In the modern political era, Democrats have nearly always outnumbered Republicans, a fact that hasn’t prevented Republicans from winning lots of elections. But over the last few years, party identification, which changes slowly, has been shifting in the Democrats’ direction. In November 2006, self-identified Democrats outnumbered Republicans by 6%. The Democrats’ edge has continued to grow since then; in May of this year, it was up to 10.1%.
Since then, however, the Democrats’ advantage has been falling steadily. As tracked by Rasmussen Reports, it declined to 9.5% in June and to 7.6% in July. My guess is that it will decline further when the August numbers come out.
This decline is probably due in part to the public figuring out that the Democrats are trying to keep petroleum prices high. Mostly, though, I think party identification is returning to more normal levels as the election season heats up, voters listen to what the candidates have to say, and old loyalties reassert themselves. What is unknowable at this point is where the numbers will be in November. While the Republicans’ gap has narrowed, it is still, as of July, greater than the margin that swept the Democrats to victory in 2006. My guess is that by November the Democrats’ advantage in party ID will be less than the 6% they enjoyed in 2006, but well above the 1.6% edge they had in 2004.
If the trend continues to move in the Republicans’ direction over the next two months, it will complicate the job of Presidential pollsters. Pollsters generally weight their results according to party affiliation, so the assumption they make about how many Democrats and Republicans there are in the electorate as of the date of the survey is critical. To the extent that party ID continues to swing toward the Republicans in the last weeks before the election, Obama will tend to underperform in the election compared to poll results.
To comment on this post, go here.