrate that English-only surveys produced a higher margin for Democrats (nearly identical to the national exit poll) than surveys that include native Asian languages. This trend plays out in party identification as well. When the survey is in an Asian language, the percentage of people who outright reject partisan labels rises to 55% (Independents plus the “don’t know/don’t think in these terms” category), while the proportion who identify as Democrats falls by 17 points.⁹⁹In total, in this survey, 47% of Asians rejected party labels, with 35% identifying as a Democrat and 18% as a Republican.

Political Contestation in 2012: Asians and the Push & Pull

Democrats have tended to do better in appealing to Asians in recent years, engaging both the community and its leaders. And prominent Asian elected officials are almost all Democratic—with Governors Nikki Haley (R-SC) and Bobby Jindal (R-LA) standing out as exceptions to the pattern. Yet for all of their importance, overwhelmingly Asian voters have not been courted by either party in recent years. In one 2012 survey, 52% of Asians said they had not been contacted by Democrats and 64% said they were not contacted by Republicans within the past 2 years. And even among those who were Independents—often targeted as swing voters—the number saying they had not been contacted was close to 6 in 10 for both parties.¹⁰⁰In a 2012 post-election survey, only 31% of Asians reported being contacted about the election.¹⁰¹

Implications for the Future

In 2008, Asian voters in Virginia supported President Obama. In one precinct where 45% of voters are of Asian heritage, the President won 68% of the total vote. In 2009, Republican gubernatorial candidate Bob McDonnell campaigned heavily in Fairfax County, Virginia, where the Asian community is centered in his state. McDonnell won 52% of the vote in the same precinct President Obama had carried with more than two-thirds of the vote the year before.¹⁰²In California, Republican Governor Schwarzenegger garnered 62% of the Asian vote; in 2010, Democratic Governor Brown won 57% of the Asian vote.¹⁰³Coupled with Asian rejection of standard partisanship and ideology, this suggests that this group is not solidly aligned in either camp but rather will remain a swing bloc of voters into the future.

Millennial Voters and the Dynamic Model

Why did Millennials, among those hit hardest by the recession, vote for President Obama in 2012? One explanation is Millennial optimism—blunting the economic effects of the recession.¹⁰⁴Another is their relative openness and progressive attitudes on many social issues—particularly gay equality—which are aligned with Democrats and President Obama, not Romney and the Republicans.¹⁰⁵Most analysts seem to agree: Republicans were the PC to Democrats’ Apple, the Mad Men to their Modern Family.¹⁰⁶And without big changes, argue those who say early voting patterns are sticky and durable, Republicans will lose an entire generation of voters.

Social Group Membership among Millennials

The Millennial Generation—those born between 1981 and 2000—received its label from the big calendar turn. But that time frame also coincided with dramatic changes in communication technologies, geopolitics, and demographics in the United States. The unique characteristics of Millennials—what sets them apart from the electorate as a whole—are likely based on shared common experiences (e.g., coming of age in a post-Cold War world). However, it is important to unpack the drivers of Millennial identity and behavior—is it simply demographics or something else?

As census data and other surveys have demonstrated, Millennials are the most diverse generation in the U.S., with approximately 40% of their members being nonwhite or Hispanic. As such, one of the biggest questions about the political values and issue positions of Millennial voters is whether they reflect something unique about Millennials, or simply the increase in Hispanic and nonwhite members in this generational cohort. After all, fully one-third of the Hispanic population is concentrated among Millennials, as is approximately one-third of the black population and 26% of the Asian population. The starkest differences are evident when comparing Millennials to the Silent generation, which is 79% non-Hispanic, white (compared to 59% of Millennials).

As is evidenced in the chart below, the Hispanic and nonwhite members of the population are concentrated among the Millennial generation, with Gen X also containing a significant portion of Hispanic and Asian members. However, the Gen X group is nearly as diverse as the Millennial generation, yet they do not display the same views on policy issues and political values. Therefore, it is likely that common experiences among Millennials have shaped their worldview at least equally to or more than the racial and ethnic composition of the cohort.

Raw population numbers in thousands, whereby 307,007 is 307,007,000.

This data suggests that Millennials’ views are not simply correlated with their race and ethnicity. However, in reviewing election data from 2008 and 2012, it is clear that President Obama’s declining support amongst the youth vote—in this case 18–29 year olds—was primarily amongst white young people.* While the President won the white youth vote by ten points in 2008, he lost it by seven points in 2012. Amongst white women ages 18–29, President Obama went from +14 in 2008 to -1 in 2012. Amongst white men ages 18–29, President Obama went from +6 in 2008 to -13 in 2012.¹⁰⁷Again, these numbers highlight the overlapping and dynamic nature of political identity, and the fact that each voter belongs to separate groups simultaneously.

Millennials were born between 1981 and 2000. In 2008, Millennials were ages 8–27. Thus in 2008, nearly all of the 18–29 year old voters were drawn from the Millennial generation. In 2012, Millennials were between the ages of 12 and 32. Thus in 2012, all of the 18–29 year old voters were drawn exclusively from the Millennials generation, but some were also counted in the 30–44 age category.

Historically, there has not been a particularly large age gap in voting patterns in support for Democratic presidential candidates. In 1972, there was a sixteen point gap between 18–29 year olds and voters 65 and over. In the subsequent elections, that gap shrank, averaging less than four points in each of the next nine presidential elections. But in 2008, there was a 21 point gap in support for President Obama between youth and seniors, and that gap persisted at 16 points in 2012.¹⁰⁸This suggests that while there may be generational differences in voting behavior, those differences wax and wane over time rather than remaining stable voting patterns.

Political Identity among Millennials

The change in the white youth vote for President Obama between 2008 and 2012 does not appear to portend a radical shift among the youth. The partisan and ideological distribution of 18–29 year olds was virtually unchanged between 2008 and 2012, according to exit polls. The number of Independents among 18–29 year olds in 2008 and 2012 was identical to the figures for the U.S. electorate as a whole. The biggest difference is that 18–29 year olds were six points more likely to be Democrats and six points less likely to be Republicans, when compared to the electorate as a whole.¹⁰⁹

Like partisan identification, the ideological breakdown of the youth vote is also remarkably stable. The number of moderates among 18–29 year olds is broadly consistent with the electorate as a whole—in 2008, 44% were moderates, and in 2012 that number was 41%. The biggest difference between younger and older voters is that fewer 18–29 year olds identified as conservatives (by eight or nine points compared to the general public) and more as liberals (by eight to ten points compared to the electorate generally). The youth vote is more liberal than the national average. A key question is whether they will remain that way.

Still, the plurality identify as moderates. And in a recent annual survey of first-time, full-time college freshman, the number claiming their ideology as “middle of the road” increased more than four points, going from 43.3% in 2008 to 47.5% in 2012.¹¹⁰Ideologically, the youth vote in recent years has been dominated by moderates, a trend that, based on the values of younger Millennials, seems likely to continue.

Political Values and Beliefs among Millennials

If the values and beliefs of Millennials cannot be read directly off of race or ethnicity, then we need to consider other unique elements of this group. There are some broad pattern of difference between Millennials and older generations. For example, Millennials are more likely to favor a larger government providing more services and are less skeptical overall towards the government (although they have trended downward on that question recently). In this section, we consider facets of Millennials’ experiences which might impact their values and beliefs.

In a 2010 survey, 61% of Millennials said their generation was distinct. In an open-ended follow-up question, 24% of those believing they were distinct cited technology as the primary reason, making it the number one response given.¹¹¹This assumption is logical based on the routines of the Millennial generation. Nine in ten are online.¹¹² Three-quarters of Millennials use social networking sites (e.g., Facebook) compared to 50% of Gen Xers, 30% of Boomers, and 6% of Silents. And 55% of Millennials check those sites daily (or multiple times per day), whereas only 38% of Gen Xers and 37% of Baby Boomers do the same.¹¹³Not surprisingly, Millennials believe technology brings them closer to friends and family (54%), with few (35%) worried that technology results in isolationism.¹¹⁴The integration of technology into Millennials’ lives is not an addendum to be learned—it is fused with their daily rhythms. They sleep with their cell phones nearby (83%), they are wirelessly connected when away from home/work (62%), and a sizeable portion now lives in cell-only households (41%).¹¹⁵

Millennials are also the most educated of any age cohort. When comparing generational cohorts when its members were ages 18–28, census data demonstrates that 54% of Millennials had at least some college education at that age, as compared to 49% of Gen Xers, 36% of Baby Boomers, and 24% of Silents.¹¹⁶Whether or not they have graduated, Millennials have been exposed to post-secondary education at a higher rate earlier in their lives. Also reflecting long-term trends, fewer Millennials live in rural areas (14%) when compared to Baby Boomers (29%) and Silents (36%) of the same age. And more than half of Millennials live in suburbs, with nearly one-third in cities.¹¹⁷

Coupled with increasing levels of education and urbanization, Millennials have views of family life that are distinct from prior generations. Only 21% of Millennials ages 18–28 are married, compared to 29% of Gen Xers, 42% of Boomers, and 54% of Silents when they were the same age.¹¹⁸And Millennials are less prone to cast moral judgment on Americans who have different family arrangements. Only 5% think interracial marriage is negative for society. Few Millennials disapprove of people living together without being married (22%) or of mothers of young children who work outside of the home (23%). By contrast, 26% of Silents and 14% of Boomers disapprove of interracial marriage, 58% of Silents and 44% of Boomers disapprove of living together without being married, and nearly 4 in ten of each older generation think mothers of young children working outside the home is a negative.¹¹⁹Some of these trends reflect lived experiences. While 61% of Millennials say they grew up in a two-parent household, that figure rises to 80% for Boomers and Silents.¹²⁰

Experiences among Millennials

It is difficult to know what the stickiest and most influential experiences will be for the Millennial generation—their “JFK moment,” if you will. But so far, the list of potentials is stunning: September 11th, wars in Iraq and Afghanistan, Hurricane Katrina, the Financial Crisis. In their own words, Millennials have summed up their early years thusly:

Our passage into adulthood has been marked by natural disasters, times of economic hardship, and the longest war fought in U.S. history. And our perspective has been fundamentally shaped by the Internet, which has made us hyper-informed and constantly connected the events around the world.¹²¹

In the face of daunting problems, Millennials’ experiences lead them to a belief that solutions require engagement from a wide range of stakeholders, including but not limited to the government. Indeed, as a result of these formidable events, many see a positive role for government (and also in going beyond the government or the private sector alone) in responding to crises.

On one political issue, it is clear that Millennials want the government to act: equality for gay and lesbian Americans. Many feel this is their generation’s civil rights issue. Seventy percent of Millennials believe gay and lesbian couples should legally be allowed to marry. The next highest level of support is from Gen Xers at 49%. Support among Baby Boomers (38%) and Silents (31%) falls well short of these astronomical numbers.¹²²Further, 74% of Millennials and 62% of Gen Xers believe homosexuality should be accepted by society (as opposed to discouraged). By contrast, Boomers (47% to 46%) and Silents (49% to 42%) are divided, with more believing homosexuality should be discouraged.¹²³

Political Contestation in 2012: Millennials and the Push & Pull

On many political issues, President Obama adopted positions and highlighted policies that aligned with Millennials—supporting marriage for gaycouples, adopting deferred action for DREAMers, ending the Iraq War, raising the age for dependents on parents’ health insurance plans, and lowering interest rates on student loans. By contrast, Romney telling young people to borrow money from their parents to pay for college struck a tone that was starkly out-of-line with most Millennials’ lived realities. And while Millennials believe the deficit is a big problem, telling young people that tax cuts and slashing discretionary spending will solve it didn’t sound like a policy, more like a prayer. Further, the anti-science tilt of the GOP in recent years, especially on climate change, appeared ludicrous to a generation who values education and prioritizes clean energy. Meanwhile, President Obama appealed to younger voters through outlets they watched (e.g., MTV), reached out through mediums they employed (e.g., YouTube, Reddit), and offered concrete policies to improve their lives.¹²⁴In short, one candidate made a play. The other bet that young voters would stay at home.

Implications for the Future

Are Millennials the most liberal generation ever, permanently aligned with Democrats? Not likely. Millennials have not adopted traditional liberal Democratic views on many issues. For example, 86% of Millennials favor private Social Security accounts and 74% favor privatizing Medicare.¹²⁵ When asked about the trade-offs between increasing Social Security taxes or cutting taxes, Millennials are split with 44% prioritizing taxes and 49% benefits. Gen Xers (61%), Boomers (63%), and Silents (62%) all prioritize benefits.¹²⁶On the deficit, Millennials prefer the Republican Party over the Democratic Party by 7 points.

Millennials are not in an enviable position. Their career trajectories are stalled at a time when more revenue is needed to finance the safety net for Baby Boomers. And when penalties for the individual mandate become real, Democrats could face a backlash on health care from younger voters. If government intervention in the economy comes with debt and higher taxes, Millennials could reevaluate their political support.

The experiences of the Millennial generation have directly impacted their political values and sense of identity. They are neither liberal culture warriors nor small government purists. Rather, their lived reality clashes with the dominant ideologies in American politics. As stewards of the land of milk and honey, Millennial Evangelicals are forcing their predecessors to reevaluate positions on climate change. Entering the workforce during the age of austerity, Millennials are forcing their predecessors to reevaluate priorities on government investments and spending on Social Security and Medicare.

Rather than a durable Democratic voting bloc, their political future likely depends on which party adapts the best.

Conclusion

Winning political coalitions are not conjured from thin air. Rather, victors successfully draw upon pre-existing narratives to appeal to heterogeneous social groups, invoking their political values and beliefs, and linking a vision of an ascendant America with the goals and aspirations of voters. Success in 2012 was part of a longer-term process of identity construction and partisan alignment. Both reassembling that coalition (for Democrats) and decomposing it (for Republicans) will require significant work in 2016 and beyond. The dominant partisan frames make it seem easy—stick with the old formula (Democrats) and embrace comprehensive immigration reform (Republicans). But just below the surface, dangers abound for both sides. And rather than a highly polarized tribal nation of liberal Democrats and conservative Republicans, we are faced with a new electorate, one that is not only more diverse and younger, but also one that rejects the typical partisan and ideological struggles associated with the second half of the 20th century. The past is prologue. What will the future bring?

Appendix A

Exit Polls