Two Electorates, Different Trends: A Reply To Jay Cost
I recently did a post on the proposition of a rightward trend in American public opinion, based in part on some new analysis by TDS Co-Editor Ruy Teixeira of the apparent upsurge in self-identifed conservatives in the midterms. This piece drew a spirited response from The Weekly Standard's Jay Cost which made some good points but unfortunately missed the main context of what we were saying and took issue with arguments we did not make.
To begin with the "arguments we did not make," Jay appears to think my denial of any "natural Republican majority" in the electorate means I think Republicans won't win majorities in the 2012 congressional elections. None of us obviously knows what will happen, but actually, I think the GOP is in a good position in congressional elections going into 2012 because of its big redistricting advantage and the surplus of Democratic Senate seats up next cycle, which means they could hang onto control of the House and make Senate gains even if they lose the presidential election and lose a majority of the national House popular vote. But those advantages may not persist over time; future Senate cycles are less skewed, and let's remember that in 2002 it was universally assumed that the last round of redistricting had given Republicans a "lock" on the House for decade--a lock that lasted exactly two cycles. Real live events matter, and districts change after they are drawn.
But the main reason for doubting a "natural Repubican majority" is the factor that we have been writing about TDS for months, and wrote about again in this piece: the very different turnout patterns that always prevail in midterm and presidential years.
To be very clear, we have two electorates in this country, which in the past has often only had a marginal effect on partisan outcomes, but now is exceptionally important because the composition of the two party's voters has become, in 2008 and in 2010, highly correlated with groups that display high (older, whiter, wealthier voters) and low (younger voters and minorities, particularly Hispanics) turnout tendencies in midterms. Jay doesn't deal with this fundamental argument about "the electorate," and instead takes issue with Ruy's analysis of under-the-surface changes in public opinion, best reachable by looking at a more stable sample of Americans, registered voters.
Jay implies that Ruy used Pew rather than Gallup data to look at changes in the ideological composition of voters over time because the former supported his position while the latter did not. I asked Ruy about this, and here was his reply:
First, Pew only provided registered voter data, not all adults data, on their ideological time trend, so I couldn't have used all adults data even if I'd wanted to. Second, I used Pew data because they provided time trends on ideology for Republicans, Democrats and the three flavors of independents: Democratic-leaning independents, pure independents and Republican-leaning independents. This was central to my analysis.
But let's look at those Gallup data. Gallup shows a conservative shift of 5 points (2006-2010) among all adults as opposed ot Pew's 3 point shift among RVs over same time period. However, two other public polls that provide time series ideology data for all adults are actually closer to Pew's 3 points than Gallup's 5 points: WaPo/ABC has 3.3 over the time period, CBS/NYT has 3.6 over the period.
Moreover, it's kind of a strange time to be appealing to the authority of Gallup after its way-off, outlier final generic ballot poll showing a 15 point Republican lead. Gallup was also reporting super-high levels of conservatives among LVs (54 percent in a one mid-October poll). Pew on the other hand nailed the result exactly. So no apologies for the data I selected.
But even if you accepted a 5 point swing among the public, a 10 point jump in conservative representation between '06 and '10 still far outstrips that swing. Bottom line: you can't account for the election results, including the high proportion of conservatives among voters, on the basis of a sharp ideological swing toward the right among registered voters or among the general public. Conservative mobilization clearly had something to do with it.
And both phenomena owed a lot to a shift in identification of Republican voters (including Republican-leaning independents) from "moderate" to "conservative," which was the main point of his analysis, and is a conclusion shared today by ProgressiveFix's Lee Drutman in a careful evaluation of the Cost/TDS exchange.
Perhaps a semantic problem clouds this issue for Jay Cost and for other readers: yes, an "electorate" in which some members are moving "to the right" while others stay where they are can be said to be "moving to the right" in an arithmetical sense, but if its main effect is simply to move the party of conservatives a bit farther away from everyone else, that's hardly an unmixed blessing.
I'll deal with one other aspect of Jay's argument: his characterization of our point-of-view as follows:
The electorate has not moved in any significant fashion, and what we saw this November is nothing for liberals to worry about.
I should hope I've dealt with the issues involved in the word "electorate;" yes, the midterm electorate moved, but partly because of the turnout patterns already discussed, and partly because of factors (ahem, the economy) which don't have much to do with ideology or much predictive value for the future. But the planted axiom that Ruy and I don't think Democrats have anything to worry about is simply wrong. Frankly, I think both parties have a lot to worry about economically, and from a structural point of view, have much to worry about if the current "two electorates" pattern persists, which is hardly conducive to stable policymaking.
Yes, Ruy and I think 2012 will likely be a better year for Democrats overall for reasons ranging from the more positive nature of the presidential electorate, to the close House districts Republicans will now have to defend, to the likely improvement in the economy, to the arguably weak GOP presidential field, to the many dilemmas Republicans face as a party that is simultaneously demanding deficit reduction, new high-end and corporate tax cuts, restoration of Medicare "cuts," and (among many GOPers) more defense spending and perhaps a new war with Iran. We also think long-range demographic trends favoring Democrats haven't suddenly gone away.
But if I were a Republican, I'd be less worried about Democratic confidence levels and be a lot more focused on restraining the exceptional triumphalism of my colleagues, many of whom seem to think that every lesson learned in past elections, including the longstanding lack of support of sizable majorities of Americans for key conservative policy positions, can now be forgotten based on one very good midterm cycle. As Democrats just learned, trends can change quickly, and so can "the electorate."