i spy ... something purple

Edward L. Glaeser and Bryce A. Ward,"Myths and Realities of American Political Geography," NBER Working Paper 11857, December 2005, http://www.nber.org/papers/w11857.

Even if you haven't considered various takes on What's the Matter with Kansas and Larry M. Bartels's explanation of what isn't the matter with Kansas, and explanations of why Bartels may not be quite right, even if Frank isn't either, and even if you're not waiting for Michael Kazin's new Bryan biography to have another whack at explaining the relation between religion and populism -- even if you haven't, and aren't, you still know that the electoral map in 2000 and 2004 looks a lot like the one from 1896, only with the party labels reversed. Which maybe sticks a little bit in your historian's craw, because it's something maybe we should have a little more, of a little higher quality, to say about.

So maybe chew a little on Glaeser and Ward's new working paper. In it they take shots at many elements of the blue state / red state cliché, including their argument that"While the geographic differences [in culture] within America are large, they are not new and they do not seem to be growing." (p. 16)

Perhaps most interesting are their analyses of why different areas are they way they are. They find that "areas where diverse populations interacted in market settings developed beliefs that reduced ethnic and religious conflict. According to this view, if ethnic groups interact at high densities they either destroy each other or eventually develop ideologies that minimize conflict." (p. 17) Or, in other words, "blueness" is a result of ethnic mixing, which in turn is a result of changes in the late c19: "Places that industrialized earlier and that attracted more immigrants at the start of the century are much more likely to have socially liberal attitudes, much less likely to take prayer seriously, and less likely to vote Republican." (p. 33)

[I confess that one of the reasons I warmed to this part of the analysis is that it corresponds to my own observations of what was going on in the Progressive era: that cities with more immigration were more likely to have certain kinds of social policies.]

They include a further analysis of the role of religion in party identification, which they find reverting to a pre-1932 pattern. (They also find, on Jeffrey Kimball's point below, "that ardent Christianity and belief in the military tend to go together." (p. 14)) They also have some interesting things to say about religious "extremes" in moderate parties, and the function of "the ability to send coded messages (this was called Dog Whistle Politics in the latest British parliamentary race)." (p. 30)

Well worth a read.