Trend

Jennifer Steinhauer writes in today's New York Times about a trend, in those states that register voters by party, away from Republican registrations, in favor of Democratic and Independent registrations.

I have no trouble believing that there is such a trend, in light of recent election results and current polls. But she might be exaggerating the trend a bit. Also, her analysis of the effect of demographic changes seems faulty.

She acknowledges that "people sometimes switch registration to vote in a primary, then flip again come Election Day". To the extent that happened this year, the switches were toward the Democratic Party, especially in the later primary states. After McCain had clinched the Republican nomination, the only contest that mattered was that between Obama and Clinton in the Democratic primaries. Therefore, some Republicans re-registered as Democrats, in order to vote in closed Democratic primaries. My guess is that, in many cases, those who changed their registration have not (yet, at least) changed it back. Most of those voters probably intend to switch back, but there is no hurry to do so, because it won't matter again until the next primary in their state.

Steinhauer cites data showing a "trend ... now in its fourth year", but she's vague about the nature of the data. Have there been Democratic gains in each year since 2005, or is she comparing the 2008 numbers to the 2005 numbers, and discerning a trend that way? If it's the latter, then those 2008 presidential primary switches would affect the figures. I'm not sure to what degree that factor would affect the data but, as I noted above, it would be a one-way effect, in favor of the Democrats. That's why I suspect that those numbers exaggerate the change.

I hope I'm not being too defensive about my Republican Party, just defensive enough.

Steinhauer mentions changes in the nature of inner-ring suburbs, and attempts to tie those changes into the question of party identification. Her argument is that more Democratic voters have moved into those areas, thereby reducing Republican dominance of suburbia. That would change the local politics, such as who gets elected mayor of a suburban town (don't knock it; some people got their political start that way). But those new suburbanites had to come from somewhere, so their moves must have reduced Democratic registration somewhere else (except to the extent that they're new voters and/or newly naturalized citizens). So the overall party split at state level would not materially change.