THE red state-blue state division has captured the pundits' imaginations, leading to much armchair theorizing about how political constituencies in the United States are evolving.
According to some, the country is splitting into two opposing camps, with political divisions becoming more polarized and more spatially segregated than they have been in the past.
A recent working paper, "Myths and Realities of American Political Geography," by two Harvard University economists, Edward L. Glaeser and Bryce A. Ward, challenges this conventional wisdom. The paper can be downloaded from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=874977.
The economists examined a number of contemporary and historical data sources on cultural, religious, economic and political attitudes and compared these responses across states.
They found that differences in political attitudes across states are nothing new: the Civil War and Roaring Twenties had much larger geographic variation in political views than we do today. Though dispersion in political attitudes has generally declined over the last 60 years, the last four years have brought a small uptick.
Though views have become somewhat less associated with geography in the 20th century, they still show strong differences. The fraction of the voters in a given area who vote Republican correlates well with the fraction who voted Republican in the last election.
Furthermore, America is not becoming more polarized. Of course, Republicans have a more positive view of the Republican Party than the Democratic Party, and vice versa, but attitudes have hardly changed since 1978. It is fair to point out, though, that attitudes seem to have become somewhat more partisan in the last few years.
The most remarkable phenomenon is the rise of religion in politics. Thirty years ago, income was a better prediction of party affiliation than church attendance, but this is no longer true. Religion also played a big role in politics a century ago, so we may well be returning to the historical norm.
Cultural and religious attitudes play a big role in voting behavior. For example, the fraction of the population who agreed with the statement "AIDS is God's punishment for immoral sexual behavior" was highly correlated with whether the state was red or blue, according to surveys conducted by the Pew Research Center. The differences in religious attitudes between Vermont and Mississippi are huge.
These cultural divisions have been around for a long time. In the 1936-37 Gallup poll, residents of New England and the Middle Atlantic states were far more likely than citizens elsewhere to support federally financed health measures aimed at venereal disease, to support a free press and to be willing to vote for Catholic or Jewish candidates.
Consumption patterns seem to be correlated with cultural attitudes. For example, the states with the largest level of wine consumption per capita also tend to have the most liberal political and social attitudes. In vino veritas?
Another peculiar connection is the strong correlation between religiosity and militarism. Respondents to Pew's survey who agree that "prayer is an important part of my daily life" also agree that the "best way to ensure peace is through military strength."
So why are these cultural and political divisions so persistent? The authors offer both some simple correlations and some more elaborate theories. It turns out that the degree of industrialization 85 years ago is an "astonishingly good predictor of Democratic support" among today's voters, as is the fraction of the population that is foreign-born.
But the biggest effect seems to be the correlation between religion and Republicanism. Among white voters who attend religious services at least once a week, 71 percent voted Republican in the last election, according to the Pew survey.
Republicans have traditionally appealed to those with higher incomes. The genius of Republicans, beginning with Ronald Reagan and continuing with Karl Rove, was to bring the religious vote into their party, forming a winning coalition of Main Street businessmen, the very wealthy and evangelical Christians. Strange bedfellows, to be sure, but they win elections.
Mr. Glaeser and Mr. Ward offer some speculation about why religion is such an attractive theme for politicians. According to their theory, direct appeals to voters on issues like abortion are tricky, because strong positions inspire groups on both sides of the issue, who then cancel each other out in votes. The trick is to send "coded messages" to different groups of voters. Strong opponents of abortion, for example, may react positively to certain religious allusions that appear innocuous to mainstream voters.
The Economist magazine characterizes American politics as a contest between the incompetence of Republicans and the incoherence of the Democrats. But there is a reason for the Democrats' incoherence: they are feverishly trying to assemble their own collection of strange bedfellows, and no one quite knows what it is.
Ultimately, both parties face the same challenge: how to keep the support of their cultural and political extremists without giving them so much power that they alienate the middle-of-the-road voters.
In this sort of game, the incumbents tend to have an advantage, unless they are perceived as having messed up so badly that even their most fervent supporters desert them. Hey, maybe the Democrats have a chance after all.