Why do we vote the way we do? So far, the authors of Democracy for Realists have shown that neither the individualist framework of the folk theory nor the managerial retrospective theory can explain historical election outcomes. What is it then? Do we vote randomly, or is there a different motivating force we have neglected?
The authors argue we need to look back to the once popular realist political tradition, which was largely discarded in favor of rationalist Enlightenment liberalism. In particular, we need a group theory of politics that accounts for “the powerful tendency of people to form groups, the ensuing construction of ‘us’ and ‘them’, and the powerful role of emotion rather than reason in directing group activity.” (p. 215)
In their view, a model of elections as being motivated by group forces will have far more explanatory power than those previously considered. Some key factors to note:
- We are socialized into groups through family and culture; our sense of identity “serves to distinguish groups to which [we] belong that are not central to [our] self-concept from those that are a more integral part of the personality” (p. 228)
- “Identities … are emotional attachments that transcend thinking” (p. 228), and this extends to membership in political parties. We identify with a party from an emotional level, not as the party that we agree with most on arts funding or foreign aid spending.
- People take their views from the groups to which they belong. As such, we get our ideas on policy and ideology from political parties, not the other way around (as in the folk theory).
- Most people “organize political thinking around social groups and their role in competing political parties. They see political and racial clashes as group conflict” (p. 221) In other words, a political fight about the marginal tax rate is not really about the appropriate level of taxation for a society; it is about Democrats versus Republicans, nothing more.
In this understanding, elections are about activating and mobilizing different parts of our identities. We all have many parts of our identity; we claim membership in many different groups simultaneously. We obtain aspects of our identity from parents, peers, regional culture, national culture, education, religion, employment, and so on.
Occasionally, these aspects will be put into conflict with one another. A classic example is that of Republican Catholics in the 1960 election. Kennedy was a Democrat, but also a Catholic. Achen and Bartels cite research that shows that reported church attendance was decisive among Republican Catholics in if they voted for Kennedy. Those whose religious identity was stronger (more church attendance) ended up voting across party lines for Kennedy; those who had weaker Catholic identity stayed true to their political party. The real kicker here is that reverse was true for Protestant Democrats as well (recall the historical animosity towards Catholics). Note that policy views were unimportant here.
(An interesting aside: nowadays, the divisiveness of Catholicism has dissipated, but a new group has replaced them in strongly activating people’s group identity: evangelical Christians. Research shows that identifying a hypothetical candidate or idea as evangelical decreases Democratic support and increases Republican support, holding all else constant).
Perhaps the best example illustrating the importance of group thinking is the post Civil Rights era realignment of the South. For the hundred years after the Civil War, the South was essentially a one party state, decisively ruled by Democrats. Being a Democrat was a part of Southern white identity. But as the Democratic coalition began to change under FDR and his successors, southern blacks started identifying as Democrats as well, especially after the reforms of the 1960s. Concurrently, whites in the South started shifting to the Republican party; now the South is as reliably Republican in national elections as it once was Democratic.
If party identification were a matter of a rational consideration of issues, we might expect this shift to have occurred rapidly among those whites who objected to integration, while those who supported integration remained in the Democratic fold. What actually happened instead was a very slow decades long shift to Republican identification among whites regardless of their views on integration. Southern whites still identified as Democrats even as they generally supported Republican presidents and policies.
In Achen and Bartels’ view, this makes sense only in light of group theory. The shift of Southern whites away from the Democratic party wasn’t driven primarily by racial animus. If it had been, we would expect differences in views on integration to manifest themselves as differences in party identification. But Southern whites who opposed integration and those who did not shifted towards a Republican identity in lockstep with each other. The motivating factor in this shift was instead white southern identity; once southern blacks started calling themselves Democrats, the party was no longer ‘theirs’, and southern whites needed a new place to call home.
This might sound silly or exaggerated, but we’ve all had similar experiences in our lives on a smaller scale. An example that comes to mind is related to that old hipster saw, “I was into them before they were cool.” For many young people, music is a powerful component of identity. The bands you like are more than just music: they are part of your personality, a way of communicating something about yourself. When the community around a band changes, the meaning of being a fan changes. If an underground band achieves mainstream success, they attract a new fan base; often, this changes the image around the artist, and those were were former listeners find themselves
Often the strength of these ties is irrespective of one’s appreciation of the music itself. I can recall an illustrative example. The internet radio company Pandora uses something called the Music Genome Project for recommendations. Songs are broken down into elements like key, tempo, instrumentation, lyrical themes, etc, and when you like a song, the algorithm will find other songs with the same characteristics. This is different from many other recommendation services which use the behavior of other listeners for matching. In this latter case, if people tend to listen to particular artists for cultural reasons, that gets passed into the algorithm, even if the music itself isn’t that similar.
This is most apparent in a story about a Pandora listener complaining that his station kept giving him Celine Dion songs. In his self conception, he wouldn’t associate himself with Celine Dion songs, so he thought something must be wrong with Pandora; he was shocked to finally realize that yes, he actually did like her music. The engine uses only song characteristics, not social cues like other internet radio; the algorithm doesn’t understand that there are aspects of identity built into listenership for many people. If the listener heard the song without knowing the artist, he may have enjoyed it absent any other connotation. But when that information was revealed, the music took an a whole new meaning, one which did not jibe with his sense of identity.
All of this is to illustrate that if we can have such powerful reactions to music, surely it’s not hard to imagine something playing out in the more consequential world of politics as well. And actually, music is a good example because it, like politics, is something that has an important identity component that is separable from its technical aspects. You can poll people’s views on a policy, and if you add information that either Republicans or Democrats support the policy you’ll get wildly different answers. Our sense of group identity is strong and overrides many other considerations.